Instrumentation [609751]

Handbook
of
Second Edition
Biomedical
Instrumentation

This page intentionally left blank

Tata McGraw-Hill Publishing Company Limited
NEW DELHI
McGraw-Hill Offices
New Delhi New York St Louis San Francisco Auck land Bogotá Caracas
Kuala Lumpur Lisbon London Madrid Mexico City Milan Montreal
San Juan Sant iago Singapore Sydney Tokyo TorontoHandbook
of
Second Edition
Biomedical
Instrumentation

Information contained in this work has been obtained by Tata McGraw-Hill Publishing Company Limited, from sources believed to
be reliable. However, neither Tata McGraw-Hill nor its authors guarantee the accuracy or completeness of any information publis hed
herein, and neither Tata McGraw-Hill nor its authors shall be responsible for any errors, omissions, or damages arising out of use of
this information. This work is published with the understanding that Tata McGraw-Hill and its authors are supplying information but
are not attempting to render engineering or other professional services. If such services are required, the assistance of an ap propriate
professional should be sought.
Copyright © 2003, 1987, Tata McGraw-Hill Publishing Company Limited. All rights reserved. Except as permitted under the
United States Copyright Act of 1976, no part of this publication may be reproduced or distributed in any form or by any means, or
stored in a database or retrieval system, without the prior written permission of the publisher.
ISBN: 978-0-07-177746-9MHID: 0-07-177746-6The material in this eBook also appears in the print version of this title: ISBN: 978-0-07-144784-3,
MHID: 0-07-144784-9.
All trademarks are trademarks of their respective owners. Rather than put a trademark symbol after every occurrence of a
trademarked name, we use names in an editorial fashion only, and to the bene fi t of the trademark owner, with no intention of
infringement of the trademark. Where such designations appear in this book, they have been printed with initial caps.
McGraw-Hill eBooks are available at special quantity discounts to use as premiums and sales promotions, or for use in corporate
training programs. To contact a representative please e-mail us at [anonimizat] OF USEThis is a copyrighted work and The McGraw-Hill Companies, Inc. (“McGrawHill”) and its licensors reserve all rights in and to th e
work. Use of this work is subject to these terms. Except as permitted under the Copyright Act of 1976 and the right to store an d
retrieve one copy of the work, you may not decompile, disassemble, reverse engineer, reproduce, modify, create derivative works
based upon, transmit, distribute, disseminate, sell, publish or sublicense the work or any part of it without McGraw-Hill’s pri or
consent. You may use the work for your own noncommercial and personal use; any other use of the work is strictly prohibited. Yo ur
right to use the work may be terminated if you fail to comply with these terms.THE WORK IS PROVIDED “AS IS.” McGRAW-HILL AND ITS LICENSORS MAKE NO GUARANTEES OR
WARRANTIES AS TO THE ACCURACY , ADEQUACY OR COMPLETENESS OF OR RESULTS TO BE OBTAINED FROM USING THE WORK, INCLUDING ANY INFORMATION THAT CAN BE ACCESSED THROUGH THE WORK VIA HYPERLINK OR OTHERWISE, AND EXPRESSLY DISCLAIM ANY WARRANTY , EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. McGraw-Hill and its licensors do not warrant or guarantee that the functions contained in the work will meet your requirements or that its operation will be uninterrupted or error free. Neither McGraw-Hill nor its licensors shall be liable to you or anyo ne else
for any inaccuracy, error or omission, regardless of cause, in the work or for any damages resulting therefrom. McGraw-Hill has
no responsibility for the content of any information accessed through the work. Under no circumstances shall McGraw-Hill and/or
its licensors be liable for any indirect, incidental, special, punitive, consequential or similar damages that result from the use of or
inability to use the work, even if any of them has been advised of the possibility of such damages. This limitation of liabilit y shall
apply to any claim or cause whatsoever whether such claim or cause arises in contract, tort or otherwise.

Preface to the Second Edition
I am very happy to present before you the second, revised and enlarged edition of my book
Handbook of Biomedical Instrumentation . Its revision and updation have become necessary not only
because of the technological changes that have taken place in the last decade, but also because ofthe immense popularity of the book among professionals in the field of biomedical instrumentation,as also students and teachers in the academic institutes. I feel honoured to have assisted theteaching community in starting numerous courses on biomedical instrumentation in theengineering colleges and polytechnics across the country, which became easier, in most of the
cases, due to the first edition of the book.
In the second edition, the existing material has been thoroughly revised taking into
consideration the developments in technology and introduction of new and improved methods of
medical diagnosis and treatment. Seven new chapters have been added including topics such asnuclear medical imaging systems covering gamma camera, PET camera and SPECT camera. Thetechnology of lithotripsy has matured and it is not only being used for destruction of kidney stonesand bile stones but also for therapeutic purposes. Description of anaesthesia machine andventilators has been included to complete the coverage of operating room equipment. Clinicallaboratory instrumentation and automated drug delivery systems are other important newchapters. A chapter on X-ray and digital radiography covers the much needed information on thisvital equipment universally used in the medical facilities.
The penetration of microcontrollers and PCs in medical instrumentation has resulted in the
integration of automation and built-in intelligence in medical instruments to a great extent. Thishas resulted in replacement of long-established recording techniques and display systems. Theadvantages of the PC architecture in terms of its high storage capacity of data and large screendisplays have been fully exploited in clinical and research applications of biomedical instruments.
Therefore, wherever it was felt necessary, reference to the use of PCs as an integral part of the
medical instruments has been made in this edition.
In order to understand linkages between the life sciences and engineering techniques, it is
necessary for engineers to have a fair understanding about the anatomy and physiology of the

human body. A brief description of the important physiological systems, namely cardiovascular
system, respiratory system and nervous system is provided in the first chapter. Specialphysiological systems are also described in other chapters, wherever it was felt necessary.
The new edition has been divided into three parts. Part one deals with measuring, recording
and monitoring systems. Part two covers modern imaging systems whereas Part three is devoted
to theraputic equipment.
The references have been thoroughly revised to include new research material from research
journals from the world over. Their inclusion in the appropriate places in the text establishes the
necessary link between the current status of technology vis-à-vis the field of research being persued.
When I wrote the first edition my children were young. They have now grown up, are married
and have children of their own. They have been urging me to update this book. While Iacknowledge their pursuation to this initiative, my heartfelt gratitude goes to my wife Mrs. RameshKhandpur who had to spend considerable time alone, watching TV, while I was working in mystudy. Often, my grand-children—Harsheen and Aashna—who are tiny-tots, would trot into mystudy to cheer me up with their pranks which made my task both pleasant and interesting. Mythanks to all my readers who have been sending in their suggestions which have mostly beenincorporated in this edition.
It is hoped that the book will enjoy the same acceptance among its readers and would prove
helpful to professionals and students working in the field of biomedical instrumentation.
Chandigarh
January 31, 2003
RS K HANDPURvi Preface to the Second Edition

Preface to the First Edition
During the last two decades, there has been a tremendous increase in the use of electronic
equipment in the medical field for clinical and research purposes. However, it is difficult to find abook which describes the physiological basis as well as the engineering principles underlying theworking of a wide variety of medical instruments. The present volume has been written to fill thisgap.
The book has been designed to cater to a wide variety of readers. The users of medical
instruments would find the text useful, as they would be able to appreciate the principle ofoperation, and the basic building blocks of the instruments they work on everyday. An attempthas been made to present the highly technical details of the instruments with descriptive and
lucid explanations of the necessary information. It thus provides a useful reference for medical or
paramedical persons whose knowledge of instrumentation is limited.
The field of biomedical engineering is fast developing and new departments are being
established in universities, technical colleges, medical institutes and hospitals all over the world.In addition to graduate engineers involved in developing biomedical instrumentation techniques,the book will find readership in the increasing number of students taking courses in physiologicalmeasurements in technical colleges.
With the widespread use and requirements of medical electronic instruments, it is essential to
have knowledgeable service and maintenance engineers. Besides having a basic knowledge of theprinciples of operation, it is important for them to know the details of commercial instrumentsfrom different manufacturers. A concise description of typical instruments from leadingmanufacturers is provided wherever deemed necessary for elucidation of the subject matter.
The book has been divided into four parts. The first part deals with recording and monitoring
instruments. This part has 11 chapters.
The first chapter begins with the explanation that the human body is a source of numerous
signals, highly significant for diagnosis and therapy. These signals are picked up from the surfaceof the body or from within. This requires electrodes of different sizes, shapes and types. Also, thereare some parameters like temperature, blood flow, blood pressure, respiratory functions etc., which

are to be routinely monitored. These parameters, which are basically non-electrical in nature, are
converted into corresponding electrical signals by various transducers. Electrodes andtransducers constitute the first building blocks of most of the diagnostic medical instruments and
are, therefore, described in the first part of this book.
After picking up the signals of interest from the body, they are processed and presented in a
form most convenient for interpretation. Display is generally on a picture tube for quick and visual
observation or a record on graph paper. Such records facilitate a detailed study by specialists at alater convenient time. Display and recording systems, and the most commonly used biomedicalrecorders are covered in the subsequent three chapters.
Next is a presentation of the various types of patient monitors. The systems aid the nurses and
the medical personnel to quickly gather information about the vital physiological parameters ofthe patient before, during and after operation, and in the intensive care ward where the patient’scondition is kept under constant surveillance.
Apart from the description of conventional equipment for monitoring heart rate, blood pressure,
respiration rate and temperature, a separate chapter has been included on arrhythmia monitoringinstruments. This class of instruments constantly scan ECG rhythm patterns and issue alarms toevents that may be premonitory or life-threatening. The chapter also includes a description ofambulatory monitoring instruments.
Foetal monitoring instrumentation is another area where considerable progress has been
reported in the last few years. Instruments for foetal heart rate monitoring based on the Dopplershift have become more reliable because of better signal processing circuitry and the use ofmicroprocessors. Intelligence is now incorporated in the cardiotocographs to provide dataprocessing for making correlation studies of the foetal heart rate and labour activity.
Wireless telemetry permits examination of the physiological data of subjects in normal
conditions and in natural surroundings without discomfort or obstruction to the person or animal
under investigation. Telemetric surveillance is the most convenient method for assessing the
condition of the patient during transportation within the hospital for making stress studies beforedischarge from the cardiac wards. The chapter on biomedical telemetry explains the techniquesand instrumentation for monitoring physiological data by telemetry in a variety of situations. Italso includes transmission of biomedical signals over the telephone lines for their study andanalysis at a distant place.
An extensive use of computers and microprocessors is now being made in medical instruments
designed to perform routine clinical measurements, particularly in those situations where datacomputing and processing could be considered as part of the measurement and diagnosticprocedure. The use of microprocessors in various instruments and systems has been explainednot only at various places in the text, but a full chapter gives a comprehensive view of computerand microcomputer applications in the medical field.
With the increasing use of monitoring and therapeutic instruments, the patient has been
included as a part of an electrical circuit and thus exposed to the possibility of providing apathway to the potentially fatal leakage currents. Such a situation particularly arises when hecarries indwelling catheters. A full chapter on patient safety describes various situations requiring
attention to avoid the occurrence of avoidable accidents. Precautions to be taken while designing
electromedical equipment from the point of view of patient safety is also discussed.viii Preface to the First Edition

The next part details the various measurement and analysis techniques in medicine and
comprises seven chapters. The first two chapters concern the measurement of blood flow and
volume.
Blood flow is one of the most important physiological parameters and is also one of the most
difficult to measure. This has given rise to a variety of techniques in an effort to meet therequirements of an ideal flow metering system. Both invasive as well as non-invasive techniqueshave been developed. The ultrasonic Doppler technique has proved to be particularly useful inblood flow measurement. A detailed description of the modern methods of blood flow measurementincluding those making use of the laser Doppler technique has been given in Chapter 12. Aseparate chapter on cardiac output measurement details out the present state of art in thisimportant area.
Pulmonary function testing equipment act as the additional means in automated clinical
procedures and analysis techniques for carrying out a complete study of the lung function fromthe respiratory process. Besides the conventional pneumotachometry, several new techniques likethe ultrasound spirometer and microprocessor based analysers are under development. Themeasurement of gases is also important for respiratory studies. Chapter 14 gives a detaileddescription of various instruments and systems for assessing pulmonary function.
The measurement of gases like oxygen and carbon dioxide in the blood, along with blood pH
form important test parameters for studying the acid-base balance in the body. Blood gas analysers
have greatly developed in the last few years. The modern microprocessor controlled instruments
include automatic sample dilution and data processing. A separate chapter on blood gas analysersgives details of modern instruments and their use in clinical practice. Oximeters are covered inChapter 16, which describes various techniques of assessing the oxygen saturation level of bloodboth by invasive and non-invasive techniques. A chapter on blood cell counters touches uponelectronic methods of blood cell counting and microprocessor based system for makingcalculations important in haematology.
The third part contains four chapters on medical imaging systems. The last decade saw an
unprecedented progress in this area and resulted in the evolution and development of ultrasonic,computerised tomography and NMR scanners. Ultrasound has proved a useful imaging modalitybecause of its non-invasive character and ability to distinguish interfaces between soft tissues.Ultrasonic imaging systems are now applied to obtain images of almost the entire range of internalorgans in the abdomen. The chapter on ultrasound covers extensive information on this technologyand includes the physics of ultrasound, pulse echo systems including M-mode echocardiography
and a variety of scanning systems and techniques. CT scanners are considered as the most
significant development since the discovery of X-rays. In spite of their inherent high cost, severalthousands of these are now installed in hospitals around the world. Keeping in view the impacton medical diagnostics, a detailed description of the various scanning techniques in CT systemshas been given in Chapter 19. The chapter also carries information on the basic X-ray machine andimage intensifiers. Thermography—the science of visualizing and interpreting the skintemperature pattern—is another technique, which stands alongside X-ray, ultrasonic and clinicalexamination as an aid to medical diagnostics. Keeping in view its usefulness and recognizing thenon-availability of information on this topic in most of the medical electronic instrumentationbooks, a separate chapter has been included in this text.Preface to the First Edition ix

The last part with six chapters is devoted to therapeutic instruments.
Two types of instruments are commonly employed to meet cardiac emergencies. These are the
pacemakers and the defibrillators. The technology of implantable pacemakers has considerablydeveloped in the past few years, resulting in the availability of pacemakers with life long guaranteeof their activity. This has become possible due to improvements in power sources, low drain
current circuits and better encapsulation techniques. The availability of programmable
pacemakers has further helped to individualise the pacemaker treatment. Similarly,microprocessor based defibrillators have appeared in the market to give the possibility of moreefficiently delivering the defibrillating discharge by appropriately adjusting the output on thebasis of patient-electrode impedance. These two topics are covered in two separate chapters.
The use of high frequency in electro-surgical procedures is well established. There has not been
very many changes in the basic design except for the availability of solid state versions with bettersafety provisions for the patients and operators. Application of lasers for bloodless surgery andfor coagulation of fine structures in the small and sensitive organs of the body is now routinelypracticed in many centres in the world. Separate chapters cover the high frequency electro-surgicalmachines and laser applications in medicine respectively.
The maintenance of renal function in acute and chronic renal failure through dialysis is a
routinely practiced technique. Haemodialysis machines for use in hospitals contain a variety ofmonitoring and control facilities, and some of these functions have also been computerised. Therehave also been attempts to bring out a wearable artificial kidney so that patients suffering fromthis disease could enjoy a near normal life during their stay away from the dialysis centre. The
chapter on haemodialysis machines includes a description of the well established machines with
an indication of the efforts on the development of portable systems.
Physiotherapy instruments like the short-wave diathermy machine, microwave diathermy
machine and ultrasonic therapy units have acquired an established role in the hospitals. Similarly,the technique of electro-diagnosis and electrotherapy are now routinely employed in thephysiotherapy departments. An extension of this technique has been the development of smallstimulators for a variety of applications like pain relief, control of micturition, epilepsy, etc. Theinformation on these techniques is usually not available in the books on the subject. The inclusionof a full chapter on these techniques fulfils this gap.
A large number of references have been included at the end. This is to help the more interested
readers to conveniently look for extra material on the subject of their interest.
I am thankful to the Director, Central Scientific Instruments Organization, Chandigarh for kind
permission to publish this book. I am also grateful to various manufacturers of medical electronicinstruments who supplied valuable information on the products along with some interestingphotographs.
Finally, I am extremely grateful to my wife Ramesh Khandpur who helped me in correcting and
comparing the typed script. I also acknowledge the assistance provided to me in this work by mychildren Vimal, Gurdial and Popila. All of them bore the brunt of uncalled for neglect over a longperiod during the preparation of the manuscript.
R S K
HANDPURx Preface to the First Edition

Contents
Preface to the Second Edition v
Preface to the First Edition vii
Part One
Measuring,Measuring,Measuring,Measuring,Measuring, RecordingRecordingRecordingRecordingRecording andandandandand MonitoringMonitoringMonitoringMonitoringMonitoring InstrumentsInstrumentsInstrumentsInstrumentsInstruments
1. Fundamentals of Medical Instrumentation 3
1.1 Anatomy and Physiology 3
1.2 Physiological Systems of the Body 4
1.3 Sources of Biomedical Signals 12
1.4 Basic Medical Instrumentation System 14
1.5 Performance Requirements of Medical
Instrumentation Systems 16
1.6 Intelligent Medical Instrumentation Systems 18
1.7 General Constraints in Design of Medical
Instrumentation Systems 26
1.8 Regulation of Medical Devices 28
2. Bioelectric Signals and Electrodes 32
2.1 Origin of Bioelectric Signals 32
2.2 Recording Electrodes 39
2.3 Silver-silver Chloride Electrodes 48
2.4 Electrodes for ECG 50
2.5 Electrodes for EEG 58
2.6 Electrodes for EMG 59
2.7 Electrical Conductivity of Electrode Jellies and Creams 61
2.8 Microelectrodes 63

xii Contents
3. Physiological Transducers 66
3.1 Introduction 66
3.2 Classification of Transducers 66
3.3 Performance Characteristics of Transducers 67
3.4 Displacement, Position and Motion Transducers 71
3.5 Pressure Transducers 78
3.6 Transducers for Body Temperature Measurement 83
3.7 Photoelectric Transducers 94
3.8 Optical Fibre Sensors 101
3.9 Biosensors 106
3.10 Smart Sensors 109
4. Recording Systems 111
4.1 Basic Recording System 111
4.2 General Considerations for Signal Conditioners 112
4.3 Preamplifiers 114
4.4 Sources of Noise in Low Level Measurements 125
4.5 Biomedical Signal Analysis Techniques 128
4.6 Signal Processing Techniques 130
4.7 The Main Amplifier and Driver Stage 131
4.8 Writing Systems 132
4.9 Direct Writing Recorders 133
4.10 The Ink Jet Recorder 142
4.11 Potentiometric Recorder 143
4.12 Digital Recorders 146
4.13 Instrumentation Tape Recorders 151
5. Biomedical Recorders 154
5.1 Electrocardiograph 154
5.2 Vectorcardiograph (VCG) 166
5.3 Phonocardiograph (PCG) 167
5.4 Electroencephalograph (EEG) 170
5.5 Electromyograph (EMG) 178
5.6 Other Biomedical Recorders 182
5.7 Biofeedback Instrumentation 183
6. Patient Monitoring Systems 186
6.1 System Concepts 186
6.2 Cardiac Monitor 187
6.3 Bedside Patient Monitoring Systems 196
6.4 Central Monitors 198
6.5 Measurement of Heart Rate 202
6.6 Measurement of Pulse Rate 204

Contents xiii
6.7 Blood Pressure Measurement 208
6.8 Measurement of Temperature 232
6.9 Measurement of Respiration Rate 232
6.10 Catheterization Laboratory Instrumentation 238
7. Arrhythmia and Ambulatory Monitoring Instruments 243
7.1 Cardiac Arrhythmias 243
7.2 Arrhythmia Monitor 244
7.3 QRS Detection Techniques 247
7.4 Exercise Stress Testing 254
7.5 Ambulatory Monitoring Instruments 256
8. Foetal Monitoring Instruments 263
8.1 Cardiotocograph 263
8.2 Methods of Monitoring Foetal Heart Rate 264
8.3 Monitoring Labour Activity 278
8.4 Recording System 281
9. Biomedical Telemetry and Telemedicine 283
9.1 Wireless Telemetry 283
9.2 Single Channel Telemetry Systems 287
9.3 Multi-channel Wireless Telemetry Systems 292
9.4 Multi-patient Telemetry 296
9.5 Implantable Telemetry Systems 298
9.6 Transmission of Analog Physiological Signals
Over Telephone 300
9.7 Telemedicine 303
10. Oximeters 312
10.1 Oximetry 312
10.2 Ear Oximeter 316
10.3 Pulse Oximeter 318
10.4 Skin Reflectance Oximeters 322
10.5 Intravascular Oximeter 323
11. Blood Flowmeters 325
11.1 Electromagnetic Blood Flowmeter 325
11.2 Types of Electromagnetic Flowmeters 328
11.3 Ultrasonic Blood Flowmeters 331
11.4 NMR Blood Flowmeter 340
11.5 Laser Doppler Blood Flowmeter 341
12. Cardiac Output Measurement 344
12.1 Indicator Dilution Method 344
12.2 Dye Dilution Method 346

xiv Contents
12.3 Thermal Dilution Techniques 347
12.4 Measurement of Continuous Cardiac Output Derived
from the Aortic Pressure Waveform 353
12.5 Impedance Technique 354
12.6 Ultrasound Method 356
13. Pulmonary Function Analysers 358
13.1 Pulmonary Function Measurements 358
13.2 Spirometry 362
13.3 Pneumotachometers 368
13.4 Measurement of Volume 370
13.5 Pulmonary Function Analyzers 375
13.6 Respiratory Gas Analyzers 379
14. Clinical Laboratory Instruments 387
14.1 Medical Diagnosis with Chemical Tests 387
14.2 Spectrophotometry 387
14.3 Spectrophotometer Type Instruments 390
14.4 Colorimeters 397
14.5 Spectrophotometers 399
14.6 Automated Biochemical Analysis Systems 403
14.7 Clinical Flame Photometers 411
14.8 Selective-ion Electrodes Based Electrolytes
Analyzer 415
15. Blood Gas Analyzers 420
15.1 Acid-base Balance 420
15.2 Blood pH Measurement 421
15.3 Measurement of Blood PCO2425
15.4 Blood pO2 Measurement 428
15.5 Intra-arterial Blood Gas Monitoring 430
15.6 A Complete Blood Gas Analyzer 433
16. Blood Cell Counters 444
16.1 Types of Blood Cells 444
16.2 Methods of Cell Counting 446
16.3 Coulter Counters 449
16.4 Automatic Recognition and Differential
Counting of Cells 457
17. Audiometers and Hearing Aids 463
17.1 Mechanism of Hearing 463
17.2 Measurement of Sound 467
17.3 Basic Audiometer 468

Contents xv
17.4 Pure Tone Audiometer 471
17.5 Speech Audiometer 471
17.6 Audiometer System Bekesy 472
17.7 Evoked Response Audiometry System 476
17.8 Calibration of Audiometers 478
17.9 Hearing Aids 479
18. Patient Safety 486
18.1 Electric Shock Hazards 486
18.2 Leakage Currents 495
18.3 Safety Codes for Electromedical Equipment 498
18.4 Electrical Safety Analyzer 499
18.5 Testing of Biomedical Equipment 500
Part Two
ModernModernModernModernModern ImagingImagingImagingImagingImaging SystemsSystemsSystemsSystemsSystems
19. X-ray Machines and Digital Radiography 507
19.1 Basis of Diagnostic Radiology 507
19.2 Nature of X-rays 509
19.3 Production of X-rays 510
19.4 X-ray Machine 513
19.5 Visualization of X-rays 526
19.6 Dental X-ray Machines 530
19.7 Portable and Mobile X-ray Units 531
19.8 Physical Parameters for X-ray Detectors 532
19.9 Digital Radiography 533
20. X-ray Computed Tomography 538
20.1 Computed Tomography 538
20.2 System Components 545
20.3 Gantry Geometry 561
20.4 Patient Dose in CT Scanners 562
21. Nuclear Medical Imaging Systems 563
21.1 Radio-isotopes in Medical Diagnosis 563
21.2 Physics of Radioactivity 564
21.3 Radiation Detectors 567
21.4 Pulse Height Analyser 570
21.5 Uptake Monitoring Equipment 571
21.6 Radio-isotope Rectilinear Scanner 572
21.7 The Gamma Camera 574
21.8 Multi-crystal Gamma Cameras 576

xvi Contents
21.9 Emission Computed Tomography (ECT) 581
21.10 Single-photon Emission Computed Tomography (SPECT) 582
21.11 Positron Emission Tomography (PET Scanner) 587
22. Magnetic Resonance Imaging System 592
22.1 Principles of NMR Imaging Systems 592
22.2 Image Reconstruction Techniques 600
22.3 Basic NMR Components 611
22.4 Biological Effects of NMR Imaging 621
22.5 Ad vantages of NMR Imaging System 621
23. Ultrasonic Imaging Systems 623
23.1 Diagnostic Ultrasound 623
23.2 Physics of Ultrasonic Waves 623
23.3 Medical Ultrasound 631
23.4 Basic Pulse-echo Apparatus 631
23.5 A-Scan 638
23.6 Echocardiograph (M-mode) 640
23.7 B-Scanner 644
23.8 Real-time Ultrasonic Imaging Systems 646
23.9 Multi-element Linear Array Scanners 649
23.10 Digital Scan Converter 667
23.11 Biological Effects of Ultrasound 668
24. Thermal Imaging Systems 670
24.1 Medical Thermography 670
24.2 Physics of Thermography 672
24.3 Infrared Detectors 675
24.4 Thermographic Equipment 675
24.5 Quantitative Medical Thermography 678
24.6 Pyroelectric Vidicon Camera 681
24.7 Thermal Camera Based on IR Sensor with Digital
Focal Plane Array 683
Part Three
TherapeuticTherapeuticTherapeuticTherapeuticTherapeutic EquipmentEquipmentEquipmentEquipmentEquipment
25. Cardiac Pacemakers 687
25.1 Need for Cardiac Pacemaker 687
25.2 External Pacemakers 688
25.3 Implantable Pacemakers 691
25.4 Recent Developments in Implantable Pacemakers 710
25.5 Pacing System Analyser 713

Contents xvii
26. Cardiac Defibrillators 714
26.1 Need for a Defibrillator 714
26.2 DC Defibrillator 715
26.3 Implantable Defibrillators 722
26.4 Pacer—cardioverter—defibrillator 725
26.5 Defibrillator Analysers 727
27. Instruments for Surgery 728
27.1 Principle of Surgical Diathermy 728
27.2 Surgical Diathermy Machine 731
27.3 Safety Aspects in Electro-surgical Units 739
27.4 Surgical Diathermy Analysers 741
28. Laser Applications in Biomedical Field 743
28.1 The Laser 743
28.2 Pulsed Ruby Laser 747
28.3 Nd-YAG Laser 749
28.4 Helium-Neon Laser 750
28.5 Argon Laser 751
28.6 CO2 Laser 755
28.7 Excimer Lasers 758
28.8 Semiconductor Lasers 758
28.9 Laser Safety 759
29. Physiotherapy and Electrotherapy Equipment 760
29.1 High Frequency Heat Therapy 760
29.2 Short-wave Diathermy 760
29.3 Microwave Diathermy 765
29.4 Ultrasonic Therapy Unit 767
29.5 Electrodiagnostic/Therapeutic Apparatus 769
29.6 Pain Relief Through Electrical Stimulation 779
29.7 Diaphragm Pacing by Radio-frequency for
the Treatment of Chronic Ventilatory Insufficiency 783
29.8 Bladder Stimulators 784
29.9 Cerebellar Stimulators 784
30. Haemodialysis Machines 785
30.1 Function of the Kidneys 785
30.2 Artificial Kidney 788
30.3 Dialyzers 789
30.4 Membranes for Haemodialysis 795
30.5 Haemodialysis Machine 797
30.6 Portable Kidney Machines 807

xviii Contents
31. Lithotriptors 810
31.1 The Stone Disease Problem 810
31.2 First Lithotriptor Machine 811
31.3 Modern Lithotriptor Systems 813
31.4 Extra-corporeal Shock-wave Therapy 822
32. Anaesthesia Machine 825
32.1 Need for Anaesthesia 825
32.2 Anaesthesia Machine 826
32.3 Electronics in the Anaesthetic Machine 835
33. Ventilators 837
33.1 Mechanics of Respiration 837
33.2 Artificial Ventilation 839
33.3 Ventilators 840
33.4 Types of Ventilators 841
33.5 Ventilator Terms 841
33.6 Classification of Ventilators 845
33.7 Pressure-volume-flow Diagrams 848
33.8 Modern Ventilators 848
33.9 High Frequency Ventilators 851
33.10 Humidifiers, Nebulizers and Aspirators 852
34. Radiotherapy Equipment 853
34.1 Use of High Voltage X-ray Machines 853
34.2 Development of Betatron 853
34.3 Cobalt-60 Machine 854
34.4 Medical Linear Accelerator Machine 859
35. Automated Drug Delivery Systems 870
35.1 Infusion Pumps 870
35.2 Components of Drugs Infusion Systems 871
35.3 Implantable Infusion Systems 874
35.4 Closed-loop Control in Infusion Systems 876
35.5 Examples of Typical Infusion Pumps 877
References 885
Index 902

/G20/G50/G41/G52/G54/G20/G4F/G4E/G45/G20/G3A /G4D/G45/G41/G53/G55/G52/G49/G4E/G47/G2C/G20 /G52/G45/G43/G4F/G52/G44/G49/G4E/G47
/G41/G4E/G44
/G4D/G4F/G4E/G49/G54/G4F/G52/G49/G4E/G47/G20 /G49/G4E/G53/G54/G52/G55/G4D/G45/G4E/G54/G53
1. Fundamentals of Medical Instrumentation
2. Bioelectric Signals and Electrodes
3. Physiological Transducers
4. Recording Systems
5. Biomedical Recorders
6. Patient Monitoring Systems
7. Arrhythmia and Ambulatory Monitoring
Instruments
8. Foetal Monitoring Instruments
9. Biomedical Telemetry and Telemedicine
10. Oximeters
11. Blood Flowmeters
12. Cardiac Output Measurement
13. Pulmonary Function Analysers
14. Clinical Laboratory Instruments
15. Blood Gas Analyzers
16. Blood Cell Counters
17. Audiometers and Hearing Aids
18. Patient Safety

This page intentionally left blank

/G46/G75/G6E/G64/G61/G6D/G65/G6E/G74/G61/G6C/G73/G20/G6F/G66
/G4D/G65/G64/G69/G63/G61/G6C/G20/G49/G6E/G73/G74/G72/G75/G6D/G65/G6E/G74/G61/G74/G69/G6F/G6E
During the last quarter of the century, there has been a tremendous increase in the use of electrical
and electronic equipment in the medical field for clinical and research purposes. In a medicalinstrumentation system, the main function is to measure or determine the presence of somephysical quantity that may be useful for diagnostic purposes. Therefore, many types of instru-mentation systems are used in hospitals and physician’s clinics.
/G20/G31/G2E/G31 /G41/G4E/G41/G54/G4F/G4D/G59/G20/G41/G4E/G44/G20/G50/G48/G59/G53/G49/G4F/G4C/G4F/G47/G59
A knowledge of the structure of the living body and its function is essential for understandingthe functioning of most of the medical instruments. The science of structure of the body is knownas“Anatomy” and that of its function, “Physiology” .
Anatomy is classified according to the following basis:
Gross anatomy deals with the study of the structure of the organs as seen by the naked eye on
dissection. It describes the shape, size, components and appearance of the organ under study.
Topographical anatomy deals with the position of the organs in relation to each other, as they are
seen in sections through the body in different planes.
Microscopic anatomy (Histology) is the study of the minute structure of the organs by means of
microscopy.
Cytology is a special field of histology in which the structure, function and development of the
cells are studied.
Similarly, physiology, which relates to the normal function of the organs of the body, can be
classified in different ways. For example:
Cell physiology is the study of the functions of the cells.HAPTER
11

4 Handbook of Biomedical Instrumentation
Pathophysiology relates to the pathological (study or symptoms of disease) functions of the organs.
In addition, classification into various sub-areas dealing with different organs can be made.
For example:
Circulatory physiology is the study of blood circulation relating to functioning of the heart.
Respiratory physiology deals with the functioning of breathing organs.
/G20/G31/G2E/G32 /G50/G48/G59/G53/G49/G4F/G4C/G4F/G47/G49/G43/G41/G4C/G20/G53/G59/G53/G54/G45/G4D/G53/G20/G4F/G46/G20/G54/G48/G45/G20/G42/G4F/G44/G59
Human body is a complex engineering marvel, which contains various types of systems such as
electrical, mechanical, hydraulic, pneumatic, chemical and thermal etc. These systems communicateinternally with each other and also with an external environment. By means of a multi-levelcontrol system and communications network, the individual systems enable the human body toperform useful tasks, sustain life and reproduce itself.
Although, the coverage of detailed information on the physiological systems is outside the
scope of this book, nevertheless a brief description of the major sub-systems of the body is givenbelow to illustrate the engineering aspects of the human body.
/G31/G2E/G32/G2E/G31 /G54/G68/G65/G20/G43/G61/G72/G64/G69/G6F/G76/G61/G73/G63/G75/G6C/G61/G72/G20/G53/G79/G73/G74/G65/G6D
The cardiovascular system is a complex closed hydraulic system, which performs the essentialservice of transportation of oxygen, carbon dioxide, numerous chemical compounds and the bloodcells. Structurally, the heart is divided into right and left parts. Each part has two chambers calledatrium and ventricle. The heart has four valves (Fig. 1.1):
• The Tricuspid valve or right atrio-ventricular valve—between right atrium and ventricle. It
consists of three flaps or cusps. It prevents backward flow of blood from right ventricle toright atrium.
• Bicuspid Mitral or left atrio-ventricular valve—between left atrium and left ventricle.
The valve has two flaps or cusps. It prevents backward flow of blood from left ventricle toatrium.
• Pulmonary valve—at the right ventricle. It consists of three half moon shaped cusps. This
does not allow blood to come back to the right ventricle.
• Aortic valve—between left ventricle and aorta. Its construction is like pulmonary valve.
This valve prevents the return of blood back to the left ventricle from aorta.
The heart wall consists of three layers: (i) The pericardium, which is the outer layer of the heart.
It keeps the outer surface moist and prevents friction as the heart beats. (ii) The myocardium is the
middle layer of the heart. It is the main muscle of the heart, which is made up of short cylindricalfibres. This muscle is automatic in action, contracting and relaxing rythmically throughout life.(iii) The endocardium is the inner layer of the heart. It provides smooth lining for the blood to flow.
The blood is carried to the various parts of the body through blood vessels, which are hollow
tubes. There are three types of blood vessels. (i) Arteries —are thick walled and they carry the
oxygenated blood away from the heart. (ii) Veins —are thin walled and carry de-oxygenated blood

Fundamentals of Medical Instrumentation 5
towards the heart. (iii) Capillaries —are the smallest and the last level of blood vessels. They are so
small that the blood cells, which make blood, actually flow one at a time through them. There are
estimated to be over 800,000 km of capillaries in human being, which include all the arteries andveins, which carry blood.
From an engineering point of view, the heart which drives the blood through the blood vessels
of the circulatory system (Fig. 1.2) consists of four chamber muscular pump that beats about 72times per minute (on an average for a normal adult), sending blood through every part of the body.The pump acts as two synchronized but functionally isolated two stage pumps. The first stage ofeach pump (the atrium) collects blood from the hydraulic system and pumps it into the secondstage ( the ventricle). In this process, the heart pumps the blood through the pulmonary circulation
to the lungs and through the systemic circulation to the other parts of the body.
In the pulmonary circulation, the venous (de-oxygenated) blood flows from the right ventricle,
through the pulmonary artery, to the lungs, where it is oxygenated and gives off carbon dioxide.The arterial (oxygenated) blood then flows through the pulmonary veins to the left atrium.
In systemic circulation, the blood is forced through blood vessels, which are somewhat elastic.
The blood flows from the left atrium to the left ventricle and is pumped through the aorta andits branches, the arteries, out into the body. Through the arterioles (small arteries), the blood isSuperior vena cava
Pulmonary
valve
Right atrium
Inferior
vena cava
Right
ventricle
Tricuspid valveMyocardiumLeft ventricleSeptumMitral valveAortic valveLeft atrium
Pulmonary
veinsPulmonary artery
Fig.1.1 Structure of the heart

6 Handbook of Biomedical Instrumentation
distributed to the capillaries in the tissues, where it gives up its oxygen and chemical compounds,
takes up carbon dioxide and products of combustion.
The blood returns to the heart along different routes from different parts of the body. It usually
passes from the venous side of the capillaries directly via the venous system to either the superiorvena cava or the inferior vena cava, both of which empty into the right atrium. The heart itself issupplied by two small but highly important arteries, the coronary arteries. They branch from theaorta just above the heart. If they are blocked by coronary thrombosis, myocardial infarction
follows, often leading to a fatal situation.
The heart rate is partly controlled by autonomic nervous system and partly by harmone action.
These control the heart pump’s speed, efficiency and the fluid flow pattern through the system.
The circulatory system is the transport system of the body by which food, oxygen, water and
other essentials are transported to the tissue cells and their waste products are transported away.
This happens through a diffusion process in which nourishment from the blood cell diffusesIntestineLiver
Kidneys
LegsAorta
Location of
sinus nodeSemilunar
valve
Right atriumLeft atrium
Tricuspid
valve
Right ventricleLung
CO2O2
O2
LungHead
Aortic valve
Mitral valve
Left ventricle
Fig.1.2 The Circulatory system

Fundamentals of Medical Instrumentation 7
through the capillary wall into interstitial fluid. Similarly, carbon dioxide and some waste
products from the interstitial fluid diffuses through the capillary wall into the blood cell.
The condition of the cardiovascular system is examined by haemodynamic measurements and
by recording the electrical activity of the heart muscle (electrocardiography) and listening to theheart sounds (phonocardiography). For assessing the performance of the heart as a pump,
measurement of the cardiac output (amount of blood pumped by the heart per unit time), blood
pressure, blood flow rate and blood volume are made at various locations throughout thecirculatory system.
/G31/G2E/G32/G2E/G32 /G54/G68/G65/G20/G52/G65/G73/G70/G69/G72/G61/G74/G6F/G72/G79/G20/G53/G79/G73/G74/G65/G6D
The respiratory system in the human body (Fig. 1.3) is a pneumatic system in which an air pump(diaphragm) alternately creates negative and positive pressures in a sealed chamber (thoraciccavity) and causes air to be sucked into and forced out of a pair of elastic bags (lungs). The lungs
are connected to the outside environment through a passage way comprising nasal cavities,
Pharynx
(throat)
Larynx
Trachea
windpipe
(air passage)
Bronchiole
(smallest air
passage)
Pleura
Lung
Diaphragm
Thorax cavity BronchiAlveoli
(branch from bronchiole
where exchange occurs)Nose
Nasal
opening
Mouth
Fig.1.3 The Respiratory system

8 Handbook of Biomedical Instrumentation
pharynx, larynx, trachea, bronchi and bronchioles. The passage way bifurcates to carry air into
each of the lungs wherein it again subdivides several times to carry air into and out of each of themany tiny air spaces (alveoli) within the lungs. In the tiny air spaces of the lungs is a membrane
interface with the hydraulic system of the body through which certain gases can defuse. Oxygen
is taken into the blood from the incoming air and carbon dioxide is transferred from the blood tothe air under the control of the pneumatic pump. Thus, the blood circulation forms the link in thesupply of oxygen to the tissues and in the removal of gaseous waste products of metabolism. Themovement of gases between blood and the alveolar air is basically due to constant molecularmovement or diffusion from points of higher pressure to points of lower pressure.
An automatic respiratory control centre in the brain maintains heart pump operation at a speed
that is adequate to supply oxygen and take away carbon dioxide as required by the system. In eachminute, under normal conditions, about 250 ml of oxygen are taken up and 250 ml of CO
2 are given
out by the body and these are the amounts of the two gases, which enter and leave the blood in thelungs. Similar exchanges occur in reverse in the tissues where oxygen is given up and CO
2 is
removed. The exact amount of CO2 expired depends upon the metabolism, the acid-base balance
and the pattern of respiration. The exchange of gases takes place in the alveoli and can be achievedby the normal 15-20 breaths/min, each one involving about 500 ml of air.
The respiratory system variables which are important for assessing the proper functioning of
the system are respiratory rate, respiratory air flow, respiratory volume and concentration of CO
2
in the expired air. The system also requires measurements to be made of certain volumes andcapacities such as the tidal volume, vital capacity, residual volume, inspiratory reserve volumeand expiratory reserve volume. The details of these are given in Chapter 13.
/G31/G2E/G32/G2E/G33 /G54/G68/G65/G20/G4E/G65/G72/G76/G6F/G75/G73/G20/G53/G79/G73/G74/G65/G6D
The nervous system is the control and communication network for the body which coordinates thefunctions of the various organs. Rapid communication between the various parts, the effective,integrated activity of different organs and tissues and coordinated contraction of muscle are
almost entirely dependent upon the nervous system. It is thus, the most highly developed and
complex system in the body. The centre of all these activities is the brain (central informationprocessor) with memory, computational power, decision making capability and a host of inputoutput channels.
The nervous system consists of a central and a peripheral part. The central nervous system is
(Fig. 1.4) made up of the encephalon (brain) and the spinal cord. The peripheral nervous systemcomprises all the nerves and groups of neurons outside the brain and the spinal cord.
The brain consists of three parts, namely, the cerebrum ,cerebellum and the brain stem.
Cerebrum: The cerebrum consists of two well demarcated hemispheres, right and left and each
hemisphere is sub-divided into two lobes: frontal lobe and temporal lobe in the left hemisphere and
parietal and occipital lobes in the right hemisphere (Fig. 1.5). The outer layer of the brain is called the
cerebral cortex. All sensory inputs from various parts of the body eventually reach the cortex,where certain regions relate specifically to certain modalities of sensory information. Variousareas are responsible for hearing, sight, touch and control of the voluntary muscles of the body.

Fundamentals of Medical Instrumentation 9
The cerebral cortex is also the centre of intellectual functions. The frontal lobes are essential for
intelligence, constructive imagination and thought. Here, large quantities of information can bestored temporarily and correlated, thus making a basis for higher mental functions.Lumbar
spinal
cord
Sacral
spinal
cordThoracic
spinal
cordSpinal
nervePosterior
nerve
rootsAnterior
nerve
rootsBrain stemCerebellumCerebrum
Cervical
spinal
cordEncephlon
Spinal cord
Fig.1.4 Central nervous system, human brain and spinal cord

10 Handbook of Biomedical Instrumentation
Each point in the motor centre in the cerebral cortex (Fig. 1.6) corresponds to a certain body
movement. In the anterior part of the parietal lobe lies the terminal station for the nerve pathwaysconducting sensation from the opposite half of the body. The sensory centre contains counterpartsof the various areas of the body in different locations of the cortex. The sensory inputs come fromthe legs, the torso, arms, hands, fingers, face and throat etc. The amount of surface allotted to each
part of the body is in proportion to the number of sensory nerves it contains rather than its actual
physical size. The visual pathways terminate in the posterior part of the occipital lobe. The rest ofthe occipital lobes store visual memories, by means of which we interpret what we see.
On the upper side of the temporal lobe, the acoustic pathways terminate making it as a hearing
centre. This is located just above the ears. Neurons responding to different frequencies of soundinput are spread across the region, with the higher frequencies located towards the front and lowfrequencies to the rear of the ear. The temporal lobes are also of importance for the storage processin the long-term memory.
Cerebellum: The cerebellum acts as a physiological microcomputer which intercepts various
sensory and motor nerves to smooth out the muscle motions which could be otherwise jerky. It alsoconsists of two hemispheres which regulate the coordination of muscular movements elicited bythe cerebrum. The cerebellum also enables a person to maintain his balance.Cerebellum
Medulla oblongataOccipital lobeHypothalamusThalamusParietal lobeCerebral cortex
Ventricle
Corpus callosum
Frontal lobe
Pituitary
Temporal lobe
Pons
Fig.1.5 Cut-away section of the human brain

Fundamentals of Medical Instrumentation 11
Brain Stem: The brain stem connects the spinal cord to the centre of the brain just below the
cerebral cortex. The essential parts of the brain stem are (i) Medulla oblongata which is the lowest
section of the brain stem and contains centres for regulating the work performed by the heart, thevasomotor centres, which control blood distribution and respiratory centre which controls theventilation of the lungs. (ii) the pons located just above the medulla and protruding somewhat in
front of the brain stem. (iii) midbrain which lies in the upper part of the brain stem (iv) the
diencephalon is located above and slightly forward of the mid brain. It has one part, the thalamus ,
which acts as a relay station for sensory pathways to the cortical sensory centre of the cerebrum. Inthe lower part of the diencephalon is the hypothalamus which has several vital centres for
temperature regulation, metabolism and fluid regulation. They include the centres for appetite,thirst, sleep and sexual drive. The hypothalamus is important for subjective feelings and emotions.
Spinal Cord: The spinal cord is a downward continuation of the medulla oblongata in the brain to
the level of first lumbar vertebra. It consists of a cylinder of nerve tissue about the thickness of thelittle finger and has a length of about 38 to 45 cms. The cord consists of white matter on the surfaceand gray matter inside. The white matter contain fibres running between the cord and brain only.The cord containing motor and sensory fibres is responsible for the link between the brain and thebody and reflex action. In the H-shaped gray matter of the spinal cord are located the neurons thatcontrol many reflexes such as the knee reflex and the bladder- emptying reflex. The reflex action isa result of the stimulation of the motor cells by stimuli brought in by sensory nerves from thetissues.
The central nervous system consists of billions of specialized cells about half of which, called
neurons, are functionally active as signal transmitters while the other half (supporting cells),maintain and nourish the neurons. The fundamental property of the neurons is the ability to
transmit electrical signals, called nerve impulses, in response to changes in their environment, i.e.
stimuli. The central nervous system controls the voluntary muscles of the body and is responsiblefor all movements and sensations.Temporal
lobeHearingVisionOccipital
lobeParietal
lobeLeg Leg
Trunk Trunk
Arm Arm
Hand Hand
Thumb Thumb
Face Face
Mouth Mouth
Throat ThroatHigher
intellectual
functionsFrontal lobeMotor
Sensory
Fig.1.6 Sites of some activity centres in the cerebral cortex

12 Handbook of Biomedical Instrumentation
The basic functional unit of the nervous system is the neuron. A typical neuron consists of a
nucleated cell body and has several processes or branches (Fig. 1.7). The size and distribution ofthese branches vary greatly at different sites and in cells with different functions, but the two main
kinds are: the axone and the dendrite . The dendrites normally conduct impulses toward the cell
body and the axons conduct away from it.
Axone Axone
Impulse transmissionDendriteCell body
Fig.1.7 Structure of the neuron and the phenomenon of impulse transmission
The neurons form an extremely complex network, which connects all parts of the body. While
the size of the central body of the nerve cell is the same as that of other cells of the body, the overallsize of the neuron structure varies from a millimetre or so in the spinal cord to over a metre inlength. For example, the axones of the foot muscle originate in the lower part of the spinal cord,where the associated nerve cells are located.
The nervous system is the body’s principal regulatory system and pathological processes in it
often lead to serious functional disturbances. The symptoms vary greatly depending upon thepart of the nervous system affected by the pathological changes. The measurements on the nervous
system include recording of electroencephalogram (EEG) and muscle’s electrical action potentials,
electromyogram (EMG), measurement of conduction velocity in motor nerves, and recording of theperipheral nerves’ action potential, electroneurogram (ENG).
/G31/G2E/G32/G2E/G34 /G4F/G74/G68/G65/G72/G20/G53/G79/G73/G74/G65/G6D/G73
There are some other important functional systems in the body, such as digestive system, excretorysystem, reproductive system and the biochemical system which perform vital functions requiredto carry out the various body functions. The coverage of all these and other systems is outside the
scope of this book.
/G20/G31/G2E/G33 /G53/G4F/G55/G52/G43/G45/G53/G20/G4F/G46/G20/G42/G49/G4F/G4D/G45/G44/G49/G43/G41/G4C/G20/G53/G49/G47/G4E/G41/G4C/G53
Biomedical signals are those signals (phenomenon that conveys information) which are used
primarily for extracting information on a biological system under investigation. The process ofextracting information could be as simple as feeling the pulse of a person on the wrist or ascomplex as analyzing the structure of internal soft tissues by an ultrasound scanner. Biomedicalsignals originate from a variety of sources (Fig. 1.8) such as:

Fundamentals of Medical Instrumentation 13
Bioelectric Signals: These are unique to the biomedical systems. They are generated by nerve cells
and muscle cells. Their basic source is the cell membrane potential which under certain conditionsmay be excited to generate an action potential. The electric field generated by the action of manycells constitutes the bio-electric signal. The most common examples of bioelectric signals are theECG (electrocardiographic) and EEG (electroencephalographic) signals.
Bioacoustic Signals: The measurement of acoustic signals created by many biomedical pheno-
mena provides information about the underlying phenomena. The examples of such signals are:flow of blood in the heart, through the heart’s valves and flow of air through the upper and lowerairways and in the lungs which generate typical acoustic signal.
Biomechanical Signals: These signals originate from some mechanical function of the biological
system. They include all types of motion and displacement signals, pressure and flow signals etc.Electroencephalogram (nervous system)
Respiratory parameters (Pulmonary system)
Esophagus temperature
Phonocardiogram (heart sounds)
Blood pressure
(cardiovascular system)
Blood flow
(cardiovascular system)
Galvanic skin
resistanceElectrooculogram
(occular system)
Electrocardiogram
(cardiovascular
system)
Impedance
pneumography
Electromyogram
(muscular system)
Pulse oximetery
(pulmonary system)
Pulse rate
(cardiovascular system)
Fig.1.8 Sources of biomedical signals

14 Handbook of Biomedical Instrumentation
The movement of the chest wall in accordance with the respiratory activity is an example of this
type of signal.
Biochemical Signals: The signals which are obtained as a result of chemical measurements from
the living tissue or from samples analyzed in the laboratory. The examples are measurement ofpartial pressure of carbon-dioxide (pCO
2), partial pressure of oxygen (pO2) and concentration of
various ions in the blood.
Biomagnetic Signals: Extremely weak magnetic fields are produced by various organs such as the
brain, heart and lungs. The measurement of these signals provides information which is notavailable in other types of bio-signals such bio-electric signals. A typical example is that ofmagneto-encephalograph signal from the brain.
Bio-optical Signals: These signals are generated as result of optical functions of the biological
systems, occurring either naturally or induced by the measurement process. For example, bloodoxygenation may be estimated by measuring the transmitted/back scattered light from a tissue atdifferent wavelengths.
Bio-impedance Signals: The impedance of the tissue is a source of important information
concerning its composition, blood distribution and blood volume etc. The measurement of galvanicskin resistance is a typical example of this type of signal. The bio-impedance signal is also obtainedby injecting sinusoidal current in the tissue and measuring the voltage drop generated by thetissue impedance. The measurement of respiration rate based on bio-impedance technique is an
example of this type of signals.
/G20/G31/G2E/G34 /G42/G41/G53/G49/G43/G20/G4D/G45/G44/G49/G43/G41/G4C/G20/G49/G4E/G53/G54/G52/G55/G4D/G45/G4E/G54/G41/G54/G49/G4F/G4E/G20/G53/G59/G53/G54/G45/G4D
The primary purpose of medical instrumentation is to measure or determine the presence of some
physical quantity that may some way assist the medical personnel to make better diagnosis andtreatment. Accordingly, many types of instrumentation systems are presently used in hospitalsand other medical facilities. The majority of the instruments are electrical or electronic systems,although mechanical systems such as ventilators or spirometers are also employed. Because ofthe predominantly large number of electronic systems used in medical practice, the conceptsexplained hereafter are mostly related to electronic medical instruments.
Certain characteristic features, which are common to most instrumentation systems, are also
applicable to medical instrumentation systems. In the broadest sense, any medical instrument(Fig. 1.9) would comprise of the following four basic functional components:
Measurand: The physical quantity or condition that the instrumentation system measures is called
themeasurand . The source for the measurand is the human body which generates a variety of
signals. The measurand may be on the surface of the body (electrocardiogram potential) or it maybe blood pressure in the chambers of the heart.
Transducer/Sensor: A transducer is a device that converts one form of energy to another. Because of
the familiar advantages of electric and electronic methods of measurement, it is the usual practiceto convert into electrical quantities all non-electrical phenomenon associated with the measurandwith the help of a transducer. For example: a piezo-electric crystal converts mechanical vibrations

Fundamentals of Medical Instrumentation 15
into an electrical signal and therefore, is a transducer. The primary function of the transducer is to
provide a usable output in response to the measurand which may be a specific physical quantity,property or condition. In practice, two or more transducers may be used simultaneously to makemeasurements of a number of physiological parameters.
Another term ‘sensor’ is also used in medical instrumentation systems. Basically, a sensor
converts a physical measurand to an electrical signal. The sensor should be minimally invasiveand interface with the living system with minimum extraction of energy.
Signal Conditioner: Converts the output of the transducer into an electrical quantity suitable for
operation of the display or recording system. Signal conditioners may vary in complexity froma simple resistance network or impedance matching device to multi-stage amplifiers and othercomplex electronic circuitry. Signal conditioning usually include functions such as amplification,filtering (analog or digital) analog-to-digital and digital-to-analog conversion or signaltransmission circuitry. They help in increasing the sensitivity of instruments by amplification ofthe original signal or its transduced form.
Display System: Provides a visible representation of the quantity as a displacement on a scale, or
on the chart of a recorder, or on the screen of a cathode ray tube or in numerical form. Although,most of the displays are in the visual form, other forms of displays such as audible signals fromalarm or foetal Doppler ultrasonic signals are also used. In addition of the above, the processedsignal after signal conditioning may be passed on to:
Alarm System —with upper and lower adjustable thresholds to indicate when the measurand goes
beyond preset limits.
Data Storage —to maintain the data for future reference. It may be a hard copy on a paper or on
magnetic or semiconductor memories.
Data Transmission —using standard interface connections so that information obtained may be
carried to other parts of an integrated system or to transmit it from one location to another.Energy source
Electric
Light
Infrared
Mechanical
UltrasoundCalibration
Sensor/
transducerPre-
amplifierSignal
processing
Control
systemAlarms
Display
Data
storage
Data
transmission
Data
recordingMeasurandSignal conditioner
Fig.1.9 General block diagram of a medical instrumentation system

16 Handbook of Biomedical Instrumentation
In most of the medical instrumentation systems, some form of calibration is necessary at regular
intervals during their operation. The calibration signal is usually applied to the sensor input or asearly in the signal conditioning chain as possible.
In many measurements in the medical field, some form of stimulus or energy is given to the
patient and the effect it has on the patient is measured. The stimulus may be visual in the form of
flash of light or audio tone or direct electrical stimulation of some part of the nervous system. A
typical example is that of recording of the evoked response with EEG machine when visual/audiblestimulus is given to the subject under test.
In some situations, it is required to have automatic control of the transducer, stimulus or signal
conditioning part of the system. This is achieved by using a feedback loop in which part of theoutput from the signal conditioning or display device is fed back to the input stage. Control andfeedback may be automatic or manual. Almost all measuring and recording equipment is nowcontrolled by microprocessors as this makes it possible to design equipment that requires minimaluser intervention, calibration and set up procedure.
Measurements on the human body can be made at several levels on the functional systems and
sub-systems. For example; it is easiest to make measurements on the human body as a whole dueto accessible environment. Examples of measurement made on the human body are recording ofelectrocardiogram and measurement of temperature. The next level of measurements can be madeon the major functional systems of the body such as the cardiovascular system, the pulmonarysystem and so on. Many of the major systems communicate with each other as well as withexternal environment. The functional systems can be further sub-divided into sub-systems and
organs and still smaller units up to the cellular and molecular level. Measurements in the medical
field are made at all these levels with specially designed instruments with appropriate degree ofsophistication.
Measurements in the medical field can be classified into two types: in vivo and in vitro .In vivo
measurement is made on or within the living organism itself, such as measurement of pressure inthe chambers of the heart. On the other hand, in vitro measurement is performed outside the body.
For example; the measurement of blood glucose level in a sample of blood drawn from the patientrepresent in vitro measurement.
/G20/G31/G2E/G35 /G50/G45/G52/G46/G4F/G52/G4D/G41/G4E/G43/G45/G20/G52/G45/G51/G55/G49/G52/G45/G4D/G45/G4E/G54/G53/G20/G4F/G46/G20/G4D/G45/G44/G49/G43/G41/G4C
/G49/G4E/G53/G54/G52/G55/G4D/G45/G4E/G54/G41/G54/G49/G4F/G4E/G20/G53/G59/G53/G54/G45/G4D/G53
Information obtained from a sensor/transducer is often in terms of current intensity, voltage level,
frequency or signal phase relative to a standard. Voltage measurements are the easiest to make,
as the signal from the transducer can be directly applied to an amplifier having a high input
impedance. However, most of the transducers produce signal in terms of current, which can beconveniently converted into voltage by using operational amplifiers with appropriate feedback.
To make an accurate measurement of voltage, it is necessary to arrange that the input impedance
of the measuring device must be large compared with the output impedance of the signal source.This is to minimize the error that would occur, if an appreciable fraction of the signal sourcewere dropped across the source impedance. Conversely, accurate measurement of current source

Fundamentals of Medical Instrumentation 17
signals necessitates that the source output impedance be large compared with the receiver
input impedance. Ideally, a receiver that exhibits a zero input impedance would not cause anyperturbation of the current source. Therefore, high-impedance current sources are more easily
handled than low-impedance current sources.
In general, the frequency response of the system should be compatible with the operating range
of the signal being measured. To process the signal waveform without distortion, the bandpass
of the system must encompass all of the frequency components of the signal that contributesignificantly to signal strength. The range can be determined quantitatively by obtaining a Fourieranalysis of the signal. The bandpass of an electronic instrument is usually defined as the rangebetween the upper and lower half-power frequencies.
The electrical signals are invariably accompanied by components that are unrelated to the
phenomenon being studied. Spurious signal components, which may occur at any frequencywithin the band pass of the system are known as noise. The instruments are designed in such away that the noise is minimised to facilitate accurate and sensitive measurement. For extraction ofinformation from noisy signals, it is essential to enhance signal-to-noise ratio, for which severaltechniques have been put in practice. The simplest method is that of bandwidth reduction,although many sophisticated methods have been developed to achieve noise reduction from thenoisy bio-medical signals.
The recent progress of digital technology in terms of both hardware and software, makes more
efficient and flexible digital rather than analog processing. Digital techniques have severaladvantages. Their performance is powerful as they are able to easily implement even complex
algorithms. Their performance is not effected by unpredictable variable such as component aging
and temperature which can normally degrade the performance of analog devices. Moreover, designparameters can be more easily changed because they involve software rather than hardwaremodifications.
The results of a measurement in medical instruments are usually displayed either on analog
meters or digital displays. Digital displays present the values of the measured quantities innumerical form. Instruments with such a facility are directly readable and slight changes in theparameter being measured are easily discernible in such displays, as compared to their analogcounterparts. Because of their higher resolution, accuracy and ruggedness, they are preferred fordisplay over conventional analog moving coil indicating meters. Different types of devices areavailable for display in numerical form.
Light emitting diodes (LED) are used in small sized seven-segment displays. These semi-
conductor diodes are made of gallium arsenide phosphide and are directly compatible with 5 Vsupplies typically encountered in digital circuitry. LEDs are very rugged and can withstand largevariations in temperature. LEDs are available in deep-red, green and yellow colours.
Liquid crystal displays (LCD) are currently preferred devises for displays as they require very
low current for their operation. LCDs with large screen sizes and full colour display capabilitiesare available commercially and are finding extensive and preferable applications in laptop
computers and many portable medical instruments.
Since computers are used increasingly to control the equipment and to implement the man-
machine interface, there is a growing appearance of high resolution colour graphic screens to

18 Handbook of Biomedical Instrumentation
display the course of vital signs relating to physiological variables, laboratory values, machine
settings or the results of image processing methods such as magnetic resonance tomography. Theanalog and digital displays have been largely replaced by video display units, which present
information not only as a list of numbers but as elegant character and graphic displays and
sometimes as a 3 dimensional colour display. Visual display units (VDU) are usually monochromeas the CRTs in these units are coated with either white or green phosphors. Coloured videodisplay units are employed in such applications as patient monitoring system and colour Dopplerechocardiography.
A keyboard is the most common device connected into almost all form of data acquisition,
processing and controlling functions in medical instruments. A keyboard can be as simple as anumeric pad with function keys, as in a calculator or complete alphanumeric and type writerkeyboard with associated group of control keys suitable for computer data entry equipment.Most available keyboards have single contact switches, which are followed by an encoder toconvert the key closures into ASCII (American Standard Code for Information Interchange) codefor interfacing with the microprocessor.
/G20/G31/G2E/G36 /G49/G4E/G54/G45/G4C/G4C/G49/G47/G45/G4E/G54/G20/G4D/G45/G44/G49/G43/G41/G4C/G20/G49/G4E/G53/G54/G52/G55/G4D/G45/G4E/G54/G41/G54/G49/G4F/G4E/G20/G53/G59/G53/G54/G45/G4D/G53
Intelligent technology is pervading every area of modern society, from satellite communications towashing machines. The medical instrumentation field is no exception from this reality. In thiscase, the goal of intelligent devices is to assure high quality of life by providing optimal health caredelivery in home care, emergency situations, diagnosis, surgical procedures and hospitalization.Medicine is now equipped with more and more signals and images taken from the human body,complex models of physiological systems and armaments of therapeutic procedures and devices.Careful observation of this process shows a congestion of the decision-making activities of medicalpersonnel. To solve this problem, some method of integrating all patient information into a conciseand interpretative form is necessary. The availability of high performance microprocessors,microcontrollers and personal computers has given powerful tools in the hands of medicalprofessionals which offers them intelligent and efficient monitoring and management of thepatients.
/G31/G2E/G36/G2E/G31 /G55/G73/G65/G20/G6F/G66/G20/G4D/G69/G63/G72/G6F/G70/G72/G6F/G63/G65/G73/G73/G6F/G72/G73/G20/G69/G6E/G20/G4D/G65/G64/G69/G63/G61/G6C/G20/G49/G6E/G73/G74/G72/G75/G6D/G65/G6E/G74/G73
The application of microprocessors in medical instrumentation, has matured following a seriesof stages. In the first stage, the microprocessors simply replaced conventional hard wiredelectronic systems that were used for processing data. This resulted in more reliable and fasterdata. This was followed soon by use of the microprocessor to control logic sequences required ininstrumentation. Thus, the microprocessor replaced programming devices as well as manualprogramming, making possible digital control of all of the functions of the medical instruments.
With the availability of more powerful microprocessors and large data storage capacity, it has
become possible to optimize the measurement conditions.
Extensive use has been made of microprocessors in medical instruments designed to perform
routine clinical measurements, particularly in those situations where data computing and

Fundamentals of Medical Instrumentation 19
processing could be considered as a part of measurement and diagnostic procedure. The
incorporation of microprocessors into instruments enables to have a certain amount of intelligenceor decision-making capability. The decision-making capability increases the degree of automation
of the instrument and reduces the complexity of the man-machine interface. Life support systems
have been designed with numerous safety back-up features and real-time self-diagnosticsand self-repair facilities. The reliability of many transducers has been improved and manymeasurements can now be made non-invasively because of the added computational ability ofmicroprocessors. The computational capability makes possible features such as automaticcalibration, operator guidance, trend displays, alarm priority and automated record keeping. Useof microprocessors in various instruments and systems has been explained at various places inthe text.
Microprocessors have been used to replace the complicated instructional procedures that are
now required in several medical instruments. Microprocessor based instrumentation is enablingto incorporate the ability to make intelligent judgement and provide diagnostic signals in case ofpotential errors, provide warnings or preferably make appropriate corrections. Already, themicroprocessors are assisting in instruction-based servicing of equipment. This is possibleby incorporating monitoring circuits that will provide valuable diagnostic information on
potential instrumentation failure modes and guide the operator in their correction. The instrument
diagnostic microprocessor programs would sense such a potential failure of the unit and theoperator is informed to remove and service the defective part while the measurement work proceedsuninterrupted.
/G31/G2E/G36/G2E/G32 /G54/G68/G65/G20/G4D/G69/G63/G72/G6F/G70/G72/G6F/G63/G65/G73/G73/G6F/G72
The microprocessor, in essence, consists of basic circuit elements such as transistors, resistors anddiodes which when combined form the basic logical elements, namely AND, OR and INVERTERS.
In principle, the complete operation of the microprocessor could be described by a combination of
these devices. More complex circuit elements such as flip-flops, counters, registers and arithmeticlogic unit are formed from these gates and go to make the complete microprocessor. Microprocessoris a single integrated circuit with 40 or even 64 or even higher connection pins.
Microprocessors are usually classified depending upon their word length. The word length of
a microprocessor defines the basic resolution and memory addressing capability. For example, an8-bit microprocessor will perform all calculations on binary numbers with 8 digits. 8 binary digitsgive a decimal number between 0-255.
The microprocessor’s most powerful asset is its enormous speed of operation. This is possible
because the microprocessor can store all the necessary instructions and data, until required inmemory. Memory which is usually external serves as a place to store instructions that direct the
activities of the Central Processing Unit (CPU) and data that are processed by the CPU (Fig. 1.10).It is arranged in two forms:
(a)Read Only M emory (ROM) to hold the program of instructions in binary digital form. The
contents of this memory cannot be altered by the functioning of the microprocessor system.
(b)Random Access M emory (RAM) to hold results and variable data, for making calculations,
remembering trends and assembling information for display devices.

20 Handbook of Biomedical Instrumentation
Many of the address locations in a typical system are storage locations in the memory. When a
memory location is addressed, the memory may store the information that resides on the data bus.This is called MEMORY WRITE. Addressing stored information to be placed on the data bus foruse by the CPU constitutes a MEMORY READ.
The microprocessor can rapidly access any data stored in memory, but often memory is not
large enough to store the entire data bank for particular application. This problem can be solvedusing external storage equipment, such as floppy disk or hard disk system. A microprocessor also
requires input/output ports, through which it can communicate its results with the outside words,
like a display or peripheral device or provide control signals that may direct another system.
As microprocessor systems are based on the binary numbering system, it is necessary to use
multiple connections generally 8,16 or 32 between each of the integrated circuits (chips). Theseinterconnections are usually referred to as buses. There are three buses in a microprocessor system.
Data Bus: A bidirectional path on which data can flow between the CPU and memory or input/
output. It carries the actual data being manipulated.
Address Bus: A unidirectional group of lines that identify a particular memory location or input/
output device.
Control Bus: It carries all the control and timing signals. It is a unidirectional set of signals that
indicate the type of activity in current process. The types of activities could be memory read,
memory write, input/output read, input/ output write and interrupt acknowledge.
The operation of the microprocessor and synchronisation of various activities under its control
is maintained by a crystal controlled clock or oscillator, which is usually at a fixed frequency,generally greater than 5 MHz.
The heart of a microprocessor based system is the central processing unit (CPU). It requests
instructions prepared by the programmer, calls for data and makes decisions related to theinstructions. Based on the data, the processor determines appropriate actions to be performed by
other parts of the system. Since there are many peripherals associated with the given system, theControl busAddress bus
Data busMemory I/O
CPU
Fig.1.10 Building blocks of a microcomputer

Fundamentals of Medical Instrumentation 21
microprocessor must be capable of selecting a particular device. It identifies each device by means
of a unique address code. A typical microprocessor has 16 binary address lines providing 65, 536addressing codes. Data to and from the processor is carried across a bi-directional 8 or 16 bit wide
data bus. Many processors also provide a serial data path. Several microprocessors use a multi-
plexed address/data bus on which both address and data are transmitted on the same signalpaths. In this case, the first portion of the bus cycle transmits the address while data transfer takesplace later in the cycle. This architecture is popular for microprocessors with an 8-bit data bus.
Another important link in the system is the set of input-output (I/O) interfaces. These interfaces
include all the information channels between the system and the real world. There are digitalports through which programs and control commands may be loaded and from which digital datamay be transmitted to peripherals such as keyboard, printers and floppy drive etc.
The assembly language of a microprocessor enables to extract the greatest run-time performance
because it provides for direct manipulation of the architecture of the processor. However, it isalso the most difficult language for writing programs, so it falls far from the optimal languageline.
The C language which is used to develop modern versions of the Unix operating system
provides a significant improvement over assembly language for implementing most applications,it is the language of choice for real time programming. It is an excellent compromise between a lowlevel assembly language and a high level language. C is standardized and structured. C programsare based on functions that can be evolved independently of one another and put together toimplement an application. These functions are to software just as black boxes are to hardware.
C programs are transportable. By design, a program developed in C on one type of processor can
be relatively easily transported to another.
/G31/G2E/G36/G2E/G33 /G54/G68/G65/G20/G4D/G69/G63/G72/G6F/G63/G6F/G6E/G74/G72/G6F/G6C/G6C/G65/G72/G73
Amicrocontroller , contains a CPU, clock circuitry, ROM, RAM and I/O circuitry on a single
integrated circuit package. The microcontroller is therefore, a self-contained device, which doesnot require a host of associated support chips for its operation as conventional microprocessorsdo. It offers several advantages over conventional multichip systems. There is a cost and space ad-
vantage as extra chip costs and printed circuit board and connectors required to support multichip
systems are eliminated. The other advantages include cheaper maintenance, decreased hardwaredesign effort and decreased board density, which is relevant in portable medical equipment.
Microcontrollers have traditionally been characterised by low cost high volume products
requiring a relatively simple and cheap computer controller. The design optimization parametersrequire careful consideration of architectural tradeoffs, memory design factors, instruction size,memory addressing techniques and other design constraints with respect to area and performance.Microcontrollers functionality, however, has been tremendously increased in the recent years.Today, one gets microcontrollers, which are stand alone for applications in data acquisition systemand control. They have analog-to-digital converters on chip, which enable them direct use ininstrumentation. Another type of microcontroller has on-chip communication controller, whichis designed for applications requiring local intelligence at remote nodes and communicationcapability among these distributed nodes. Advanced versions of the microcontrollers in 16-bit

22 Handbook of Biomedical Instrumentation
configuration have been introduced for high performance requirements particularly in appli-
cations where good arithmetical capabilities are required.
/G31/G2E/G36/G2E/G34 /G49/G6E/G74/G65/G72/G66/G61/G63/G69/G6E/G67/G20/G41/G6E/G61/G6C/G6F/G67/G20/G53/G69/G67/G6E/G61/G6C/G73/G20/G74/G6F/G20/G4D/G69/G63/G72/G6F/G70/G72/G6F/G63/G65/G73/G73/G6F/G72/G73
It is well-known that we live in an analog world. Virtually, all information we need to acquire fromthe human body and eventually analyze is in the analog form i.e. the signals consist of manywaveforms that continuously vary as a function of time. Examples include electrocardiograph,pressure signals and pulse waveform.
For interfacing analog signals to microprocessors/microcomputers, use is made of some kind
of data acquisition system. The function of this system is to acquire and digitize data, often fromhostile clinical environments, without any degradation in the resolution or accuracy of the signal.Since software costs generally far exceed the hardware costs, the analog/digital interface structuremust permit software effective transfers of data and command and status signals to avail of the
full capability of the microprocessor.
The analog interface system, in general, handles signals in the form of voltages. The physical
parameters such as temperature, flow, pressure, etc. are converted to voltages by means of
transducers. The choice and selection of appropriate transducers is very important, since the datacan only be as accurate as the transducer.
Figure 1.11 shows a block diagram of a universal interface circuit for connecting analog signals
to microprocessors. It basically comprises a multiplexer, instrumentation (buffer) amplifier, asample-and-hold circuit, analog-to-digital converter (ADC), tristate drivers and control logic. Thesecomponents operate under the control of interface logic that automatically maintains the correctorder of events.
Multiplexer
Multiplexer
address
Address
decoderControl
logicA-D
converter
Interrupt request
Wait/go controlControl busAddress bus
Control
busEnable SC EOCHold
capacitorHold
AwAnalog
inputsBuffer
amp.Sample and
hold amp.Tristate
drivers
Data
bus
Fig.1.11 Interfacing analog signals to microcomputers
Multiplexer: The function of the multiplexer is to select under address control, an analog input
channel and connected to the buffer amplifier. The number of channels is usually 8 or 16.

Fundamentals of Medical Instrumentation 23
Depending on its input configuration, the multiplexer will handle either single ended or
differential signals.
The address logic of most multiplexers can perform both random and sequential channel
selection. For real time systems, the random mode permits the multiplexer to select any channelwhen the program responds to a peripheral service request. Sequential channel selection, as the
name imploys involves addressing each channel in order.
Buffer Amplifier: The buffer amplifier conditions the selected input signal to a suitable level for
application to the A/D converter. Driven by the multiplexer, the buffer amplifier, which is usuallyan instrumentation amplifier, provides impedance buffering, signal gain and common mode
rejection. It has a high input impedance, 100 Mohms or more to reduce the effects of any signal
distortion caused by the multiplexer. The high input impedance also minimizes errors due to thefinite on-resistance of the multiplexer channel switches.
To improve system sensitivity, the amplifier boosts the input signal. If it is required to have
analog signals of differing ranges, connected to the multiplexer input, then a programmable gainamplifier would be preferable where the gain would be set in accordance with the multiplexerselection address. The use of programmable gain amplifiers removes the necessity to standardizeon the analog input ranges.
Sample and Hold Circuit: The A/D converter requires a finite time for the conversion process,
during which time the analog signal will still be hanging according to its frequency components.It is therefore necessary to sample the amplitude of the input signal, and hold this value on theinput to the A/D converter during the conversion process. The sample and hold circuit freezes itsoutput on receipt of a command from the control circuit, thereby providing an essentially constantvoltage to the A/D converter throughout the conversion cycle.
The sample hold is essentially important in systems having resolution of 12-bits or greater, or in
applications in which real time inputs are changing rapidly during a conversion of the sampledvalue. On the other hand, a sample hold may not be required in applications where input variationis low compared to the conversion time.
A/D Converter: The A/D converter carries out the process of the analog-to-digital conversion. It is
a member of the family of action/status devices which have two control lines—the start conversionor action input line and the end of conversion or status output line.
An A/D converter is a single chip integrated circuit having a single input connection for
the analog signal and multiple pins for digital output. It may have 8,12,16 or even more outputpins, each representing an output bit. The higher the number of bits, the higher the precision ofconversion. Each step represents a change in the analog signal: 8-bits gives 256 steps, 12-bitsprovides 4096 steps and we get 32768 steps with 16 output bits.
The key parameters in A/D converters are:
•Resolution of the A/D converter is a measure of the number of discrete digital code that it
can handle and is expressed as number of bits (binary). For example, for an 8-bit converter,the resolution is 1 part in 256.
•Accuracy is expressed as either a percentage of full scale or alternatively in bits of resolu-
tion. For example, a converter may be termed 12-bit accurate if its error is 1 part in 4096. The

24 Handbook of Biomedical Instrumentation
sources of error contributing to the inaccuracy of a converter or linearly, gain, error and off-
set error.
•Integral non-linearity is a measure of the deviation of the transfer function from a straight
line.
•Off-set error is a measure of the difference of the analog value from the ideal at a code of all
zeroes.
•Gain error represents the difference in slope of the transfer function from the ideal.
•Speed of an A/D converter is generally expressed as its conversion time, i.e. the time elapsed
between application of a convert command and the availability of data at its outputs. Thespeed of D/A converter is measured by its settling time for a full scale digital input change.
Each of the above parameters is temperature-dependent and they are usually defined at 25° C.
Tri-state Drivers: The tri-state drivers provide the necessary isolation of the A/D converter output
data from the microprocessor data bus and are available as 8-line units. Thus, for the 10 or 12 bitconverters, two drivers would be required which would be enabled by two different read addressesderived from the address decoder.
Some A/D converters have in-built tri-state drivers. However, because of their limited drive
capability, they can be used only on lightly loaded buses. For heavily loaded systems, as inmicrocomputers, the built-in drivers are permanently enabled and separate tri-state driversemployed for the data bus isolation.
Control Logic: The control logic provides the necessary interface between the microprocessor
system but the elements of the acquisition unit in providing the necessary timing control. It is toensure that the correct analog signal is selected, sampled at the correct time, initiate the A/Dconversion process (start-conversion = SC) and signals to the microprocessors on completion ofconversion (End of conversion = EOC).
Output Interface: Digital output signals often have to be converted into analog form so that they
can be used and acted upon by external circuits, e.g., oscilloscope, chart recorder, etc. Therefore,digital-to-analog (D/A) converters are used for converting a signal in a digital format into ananalog form. The output of the D/A converter is either current or voltage when presented with abinary signal at the input.
The input coding for the D/A converter is similar to the output coding of the A/D converter,
while full-scale outputs are jumper-selectable for 0 to ± 1, ± 5 and ± 10 V. D/A converters generallydeliver the standard 4 to 20 mA output and loading can range from 50 W to 4 kW. The importantparameters which govern the choice of an A/D converter or D/A converter are resolution,measurement frequency, input characteristics, offset error, noise, microprocessor compatibilityand linearity, etc.
/G31/G2E/G36/G2E/G35 /G50/G43/G20/G42/G61/G73/G65/G64/G20/G4D/G65/G64/G69/G63/G61/G6C/G20/G49/G6E/G73/G74/G72/G75/G6D/G65/G6E/G74/G73
An area of intense commercial activity in the field of computers is due to the popularity of the socalled personal computers (PC) or home computers. The low cost and increasing power of the

Fundamentals of Medical Instrumentation 25
personal computers are making them popular in the medical field. Also, software for personal
computers is largely commercially available and the users can purchase and use it conveniently.Personal computers are now well established and widely accepted in the medical field for data
collection, manipulation and processing and are emerging as complete workstations for a variety
of applications. A personal computer becomes a workstation with the simple installation of one ormore “instruments-on-a-board” in its accessory slots, and with the loading of the driver softwarethat comes with each board. The concept has proven to be ideal instrument, providing a low costyet highly versatile computing platform for the measurement, capture, analysis, display andstorage of data derived from a variety of sources.
Fig. 1.12 illustrates the typical configuration of a PC based workstation. It is obvious that the
system is highly flexible and can accommodate a variety of inputs, which can be connected to a PCfor analysis, graphics and control. Basic elements in the system include sensors or transducersthat convert physical phenomena into a measurable signal, a data acquisition system (a plug-ininstrument/acquisition board), an acquisition/analysis software package or programme andcomputing platform. The system works totally under the control of software. It may operate fromeither the PC’s floppy and/or hard disk drive. Permanent loading or unloading of driver files canbe accomplished easily. However, for complex applications, some programming in one or several
of the higher level programming languages such as ‘C language’ may be needed. Data received
from the measurements can be stored in a file or output to a printer, plotter or other device via oneof the ports on the computer.
Physical
phenomena
Biomedical
signalsTransducers Signal conditioning Data acquisition board
or modulePersonal computer
Software
Fig.1.12 Typical configuration of PC based medical instrument
PC based medical instruments are gaining in popularity for several reasons including price,
programmability and performance specifications offered. Software development, rather thanhardware development, increasingly dominates new product design cycles. Therefore, one of themost common reasons why system designers are increasingly choosing PC and PC architecture isfor its rich and cost effective software tool set. This includes operating systems, device drivers,
libraries, languages and debudging tools. Several examples of PC based medical instruments can
be found at various places in the book.

26 Handbook of Biomedical Instrumentation
/G20/G31/G2E/G37 /G47/G45/G4E/G45/G52/G41/G4C/G20/G43/G4F/G4E/G53/G54/G52/G41/G49/G4E/G54/G53/G20/G49/G4E/G20/G44/G45/G53/G49/G47/G4E/G20/G4F/G46/G20/G4D/G45/G44/G49/G43/G41/G4C
/G49/G4E/G53/G54/G52/G55/G4D/G45/G4E/G54/G41/G54/G49/G4F/G4E/G20/G53/G59/G53/G54/G45/G4D/G53
Medical equipment are primarily used for making measurements of physiological parameters of
the human body and also in some cases a stimulus or some kind of energy is applied to the human
body for diagnosis and treatment. Some of the important factors, which determine the design of amedical measuring instrument, are:
• Measurement Range: Generally the measurement ranges are quite low compared with
non-medical parameters. Most signals are in the microvolt range.
• Frequency Range: Most of the bio-medical signals are in the audio frequency range or
below and that many signals contain dc and very low frequency components.
These general characteristics of physiological signals limit the practical choices available to
designers of medical instruments. Besides, there are some additional constraints, which need to beconsidered while designing a measurement system for medical applications. Some of these are:
Inaccessibility of the Signal Source: One of the major problems in making measurements from a
living system is the difficulty in gaining access to the source of the physiological variable beingmeasured. For example; measurement of intracranial pressure in the brain requires the placementof a sensor in the brain, which is quite a difficult task. Besides, the physical size of many sensorsmay put a constraint for its use on the area of interest. Evidently, such inaccessible physiologicalvariables must be measured indirectly. The typical example of making indirect measurement ofblood pressure on the brachial artery is that of using cuff based Korotoff method. In such cases,corrections need to be applied to data that might have been affected due to the indirect measuringprocess.
Variability of Physiological Parameters: Physiological variables of interest for measurement from
the human body are rarely deterministic as they are generally time-variant. In other words, manymedical measurements vary widely among normal patients even when conditions are similar.Therefore, the physiological variable must be represented by some kind of empirical, statisticaland probabilistic distribution function.
Many internal anatomical variations exist among patients and therefore, the variability of
physiological parameters from one patient to another is a normal observation. Therefore, statistical
methods are employed in order to establish relationships among variables.
Interference among Physiological Systems: Many feedback loops exist among physiological
systems and many of the interrelationships amongst them contribute to this inherent variability ofphysiological signals. In other words, stimulation of one part of a given system generally affects
all other parts of that system in some way. Also, unlike many complex non-medical systems, a
biological system is of such a nature that it is not possible to turn it off and remove parts of it duringmeasurement procedure to avoid interference from undesirable physiological signals.
Transducer Interface Problems: All measurement systems are affected in some way by the presence
of the measuring transducer. The problem gets compounded while making measurement on the
living system where the physical presence of the transducer may change the reading significantly.Also, the presence of a transducer in one system can affect responses in other systems.

Fundamentals of Medical Instrumentation 27
Adequate care needs to be taken while designing a measuring system to ensure that the loading
effect of the transducer is minimal on the source of the measured variable.
High Possibility of Artifacts: The term artifact refers to an undesirable signal that is extraneous to
the physiological variable under measurement. The examples of artifacts are: 50 Hz electricalinterference, cross talk and noise generated within the measuring instrument. A major source ofartifacts in medical instruments is due to the movement of the subject. Many of the transducers aresensitive to the movement and therefore, the movement of the subject result in generating spurious
signals, which may even be large enough to obscure the signal of interest. This type of situation
puts a heavy demand on the signal conditioning part of the measurement system.
Safe Levels of Applied Energy: Nearly all biomedical measurements require some form of energy
to be applied to the living tissue or some energy gets applied as an incidental consequence of
transducer operation. For example, ultrasonic imaging techniques depend upon externally applied
ultrasound energy to the human body. Safe levels of the various types of energy on the humansubjects are difficult to establish. However, designers of medical instruments depend upon a largenumber of studies carried out by numerous researchers, which establish the threshold of adverseaffects by the applied energy.
Patient Safety Considerations: Medical instruments have to be physically connected to the patient
in some way or the other. In case it happens to be an electric or electronic equipment, the possibilityof an electric shock hazard is very strong unless adequate measures have been taken in the designof the equipment. In addition, the equipment is used by non-technical medical and paramedicalstaff and their safety needs also to be ensured. Various organizations at national and internationallevel have laid down specific guidelines to provide for the safety and effectiveness of the medicaldevices intended for use on human subjects.
Reliability Aspects: In case of life saving equipment like defibrillators, their failure to operate or
provide desired output can become a potential life threat for the patient. Therefore, equipmentmust be reliable, simple to operate and capable of withstanding physical abuse due to trans-portation within the hospital or in the ambulances and exposure to corrosive chemicals.
Human Factor Considerations: As a result of the increasing complexity of medical devices and
systems, the demand on physicians and paramedical staff using the equipment have continued togrow. The equipment requires a high amount of information exchange between itself and the userin order to monitor and control the technical functions of the system. Further more, medical staffgenerally have only little experience in working with complex technical system. There is a risk thatthe medical staff is not able to master the equipment adequately for every task. This inadequacycan increase the probability of error and reduce the quality and reliability of a clinical procedure.As a result, the desired or intended performance of the whole system may not be achieved due to
deficiencies in man-machine interaction. The user interface design issues therefore assume more
and more importance in case of medical equipment.
Government Regulations: During the initial stages of introduction of technology and a range of
diagnostic and therapeutic devices in the medical field, there was almost no government control
on their design, testing and sales. Situation is rapidly changing and government regulations are
being introduced to ensure that the equipment perform their intended function and are safe to

28 Handbook of Biomedical Instrumentation
operate and function. Designers of medical instruments should therefore be fully conversant with
all such regulations on a particular product or system issued by national and internationalagencies.
It is thus obvious that there are many factors that impose constraints on the design of medical
instruments. In addition to these, there are general considerations, which need to be considered
into the initial design and development of a medical instrument. These factors are:
Signal Considerations: Type of sensor, sensitivity, range, input impedance, frequency response,
accuracy, linearity, reliability, differential or absolute input.
Environmental Considerations: Signal-to-noise ratio, stability with respect to temperature, pressure,
humidity, acceleration, shock, vibration, radiation etc.
Medical Considerations: Invasive or non-invasive technique, patient discomfort, radiation and heat
dissipation, electrical safety, material toxicity etc.
Economic Considerations : Initial cost, cost and availability of consumables and compatibility with
existing equipment.
Obviously, a project for a commercial medical instrument is quite complex which must take into
consideration several factors before it is launched for design and development. In addition, theassociation of the engineering design team with motivated medical professionals is essential forthe success of the project. This association is useful not only during the development process, butalso for the clinical trials of the product so developed.
/G20/G31/G2E/G38 /G52/G45/G47/G55/G4C/G41/G54/G49/G4F/G4E/G20/G4F/G46/G20/G4D/G45/G44/G49/G43/G41/G4C/G20/G44/G45/G56/G49/G43/G45/G53
The medical instrumentation industry in general and hospitals in particular are required to bemost regulated industries. This is because when measurements are made on human beings and bythe human beings, the equipment should not only be safe to operate but must give intendedperformance so that the patients could be properly diagnosed and treated. Adequate measuresneed to be evolved so that the users of medical equipment are not subject to legal, moral and ethicalissues in their practice since they deal with the health of the people which could at times be asvital as the question of life and death. To minimize such type of problems, various countries haveintroduced a large number of codes, standards and regulations for different types of equipmentand facilities. It is therefore, essential that engineers understand their significance and be aware ofthe issues that are brought about by technological and economic realities.
Regulations: A regulation is an organization’s way of specifying that some particular standard
must be adhered to. These are rules normally promulgated by the government.
Standards: A standard is a multi-party agreement for establishment of an arbitrary criterion for
reference. Alternatively, a standard is a prescribed set of rules, conditions or requirements concernedwith the definition of terms, classification of components, delineation of procedures, specificationsof materials, performance, design or operations, measurement of quality and quality in describingmaterials, products, systems, services or practice. Standards exist that address systems (protectionof the electrical power distribution system from faults), individuals (measures to reduce potential
electric shock hazards) and protection of the environment (disposal of medical waste).

Fundamentals of Medical Instrumentation 29
Codes: A system of principles or regulations or a systematized body of law or an accumulation
of a system of regulations and standards. The most familiar code in USA is the National ElectricCode issued by National Fire Protection Association (NFPA). In India, it is the National Electric
Code issued by the Bureau of Indian Standards. In general, a code is a compilation of standards
relating to a particular area of concern. For example; a state government health codes containstandards relating to providing health care to the state population.
Specifications: Documents used to control the procurement of equipment by laying down the
performance and other associated criteria. These documents usually cover design criteria, system
performance, materials and technical data.
Standards, codes and regulations may or may not have legal implications depending upon
whether the promulgating organization is government or private.
/G31/G2E/G38/G2E/G31 /G54/G79/G70/G65/G73/G20/G6F/G66/G20/G53/G74/G61/G6E/G64/G61/G72/G64/G73
There are in general three type of standards for medical devices:
Voluntary Standards: Developed through a consensus process where manufacturers, users,
consumers and government agencies participate. They carry no inherent power of enforcement
but provide a reference point of mutual understanding.
Mandatory Standards: Required to be followed under law. They are incumbent on those to whom
the standard is addressed and enforceable by the authority having jurisdiction.
Proprietary Standards: Developed either by a manufacturer for its own internal use or by a trade
association for use by its members. They can be adopted as voluntary or mandatory standardswith the consensus/approval of the concerned agencies.
/G31/G2E/G38/G2E/G32 /G52/G65/G67/G75/G6C/G61/G74/G6F/G72/G79/G20/G52/G65/G71/G75/G69/G72/G65/G6D/G65/G6E/G74/G73
In 1976, the US Congress approved Medical Device Amendments to the Federal Food, Drugand Cosmetic Act which empowered the Food and Drug Administration (FDA) to regulatenearly every facet of the manufacture and sale of medical and diagnostic devices. The term“Medical Device” in this law means “any item promoted for a medical purpose that does not relyon chemical action to achieve its intended effect”. Further amendments to the Act have beenmade with the primary purpose to ensure the safety and efficacy of new medical devices prior tomarketing of the devices. This is accomplished by classifying the devices into three classes based
on the principle that devices that pose greater potential hazards should be subject to more
regulatory requirements.
/G43/G6C/G61/G73/G73/G96/G49
General Controls: A device for which the controls authorized by law are sufficient to provide
reasonable assurance of the safety and effectiveness of the device. Manufacturers are required toperform registration, pre-marketing notification, record keeping, labeling, reporting of adverseexperiences and good manufacturing practices. Obviously, these controls apply to all three classes.

30 Handbook of Biomedical Instrumentation
/G43/G6C/G61/G73/G73/G96/G49/G49
Performance Standards: Apply to devices for which general controls alone do not provide
reasonable assurance of safety and efficacy, and for which existing information is sufficient toestablish a performance standard that provides this assurance. However, until performancestandards are developed by regulation, only general controls apply.
/G43/G6C/G61/G73/G73/G96/G49/G49/G49
Pre-market Approval: Apply to devices which are used to support or sustain human life or to
prevent impairment of human health, devices implanted in the body and devices which present apotentially un-reasonable risk of illness or injury. These are highly regulated devices and requiremanufacturers to prove their safety and effectiveness prior to their market release.
It may be of interest to note that software which is being increasingly used in medical devices
has become an area of utmost importance because several serious accidents have been traced tosoftware bugs and problems. In view of this, there is an increased requirement for maintainingtraceability of devices to the ultimate customer, post marketing surveillance for life-sustaining andlife-supporting implants, and hospital reporting system for adverse incidents.
New medical devices can be introduced in the market by two path ways. For devices that
perform a new function or operate on a new principle, the FDA requires premarket approval. Fora device that duplicates the function of a device already in the market and if the device issubstantially equivalent to the existing device, the approval is granted. Such type of regulatory
requirements exist in some other countries also.
/G31/G2E/G38/G2E/G33 /G53/G74/G61/G6E/G64/G61/G72/G64/G73/G20/G52/G65/G6C/G61/G74/G65/G64/G20/G41/G67/G65/G6E/G63/G69/G65/G73
Most countries of the world have their own internal agencies to set and enforce standards. For
example, in India, the agency responsible for laying down standards for various products andservices is Bureau of Indian Standards (BIS). However, in the present world of internationalcooperation and trade, it has become necessary to adopt uniform standards which could beapplicable across national boundaries. There are two organisations at the international level
which are active in this area.
International Electro-technical Commission (IEC): Deals with all matters relating to standards for
electrical and electronic items. Membership in the IEC is held by a national committee for eachnation. One of the notable standards developed under IEC is 60601–1, Safety of Medical Electrical
Equipment, Part–I: General Requirements for Safety (1988) and its Amendment (1991) and the
document 60601-1-1, Safety Requirements for Medical Electrical Systems.
International Organization for Standardization (ISO): This organization oversees aspects of device
standards other than those related to electro-technology. The purpose of the ISO is to facilitate
international exchange of goods and services and to develop mutual cooperation in intellectual,
scientific, technological and economic ability.
In addition, many agencies promulgate regulations and standards in the areas of electrical
safety, fire safety, technology management, occupational safety, radiology, nuclear medicine,clinical laboratories, bio-safety, infection control, anaesthesia equipment, power distribution

Fundamentals of Medical Instrumentation 31
and medical gas systems. There are thousands of applicable standards, clinical practice
guidelines, laws and regulations. In addition, voluntary standards are issued by a large numberof organizations and mandatory standards by numerous government agencies. Biomedical
Engineers are therefore, advised to consult the relevant international/national standards for
effective discharge of their professional duties.

32 Handbook of Biomedical Instrumentation
/G42/G69/G6F/G65/G6C/G65/G63/G74/G72/G69/G63/G20/G53/G69/G67/G6E/G61/G6C/G73/G20/G61/G6E/G64/G20/G45/G6C/G65/G63/G74/G72/G6F/G64/G65/G73
/G20/G32/G2E/G31 /G4F/G52/G49/G47/G49/G4E/G20/G4F/G46/G20/G42/G49/G4F/G45/G4C/G45/G43/G54/G52/G49/G43/G20/G53/G49/G47/G4E/G41/G4C/G53
The association of electricity with medical science dates back to the 18th century when Galvani
demonstrated that most of the physiological processes were accompanied with electrical changes.This discovery formed the basis of the explanation of the action of living tissues in terms ofbioelectric potentials. It is now well established that the human body, which is composed of livingtissues, can be considered as a power station generating multiple electrical signals with two
internal sources, namely muscles and nerves. Normal muscular contraction is associated with the
migration of ions which generates potential differences measurable with suitably placed electrodes.For example, the heart and the brain produce characteristic patterns of voltage variations whichwhen recorded and analyzed are useful in both clinical practice and research. Potential differencesare also generated by the electrochemical changes accompanied with the conduction of signalsalong the nerves to or from the brain. These signals are of the order of a few microvolts and give riseto a complicated pattern of electrical activity when recorded. The fact that the activity of the livingtissues is due to the potential changes in them suggested the use of external electricity for thediagnosis of certain diseases affecting muscles and nerves, for the augmentation or replacement ofa deficient natural activity or for the restoration of a palsied muscle.
Bioelectric potentials are generated at a cellular level and the source of these potentials is ionic
in nature. A cell consists of an ionic conductor separated from the outside environment by asemipermeable membrane which acts as a selective ionic filter to the ions. This means that someions can pass through the membrane freely where as others cannot do so. All living matter is
composed of cells of different types. Human cells may vary from 1 micron to 100 microns in
diameter, from 1 mm to 1 m in length, and have a typical membrane thickness of 0.01 micron (PeterStrong, 1973).
Surrounding the cells of the body are body fluids, which are ionic and which provide a
conducting medium for electric potentials. The principal ions involved with the phenomena ofproducing cell potentials are sodium (Na
+), potassium (K+) and chloride (Cl–). The membrane of
excitable cells readily permits the entry of K+ and Cl–but impedes the flow of Na+ even though there
may be a very high concentration gradiant of sodium across the cell membrane. This results in theHAPTER
22

Bioelectric Signals and Electrodes 33
concentration of the sodium ion more on the outside of the cell membrane than on the inside. Since
sodium is a positive ion, in its resting state, a cell has a negative charge along the inner surface ofits membrane and a positive charge along the outer portion. The unequal charge distribution is a
result of certain electrochemical reactions and processes occurring within the living cell and
the potential measured is called the resting potential. The cell in such a condition is said to bepolarized. A decrease in this resting membrane potential difference is called depolarization.
The distribution of positively charged ions on the outer surface and negatively charged ions
inside the cell membrane results in the difference of potential across it and the cell becomes,in effect, a tiny biological battery. Experiments have shown that the internal resting potentialwithin a cell is approximately –90 mV with reference to the outside of the cell. When the cell isexcited or stimulated, the outer side of the cell membrane becomes momentarily negative withrespect to the interior. This process is called depolarization and the cell potential changes toapproximately +20 mV. Repolarization then takes place a short time later when the cell regains itsnormal state in which the inside of the membrane is again negative with respect to the outside.Repolarization is necessary in order to re-establish the resting potential. This discharging andrecharging of the cell produces the voltage waveforms which can be recorded by suitable methodsusing microelectrodes. A typical cell potential waveform so recorded is shown in Fig. 2.1.
–50 mV
–60 mV Threshold
–90 mV Polarized cell
resting potential0 mV Baseline+20 mVCurrent
stimulus
Minimum width of
current stimulus for
action potential
generation
2 msDepolarization Repolarization
Fig.2.1 A typical cell potential waveform

34 Handbook of Biomedical Instrumentation
The wave of excitation while propagating in the muscle causes its contraction. The contraction
wave always follows the excitation wave because of its lower velocity. This phenomenon is foundwith the skeletal muscles, the heart muscle and the smooth muscles. In its turn, every contraction
(movement) of a muscle results in the production of an electric voltage. This voltage occurs in the
muscle in such a way that the moving muscle section is always negative with respect to itssurroundings. These voltages are called action potentials because they are generated by the actionof the muscles. After complete contraction, repolarization takes place resulting in the relaxation ofthe muscle and its returning to the original state. Figure 2.2 shows electrical activity associatedwith one contraction in a muscle.
++ ++
++ +++++++++++
++
++Resting state
(polarized)
Depolarization completed++
++
+
++++
++++
++++ ––––––
––––
––––––
––– – – –– – – ––––– –– ––
––
––
––
––
+
+++
++
++
++
++–
–– –– ––
––
––
––
–UnexcitedExcitedStimulation
Depolarization
startedRepolarizationRepolarization
completed
Fig.2.2 Electrical activity associated with one contraction in a muscle
The currents involved in bioelectricity are unlike the currents involved in electronics. Bioelectric
currents are due to positive and negative ion movement within a conductive fluid. The ions possessfinite mass and encounter resistance to movement within the fluid for they have limited speeds.The cell action potential, therefore, shows a finite rise time and fall time. It may be noted that a cellmay be caused to depolarize and then repolarize by subjecting the cell membrane to an ioniccurrent. However, unless a stimulus above a certain minimum value is applied, the cell will not bedepolarized and no action potential is generated. This value is known as the stimulus threshold.After a cell is stimulated, a finite period of time is required for the cell to return to its pre-stimulus
state. This is because the energy associated with the action potential is developed from metabolic
processes within the cell which take time for completion. This period is known as refractory period.
The bioelectric signals of clinical interest, which are often recorded, are produced by the
coordinated activity of large groups of cells. In this type of synchronized excitation of many cells,

Bioelectric Signals and Electrodes 35
the charges tend to migrate through the body fluids towards the still unexcited cell areas. Such
charge migration constitutes an electric current and hence sets up potential differences betweenvarious portions of the body, including its outer surface. Such potential differences can be
conveniently picked up by placing conducting plates (electrodes) at any two points on the surface
of the body and measured with the help of a sensitive instrument. These potentials are highlysignificant for diagnosis and therapy. The primary characteristics of typical bioelectric signals aregiven in Table 2.1.
/G95/G20/G54/G61/G62/G6C/G65/G20/G32/G2E/G31 Bioelectric Signals
Parameter Primary signal characteristics Type of Electrode
Electrocardiography Frequency range: 0.05 to 120 Hz Skin electrodes
(ECG) Signal amplitude: 0.1 to 5 mV
Typical signal: 1 mV
Electroencephalo- Frequency range: 0.1 to 100 Hz Scalp electrodes
graphy (EEG) Signal amplitude: 2 to 200 mV
Typical signal: 50 mV
Electromyography Frequency range: 5 to 2000 Hz Needle electrodes
(EMG) Signal amplitude: 0.1 to 5 mV
Electroretinography Frequency range: dc to 20 Hz Contact electrodes
(ERG) Signal amplitude: 0.5 mV to 1 mV
Typical signal: 0.5 mV
Electro-oculography Frequency range: dc to 100 Hz Contact electrodes
(EOG) Signal amplitude: 10 to 3500 mV
Typical signal: 0.5 mV
/G32/G2E/G31/G2E/G31 /G45/G6C/G65/G63/G74/G72/G6F/G63/G61/G72/G64/G69/G6F/G67/G72/G61/G6D/G20/G28/G45/G43/G47/G29
The recording of the electrical activity associated with the functioning of the heart is known as
electrocardiogram. ECG is a quasi-periodical, rhythmically repeating signal synchronized by thefunction of the heart, which acts as a generator of bioelectric events. This generated signal can bedescribed by means of a simple electric dipole (pole consisting of a positive and negative pair ofcharge). The dipole generates a field vector, changing nearly periodically in time and space andits effects are measured on the surface. The waveforms thus recorded have been standardized interms of amplitude and phase relationships and any deviation from this would reflect the presenceof an abnormality. Therefore, it is important to understand the electrical activity and the associatedmechanical sequences performed by the heart in providing the driving force for the circulation ofblood.
The heart has its own system for generating and conducting action potentials through a
complex change of ionic concentration across the cell membrane. Located in the top right atriumnear the entry of the vena cava, are a group of cells known as the sino-atrial node (SA node) that
initiate the heart activity and act as the primary pace maker of the heart (Fig. 2.3). The SA node is
25 to 30 mm in length and 2 to 5 mm thick. It generates impulses at the normal rate of the heart,

36 Handbook of Biomedical Instrumentation
about 72 beats per minute at rest. Because the body acts as a purely resistive medium, the potential
field generated by the SA node extends to the other parts of the heart. The wave propagatesthrough the right and left atria at a velocity of about 1 m/s. About 0.1 s are required for theexcitation of the atria to be completed. The action potential contracts the atrial muscle and theimpulse spreads through the atrial wall about 0.04s to the AV (atrio-ventricular) node. This nodeis located in the lower part of the wall between the two atria.
The AV node delays the spread of excitation for about 0.12 s, due to the presence of a fibrous
barrier of non-excitable cells that effectively prevent its propagation from continuing beyond thelimits of the atria. Then, a special conduction system, known as the bundle of His (pronounced as
hiss) carries the action potential to the ventricles. The atria and ventricles are thus functionally
linked only by the AV node and the conduction system. The AV node delay ensures that the atriacomplete their contraction before there is any ventricular contraction. The impulse leaves the AVnode via the bundle of His. The fibres in this bundle, known as Purkinje fibres, after a shortdistance split into two branches to initiate action potentials simultaneously in the two ventricles.
Conduction velocity in the Purkinje fibres is about 1.5 to 2.5 m/s. Since the direction of the
impulse propagating in the bundle of His is from the apex of the heart, ventricular contractionbegins at the apex and proceeds upward through the ventricular walls. This results in thecontraction of the ventricles producing a squeezing action which forces the blood out of theventricles into the arterial system. Figure 2.3 shows the time for action potential to propagate tovarious areas of the heart.A-V
nodeS-A
node.0
.025
0.19
.14.2
0.18
0.15
0.1550.145
0.16.17.12.045.03
RA
RV
LVLA.06
Right bundle
branchLeft bundle
branchPurkinje
networkPurkinje
network
MyocardiumBundle
of His
Fig.2.3 The position of the sino-atrial node in the heart from where the impulse re-
sponsible for the electrical activity of the heart originates. The arrow showsthe path of the impulse.
Note: The numbers like 0.18, 0.145, 0.15, 0.2 … etc. indicate the time taken for
the impulse to travel from the S-A node to various parts of the heart

Bioelectric Signals and Electrodes 37
The normal wave pattern of the electrocardiogram is shown in Fig. 2.4. The PRandPQinterval,
measured from the beginning of the Pwave to the onset of the R or Q wave respectively, marks the
time which an impulse leaving the SA node takes to reach the ventricles. The PRinterval normally
lies between 0.12 to 0.2 s. The QRS interval, which represents the time taken by the heart impulse
to travel first through the interventricular system and then through the free walls of the ventricles,normally varies from 0.05 to 0.10s.
Rwave
PwaveTwave
TPinterval
(ventricular section)QTintervalPQintervalPQ
intervalST
intervalQRS
complex
Heart cyclewave waveUwave Pwave S Q
(Auricular
section)Condition time(intermediate
interval)
Fig.2.4 Normal wave pattern of an ECG waveform recorded in the standard
lead position
TheT wave represents repolarization of both ventricles. The QT interval, therefore, is the period
for one complete ventricular contraction (systole). Ventricular diastole, starting from the end of theT wave extends to the beginning of the next Qwave. Typical amplitude of QRS is 1 mV for a normal
human heart, when recorded in lead 1 position.
/G32/G2E/G31/G2E/G32 /G45/G6C/G65/G63/G74/G72/G6F/G65/G6E/G63/G65/G70/G68/G61/G6C/G6F/G67/G72/G61/G6D/G20/G28/G45/G45/G47/G29
The brain generates rhythmical potentials which originate in the individual neurons of the brain.
These potentials get summated as millions of cell discharge synchronously and appear as a surfacewaveform, the recording of which is known as the electroencephalogram (Fig. 2.5).
The neurons, like the other cells of the body, are electrically polarized at rest. The interior of the
neuron is at a potential of about –70 mV relative to the exterior. When a neuron is exposed to astimulus above a certain threshold, a nerve impulse, seen as a change in membrane potential, isgenerated which spreads in the cell resulting in the depolarization of the cell. Shortly afterwards,
repolarization occurs.

38 Handbook of Biomedical Instrumentation
Fig.2.5 Typical EEG signal waveform
The EEG signal can be picked up with electrodes either from the scalp or directly from the
cerebral cortex. The peak-to-peak amplitude of the waves that can be picked up from the scalp isnormally 100 mV or less while that on the exposed brain, is about 1 mV. The frequency varies
greatly with different behavioural states. The normal EEG frequency content ranges from 0.5 to
50 Hz. The nature of the wave varies over the different parts of the scalp.
The variations in EEG signals both in terms of amplitude and frequency are of diagnostic value.
Frequency information is particularly significant since the basic frequency of the EEG range isclassified into the following five bands for purposes of EEG analysis:
Delta ( d) 0.5–4 Hz
Theta ( q) 4–8 Hz
Alpha ( a) 8–13 Hz
Beta ( b) 13–22 Hz
Gamma ( g) 22–30 Hz
The alpha rhythm is one of the principal components of the EEG and is an indicator of the state
of ‘alertness’ of the brain. It serves as an indicator of the depth of anaesthesia in the operating
room. The frequency of the EEG seems to be affected by the mental activity of a person. The wide
variation among individuals and the lack of repeatability in a given person from one occasion toanother makes the analysis a difficult proposition. However, certain characteristic EEG waveformscan be conveniently related to gross abnormalities like epileptic seizures and sleep disorders.
Besides the importance of the frequency content of the EEG pattern, phase relationships between
similar EEG patterns from different parts of the brain are also being studied with great interest inorder to obtain additional knowledge regarding the functioning of the brain. Another importantmeasurement is the recording of ‘evoked response’, which indicates the disturbance in the EEGpattern resulting from external stimuli. The stimuli could be a flash of light or a click of sound.Since the responses to the stimuli are repeatable, the evoked response can be distinguishedfrom the rest of the EEG activity by averaging techniques to obtain useful information about thefunctioning of particular parts of the brain.
/G32/G2E/G31/G2E/G33 /G45/G6C/G65/G63/G74/G72/G6F/G6D/G79/G6F/G67/G72/G61/G6D/G20/G28/G45/G4D/G47/G29
The contraction of the skeletal muscle results in the generation of action potentials in the individualmuscle fibres, a record of which is known as electromyogram. The activity is similar to thatobserved in the cardiac muscle, but in the skeletal muscle, repolarization takes place much more

Bioelectric Signals and Electrodes 39
rapidly, the action potential lasting only a few milliseconds. Since most EMG measurements are
made to obtain an indication of the amount of activity of a given muscle, or a group of muscles,rather than of an individual muscle fibre, the EMG pattern is usually a summation of the individual
action potentials from the fibres constituting the muscle or muscles being studied. The electrical
activity of the underlying muscle mass can be observed by means of surface electrodes on the skin.However, it is usually preferred to record the action potentials from individual motor units forbetter diagnostic information using needle electrodes.
In voluntary contraction of the skeletal muscle, the muscle potentials range from 50 mV to 5 mV
and the duration from 2 to 15 ms. The values vary with the anatomic position of the muscle and thesize and location of the electrode. In a relaxed muscle, there are normally no action potentials. Atypical EMG signal is shown in Fig. 2.6.
Fig.2.6 Waveshape of a typical EMG signal
/G20/G32/G2E/G32 /G52/G45/G43/G4F/G52/G44/G49/G4E/G47/G20/G45/G4C/G45/G43/G54/G52/G4F/G44/G45/G53
Bioelectric events have to be picked up from the surface of the body before they can be put into the
amplifier for subsequent record or display. This is done by using electrodes. Electrodes make atransfer from the ionic conduction in the tissue to the electronic conduction which is necessary formaking measurements. Electrodes are also required when physiological parameters are measuredby the impedance method and when irritable tissues are to be stimulated in electrotherapy. Twotypes of electrodes are used in practice-surface electrodes and the deep-seated electrodes. The
surface electrodes pick up the potential difference from the tissue surface when placed over it
without damaging the live tissue, whereas the deep-seated electrodes indicate the electric potentialdifference arising inside the live tissue or cell. The same classification can be applied to electrodesused for the stimulation of muscles.

40 Handbook of Biomedical Instrumentation
Electrodes play an important part in the satisfactory recording of bioelectric signals and their
choice requires careful consideration. They should be comfortable for the patients to wear overlong periods and should not produce any artefacts. Another desirable factor is the convenience of
application of the electrodes.
/G32/G2E/G32/G2E/G31 /G45/G6C/G65/G63/G74/G72/G6F/G64/G65/G2D/G54/G69/G73/G73/G75/G65/G20/G49/G6E/G74/G65/G72/G66/G61/G63/G65
The most commonly used electrodes in patient monitoring and related studies are surface
electrodes. The notable examples are when they are used for recording ECG, EEG and respiratoryactivity by impedance pneumography. In order to avoid movement artefacts and to obtain a clearlyestablished contact (low contact impedance) an electrolyte or electrode paste is usually employedas an interface between the electrode and the surface of the source of the event. Figure 2.7 (a, b)
represent the electrode-tissue interface.
Electrolyte-skin
interfaceElectrolyte TissueTo
instrumentMetal-electrolyte
interface
Fig.2.7(a) Electrode-tissue interface for surface electrodes used with electrode jelly
The characteristic of a surface electrode
composed of a metal electrode and attached tothe surface of the body through an electrolyte(electrode jelly) are dependent upon the condi-tions at the metal-electrolyte interface, the
electrolyte-skin interface and the quality of the
electrolyte.
Metal-Electrolyte Interface: At the metal-electrolyte
transition, there is a tendency for each electrode
to discharge ions into the solution and for ions in
the electrolyte to combine with each electrode.The net result is the creation of a charge gradient(difference of potential) at each electrode, thespatial arrangement of which is called theelectrical double layer (Fig. 2.7(c)). The doublelayer is known to be present in the regionimmediately adjacent to the electrode and can berepresented, in its simplest form, as two parallelsheets of charge of opposite sign separated bya thin film of dielectric. Therefore, the metal-+
+
+
+++
+
+
++–
–
––
––
–
––
–Electrolyte ElectrodeHelmholtz double layer
Fig.2.7(b) Electrode tissue interface
circuit involves transferof electrons from the metalphase to an ionic carrierin the e lectrolyte, a charge
double layer (capacitance)forms at the interface

Bioelectric Signals and Electrodes 41
electrolyte interface appears to consist of a voltage source in series with a parallel combination of
a capacitance and reaction resistance. The voltage developed is called the half-cell potential.
To a first-order approximation, the half-cell potential is equal to the electrode potential of the
metal, if the electrodes were used in a chemical measuring application. All electrode potentials aremeasured with respect to a reference electrode, usually that of hydrogen absorbed on platinumblack. This is an inconvenient electrode to make and, therefore, other alternative electrodes whichmay have fairly stable and repeatable potential (e.g. calomel electrode) are employed. Electrodepotentials of some of the commonly used metals in the electrochemical series with respect to
Hydrogen are given in Table 2.2.
/G95/G20/G54/G61/G62/G6C/G65/G20/G32/G2E/G32 Electrode Potentials of some Metals with Respect to Hydrogen
Metal Ionic symbol Electrode potential
Aluminium AI+++–1.66 V
Iron Fe++–0.44 V
Lead Pb++–0.12 V
Hydrogen H+ 0
Copper C+++0.34 V
Silver Ag+ +0.80 V
Platinum Pt+ + 1.2 V
Gold Au+ + 1.69 V
This table shows that the electrode potentials are appreciable when dissimilar metals are used.
They also exist, though of smaller magnitude, even if electrodes of similar materials are employed.The lowest potential has been observed to be in the silver-silver chloride electrodes. The values ofthe capacitance and the resistance depend upon many factors which include the current density,temperature, type and concentration of the electrolyte and the type of metal used.
The difference in half-cell potentials that exists between two electrodes is also called ‘offset
potential’. The differential amplifiers used to measure potentials between two electrodes are+
++
+
+++
++
+
++
+
++++
+
+
+++–
–
–
––––
––
–
––
–
–
–
–
ElectrodeElectrodeCharge layer
TerminalElectrolyte
RCE
(i) (ii)
Fig.2.7(c) (i) Charge distribution at electrode-electrolyte interface
(ii)Three components representing the interface

42 Handbook of Biomedical Instrumentation
generally designed to cancel the electrode offset potential so that only the signals of interest
are recorded. The electrode offset potential produced between electrodes may be unstable andunpredictable. The long-term change in this potential appears as baseline drift and short-term
changes as noise on the recorded trace. If electrodes are used with ac-coupled amplifiers, the long
term drift may be partially rejected by the low frequency characteristics of the amplifier. But it willdepend upon the rate of change of electrode offset potential in relation to the ac-coupling timeconstant in the amplifier. For example, if the electrode offset potential drift rate is 1 mV/s,satisfactory results can only be obtained if the low frequency response of the amplifier is 1 Hz.
Also, the absolute value of the electrode offset potential is rarely significant except when it may
exceed the maximum dc differential offset of the amplifier. In such a case, the trace may go out ofthe monitor screen or the pen in a recording instrument shifts to the extreme end of the chart paper,and then it will not be possible to bring them back. Silver-silver chloride electrodes have beenfound to give almost noise free characteristics. They are also found to be acceptable from the pointof view of long-term drift. Electrodes made of stainless steel are generally not acceptable for highsensitivity physiological recordings. This is because stainless steel electrodes in contact with asaline electrolyte produce a potential difference of 10 mV between the electrodes, whereas thisvalue is 2.5 mV for silver-silver chloride electrodes. Some representative values of potential between
electrodes in electrolytes are given in Table 2.3. Staewen (1982) discusses various aspects
concerning dc offset voltage standard for pregelled ECG disposable electrodes.
/G95/G20/G54/G61/G62/G6C/G65/G20/G32/G2E/G33 Potential between Electrolytes in Electrodes (Courtesy: Geddes and Baker 1975)
Electrode metal Electrolyte Potential difference
between electrodes
Stainless steel Saline 10 mV
Silver Saline 94 mV
Silver-silver chloride Saline 2.5 mV
Silver-silver chloride(11 mm disc) ECG paste 0.47 mV
Silver-silver chloride (sponge) ECG paste 0.2 mV
Warburg (1899) in his pioneering studies discovered that a single electrode/electrolyte interface
can be represented by a series capacitance Cand resistance R as shown in Fig. 2.8(a). However, C
and Rare unlike real capacitors and resistors because their values are frequency and current-
density dependent. Often, these components are called the polarization capacitance and resistance.Warburg found that, for low current density, the reactance X of C (1/2 pfC) equals R; both varied
almost inversely as the square root of frequency, i.e. R= X = k/
f, where kis a constant. The
consequence of this relationship is that the phase angle q is constant at p/4 for all frequencies.
However, only a limited number of studies have tested the accuracy of the Warburg model (Ragheb
and Geddes, 1990).
It has been observed that the Warburg series RCequivalent does not adequately represent the
behaviour of an electrode/electrolyte interface as this equivalent does not truly account for thevery low-frequency behaviour of the interface. It is well known that such an interface can passdirect current. Therefore, a resistance R
fplaced in parallel with the Warburg equivalent is more

Bioelectric Signals and Electrodes 43
appropriate. Figure 2.8(b) shows this equivalent circuit in which Rf represents the Faradic leakage
resistance. The value of Rf is high in the low-frequency region and is dependent on current density,
increasing with a increase in current density.
To complete the equivalent circuit of an electrode/electrolyte interface, it is necessary to add the
half-cell potential E. This is the potential developed at the electrode/electrolyte interface. The
value of Edepends on the species of metal and the type of electrolyte, its concentration and
temperature. Figure 2.8(c) illustrates the complete equivalent circuit of a single electrode/electrolyteinterface.
Electrolyte-Skin Interface: An approximation of the electrolyte-skin interface can be had by
assuming that the skin acts as a diaphragm arranged between two solutions (electrolyte and bodyfluids) of different concentrations containing the same ions, which is bound to give potentialdifferences. The simplest equivalent representation could then be described as a voltage source in
series with a parallel combination of a capacitance and resistance. The capacitance represents the
charge developed at the phase boundary whereas the resistance depends upon the conditionsassociated with ion-migration along the phase boundaries and inside the diaphragm.
The above discussion shows that there is a possibility of the presence of voltages of non-
physiological origin. These voltages are called contact potentials.
The electrical equivalent circuit of the surface electrode suggests that the voltage presented to
the measuring instrument from the electrode consists of two main components. One is the contactpotential and the other is the biological signal of interest. The contact potential depends uponseveral factors and may produce an interference signal which exceeds several times the usefulsignal. The contact potential is found to be a function of the type of skin, skin preparation andcomposition of the electrolyte.
When bioelectric events are recorded, interference signals are produced by the potential
differences of metal-electrolyte and the electrolyte-skin interface. Normally, these potentialdifferences are connected in opposition during the recording procedure, and in the case of atruly reversible and uniform electrode pair, their difference would be nil. However, in practice, aR
RR
C
CEC
R1
R1Electrode
terminal
Electrode
terminalElectrode
terminal
(b)
(c)(a)
Electrolyte
ElectrolyteElectrolyte
Fig.2.8 (a) Warburg equivalent for an electrode-electrolyte interface
(b)Addition of the faradic leakage resistance Rf to account for the direct
current properties
(c)Half-cell potential E of the electrode-electrolyte interface

44 Handbook of Biomedical Instrumentation
difference of potential—may be extremely small—is found to exist between electrodes produced
even under conditions of utmost care during manufacture. Also, some of the elements in theequivalent circuit are time-dependent and are bound to show slow variations with time.
The main reason for this rate of change is due to a relative displacement affecting chiefly the
potential of the metal-electrolyte transition. Other factors responsible for variations of potential
difference with time can possibly be temperature variations, relative displacement of the
components in the system and changes in the electrolyte concentration, etc. (Odman and Oberg,1982).
If ac signals are to be recorded, the potential difference between the two electrodes will not
interfere with the useful signals, provided that the contact potential difference between theelectrodes is constant. However, if the rate of change with time of the contact potential falls withinthe frequency spectrum of the signal under test, an error will be produced. The problem of differenceof contact potentials becomes serious in case dc signals such as EOG are to be recorded. Anyvariation in the contact potential would greatly alter the character of the signal to be recordedwhich may itself be of extremely low amplitude—of the order of a few microvolts.
Based on the above mentioned considerations, it is possible to construct the circuit in which a
pair of electrodes is placed in electrolytic contact with a subject. The electrodes are used to measurea bioelectric event and are connected to a differential amplifier. Three potentials are found to existin this circuit (Fig. 2.9), one is due to the bioelectric event ( E
b) and the other two are non-physiologic
Rf1
Rf2R1
R2Z2Z1
RC1
C2E1
E21
2EbRtCtElectrode 1
Electrode 2Body tissue
and
fluidsRR12and : Resistance
EE12and : Half-cell potential
CC12and : Capacitance
ZZ12and : Skin contact impedance
Rt: Tissue resistance
Ct: Tissue capacitance
RRff12and : Faradic leakage resistance
Eb: Bioelectic event
Fig.2.9 Equivalent circuit for a pair of electrodes (1,2) on a subject represented by
R Rt, Ct. Embedded in the subject is a bioelectric generator Eb (after Tacker and
Geddes, 1996)

Bioelectric Signals and Electrodes 45
and represent the half-cell potentials ( E1 and E2) of the electrodes. Z1 and Z2 are the skin contact
impedances of these electrodes and R is the tissue resistance or resistance of the bioelectric
generator. This circuit shows that the impedance of the electrodes would be high in the low-
frequency region and it would decrease with increasing frequency. It is further clear that in the
measurement of a bioelectric signal, it is essential to minimize potential drops across the electrodeimpedance. This is achieved by making the skin-contact impedance as low as possible and makingthe input impedance of the measuring device as high as possible.
/G32/G2E/G32/G2E/G32 /G50/G6F/G6C/G61/G72/G69/G7A/G61/G74/G69/G6F/G6E
If a low voltage is applied to two electrodes placed in a solution, the electrical double layers aredisturbed. Depending on the metals constituting the electrodes, a steady flow of current may or
may not take place. In some metal/liquid interfaces, the electrical double layer gets temporarily
disturbed by the externally applied voltage, and therefore, a very small current flows after the firstsurge, thus indicating a high resistance. This type of electrode will not permit the measurement ofsteady or slowly varying potentials in the tissues. They are said to, be polarized or nonreversible.Thus, the phenomenon of polarization affects the electro-chemical double layer on the electrodesurface and manifests itself in changing the value of the impedance and voltage sourcerepresenting the transition layer. Parsons (1964) stated that electrodes in which no net transfer ofcharge takes place across the metal-electrolyte interface can be termed as perfectly polarized.Those in which unhindered exchange of charge is possible are called non-polarizable or reversibleelectrodes. The ionic double layer in metals of these electrodes is such that they allow considerablecurrent to flow when a small voltage is applied, thus offering a low resistance.
Although polarizable electrodes are becoming less common, they are still in use. They usually
employ stainless steel and are used for resting ECGs or other situations where there is smalllikelihood that the electrodes would be exposed to a large pulse of energy (such as a defibrillation
discharge) in which case they would retain a residual charge, become polarized, and will no
longer transmit the relatively small bioelectric signals, thus becoming useless.
Non-polarizing electrodes on the other hand, are designed to rapidly dissipate any charge
imbalance induced by powerful electrical discharges such as a defibrillation procedure. Rapiddepolarization enables the immediate reappearance of bioelectric signals on the monitor afterdefibrillation. For this reason, non-polarizing electrodes have become the electrodes of choice formonitoring in the intensive care units and stress testing procedures. Historically, these electrodesemploy a conducting metal with a silver/silver-chloride (Ag/AgCl) surface in contact with theconducting gel.
The choice of metals for electrodes is not determined only by their susceptibility to polarization,
but other factors such as mechanical properties, skin irritation or skin staining, etc. have also to betaken into consideration. A detailed comprehensive review of electrodes for measurement ofbioelectronic events is given by Geddes and Baker (1975).
The LeeTec Corporation, USA has devised a tin-backed electrode, Tracets MP-3000 (Fig. 2.10)
which is non-polarizable and performs electrically as well as or better than similar electrodesemploying silver/silver-chloride (Montecalvo and Rolf, 1990). U.S. Patent 4,674,512 describes theconstruction of this non-polarizing ECG electrode, which employs no silver. This represents a

46 Handbook of Biomedical Instrumentation
new era for electrocardiology where silver is no longer a critical electrode component for quality
performance.
/G32/G2E/G32/G2E/G33 /G53/G6B/G69/G6E/G20/G43/G6F/G6E/G74/G61/G63/G74/G20/G49/G6D/G70/G65/G64/G61/G6E/G63/G65
The bioelectrical events are usually recorded by means of metallic electrodes placed on the surface
of the body. The electrical activity generated by various muscles and nerves within the body isconducted to the electrode sites through the body tissues, reaches the electrodes through the skin-electrode transition and is then conducted by direct wire connection to the input circuit of therecording machine. The impedance at the electrode-skin junction comes in the overall circuitry of
the recording machine and, therefore, has significant effect on the final record. Skin electrode
impedance is known as the contact impedance and is of a value much greater than the electricalimpedance of the body tissue as measured beneath the skin. The outer horny layer of the skin isresponsible for the bulk of the skin contact impedance and, therefore, a careful skin preparation isessential in order to obtain best results.
Measurement of Skin Contact Impedance: A convenient method to measure the contact impedance
at any individual electrode is shown in Fig. 2.11. This method has been suggested by Miller (1969).The three electrodes, A, B and C, have contact impedance respectively of Z
a, ZbandZc. An oscillator
provides a constant current in the frequency range of 0.1–100 Hz through the 47 k W series resistor.
By suitably positioning the switch, a sensitive oscilloscope can be used to monitor either thevoltage dropped across the 1 k W resistor or the voltage dropped across Z
b. The voltage drop across
Zb can be neglected since the input impedance of the oscilloscope used with an input probe is
usually high. From the voltage dropped across the 1 k W resistor it is possible to calculate the circuit
current and thus to obtain a value for Zc. Using this technique, the skin contact impedance of the
following types of electrodes were measured by Hill and Khandpur (1969).
• Plastic cup self-adhesive electrodes (Boter et al, 1966)
• Metal plate limb electrodes used with conducting jelly• Metal plate electrodes used with conducting plastic (Jenkner, 1967)Printer PVC
label
Tin
LecTec’s
MS-3000 conductive
adhesive
Fig.2.10 Resting ECG electrode Lec Tec MP-3000-a multipurpose monitoring and
diagnostic non-polarizable electrode

Bioelectric Signals and Electrodes 47
• Dry multi-point limb electrodes (Lewes, 1966)
• Dry multi-point suction chest electrodes• Self-adhesive multi-point chest electrodes used with conducting jelly• Self-adhesive gauze electrodes• Self-adhesive dry multi-point chest electrodes (Lewes and Hill, 1967)
Representative plots of contact impedance versus frequency are shown in Fig. 2.12.
Usually the contact impedance in respect of surface electrodes used for recording of ECG
is measured (Grimnes, 1983) at 10–20 Hz because most of the energy content of the ECG isconcentrated below 30 Hz. Geddes and Baker (1968) used a synchronous rectifier with a phase-sensitive detector to continuously measure the resistive and reactive components of the impedance.
/G32/G2E/G32/G2E/G34 /G4D/G6F/G74/G69/G6F/G6E/G20/G41/G72/G74/G65/G66/G61/G63/G74/G73
Motion artefact is a problem in biopotential measurements. The problem is greatest in cardiacstress laboratories where the exercise ECG is recorded. The problem is also serious in coronarycare units where patients are monitored for relatively long periods. Motion of the subject undermeasurement creates artefacts which may even mask the desired signal or cause an abrupt shift inthe baseline. These artefacts may result in a display being unreadable, a recording instrumentexceeding its range, a computer yielding incorrect output or a false alarm being triggered by themonitoring device. Tam and Webster (1977) concluded that the skin-electrolytic paste interface isthe major source of motion artefact. When a metal electrode contacts an electrolytic paste, a half-cell potential is generated at the electrode-paste interface. Kahn (1965) demonstrated that when
polarizable metal-plate electrodes are used, the electrode-paste interface can be a source of motionOscilloscopeAudio
oscillatorPatient’s
body
AB
SC
1 K47 KZaZbZc
Fig.2.11 Arrangement for measurement electrode skin-contact impedance for
surface electrodes

48 Handbook of Biomedical Instrumentation
artefact. When the paste is agitated, the half-cell potential varies because of the altered metallic ion
gradient at the interface. He recorded a 1 mV offset potential change from a silver-silver chlorideelectrode exposed to a flowing stream of saline solution, as contrasted to 30 mV change for some
silver electrodes.
Motion artefact is reduced to a negligible magnitude by skin abrasion. However, when the skin
is abraided, it is more susceptible to irritants. The possible sources for skin irritation include the
electrode, the paste and the adhesive. When large currents flow through metallic electrodes,migration of some ions into the skin can cause irritation. However, silver-silver chloride electrodesdo not cause much problem since silver chloride is almost insoluble in a chloride containingsolution. Therefore, when these electrodes are used, the skin irritation is mostly caused by thepaste and or the adhesive. Most commercial pastes produce about the same irritation when usedon unprepared skin. They cause itching due to restricted perspiration, and reddening of the skindirectly under the electrodes appears in 2–4 days. Thakor and Webster (1985) studied the sourcesof artefacts, means of reducing them using skin preparations, the electrode designs and theirplacement on the chest for long-term ambulatory ECG.
/G20/G32/G2E/G33 /G53/G49/G4C/G56/G45/G52/G2D/G53/G49/G4C/G56/G45/G52/G20/G43/G48/G4C/G4F/G52/G49/G44/G45/G20/G45/G4C/G45/G43/G54/G52/G4F/G44/G45/G53
One of the important desirable characteristics of the electrodes designed to pick up signals frombiological objects is that they should not polarize. This means that the electrode potential must notvary considerably even when current is passed through them. Electrodes made of silver-silverchloride have been found to yield acceptable standards of performance. By properly preparing48121620242832
0.1 1 10 100ABCDEFGH
Electrode contact impedance(kilo-ohms)
Fig.2.12 Electrode skin-contact impedance versus signal frequency for different types
of electrodes (after Hill and Khandpur, 1969)

Bioelectric Signals and Electrodes 49
and selecting the electrodes, pairs have been produced with potential differences between them
of only fractions of a millivolt (Feder, 1963). Standing voltage of not more than 0.1 mV with adrift over 30 min. of about 0.5 mV was achieved in properly selected silver-silver chloride
electrodes by Venables and Sayer (1963). Silver-silver chloride electrodes are also nontoxic and are
preferred over other electrodes like zinc-zinc sulphate, which also produce low offset potentialcharacteristics, but are highly toxic to exposed tissues. Silver-silver chloride electrodes meet thedemands of medical practice with their highly reproducible parameters and superior propertieswith regard to long-term stability.
Production of Silver-Silver Chloride Electrodes: Silver-silver chloride electrodes are normally
prepared by electrolysis. Two silver discs are suspended in a saline solution. The positive pole ofa dc supply is connected to the disc to be chlorided and the negative pole goes to the other disc. Acurrent at the rate of 1 mA/cm
2 of surface area is passed through the electrode for several minutes.
A layer of silver chloride is thus deposited on the surface of the anode. The chemical changes thattake place at the anode and cathode respectively are:
NaCl = Na
+ + Cl–
Cl–+ Ag+Æ AgCl
The positively charged sodium ions generate hydrogen when they reach the cathode surface.
2Na+ + 2H2O + 2 electrons Æ 2NaOH + H2
To prepare silver-silver chloride electrodes of good quality, only pure silver should be used and
the saline solution should be made from analar grade sodium chloride. Before chloriding, silvermust be cleaned—preferably by the electrolytic method.
Geddes et al. (1969) investigated the effect of the chloride deposit on the impedance-frequency
characteristics of the silver-silver chloride electrodes. They demonstrated that the impedance wasdifferent for different layers of chloride and that there is an optimum chloriding, which gives thelowest impedance. They concluded that the lowest electrode-electrolyte impedance in thefrequency range of 10 Hz to 10 kHz was found to occur with a chloride deposit ranging between100 and 500 mAs/cm
2 of electrode area. To achieve this deposit by manipulation of current and
time, the minimum constant chloriding current density should be 5 mA/cm2 of electrode area.
Higher values may be used with a corresponding reduction in time to achieve the 100–500
mAs/cm2 chloride deposit. With this chloride deposit, the electrode electrolyte impedance was
found to be resistive. The use of a chloride deposit in excess of this range did not alter the resistivenature of the electrode-electrolyte impedance although it increased its magnitude. Cole andKishimoto (1962), however found that the chloride deposit for achieving the lowest impedance is2000 mAs/cm
2. Geddes (1972) confirmed that an optimal coating of silver chloride applied to a
silver electrode minimizes the electrical impedance. This is supported by Getzel and Webster(1976) who concluded that silver chloride may be applied to cleaned silver electrodes in theamount of 1050–1350 mA s/cm
2 in order to reduce the impedance of the electrodes. However, to
further reduce the impedance of the electrodes, they should be coated with at least 2000 mAs/cm2
of silver chloride followed by immersion in a photographic developer for 3 minutes. A secondlayer of silver chloride, however, did not result in any further reduction in impedance. Grubbs andWorley (1983) obtained a lower and more stable impedance electrode by placing a heavier initial
chloride coat on an etched silver electrode, and then electrolytically removing a portion of that coat.

50 Handbook of Biomedical Instrumentation
/G20/G32/G2E/G34 /G45/G4C/G45/G43/G54/G52/G4F/G44/G45/G53/G20/G46/G4F/G52/G20/G45/G43/G47
/G32/G2E/G34/G2E/G31 /G4C/G69/G6D/G62/G20/G45/G6C/G65/G63/G74/G72/G6F/G64/G65/G73
The most common type of electrodes routinely used for recording ECG are rectangular or circular
surface electrodes (Fig. 2.13). The material used is german silver, nickel silver or nickel plated steel.
They are applied to the surface of the body with electrode jelly. The typical value of the contactimpedance of these electrodes, which are of normal size, is nearly 2 to 5 k W when measured at
10 Hz. The electrodes are held in position by elastic straps. They are also called limb electrodes asthey are most suitable for application on the four limbs of the body. The size of the limb electrodesis usually 3 ¥ 5 cm and they are generally made of german silver, an alloy of zinc, copper and
nickel. They are reusuable and last several years.
Arm ElectrodeRubber strap
Fig.2.13 ECG plate e lectrode. The electrode is usually fastened to the arm or leg with a
perforated rubber strap which keeps it in position during ECG recording
Limb electrodes are generally preferred for use during surgery because the patient’s limbs are
relatively immobile. Moreover, chest electrodes cannot be used as they would interfere with thesurgery.
Limb electrodes are not suitable for use in long-term patient monitoring because the long flowing
leads are inconvenient to the patient. Also, the electromyographic voltages generated by the activity
of the limb muscles makes them unsuitable for use when monitoring conscious and semi-conscious
patients.
Suction-cup electrode is commonly used to record the unipolar chest leads. It has a high contact
impedance as only the rim of the electrode is in contact with the skin. The electrode is popular forits practicality, being easily attachable to fleshy parts of the body. Electrode jelly forms the vacuumseal.
/G32/G2E/G34/G2E/G32 /G46/G6C/G6F/G61/G74/G69/G6E/G67/G20/G45/G6C/G65/G63/G74/G72/G6F/G64/G65/G73
Limb electrodes generally suffer from what is known as motion artefacts caused due to the relativemotion at the interface between the metal electrode and the adjacent layer of electrode jelly, Kahn

Bioelectric Signals and Electrodes 51
(1965) and Boter et al (1966). The interface can be stabilized by the use of floating electrodes in
which the metal electrode does not make direct contact with the skin. The electrode (Fig. 2.14)consists of a light-weight metalled screen or plate held away from the subject by a flat washer
which is connected to the skin. Floating electrodes can be recharged, i.e. the jelly in the electrodes
can be replenished if desired.
Fig.2.14 Light weight floating electrode with press stud for long-term monitoring of
ECG
Patten et al (1966) have described spray-on chest electrodes where a conducting spot is
developed on the skin by spraying a film of conducting adhesive (mixture of Duco cement, silverpowder and acetone). Connection with the instrument is established with silver-plated copperwires fixed in the conducting adhesive. The type of electrodes are extremely light-weight and donot make use of electrode jelly. This makes them ideal for use in monitoring the ECG of exercisingsubjects and aeroplane pilots as they give rise to minimal motion artefacts. The contact impedanceshown by these electrodes is of the order of 50 k W.
Completely flexible ECG electrodes for the long-term monitoring of ECG during space flight are
reported by Sandler et al (1973). These electrodes were made of silver-impregnated silastic rubber
and were found to be comfortable to wear. They were also evaluated for use during exercise orprolonged monitoring as may be necessary in an intensive care or coronary care unit.
/G32/G2E/G34/G2E/G33 /G50/G72/G65/G67/G65/G6C/G6C/G65/G64/G20/G44/G69/G73/G70/G6F/G73/G61/G62/G6C/G65/G20/G45/G6C/G65/G63/G74/G72/G6F/G64/G65/G73
Electrodes which are employed in stress testing or long term monitoring, present additionalproblems because of the severe stresses, perspiration and major body movement encountered insuch studies. Both design considerations and application techniques of electrodes used inelectrocardiography are necessary to prevent random noise on the baseline, baseline wanderingand skin contact over extended periods causing a loss of signal. To overcome problems dueto prolonged application, special disposable electrodes have been developed. Figure 2.15(a)illustrates the principle of a pregelled electrode while Fig. 2.15(b) shows a cross-section of such anelectrode. The main design feature of these electrodes which helps in reducing the possibility ofartefacts, drift and baseline wandering is the provision of a high-absorbancy buffer layer with
isotonic electrolyte. This layer absorbs the effects of movement of the electrode in relationship to

52 Handbook of Biomedical Instrumentation
the skin, and attempts to maintain the polarization associated with the half-cell potential constant.
Since perspiration is the most common cause of electrode displacement, the use of an additionalporous overlay disc resists perspiration and ensures secure placement of the electrode on the skineven under stress conditions. Figure 2.16 show a typical pregelled electrode.
Various manufacturers offer common features in pregelled electrode construction. The lead
wire’s female connector “snaps” on, allowing a convenient snap-on pull-off connection with a360 rotation providing mechanical and electrical connection. The plastic eyelet or sensor has adiameter of 0.5–1.5 cm and is electroplated with silver up to a thickness of 10 mm. The surface of the
Ag layer is partially converted to AgCl.
The tape is made from one of the adhesive coated occulusive foams made from a plastic, such as
polyethylene, or a porous backing, such as non-woven cloth. Tapes used for first aid dressings are
suitable. The electrode diameters range from 4–6 cm.
Some advantages of these electrodes as compared to plates and welch bulbs are : as there is no
risk of infection which is possible with reusable electrodes, their smaller size makes them lessprone to detachment and also less time is required per ECG procedure.Silver-silver chloride (AgCI)Silver (Ag)
++++++++++++––––––––––––Connecting wire
Electrolyte with chloride ionDouble yeat
(a)
SkinAdhesiveFlexible lipPorous
overlay ringPolyester sponge with
isotonic electrolyteSoft thermoplastic
rubber bodyElectrode snap contact
Silver-silver chloride
conductor
Isotonic electrolyte
Buffer layer with
isotonic electrolyte
(b)
Fig.2.15 (a) Principle of pre-gelled ECG electrode made of silver-silver chloride. The
electrode has electrolyte layers that are made of a firm gel which hasadhesive properties. The firm gel minimizes the disturbance of thecharge double layer
(b) Cross-section of a typical pre-gelled electrode

Bioelectric Signals and Electrodes 53
Skin-electrode adhesion is an important performance criterion. Partial electrode lift would
cause a gel dry out and intermittent contact artefacts, while premature electrode fall off wouldincrease monitoring costs. The adhesives used to secure electrodes to the skin are usually pressure-sensitive adhesives which implies that a force must be applied to achieve adhesion. The adhesiveshould have good bonding capability; internal strength so that upon removal, objectionable“stringy” residue will not remain on the skin; good temperature stability; immunity to oxidationfrom air pollutants, saline and other common solutions; resistance to water, isopropyl alcohol,saline and other common solutions used in hospitals and low potential for skin irritation.
Trimby (1976) describes an alternative design to reduce motion artefacts generated by
mechanical shocks. Ideally, the thin layer (critical layer) of electrolyte just under the metal pieceshould be cushioned from all mechanical shocks. Figure 2.17 shows the construction of fluidcolumn electrodes (HP 14245A, HP 14248A) in which the critical layer is protected by a semi-rigid
collar. These and several other commercially available electrodes are surrounded on the first few
tenths of a millimetre by a plastic collar. However, even with this mechanical stabilization, astrong force in the vertical direction will be transmitted up the electrolyte column and will disturbthe interface area. Motion artefact will still be seen in the trace. The rigid collar serves to minimizeinterference from minor pulls on the lead wire.
Pregelled ECG surface electrodes are manufactured with 0.3–1.5 g of electrolyte paste (gel,
cream, or jelly) in contact with the sensor which forms a conducting bridge with the skin. A highvalue of [CI
-] common in pastes renders the electrode more non-polarizable and decreases the
Fig.2.16 Disposable pre-gelled ECG electrode. A porous tape overlaying ring placed
over the electrode resists perspiration and ensures positive placement understress conditions (Courtesy: Mar Avionics, USA)

54 Handbook of Biomedical Instrumentation
electrolyte skin impedance. However, it must not be high enough to cause skin irritation. Ideally,
gels should possess the following properties:
• Stay moist for the intended shelf life and during use. This is controlled by including a
humectant in the gel.
• Prevent micro-organism and mould growth. Generally, the gels contain a bactericide/
fungicide and may be disinfected using gamma radiation.
• Provide low electrolyte skin impedance by having ionic salts and surfactants.• Cause minimum skin irritation, for which gels should have a pH range of 3.5–9.
Electrode-skin contact impedance with respect to pregelled disposable ECG electrodes was
measured by Klingler et al (1979 b) using a 10 Hz ac current source. It was found that the skin on
the average contributed a resistance of 564 W to the equivalent electrode impedance if mildly
abraided, but contributed 54.7 k W if applied to clean, dry skin. They further found that over 90% of
electrodes on abraided skin will have less than 5 k W impedance imbalance. The curve for clean,
dry skin shows a very significant imbalance. In fact 20% of the electrodes will exceed 15 k W
imbalance.
ECG electrodes are used in conjunction with cardiac monitors or ECG recorders which
invariably have dc-input bias currents. Klingler et al (1979 b) studied the effects of these small dc
currents on the offset potentials of disposable ECG electrodes. They found that after periodsranging from a few minutes to several days, the electrodes tested (four brands of silver-silverchloride and two brands of stainless steel) exhibited offset potentials exceeding 200 mV aftersubjection to dc bias currents over 200 nA. All silver-silver chloride electrodes were able towithstand bias currents of 200 nA, with minimal changes in offset for periods up to seven days.On the other hand, the stainless steel electrodes exhibited large offset potentials within minutesafter subjection to bias currents of only 100 nA. Based on this study, a 200 nA limit on the dc-inputbias current for cardiac monitors is suggested.
The modern ECG monitors are generally provided with inputs protected against defibrillator
overloads. The high defibrillating currents are harmlessly bypassed through neon or diode break-
down circuits. However, this unidirectional current passes through and tends to polarize the
electrodes. Usually, the standards on ECG monitors require that the trace be readable within 5 safter three or fewer defibrillator discharges. This implies that the electrode polarization voltagemust return to below 300 mV within a few seconds after application of the defibrillating voltages.FoamElectrolyteSnap
connectorVertical direction
Fluid column
protective cop
Foam
Critical
layer Sponge
Fig.2.17 Construction of a fluid column electrode

Bioelectric Signals and Electrodes 55
Schoenberg et al (1979) developed a standard test method for evaluating defibrillation recovery
characteristics of disposable ECG electrodes.
ECG pregelled electrodes can be characterized electrically by tests developed by the Association
for the Advancement of Medical Instrumentation (AAMI), USA to establish a reasonable safetyand efficacy level in clinical use of electrodes. In abridged form the standards are as follows:
1.Direct-current offset voltage: A pair of electrodes connected gel-to-gel, after 1 min stabiliza-
tion period must exhibit offset voltage no greater than 100 mV.
2.Combined offset instabiliy and internal noise: A pair of electrodes connected gel-to-gel shall
after 1 min. stablization generate a voltage no greater than 150 mV in the passband.
3.Alternating-current impedance: The average value of 10 Hz for at least 12 electrode pairs
connected at a level of impressed current not exceeding 100 mA peak to peak shall notexceed 2 k W. None of the individual pair impedances shall exceed 3 k W.
4.Defibrillation overload recovery: The absolute value of polarization potential of a pair of
electrodes connected gel-to-gel shall not exceed 100 mV, 5s after each of four capacitordischarges. The capacitor should be 10 mF charged to 200 V and discharged through the
electrode pair with 100 W in series.
5.Bias current tolerance: The observed dc voltage offset change across an electrode pair
connected gel-to-gel shall not exceed 100 mV when subjected to a continuous 200 mA dc
current over the period recommended by the manufacturer for the clinical use of theelectrodes. In no case shall this period be less than 8 hours.
/G32/G2E/G34/G2E/G34 /G50/G61/G73/G74/G65/G6C/G65/G73/G73/G20/G45/G6C/G65/G63/G74/G72/G6F/G64/G65/G73
ECG monitoring electrodes, in a majority of the cases, are metal plates applied to the skin aftercleaning and application of a coupling-electrolyte in the form of an electrode paste or jelly. Suchpreliminary preparation can be sometimes irritating and time consuming. Also, it is often not donesatisfactorily, resulting in problems like poor quality signals and baseline drift, etc. Anotherdisadvantage of using electrode jelly is that during long-term monitoring there is likely to bepatient-skin reactions as the electrode-skin interface dries out in a matter of a few hours. Theelectrodes need to be periodically removed for jelly replenishments, thus causing further discomfortdue to repetitive skin preparation. In addition, bacterial and fungal growth can take place underelectrodes worn over long periods. Also, in conductive electrodes, shifts in electrode position atthe electrode-skin interface appear as baseline drift, particularly when the subject moves.Therefore, any attempt of using a dry electrode that may dispense with the practice of skinpreparation would look attractive.
Capacitive Electrodes: A metal plate electrode in direct contact with the skin though makes a very
high resistive contact and has a considerable capacitive contact too with the skin (Stevens, 1963).By using a very high input impedance amplifier, it is possible to record a signal through the tissue-electrode capacitance. Lopez and Richardson (1969) describe the construction of electrodes which
can be capacitively coupled to the subject. The electrode consists of an aluminium plate which is
anodized on the surface to be placed in contact with the skin. The ohmic resistance of the anodizedelectrode is about 1 to 30 G W (1000–30,000 M W). Two such electrodes are applied to the subject

56 Handbook of Biomedical Instrumentation
without any preparation of the skin and the output of the source followers is connected to a
conventional electrocardiograph. Wolfson and Neuman (1969) also designed a capacitivelycoupling electrode and used a high input impedance amplifier having a MOSFET in the inputstage arranged in a source-follower configuration. The capacitances encountered in such type of
electrodes range from about 5000 to 20,000 pf/cm
2 of the electrode area (Geddes, 1972).
Conrad (1990) illustrates the construction of a capacitive electrode which is formed from
conductive silicon wafer, oxygen diffused into one surface produces a silicon dioxide layer, which
serves as the dielectric. A high performace FET operational amplifier in unity gain configurationacts as an impedance changer to permit use with systems designed for paste type electrodes. Theinsulated, capacitively coupled electrode is used on unprepared skin, which acts as one plate ofthe capacitor, the substrate acts as the other plate.
Luca et al (1979) designed an electrode and amplifier as an integrated unit, so that the assembly
could be used in the front end of the commonly used biomedical recorders. The arrangement(Fig. 2.18) basically comprises a metal shell which performs a dual function as a housing for theelectrode and as the ground contact. The shell is made of highly pure titanium metal measuring30¥ 1 5 ¥ 7 mm. Two FETs are cemented with epoxy glue in the middle of the shell, their centres
spaced 10 mm apart. The recording surfaces are formed by the cases of the two FETs. The canshave a diameter of 4.5 mm and are made of stainless steel. The rectangular border of the shell actsas the ground contact and the remainder of the shell forms a shield against interfering radiation.The source leads of the two FETs are connected to the differential inputs of an instrumentation
amplifier. The amplifier (Analog Devices 521) has a high ac input impedance (> 100 M W).
CanInstrumentation
amplifierCable Electrode
unitCurrent
sinks
10 K
10 K47 K
Shield
Shell
+V –VLED+
Fig.2.18 Schematic diagram of integrated electrode and amplifier arrangement for
pasteless operation (after Luca et al, 1979; reproduced by permission of Med.and Biol. Eng. and Comp.)
Among the drawbacks of the dry or pasteless electrodes are high electrode-skin resistance and
the strong sensitivity to motion. Whereas a high electrode-skin resistance can be overcome bysufficiently high values of the amplifier input resistance, no universal remedy for eliminatingmotion artefacts from dry bioelectrodes has so far been suggested. Skin preparation certainly

Bioelectric Signals and Electrodes 57
minimises motion artefacts (Burbank and Webster, 1978), but this time consuming manipulation
loses the practicability of dry electrodes for routine clinical use and is also less acceptable to thepatient.
An important characteristic observed with dry surface electrodes is that with the passage of
time, the resistance of a dry electrode placed on dry human skin decreases exponentially. Geddes
et al (1973) report that measurements made by them on silver disc electrodes (1.7 cm diameter)
showed that for 10 determinations, the average initial resistance was about 1.36 M W. After about
20 min. the resistance of the electrode-subject circuit had dropped to about one-tenth of its initialvalue. The fall in impedance is attributed to the presence of small amounts of perspirationaccumulated under the electrode.
Air-Jet ECG Electrodes: Wohnhas (1991) describes a novel air-jet electrode which employs
Bernoulli technology to achieve constant and secure electrode contact resulting in quality tracingswhile minimizing artefact and maximizing baseline stability.
Air-jet electrodes (Fig. 2.19) are Ag-AgCl electrodes encased within a contoured medical silicon
cup bounded by a skin-engaging rim. The contact area (pill) is anchored to a layer of synthetic,sintered carbon by a titanium screw. The miniature silver venturi air-jet bisects the sintered layerof synthetic carbon.
Fig.2.19 Air-jet electrodes (Courtesy: M/s Medi-Globe, USA)

58 Handbook of Biomedical Instrumentation
Fig.2.20 EEG electrode which can be applied to the surface of the skin by an adhesive
tape (Courtesy: In Vivo Metrics, USA)Ambient air is drawn to a small compressor and passed to the electrodes at a constant pressure
of 3.8 psi. As the air passes through the venturi air-jets, the electrodes produce a constant vacuum
of 2.0 psi at the point of contact. The constant vacuum allows the electrodes to remain securelyattached to the body. The fact that the air-jet functions with a positive flow of air ensures that nohair, skin particles or liquids clog the system. The electrode system eliminates gels, tapes and glueresidues and generally reduces the need to sand or shave the patient’s skin.
The combination of two special features—silver/silver chloride pills and sintered carbon
base-plates—optimize the effectiveness of the air-jet electrodes. The electrode pills possess theadvantage of low contact resistance. The micro-pressed carbon base-plate located behind theelectrode pill serves to accurately transmit the electronic signal acquired from the electrode pill tothe venturi air-jet and on to the stainless steel leadwire, and to securely position the air-jets withinthe electrode housing.
The venturi air-jet electrodes are integral components of the system comprising the ECG
electrodes and patient cable in one device and is compatible with any ECG recorder on the market.The system uses a compressor to pump ambient air into a distribution box. From the distributionbox, ten PVC hoses carry air and the stainless steel leadwires to the air jet electrodes. The electrodesystem is effective for both resting and exercise ECG applications. Heavy skin moisture is not a
problem for the air jet electrodes—all excess liquids are literally blown out of the system via the air-
jets. Also, the electrodes clean themselves when sprayed with a suitable disinfectant andsubmerged into water or alcohol while the system is running.
/G20/G32/G2E/G35 /G45/G4C/G45/G43/G54/G52/G4F/G44/G45/G53/G20/G46/G4F/G52/G20/G45/G45/G47
Among the most commonly used electrodes for EEG (electroencephalogram) recording are thechlorided silver discs (Fig. 2.20) having approximately 6–8 mm diameters. Contact with the scalp

Bioelectric Signals and Electrodes 59
is made via an electrolytic paste through a washer of soft felt. They have ac resistance varying from
3–20 k W. Small needle electrodes are sometimes used for carrying out special EEG studies when
they are inserted subcutaneously. Silver ball or pellet electrodes covered with a small cloth pad are
useful when electrical activity is to be recorded from the exposed cortex, but they have high dc
resistances.
Hector (1968) describes a pad electrode (Fig. 2.21(a)) which is made from a silver rod belled out
at the end and padded with a sponge, or a similar material, contained in gauze. It is screwed intoan insulated mount and held in place on the head with a rubber cap. To hold three such electrodes,an adjustable tripod mount is employed. Another type of EEG electrode consists of multiple finechlorided silver wires (Fig. 2.21(b)) fixed in a rigid plastic cup. The plastic cup is fixed to the scalpwith an adhesive. It is filled with jelly through a hole in the top. In this electrode, contact with thetissue is made via an electrolyte bridge so that jelly in contact with the electrode metal is notdisturbed by scalp movement. To avoid metal junctions which may get corroded with electrolyte,the silver wires are used as the output lead. The large surface area and excess of silver chloridefavour stability.
/G20/G32/G2E/G36 /G45/G4C/G45/G43/G54/G52/G4F/G44/G45/G53/G20/G46/G4F/G52/G20/G45/G4D/G47
Electrodes for electromyographic work are usually of the needle type (Fig. 2.22(a)). Needleelectrodes are used in clinical electromyography, neurography and other electrophysiologicalinvestigations of the muscle tissues underneath the skin and in the deeper tissues. The material ofthe needle electrode is generally stainless steel. In spite of the fact that stainless steel is unfavourableelectrode material from the point of view of noise, it is preferred in EMG work due to its mechanicalsolidity and low price. Needle electrodes are designed to be fully autoclavable and in any casethey should be thoroughly sterilized before use.
Chlorided
silver wire
Plastic
cup
SupportPad
electrode
(a) (b)
Fig.2.21 (a) Pad electrode for recording EEG (after Hector, 1968)
(b)EEG electrode consisting of chlorided silver wire in plastic cup. Jelly is
inserted from the hole kept for the purpose

60 Handbook of Biomedical Instrumentation
Needle electrodes come in various forms. The monopolar needle electrode usually consists of a
Teflon coated stainless steel wire which is bare only at the tip. It is found that after the needle hasbeen used a number of times, the Teflon coating will recede, increasing the tip area. The needle
must be discarded when this occurs. Bipolar (double coaxial) needle electrodes contain two
insulated wires within a metal cannula. The two wires are bared at the tip and provide the contactsto the patient. The cannula acts as the ground. Bipolar electrodes are electrically symmetrical andhave no polarity sense.
A concentric (coaxial) core needle electrode contains both the active and reference electrode
within the same structure. It consists of an insulated wire contained within a hypodermic needle(Fig. 2.22(b)). The inner wire is exposed at the tip and this forms one electrode. The concentricneedle is very convenient to use and has very stable electrical characteristics. Care should be takento maintain the surface electrode in good condition in order to avoid artefacts. Concentric needleelectrodes are made by moulding a fine platinum wire into a hypodermic needle having an outsidediameter less than 0.6 mm. One end of the needle is bevelled to expose the end of the wire and toprovide easy penetration when the needle is inserted. The surface area of the exposed tip of thewire may be less than 0.0005 mm
2.
(a) (b)
Fig.2.22 (a) Needle type EMG electrode
(b)Hypodermic needle type EMG electrode
Multi-element needle electrodes are used to pick up the signals from individual fibres of muscle
tissue. Special needles are available using 25-micron diameter electrode surfaces and having up to
14 pick up surfaces down the side of one needle. From the point of view of construction, needleelectrodes are the simplest. However, edging of the needle point to the suitable angle, providing aproper plastic coating, making them resistant against thermal and chemical stresses and ensuringhistological suitability is a difficult manufacturing process.
For the measurement of potentials from a specific part of the brain, longer needles are actually
inserted into the brain. The needles are precisely located by means of a map or atlas of the brain.Generally, a special instrument called a stereotaxic instrument is used to hold the subject’s headand guide the placement of electrodes. Often, these electrodes are implanted to permit repeatedmeasurements over an extended period of time.

Bioelectric Signals and Electrodes 61
The ground electrode for EMG studies usually consists of a conducting strip which is inserted
into a saline soaked strap and wrapped around the patient’s limb. The ground electrode is usuallypositioned over bony structures rather than over large muscle masses, in the vicinity of the
recording and stimulating electrodes, and where possible, equidistant from them. Surface
electrodes are employed for recording gross electrical activity from a particular group of underlyingmuscles in nerve-conduction velocity measurements. A single surface electrode may also be usedas the reference (indifferent) electrode with monopolar needle electrodes. Surface electrodes can beeasily and quickly attached and are generally comfortable to wear over long periods. Surfaceelectrodes usually consist of square or circular metal (chlorided silver) plates with leadoff wiresattached. They are held in place by straps or adhesive tapes. To reduce electrical resistance betweenthe skin and the electrode, the use of saline soaked felt pads or a small amount of electrode gelbetween the electrode surface and the skin is recommended. Disposable, adhesive type electrodesare also used for EMG work.
Bhullar et al (1990) describe the design and
construction of a selective non-invasive surfaceelectrode to study myoelectric signals. The electroderecording surfaces are two concentric steel rings. A
third ring attached to the casing of the electrode is
the earth contact. The rings are separated from eachother by Teflon, the insulating material. Figure 2.23shows a schematic diagram of the electrode con-figuration. The small surface area of the electrodeplates, the small physical size and the concentricarrangement produce the effect of recording signalsmainly from fibres near to the axis of the electrodeand thereby make the electrode much more selective.The concentric ring instead of the normal passiveelectrode configuration also obviates the problemof electrode alignment relative to the direction of the
muscle fibres. The results of tests undertaken with
these electrodes showed that it was able to pick upindividual motor unit action potentials at moderate force levels.
Crenner et al (1989) constructed a special electrode which allows recording of electrical
signals from their muscular layers, specifically to collect electromyographic signals of thegastrointestinal tract. The active electrode is surrounded by a ring which avoids the recording ofinterfering signals.
/G20/G32/G2E/G37 /G45/G4C/G45/G43/G54/G52/G49/G43/G41/G4C/G20/G43/G4F/G4E/G44/G55/G43/G54/G49/G56/G49/G54/G59/G20/G4F/G46/G20/G45/G4C/G45/G43/G54/G52/G4F/G44/G45/G20/G4A/G45/G4C/G4C/G49/G45/G53/G20/G41/G4E/G44/G20/G43/G52/G45/G41/G4D/G53
Conducting creams and jellies have for long been used to facilitate a more intimate contact betweenthe subject’s skin and the recording electrodes. The outer horny layer of the skin is responsible forthe bulk of the skin contact impedance, and for this reason careful skin preparation is essential inInverting electrode Earth ringNoninverting electrode
13 mm
Fig.2.23 Schematic diagram of the
electrode configuration to
study myoelectric signals(after Bhullar et al., 1990)

62 Handbook of Biomedical Instrumentation
order to obtain the best results. The recording site should first be cleaned with an ether-meth
mixture. In addition to having good electrical conductivity, the electrode jelly must have aparticular chloride ion concentration (about 1%) close to the physiological chloride concentration.
This is primarily important for long-term monitoring because it should not produce a harmful
diffusion between the jelly and the body. It is to be particularly ensured that the jelly chosen is ofa bland nature and does not contain soap or phenol which can produce a marked irritation of theskin after a few hours.
The electrical conductivity of different makes of electrode cream can be measured (Hill and
Khandpur, 1969) by means of the Schering ac bridge circuit. The cream is placed in a perspexconductivity cell of known dimensions and the resistive component of the cell impedance ismeasured at 10 Hz, the conductivity being calculated from the cell dimensions.
The contact impedance of the skin depends upon the type of electrolyte used and the time
(Trimby, 1976). Figure 2.24 shows the effect of these parameters. A low concentration sodiumchloride electrolyte has 0.5% NaCI and a high concentration electrolyte has a concentration in therange of 5 to 10% NaCI. The impedance is found to fall rapidly to 40% between 7 to 30 min.Stabilization occurs at about 30 to 45 min. An interesting observation from this figure is that whilepre-rubbing the skin will lower the initial impedance value, the final value after using a highconcentration electrolyte becomes nearly the same.
500 K
100 K
5 K
15 30
Time (minutes)45Rs(ohms)High concentration
electrolyteLow concentration
electrolyte (NaCl)
Abraided
skin
Fig.2.24 Variation of contact impedance with electrolyte concentration and time (re-
drawn after Trimby, 1976; Courtesy: Hewlett Packard, USA)
Electrode jelly can be replaced in certain cases by using a conducting plastic as an interface
between the electrode and the surface of the body. Jenkner (1967) used silastic S-2086 by DowCorning with EEG electrodes and showed that contact resistance was almost the same as with aconventional electrode which make use of electrode jelly.

Bioelectric Signals and Electrodes 63
/G20/G32/G2E/G38 /G4D/G49/G43/G52/G4F/G45/G4C/G45/G43/G54/G52/G4F/G44/G45/G53
To study the electrical activity of individual cells, microelectrodes are employed. This type of
electrode is small enough with respect to the size of the cell in which it is inserted so thatpenetration by the electrode does not damage the cell. The size of an intracellular microelectrode is
dictated by the size of the cell and the ability of its enveloping membrane to tolerate penetration by
the microelectrode tip. Single-living cells are rarely larger than 0.5 mm (500 microns) and areusually less than one-tenth of this size. Typical microelectrodes have tip dimensions ranging from0.5 to 5 microns. The tips of these electrodes have to be sufficiently strong to be introduced throughlayers of tissues without breaking.
Two types of microelectrodes are generally used: metallic (Fig. 2.25(a)) and glass microcapillaries
(Fig. 2.25(b)). Metallic electrodes are formed from a fine needle of a suitable metal drawn to afine tip. On the other hand, glass electrodes are drawn from Pyrex glass of special grade. Thesemicrocapillaries are usually filled with an electrolyte. The metal microelectrodes are used in directcontact with the biological tissue and, therefore, have a lower resistance. However, they polarizewith smaller amplifier input currents. Hence, they tend to develop unstable electrode offsetpotentials and are therefore not preferred for steady state potential measurements. On the otherhand, in case of glass microelectrodes, improved stability can be obtained by properly choosingthe metal and the electrolyte so that the small current passing through their junction may not be
able to modify the electrical properties of the electrodes. Also, the glass microelectrode has a
substantial current carrying capacity because of the large surface contact area between the metaland the electrolyte.
Metal
Glass TipLead wire
CapElectrolyteMetal electrode
Tip
(a) (b)
Fig.2.25 (a) Microelectrodes—metal microelectrodes
(b)Microelectrodes—micropipette or micro capillaries electrode
The microelectrodes have a very high impedance as compared to conventional electrodes
used for recording ECG, EEG, etc. The high impedance of a metal microelectrode is due to thecharacteristics of the small area metal-electrolyte interface. Similarly, a micropipet tip is filled withan electrolyte which substitutes an electrolytic conductor of small cross-sectional area, whichgives a micropipet its high resistance. Because of high impedance of microelectrodes, amplifiers

64 Handbook of Biomedical Instrumentation
with extremely high input impedances are required to avoid loading the circuit and to minimize
the effects of small changes in interface impedance.
/G32/G2E/G38/G2E/G31 /G47/G6C/G61/G73/G73/G20/G4D/G69/G63/G72/G6F/G63/G61/G70/G69/G6C/G6C/G61/G72/G79/G20/G45/G6C/G65/G63/G74/G72/G6F/G64/G65/G73
Several methods exist for producing microelectrodes of wide variety and shapes. For drawingelectrodes of uniform and accurate diameter, it is essential to maintain constant timing, tem-perature, strength and direction of pull. These factors are difficult to control when the electrodesare drawn manually. The mechanical method employs gravitational force for extension and theelectrodes which are drawn in one or more stages can readily produce capillary tubes between3–30 mm diameter, but great difficulty is encountered in producing electrodes of less than 1 mm.
The most commonly used method for making small tip micropipet consists of the circumferential
application of heat to a small area of glass tubing which is placed under some initial tension.
When the glass softens, the tension is increased very rapidly and the heat is turned off. Propertiming, controlled adjustment of the amount of heat as well as the initial and final tensions andcooling result in the production of microcapillaries with controlled dimensions.
/G32/G2E/G38/G2E/G32 /G4D/G65/G74/G61/G6C/G20/G4D/G69/G63/G72/G6F/G65/G6C/G65/G63/G74/G72/G6F/G64/G65/G73
Metal electrodes with very fine tips used for recording from single cells have the advantage overglass micropipetes of being relatively robust. Steel microelectrodes can be made from ordinary
darning needles but preferably they should be of good stainless steel wire. They can be easily
made up to 10 mm diameter but great care has to be taken for diameters as small as 1 mm. These
very small tips are not very satisfactory as they are extremely brittle and have very high inputimpedance. Hubel (1957) described a method to make tungsten microelectrodes with a tip diameterof 0.4 mm. He used electropointing technique which consists in etching a metal rod while the metal
rod is slowly withdrawn from the etching solution, thus forming a tapered tip on the end of therod. The etched metal is then dipped into an insulating solution for placing insulation on all butthe tip.
Figure 2.26 shows the cross-section of a metal
microelectrode. In this electrode, a thin film ofprecious metal is bonded to the outside of a drawnglass microelectrode. This arrangement offers lowerimpedance than the microcapillary electrode, infinite
shelf life and reproduciable performance, with ease
of cleaning and maintenance. The metal—electrolyte
interface is between the metal film and the electrolyteof the cell.
Skrzypek and Keller (1975) illustrated a new
method of manufacturing tungsten microelectrodepermitting close control of microelectrode para-
meters. In this technique, the tips are dc electroetched
to diameters below 500
° A and completely coveredGlass Metal film
InsulationTip
Fig.2.26 Microelectrode—thin metal
film coated micropipette

Bioelectric Signals and Electrodes 65
with polymethyl methacrylate. An electron beam from a scanning electron microscope is then
used to expose a precise area on the tip for later removal by chemical methods. Recording results
with these electrodes suggested good desirable recording characteristics, i.e. ability to isolate and
hold single cells.
Tungsten is preferred for constructing micro-electrodes due to its mechanical strength and its
apparent inertness. Although tungsten itself is reactive, a surface layer of tungsten oxide will,in most situations, protect the metal against corrosion. The electrical properties of tungstenmicroelectrodes made with a taper of the tip of about 1:10 and insulated with lacquer leaving a tiplength of about 10–100 μm were studied by Zeuthen (1978). The resting potential in saline wasfound to be –0.3 V relative to a silver-silver chloride reference electrode for input currents less than10
–12 A. The small signal impedance was ideally that of a capacitor 0.4 pF/ mm2between 10 and
1000 Hz. Imperfect insulation at the tip caused this impedance to be increasingly resistive. Theelectrochemical properties of tungsten showed that it behaves as an inert metal within thepotentials where it is usually used in biological experiments.
Jobling et al (1981) constructed an active microelectrode array using IC fabrication technology to
record extracellular potentials in nervous tissues. The array substantially reduces the noise causedby electrostatic pick-up with a good long-term stability.

66 Handbook of Biomedical Instrumentation
/G50/G68/G79/G73/G69/G6F/G6C/G6F/G67/G69/G63/G61/G6C/G20/G54/G72/G61/G6E/G73/G64/G75/G63/G65/G72/G73
/G20/G33/G2E/G31 /G49/G4E/G54/G52/G4F/G44/G55/G43/G54/G49/G4F/G4E
Transducers are devices which convert one form of energy into another. Because of the familiar
advantages of electric and electronic methods of measurement, it is the usual practice to convert allnon-electric phenomenon associated with the physiological events into electric quantities.Numerous methods have since been developed for this purpose and basic principles of physicshave extensively been employed. Variation in electric circuit parameters like resistance, capaci-
tance and inductance in accordance with the events to be measured, is the simplest of such
methods. Peizo-electric and photoelectric transducers are also very common. Chemical events aredetected by measurement of current flow through the electrolyte or by the potential changesdeveloped across the membrane electrodes. A number of factors decide the choice of a particulartransducer to be used for the study of a specific phenomenon. These factors include:
• The magnitude of quantity to be measured.
• The order of accuracy required.• The static or dynamic character of the process to be studied.• The site of application on the patient’s body, both for short-term and long-term monitoring.• Economic considerations.
/G20/G33/G2E/G32 /G43/G4C/G41/G53/G53/G49/G46/G49/G43/G41/G54/G49/G4F/G4E/G20/G4F/G46/G20/G54/G52/G41/G4E/G53/G44/G55/G43/G45/G52/G53
Many physical, chemical and optical properties and principles can be applied to construct
transducers for applications in the medical field. The transducers can be classified in many ways,such as:
(i) By the process used to convert the signal energy into an electrical signal. For this, trans-
ducers can be categorized as:
Active Transducers —a transducer that converts one form of energy directly into another. For
example: photovoltaic cell in which light energy is converted into electrical energy.HAPTER
33

Physiological Transducers 67
Passive Transducers —a transducer that requires energy to be put into it in order to translate
changes due to the measurand. They utilize the principle of controlling a dc excitationvoltage or an ac carrier signal. For example: a variable resistance placed in a Wheatstone
bridge in which the voltage at the output of the circuit reflects the physical variable. Here,
the actual transducer is a passive circuit element but needs to be powered by an ac or dcexcitation signal.
(ii) By the physical or chemical principles used. For example: variable resistance devices, Hall
effect devices and optical fibre transducers.
(iii) By application for measuring a specific physiological variable. For example: flow
transducers, pressure transducers, temperature transducers, etc.
Biomedical parameters which are commonly encountered are listed alongwith their characteristics
and corresponding transducers in Tables 3.1 to 3.3.
/G20/G33/G2E/G33 /G50/G45/G52/G46/G4F/G52/G4D/G41/G4E/G43/G45/G20/G43/G48/G41/G52/G41/G43/G54/G45/G52/G49/G53/G54/G49/G43/G53/G20/G4F/G46/G20/G54/G52/G41/G4E/G53/G44/G55/G43/G45/G52/G53
A transducer is normally placed at the input of a measurement system, and therefore, its
characteristics play an important role in determining the performance of the system. The
characteristics of a transducer can be categorized as follows:
/G33/G2E/G33/G2E/G31 /G53/G74/G61/G74/G69/G63/G20/G43/G68/G61/G72/G61/G63/G74/G65/G72/G69/G73/G74/G69/G63/G73
Accuracy: This term describes the algebraic difference between the indicated value and the true or
theoretical value of the measurand. In practice, accuracy is usually expressed as a percentage of
full scale output or percent of reading or ± number of digits for digital readouts.
Precision: It refers to the degree of repeatability of a measurement. Precision should not be confused
with accuracy. For example: an uncompensated offset voltage in an operational amplifier maygive very reproducible results (high precision), but they may not be accurate.
Resolution: The resolution of a transducer indicates the smallest measureable input increment.
Sensitivity: It describes transfer ratio of output to input.
Drift: It indicates a change of baseline (output when input is zero) or of sensitivity with time,
temperature etc. It may be noted that the sensitivity of the device does not change the calibrationcurve if shifted up or down.
Linearity: It shows closeness of a transducer’s calibration curve to a specified straight line with
in a given percentage of full scale output. Basically, it reflects that the output is in some wayproportional to the input.
Threshold: The threshold of the transducer is the smallest change in the measurand that will result
in a measureable change in the transducer output. It sets a lower limit on the measurementcapability of a transducer.
Noise: This is an unwanted signal at the output due either to internal source or to interference.

68 Handbook of Biomedical Instrumentation
/G95/G20/G54/G61/G62/G6C/G65/G20/G33/G2E/G31 Signals from Cardiovascular System
Parameter Primary signal characteristics Transducer required
Blood pressure Frequency range: dc to 200 Hz Unbonded wire strain gauge pressure
(arterial, direct) (dc to 40 Hz usually adequate) transducer
Pressure range: 20 to 300 mm Hg; Bonded semiconductor strain gauge
slightly negative in left ventricle transducer
Capacitance type pressure transducerDifferential transformer
Blood pressure Frequency range: dc to 5 Hz Low frequency microphone for pick-(arterial, indirect) Pressure range: 20 to 300 mm Hg ing up Korotkoff sounds
Microphone: 10 to 100 Hz, signal
voltage depends upon the type ofmicrophone used
Blood pressure Frequency range: dc to 40 Hz Strain gauge pressure transducer
(venous, direct) Pressure range: –5 to +20 mm Hg with higher sensitivity
Peripheral arterial blood Frequency range: 0.1 to 50 Hz Finger or ear lobe pick-up (light source
pressure pulse wave usually adequate and photocell) and
Pulse trace similar to arterial blood piezo-electric pick-up or by impedance
pressure waveform method
Blood flow Rate: 0 to 300 ml/s Tracer methods
(aortic or venous) Frequency range: 0 to 100 Hz Electromagnetic flowmeter
Ultasonic Doppler flowmeter
Cardiac output Frequency range: 0 to 60 Hz Dye dilution methods
(blood flow) (0 to 5 Hz usually adequate) Integration of aortic blood flow func-
Blood flow measured as litres tion
per minute (4 to 25 litre/min) Thermal dilution method
Electromagnetic flowmeter and Inte-
grator.
Heart rate Rate: 25 to 300 beats per min. Derived from ECG or arterial
Normal human heart rate at rest blood pressure waveform or photo-
60 to 90 beats/min. electric plethysmograph
Normal foetal heart rate: 110 to 175 From foetal phonocardiogram ultra-
beats/min. sonic method, scalp electrodes for
foetal ECG
Phonocardiogram Frequency range: 20 to 2000 Hz Crystal or moving coil microphone
(heart sounds) Signal voltage depends upon the
type of microphone used
Ballistocardiogram Frequency range: dc to 40 Hz Infinite period platform with strain
(BCG) gauge accelerometerImpedance cardiogram Frequency range: dc to 50 Hz Surface or needle electodes
(Rheocardiography) Impedance range: 10 to 500 W
Typical frquency employed formeasuring impedance: 20 kHz
to 50 kHz
Oximetry Frequency range: 0 to 60 Hz; Photoelectric pulse pick-up
0 to 5 Hz usually adequate Photoelectric flow-through
cuvette

Physiological Transducers 69
/G95/G20/G54/G61/G62/G6C/G65/G20/G33/G2E/G32 Signals from Respiratory System
Parameter Primary signal characteristics Transducer required
Respiration rate Normal range: 0 to 50 breaths/min Thermistor
Impedance pneumography
electrodes
Microswitch method
CO2 detectors
Strain gauge transducer
Doppler shift transducer
Pneumotachogram Frequency range: dc to 50 Hz Fleisch pneumotachograph
(respiratory flow rate) BMR spirometersImpedance pneumograh Frequency range: dc to 30 Hz Surface or needle electodes
(from demodulated carrier)
Typical carrier frequency used:
20 kHz to 50 kHz
Tidal volume Frequency range: dc to 5 Hz Direct from spirometer
(Volume/breath) Typical value for adult human: Integrated from pneumota-
600 ml/breath chogram
Minute ventilation 6 to 8 l/min Integrated from pneumota-
(vol/min) chogram
Gases in expired air CO
2–normal range: 0 to 10% Infrared sensors
End tidal
CO2 (human) 4 to 6% Mass spectrometery
N2O – Range 0 to 100%
Halothane–Range 0 to 3%
Diffusion of inspired Normal range of nitrogen concen- Discharge tube nitrogen ana-
gases (nitrogen wash- tration differential: 0 to 10% lyzer
out technique)
Pulmonary diffusing Normal range(human): 16 to 35 ml Infrared absorption by CO
capacity (using carbon CO/mmHg/min
monoxide)
Dissolved Gases and pH
pH Signal range: 0 to ± 700 mV Glass and calomel electrodes
Partial pressure of Range: 1 to 1000 mm Hg Direct recording CO2 electrode
dissolved CO2(PCO2) Normal signal range: 0 to ± 150 mV
Partial pressure of Normal measurement range: 0 to Polarographic electrodes
dissolved oxygen ( PO2) 800 mm Hg
Hyperbaric PO2 range: 800 to 3000
mm Hg
Na+, CI–Cations Signal range: 60 mV/decade Specific ion-sensitive electrodes
Hysteresis: It describes change of output with the same value of input but with a different history
of input variation. For example: hysteresis is observed when the input/output characteristics fora transducer are different for increasing inputs than for decreasing inputs. It results when some ofthe energy applied for increasing inputs is not recovered when the input decreases.
Span: It indicates total operating range of the transducer.

70 Handbook of Biomedical Instrumentation
Saturation: In a transducer the output is generally proportional to the input. Sometimes, if the
input continues to increase positively or negatively, a point is reached where the transducer willno longer increase its output for increased input, giving rise to a non-linear relationship. The
region in which the output does not change with increase in input is called the saturation region.
Conformance: Conformance indicates closeness of a calibration curve to a specified curve for an
inherently non-linear transducer.
/G33/G2E/G33/G2E/G32 /G44/G79/G6E/G61 /G6D/G69/G63/G20/G43/G68/G61/G72/G61/G63/G74/G65/G72/G69/G73/G74/G69/G63/G73
Only a few signals of interest in the field of medical instrumentation, such as body temperature,
are of constant or slowly varying type. Most of the signals are function of time and therefore,it is the time varying property of biomedical signals that is required to consider the dynamiccharacteristics of the measurement system. Obviously, when a measurement system is subjected tovarying inputs, the input-output relation becomes quite different form that in the static case. Ingeneral, the dynamic response of the system can be expressed in the form of a differential equation.
For any dynamic system, the order of the differential equation that describes the system is called
the order of the system. Most medical instrumentation systems can be classified into zero-, first-,second-, and higher-order systems.
Azero-order system has an ideal dynamic performance, because the output is proportional to
the input for all frequencies and there is no amplitude or phase distortion. A linear potentiometer
used as a displacement trasduser is a good example of a zero-order transducer.
Thefirst-order transducer or instrument is characterized by a linear differential equation. The
temperature transducers are typical examples of first order measuring devices since they can
be characterized by a single parameter, i.e. time constant ‘ T’. The differential equation for the
first-order system is given by/G95/G20/G54/G61/G62/G6C/G65/G20/G33/G2E/G33 Physical Quantities
Parameter Primary signal characteristics Transducer required
Temperature Frequency range: dc to 1 Hz Thermocouples
Electrical resistance thermometerThermistor
Silicon diode
Galvanic skin resistance Resistance range: 1 k to 500 kW Surface eletrodes similar to that
(GSR) of ECG electrodes
Plethysmogram (volume Frequency range: dc to 30 Hz Qualitative Plethysmography by:measurement) -photoelectric method
-impedance method-piezo-electric method
Tocogram (uterine contr- Frequency range: dc to 5 Hz Strain gauge transducersations measurement
during labour)
Isometric force, dimensional Strain gauge (resistance or
change semiconductor

Physiological Transducers 71
y+T dy/dx = x(t)
where x is the input, y the output and x(t), the time function of the input.
A transducer or an instrument is of second-order if a second-order differential equation is
required to describe the dynamic response. A typical example a second-order system is thespring—mass system of the measurement of force. In this system, the two parameters thatcharacterize it are the natural frequency f
n(or angular frequency wn= 2pfn), and the damping ratio
‘z’. The second-order differential equation for this system is given by
(1/w2
n)d2y/dx2 + (2z/wn)dy/dx+y=x(t)
where wn is in radians per second and ‘ z’ is a non-dimensional quantity. It may be noted that in
this system, mass, spring and viscous-damping element oppose the applied input force and theoutput is the resulting displacement of the movable mass attached to the spring.
In the second-order systems, the natural frequency is an index of speed of response whereas the
damping ratio is a measure of the system stability. An under-damped system exhibits oscillatory
output in response to a transient input whereas an over-damped system would show sluggish
response, thereby taking considerable time to reach the steady-state value. Therefore, such systemsare required to be critically damped for a stable output.
Another important parameter to describe the dynamic performance of a transducer is the
response time . It characterizes the response of a transducer to a step change in the input
(measurand) and includes rise time, decay time and time constant.
/G33/G2E/G33/G2E/G33 /G4F/G74/G68/G65/G72/G20/G43/G68/G61/G72/G61/G63/G74/G65/G72/G69/G73/G74/G69/G63/G73
There are many other characteristics which determine the performance and choice of a transducerfor a particular application in medical instrumentation systems. Some of these characteristics are:
• Input and output impedance
• Overlead range• Recovery time after overload• Excitation voltage• Shelf life• Reliability• Size and weight
/G20/G33/G2E/G34 /G44/G49/G53/G50/G4C/G41/G43/G45/G4D/G45/G4E/G54/G2C/G20/G50/G4F/G53/G49/G54/G49/G4F/G4E/G20/G41/G4E/G44/G20/G4D/G4F/G54/G49/G4F/G4E/G20/G54/G52/G41/G4E/G53/G44/G55/G43/G45/G52/G53
These transducers are useful in measuring the size, shape and position of the organs and tissues
of the body. Specifically, the following measurements are made:
Position: Spatial location of a body or point with respect to a reference point.
Displacement: Vector representing a change in position of a body or point with respect to a
reference point. Displacement may be linear or angular.
Motion: Change in position with respect to a reference system.

72 Handbook of Biomedical Instrumentation
Displacement transducers can be used in both direct and indirect systems of measurement.
For example: direct measurements of displacement could be used to determine the change indiameter of the blood vessels and the changes in volume and shape of the cardiac chambers.
Indirect measurements of displacement are used to quantify the movement of liquids through the
heart valves. For example: detection of movement of the heart indirectly through movement ofa microphone diaphragm. Displacement measurements are of great interest because they formthe basis of many transducers for measuring pressure, force, acceleration and temperature, etc.The following types of transducers are generally used for displacement, position and motionmeasurements.
/G33/G2E/G34/G2E/G31 /G50/G6F/G74/G65/G6E/G74/G69/G6F /G6D/G65/G74/G72/G69/G63/G20/G54/G72/G61/G6E/G73/G64/G75/G63/G65/G72/G73
One of the simplest transducers for measuring displacement is a variable resistor (potentiometer)
similar to the volume control on an audio electronic device. The resistance between two terminalson this device is related to the linear or angular displacement of a sliding tap along a resistanceelement (Fig. 3.1).
Rotary
input
AC B
Rotational displacement Linear displacement
(b) (a)CLinear inputAB
Fig.3.1 Principle of a potentiometric displacement transducer for the measurement
of (a) linear displacement (b) angular displacement
When the fixed terminals of the potentiometer are connected to the power supply, either
ac or dc, output voltage at the wiper varies with the displacement of the object. Single turn, multi-turn (2 to 10 turns), linear, logarthmic and sectored potentiometers are available for differentapplications.
/G33/G2E/G34/G2E/G32 /G56/G61/G72/G69/G61 /G62/G6C/G65/G20/G43/G61/G70/G61/G63/G69/G74/G61/G6E/G63/G65
When the distance between a pair of metallic plates forming an electrical capacitance is altered,
there is a change in the capacitance according to the relation:
C= 0.0885 k ◊A/d
where C = capacitance in micro-microfarads

Physiological Transducers 73
d= distance between the plates in cm
A= area of each identical plate in cm2
k= dielectric constant of the medium separating the two plates
Each of the quantities in this equation can be varied to form a displacement transducer. By
moving one of the plates with respect to the other, the capacitance will vary inversely with respect
to the plate separation. This will give a hyperbolic displacement—capacitance characteristic.
However, if the plate separation is kept constant and the plates are displaced laterally withrespect to one another so that the area of overlap changes, a displacement can be produced—capacitance characteristic that can be linear, depending upon the shape of the actual plates.
The inverse relationship between Cand dmeans that the sensitivity increases as the plate
separation decreases. Hence, it is desirable to design a capacitor transducer with a small plateseparation and a large area. However, in the case of a displacement transducer, it is the change incapacitance that is proportional to the displacement, the sensitivity will be independent of platearea but increases as dapproaches zero.
For measurement of displacement, the capacitor is made a part of an LC oscillator wherein the
resulting frequency change can be converted into an equivalent output voltage. However, lack ofrepeatability and the difficulty in properly positioning the transducer makes the measurement adifficult task.
Capacitance transducers are very sensitive displacement transducers. Therefore, they are
required to be thermally insulated and the connecting cables have to be made as small as possiblein length to avoid involvement of the cable capacitance in the measuring circuit. They have thedisadvantage that a special high frequency energizing system is needed for their operation.
/G33/G2E/G34/G2E/G33 /G56/G61/G72/G69/G61 /G62/G6C/G65/G20/G49/G6E/G64/G75/G63/G74/G61/G6E/G63/G65
Changes in the inductance can be used to measure displacement by varying any of the three coil
parameters given in the following equation
L=n2Gm
Where L= inductance of the coil
n= number of turns in a coil
m= permeability of the medium
G= geometric form factor
The inductance can be changed either by varying its physical dimensions or by changing
the permeability of its magnetic core. The core having a permeability higher than air can be madeto move through the coil in relation to the displacement. The changes in the inductance can bemeasured using an ac signal which would then correspond to the displacement.
Another transducer based on inductance is the variable reluctance transducer. In this
transducer, the core remains stationary but the air gap in the magnetic path of the core is varied tochange the effective permeability. It may however be noted that the inductance of the coil in thesecases is usually not related linearly to the displacement of the core or the size of the air gap,especially if large displacements are to be measured.

74 Handbook of Biomedical Instrumentation
/G33/G2E/G34/G2E/G34 /G4C/G69/G6E/G65/G61/G72/G20/G56/G61/G72/G69/G61 /G62/G6C/G65/G20/G44/G69/G66/G66/G65/G72/G65/G6E/G74/G69/G61/G6C/G20/G54/G72/G61/G6E/G73/G66/G6F/G72 /G6D/G65/G72/G20/G28/G4C/G56/G44/G54/G29
An extremely useful phenomenon frequently
utilized in designing displacement pick up unitsis based upon variations in coupling between
transformer windings, when the magnetic core of
the transformer is displaced with respect to theposition of these two windings. These transducerscan be conveniently used for the measurement ofphysiological pressures. Generally, a differentialtransformer (Fig. 3.2) designed on this principleis employed for this purpose. The central coilis the energizing or primary coil connectedto a sinewave oscillator. The two other coils(secondary coils) are so connected that theiroutputs are equal in magnitude but opposite inphase.
With the ferromagnetic core symmetrically
placed between the coils, and the two secondary
coils connected in series, the induced output voltage across them would be zero.
When the core is moved, the voltage induced in one of the secondary windings exceeds that
induced in the other. A net voltage output then results from the two secondaries. The phase of the
output will reverse if the core moves past the central position. A simple bridge circuit can beemployed to detect the differential signals thus produced. The signal can be further processed todirectly display a calibrated output in terms of mm of displacement. Since there is always somecapacitive coupling between windings of differential transducers, it produces a quadraturecomponent of induced voltage in the secondary windings. Because of the presence of thiscomponent, it is usually not possible to reduce the output voltage to zero unless the phase of thebacking voltage is also altered.
Differential transformer displacement transducers generally work in conjunction with carrier
amplifiers. Typical operating excitation of these transducers is 6 V at 2.5 kHz. Since the outputvoltage of the LVDT is proportional to the excitation voltage, the sensitivity is usually defined fora 1 Volt excitation. Commercial devices typically have a sensitivity of 0.5 to 2.0 mV per 0.001 cmdisplacement for a 1 Volt input. Full scale displacements of 0.001 to 25 cm with a linearity of ± 0.25% are available.
/G33/G2E/G34/G2E/G35 /G4C/G69/G6E/G65/G61/G72/G20/G6F/G72/G20/G41/G6E/G67/G75/G6C/G61/G72/G20/G45/G6E/G63/G6F/G64/G65/G72/G73
With the increasing use of digital technology in biomedical instrumentation, it is becomingdesirable to have transducers which can give the output directly in the digital format. Transducersare now available for the measurement of linear or angular displacements, which provide outputin the digital form. These transducers are basically encoded disks or rulers with digital patternsphotographically etched on glass plates. These patterns are decoded using a light source and anarray of photodetectors (photo-diodes or photo-transistors). A digital signal that indicates theAmplifierRR
Zero
adjustmentSec. 1
Sec. 2Primary
Movable
coreDiaphragm
Fig.3.2 Principle of LVDT pressure trans-
ducer. The diagram shows thediffrential transformer and bridgecircuit for detection of diffrentialsignals

Physiological Transducers 75
position of the encoding disk is obtained, which thereby represents the displacement being
measured. Figure 3.3 shows typical patterns on the digital encoders.
248 1 6 1
+–
Fig.3.3 Spatial encoder using a binary counting system
The encoder consists of a cylinderical disk with the coding patterns arranged in concentric
rings, having a defined number of segments on each ring. The number of segments on the concentricrings decrease, in a binary count (32-16-8-4-2), from a total of 32 (16 conductive and 16 non-conductive) on the outside ring to two on the inside ring. Each angular position of the disk wouldhave a different combination of segments which will indicate the position of the shaft on whichthe disk is mounted. Since the outer ring of the disk encoder has 32 distinct areas, the resolutionwould be 1/32 ¥ 360° = 11–1/4° (1 digit). The resolution can be improved by increasing the
number of segments on each ring, thereby decreasing the angle corresponding to each segment.
/G33/G2E/G34/G2E/G36 /G50/G69/G65/G7A/G6F/G2D/G65/G6C/G65/G63/G74/G72/G69/G63/G20/G54/G72/G61/G6E/G73/G64/G75/G63/G65/G72/G73
The piezo-electric effect is a property of natural crystalline substances to develop electric potentialalong a crystallographic axis in response to the movement of charge as a result of mechanicaldeformation. Thus, piezo-electricity is pressure electricity. The reverse effect is also present. Whenan electrical field is applied, the crystal changes shape. On application of pressure, the charge Qdeveloped along a particular axis is given by
Q=dF coulomb
where dis the piezo-electric constant (expressed in Coulombs/Newton, i.e. C/N) and F is the
applied force. The change in voltage can be found by assuming that the system acts like a parallel—plate capacitor where the voltage E
oacross the capacitor is charge Q divided by capacitance C.
Therefore
Eo=Q
C=dF
C

76 Handbook of Biomedical Instrumentation
The capacitance between two parallel plates of area ‘ a’ separated by distance ‘ x’ is given by
C=Œa
x
where Πis the dielectric constant of the insulator between the capacitor plates. Hence,
Eo=d
Œ◊F
a◊x = g◊P◊x
where d/Π= g is defined as the voltage sensitivity in volts, P is the pressure acting on the crystal
per unit area and x is the thickness of the crystal.
Typical values for d are 2.3 C/N for quartz and 140 C/N for barium titanate. For a piezo-electric
transducer of 1 cm2 area and 1 mm thickness with an applied force due to 10 g weight, the output
voltage is 0.23 mV and 14 mV for the quartz and barium titanate crystals respectively.
The priniciple of operation is that when an asymmetrical crystal lattice is distorted, a charge
re-orientation takes place, causing a relative displacement of negative and positive charges. Thedisplaced internal charges induce surface charges of opposite polarity on opposite sides of thecrystal. Surface charge can be determined by measuring the difference in voltage between
electrodes attached to the surfaces.
Piezo-electric materials have a high resistance, and therefore, when a static deflection is applied,
the charge leaks through the leakage resistor. It is thus important that the input impedance of the
voltage measuring device must be higher than that of the piezo-electric sensor.
In order to describe the behaviour of piezo-electric elements, equivalent circuit techniques have
been utilized. The analogies for electro-mechanical system which have been used for piezo-electricstructures are:
Electrical Units Analogous Mechanical Units
Voltage Force
Current Velocity
Charge Displacement
Capacitance ComplianceInductance Mass
Impedance Mechanical impedance
Figure 3.4 shows the two basic equivalent circuits. If the constants in the two circuits are
suitably related, the circuits are equivalent at all frequencies. The mechanical terminals representthe face or point of mechanical energy transfer to or from the piezo-electric element. The inductancesymbol represents the effective vibrating mass of the element. The transformer symbol representsan ideal electro-mechanical transformer—a device that transforms voltage to force and vice versa,and current to velocity and vice versa. The transformation ratio N:1 in the circuit A is the ratio of
voltage input to force output of the ideal transformer and also the ratio of velocity input to current
output. The transformation ratio 1:N in circuit B has similar significance. The capacitance symbolson the electrical side represent electrical capacitances. The capacitance symbol on the mechanicalside represent mechanical compliances.
A large number of crytsallographic substances are known but the most extensively used piezo-
electric crystals are quartz, tourmaline, ammonium dihydrogen phosphate, Rochelle salt, lithiumsulphate, lead zirconate and barium titanate.

Physiological Transducers 77
Quartz is the most stable natural crystal with high mechanical and thermal stability and has
volume resistivity higher than 104 ohm-cm and small internal electric loss. Rochelle salt has low
mechanical strength, high volume resistivity of 1012 ohm–cm and is affected by humidity.
Ammonium dihydrogen phosphate has high thermal stability, volume resistivity of 1012 ohm– cm
and high sensitivity. Barium titanate ceramic is a ferroelectric crystal and has small voltage output.The voltage output versus strain input characteristic is linear only on a very small range and isaffected by humidity. They have a range of measurement from 1N to 2000N, linearity within ± 1%,are simple in structure, stiff in nature, robust in use and preferred for dynamic measurements.
The piezo-electric transducers manufactured by M/S Brush Clevite, UK are sold under the
trade name PZT which indicates a Lead Zirconate–Lead Titanate composition. They are available
in many types for specific applications such as:
• PZT-2 is used for high frequency ultrasonic transducer applications and delay lines.
• PZT- 4 is used for high power acoustic radiating transducers for use in deep–submersion
type applications and as the active element in electrical power generating systems.
• PZT-5A has high sensitivity, high time stability and resistivity at elevated temperatures.
They are mostly used in hydrophones and instrument applications.
• PZT-5H has higher sensitivity and higher diaelectric constant than PZT-5A. However, it
has a lower temperature stability and limited working temperature range.
• PZT-8 is a high power material which has much lower diaelectric losses under high
electric drive.
Piezo-electric materials are also available as polymeric films such as polyvinylidene fluoride
(PVDF). These films are very thin, light weight and flexible and can be cut easily for adaptation touneven services. These films are not suitable for resonance applications due to the low mechanicalquality factor but are extensively used in acoustical broad band applications for microphones andloudspeakers.RL
R¢LCm
C¢mM
MCe
C¢eElectrical
ElectricalMechanical
MechanicalN:I
I:NCircuit A
Circuit B
Fig.3.4 Equivalent circuits for piezo-electric elements

78 Handbook of Biomedical Instrumentation
Piezo-electric transducers find numerous applications in the medical instrumentation field.
They are used in ultrasonic scanners for imaging and blood flow measurements. They are used inthe detection of Korotkoff sounds in non-invasive blood pressure measurements and in external
and internal phonocardiography. The details of their construction and associated requirements
and characteristics are given in the respective chapters.
/G33/G2E/G34/G2E/G37 /G4F/G74/G68/G65/G72/G20/G44/G69/G73/G70/G6C/G61/G63/G65 /G6D/G65/G6E/G74/G20/G53/G65/G6E/G73/G6F/G72/G73
The position and motion can be detected using optical transducers. Both amplitude and position
of the transmitted and reflected light can be used to measure displacement. An optical fibre canalso be used to detect displacement by measuring the transmitted light intensity or phase differencebetween the measuring beam and a reference beam.
Similarly, ultrasonic, microwave, X-rays, and nuclear radiations can be used to sense position.
These are covered in the respective chapters at other places in the book.
/G20/G33/G2E/G35 /G50/G52/G45/G53/G53/G55/G52/G45/G20/G54/G52/G41/G4E/G53/G44/G55/G43/G45/G52/G53
Pressure is a very valuable parameter in the medical field and therefore many devices have beendeveloped to effect its transduction to measurable electrical signals. The basic principle behind allthese pressure transducers is that the pressure to be measured is applied to a flexible diaphragmwhich gets deformed by the action of the pressure exerted on it. This motion of the diaphragm isthen measured in terms of an electrical signal. In its simplest form, a diaphragm is a thin flat plateof circular shape, attached firmly by its edge to the wall of a containing vessel. Typical diaphragmmaterials are stainless steel, phosphor bronze and beryllium copper.
Absolute pressure is pressure referred to a vacuum. Gauge pressure is pressure referred to
atmospheric pressure. The commonly used units for pressure are defined at 0°C as
P = 1 mm Hg = 1 torr = 12.9 mm blood = 13.1 mm saline = 13.6 mm H
2O = 133.0 dyn/cm2 =
1330 bar = 133.31 Pa = 0.133 kPa (Pa = Pascal)
For faithful reproduction of the pressure contours, the transducing system as a whole must
have a uniform frequency response at least up to the 20th harmonic of the fundamental frequencyof the signal. For blood pressure recording (which is at a rate of say 72 bpm or 1.2 Hz), the systemshould have a linear frequency response at least up to 30 Hz.
The most commonly employed pressure transducers which make use of the diaphragm are of
the following types:
•Capacitance manometer— in which the diaphragm forms one plate of a capacitor.
•Differential transformer —where the diaphragm is attached to the core of a differential
transformer.
•Strain gauge —where the strain gauge bridge is attached to the diaphragm.
Displacement transducers can be conveniently converted into pressure transducers by attaching
a diaphragm to the moving member of the transducer such that the pressure is applied to thediaphragm. The following pressure transducers are commercially available.

Physiological Transducers 79
/G33/G2E/G35/G2E/G31 /G4C/G56/G44/G54/G20/G50/G72/G65/G73/G73/G75/G72/G65/G20/G54/G72/G61/G6E/G73/G64/G75/G63/G65/G72
LVDT pressure transducer consists of three parts: a plastic dome with two female Luer-Lok fittings,
a stainless steel diaphragm and core assembly and a plastic body containing the LVDT coilassembly. The transducers are available commercially in a complete array of pressure ranges with
corresponding sensitivities, volume displacements and frequency response characteristics. LVDT
pressure transducers are available in two basic diaphragm and core assemblies. The first forvenous and general purpose clinical measurements has a standard-size diaphragm with internalfluid volume between the dome and diaphragm of less than 0.5 cc. The second design with higherresponse characteristics for arterial pressure contours has a reduced diaphragm area and internalvolume of approximately 0.1 cc. The gauge can be sterilized by ethylene oxide gas or cold liquidmethods. The LVDT gauges offer a linearity better than ±1%.
Arterial pressure transducers consist of very small differential transformers having tiny cores.
A typical commercially available transducer has an outside diameter of 3.2 mm and a length of9 mm, the movement of the movable core is ±0.5 mm and gives the full range of electrical output as125 mV/mm (62.5 mV in either direction). Baker et al (1960) designed a miniature differential
transformer pressure transducer which could be mounted directly on a dog’s heart. The size of thetransducer was 12.5 ¥ 6.25 mm and gave an average sensitivity of 38 mV per mmHg per volt
applied to the input coils. The zero drift has been measured at less than 0.2 mmHg/°C.
/G33/G2E/G35/G2E/G32 /G53/G74/G72/G61/G69/G6E/G20/G47/G61/G75/G67/G65/G20/G50/G72/G65/G73/G73/G75/G72/G65/G20/G54/G72/G61/G6E/G73/G64/G75/G63/G65/G72/G73
Nearly all commercially available pressure monitoring systems use the strain gauge type pressuretransducers for intra-arterial and intravenous pressure measurements. The transducer is basedupon the changes in resistance of a wire produced due to small mechanical displacements. Alinear relation exists between the deformation and electric resistance of a suitably selected gauge(wire, foil) over a specified range.
The figure of merit which describes the overall behaviour of the wire under stress is determined
from the “Gauge Factor” which is defined as
g=
D
DRL
LL/
/
where DR= Incremental change in resistance due to stress
R= Resistance of an unstretched wire
DL= Incremental change in length
L= Unstretched length of wire
Accordingly, the gauge factor gives information on the expected resistance change or output
signal at maximum permissible elongation. The gauge factor determines to a large extent thesensitivity of the wire when it is made into a practical strain gauge. The gauge factor varies withthe material. So, it is advisable to select a material with a high gauge factor. But the wire usedshould be selected for the minimum temperature coefficient of resistance.
Table 3.4 shows the gauge factor for different materials along with their temperature coefficient
of resistance.

80 Handbook of Biomedical Instrumentation
/G95/G20/G54/G61/G62/G6C/G65/G20/G33/G2E/G34 Gauge Factors for Different Materials
Material Gauge factor Temperature coefficient
of resistance ( W/W/°C)
Constantan 2.0 2 ¥ 10–6
Platinum 6.1 3 ¥ 10–3
Nickel 12.1 6 ¥ 10–3
Silicon 120 6 ¥ 10–3
From sensitivity considerations, semiconductor silicon seems to be the obvious choice but
it has been seen that it is highly temperature dependent. Techniques have, however, been
developed to partially compensate for this temperature effect by the use of thermistors and
combinations of suitably chosen p-type and n-type gauges. Silicon strain gauges can be made
to have either positive or negative gauge factors by selectively doping the material. In this way,for a strain which is only compressive, both positive and negative resistance changes can beproduced. Double sensitivity can be achieved by using two gauges of each type arranged inthe form of a bridge circuit and can be made to respond to strains of the same sign in each of thefour arms.
Unbonded Strain Gauges: Most of the pressure transducers for the direct measurement of blood
pressure are of the unbonded wire strain gauge type. The arrangement consists in mounting strainwires of two frames which may move with respect to each other.
The outer frame is fixed and the inner frame which is connected to the diaphragm upon which
the pressure acts, is movable. A pressure Papplied in the direction shown (Fig. 3.5) stretches wires
B and C and relaxes wires A and D. These wires form a four arm active bridge. The moving frameis mounted on springs which brings it to the central reference position when no pressure isapplied to the diaphragm. It has been observed that even after employing utmost care duringmanufacture, it is not possible to produce a zero output signal of the transducer at zero pressure.This may be due to the inhomogeneous character of the wire and the inevitable dimensionalinaccuracies of machining and assembly. Extra resistors are, therefore, placed in the gauge housing
Fixed
frameMovable
frameSapphire
pinsDiaphragm
PCBA
D
Fig.3.5 Schematic diagram of an unbonded strain gauge pressure transducer

Physiological Transducers 81
(Fig. 3.6) to adjust the same electrically. Zeroing of the bridge can be accomplished by a resistor Rx
connected in series thereto. Similarly, a series connected resistor Rt in the other arm of the bridge
serves to compensate for zero point drift, caused by temperature changes.
DE E= DR
R
R DR–R DR–
RD R+RD R+RxRt
ExcitationE
Fig.3.6 Resistors mounted in the pressure gauge housing to enable adjustments of
electrical zero at zero pressure conditions
The unbonded strain gauge transducers are preferred when low pressure measurements are to
be made since hysteresis errors are much less than would be the case if wire gauges were bound tothe diaphragm. Unbonded strain gauge transducers can be made sufficiently small, which areeven suitable for mounting at the tip of a cardiac catheter.
Bonded Strain Gauges: The bonded strain gauge consists of strain-sensitive gauges which are
firmly bonded with an adhesive to the membrane or diaphragm whose movement is to be recorded.In practice, it is made by taking a length of very thin wire (for example, 0.025 mm dia) or foil whichis formed into a grid pattern (Fig. 3.7 a,b) and bonded to a backing material. The backing materialcommonly used is paper, bakelized paper or and similar material. For pressure measurementpurposes, a strain gauge constructed as above, is attached to a diaphragm. The deflection of thediaphragm under pressure causes a corresponding strain in the wire gauge. Since the deflection isproportional to pressure, a direct pressure resistance relation results.
By using a pair of strain gauges and mounting them one above the other, the changes in the
resistances of the two gauges arising from the changes in the ambient temperature can be cancelled.Also, one strain gauge would increase while the other would decrease in resistance when thepressure is applied to them.

82 Handbook of Biomedical Instrumentation
Silicon Bonded Strain Gauges: In recent years, there has been an increasing tendency to use bonded
gauges made from a silicon semiconductor instead of from bonded wire or foil strain gauges.This is because of its higher gauge factor resulting in a greater sensitivity and potential for
miniaturization.
Figure 3.7(c) shows an arrangement in which
the positive-doped ( p-doped) silicon elements
of a Wheatstone bridge are diffused directlyon to a base of negative-doped ( n-doped) silicon.
Although, semiconductor strain gauges arevery sensitive to variations in temperature, theinclusion of eight elements to form all fourresistive arms of a bridge eliminates this problemby exposing all of the elements to the sametemperature.
In the bonded silicon semiconductor strain
gauge, the conventional wire or foil element isreplaced by a single chip of silicon, processed toa finished size less than 0.0125 mm thick and0.25 mm wide. The strip is then mounted on a
substrate of epoxy resin and glass fibre with
nickel strips for the electrical connections. Goldwires are used for interconnection betweensilicon and nickel. The complete assembly iscoated with epoxy resin to protect it fromSoldering
tagsSoldering
tags
Wire
Felt Foil
Backing
(a) (b)
Fig.3.7(a) Wire strain gauge a bonded type strain guage pressure tranducer
(b) Etched foil strain gauge
(+)
(–) (–)(–)
(–)(+)
(+)(+)
R1S1S2Q1
Q2T1
T2
T2Q1
Q2
R2
R1S1
S2T1R2P2P1SiliconClamp
Diffused
regionPn-type Si
plane
Fig.3.7(c) Diffused p-type strain gauge

Physiological Transducers 83
environmental conditions. Lead wire resistance and capacitance also change with temperature.
Therefore, lead wires from the strain gauge to the Wheatstone bridge must also be temperaturecompensated. Compensation for temperature variation in the leads can be provided by using the
three lead method. In this method, two of the leads are in adjacent legs of the bridge which cancels
their resistance changes and does not disturb the bridge balance. The third lead is in series withthe power supply and is, therefore, independent of bridge balance.
The bridge power supply is regulated and temperature compensated. The piezo-resistive
elements are the four arms of a Wheatstone bridge and consist of four p-doped (boron) regions,
diffused into the edged chip of the n-type silicon. The transducer has a range of 0–1 atmosphere.
It has hysteresis error of ±0.1% full scale and a temperature coefficient of ±2–3 mV/°C. It gives an
output of ±0.75 V per 100 mmHg. The high output permits the use of this transducer with practicallyany dc recorder.
/G20/G33/G2E/G36 /G54/G52/G41/G4E/G53/G44/G55/G43/G45/G52/G53/G20/G46/G4F/G52/G20/G42/G4F/G44/G59/G20/G54/G45/G4D/G50/G45/G52/G41/G54/G55/G52/G45/G20/G4D/G45/G41/G53/G55/G52/G45/G4D/G45/G4E/G54
The most popular method of measuring temperature is by using a mercury-in-glass thermometer.However, they are slow, difficult to read and susceptible to contamination. Also, reliable accuracycannot be attained by these thermometers, especially over the wide range which is now found to be
necessary. In many of the circumstances of lowered body temperature, continuous or frequent
sampling of temperature is desirable, as in the operating theatre, post-operative recovery roomand intensive care unit, and during forced diuresis, massive blood transfusion, and accidentalhypothermia. The continuous reading facility of electronic thermometers obviously lends itself tosuch applications. Electronic thermometers are convenient, reliable and generally more accuratein practice than mercury-in-glass thermometers for medical applications. They mostly use probesincorporating a thermistor or thermocouple sensor which have rapid response characteristics.The probes are generally reusable and their covers are disposable.
Small thermistor probes may be used for oesophageal, rectal, cutaneous, subcutaneous,
intramuscular and intravenous measurements and in cardiac catheters. Thermocouples arenormally used for measurement of surface skin temperature, but rectal thermocouple probesare also available. Resistance thermometers are usually used for rectal and body temperaturemeasurement. The resistance thermometer and thermistor measure absolute temperature, whereasthermocouples generally measure relative temperature.
/G33/G2E/G36/G2E/G31 /G54/G68/G65/G72 /G6D/G6F/G63/G6F/G75/G70/G6C/G65/G73
When two wires of different materials are joined together at either end, forming two junctions
which are maintained at different temperatures, a thermo-electromotive force (emf) is generatedcausing a current to flow around the circuit. This arrangement is called a thermocouple. Thejunction at the higher temperature is termed the hot or measuring junction and that at the lowertemperature the cold or reference junction. The cold junction is usually kept at 0°C. Over a limitedrange of temperature, the thermal emf and hence the current produced is proportional to the
temperature difference existing between the junctions. Therefore, we have a basis of temperature

84 Handbook of Biomedical Instrumentation
measurement, since by inserting one junction in or on the surface of the medium whose temperature
is to be measured and keeping the other at a lower and constant temperature (usually O°C), ameasurable emf is produced proportional to the temperature difference between the two junctions.
The reference junction is normally held at O°C inside a vacuum flask containing melting ice.
The amount of voltage change per degree of temperature change of the junction varies with the
kinds of metals making up the junction. The voltage sensitivities of thermocouples made of various
metals are given in Table 3.5.
/G95/G20/G54/G61/G62/G6C/G65/G20/G33/G2E/G35 Thermal emf for Various Types of Thermocouples
Useful range Sensitivity at 20°C
Type Thermocouple °C (m V/°C)
T Copper-constantan – 150 to + 350 45
J Iron-constantan – 150 to + 1000 52
K Chromel-alumel – 200 to + 1200 40
S Platinum-platinum (90%) 0 to + 1500 6.4
rhodium (10%)
Figure 3.8 illustrates the emf output versus temperature in degrees for each of the commonly
used thermocouples. Two important facts emerge from this graph: (i) the sensitivity (slope) of eachcurve is different, and (ii) none of the curves has a perfectly linear rate of change of emf output perdegree F. This shows that each type of thermocouple has a unique, non-linear response to
temperature. Since most recording or display devices are linear, there is a need to linearize theoutput from the thermocouples so that the displayed output gets correlated with actualtemperature. The sensitivity of a thermocouple does not depend upon the size of the junction or thewires forming it as the contact potentials developed are related to the difference in the workfunction of the two metals.
For medical applications, a copper-constantan combination is usually preferred. With the
reference junction at 0°C and the other at 37.5°C, the output from this thermocouple is 1.5 mV.Two types of measuring instruments can be used with thermocouples to measure potentialdifferences of this order. In one, moving coil movements are used as millivoltmeters to measure thethermocouple emf. They are directly calibrated in temperature units. Usually in clinical thermo-couple instruments, reflecting galvanometers or light spot galvanometers are preferred to measure
and display temperature. If the thermocouple voltages are small (less than 1 mv), they can be
readily measured with a precision dc potentiometer having a Weston cadmium-mercury cell as areference. They can also be read directly on a digital voltmeter or by using a chopper stabilized dcamplifier followed by a panel meter of the analog or digital type.
Much experimentation with thermocouple circuits has led to the formulation of the following
empirical laws which are fundamental to the accurate measurement of temperature by thermo-electric means: (i) the algebraic sum of thermoelectric emf’s generated in any given circuitcontaining any number of dissimilar homogeneous metals is a function only of the temperaturesof the junctions, and (ii) if all but one of the junctions in such a circuit are maintained at some

Physiological Transducers 85
reference temperature, the emf generated depends only upon the temperature of that junction and
can be used as a measure of temperature. It is thus evident that the junction temperature can bedetermined if the reference junction is at a different, but known temperature. It is also feasible inthe use of oven-controlled and electrically simulated reference junctions.
The temperature of the measuring junction can be determined from the thermo emf only if the
absolute temperature of the reference junction is known. This can be done by either of the followingmethods:
1. By measuring the reference temperature with a standard direct reading thermometer.
2. By containing the reference junction in a bath of well defined temperature, e.g. a carefully
constructed and used ice bath which can give an accuracy of 0.05°C with a reproducibilityof 0.001°C.
A simpler method is to use a reference-temperature compensator which generates an emf that
will exactly compensate for variations in the reference-junction temperature. Figure 3.9 illustrates
a thermally sensitive bridge which is designed to generate an emf that varies with the enclosure
temperature (normally ambient) in such a way that variations in the cold junction are nullified.0102040
3050607080
1000 0 2000 3000JK
T
S
Temperature, degrees FOutput millivolts
Fig.3.8 EMF versus temperature characteristics of major types of thermocouples
(Courtesy: Gould Inc., USA)

86 Handbook of Biomedical Instrumentation
In this circuit, R2is a temperature-sensitive component that is thermally bonded to the cold
junction thermocouple. The resistance-temperature curve of R2 matches the emf temperature
characteristic of the thermocoupled material. The voltage change across R2 is equal and opposite
to the cold junction thermal volt-age over a limited ambient temperature range. This systemintroduces errors if the enclosure temperature undergoes wide variations. How-ever, for moderatefluctuations (±5°C), it enables an effective reference temperature stability of ±0.2°C to be achieved.This type of reference temperature simulation is well adapted to zero suppression of a large
temperature. By specifying the reference temperature of or near the temperature to be recorded,
small temperature changes can be recorded.
Precise, easily calibrated, integrated circuit (LM 135 series from National Semiconductors)
temperature sensors can be conveniently used in cold junction compensation circuits. They aresuitable in compensation circuits over a –55° to + 150°C temperature range.
The small size and very fast response of thermocouples make them suitable for intracellular
transient temperature measurements and for measuring temperatures from within the body atsites like the oesophagus, rectum, etc. They can be inserted into catheters and hypodermic needles.Special needles are commercially available that enable a thermocouple to be subcutaneouslyimplanted, the needle being withdrawn to leave the thermocouple in place. Mekjavic et al (1984)
have illustrated the construction and evaluation of a thermocouple probe for measuringoesophageal temperature.
Cain and Welch (1974a) developed thin film Cu-Ni micro-thermocouple probes for dynamic as
well as static temperature measurements in biological tissue. These probes which use a quartzsubstrate exhibit response times of less than a millisecond with thermal properties similar totissues. Their thermoelectric emf of 21 mmV/°C is linearly dependent on temperature over the
range normally encountered in biological measurements. Probe tip diameters as small as 10 and30mm have been fabricated. These small sizes enable the measurement of temperature profiles of
the retina of the eye. These probes have been experimentally inserted through the sclera of the eye
and kept in contact with the biological tissue for periods of 4 to 6 hours. Their linearity, thermalemf and response time make it possible to record the response to laser irradiation for temperaturerises from 1 to 45°C and from 1 ms to 10s (Cain and Welch, 1974b).OutputCold junction
Measure
junction+V
–VR1
R2R3
R4
Fig.3.9 Bridge type reference j unction com pensator

Physiological Transducers 87
/G33/G2E/G36/G2E/G32 /G45/G6C/G65/G63/G74/G72/G69/G63/G61/G6C/G20/G52/G65/G73/G69/G73/G74/G61/G6E/G63/G65/G20/G54/G68/G65/G72 /G6D/G6F/G6D/G65/G74/G65/G72
The temperature dependence of resistance of certain metals makes it convenient to construct a
temperature transducer. Although most of the metals can be used, the choice, however, dependsupon the linearity and sensitivity of the temperature resistance characteristics. The resistance
R
tof a metallic conductor at any temperature t is given by:
Rt= Ro(1 +at)
where Ro= resistance at 0°C and
a= temperature coefficient of resistivity
IfRo and a are known, a measurement of Rtshall directly give the value of temperature. The
increase in resistance is linear over the range 0–100°C for commonly used materials in resistance
thermometry.
Normally platinum or nickel are used for resistance thermometry since they can be readily
obtained in a pure form and are comparatively stable. Thermometers constructed from a coil ofthese metals have been used for the measurement of skin, rectal and oesophageal temperature. Thecoefficient of resistivity for platinum is 0.004 W/W °C. The measurement of resistance is generally
carried out by using a Wheatstone bridge in which all leads must be of constantan to keeptheir own temperature resistance changes minimal. The comparator resistances must also betemperature stable. The bridge can be operated from either direct or alternating current, but toneglect any electrochemical changes or polarization in the circuit, an ac bridge is recommended.
Figure 3.10 shows the simple bridge circuit employed in resistance thermometry. A and B
are fixed resistances. C is a variable resistance made from constantan which has a very low
temperature coefficient of resistance. The measuring coil and its connecting leads are placed inone arm of the bridge circuit with a dummy pair of leads connected in the opposite arm. In thismanner, changes in resistance of the coil leads with ambient temperature are cancelled out by the
corresponding changes in the dummy or compensating leads.
The simple Wheatstone bridge circuit is basically a non-linear device when it is operated away
from its null-balance point; therefore, it is important to understand the degree of non-linearity to be
AB
CRatio arms
Compensating
leadsMeasuring coil
(Pt. or nickel)+
–
Fig.3.10 Circuit arrangement of a metal resistance thermometer

88 Handbook of Biomedical Instrumentation
encountered. Platinum resistance sensors are linear within ±10% between –200 and +600°C. To
achieve better linearization, Foster (1974) describes a circuit which makes use of a nonlinearamplifier. Positive feedback is added around the input amplifier, making its effective gain increase
as the sensor resistance increases with temperature. The linearization achieved is typically
accurate within ±0.5°C in the worst case over the specified range. The analog linearization methodhas the advantage that there are no discontinuities in the displayed temperature as the temperaturechanges. This is sometimes a problem when piece-wise linear compensation networks or readonly memories are used for linearization.
Resistance thermometers are particularly suitable for remote reading and can be made of inert
materials like platinum. They are very stable and show almost no hysteresis with large temperatureexcursions. However, the size of the coils used presently is such that in the medical field, theresistance thermometer probes are best suited only for rectal probe use. They are bulky and are notsuitable for mounting in needles for the measurement of tissue temperature.
/G33/G2E/G36/G2E/G33 /G54/G48/G45/G52/G4D/G49/G53/G54/G4F/G52/G53
Thermistors are the oxides of certain metals like manganese, cobalt and nickel which have largenegative temperature coefficient of resistance, i.e. resistance of the thermistor shows a fall with
increase in temperature. The general resistance-temperature relation for a thermistor is given by:
R= A e
BIT
where R= resistance of the thermistor in W
T= absolute temperatue
A and B are constants
Thermistors when used for measuring temperature have many advantages over thermocouples
and resistance thermometers. They are summarized as follows:
• They show a considerably high sensitivity—about 4% change in resistance per degree
centigrade.
• Since the thermistors themselves are of high resistance, the resistances of the connecting
leads are of small influence and therefore, no compensating leads are necessary.
• The time constant can be made quite small by easily reducing the mass of the thermistor.
Hence the measurements can be taken rather quickly.
• They can be had in a variety of shapes making them suitable for all types of applications.• They are small in size—about the size of the head of a pin and can be mounted on a
catheter or on hypodermic needles.
• Since the thermistors are available in a large range of resistance values, it makes it much
easier to match them in the circuits.
The large change in resistance with temperature means that a comparatively simple bridge
circuit is sufficient. The thermistors have inherently non-linear resistance-temperature characteristics.But by the proper selection of the values of the bridge resistances (bridge in itself is a non-lineararrangement), it is possible to get nearly linear calibration of the indicating instrument over alimited range.

Physiological Transducers 89
There is drawback in using thermistors when multichannel temperature indicators are made.
This is due to the fact that it is not possible to make a batch of them with uniform characteristics.The normal tolerance on resistance is ±20% but matching down to ±1% is available at higher cost.
Recalibration of the instrument is thus required whenever a particular probe has to be replaced.
For linearization of thermistors over a limited temperature range (Sapoff, 1982), two approaches
are usually employed. Figure 3.11(a) and (b) show these techniques, which are:
(i) If the thermistor is supplied with a constant current and the voltage across the thermistor
is used to indicate the temperature, linearization can be obtained by shunting the ther-mistor with a selected resistor R
P. The objective is to make the point of inflection of the
parallel combination coincide with the mid-scale temperature (Fig. 3.11(a)).
(ii)When the current through the thermistor, for a fixed applied difference, is used to indicate
the temperature, the series arrangement is employed (Fig. 3.11(b)).
RP
RP
TiRTRT
TemperatureResistanceR
(a)
TiGT
–GSGS
TemperatureConductor GRT=1
GT
(b)
Fig.3.11(a) Two schemes for linearization of thermistor for a short range and their
characteristics. (a) Parallel combination of a thermistor and a resistor
(b) Series combination of as thermistor and a resistor
By using this technique, the maximum deviation from linearity observed is 0.03°C for every
10°C. It may, however, be noted that improved linearity achieved results in a decrease in the

90 Handbook of Biomedical Instrumentation
effective temperature coefficient of the combination. For example, the temperature coefficient is
3.3%°C for a thermistor, when optimally linearized with a series resistance. More complex circuitarrangements are needed to achieve better linearization over wider ranges of temperature.
Cronwell (1965) has presented a review of the techniques used for linearizing the thermistors.
Thermistors which provide a linear change in resistance with a change in temperature have been
given the name ‘Linistors’.
An operational amplifier circuit designed by Stockert and Nave (1974) provides a linear relation
between the output voltage and temperature from 10 to 50°C using a non-linear thermistor
as the temperature transducer. The paper discusses the method for calculating the circuit values.Allen (1978) suggests the use of a microprocessor as a ‘look-up’ table of thermistor resistance-temperature values, using linear interpolation between the values to determine the temperature.
Thermistors are made in a wide variety of forms suitable for use in medical applications. They
are available as wafers required for applying to the skin surface, rods which can be used for rectal,oral, or similar insertions, and tiny beads sosmall that they can be mounted at the tip of ahypodermic needle for insertion into tissues.These tiny thermistors have very small thermal‘time constants’ when they are properly mount-ed. ‘Time constant’, the standard measure ofthermistor probe response time, is the timerequired for a probe to read 63% of a newly
impressed temperature. Generally, the time
constants are obtained by transferring theprobe from a well stirred water bath at 20°C toa like bath at 43°C. Approximately, five ‘timeconstants’ are required for a probe to read 99%of the total change.
Thermistors with positive thermo-resistive
coefficients are called Posistors (PTC). They ex-hibit a remarkably high variation in resistancewith increase in temperature. The posistors aremade from barium titanate ceramic.
PTCs are characterized by an extremely large
resistance change in a small temperature span.Figure 3.12 shows a generalized resistancetemperature plot for a PTC (Krelner, 1977). Thetemperature at which the resistance begins to
increase rapidly is referred to as the switching
temperature. This point can be changed frombelow 0°C to above 120°C. Thus, PTCs have anearly constant resistance at temperaturesbelow the switching temperature but show arapid increase in resistance at temperatures above the switch temperature.
30 0 60 90 120 150 180 2101.010102103104
0.1
Temperature °C TsResistance (ohms)
Fig.3.12 A PTC thermistor switches a bruptly
when heated to the switching tem-
perature, resistance changes by fourdecades (Courtesy: Keystone CarbonCo., USA)

Physiological Transducers 91
Thermistor probes for measurement of body temperature consist of thermistor beads sealed into
the tip of a glass tube. The bead is protected by this glass housing and the rapid response is alsomaintained. Probes suitable for rectal and oesophageal applications usually contain a thermistor
bead mounted inside a stainless steel sheath. Figure 3.13 shows a variety of probes suitable for
different applications in the medical field.
Fig.3.13 Different shapes of thermistor probes (Courtesy: Yellow Springs
Instruments Co., USA)
Probes should be sterilized by using a chemical solution such as 70% alcohol or Dakin’s
solution (sodium hypochlorite in neutral buffer). The cleaning agents Zephiran and Haemo-solare likewise suitable. Probes should not be boiled or actoclaved except where it is specificallymentioned.
The instruments which give direct reading of the temperature at the thermistor position are
known by the name telethermometers because of their ability to use leads which are hundreds offeet long without a significant decrease in accuracy. The continuous signal is also suitable forrecording without amplification. Figure 3.14 is a typical circuit diagram of a telethermometer. The
instrument works on a 1.5 V dry cell which has an operating life of 1000 h. The addition of a
second thermistor in the opposite arm of the bridge can make the circuit twice as sensitive andpermits the use of a lower sensitivity meter.
For specific applications like pyrogenic studies, it is often necessary to have a multichannel
system, which should automatically scan 3, 7 or 11 probes in sequence with 20 s, 1 min. or 5 min.readings per probe and record the measured readings on a recorder. For this, a motorized rotaryswitch can be employed. The recorder output in these instruments is 0 to 100 mV with zero outputat the highest temperature.

92 Handbook of Biomedical Instrumentation
Probe
R9R4R7
R3(Calibration mark)On1.5 V
Dry cell
R6
R1
R11
R8R2R5
MThermistor Balance
Recorder
Fig.3.14 Circuit diagram of a telethermometer (Courtesy: Yellow Springs
Instruments Co. USA)
/G33/G2E/G36/G2E/G34 /G52/G61/G64/G69/G61/G74/G69/G6F/G6E/G20/G54/G68/G65/G72 /G6D/G6F/G6D/G65/G74/G65/G72/G79
Any material placed above absolute zero temperature emits electromagnetic radiation from its
surface. Both the amplitude and frequency of the emitted radiation depends on the temperature ofthe object. The cooler the object, the lower the frequency of the emitted electromagnetic waves, and
lesses the power emitted. The temperature of the object can be determined from the power emitted.
Infrared thermometers measure (Ring, 1988) the magnitude of infrared power (flux) in a broadspectral range, typically from 4 to 14 micrometers. They make no contact with the object measured.The detectors used for measuring the emitted infrared radiation are thermopiles (series connectedthermocouple pairs), pyroelectric sensors, Golay cells, photoconductive cells and photovoltaiccells. All these devices can be mounted in specially designed housing to measure surfacetemperature without direct contact with the body. Table 3.6 summarizes salient specifications forvarious types of temperature measuring devices.
Hand-held infrared scanners are now available for monitoring the pattern of skin temperature
changes, particularly for tympanic membrane temperature measurements. This measurement isbased on the development of a new type of sensor called “Pyroelectric Sensor”.
A pyroelectric sensor develops an electric charge that is a function of its temperature gradient.
The sensor contains a crystalline flake which is preprocessed to orient its polarized crystals.Temperature variation from infrared light striking the crystal changes the crystalline orientation,resulting in development of an electric charge. The charge creates a current which can be accuratelymeasured and related to the temperature of the tympanic membrane.
Infrared thermometers have significant advantages over both glass and thermistor thermometers
used orally, rectally or axillarily. They eliminate reliance on conduction and instead measure the

Physiological Transducers 93
body’s natural radiation. They use an ideal measurement site—the tympanic membrane of the ear
which is a function of the core body temperature. It is a dry, non-mucous membrane site thatminimizes risks of cross-contamination.The disadvantage of infrared thermometers is their highcost as compared to other types of thermometers.
/G33/G2E/G36/G2E/G35 /G53/G69/G6C/G69/G63/G6F/G6E/G20/G44/G69/G6F/G64/G65
The voltage drop across a forward biased silicon diode is known to vary at the rate of 2 mV/°C. This
suggests that a silicon diode can be used as a temperature sensor. Griffths and Hill (1969) describe
the technique and circuit diagram for the measurement of temperature using a silicon diode.
The circuit (Fig. 3.15) is designed to monitor body temperature in the range of 34–40°C with an
accuracy of 2.5%. The temperature-sensing diode D1 is connected to the non-inverting input of an
operational amplifier. The gain of the amplifier is 500. With D1 mounted in a water bath at 37°C, R1
is adjusted to give an amplifier output voltage of zero. Using a centre zero meter, the scale can be
calibrated to ±3°C.
+12 V
–12 VD1
R110 K
18 K 1 K500 K2.2 KOutput
Fig.3.15 Use of a silicon diode as a temperature sensor (redrawn after Griffths and
Hill, 1969)/G95/G20/G54/G61/G62/G6C/G65/G20/G33/G2E/G36 Comparison of Electrical Temperature-Sensing Techniques
RTD Thermistor Thermocouple Radiation
Accuracy 0.01° to 0.1°F 0.01° to 1°F 1° to 10°F 0.2°F
Stability Less than 0.1% 0.2°F drift/year 1° drift/year Same as
drift in 5 years thermocouple
Sensitivity 0.1 to 10 ohms/°F 50 to 500 ohms/F° 50 to 500 m volts/°F Same as
thermo-couple
Features Greatest accuracy Greatest Greatest economy; Fastest response;
over wide spans; sensitivity highest range no contaminations;
greatest stability easiest to use

94 Handbook of Biomedical Instrumentation
The disadvantage of using a diode as a temperature sensor is the requirement of a stable
calibration source. Soderquist and Simmons (1979) explain the use of a matched transistor pairwhich has predictable differential base-emitter voltage relationship which can be exploited as a
temperature sensor. It is advantageous to have a diode or transistor sensor fabricated on a chip
with interfacing electronics by integrated circuit (IC) technology. Several integrated temperaturesensors have been developed and some of these are available commercially. AD 599 is an integratedcircuit temperature transducer manufactured by Analog Devices. The transducer produces anoutput proportional to absolute temperature. For supply voltages between +4 and +30 V, it acts asa high impedance, constant-current regulator passing 1 mA/K. The transducer uses a fundamental
property of the silicon transistor from which it is made to realise its temperature proportionalcharacteristics.
/G33/G2E/G36/G2E/G36 /G43/G68/G65 /G6D/G69/G63/G61/G6C/G20/G54/G68/G65/G72 /G6D/G6F/G6D/G65/G74/G65/G72/G79
Liquid crystals which are commonly used for digital displays are composed of materials which
have liquid like mechanical properties but possess the optical properties of single crystals. Thesecrystals demonstrate remarkable changes in their optical properties when temperature is varied.The change in the colour of the crystal related to temperature is used to measure the surfacetemperature. A liquid crystal film encapsulated in a plastic housing can be attached to thebody surface. Some systems use flexible liquid crystal plates so that better contact over thebody surface can be achieved. These are single use thermometers and there is less likelihood ofcross-contamination.
The clinical use of liquid crystal thermometery has not become popular because of its low
thermal resolution (± 0.5°C) and slow response time (> 60 sec.). In addition, contact thermometerymight modify the temperature of the measured skin due to undue pressure over the skin surface.
/G20/G33/G2E/G37 /G50/G48/G4F/G54/G4F/G45/G4C/G45/G43/G54/G52/G49/G43/G20/G54/G52/G41/G4E/G53/G44/G55/G43/G45/G52/G53
Photoelectric transducers are based on the principle of conversion of light energy into electricalenergy. This is done by causing the radiation to fall on a photosensitive element and measuringthe electrical current so generated with a sensitive galvanometer directly or after suitableamplification. There are two types of photoelectric cells—photovoltaic cells and photomissivecells.
/G33/G2E/G37/G2E/G31 /G50/G68/G6F/G74/G6F/G76/G6F/G6C/G74/G61/G69/G63/G20/G6F/G72/G20/G42/G61/G72/G72/G69/G65/G72/G20/G4C/G61/G79/G65/G72/G20/G43/G65/G6C/G6C/G73
Photovoltaic or barrier layer cells usually consist of a semiconducting substance, which isgenerally selenium deposited on a metal base which may be iron and which acts as one of theelectrodes. The semiconducting substance is covered with a thin layer of silver or gold depositedby cathodic deposition in vacuum. This layer acts as a collecting electrode. Figure 3.16 shows theconstruction of the barrier layer cell. When radiant energy falls upon the semiconductor surface, itexcites the electrons at the silver-selenium interface. The electrons are thus released and collectedat the collector electrode.

Physiological Transducers 95
Selenium
IronCollector electrode
Lacquer
GlassNegative
contact
strip
Plastic
case
Spring contact
for +ve terminal
Fig.3.16 Construction of a barrier layer cell
The cell is enclosed in a housing of insulating material and covered with a sheet of glass. The
two electrodes are connected to two terminals which connect the cell with other parts of theelectrical circuit.
Photovoltaic cells are very robust in construction, need no external electrical supply and
produce a photocurrent sometimes stronger than other photosensitive elements. Typicalphotocurrents produced by these cells are as high as 120 mA/lumen. At constant temperature, the
current set up in the cell usually shows a linear relationship with the incident light intensity.Selenium photocells have very low internal resistance, and therefore, it is difficult to amplify thecurrent they produce by dc amplifiers. The currents are usually measured directly by connectingthe terminals of the cell to a very sensitive galvanometer.
Selenium cells are sensitive to almost the entire range of wavelengths of the spectrum. However,
their sensitivity is greater within the visible spectrum and highest in the zones near to the yellowwavelengths. Figure 3.17 shows spectral response of the selenium photocell and the human eye.
Selenium cells have a high temperature coefficient and therefore, it is very necessary to allow
the instrument to warm up before the readings are commenced. They also show fatigue effects.
When illuminated, the photocurrent rises to a value several percent above the equilibrium value
and then falls off gradually. When connected in the optical path of the light rays, care should betaken to block all external light and to see that only the light from the source reaches the cell.
Selenium cells are not suitable for operations in instruments where the levels of illumination
change rapidly, because they fail to respond immediately to those changes. They are thus notsuitable where mechanical choppers are used to interrupt light 15–60 times a second.
/G33/G2E/G37/G2E/G32 /G50/G68/G6F/G74/G6F/G65 /G6D/G69/G73/G73/G69/G76/G65/G20/G43/G65/G6C/G6C/G73
Photoemissive cells are of three types: (a) high vacuum photocells, (b) gas-filled photocells and (c)
photomultiplier tubes. All of these types differ from selenium cells in that they require an external

96 Handbook of Biomedical Instrumentation
power supply to provide a sufficient potential difference between the electrodes to facilitate the
flow of electrons generated at the photosensitive cathode surface. Also, amplifier circuits areinvariably employed for the amplification of this current.
High Vacuum Photoemissive Cells: The vacuum photocell consists of two electrodes—the cathode
having a photosensitive layer of metallic cesium deposited on a base of silver oxide and the anodewhich is either an axially centered wire or a rectangular wire that frames the cathode. Theconstruction of the anode is such that no shadow falls on the cathode. The two electrodes aresealed within an evacuated glass envelope.
When a beam of light falls on the surface of the cathode, electrons are released from it, which
are drawn towards the anode which is maintained at a certain positive potential. This givesrise to a photocurrent which can be measured in the external circuit. The spectral response of aphotoemissive tube depends on the nature of the substance coating the cathode, and can be variedby using different metals or by variation in the method of preparation of the cathode surface.Cesium-silver oxide cells are sensitive to the near infrared wavelengths. Similarly, potassium-silver oxide and cesium-antimony cells have maximum sensitivity in the visible and ultravioletregions. The spectral response also depends partly on the transparency to different wavelengthsof the medium to be traversed by the light before reaching the cathode. For example, the sensitively
of the cell in the ultraviolet region is limited by the transparency of the wall of the envelope. For this
region, the use of quartz material can be avoided by using a fluorescent material, like sodiumsalicylate, which when applied to the outside of the photocell, transforms the ultraviolet intovisible radiations.+Human eye
Photocell response
300 400 500 600 700 800020406080100
Spectral response
Wavrlength m m
Fig.3.17 Spectral response of a human eye and a selenium photocellWavelength m m

Physiological Transducers 97
Figure 3.18 shows the current-voltage characteristics of a vacuum photoemissive tube at
different levels of light flux. They show that as the voltage increases, a point is reached where allthe photoelectrons are swept to the anode as soon as they are released which results in a saturation
photocurrent. It is not desirable to apply very high voltages, as they would result in a excessive
dark current without any gain in response.
50 0 100
Anode voltage (V)150 2001234
Light flux 0.12 lumens
0.020.040.060.080.10
Anode current am
Fig.3.18 Current voltage characteristics of a vacuum photoemissive tube at
different levels of light flux
Figure 3.19 shows a typical circuit configuration usually employed with photoemissive tubes.
Large values of phototube load resistor are employed to increase the sensitivity up to the practical
limit. Load resistances as high as 10,000 M W have been used. This, however, almost puts a limit,
as further increase of sensitivity induces difficulties in the form of noise, non-linearity and slowresponse. At these high values of load resistors, it is very essential to shield the circuit frommoisture and electrostatic effects. Therefore, special type of electrometer tubes, carefully shieldedand with a grid cap input are employed in the first stage of the amplifier.
Gas-filled Photoemissive Cells: This type of cell contains small quantities of an inert gas like
argon, whose molecules can be ionized when the electrons present in the cell posses sufficientenergy. The presence of small quantities of this gas prevent the phenomenon of saturation current,when higher potential differences are applied between the cathode and anode. Due to repeatedcollisions of electrons in the gas-filled tubes, the photoelectric current produced is greater even atlow potentials.
Photomultiplier Tubes: Photomultiplier tubes are used as detectors when it is required to detect
very weak light intensities. The tube consists of a photosensitive cathode and has multiple cascade

98 Handbook of Biomedical Instrumentation
stages of electron amplification in order to achieve a large amplification of the primary p hoto-
current within the envelope of the phototube itself. The electrons generated at the photocathode areattracted by the first electrode, called dynode, which gives out secondary electrons. There may be9–16 dynodes (Fig. 3.20). The dynode consists of a plate of material coated with a substancehaving a small force of attraction for the escaping electrons. Each impinging electron dislodgessecondary electrons from the dynode. Under the influence of positive potential, these electrons are
accelerated to the second dynode and so on. This process is repeated at the successive dynode,
which are operated at voltages increasing in steps of 50–100 V. These electrons are finally collectedat the collector electrode.
DynodesIncident lightPhoto
cathode Anode
RM Out
Fig.3.20 Principle of photomultiplier tube
The sensitivity of the photomultiplier tube can be varied by regulating the voltage of the first
amplifying stage. Because of the relatively small potential difference between the two electrodes,the response is linear. The output of the photomultiplier tube is limited to an anode current of a fewmilliamperes. Consequently, only low intensity radiant energy can be measured without causingany appreciable heating effect on the electrodes surface. They can measure light intensities aboutLight
sourceCollector electrode
(anode) Photocathode
RL
Fig.3.19 Typical circuit configuration employed with photoemissive tubes

Physiological Transducers 99
107times weaker than those measurable with an ordinary phototube. For this reason, they should
be carefully shielded from stray light. The tube is fairly fast in response to the extent that they areused in scintillation counters, where light pulses as brief as 10
–9 s duration are encountered. A
direct current power supply is required to operate a photomultiplier, the stability of which must be
at least one order of magnitude better than the desired precision of measurement; for example, toattain precision of 1%, fluctuation of the stabilized voltage must not exceed 0.1%.
Fatigue and saturation can occur at high illumination levels. The devices are sensitive to
electromagnetic interference and they are also more costly than other photoelectric sensors.Photomultipliers are not uniformly sensitive over the whole spectrum and in practice, manu-facturers incorporate units best suited for the frequency range for which the instrument is designed.In the case of spectrophotometers, the photomultipliers normally supplied cover the range of 185to 650 nm. For measurements at long wavelengths, special red sensitive tubes are offered. Theycover a spectral range from 185 to 850 nm but are noticeably less sensitive at wavelengths below450 nm than the standard photomultipliers.
Photomultiplier tubes may be damaged if excessive current is drawn from the final anode. Since
accidental overload may easily occur in a laboratory and tubes are expensive to replace, it isadvisable to adopt some means of protection from overloads.
/G33/G2E/G37/G2E/G33 /G53/G69/G6C/G69/G63/G6F/G6E/G20/G44/G69/G6F/G64/G65/G20/G44/G65/G74/G65/G63/G74/G6F/G72/G73
The photomultiplier which is large and expensive and requires a source of stablized high voltagecan be replaced by a silicon diode detector (photodiode, e.g. H.P.5082–4220). This diode is useablewithin a spectral range of 0.4–1.0 mm, in a number of instruments (spectrophotometers, flame
photometers). The photodiode can be powered from a low voltage source.
These detectors when integrated with an operational amplifier have performance charac-
teristics which compare with those of a photomultiplier over a similar wavelength range.Figure 3.21 shows the spectral response of silicon diode detectors. The devices being solid state aremechanically robust and consume much less power. Dark current output and noise levels aresuch that they can be used over a much greater dynamic range.
/G33/G2E/G37/G2E/G34 /G44/G69/G6F/G64/G65/G20/G41/G72/G72/G61/G79/G73
Diode arrays are assemblies of individual detector elements in linear or matrix form, which in aspectrophotometer can be mounted so that the complete spectrum is focussed on to an array ofappropriate size. The arrangement does not require any wavelength selection mechanism and theoutput is instantaneously available. However, resolution in diode arrays is limited by the physicalsize of individual detector elements, which at present is about 2 nm.
The diode array photodetectors used in the Hewlett Packard spectrophotomer Model 8450A
consists of two silicon integrated circuits, each containing 211 photosensitive diodes and 211storage capacitors. The photodiode array is a PMOS (p-channel metal-oxide semiconductor)integrated circuit that is over 1.25 cm long. Each photosensitive diode in the array is 0.05 by
0.50 mm and has a spectral response that extends well beyond the 200–800 nm range.

100 Handbook of Biomedical Instrumentation
A functional block diagram of the diode array chip is shown in Fig. 3.22. In parallel with
each of the 211 photodiodes is a 10pF storage capacitor. These photo diode-capacitor pairs aresequentially connected to a common output signal line through individual MOSFET switches.When a FET switch is closed, the preamplifier connected to this signal line forces a potential of– 5V on to the capacitor-diode pair. When the FET switch is opened again, the photocurrent causes
the capacitor to discharge towards zero potential. Serial read-out of the diode array is accom-
plished by means of a digital shift register designed into the photodiode array chip.
The diode arrays typically exhibit a leakage current less than 0.1 pA. This error increases
exponentially with temperature, but because the initial leakage value is so low, there is no need tocool the array at high ambient temperatures.300 400 500 600 700 800 900 1000 2000.10.20.30.40.5
0Normal
photovoltaicBlue
enhanced
UV enhanced
Wavelength (nanometers)Responsivity-amperes/watt
Fig.3.21 Spectral response of silicon diode detectors
10 pF 10 pF 10 pFOutput
Reset
Drain
CommonIntegrator
reset FET
switch
Diode 1 Diode 2 Diode 211Photodiode
capacitor
pairsMOSEFT
switcesStartClocks
Shift register
Fig.3.22 Functional diagram of a photodiode arrayMOSFET
switches

Physiological Transducers 101
/G20/G33/G2E/G38 /G4F/G50/G54/G49/G43/G41/G4C/G20/G46/G49/G42/G52/G45/G20/G53/G45/G4E/G53/G4F/G52/G53
The development of optical fibres has given rise to a number of transducers which find applications
in the medical field. The ability of these fibres to transmit light over great distances with low powerloss and the interaction of light with a measured system provide the basis of these sensing devices.
These sensors are electrically passive and consequently immune to electromagnetic disturbances.
They are geometrically flexible and corrosion resistant. They can be miniaturized and are mostsuitable for telemetry applications.
A great number of optical fibre sensors have been developed for biomedical applications (Martin
et al, 1989). The potential of optical fibres in the sensing of chemical species has led to thedevelopment of a number of optical fibre chemical sensors. They have also been devised for themeasurement of physical parameters such as temperature, pressure and displacement. However,the most direct use of optical fibres in medicine is for gaining access to otherwise inaccessibleregions whether for imaging these areas, as in endoscopy or as the delivery system for light in lasersurgery. However, fibre-optic sensors (Walt, 1992) which are predominantly used for physio-logical measurements are included in the following coverage.
The optical transducers are based on glass or plastic fibres, about 100 to 250 mm in diameter, as
found in fibre-optic communication systems. The initial optical fibre had poor transmissioncharacteristics, but within a decade, fibre losses were reduced from 1000 db/km in 1966 to below1 db/km in 1976. Such improvements in the manufacture and theoretical understanding of lighttransmission in optical beam have lead to their wide spread use in a variety of applications,
including that of clinical medicine.
/G33/G2E/G38/G2E/G31 /G41/G64/G76/G61/G6E/G74/G61/G67/G65/G73/G20/G6F/G66/G20/G4F/G70/G74/G69/G63/G61/G6C/G20/G46/G69 /G62/G72/G65/G20/G53/G65/G6E/G73/G6F/G72/G73
• Optical fibre sensors are non-electrical and hence are free from electrical interference usu-
ally associated with electronically based sensors.
• They are immune from cross-talk.• There is a high degree of mechanical flexibility associated with the fibre optic and this
combined with its reduced size, allows access to otherwise inaccessible areas of thebody.
• They are suitable for telemetry applications as bulk of the instrumentation can be at a
reasonable distance from the patient.
• These sensors do not involve any electrical connection to the patient body, thereby ensuring
patient safety.
• More than one chemical species can be measured with a single sensor by employing more
than one probe detection wavelength offering substantial economic advantage.
• These devices are intrinsically safe, involving low optical power—generally a few
milliwatts.
• The sensors are capable of observing a sample in its dynamic environment, no matter how
distant, difficult to reach or hostile the environment.
• The cost is low enough to make the sensors disposable for many applications.

102 Handbook of Biomedical Instrumentation
/G33/G2E/G38/G2E/G32 /G54/G79/G70/G65/G73/G20/G6F/G66/G20/G4F/G70/G74/G69/G63/G61/G6C/G20/G46/G69 /G62/G72/G65/G20/G53/G65/G6E/G73/G6F/G72/G73
Optical fibre sensors are generally classified into three types as follows:
• Photometric sensors
• Physical sensors• Chemical sensors
Photometric Sensors: Several types of measurements can be made by using the optical fibre as a
device for highly localized observation of the spectral intensity in the blood or tissue. Light
emanating from a fibre end will be scattered or fluoresced back into the fibre, allowing measure-ment of the returning light as an indication of the optical absorption or fluorescence of the volumeat the fibre tip. The variations in the returning light are sensed using a photodetector. Such sensorsmonitor variations either in the amplitude or frequency of the reflected light.
Amplitude Measurements: The most widely used photometric sensor in the amplitude measurement
category is the oximeter. This device measures the oxygen saturation of blood based on the factthat haemoglobin and oxyhaemoglobin have different absorption spectra. The details of this typeof sensor are given in Chapter 10. The use of fibre-optic catheters allows oxygen saturation to bemonitored intra-arterially.
Blood flow measurement based on dye densitometry is closely related to oximetry. A dye,
commonly indocyanine green is injected into the blood and its concentration monitored by itsabsorption at an appropriate wavelength. The time variation of dye concentration can then beused to calculate cardiac output by dilution techniques. The details of these devices are given inChapter 12.
Monitoring the amplitude of the reflected or transmitted light at specific wavelengths can
provide useful information concerning the metabolic state of the tissue under investigation. Thetechnique is non-invasive and fibre-optics play an important role as the technique enables verysmall areas of tissue to be examined so that metabolism at a localized level can be followed. Themethod is based on fluorometry and depends upon the direct observation of tissue and blood
luminescence using fibre-optic light guide to connect the instrument to the tissue.
Frequency Measurements: The second category of photometric sensors using fibre-optic light guide
is based on frequency changes in the signal. The most common example is that of laser Dopplervelocimetry. In this method, light from a laser, normally helium/neon, is sent via a fibre onto the
skin surface. The moving red blood cells scatter the light and produce a Doppler frequency shift
because of their movement. When the light, shifted and unshifted in frequency is mixed, a spectrumof beat frequencies is obtained. Using a number of different processing techniques on the beatfrequency spectrum, the information on the blood flow can be obtained. This technique is given indetail in Chapter 11.
Physical Sensors: Two of the most important physical parameters that can be advantageously
measured using fibre optics are temperature and pressure. These sensors are based on theattachment of an optical transducer at the end of an optical fibre.
Temperature Sensors: The production of localized and controlled hyperthermia (elevated tempera-
tures in the range of 42–45°C or higher) for cancer treatment by electromagnetic energy, either in

Physiological Transducers 103
the radio frequency or microwave frequency range, poses a difficult temperature measurement
problem. Traditional temperature sensors, such as thermistors or thermocouples, have metalliccomponents and connecting wires which perturb the incident electromagnetic (EM) fields and
may even cause localized heating spots and the temperature readings may be erratic due to
interference. This problem is overcome by using temperature sensors based on fibre-optics. Thesedevices utilise externally induced changes in the transmission characteristics of the optical fibresand offer typical advantages of optical fibres such as flexibility, small dimensions and immunityfrom EM interference.
One of the simplest types of temperature sensors consists of a layer of liquid crystal at the end
of optical fibres, giving a variation in light scattering with temperature at a particular wavelength.Figure 3.23 shows ray-path configuration of a temperature sensor which utilizes a silica-coresilicon-clad fibre, with an unclad terminal portionimmersed in a liquid which replaces the clad. Atemperature rise causes a reduction in the refractiveindex of the liquid clad fibre section. Therefore, thelight travelling from the silicon-clad fibre to the liquid-clad fibre undergoes an attenuation which decreases
by increasing temperature. The light from an 860 nm
light-emitting diode (LED) is coupled into the fibre.The light reflected backwards is sent along the samefibres and the light amplitude modulation inducedby the thermo-sensitive clading applied on the distalend of the fibre is detected and processed. Scheggi et al
(1984) constructed a miniature temperature probe formedical use with a 0.8 mm external diameter and 0.5mm internal diameter. The sensitivity achieved was±0.1°C in the temperature interval 20–50 °C.
Another type of temperature sensor is based upon the temperature dependence of the band edge
absorption of infrared light in GaAs (gallium arsenide) crystal, proposed and developed byChristensen (1977). The variation of band-gap energy with temperature (thermal wavelength shift)provides a measureable variation in the transmission efficiency of infrared light through the
crystal.
In the temperature measuring system (Fig. 3.24(a)) based on this principle, light is emitted by
an LED, transmitted to and from the crystal via optical fibres and measured by a photodetector.
No metal parts are used in the temperature probe design, resulting in transparency of the probeto elecromagnetic fields. Single sensor probe with an outer diameter of 0.6 mm and a four pointtemperature sensor probe (Fig. 3.24(b)) of 1.2 mm-diameter based on this technique are commerciallyavailable.
Fluoroptic temperature sensors (Culshaw, 1982) are other useful devices which can be used for
tissue temperature measurement. They contain a rare earth phosphor which is illuminated by awhite light along a short length of large core optical fibre. The light excites the phosphor whichemits a number of lines. By using filters, two of these lines at 540 and 630 nm are selected, andthe ratio of their intensities is a single valued function of the temperature of the phosphor. ByLaLiquid
nt
n1n2Clad
Core
Fig.3.23 Principle of a temperature
sensor based on variationof refractive i ndex with
temperature. The diagramshows optical ray pathconfiguration

104 Handbook of Biomedical Instrumentation
Catheter
Normal
tissue
Tumour
Teflon tubing
(shield)
GaAs
Crystal
Optical fibres
(a)0.6 mm
4.5 cm max.
1.1 mm
(b)
Fig.3.24 Temperature sensor based on change of wavelength of infrared light in gal-
lium arsenide crystal (a) single point sensor (b) multipoint sensor (Courtesy:
M/s Chlinitherm, USA)

Physiological Transducers 105
measuring this ratio, an exact measure of the temperature may be made. The measurement is
independent of the output light intensity. Resolution of 0.1 °C over the range –50 to +250 °C is
reported with this technique.
Pressure Sensors: Measurement of intracranial and intracardiac pressure are both important and
can be performed using fibre optic sensors. For intracranial pressure measurement, the device isbased on a pressure balancing system. Here static pressure is to be monitored and a sensor basedon the deflection of a cantilever mirror attached to a membrane has been demonstrated. Deflection
of the membrane causes the light emitted frame centre optical fibre to be reflected differentially
towards either of the two light-collecting fibres located on each side of the control fibre. The ratioof the light collected by two different fibres is sensed and suitable feedback air pressure is appliedto the interior of the probe through a pneumatic connecting tube, balancing the membrane to itsnull position and providing a readout of the balancing pressure.
A similar sensor based on the deflection of a mirror has been developed for monitoring
intravascular pressure. For intravascular use, dynamic pressure measurement is needed andhence the sensors should not only be small but also have good frequency response in order tofollow the pressure waveforms faithfully.
Chemical Sensors: The development of optical fibre sensors for chemical species has attracted
much interest. The ability of these fibres to transmit light over great distances with low power lossand the interaction of light with a measurand provide the basis of these sensing devices.
The basic concept of a chemical sensor based on optical fibres (Sertz, 1984) is illustrated in
Fig. 3.25. Light from a suitable source is applied to the fibre and is directed to a region where thelight interacts with the measurement system or with a chemical transducer. The interaction resultsin a modulation of optical intensity and the modulated light is collected by the same or anotheroptical fibre and measured by photo-detection system.
Light source
Feed fibre Return fibrePhoto detector
Fig.3.25 An optical fibre chemical sensor (after Narayana Swamy and Sevilla,
1988)

106 Handbook of Biomedical Instrumentation
The optical sensing of chemical species is based on the interaction of these entities with light.
When light strikes a substance, a variety of interaction may occur between the photons of theelectromagnetic radiation and the atoms and molecules of the substance. These interactions
involve an exchange of energy and may lead to absorption, transmission, emission, scattering or
reflection of light. The quantized nature of this energy transfer produces information about thecomposition of the system and forms the basis of the spectrosopic method of chemical analysis(Narayana Swamy and Sevilla, 1988).
Two types of optical fibre sensors have been developed for measuring chemcial species:
•Spectroscopic Sensors: This type of sensor detects the analyte species directly through their
characteristic spectral properties. In these sensors, the optical fibre functions only as a
light guide, conveying light from the source to the sampling area and from the sample tothe detector. Here, the light interacts with the species being sensed.
•Chemical Sensors: In the chemical sensors, a chemical transduction system is interfaced to
the optical fibre at its end. In operation, interaction with analyte leads to a change inoptical properties of the reagent phase, which is probed and detected through the fibreoptic. The optical property measured can be absorbance, reflectance or luminescence. Thesesensors have a great specificity as a consequence of the incorporation of the chemicaltransduction system.
/G20/G33/G2E/G39 /G42/G49/G4F/G53/G45/G4E/G53/G4F/G52/G53
Biosensors combine the exquisite selectivity of biology with the processing power of modernmicroelectronics and optoelectronics to offer powerful new analytical tools with major applicationsin medicine, environmental studies, food and processing industries.
All chemical sensors consist of a sensing layer that interacts with particular chemical substances,
and a transducer element that monitors these interactions. Biosensors are chemical sensors in
which the sensing layer is composed of biological macro-molecules, such as antibodies or enzymes.
Today, the term biosensor is used to describe a wide variety of analytical devices based on the
union between biological and physico-chemical components. The biological component can
consist of enzymes, antibodies, whole cells or tissue slices and is used to recognize and interactwith a specific analyte. The physico-chemical component, often referred to as the transducer,converts this interaction into a signal, which can be amplified and which has a direct relationshipwith the concentration of the analyte. The transducer may use potentiometric, amperometric,optical, magnetic, colorimetric or conductance change properties.
Biosensors offer the the specificity and sensitivity of biological-based assays packaged into
convenient devices for an in situ use by lay personnel. For example, in the medical field biosensorsallow clinical analyses to be performed at the bedside, in critical care units and doctors’ officesrather than in centralized laboratories. The subsequent results of the test can then be acted uponimmediately thus avoiding the delays associated with having to send samples to and having towait for results from centralized laboratories
The critical areas of biosensor construction are the means of coupling the biological component to
the transducer and the subsequent amplification system. Most of the early biosensors immobilized

Physiological Transducers 107
enzymes on selective electrodes, such as the Clark O2 electode, which measured one of the reaction
products (e.g. O2) of the enzyme-analyte interaction. The most successful biosensor to-date is the
home blood glucose monitor for use by people suffering from diabetes. The biosensor in this
instrument relies upon enzymes that recognise and catalyze reactions of glucose with the
generation of redox-active species that are detected electrochemically.
Figure 3.26 shows the construction of this type of sensor. If the immobilized enzyme is soluble
glucose oxidase between the two membranes, it becomes a glucose sensor. It works on the principlethat in the presence of glucose, oxygen is consumed, providing a change in the signal from aconventional oxygen electrode.
7
654321
(Platinum cathode)
(Immobilized enzyme)(Outer device
membrane)(Inner electrode
membrane)(Rubber ring)(Silver anode)(Electrolyte)
Fig.3.26 Constructional details of an enzyme utilizing sensor with
oxygen electrode as the underlying analytical tool
The chemical reaction of glucose with oxygen is catalyzed in the presence of glucose oxidase.
This causes a decrease in the partial pressure of oxygen (pO2), and the production of hydrogen
peroxide by the oxidation of glucose to gluconic acid as per the following reaction:
Glucose + O2Glucose oxidaseæÆææææ æ Gluconic acid + H2 O2

108 Handbook of Biomedical Instrumentation
The changes in all of these chemical components can be measured in order to determine the
concentration of glucose.
For constructing the sensor, glucose oxidase entrapped in a polyacrylamide gel was used. In
general, the response time of such types of bioelectrodes as slow and subsequent work hasconcentrated on closer coupling of the biological component to the transducer. The present
technology in biosensors disposes of the coupling agent by direct immobilization of the enzyme
onto an electrode surface, making the bio-recognition component an integral part of the electrodetransducer. The major disadvantage of enzymatic glucose sensors is the instability of theimmobilized enzyme. Therefore, most glucose sensors operate effectively only for short periodsof time.
A number of alternative approaches have been investigated to develop a glucose sensor. An
important principle that can be used for this purpose depends upon the fluorescence-based,reversible competitive affinity sensor. The sensing element consists of a 3 mm hollow dialysis tuberemotely connected to a fluorimeter via a single optical fibre (Fig. 3.27). It contains a carbohydratereceptor, Canavalin A, immobilized on its inner surface and a fluorescein-labelled indicator as acompeting agent.
Glucose
CanavalinFITC dextran
Immobilised can
A
FITC-dextranHollow dialysis
fibreOptical
fibre
Excitation
Emission
Seal Seal 3 mm
Fig.3.27 Miniature optical glucose sensor (redrawn after Mansouri and Schultz, 1984)
The analyte glucose in the external medium diffuses through the dialysis membrane and
competes for binding sites on a substrate (Canavalin A), with FITC-dextran. The sensor is arrangedso that the substrate is fixed in a position out of the optical path of the fibre end. It is bound tothe inner wall of a glucose-permeable hollow fibre fastened to the end of an optical fibre. Thehollow fibre acts as the container and is impermeable to the large molecules of the fluorescentindicator. Increasing glucose concentration displaces labelled FITC-dextran from the CanavalinA, causing it to be free to diffuse into the illuminated solution volume. The optical field thatextends from the fibre sees only the unbound indicator. At equilibrium, the level of free fluorescein

Physiological Transducers 109
in the hollow fibre is measured as fluorescence intensity and is correlated to the external glucose
concentration.
Biosensors are the most appropriate technology in areas where traditional laboratory analyses,
with their associated cost and time requirement, are not a suitable soluation. One such area inwhich sampling and laboratory analysis are clearly approriate is the detection of hazardous gas
escapes in the workplace. The protection of workers in the chemical industry is ideally achieved
with personal monitors based on biosensors technology that can be worn on the clothing andwhich will give an immediate audible warning of gas escape. For this purpose, biosensors fordetection of hydrogen cyanide have been developed.
Besides the medical field, biosensors have tremendous applications in the food and beverage
industries. Although several biosensors have been developed over the past few years and there arealready numerous working biosensors, various problems still need to be resolved. Most complexproblems awaiting solution are their limited lifetime, which restrict their commercial viability,necessitating improvements in their stability.
/G20/G33/G2E/G31/G30 /G53/G4D/G41/G52/G54/G20/G53/G45/G4E/G53/G4F/G52/G53
Although an accepted industry definition for a smart sensor does not currently exist, it isgenerally agreed that they have tight coupling between sensing and computing elements. Theircharacteristics, therefore, include: temperature compensation, calibration, amplification, some
level of decision-making capability, self-diagnostic and testing capability and the ability to
communicate interactively with external digital circuitry. Currently available smart sensors areactually hybrid assemblies of semiconductor sensors plus other semiconductor devices. In somecases, the coupling between the sensor and computing element is at the chip level on a singlepiece of silicon in what is referred to as an integrated smart sensor. In other cases, the term isapplied at the system level. The important role of smart sensors are:
Signal Conditioning: The smart sensor serves to convert from a time-dependent analog variable to
a digital output. Functions such as linearization, temperature compensation and signal processingare included in the package.
Tightening Feedback Loops: Communication delays can cause problems for systems that rely on
feedback or that must react/adapt to their environment. By reducing the distance betweensensor and processor, smart sensors bring about significant advantages to these types ofapplications.
Monitoring/Diagnosis: Smart sensors that incorporate pattern recognition and statistical techniques
can be used to provide data reduction, change detection and compliation of information formonitoring and diagnostic purposes, specially in the health sector.
Smart sensors divert much of the signal processing workload away from the general purpose
computers. They offer a reduction in overall package size and improved reliability, both of whichare critical for in situ and sample return applications. Achieving a smart sensor depends onintegrating the technical resources necessary to design the sensor and the circuitry, developing amanufacturable process and choosing the right technology.

110 Handbook of Biomedical Instrumentation
A typical example of a smart sensor is a pressure sensor (MPX5050D) with integrated amplifi-
cation, calibration and temperature compensation introduced by Motorola (Frank, 1993). Thesensor typically uses piezioresistive effect in silicon and employs bipolar integrated circuit
processing techniques to manufacture the sensor.
By laser trimming thin film resistors on the pressure sensor, the device achieves a zero-pressure
offset-voltage of 0.5 V nominal and full-scale output voltage of 4.5 V, when connected to a 5.0 V
supply. Therefore, the output dynamic range due to an input pressure swing of 0–375 mmHg is
4.0 V. The performance of the device compares favourably to products that are manufactured withdirect components.

Recording Systems 111
/G52/G65/G63/G6F/G72/G64/G69/G6E/G67/G20/G53/G79/G73/G74/G65/G6D/G73
/G20/G34/G2E/G31 /G42/G41/G53/G49/G43/G20/G52/G45/G43/G4F/G52/G44/G49/G4E/G47/G20/G53/G59/G53/G54/G45/G4D
Recorders provide a permanent visual trace or record of an applied electrical signal. There are
many types of recorders utilizing a variety of techniques for writing purposes. The most elementaryelectronic recording system is shown in Fig. 4.1. It consists of three important components. Firstly,the electrode or the transducer. The electrode picks up the bioelectrical potentials whereas thetransducer converts the physiological signal to be measured into a usable electrical output. The
signal conditioner converts the output of the electrode/transducer into an electrical quantity
suitable for operating the writing system. The writing system provides a visible graphic representa-tion of the quantity of the physiological variable of interest.
Electrodes
transducerSignal
conditionerWriting
system
Fig.4.1 A basic electronic recording system
In medical recorders, the signal conditioners usually consist of a preamplifier and the main
amplifier. Both these types of amplifiers must satisfy specific operating requirements such asinput impedance, gain and frequency response characteristics for a faithful reproduction of theinput signal.
To make the signal from any transducer compatible with the input signal required for the driver
amplifier of the display or recording system, it is usual to arrange to normalize the electricalsignals produced by each transducer. This is done in the signal conditioner which adjusts itsoutput to a common signal level, say one volt. The necessary adjustments of gain and frequencyresponse are provided by the signal conditioners. By this means, it is possible to interchange theHAPTER
44

112 Handbook of Biomedical Instrumentation
signal conditioners to record any one of the physical or bioelectric events on the same writing
channel.
The writing systems which are available in many forms constitute the key portion of the
recording instrument. The commonly used writing systems are the galvanometer type pen recorder,the inkjet recorder and the potentiometric recorder. While the electrodes and transducers have
been described in Chapters 2 and 3, the writing methods and signal conditioners are illustrated in
this chapter.
/G20/G34/G2E/G32/G47/G45/G4E/G45/G52/G41/G4C/G20/G43/G4F/G4E/G53/G49/G44/G45/G52/G41/G54/G49/G4F/G4E/G53/G20/G46/G4F/G52/G20/G53/G49/G47/G4E/G41/G4C/G20/G43/G4F/G4E/G44/G49/G54/G49/G4F/G4E/G45/G52/G53
Information obtained from the electrodes/transducers is often in terms of current intensity, voltagelevel, frequency or signal phase relative to a standard. In addition to handling specific outputsfrom these devices, signal conditioners used in biomedical instruments perform a variety ofgeneral purpose conditioning functions to improve the quality, flexibility and reliability of themeasurement system. Important functions performed by signal conditioners, before the signal isgiven to a display or recording device, are illustrated below:
Signal Amplification: The signals available from the transducers are often very small in magnitude.
Amplifiers boost the level of the input signal to match the requirements of the recording/displaysystem or to match the range of the analog-to-digital convertor, thus increasing the resolution andsensitivity of the measurement.
Bioelectric measurements are basically low-level measurements, which involve amplifying
and recording of signals often at microvolt levels. The problem of electrical noise makes thesemeasurements a difficult proposition and calls for both expert technical knowledge and skillfulhandling of the signal in the circuit design. Noise can produce errors in measurements and
completely obscure useful data. It is a special problem in applications where low-level signals are
recorded at high off-ground voltages, or transmitted over distance or obtained in electromagneticnoise environments. Using signal conditioners located closer to the signal source, or transducer,improves the signal-to-noise ratio of the measurement by boosting the signal level before it isaffected by the environmental noise.
Frequency Response: Modern biomedical instruments are designed to handle data with bandwidths
from dc up to several hundred cycles per second. Electrical or mechanical filters cannot separateuseful signals from the noise when their bandwidths overlap. Instruments and recording systemsthat work satisfactorily for steady state or low frequency data are generally inadequate to meet thisrequirement. On the other hand, recording systems that will faithfully reproduce such data areinherently more susceptible to external noise and, therefore, they must be designed to eliminate thepossibility of signal contamination from noise.
The bioelectric signals often contain components of extremely low frequency. For a faithful
reproduction of the signal, the amplifiers must have excellent frequency response in the sub-audiofrequency range. This response should be down to less than one hertz which is a very frequentrequirement.
In all RC-coupled amplifiers, low frequency response is limited by the reactance of the interstage
coupling capacitors. To achieve the low frequency response required for medical applications, the

Recording Systems 113
amplifier must have large values of coupling capacitance. The disadvantage of using large
capacitors is that they can cause blocking of the amplifier in cases of high-level input, arising dueto switching transients or other high-level inputs. Because of the long time constant introduced by
these large coupling capacitors, several seconds may elapse before the capacitors have discharged
back to the normal levels. The amplifier, therefore, becomes momentarily unreceptive followingeach occurrence of overdriving signals. This type of problem does not exist in direct coupledamplifiers simply because there are no coupling capacitors.
Although the direct coupled amplifier gives an excellent frequency response at low frequencies,
it tends to drift. The drift is a slow change of output having no relation with the input signalapplied to the amplifier. Since the frequency response of the RC-coupled amplifier does not extendall the way down to dc, it does not drift. In medical amplifiers, the advantages of both types of
coupling can be obtained in one amplifier. Typically, all stages except one are direct coupled. The
one RC coupled stage prevents the drifting of the output. Suitable measures are taken to preventblocking of the amplifier due to overdrive by quickly discharging the coupling capacitorautomatically after occurrence of the overdriving input.
It is not desirable to have the frequency response of the amplifiers much above the highest
signal frequency of interest. Excessive bandwidth allows passage of noise voltages that tend to
obscure the bioelectric signal. Another reason for limiting the response of an amplifier to the signalbandwidth is to minimize the tendency towards oscillation due to stray feedback. High frequency
response, therefore, is deliberately limited as a means of noise reduction.
Filtering: A filter is a circuit which amplifies some of the frequencies applied to its input and
attenuates others. There are four common types: high-pass, which only amplifies frequenciesabove a certain value; low-pass, which only amplifies frequencies below a certain value; band-pass, which only amplifies frequencies within a certain band; and band stop, which amplifies allfrequencies except those in a certain band.
Filters may be designed using many different methods. These include passive filters which use
only passive components, such as resistors, capacitors and inductors. Active filters use amplifiersin addition to passive components in order to obtain better performance, which is difficult withpassive filters. Operational amplifiers are frequently used as the gain blocks in active filters.Digital filters use analog-to-digital converters to convert a signal to digital form and then use
high-speed digital computing techniques for filteration.
Additionally, signal conditioners can include filters to reject unwanted noise within a certain
frequency range. Almost all measuring and recording applications are subject to some degree of50 Hz noise picked up from power lines or machinery. Therefore, most signal conditioners includelow-pass filters designed specifically to provide maximum rejection of 50 Hz noise. Such filters are
called “notch” filters.
Filters can be classified as digital and analog filters. They differ by the nature of the input and
the output signals. An analog filter processes analog inputs and generates analog outputs. Adigital filter processes and generates digital data. The processing techniques followed are alsodifferent. Analog filters are based on the mathematical operator of differentiation and digitalfilters require no more than addition, multiplication and delay operators.
Digital filters have several advantages over analog filters. They are relatively insensitive to
temperature, ageing, voltage drift and external interference as compared to analog filters. Their

114 Handbook of Biomedical Instrumentation
response is completely reproducible and predictable, and software simulation can exactly reflect
product performance.
Isolation: Improper grounding of the system is one of the most common causes of measurement
problems and noise. Signal conditioners with isolation can prevent these problems. Such devicespass the signal from its source to the measurement device without a physical or galvanicconnection by using transformer, optical or capacitive coupling techniques. Besides breakingground loops, isolation blocks high voltage surges and rejects high common mode voltages.
Excitation : Signal conditioners are sometimes also required to generate excitation for some
transducers. Strain gauges, thermistors, for example, require external voltage or current excitation.Signal conditioning part of the measurement system usually provides the excitation signal. Straingauges are resistance devices in a Wheatstone bridge configuration, which require bridge
completion circuitry and an excitation source .
Linearizaion : Another common signal conditioning function is linearization. Many transducers
such as thermocouples and thermistors have a non-linear response to changes in the phenomenonbeing measured. Signal conditioners include either hardware-based or software-based linearization
routines for this purpose.
Signal conditioners, therefore, perform an extremely useful function in a measuring and
recording system as they determine the range, accuracy and resolution of the system.
/G20/G34/G2E/G33 /G50/G52/G45/G41/G4D/G50/G4C/G49/G46/G49/G45/G52/G53
Modern multi-channel biomedical instruments and recorders are usually modularly designed to
meet both existing and anticipated requirements. Numerous configurations provide for everymeasurement need, with or without interchangeable plug-in preamplifiers, which provides achoice of signal conditioners for a large selection of analog measurements. Conventionalpreamplifiers offer a wide range of input sensitivities to cover virtually all signal sources.Calibrated zero suppression to expand desired portions of an input signal, and selectable low-pass filtering facilities to reject noise or unwanted signal components are available on theseamplifiers.
For biophysical measurements, the amplifiers employed include: (i) ac/dc universal amplifier
with special features such as capacity neutralization, current injection, low leakage current andlow dc drift suitable for intracellular measurements through high resistance fluid-filled electrodes
or to make extracellular recordings through metal microelectrodes for EMG, EEG, EOG, etc. (ii) an
ECG amplifier with full 12 lead selection and patient isolation (iii) a transducer amplifier suitedfor bridge measurements on strain gauges, strain gauge based blood pressure transducers, forcetransducers, resistance temperature devices and direct low level dc input signals and (iv) a dcamplifier used in conjunction with standard thermistor probes for the accurate measurement oftemperature within the range of medical applications.
Various types of amplifiers which are generally used are as follows:
Differential amplifier is one which will reject any common mode signal that appears simultaneously
at both amplifier input terminals and amplifies only the voltage difference that appears across its

Recording Systems 115
input terminals. Most of the amplifiers used for measuring bioelectric signals are of the differential
type.
Ac coupled amplifiers have a limited frequency response and are, therefore, used only for special
medical applications such as electrocardiograph machine. For electrocardiograms, an ac amplifierwith a sensitivity, giving 0.5 mV/cm, and frequency response up to 1 kHz and an input impedanceof 2 to 5 M W is used. For such applications as retinography, EEG and EMG, more sensitive ac
amplifiers are required, giving a chart sensitivity of say 50 mV/cm with a high input impedance of
over 10 M W.
Carrier amplifiers are used with transducers which require an external source of excitation. They
are characterized by high gain, negligible drift, extremely low noise and the ability to operate withresistive, inductive or capacitive type transducers. They essentially contain a carrier oscillator, a
bridge balance and calibration circuit, a high gain ac amplifier, a phase-sensitive detector and a dc
output amplifier.
DC amplifiers are generally of the negative feedback type and are used for medium gain
applications down to about 1 mV signal levels for full scale. They are not practical for very low
level applications because of dc drift and poor common-mode rejection capabilities. They are
usually employed as pen drive amplifiers in direct writing recorders.
Chopper input dc amplifiers are preferred for low level inputs to instrumentation systems because
of their high sensitivity, negligible drift and excellent common mode rejection capability. Their
high frequency response is limited to about one half of the input chopper frequency.
Chopper stabilized dc amplifiers are used for low level but preferably wideband applications such
as oscilloscopes, tape recorders and light beam oscilloscope recorders. These are complexamplifiers having three amplifiers incorporated in the module. This includes an ac amplifier for
signals above about 20 Hz, a dc chopper input amplifier for signals from about 20 Hz down to dc
plus a wideband feedback stabilized dc amplifier.
DC bridge amplifiers are employed with resistive transducers which require an external source of
excitation. Essentially, the amplifier comprises of a stable dc excitation source, a bridge balance
and calibration unit, a high gain differential dc amplifier and a dc output amplifier. They can be
used as conventional dc high gain amplifiers and offer operating simplicity and high frequencyresponse. These amplifiers are necessary for transducers used to measure temperature and bloodpressure. The sensitivity in these cases may be 50 mV/cm with an input impedance of 50 k W.
/G34/G2E/G33/G2E/G31 /G44/G69/G66/G66/G65/G72/G65/G6E/G74/G69/G61/G6C/G20/G41/G6D/G70/G6C/G69/G66/G69/G65/G72
Medical amplifiers designed for use in the input stage (preamplifiers) are mostly of the differential
type. These type have three input terminals out of which one is arranged at the reference potential
and the other two are live terminals. The differential amplifier is employed when it is necessary to
measure the voltage difference between two points, both of them varying in amplitude at differentrates and in different patterns. Heart-generated voltages picked up by means of electrodes on thearms and legs, and brain-generated voltages picked up by the electrodes on the scalp are typicalexamples of signals whose measurement requires the use of differential amplifiers.

116 Handbook of Biomedical Instrumentation
The differential amplifier is an excellent device for use in the recording systems. Its excellence
lies in its ability to reject common-mode interference signals which are invariably picked up byelectrodes from the body along with the useful bioelectric signals. Also, as a direct coupled
amplifier, it has good stability and versatility. High stability is achieved because it can be
insensitive to temperature changes which is often the source of excessive drift in otherconfigurations. It is versatile in that it may be adapted for a good many applications, e.g.applications requiring floating inputs and outputs or for applications where grounded inputsand/or outputs are desirable.
The working of a differential amplifier can be explained with the help of Fig. 4.2 where the two
transistors with their respective collector resistances ( R
1 and R2) form a bridge circuit. If the two
resistors and the characteristics of the two transistors are identical, the bridge is perfectly balancedand the potential difference across the output terminals is zero.
OutputR1
Ri1
Ri2T1 T2R2+V
–VCommon
mode
input
21
Differential
input
Fig.4.2 Typical differential amplifier configuration
Let us now apply a signal at the input terminals 1 and 2 of this circuit. The signal is to be such
that at each input terminal, it is equal in amplitude but opposite in phase with reference to theground. This signal is known as the differential mode signal. Because of this signal, if the collectorcurrent of T
1 increases, the collector current of T2 will decrease by the same amount, and the
collector voltage of T1 will decrease while that of T2 will increase. This results in a difference
voltage between the two output terminals that is proportional to the gain of the transistors.
On the other hand, if the signal applied to each input terminal is equal in amplitude and is in
the same phase (called the common-mode input signal), the change in current flow through bothtransistors will be identical, the bridge will remain balanced, and the voltage between the output
terminals will remain zero. Thus, the circuit provides high gain for differential mode signals and

Recording Systems 117
no output at all for common-mode signals. Resistances Ri1 and Ri2 are current limiting resistances
for common-mode signals.
The ability of the amplifier to reject these common voltages on its two input leads is known as
common-mode rejection and is specified as the ratio of common-mode input to differential input toelicit the same response. It is abbreviated as CMRR (Common-mode rejection ratio). CMRR is an
important specification referred to the differential amplifier and is normally expressed as decibels.
CMRR of the preamplifiers should be as high as possible so that only the wanted signals find away through the amplifier and all unwanted signals get rejected in the preamplifier stage. A highrejection ratio is usually achieved by the use of a matched pair of transistors in the input stage ofthe preamplifier and a large ‘tail’ resistance in the long-tailed pair to provide maximum negativefeedback for inphase signals. The technique of long-tailing (a technique used to current drive anactive device) improves the CMRR in differential amplifiers without upsetting the gain for thedesired signal. Very high CMRR can be achieved with the use of an active long-tail.
In order to be able to minimize the effects of changes occurring in the electrode impedances, it is
necessary to employ a preamplifier having a high input impedance. It has been found that a lowvalue of input impedance gives rise to a considerable distortion of the recordings (Hill andKhandpur, 1969).
Figure 4.3 is an equivalent circuit for the input of an ECG amplifier, in which:
Z1
Z2 rr
ZI
2
ZI
2VeVh
VbVaElectrode
Electrode
Fig.4.3 Equivalent circuit for the input of an ECG amplifier
Vh represents the voltage signal generated by the heart.
Ve represents unwanted inphase signal picked up from the mains wiring and other sources.
ZI is the total input impedance of the preamplifier.
Z1 and Z2 are the skin contact impedances of the electrodes.
The resistance r represents tissue and blood resistance which is negligibly low as compared
with other impedances.
If the amplifier is perfectly balanced by equal inphase voltages, Va and Vb, at the electrodes
would give rise to a zero output signal. However, the voltages Va and Vb depend, in practice, on the
values of Z1 and Z2. It can be shown that the electrical interference signal Ve will give rise to the
same output signal as would a desired signal, from the patient, of amplitude

118 Handbook of Biomedical Instrumentation
ZZ
ZV
Ie21
2-◊/
Hence, the discrimination factor between desired and undesired signals is given by
Z
ZZI/2
21-
Assuming a common mode rejection ratio of 1000:1 and a difference of electrode skin contact
impedance as 1 k W (Z2–Z1 = 1 k W), then
CMRR = Z
ZZI
221()-
1000 = ZI
2 1000¥
ZI= 2 M W
IfZ2–Z1 = 5 k W.
then, ZI = 10 M W
This shows that a high input impedance is very necessary in order to obtain a high CMRR.
Also, the electrode skin resistance should be low and as nearly equal as possible. Winterand Webster (1983) reviewed several design approaches for reducing common-mode voltageinterference in biopotential amplifiers.
The design of a good differential amplifier essentially implies the use of closely matched
components which has been best achieved in the integrated circuit form. High gain integrated dcamplifiers, with differential input connections and a provision for external feedback have beengiven the name operational amplifiers because of their ability to perform mathematical operations.These amplifiers are applied for the construction of ac or dc amplifiers, active filters, phaseinverters, multi-cvibrators and comparators, etc. by suitable feedback arrangement, and thereforefind a large number of applications in the medical field.
Figure 4.4 shows a single op-amp in a differential configuration. The common mode rejection
for most op-amps is typically between 60 dB and 90 dB. This may not be sufficient to reject commonmode noise generally encountered in biomedical measurements. Also, the input impedance is notvery high to handle signals from high impedance sources. One method to increase the input
+–
R2R2
R1R1
V1
V2Input Output ( ) V0
Gain = / RR21
Output = ( – ) VV V01 2R
R2
1
Fig.4.4 A single op-amp in a differential configuration

Recording Systems 119
impedance of the op-amp is to use field effect transistors (FET) in the input differential stage. A
more common approach is to use an instrumentation amplifier in the preamplifier stage.
/G34/G2E/G33/G2E/G32/G49/G6E/G73/G74/G72/G75/G6D/G65/G6E/G74/G61/G74/G69/G6F/G6E/G20/G41/G6D/G70/G6C/G69/G66/G69/G65/G72
The differential amplifier is well suited for most of the applications in biomedical measurements.However, it has the following limitations:
• The amplifier has a limited input impedance and therefore, draws some current from the
signal source and loads them to some extent.
• The CMRR of the amplifier may not exceed 60 dB in most of the cases, which is usually
inadequate in modern biomedical instrumentation systems.
These limitations have been overcome with the availability of an improved version of the
differential amplifier, whose configuration is shown in Fig. 4.5. An instrumentation amplifier is aprecision differential voltage gain device that is optimized for operation in an environment hostileto precision measurement. It basically consists of three op-amps and seven resistors. Basically,connecting a buffered amplifier to a basic differential amplifier makes an instrumentation
amplifier.
+–
+–+
–A3A1
A2R
RR
RR
RgOutput
terminal+
+–
–V2
V1(–ve) input
(+ve) inputRvar
to balance
out common-
mode voltageV
VV0
12–= 1 +2
a
Fig.4.5 Schematic diagram of an instrumentation amplifier
In the figure shown above, op-amp A3 and its four equal resistors R form a differential amplifier
with a gain of 1. Only A3 resistors have to be matched. The variable resistance Rvar is varied to
balance out any common-mode voltage. Another resistor Rg, is used to set the gain using the
formula

120 Handbook of Biomedical Instrumentation
V
VV0
12-= 1 + 2
a
Where a=Rg/R
V1 is applied to the +ve input terminal and V2 to the –ve input terminal. V0 is proportional to the
difference between the two input voltages.
The important characteristics of the instrumentation amplifier are:
• Voltage gain from differential input ( V1– V2) to single ended output, is set by one resistor.
• The input resistance of both inputs is very high and does not change as the gain is varied.•V
0 does not depend on common-mode voltage, but only on their difference.
If the inputs are prone to high voltage spikes or fast swings, which the op-amps cannot cope
with, they may be protected using back-to-back connected diodes at their inputs. However, thisreduces the input impedance value substantially and also limits the bandwidth.
The instrumentation amplifier offers the following advantages for its applications in the
biomedical field:
• Extremely high input impedance
• Low bias and offset currents• Less performance deterioration if source impedance changes• Possibility of independent reference levels for source and amplifier
• Very high CMRR
• High slew rate• Low power consumption
Good quality instrumentation amplifiers have become available in single IC form such as
mA725, ICL7605, LH0036, etc.
/G34/G2E/G33/G2E/G33 /G43/G61/G72/G72/G69/G65/G72/G20/G41/G6D/G70/G6C/G69/G66/G69/G65/G72/G73
To obtain zero frequency response of the dc amplifier and the inherent stability of the capacitance
coupled amplifier, a carrier type of amplifier is generally used. The carrier amplifier consists of an
oscillator and a capacitance coupled amplifier. The oscillator is used to energize the transducer
with an alternating carrier voltage. The transducers, which require ac excitation, are those whoseimpedance is not purely resistive. Example can be of a capacitance based pressure transducerwhose impedance is mainly capacitative with a small resistive component. The frequency of theexcitation voltage is usually around 2.5 kHz. The transducer shall change the amplitude of thecarrier voltage in relation to the changes in the physiological variable being measured. The outputof the transducer therefore, would be an amplitude modulated (AM) signal (Fig. 4.6). Themodulated ac signal can then be fed to a multi-stage capacitance coupled amplifier. The first stageproduces amplification of the AM signal. The second stage is so constructed that it can respondonly to signal frequency of the carrier. It can be further amplified in the following stage. Afteramplification, the signal is demodulated in a phase-sensitive demodulator circuit. This helps toextract amplified signal voltage after the filter circuit. The voltage produced by the demodulator
can then be applied to the driver stage of the writing system.

Recording Systems 121
Amplifier
Direct
writing
recorderRectifier
Phase
sensitive
detector
Carrier
oscillatorStrain gauge
transducer
Fig.4.6 Block diagram of carrier amplifier
Carrier amplifiers can be used with a resistance strain gauge transducer such as a semi-
conductor strain gauge. When used with pressure gauges, a calibration control is provided on the
carrier amplifier. This enables direct measurements of the blood pressure from the calibratedgraphic recorder.
Lock-in amplifier is a useful version of the carrier technique designed for the measurement of
low-level signals buried in noise. This type of amplifier, by having an extremely narrow-widthoutput band in which the signal is carried, reduces wideband noise and increases the signal-to-noise ratio. Thus, the difference between carrier amplifier and lock-in amplifier is that the formeris a general purpose instrument amplifier while the latter is designed to measure signals in a noisybackground.
In principle, the lock-in amplifier works by synchronizing on a single frequency, called the
reference frequency. This frequency is made to contain the signal of interest. The signal ismodulated by the reference frequency in such a way that all the desired data is at the singlereference frequency whereas the inevitable noise, being broadband, is at all frequencies. Thispermits the signal to be recovered from its noisy background.
/G34/G2E/G33/G2E/G34 /G43/G68/G6F/G70/G70/G65/G72/G20/G41/G6D/G70/G6C/G69/G66/G69/G65/G72
The chopper amplifier is a useful device in the field of medical electronics as it gives anothersolution to the problem of achieving adequate low frequency response while avoiding the driftproblem inherent in direct coupled amplifiers. This type of amplifier makes use of a choppingdevice, which converts a slowly varying direct current to an alternating form with amplitudeproportional to the input direct current and with phase dependent on the polarity of the original

122 Handbook of Biomedical Instrumentation
signal. The alternating voltage is then amplified by a conventional ac amplifier whose output is
rectified back to get an amplified direct current. A chopper amplifier is an excellent device forsignals of narrow bandwidth and reduces the drift problem.
Figure 4.7 shows a simplified block diagram of a single-ended chopper stabilized amplifier.
The amplifier achieves its ultra low dc offset voltage and bias current by chopping the low
frequency components of the input signal, amplifying this chopped signal in an ac amplifier ( A
1)
and then demodulating the output of the ac amplifier. The low frequency components are derivedfrom the input signal by passing it through the low-pass filter, consisting of R
2,C2 and R2. The
chopping signal is generated by the oscillator. The filtered output is then further amplified in asecond stage of dc amplification ( A
2). High frequency signals, which are filtered out at the input
of the chopper channel, are coupled directly into the second stage amplifier. The result of thistechnique is to reduce the dc offsets and drift of the second amplifier by a factor equal to the gainof the chopper channel. The ac amplifier introduces no offsets. Minor offsets and bias currentsexist due to imperfect chopping, but these are extremely small. The amplifier modules contain thechopper channel, including switches and switch-driving oscillator built on the module; only thedc power is supplied externally.
DemodulatorA2
A1
Oscillator
Chopper channelC2C1
R2R2R1 R3R3
C3Input Output2nd stage amplifier
AC
amplifierLow-pass
filter
Low-pass
filter
Fig.4.7 Simplified block diagram of a single ended chopper-stabilized operational
amplifier
Due to the extremely low dc offset and dc drift associated with the chopper-stabilized amplifier,
the signal resolution is limited only by the noise present in the circuit. Thus, it is desirable todesign the feedback networks and external wiring to minimize the total circuit noise. When thefull bandwidth of the amplifier is not required, it is advisable that a feedback capacitor be used tolimit the overall bandwidth and eliminate as much high frequency noise as possible. Shielding offeedback components is desirable in chopper amplifiers. It is particularly necessary in electricallynoisy environments. Use of shielded wire for summing junction leads is also recommended.
Typical voltage drift in chopper-stabilised amplifiers is 0.1 mV/
0C and current drift as 0.5 pA/°C.

Recording Systems 123
The great strength of the chopper-stabilized amplifier is its insensitivity to component changes
due to ageing, temperature change, power supply variation or other environmental factors. Thus,it is usually the best choice where both offset voltage and bias current must be small over long
periods of time or when there are significant environmental changes. Both bias current and offset
voltage can be externally nulled. Chopper amplifiers are available in both single-ended as well asdifferential input configurations. Chopper amplifiers find applications in the medical field inamplification of small dc signals of a few microvolts. Such order of amplitudes are obtainable fromtransducers such as strain gauge pressure transducers, temperature sensors such as thermistorsand strain gauge myographs, when they are used as arms of a dc Wheatstone bridge. A chopperamplifier is also suitable for use with a thermocouple.
/G34/G2E/G33/G2E/G35 /G49/G73/G6F/G6C/G61/G74/G69/G6F/G6E/G20/G41/G6D/G70/G6C/G69/G66/G69/G65/G72/G73
Isolation amplifiers are commonly used for providing protection against leakage currents. Theybreak the ohmic continuity of electric signals between the input and output of the amplifier. Theisolation includes different supply voltage sources and different grounds on each side of theisolation barrier. Three methods are used in the design of isolation amplifiers: (i) transformerisolation (ii) optical isolation (iii) capacitive isolation.
The transformer approach is shown in Fig. 4.8. It uses either a frequency-modulated or a pulse-
width-modulated carrier signal with small signal bandwidths up to 30 kHz to carry the signal. Ituses an internal dc–to-dc converter comprising of a 20 kHz oscillator, transformer, rectifier andfilter to supply isolated power.
Elect.
circuit
Internal
shield
Instrument case
Patient’s ground driven
to zero by amplifierPatient
signal
amplifiersPatient
Isolation
transformerMains
transformer
L
N
Fig.4.8 Isolation amplifier (transformer type)
Isolation could also be achieved by optical means in which the patient is electrically connected
with neither the hospital line nor the ground line. A separate battery operated circuit suppliespower to the patient circuit and the signal of interest is converted into light by a light source (LED).

124 Handbook of Biomedical Instrumentation
This light falls on a phototransistor on the output side, which converts the light signal again
into an electrical signal (Fig. 4.9), having its original frequency, amplitude and linearity. Nomodulator/demodulator is needed because the signal is transmitted optically all the way.
R11R10R8R9
R7
R5R6R3
R1
R2
R4
–Ve+V +Ve
CommonOpto
isolatorECG output
T1A1
Phototransistor
Fig.4.9 Optically isolated isolation amplifier
The capacitive method (Fig. 4.10) uses digital encoding of the input voltage and frequency
modulation to send the signal across a differential capacitive barrier. Separate power supply isneeded on both sides of the barrier. Signals with bandwidths up to 70 kHz can be convenientlyhandled in this arrangement.
Modulator DemodulatorInput Output
signal signal
Capacitive
isolation
Fig.4.10 Capacitively coupled isolation amplifier
The relative merits of the three types of isolation techniques are:
• All three types are in common use, though the transformer isolation amplifier is more
popular.
• Opto-coupled amplifier uses a minimum number of components and is cost effective,
followed by the transformer coupled amplifier. The capacitor coupled amplifier is the mostexpensive.
• Opto-isolated amplifiers offer the lowest isolation voltage (800 V continuous) between
input and output; transformer coupled 1200 V and capacitance coupled 2200 V.
• Isolation resistance levels are of the order of 10
10, 1012 and 1012 ohms for transformer
coupled, opto-coupled and capacitance coupled amplifiers respectively.

Recording Systems 125
• Gain stability and linearity are best for capacitance coupled versions—0.005%, and on par
for the transformer and opto-coupled amplifier—0.02%.
Electrical isolation is the most commonly used technique to ensure patient protection
against electrical hazards. Instruments such as electrocardiographs, pressure monitors, pressuretransducers, pacemakers and others have been designed to electrically separate the portion of thecircuit to which the patient is connected from the portion of the circuit connected to the ac powerline and ground.
/G20/G34/G2E/G34 /G53/G4F/G55/G52/G43/G45/G53/G20/G4F/G46/G20/G4E/G4F/G49/G53/G45/G20/G49/G4E/G20/G4C/G4F/G57/G20/G4C/G45/G56/G45/G4C/G20/G4D/G45/G41/G53/G55/G52/G45/G4D/G45/G4E/G54/G53
/G34/G2E/G34/G2E/G31 /G45/G6C/G65/G63/G74/G72/G6F/G73/G74/G61/G74/G69/G63/G20/G61/G6E/G64/G20/G45/G6C/G65/G63/G74/G72/G6F/G6D/G61/G67/G6E/G65/G74/G69/G63/G20/G43/G6F/G75/G70/G6C/G69/G6E/G67/G20/G74/G6F/G20/G61/G63/G20/G53/G69/G67/G6E/G61/G6C/G73
The distributed capacitance between the signal conductors and from the signal conductors to the
ground provides a low impedance ac path, resulting in signal contamination from external sourceslike power lines and transformers. Similarly, the alternating magnetic flux from the adjacentpower line wires induces a voltage in the signal loop which is proportional to the rate of change ofthe disturbing current, the magnitude of the disturbing current and the areas enclosed by thesignal loop. It is inversely proportional to the distance from the disturbing wire to the signalcircuit. Unequal distances of the two signal carrying conductors from the disturbing current wireresult in unequal mutual inductances, which cause the magnetic field to produce a noise voltageacross the amplifier input terminals.
Low-level signals are sensitive to external contamination especially in the case of high source
impedance. Referring to Fig. 4.11, it is obvious that the currents generated by various noise signalswill flow through the signal source impedance Z and result in an unwanted addition to the
Recorder Amp.Signal source
z
Fig.4.11 Currents produced by various forms of noise flow through the signal
source impedance and become an unwanted addition to the useful signal. The
noise amplitude is directly proportional to signal source impedance (Cour-tesy: Gould Inc. USA)

126 Handbook of Biomedical Instrumentation
bioelectric or transducer signal. This may include electromagnetic noise pick-up, electrostatic
pick-up and the unwanted current generated by a ground loop between two separate grounds onthe same signal circuit. The magnitude of these unwanted signals will be directly proportional to
the signal source impedance as shown by the relationship:
Amplifier input signal = E + I Z
where E= normal signal amplitude
Z= impedance of signal source
I= current generated by noise
It is obvious that as the signal source impedance approaches zero, so will the noise input to the
amplifier. In fact, low-source impedance effectively shunts out the noise.
To prevent noise pick-up from electrostatic fields, low-level signal conductors are surrounded
by an effective shield. This is usually a woven metal braid around the signal pair, which is placedunder an outside layer of insulation. A more effective shielding is provided by a special typeof signal cable, which has lapped foil shields, plus a low resistance drain wire instead of theconventional braided wire shield.
The easiest and generally the best way to protect a signal cable against external electromagnetic
disturbances is to twist the circuit conductors closely together to electrically cancel the effect of an
external magnetic field. The shorter the lay of the twist, the greater the noise rejection. Thus,
electromagnetic coupling is reduced by shielding, wire twisting and proper grounding whichprovide a balanced signal pair with satisfactory noise rejection characteristics.
/G34/G2E/G34/G2E/G32/G50/G72/G6F/G70/G65/G72/G20/G47/G72/G6F/G75/G6E/G64/G69/G6E/G67/G20/G28/G43/G6F/G6D/G6D/G6F/G6E/G20/G49/G6D/G70/G65/G64/G61/G6E/G63/G65/G20/G43/G6F/G75/G70/G6C/G69/G6E/G67/G29
Placing more than one ground on a signal circuit produces a ground loop which may generate somuch noise that it may completely obscure the useful signal. The term ‘grounding’ means a low-impedance metallic connection to a properly designed ground grid, located in the earth. Stable
grounding is necessary to attain effective shielding of low level circuits to provide a stable reference
for making voltage measurements and to establish a solid base for the rejection of unwantedcommon-mode signals. There are generally two grounding systems—a system ground and a signalground. All low-level measurements and recording systems should be provided with a stablesystem ground to assure that electronic enclosures and chassis operating in an electromagneticenvironment are maintained at zero potential. In most instances, the third copper conductor in allelectrical circuits, which is firmly tied to both electric power ground—the building ground and thewater system, will provide a satisfactory system ground. In the signal ground, on the other hand,it is necessary to ensure a low noise signal reference to the ground. This ground should be a low-impedance path to wet earth to minimize the introduction of spurious voltages into the signalcircuitry. It is important to note that a signal circuit should be grounded at one point only.
Two separate grounds are seldom at the same absolute voltage. If we connect more than one
ground to the same signal circuit, an unwanted current will flow in the ground loop thus created.This current combines itself with the useful signal (Fig. 4.12). Also, there is a second ground loop
through the signal cable-shield from the signal source to the amplifier. The current in the shield is
coupled to the signal pair through the distributed capacitance in the signal cable. This current

Recording Systems 127
then flows through the output impedance of the signal source and back to the ground, thus adding
a second source of noise to the useful signal. Either one of these ground loops generates a noisesignal that is larger than a typical millivolt useful signal. Ground loops are eliminated by thefloating lower input terminal of the amplifier. The amplifier enclosure is still solidly grounded toearth-round No. 2 but this will not create a ground loop, since the amplifier enclosure is insulatedfrom the signal circuit. The ground-loop through the signal cable is removed by grounding theshield only at the signal source which is the proper configuration for minimum noise pick-up(Fig. 4.13).
Galv. Amp.
Potential differenceShieldSignal
sourceRecorder chasis
Earth ground
no. 1Earth ground
no. 2No path for
current to circulate
Fig.4.13 Eliminating multiple grounds. The ground loop in the lower signal lead has
been broken by removing the jumper wire to earth ground no. 2.The ground loop in the cable shield has been broken by removing itsconnection to earth ground no. 2. (Courtesy: Gould Inc., USA)Galv. Amp.
Potential differenceCirculating currentsShieldSignal
sourceRecorder chasis
Earth ground no. 1 Earth ground no. 2
Fig.4.12 Ground loop created by more than one ground on a signal circuit. The poten-
tial difference between earth ground no. 1 and earth ground no. 2 causescurrent to circulate in the signal cable shield and also in the lower signalconductor, producing two separate ground loopsRecorder chassis
Recorder chassis

128 Handbook of Biomedical Instrumentation
/G20/G34/G2E/G35 /G42/G49/G4F/G4D/G45/G44/G49/G43/G41/G4C/G20/G53/G49/G47/G4E/G41/G4C/G20/G41/G4E/G41/G4C/G59/G53/G49/G53/G20/G54/G45/G43/G48/G4E/G49/G51/G55/G45/G53
/G34/G2E/G35/G2E/G31 /G46/G6F/G75/G72/G69/G65/G72/G20/G54/G72/G61/G6E/G73/G66/G6F/G72/G6D
In biomedical instrumentation, sensors/transducers pick up signals from biologic tissue with the
objective of finding out their health and well being. Signal processing employs sophisticated
mathematical analysis tools and algorithm to extract information burried in these signals receivedfrom various sensors and transducers. Signal processing algorithms attempt to capture signalfeatures and components that are of diagnostic value. Since most signals of bio-origin are timevarying, there is a need to capture transient phenomena when studying the behaviour of bio-signals.
There are two different presentations of the same experimental data, known as domains. These
are the time-domain in which the data are recorded as a series of measurements at successive timeintervals and the frequency domain in which the data are represented by the amplitude of its sineand cosine components at different frequencies. For example; for recording and display purposes,the biomedical signals are represented in the time domain, i.e. the signal is represented by meansof its value ( y-axis) on the time axis ( x-axis). In the frequency domain any biomedical signal may
be described as consisting of sine waves and having different amplitudes and phases ( y-axis) as a
function of the frequency ( x-axis). The transformation between the two representations is given by
the Fourier Transform (FT).
The basic motivation for developing frequency analysis tools is to provide a mathematical and
pictorial representation for the frequency components that are contained in any given signal.
The term spectrum is used when referring to the frequency content of a signal. The process of
obtaining the spectrum of a given signal using thebasic mathematical tools is known as frequency
orspectral analysis . Most biomedical signals of
practical interest can be decomposed into a sum of
sinusoidal signal components. For the class of
periodic signals, such a decomposition is called aFourier series . For the class of finite energy signals,
the decomposition is called the Fourier transform .
Referring to Fig. 4.14, a system’s response to a
varying input signal s(t), with a frequency spec-
trum S(f), can essentially be described inter-
changeably, by the response r(t) in the time domain
(as a time history) or R(f) in the frequency domain
(as a frequency spectrum).
Key system characteristics operate on the signal
to produce the response to the stimulus. In thefrequency domain, the operation can be expressed as a simple product. The ratio of response tostimulus is called the transfer function ,H(f)
H(f)= R(f)/S(f)
The time domain and frequency domains are closely connected via the Fourier transform.ht()
Hf()st()
Sf()rt()
Rf()Time domain
Frequency
domainFourier
transform
Fig.4.14 A systems transfer function,
h(t) and H(f) characterizes itsresponse r(t) in the timedomain and R(f) in the fre-quency domain, to receivestimuli, s(t) and S(f). The twodomains are related by theFourier transforms

Recording Systems 129
/G34/G2E/G35/G2E/G32/G46/G61/G73/G74/G20/G46/G6F/G75/G72/G69/G65/G72/G20/G54/G72/G61/G6E/G73/G66/G6F/G72/G6D/G20/G28/G46/G46/G54/G29
Using the Fourier Transform can become tedious and time consuming even when using computers
and especially when a large number of points have to be considered. The situation has consi-derably been eased with the introduction of the Fast Fourier Transform (FFT) algorithm, which
expands the signal into sine and cosine functions. The frequency spectrum is computed for each
discrete segment of the signal. The output of the FFT algorithm is a set of coefficients, two for eachfrequency component in the signal’s spectrum. One coefficient ( A) is multiplied by the cosine, or
amplitude portion of the component. The other ( B) is multiplied by the sine, or phase portion, of the
component. Each component in the FFT series can then be represented as:
A cos ( wt) + iB sin ( f)
Where w= Angular frequency of the component
A= An FFT coefficient
B= An FFT coefficient
f= The phase angle of the component
i= The imaginary number, ÷– 1
t= time
The number of frequency components and pairs of FFT coefficients necessary to represent a
given waveform is a function of the highest frequency to be resolved and the sample rate used.
Figure 4.15 illustrates the decomposition of a typical bioelectric signal (EEG waveform) into its
basic frequency components and then displays them as a frequency spectrum. The diagram showsfrequency components along with the amplitude present in each frequency. Once the frequencyspectrum of a given segment of the signal has been calculated, a number of techniques are available
to display the information.
FFT
FFT
FFTHz
Hz
Hz10 Hz
20 Hz
EEG signal30
30
30 HzAmplitude
Amplitude
Amplitude0
0
0
Fig.4.15 EEG signal decomposed into basic frequency components
The FFT method assumes the signal to be stationary and is therefore insensitive to its varying
features. However, most biomedical signals are non-stationary and have highly complex time

130 Handbook of Biomedical Instrumentation
frequency characteristics. The stationary condition for the non-stationary signal can be satisfied
by dividing the signals into blocks of short segments, in which the signal segment can be assumedto be stationary. This method, is called the short-time Fourier transform (STFT). However, the
problem with this method is the length of the desired segment. Choosing a short analysis window
may cause poor frequency resolution. On the other hand, a long analysis window may improvethe frequency resolution but compromises the assumption of stationarity within the window.
/G34/G2E/G35/G2E/G33 /G57/G61/G76/G65/G6C/G65/G74/G20/G54/G72/G61/G6E/G73/G66/G6F/G72/G6D
An emerging method to analyze non-stationary biomedical signals is the wavelet transform . The
wavelet method acts as a mathematical microscope in which we can observe different parts of thesignal by just adjusting the focus. In practice, it is not necessary for the wavelet transform to have
continuous frequency (scale) parameters to allow fast numerical implementations; the scale can
be varied as we move along the sequence. Therefore, the wavelet transform has very good timeresolution at high frequencies and good frequency resolution at low frequencies. In biomedicalengineering, wavelet transform have been widely used in many research areas including spatialfiltering, edge detection, feature extraction, data compression, pattern recognition, speechrecognition, image compression and texture analysis. For example, Sahambi et al (2000) has
developed an algorithm for beat-by-beat QT interval variability using wavelet approach.
Wavelets are a relatively new signal processing method. A wavelet transform is almost always
implemented as a bank of filters that decompose a signal into multiple signal bands. It separatesand retains the signal features in one or a few of these bands. Thus, one of the biggest advantagesof using the wavelet transform is that signal features can be easily extracted. In many cases, awavelet transform outperforms the conventional FFT when it comes to feature extraction andnoise reduction.
Another method of signal analysis is that of adaptive filter that can continuously adjust itself to
perform optimally under the changing circumstances. This is achieved by correcting the signalaccording to the specific application. The correction may be enhancement or some reshaping, for
which a correction algorithm is required. This can be best implemented digitally. Most adaptive
filters, therefore, are implemented by means of computers or special digital signal processingchips.
/G20/G34/G2E/G36 /G53/G49/G47/G4E/G41/G4C/G20/G50/G52/G4F/G43/G45/G53/G53/G49/G4E/G47/G20/G54/G45/G43/G48/G4E/G49/G51/G55/G45/G53
When we pass a signal through a device that performs an operation, as in filtering, we say we haveprocessed the signal. The type of operation performed may be linear or non-linear. Such operationsare usually referred to as signal processing.
The operations can be performed with a physical device or software. For example a digital
computer can be programmed to perform digital filtering. In the case of digital hardware operations
(logic circuits), we have a physical device that performs a specified operation. In contrast, in the
digital processing of signals on a digital computer, the operations performed on a signal consist aof number of mathematical operations as specified by the software.
Most of the signals encountered in biomedical instrumentation are analog in nature, i.e. the
signals are functions of a continuous variable such as time or space. Such signals may be processed

Recording Systems 131
by analog systems such as filters or frequency analyzers or frequency multipliers. Until about
two decades ago, most signal processing was performed using specialized analog processors.As digital systems became available and digital processing algorithms were developed, digital
processing became more popular. Initially, digital processing was performed on general purpose
microprocessors. However, for more sophisticated signal analysis, these devices were quite slowand not found suitable for real-time applications. Specialized designs of microprocessors haveresulted in the development of digital signal processors, which although perform a fairly limitednumber of functions, but do so at very high speeds.
Digital signal processor (popularly known as DSP) requires an interface (Fig. 4.16) between the
analog signal and the digital processor, which is commonly provided by an analog-to-digitalconverter. Once the signal is digitized, the DSP can be programmed to perform the desiredoperations on the input signal. The programming facility provides the flexibility to change thesignal processing operations through a change in the software, whereas hardwired machines aredifficult to configure. Hence programmable signal processors are common in practice. On theother hand when the signal processing operations are well defined, as in some applications, ahardwired implementation of the operations can be optimized so that it results in cheaper andfaster signal processors. In cases when the digital output from a processor is to be given to a user
in the analog form, a DA converter is required.
Analog-to-
digital
converterDigital
signal
processorDigital-to-
analog
converterAnalog Digital Digital Analog
input
signalsignal output
signaloutput
signal
Fig.4.16 Basic elements of a digital signal processor (DSP) system
DSPs are available as single chip devices and are commercially available. The most widely
used DSP family is the TMS320 from Texas Instruments. Another range of processors availablefrom Motorola is the DSP56001. For the sake of comparison of speed, the 16-bit Motorola 68,000microprocessor can handle 2,70,000 multiplications per second while the DSP56001 is capable of10,000,000 multiplications per second, thus giving an increase in speed of 37 times. Because ofthe flexibility to reconfigure the DSP operations, they are used in most of the modern biomedicalinstruments for signal processing applications like transformation to the frequency domain,averaging and a variety of filtering techniques.
/G20/G34/G2E/G37 /G54/G48/G45/G20/G4D/G41/G49/G4E/G20/G41/G4D/G50/G4C/G49/G46/G49/G45/G52/G20/G41/G4E/G44/G20/G44/G52/G49/G56/G45/G52/G20/G53/G54/G41/G47/G45
Normal practice in the design of recorders is to have a preamplifier followed by the main amplifier
which in turn is connected to the driving stage. The main amplifier thus forms an intermediate
stage between the preamplifier and the output stage. The gain and frequency response of the mainamplifier is adjusted to give adequate pen deflection and frequency response.
The pen motor is driven by a dc driver stage feeding a four transistor output stage operating
the galvanometer. A bridge arrangement is preferred because of the low power efficiency ofconventional push-pull power amplifiers. Referring to Fig. 4.17, the current through T
1 and T4
increases as that through T2andT3 decreases. Thus when T1 and T4 approximate to short circuits,

132 Handbook of Biomedical Instrumentation
T2 and T3 are nearly cut-off and almost all the
circuit current passes through the galvanometercoil. T
1 and T2 function as emitter loads. T1 and T4
operate as amplifiers having T1 and T3as collector
loads. Resistors R1to R4 secure the correct biasing
ofT2 and T4.
The output stages of the amplifier are required to
produce more output for a given input signal levelas the signal frequency increases. This is achievedin the output amplifiers by the use of a frequencyselective network in the negative feedback line ofthe driver and output stages so that the gain of theamplifier increases with frequency.
Also, the signal fidelity can be maintained in a
linear recording system only when it has a linearphase shift characteristic and constant gain overthe band of frequencies in the signal of interest. The pen motors need to be protected from excessivetransient energy being applied during switch on or switch off. A delay circuit holds the pen motorgain low for a short time after switch on. When the power supply is switched off, the output goes
rapidly to zero, reducing pen motor gain in advance of power fall. Pen over-deflection is limited
both electrically by adjustable output swing limits and mechanically by pen stops.
/G20/G34/G2E/G38 /G57/G52/G49/G54/G49/G4E/G47/G20/G53/G59/G53/G54/G45/G4D/G53
The final and most important stage in any recorder is the writing system. For a faithful repro-duction of the input signal, three basic conditions must be satisfied by the individual parts of thesystem. These requirements are linearity over the required range of signal amplitudes and anadequate passband for the frequencies involved without producing any phase shift between theinput and recorded signal. Recorders are selected according to the frequency response of the data,accuracy requirements, the type of chart record that is desired and the number of data channelsthat must be recorded on a single piece of chart paper. According to frequency response, recordersfall into the following groups:
Potentiometric recorders usually provide a frequency response of 1 Hz at 25 cm peak-to-peak or
up to 6 Hz at reduced amplitude. Chart paper is inexpensive and the writing method is generally
capillary ink or fibre-tip pen.
Direct-writing ac recorders provide a frequency response up to 60 Hz at 40 mm peak-to-peak or
up to 200 Hz at reduced amplitudes. The most common type of direct recorder is the stylus type
which directly writes on the paper moving beneath it. The stylus can be made to write by severalmethods, but the most commonly employed is a heated stylus writing on specially prepared heatsensitive paper. An ink system can also be used but difficulty is experienced in obtaining a uniformflow with the additional problem of clogging of ink in the pen if the recorder is kept unused overlong periods.+V
T1
T2R1R3
R5R2R4T3
T4Pen coil
Fig.4.17 Bridge output stage for driv-
ing galvanometer coil in di-rect writing recorders

Recording Systems 133
Ink-jet recorder gives a frequency response up to 1000 Hz. It employs inexpensive plain paper as
the writing system makes use of a jet of ink.
Electrostatic recorder employs an electrostatic writing process and works for frequencies up to
10 kHz. The accuracy of the peak-to-peak amplitude is 0.1% at 1000 Hz, 0.2% at 1500 Hz, 2% at5000 Hz, 20% at 15 kHz. It is independent of signal amplitude and the number of channels.
Thermal array recorder uses a specially designed linear thermal array head and thus dispenses
with pens, pen motor or linkages. Frequency response is independent of trace amplitude and the
number of channels. A 20 Hz sine wave is defined by 10 adjacent segments. The accuracy of peak-
to-peak amplitude is 0.2% up to 32 Hz, 2% at 100 Hz and 20% at 320 Hz. A single pulse of durationgreater than 625 ms will be recorded with full amplitude. It uses a plain, heat sensitive paper.
Ultra-violet recorders with mirror galvanometer arrangement and an ultra-violet light source
gives-frequency response up to 2000 Hz. They make use of special UV sensitive paper, whichrequires careful storage. For higher frequency records, UV recorders using a fibre-optic cathoderay tube are used. Their response goes up to several MHz. These recorders are no longer in use.
Cathode ray oscilloscopes are widely used for the display of waveforms encountered in the medical
field. These waveforms can be recorded from the CRO screen by running a photographic filmthrough a recording camera fixed in front of the screen. Recorders are either of the single channeltype or of the type which record several channels simultaneously. Each channel usually carries itsown independent recording system to avoid interference from other channels.
/G20/G34/G2E/G39 /G44/G49/G52/G45/G43/G54/G20/G57/G52/G49/G54/G49/G4E/G47/G20/G52/G45/G43/G4F/G52/G44/G45/G52/G53
In the most commonly used direct writing recorders, a galvanometer activates the writing armcalled the pen or the stylus. The mechanism is a modified form of the D’Arsonval meter movement.
This arrangement owes its popularity to its versatality combined with reasonable ruggedness,
accuracy and simplicity.
A coil of thin wire, wound on a rectangular aluminium frame is mounted in the air space
between the poles of a permanent magnet (Fig. 4.18). Hardened-steel pivots attached to the coilframe fit into jewelled bearings so that the coil rotates with a minimum of friction. Most often, thepivot and jewel is being replaced by a taut band system. A light-weight pen is attached to the coil.Springs attached to the frame return the pen and coil always to a fixed reference point.
When current flows through the coil, a magnetic field is developed which interacts with the
magnetic field of the permanent magnet. It causes the coil to change its angular position as in anelectric motor. The direction of rotation depends upon the direction of flow of current in the coil.The magnitude of pen deflection is proportional to the current flowing through the coil. Thewriting stylus can have an ink tip or it can have a tip that is the contact for an electro-sensitive,pressure sensitive or heat sensitive paper. If a writing arm of fixed length is used, the ordinate willbe curved. In order to convert the curvilinear motion of the writing tip into a rectilinear motion,various correcting mechanisms have been devised to change the effective length of the writing armas it moves across the recording chart.
Taut band instruments are preferred over pivot and jewel type instruments because they have
the advantages of increased electrical sensitivity, elimination of friction, better repeatability and
increased service life.

134 Handbook of Biomedical Instrumentation
Of the several writing methods available, the ink recording method is widely used in slow
speed as well as in high speed recorders. The writing pen depends upon the capillary action of theink for its performance. The pen tip may be of stainless steel or tungsten carbide or a small glassnozzle. The ink reservoir is usually placed slightly above the plane of writing to facilitate the flowof ink. The ink used must flow out from the pen tip in an even manner so that the trace is continuousand no gaps are produced.
Writing quality of a recorder is important to ensure clear non-smudging traces of uniform
width in both steady state and dynamic recording. The trace should be thin, yet well defined anduninterrupted, to allow best resolution of measurements. These problems spurred the developmentof the pressurised ink system. This system uses high pressure, high viscosity ink to overcomeinertial effects within the pen tube and assure a continuous flow of ink even at high pen velocities.Writing fluid, in these systems, is supplied by a central ink reservoir that produces pressures of 15
to 20 psi and forces viscous ink into the microscopic pores in the chart paper surface. Any ink
above the surface is sheared off by the pen tip leaving a permanent trace that is dry and uniformat all writing speeds. Typically, two ounces of viscous ink will provide up to 500 h of recordingwithout refilling. This system is used in the Brush Mark 200 series direct writing recorders.
Most of the portable recorders use the heated stylus writing system wherein recording paper
moves over a steel knife edge kept at right angles to the paper motion and the stylus moves alongthe knife edge. The hot stylus burns off the white cellulose covering of the heat sensitive paper,exposing the black under-surface of the paper thus forming the trace.
Paper Drive: The usual paper drive is by a synchronous motor and a gear box. The speed of the
paper through the recorder is determined by the gear ratio. If it is desired to change the speed of theWriting
arm
Moving coil
Chart
motorPermanent
magnet
Fig.4.18 Principle of a direct writing galvanometric recorder

Recording Systems 135
paper, one or more gears must be changed. Certain instruments are of the fixed speed variety,
i.e. there is no provision for changing the rate at which the paper moves under the pen. Manyapplications, for example, ECG machines, have a single speed.
A constant speed is the basic requirement of the paper drive because the recorded events are
time correlated. The frequency components of the recorded waveform can be determined if it is
known how far the paper moved past the pen position as the record was being taken. Some types
of medical recorders incorporate arrangements for several chart speeds. In such cases, the gear boxhas a single fixed reduction ratio and the speed variation over the total range is achieved by digitalelectronic means. This technique gives a wide ranging crystal controlled accuracy with exactspeed ratios, without the high power consumption and relatively large steps of the steppingsystem, and the added reliability of fewer moving parts compared to a mechanical gear changesystem. Most of the recorders contain an additional timing mechanism that prints a series of smalldots on the edge of the paper as it moves through the recorder. This “time marker” produces onemark each second or at some other convenient time interval.
/G34/G2E/G39/G2E/G31 /G50/G65/G72/G66/G6F/G72/G6D/G61/G6E/G63/G65/G20/G43/G68/G61/G72/G61/G63/G74/G65/G72/G69/G73/G74/G69/G63/G73/G20/G6F/G66/G20/G47/G61/G6C/G76/G61/G6E/G6F/G6D/G65/G74/G65/G72/G73/G20/G55/G73/G65/G64
/G69/G6E/G20/G44/G69/G72/G65/G63/G74/G20/G57/G72/G69/G74/G69/G6E/G67/G20/G52/G65/G63/G6F/G72/G64/G65/G72/G73
The galvanometers used in the recorders generally resemble the corresponding types in the
indicating instruments except that they have a lighter arm carrying a pen in place of the pointer.The pen rests lightly on a chart which moves at a uniform speed in a direction perpendicular tothat of the deflection of the pen. Owing to the friction of the pen on the chart and due to thenecessarily greater weight of the moving system, the design of the galvanometer must be somewhatmodified if it is to be used for recording purposes.
In the direct writing recorders, there are several forces which act upon the moving system. The
three basic forces are: (i) the deflecting force, (ii) the controlling force, (iii) the damping force.
The deflecting force results from the current which flows in the coil and is supplied to it from the
driving amplifier. This force causes the moving system of the recorder to move from its zeroposition. A controlling force applied to it will limit the otherwise indefinite movement of the pen
motor and ensures that the same magnitude of the deflection is always obtained for a given value
of the quantity to be recorded. The damping force is necessary in order to bring the movementsystem to rest in its deflected position quickly. In the absence of damping, owing to the inertia ofthe moving parts, the pen would oscillate about the final deflecting position for some time beforecoming to rest. The function of damping is to absorb energy from the oscillating system and tobring it promptly in its equilibrium position.
Damping Control and Frequency Response of Pen Motors: Galvanometers are characterized by five
important parameters: frequency response, sensitivity (current, voltage and wattage), phase angle,damping and power dissipation. Each one of these parameters is not only important individually,but is also dependent upon the others. Thus, the galvanometer selected to obtain the desiredresults under certain specified conditions must therefore be a compromise of all thesecharacteristics (Mercier, 1973).
Step Function Response: A pen galvanometer is unable to instantaneously follow the deflection
arising when a step change in current is applied to it as would occur when a calibration signal is

136 Handbook of Biomedical Instrumentation
applied. The response of the moving coil lags behind the driving signal due to the mechanical
inertia of the movement. With a good direct writing recorder, the rise time is of the order of 3 ms toattain a 1 cm deflection when the 1 mV calibration button is pressed. The effect of inertia also
causes the coil to tend to overshoot its final deflection. The amount of overshoot depends upon the
value of the damping factor. This is taken as unity when the galvanometer is ‘critically’ damped.Under these conditions, the coil will deflect smoothly to take up its final position in the shortestpossible time without an overshoot. If the damping factor is substantially less than the criticalvalue, the writing arm will overshoot its final deflected position and execute a damped simpleharmonic frequency of almost the natural frequency of the coil and arm assembly.
Figure 4.19 illustrates the step function response for a galvanometer with nominal 65%
damping. To standardize the curve, response is shown versus free period of 1/natural frequencyso that the difference in galvanometer sensitivities can be eliminated. The deflections are shown asratios of the direct current deflections. The manner in which a galvanometer follows a suddencurrent rise depends upon the damping ratio.
0.2 00.00.20.40.60.81.01.21.4
0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.80.690.64
1.0 (Critical)n= 0.59
Free period n= the damping factorRelative amplitude
Fig.4.19 The effect of applying a step input current to a galvanometer with varying
degrees of damping. The ordinate is the ratio of the deflection at a particulartime to the steady-state deflection. The abcissa is the ratio of time to theperiodic time of the galvanometer natural frequency.
The degree of damping depends upon the resistance in the coil circuit. The value of the
resistance required to be placed in parallel with a given galvanometer for correct damping isspecified by the manufacturers. However, when such a galvanometer is to be used in a recordingsystem and it is driven by the power amplifier, it is necessary to readjust the damping resistance sothat the output impedance of the amplifier is taken care of.

Recording Systems 137
Phase Angle: Figure 4.20 shows the theoretical phase angle between a sinusoidal current applied
to the galvanometer of any undamped natural frequency. The frequency of the current is shown asthe ratio of this frequency to the undamped natural frequency of the galvanometer. Phase angle is
also dependent upon the amount of damping. For any frequency ratio and amount of damping,
the phase angle is shown directly.
30
06090120
1.0 1.5 0.5 02.000.800.64n= 0.5
1.0
0.69
0.59
Frequency ratio ( ) fPhase shift
Fig.4.20 The phase shift between the sinusoidal driving signal and the resultant coil
motion as a function of the signal frequency to the natural frequency
Frequency Response and Sensitivity: Galvanometers when damped should give a frequency
response flat within ±5% from dc to 60% of their undamped natural frequency. The naturalfrequency is calculated by multiplying the required frequency response by 1.6. Under certainconditions, it may be desirable to utilize a galvanometer with higher natural frequency than that
calculated as outlined above, but it may be noted, a lower sensitivity will be the result of the
accompanying higher frequency response.
Figure 4.21 represents the theoretical deflection of a galvanometer to a sinusoidal current,
constant in magnitude but variable in frequency. To make the curves applicable to a galvanometerof any undamped frequency and sensitivity, the frequency of the current is shown as the ratio of

138 Handbook of Biomedical Instrumentation
this frequency to the undamped natural frequency of the galvanometer. The deflection is also
dependent upon the amount of damping in the system. The effective damping of the moving coil
assembly can be controlled by arranging the output impedance of the deflection amplifier to fallwith frequency in such a way as to provide the damping required.
The maximum band of flat frequency response for any direct writing recorder is always limited
by some specific peak-to-peak amplitude on the chart. Figure 4.22 shows a typical frequencyversus amplitude curve for a direct writing recorder. The recorder is capable of writing full scaleamplitude (50 mm chart width) up to approximately 40 Hz, 50% of full scale amplitude up to60 Hz, 20% of full scale amplitude at 100 Hz and about 10% of full amplitude at 200 Hz. Thischaracteristic curve is typical of all direct recorders and indicates that increasing the pen motorpower would not help very much. Therefore, for higher frequency response, it becomes imperativeto go to some form of light beam recorder where the mass of the moving element is much smaller.
/G34/G2E/G39/G2E/G32/G4C/G69/G6E/G65/G61/G72/G69/G74/G79/G20/G43/G6F/G6E/G73/G69/G64/G65/G72/G61/G74/G69/G6F/G6E/G73/G20/G69/G6E/G20/G44/G69/G72/G65/G63/G74/G20/G57/G72/G69/G74/G69/G6E/G67/G20/G52/G65/G63/G6F/G72/G64/G65/G72/G73
Ideally, a linear writing system should produce a pen deflection that is directly proportional to theinput current given to the pen motor. This implies that equal increments of input would causeequal changes of trace amplitude in any region of the chart.
In many industrial and general purpose recorders, the pen generates curvilinear traces. The
tip of the pen in such cases traces out an arc on the recording chart with the deflection of then= 0.50.69
0.64
0.59
0.8
1.0
1.5
2.01.0
0.5
00 . 5 1 . 0
Frequency ratio ( ) fRelative amplitude ( )x
1.5 2.0
Fig.4.21 The ratio of galvanometer deflection at any signal frequency to the static
deflection as a function of the ratio of the signal frequency to the natural
frequency for various degrees of damping

Recording Systems 139
galvanometer coil. The system is simple, reliable and cheap but may not be acceptable due to the
following disadvantages: (i) special chart paper is required, ruled to match the pen radius, and(ii) curvilinear traces are often difficult to correlate to their mathematical shapes. It makes theanalysis somewhat more difficult.
The curvilinear arc at the pen tip can be converted into a rectilinear trace with more easily
interpretable rectangular coordinates. Two methods are in common use. In one, an electricallyheated stylus moves on an arc across a special heat-sensitive paper as the paper is drawn over astraight fixed knife (Fig. 4.23) edge. A different portion of the stylus end contacts the paper for eachangle of displacement. The arrangement gives a straight line across the width of the paper with thepaper stationary. However, the method is subject to a degree of geometric non-linearity, whichmay be around 30% for a coil rotation of 18 degrees from the chart centre line to the chart edge.
The second method (Fig. 4.24) makes use of a rotary to linear linkage, which eliminates geometric
errors inherent with conventional knife edge systems. When the pen motor armature rotates, itdrives the pen support arm. The coupling provided by a flexible metal band causes the pen shaftat the upper end of the pen support arm to rotate in the opposite direction since the lower end of theflexible band is firmly attached to a stationary ring on the top of the pen motor housing. As the pen
support arm swings in an arc about the axis of the pen motor armature, the pen tip travels in a
straight line and produces a deflection that is directly proportional to the angular movement of thepen motor armature. The angular displacement of the pen motor coil is translated into a straightline motion at the pen tip by the equal and opposite curvilinear path of the pen support arm. Thisis achieved to an accuracy of 0.2% in recorders, which employ this type of arrangement.1 2 3 5 10 20 30 50 100 200234520304050100
10
Chart Divisions
Frequency in Hz
Fig.4.22 Frequency versus amplitude curve for a direct writing recorder

140 Handbook of Biomedical Instrumentation
A high degree of linearity, accuracy and rapid response is achieved by using closed loop pen
position feedback systems, such as the one employed in Brush Recorder Mark 200. This is theresult of four separate but mutually dependent devices. These are:
(i) A position feedback pen motor that develops more than 250 G’s at the pen tip and main-
tains position accuracy within 0.25%.R
y
Stylus
Rotating coilq
Knife edge
yR=t a n q
Fig.4.23 A heated stylus system for rectilinear writing on a heat-sensitive paper over
a knife edge
Drive bandR2R1
aqb
bKsinq
asinqy
K= –1R
R2
1
ya b K= sin + sin qq
Fig.4.24 Parallel motion mechanism for rectilinear writing. The mechanism keeps the
pen tip nearly perpendicular to the direction of chart motion

Recording Systems 141
(ii) A precise non-contact pen position sensing transducer with infinite resolution called the
Metrisite.
(iii) A rectilinear transfer linkage which translates angular pen motor movement into rectilin-
ear pen travel without introducing hysteresis.
(iv) A pressurized fluid writing system that ensures uniform traces at all writing speeds.
Figure 4.25 illustrates the principle of closed-loop pen position feedback, which automatically
eliminates virtually all sources of error between the incoming signal and pen position. Actual penposition on the chart is continuously sensed by the contact-free Metrisite transducer located insidethe pen-motor. Output from the Metrisite is fed to a summing junction on the pen drive amplifierwhere it is compared to the input signal. The difference between the Metrisite output and theincoming signal produces an “error” voltage, which is amplified and instantaneously corrects the
Fig.4.25 Servoloop feedback pen motor system for fast and accurate recordings
(Courtesy Gould Inc., USA)

142 Handbook of Biomedical Instrumentation
pen position to correspond with the incoming signal. An important advantage of the closed-loop
pen position feedback is that it can produce position stiffness at the pen tip, which is 10 to 100times greater than the conventional spring restored recorders. Errors due to friction are virtually
eliminated and the pen motor heating is reduced. Absolute linearity is maintained across the
entire chart since the pen motor has no torsion spring, never works against mechanical restraintand uses operating power only for changing pen position. Thus, recording accuracy at the pen tipis limited only by the output accuracy of the Metrisite transducer and the ability of the pen motorto apply the necessary corrective forces.
Much depends on the character of the pen position transducer. In a rapidly moving
system such as a recorder, any tendency to wear will reduce accuracy because of dirt, or othercontaminants would quickly destroy reliability. A good pen position sensor is essential in order tocombine fast dynamic response and good recording accuracy in a single instrument.
/G20/G34/G2E/G31/G30 /G54/G48/G45/G20/G49/G4E/G4B/G20/G4A/G45/G54/G20/G52/G45/G43/G4F/G52/G44/G45/G52
Certain biomedical applications like phonocardiography and electromyography require recordingof signals of which the frequency spectrum extends much above the high frequency response ofthe conventional pen recorders. The high frequency response of the stylus direct recorders is
limited by the large moment of inertia of the moving parts and friction of the writing arm over
paper. Moreover, stylus systems can usually only record amplitudes up to 30 mm without collidingwith each other and causing damage. Overlapping of the adjacent tracings is thus not possible.This means that with multi-channel recorders, the stylus recording system requires very widepaper for tracing large amplitudes since the maximum deflection must be recorded withoutoverlapping.
An elegant method of increasing the frequency response extending to several hundred cycles
and combining the advantages of the direct recorder input signal with those of the photographicrecorder is embodied in the jet recording system.
The technique consists of a very fine jet of ink, which replaces the light beam of a photographic
recorder or the stylus of a direct recorder. The jet of ink is produced when the ink is expelled froma nozzle of an extremely fine bore at high pressure. The ink is squirted over the chart movingbeneath the jet. The method is usable up to frequencies of about 1000 Hz. The arrangement(Fig. 4.26) consists of a glass capillary tube placed between the poles of an electromagnet. The coilsof this electromagnet are connected to the output amplifier and are driven by the amplified signal.Attached to this capillary is a very small cylindrical permanent magnet. The variations of current
corresponding to the signal variation in the electromagnet coil produce a varying magnetic field,
which interacts with the field of the permanent magnet. This interaction results in the deflection ofthe capillary tube. The capillary tube is shaped in the form of a nozzle at one end and ink issupplied to this tube at a high pressure from the other end. The ink coming out of the nozzle strikesthe paper and the signal waveform is traced.
Reliable functioning of the jet system demands that the ink be properly filtered to ensure that
even small particles are kept back so that they do not block the nozzle. The high pressure necessaryfor jet recording is produced by a pump and is adjustable between 20–50 atmospheres.

Recording Systems 143
The inherent high frequency response of the jet
recording system is due to the significant reduc-tion in the inert mass of the moving assembly. In
fact, this mass is so low that it compares with that
of the photographic system used in earlier days.With very little mass of the moving parts, the jetrecording mechanism needs only very low drivingpower despite the high inherent frequency.Therefore, it is possible to record deflections of ±30mm measured from the base line. From this it isseen that amplitudes of more than 60 mm can betraced with the jet recorder. Overlapping ofadjacent tracings causes no difficulty becauseliquid jets cannot penetrate each other as is thecase with light beams. Theoretically, the points of
impact of the individual jets do not lie on the same
time ordinate of the recording paper. In practice,however, the shifting is not noticed since it isless than the evaluation accuracy. Since the jetrecorder can write on the recording paper without friction, linear tracing is ensured even with verysmall amplitudes.
Recording of large amplitudes means that there would be considerable deviation of linearity at
the two extreme ends on the recording paper. This is because of the tangent error. Suitabletechniques are adopted to compensate for this error at larger deflections. Compensation is achievedby employing a linear magnetic field rather than the usual radial field. The effective magnetic fieldis then proportional to the cosine of the angle of deflection and thus reduces the sensitivity atlarger deflections.
As with photographic recorders, the recording mechanism of the jet recorder is liquid damped.
This is to obtain balanced tracing over the whole frequency range. The damping medium isconducted in a tube sealed at one end. Its viscosity is so high that it does not run out even if therecording mechanism is turned upside down.
The jet recorder makes use of normal untreated paper, which is much cheaper than the heat
sensitive paper used in the heated-stylus paper recorders. Therefore, it provides an economic
recording method, particularly for multi-channel applications.
/G20/G34/G2E/G31/G31 /G50/G4F/G54/G45/G4E/G54/G49/G4F/G4D/G45/G54/G52/G49/G43/G20/G52/G45/G43/G4F/G52/G44/G45/G52
For the recording of low frequency phenomena, strip chart recorders based on the potentiometric
null-balance principle are generally used. The operating principle of a potentiometer recorder isshown in Fig. 4.27. A slide wire ABis supplied with a constant current from a battery S. The slide
wire is constructed from a length of resistance wire of high stability and uniform cross sectionsuch that the resistance per unit length is constant. The unknown dc voltage is fed between theJet dia 0.01 mm
Magnet needle
Capillary
FilterInput signal
Fig.4.26 An inkJet writing mechanism

144 Handbook of Biomedical Instrumentation
moving contact C and one end A of the slide wire. The moving contact is adjusted so that the
current flowing through the detector is zero. At that moment the unknown input voltage isproportional to the length of the wire AC.In practice, the slide wire is calibrated in terms of span
voltage, the typical span being 100, 10 or 1 mV. The moving contact of the slide wire is made to
carry a pen, which writes on a calibrated chart moving underneath it.
Tacho
generatorChopperVoltage
amplifierPower
amplifier
Phase shiftnetworkLine
voltage
AC
servomotorShaft linkageStabilized
power supplySB
ACSignal input
PenSignal controller
Fig.4.27 Schematic diagram of a self-balancing potentiometric recorder
For obtaining the null-balance, a self-balancing type potentiometer is generally used. The
balancing of the input unknown voltage against the reference voltage is achieved by using a servo-system. The potential difference between the sliding contact C and the input dc voltage is given to
a chopper type dc amplifier in place of a galvanometer. The chopper is driven at the mainsfrequency and converts this voltage difference into a square wave signal. This is amplified by theservo-amplifier and then applied to the control winding of a servo motor. The servo motor is atwo-phase motor whose second winding is supplied with a 50 Hz mains supply that works as areference phase winding. The motor is mechanically coupled to the sliding contact. The motor
turns to move the pen and simultaneously varies the voltage of the sliding contact such that the
potential difference between the input voltage and reference voltage is zero. The circuit operates insuch a manner that the motor moves in one direction if the voltage across the ac is greater and inthe opposite direction if it is less than the input voltage. The servo motor is shaft-coupled to atechno-generator, which provides the necessary damping to the servo motor. It slows down as itapproaches balance position and thus minimizes the overshoot.
The servo motor generally used to drive the pen in the self-balancing potentiometric recorders
is the ac two-phase induction motor. The motor has two separate stator windings, which arephysically perpendicular to each other. The out-of-phase alternating currents in the two statorwindings produce a rotating magnetic field. This rotating magnetic field induces a voltage in therotor and the resulting current in the armature produces an interacting field, which makes therotor to turn in the same direction as the rotating magnetic field. To produce a rotating magneticfield, the ac voltage applied to one stator winding should be 90° out-of-phase with the voltageapplied to the other winding. It can be done either in the power amplifier, which supplies the
control winding, or in a phase-shift network for the line winding. For no input signal, obviously,

Recording Systems 145
the rotor does not turn. When a voltage is applied to the input, the rotor would turn slowly and its
speed would increase with the magnitude of the input voltage. Power amplifiers are required tosupply the necessary alternating current to the control winding of the servo motor. They are
essentially class B push-pull amplifiers.
The chart is driven by a constant speed motor to provide a time axis. Therefore, the input signal
is plotted against time. The recorders of this type are called T-Y recorders. If the chart is made to
move according to another variable, then the pen would move under the control of the secondvariable in the X direction. Such type of recorders are called X-Y recorders. The chart drive can alsobe provided by a stepper motor, which is controlled by a reference frequency from a stableoscillator. Choice of speeds is achieved by a programmable frequency divider and applied to themotor. The chart drive accuracy is thus independent of the power line frequency. Using a steppermotor instead of a synchronous motor eliminates the need for a mechanical transmission toprovide different chart speeds and substantially increases reliability.
Wire wound or conductive plastic potentiometers are used as position feedback elements in
most servo recorders. However, the electromechanical contact between slider and slidewire limitsthe reliability of such recorders and requires regular maintenance. A non-contacting ultrasonicpen position transducer permits high reliability and maintenance-free operation of a recordingpotentiometer. The ultrasonic transducer measures distance ratios as the ratio of the propagationtimes of an ultrasonic pulse travelling from a transmitter past two sensors on a magnetostrictive
delay line. Figure 4.28 shows the basic construction of the position feedback ultrasonic transducer.
Ultrasonic pulses are generated and detected using coils through which the magnetostrictive linepasses. Three coils ( N
1, N0, N2) are used. Coil N0is attached to the recorder pen and is used to
generate magnetostrictive pulses in the line. When an electrical pulse is applied to N0,its magnetic
field produces a magnetostrictive pulse in the line. This stress impulse propagates at ultrasonicspeed along the line. This impulse reaches fixed coils N
1 and N2 after times T1 and T2 respectively
and pulses are induced in the coils as shown in the diagram.
The propagation times T1 and T2correspond to the distances D1 and D2 divided by the speed of
propagation V of the ultrasonic pulse in the line. The ratio of propagation times equals the ratio of
corresponding distances derived as follows:
X=()
()TT
TT12
12-
+ = 2
12x
DD+
where T1=D1/V, T2 = D2/V
x= deviation of N0from centre point between N1 and N2
X= computed time (and distance) ratio
Stegenga (1980) explains the circuit used with an ultrasonic feedback transducer used in
recording potentiometers. The ultrasonic transducer has been found to be a practicable positionfeedback element.
Potentiometric recorders can be conveniently used to record a number of slowly varying
physiological signals. In particular, it is possible to produce a record of the patient’s conditionover a period of 24 h. An electrically driven switch connects the input signals in turn to therecorder, which would then print a dot on the calibrated chart. The position of the dot on the chartgives the value of the parameter at that moment.

146 Handbook of Biomedical Instrumentation
/G20/G34/G2E/G31/G32/G44/G49/G47/G49/G54/G41/G4C/G20/G52/G45/G43/G4F/G52/G44/G45/G52/G53
Digital recorders use a linear fixed array of small recording elements under which the paper
moves. This is in contrast with the conventional recorders that use a moving pen or stylus. Thestylus in these recorder is a large number of fixed stylii, each one of which corresponds to oneamplitude of signal to be recorded. Signals are thus reproduced as discrete values at discretetimes. Analog as well as digital signals can be processed. In the analog case, sampling and
digitization are part of the recording process. The accuracy of array recorders is determined by the
act of sampling and digitization.
As all processing occurs in digital form and recording is a matter of generating the correct
addresses of writing points to be activated, there is no problem in writing figures alongwith signaltracings. Alphanumeric information to be added to the chart can be supplied via the keyboard. Adisplay unit that visualizes the traces stored in memory before they are recorded on the chart canbe provided, thus, giving an opportunity for saving the paper. Time compression or expansion ispossible as data can be sampled at a high rate (expansion), stored in memory and then recorded asthey are released at a reduced rate. A bandwidth higher than that determined by the writingfrequency can be achieved. In this way, fast transients that cannot be handled online can beN1N0N2
D1D2xDD12+
2
N0Input
N1Output
N2OutputT2T1
Fig.4.28 Position feedback ultrasonic transducer for potentiometric recorders

Recording Systems 147
accurately reproduced. A large number of input channels can be made available through a multi-
channel data acquisition system (Collier, 1991).
The obvious advantage of a linear array recorder is the absence of moving parts—nothing
moves but the paper. Inconveniences resulting from mechanical moving parts are not present andthere is no problem of overshoot. The two commonly used techniques in linear array recording
technology are: thermal and electrostatic.
/G34/G2E/G31/G32/G2E/G31 /G54/G68/G65/G72/G6D/G61/G6C/G20/G41/G72/G72/G61/G79/G20/G52/G65/G63/G6F/G72/G64/G65/G72
The thermal array writing technique helps to record analog traces, grid lines, trace identification
and alphanumerics on plain thermal sensitive paper. The array is composed of 512 thermal styliispaced at four per millimetre in a linear array. The array is 128 mm long by 0.25 mm wide.
Array writing is a three-step process. First, the individual drivers for the stylii to be heated are
selected and latched. Next, all stylii to be driven are heated simultaneously by an applied voltage.Finally, the paper is moved an increment equal to or less than the width of one element. Thisprocess is repeated up to 200 times per second. This writing technique makes it possible to write
any combination of dots anywhere on the record. If during the writing period, the analog input
signal changes value, a line segment is printed connecting the lowest and highest amplitudes ofthe signal during the next writing period. Thus, a sine wave of 20 Hz is actually made up of 10discrete line segments. A 20 Hz square wave, on the other hand, appears as an exact reproductiondue to its ability to lay down the transient line all at once.
Thermal writing occurs when specially treated, thermally sensitive paper is heated to 90°–
110°C. The paper, which is initially white or off-white, turns a dark colour (usually blue or black)where heated. The dark colour is caused by a chemical reaction resulting from heating the paper.The functioning of the thermal array writing system is basically digital. A raster line (a series of 1’sand 0’s, one digit corresponding to each of the 512 thermal stylii, a 1 indicating on, a 0 off) is fed tothe driver circuitry, which activates the individual stylii to be heated. Each raster line fed to thedrive circuitry is a summation of raster lines from several sources. The grid generator raster linemakes up the dots forming the grid lines. The raster line from the annotation section allows theprinting of preamp settings and chart speed. Each analog plug-in channel accepts one input and
converts it to a raster line. The analog signal is sampled by a discrete analog-to digital converter
1600 times per second. Each sample generates a 512 bit raster line with all zeroes except for asingle 1, representing the position on the thermal array corresponding to the value of the analogsignal. The plug-in makes at least eight conversions between prints, combining each new rasterline with the old raster line and filling between the highest and lowest values with ONES. Aftereight or more conversions, the raster line from this plug-in is combined with all other raster linesand fed to the thermal array drive circuitry.
The heart of the system is the unique thermal array head (Fig. 4.29) and attendant drive circuitry.
The head is made up of three components; the heater bar, the heat exchanger and the drive circuitry.The heater bar is basically a thick film microcircuit. It comprises 5 mil conductors on 10 mil centresconnected to a common bus by thick film resistors. The resistors form a linear array of 512 activeheating elements spaced at 4 per mm. Utilization of a high heat conductivity substrate allowsquick conduction of excess heat to the exchanger attached to the back of the substrate.

148 Handbook of Biomedical Instrumentation
Heat exchanger
Substrate (heater bar)
Mass interconnect
Flex connector
Driver boardsCommon power return
Heater element
Conductive lines
Fig.4.29 Principle of a thermal array recorder (Courtesy: Gould Inc., USA)
The heat exchanger is cooled using forced air. The efficiency of the heat exchanger allows the
head to have a 25% duty cycle. Even when the recorder is printing all black at the highest chartspeed, there is no problem of overheating the head. The head drive circuitry is contained on twoboards, which are half flex, half rigid. The interconnect pattern is etched directly on the flexportion of the board which is then connected directly to the heater bar. Thus, there is only one massinterconnect point improving reliability and providing ease of head replacement.
The heater bar may be easily removed from the heat sink and disconnected from the drive
circuitry. A replacement can be just as easily installed and connected without special tools orfixtures making the head field replaceable. The thermal array writing system has no pens to wear
out, no ink to fuss with and no pen motors or linkages to cause problems. The only moving part
is the paper drive mechanism. Hence, shortcomings inherent to mechanical linkages such asovershoot, hysteresis and poor reliability are avoided. For the same reason, a fast transient responsetime of 625 ms full scale is possible. With no linkages in the way, all traces can go full scale and
overlap in any desired relationship. Because grids are selected and printed with the traces, lowcost plain thermal paper can be used.
With the advent of array recorders, frequency response needs to be redefined. In moving pen
recorders, frequency response is defined for a sine wave at some specified amplitude. A sine waveis used because it is most easily reproduced by a writing system with moving parts with associatedinertia. With array recorders, there is no inertia and in fact they can reproduce square waves at ahigher frequency than sine waves.

Recording Systems 149
In defining frequency response for array recorders, three parameters must be used. Out of these
parameters which one is more important depends on the specific application. The parameters are“peak capture”, “bandwidth” and “waveform response”. Peak capture is defined as the shortest
duration pulse which can be represented at true value by the recorder. The array recorder has a
peak capture rating of 625 ms, that is any pulse lasting longer than 625 ms will be recorded at full
value. Bandwidth is defined as the maximum sine wave frequency which can be recorded with aspecified accuracy. From the classical sampling theory, an envelope ripple can occur before theusual-3 dB point is reached, and the envelope ripple is regarded as highly objectionable. So, wemust speak of bandwidth as the maximum sine wave frequency without envelope ripple morethan ±0.5 stylus. In this case only a black strip whose width equals the signal amplitude will beprinted because even at maximum paper speed the line segments are contiguous. No waveform orfrequency analysis will be possible but envelope (and transient peak) response is still very useful.The bandwidth of a recorder is 100 Hz with an accuracy of 2%. Waveform response is defined asthat frequency at which a sine wave contains enough line segments to be aesthetically pleasing. Itis specified as a frequency at which each cycle is made up of some minimum number of line
segments (usually 10 or 20). As described earlier, each waveform is made up of line segments.
Typically, a 20 Hz signal would have 10 line segments per cycle. Brimbal and Robillard (1990)describe thermal array recorders for high speed applications, whereas Gaskill (1991) explains therelationship between recorder resolution and print head density.
All medical instruments that require oscillographic recorders are presently preferring thermal
array printers. They aim to create and store arbitrary backgrounds, print multi-channels with fullscale full overlapping of waveforms, store and print annotated channels and create graphics, allat the same time. Embedded software provides freedom to create output anyway the user wishes tosee it and change the way it looks, instantly.
/G34/G2E/G31/G32/G2E/G32 /G56/G69/G64/G65/G6F/G20/G50/G72/G69/G6E/G74/G65/G72/G73
Hardly two decades ago, capturing medical images was limited to X-ray photographic printstaken by a camera mounted on the front of a display monitor. The introduction of video technologyto medicine has revolutionized medical imaging with both video tape and video stills beingused to record almost every medical procedure. The early users of electronic medical imaging wereobstetricians and gynaecologists, who used the technology to generate ultrasound images of foetus.Video technology was introduced in the early 1980s with the availability of low cost black andwhite thermal video printers. The printers displaced the film cameras by providing quality imagesat a lower cost.
Black and white printers were designed specifically for use with medical equipment, such as
ultrasound systems and cardiac cath labs. They can be easily interfaced to medical devices with astandard video signal and produce hard copy images in 4 to 25 seconds. Each heat element on thethermal head of a black and white printer can apply 256 different steps of heat that produce 256different densities for each spot (pixel), resulting in a 256 gradation for each pixel.
The next major step came in the late 1980s with the development of the colour video printer.
They were particularly used in recording colour flow Doppler ultrasound and surgical endoscopy.
Video printers use a dye transfer sublimation thermal printing method in which thousands of

150 Handbook of Biomedical Instrumentation
heating elements come into contact with yellow, magenta and cyan pigments on chemical coated
plastic film. The amount of heat emitted by each heating element controls the amount of dye trans-ferred to the paper. Most video printers support 256 shades of colour per primary colour, which
translates to over 16 million (256
3) colours. The result is a photo realistic, full 24-bit colour image.
First generation printers offered single–frame memory only. This meant that one could capture
images and print them, but the printer remained useless during print time. Modern printers come
with multi-frame memory in which the printers continue to capture images while other images arebeing printed. This enable the doctor to view several images and print only those which areclinically most relevant.
/G34/G2E/G31/G32/G2E/G33 /G45/G6C/G65/G63/G74/G72/G6F/G73/G74/G61/G74/G69/G63/G20/G52/G65/G63/G6F/G72/G64/G65/G72
Electrostatic recorders are high frequency analog recorders which employ a high resolutionelectrostatic device to produce records on a wide, low cost paper at chart speeds up to 250 mm per
second. By eliminating moving writing parts, the electrostatic writing process disposes of the
characteristic moving pen problems like: inertia effects such as overshoot or low-frequencyresponse limits, linkage effects such as non-linearity, hysteresis, and the inability to overlap traces,and preprinted grids that move with paper movement and expand or shrink with changinghumidity conditions.
The Gould ES 1000 (Fig. 4.30) electrostatic writing system is composed of three elements: the
imaging head, the toning head, and the vacuum knife. The imaging head is composed of a lineararray of 1000 wire elements, spaced 4 per mm, for a total length of 250 mm. On each side of thearray are 32 copper bars called shoes. As the paper moves over the image head, a negative voltageis applied to selected wire elements and a positive voltage is applied to the closest shoes. Thisplaces a negative point charge on the paper at the point where the wire element was. The paperthen passes the toner head and positively charged ink particles adhere to where the paper hadnegative charge. A vacuum knife finally removes all excess toner and particles, making the imagewith charged particles. Exposure to air causes the adhesive-coated particles to permanently bond
to the paper and the record emerges from the machine completely dry.
+
PaperWriting
headToner
headVacuum
knife
Fig.4.30 Principle of electrostatic recorder (Courtesy: Gould Inc., USA)

Recording Systems 151
The paper advances in increments less than or equal to the width of a single element, or
0.25 mm. The imaging elements are energized at a constant rate of 1000 times/s. If during theincremental period the analog signal changes value, then a line segment is printed connecting the
lowest and highest amplitudes during that period. Thus, a sine wave of 100 Hz is actually made
up of 10 discrete line segments. A 100 Hz square wave, on the other hand, appears as an exactreproduction due to the ability of the system to lay down the transient line all at once. Theelectrostatic writing system allows the production of any combination of dots anywhere on thechart. Plain electrostatic paper is used, with grid lines printed as selected, along with traceidentification and all desired alphanumerics.
Dot spacing along the amplitude axis is 0.25 mm. This allows 0.1% resolution. Along the time
axis, the top chart speed of 250 mm/s produces dot spacing of 0.25 mm. Because the printing rateis constant, this resolution increases to 40 dots/mm at a chart speed of 25 mm/s, for example. Thechart is driven at continuous selected speed by a servo-controlled dc motor. Six speed choices areprovided from 5 to 250 mm/s. Speeds can be controlled either manually or remotely by a 3-bitword signal.
In defining frequency response for the electrostatic recorders, three measures similar to thermal
array recorders must be used. The measures are “peak capture”, “bandwidth” and “waveformresponse”.
Peak capture rating of the electrostatic recorder is 40 microsecond. The bandwidth of the
recording system is 5000 Hz with an accuracy of 2% or 15 kHz with an accuracy of 20%. However,in this case, only a black strip whose width equals the signal amplitude will be printed. No
waveform or frequency analysis will be possible. Each waveform is made up of line segments and
a 100 Hz signal would have 10 line segments per cycle.
/G20/G34/G2E/G31/G33 /G49/G4E/G53/G54/G52/G55/G4D/G45/G4E/G54/G41/G54/G49/G4F/G4E/G20/G54/G41/G50/G45/G20/G52/G45/G43/G4F/G52/G44/G45/G52/G53
Magnetic tape recording techniques and equipment have found extensive use in the hospital set-up. The fact that the signal is always available in electrical form, makes it possible to record thewhole of an experimental procedure on a tape and then play it back for display on the CRT at alater time. The use of computers in the medical field has further broadened the field of magnetictape recording. The information fed into the computers is coded and stored on magnetic tape, thusforming the memory banks for the digital computers.
Magnetic tape recording offers some useful features over other methods of recording. It permits
the recording of signals, with suitable techniques, from dc up to several MHz. As the recordings ofthe tape can be erased any number of times, the tape becomes re-usable, thus offering economy inthe recording process. The ability to alter the time base of the recorded events on the tape is
something which no other recording medium provides. The events can be played back either faster
or slower than they actually occurred. This permits the use of miniature tape recorders forambulatory monitoring. Since the tape can be played back any number of times, it permitsextracting every bit of useful information from the recording. It is also possible to have a very widedynamic range of recording which may be in excess of 50 decibels. This permits an accurate andlinear recording from full-scale signal level down to its 1/3%.

152 Handbook of Biomedical Instrumentation
The most familiar method of recording signals on the magnetic tape, is the direct recording
process. The electrical signal to be recorded is amplified by the recording amplifier and it is thenfed into the recording head where corresponding magnetic fields are produced. The varying
magnetic field is transferred to the tape in the form of magnetic patterns in accordance with the
signal variations with the tape moving past the head. On replay, the recording process is repeatedin the reverse order. The magnetized tape is pulled past and the magnetic field induces a currentin the playback head coil. This is then amplified before being passed to a loud speaker, a recorderor any other display device.
/G34/G2E/G31/G33/G2E/G31 /G54/G61/G70/G65/G20/G52/G65/G63/G6F/G72/G64/G69/G6E/G67/G20/G77/G69/G74/G68/G20/G46/G72/G65/G71/G75/G65/G6E/G63/G79/G20/G4D/G6F/G64/G75/G6C/G61/G74/G69/G6F/G6E
The inherent limitation of the direct recording process on the magnetic tape is its inability to record
signals of very low frequencies. The frequency range of signals which we come across in the
medical field varies from dc to several hundred Hz. There is a method to handle these signals byemploying a carrier frequency which is frequency-modulated by the signal to be recorded. Onreplay, the signal has to be demodulated and passed through a low-pass filter for removing thecarrier and other unwanted frequencies. The circuit details of an FM tape recording system forbiomedical application are given by Smith et al. (1979).
An important application of the FM recording technique is its flexibility in using “frequency-
deviation multiplexing” where a number of carrier frequencies are separately modulated bydifferent input signals. The resulting signals can then mix linearly and the composite signal canbe recorded using the direct recording process. This method offers the facility of recording manychannels of signal information on one track of the tape and utilizing the wide bandwidth and thelinearity of the direct recording technique.
FM recording technique requires that the tape should move across the heads at a precisely
uniform speed, because any speed variations introduced into the tape will cause an unwantedmodulation of the carrier frequency. This will result in noise in the output signal and wouldlimit the accuracy and dynamic range of the FM system. The variation of speed of tape movement
would cause wow and flutter and the problem is particularly acute in the frequency multiplexed
system.
/G34/G2E/G31/G33/G2E/G32/G54/G61/G70/G65/G20/G52/G65/G63/G6F/G72/G64/G69/G6E/G67/G20/G77/G69/G74/G68/G20/G44/G69/G67/G69/G74/G61/G6C/G20/G54/G65/G63/G68/G6E/G69/G71/G75/G65
Another method of recording information on the magnetic tape is the digital recording process.This process has been growing rapidly in importance as a result of the widespread application ofdigital computers. In this method, a sampling technique is used to measure a varying signal. Thesampled readings are then converted into a code consisting of a group of binary digits. Recording
is accomplished by magnetizing the tape to saturation in either of its two possible directions at
discrete points along its length.
The advantages of the digital recording process are its inherent capability of extremely
high orders of accuracy, its insensitivity to tape speed variations and a simple recording andreproducing electronic circuitry.

Recording Systems 153
The major problem of digital recording process is its sensitivity to tape drop-out errors. Since all
information is contained in the presence or absence of pulses, on replay we cannot tolerate the lossof pulses or the generation of spurious pulses caused by tape imperfections. This puts a limitation
on the practical minimum duration for pulses, thus affecting the pulse packing density. The
process also needs the data to be digitized at the source or special digital transducers must beemployed.

154 Handbook of Biomedical Instrumentation
/G42/G69/G6F/G6D/G65/G64/G69/G63/G61/G6C/G20/G52/G65/G63/G6F/G72/G64/G65/G72/G73
/G20/G35/G2E/G31 /G45/G4C/G45/G43/G54/G52/G4F/G43/G41/G52/G44/G49/G4F/G47/G52/G41/G50/G48
The electrocardiograph (ECG) is an instrument, which records the electrical activity of the heart.
Electrical signals from the heart characteristically precede the normal mechanical function andmonitoring of these signals has great clinical significance. ECG provides valuable informationabout a wide range of cardiac disorders such as the presence of an inactive part (infarction) or anenlargement (cardiac hypertrophy) of the heart muscle. Electrocardiographs are used in catheteriza-
tion laboratories, coronary care units and for routine diagnostic applications in cardiology.
Although the electric field generated by the heart can be best characterized by vector quantities,
it is generally convenient to directly measure only scalar quantities, i.e. a voltage difference of mV
order between the given points of the body. The diagnostically useful frequency range is usuallyaccepted as 0.05 to 150 Hz (Golden et al 1973). The amplifier and writing part should faithfully
reproduce signals in this range. A good low frequency response is essential to ensure stability ofthe baseline. High frequency response is a compromise of several factors like isolation between auseful ECG signal from other signals of biological origin (myographic potentials) and limitationsof the direct writing pen recorders due to mass, inertia and friction. The interference of non-biological origin can be handled by using modern differential amplifiers, which are capable ofproviding excellent rejection capabilities. CMRR of the order of 100–120 dB with 5 k W unbalance
in the leads is a desirable feature of ECG machines. In addition to this, under specially adversecircumstances, it becomes necessary to include a notch filter tuned to 50 Hz to reject hum dueto power mains. The instability of the baseline, originating from the changes of the contact
impedance, demands the application of the automatic baseline stabilizing circuit. A minimum of
two paper speeds is necessary (25 and 50 mm per sec) for ECG recording.
/G35/G2E/G31/G2E/G31 /G42/G6C/G6F/G63/G6B/G20/G44/G69/G61/G67/G72/G61/G6D/G20/G44/G65/G73/G63/G72/G69/G70/G74/G69/G6F/G6E/G20/G6F/G66/G20/G61/G6E/G20/G45/G6C/G65/G63/G74/G72/G6F/G63/G61/G72/G64/G69/G6F/G67/G72/G61/G70/G68
Figure 5.1 shows the block diagram of an electrocardiograph machine. The potentials picked upby the patient electrodes are taken to the lead selector switch. In the lead selector, the electrodes areHAPTER
55

Biomedical Recorders 155
selected two by two according to the lead program. By means of capacitive coupling, the signal is
connected symmetrically to the long-tail pair differential preamplifier. The preamplifier is usuallya three or four stage differential amplifier having a sufficiently large negative current feedback,
from the end stage to the first stage, which gives a stabilizing effect. The amplified output signal is
picked up single-ended and is given to the power amplifier. The power amplifier is generally ofthe push-pull differentical type. The base of one input transistor of this amplifier is driven by thepreamplified unsymmetrical signal. The base of the other transistor is driven by the feedbacksignal resulting from the pen position and connected via frequency selective network. The outputof the power amplifier is single-ended and is fed to the pen motor, which deflects the writing armon the paper. A direct writing recorder is usually adequate since the ECG signal of interest haslimited bandwidth. Frequency selective network is an R–C network, which provides necessary
damping of the pen motor and is preset by the manufacturer. The auxiliary circuits provide a 1 mVcalibration signal and automatic blocking of the amplifier during a change in the position of thelead switch. It may include a speed control circuit for the chart drive motor.
Lead
selectorPreampPower
amplifierBridge
output
circuit
Pen motorFrequency
selective
feedback
networkAuxiliary
circuits
Chart
transport
motorECG
electrodes
Fig.5.1 Block diagram of an ECG machine
A ‘stand by’ mode of operation is generally provided on the electrocardiograph. In this mode,
the stylus moves in response to input signals, but the paper is stationary. This mode allows the
operator to adjust the gain and baseline position controls without wasting paper.
Electrocardiograms are almost invariably recorded on graph paper with horizontal and vertical
lines at 1 mm intervals with a thicker line at 5 mm intervals. Time measurements and heart ratemeasurements are made horizontally on the electrocardiogram. For routine work, the paperrecording speed is 25 mm/s. Amplitude measurements are made vertically in millivolts. Thesensitivity of an electrocardiograph is typically set at 10 mm/mV.
Isolated Preamplifier: It had been traditional for all electrocardiographs to have the right leg (RL)
electrode connected to the chassis, and from there to the ground. This provided a ready path forany ground seeking current through the patient and presented an electrical hazard. As themicroshock hazard became better understood, particularly when intracardiac catheters areemployed, the necessity of isolating the patient from the ground was stressed. The American HeartAssociation guidelines state that the leakage current should not be greater than 10 microampereswhen measured from the patient’s leads to the ground or through the main instrument grounding

156 Handbook of Biomedical Instrumentation
wire with the ground open or intact. For this, patient leads would have to be isolated from the
ground for all line operated units.
Figure 5.2 shows a block diagram of an isolation preamplifier used in modern electro-
cardiographs. Difference signals obtained from the right arm (RA), left arm (LA) and right leg (RL)is given to a low-pass filter. Filtering is required on the input leads to reduce interference caused by
electrosurgery and radio frequency emissions and sometimes from the 50 kHz current used for
respiration detection. The filter usually has a cut off frequency higher than 10 kHz. A multistagefilter is needed to achieve a suitable reduction in high frequency signal.
STD
1 mV
DC amp.Synchronous
modulatorSynchronous
demodulatorModulatorGuard
DriverTransformer
Output
100 kHz
oscillator
Isolated power
transformerRect.
&
filt.+V
–VFloating
common
Electrosurgery
filterHigh-voltage and
over-voltage
protectionLead selector
RA
LA
LL
RL
Fig.5.2 Block diagram of an isolation preamplifier (transformed-coupled)
commonly used in modern ECG machines
The filter circuit is followed by high voltage and over voltage protection circuits so that the
amplifier can withstand large voltages during defibrillation. However, the price of this protectionis a relatively high amplifier noise level arising from the high series resistance in each lead.
The lead selector switch is used to derive the required lead configurations and give it to a
dc-coupled amplifier. A dc level of 1 mV is obtained by dividing down the power supply, whichcan be given to this amplifier through a push button for calibration of the amplifier. Isolation of thepatient circuit is obtained using a low capacitance transformer whose primary winding is drivenfrom a 100 kHz oscillator. The transformer secondary is used to obtain an isolated power supplyof ±6 V for operating the devices in the isolated portion of the circuit and to drive the synchronousmodulator at 100 kHz, which linearly modulates an ECG signal given to it. The oscillator frequencyof 100 kHz is chosen as a compromise so that reasonable size transformers (higher the frequencythe smaller the transformer) could be used and that the switching time is not too fast, so thatinexpensive transistors and logic circuitry can be utilized. A square wave is utilized to minimize

Biomedical Recorders 157
the power requirements of the driven transistors. A synchronous demodulator is chosen to give
low noise performance utilizing switching FET’s.
Isolation of the patient preamplifier can also be obtained using an optical isolator. The high
common-mode rejection of the amplifier is obtained by proper shielding. The effective capacitancefrom the input leads to the earth is made negligible. The preamplifier circuitry should is preferably
be shielded in a separate case.
To minimize the common-mode signal between the body of the patient and the floating ground,
a right leg drive circuit (Fig. 5.3) is used. The common-mode signals after amplification in a
preamplifier are inverted and fed back to the right leg electrode, reducing the common modevoltage on the input with respect to the floating ground. Winter and Webster (1983) examinedoptimal design parameters for a driven-right-leg circuit.
LA
RLRALine
B
A
C
ZLAZRLZRA
C2
C1
Common-mode rejection amplifierECGRefRL-drive circuit
Fig.5.3 Improvement in CMRR using right leg drive (Courtesy: Hewlett Packard,
USA)
The presence of stray capacitance at the input of the preamplifier causes common-mode currents
to flow in LA and RA, resulting in a voltage drop at the electrode resistors. An imbalance of thestray capacitance or the electrode resistors causes a difference signal. This difference signal can bealmost eliminated, in that the common-mode currents of stray capacitances are not allowed toflow through the electrode resistors but are neutralized by currents delivered to stray capacitancesfrom the common-mode rejection amplifier. In other words, the potentials at A, B and C are

158 Handbook of Biomedical Instrumentation
equalised through an in-phase component of the common-mode voltage, which the amplifier
delivers via C1 and C2 to LA and RA. As a result, the potentials at A, B and C are kept equal,
independent of the imbalance in the electrode resistors and stray capacitance.
The modern ECG machines with their completely shielded patient cable and lead wires and
their high common-mode rejection, are sufficiently resistant to mains interference. However, there
could be locations where such interference cannot be eliminated by reapplying the electrodes or
moving the cable, instrument or patient. To overcome this problem, some ECG machines have anadditional filter to sharply attenuate a narrow band centred at 50 Hz. The attenuation providedcould be up to 40 dB. In this way, the trace is cleaned up by the substantial reduction of linefrequency interference.
Isolation amplifiers are available in the modular form. One such amplifier is Model 274 from
Analog Devices. This amplifier has the patient safety current as 1.2 mA at 115 V ac 60 Hz and offers
a noise of 5 mV pp. It has a CMRR of 115 dB, differential input impedance of 10
12W paralleled with
3 pF and common mode impedance as 1011W and a shunt capacitance of 20 pF. It is optimized for
signal frequencies in the range of 0.05 to 100 Hz. Metting van Rijn et al (1990) detail out methods
for high-quality recording of bioelectric events with special reference to ECG.
/G35/G2E/G31/G2E/G32 /G54/G68/G65/G20/G45/G43/G47/G20/G4C/G65/G61/G64/G73
Two electrodes placed over different areas of the heart and connected to the galvanometer willpick up the electrical currents resulting from the potential difference between them. For example,if under one electrode a wave of 1 mV and under the second electrode a wave of 0.2 mV occur at thesame time, then the two electrodes will record the difference between them, i.e. a wave of 0.8 mV.The resulting tracing of voltage difference at any two sites due to electrical activity of the heart iscalled a “LEAD” (Figs 5.4 (a)-(d)).
Bipolar Leads: In bipolar leads, ECG is recorded by using two electrodes such that the final trace
corresponds to the difference of electrical potentials existing between them. They are calledstandard leads and have been universally adopted. They are sometimes also referred to asEinthoven leads (Fig. 5.4(a)).
In standard lead I, the electrodes are placed on the right and the left arm (RA and LA). In lead II,
the electrodes are placed on the right arm and the left leg and in lead Ill, they are placed on theleft arm and the left leg. In all lead connections, the difference of potential measured betweentwo electrodes is always with reference to a third point on the body. This reference point isconventionally taken as the “right leg”. The records are, therefore, made by using three electrodes
at a time, the right leg connection being always present.
In defining the bipolar leads, Einthoven postulated that at any given instant of the cardiac
cycle, the electrical axis of the heart can be represented as a two dimensional vector. The ECG
measured from any of the three basic limb leads is a time-variant single-dimensional componentof the vector. He proposed that the electric field of the heart could be represented diagrammaticallyas a triangle, with the heart ideally located at the centre. The triangle, known as the “Einthoven
triangle ”, is shown in Fig. 5.5. The sides of the triangle represent the lines along which the three
projections of the ECG vector are measured. It was shown that the instantaneous voltage measured

Biomedical Recorders 159
Bipolar Limb Leads
(a)
(b)Unipolar limb leads
Lead AVR** Lead AVL** Lead AVF**C.M
AMPLC.M
AMPL
C.M means “common mode”
Lead I Lead II Lead IIIC.M
AMPLC.M
AMPL
C.M
AMPLC.M
AMPLBuffersBuffers Buffers Buffers
Buffers Buffers
+++ +
++
RLRL RL RL
RL RLRARA RA RA
RA RA LALA LA LA
LA LA––– –
––
Fig.5.4 Types of lead connections with typical ECG waveforms (a) bipolar
limb leads (b) unipolar limb leads (Courtesy: Hewlett Packard, USA)

160 Handbook of Biomedical Instrumentation
Buffers(c) Unipolar chest leads
Fourth intercostal space,
at right sternal margin.
Fourth intercostal space,
at left sternal margin.
Midway between V and V .24
Fifth intercostal space, at
mid-calvicular line.
Same level as V , on anterior
axilliary line.4
Same level as V , on midaxilliary line. 4
Ensiform, base of sternum.V1
V2
V3
V4
V5
V6
EC.M.
AMPL
Lead V ** CH positions CH positionsRLRACH
LA
+
V1 V2
V3V R3
V4V5V6 VE–
LEAD
LEAD
LEADLEAD
LEAD
LEAD
(d)V1
V2
V3V4
V5
V6
Fig.5.4 Types of lead connections with typical ECG waveforms (c) position
of the chest lead in unipolar precordial lead recording (d) C leads(Courtesy: Hewlett Packard, USA)

Biomedical Recorders 161
from any one of the three limb lead positions is approximately equal to the algebraic sum of the
other two or that the vector sum of the projections on all three lines is equal to zero.
In all the bipolar lead positions, QRS of a normal heart is such that the R wave is positive and
is greatest in lead II.
Unipolar Leads (V Leads): The standard leads record the difference in electrical potential between
two points on the body produced by the heart’s action. Quite often, this voltage will show smaller
changes than either of the potentials and so better sensitivity an be obtained if the potential of a
single electrode is recorded. Moreover, if the electrode is placed on the chest close to the heart,higher potentials can be detected than normally available at the limbs. This lead to the developmentof unipolar leads introduced by Wilson in 1894. In this arrangement, the electrocardiogram isrecorded between a single exploratory electrode and the central terminal, which has a potentialcorresponding to the centre of the body. In practice, the reference electrode or central terminal isobtained by a combination of several electrodes tied together at one point. Two types of unipolarleads are employed which are: (i) limb leads, and (ii) precordial leads.
(i)Limb leads In unipolar limb leads (Fig. 5.4(b)), two of the limb leads are tied together and
recorded with respect to the third limb. In the lead identified as AVR, the right arm isrecorded with respect to a reference established by joining the left arm and left leg elec-trodes. In the AVL lead, the left arm is recorded with respect to the common junction of theright arm and left leg. In the AVF lead, the left leg is recorded with respect to the two armelectrodes tied together.
They are also called augmented leads or ‘averaging leads’. The resistances inserted
between the electrodes-machine connections are known as ‘averaging resistances’.
(ii)Precordial leads The second type of unipolar lead is a precordial lead. It employs an explor-
ing electrode to record the potential of the heart action on the chest at six different posi-tions. These leads are designated by the capital letter ‘V’ followed by a subscript numeral,which represents the position of the electrode on the pericardium. The positions of thechest leads are shown in Fig. 5.4(c).++
LLII IIII
LA RA+
ECG
vectorLead I
Lead IILead III+ –RA LA
++–
–
LL
Fig.5.5 The Einthoven triangle for defining ECG leads

162 Handbook of Biomedical Instrumentation
/G35/G2E/G31/G2E/G33 /G45/G66/G66/G65/G63/G74/G73/G20/G6F/G66/G20/G41/G72/G74/G65/G66/G61/G63/G74/G73/G20/G6F/G6E/G20/G45/G43/G47/G20/G52/G65/G63/G6F/G72/G64/G69/G6E/G67/G73
Abnormal patterns of ECG may be due to pathological states or on occasion they may be due to
artefacts. To diagnose the presence of undesirable artefacts on the ECG trace, a few recordings areillustrated below:
Interference from the Power Line: Power line interference is easily recognizable since the inter-
fering voltage in the ECG would have a frequency of 50 Hz (Fig. 5.6(a)). This interference may bedue to the stray effect of the alternating current on the patient or because of alternating currentfields due to loops in the patient cable. Other causes of interference are loose contacts on the
patient cable as well as dirty electrodes. When the machine or the patient is not properly grounded,
power line interference may even completely obscure the ECG waveform.
(a)
(b)
(c)
Fig.5.6 (a) ECG recording with regular spreading of the curve with super
imposed 50 Hz power line interference signals
(b)Recording with irregular trembling of the ECG trace without
wandering of the base line but otherwise normal ECG trace
(c)ECG trace without wandering of the base line
The most common cause of 50 Hz interference is the disconnected electrode resulting in a very
strong disturbing signal. It is often strong enough to damage the stylus of an unprotected direct
writing recorder, and therefore needs quick action.
Sometimes static charges on the synthetic uniform of the operator may result in a random noise
on the trace. This noise is very difficult to remove except in those machines which have very high
CMRR. The noise can be reduced by partially shielding the patient by means of the bed springs.Connection of the springs to the instrument case helps to compensate for a poor CMRR (Spooner,1977).

Biomedical Recorders 163
Electromagnetic interference from the power lines also results in poor quality tracings. Electrical
equipment such as air-conditioners, elevators and X-ray units draw heavy power-line current,which induce 50 Hz signals in the input circuits of ECG machines. Due to unbalanced linkages,
common mode rejection circuits almost prove ineffective against them. A practical solution to
minimize this problem is physical separation between the interference causing sources and thepatient. Levkov et al (1984) developed a method of digital 50 Hz interference elimination by
computing the interference amplitudes and subtracting these data from the original signal, therebygreatly reducing the requirements of amplifiers, shielding, earthing, electrode quality andapplication procedures.
Electrical power systems also induce extremely rapid pulses or spikes on the trace, as a result of
switching action. Use of a transient suppressor in the mains lead of the machines helps to solvethis problem.
Shifting of the Baseline: A wandering baseline (Fig. 5.6(b)) but otherwise normal ECG trace is
usually due to the movement of the patient or electrodes. The baseline shift can be eliminated byensuring that the patient lies relaxed and the electrodes are properly attached. Baseline wander isusually observed immediately after application of the electrodes. It is due to a relatively slowestablishment of electrochemical equilibrium at the electrode-skin interface. This can be minimizedby selecting the proper electrode material, which will reach equilibrium quickly with a goodelectrode jelly.
Muscle Tremor: Irregular trembling of the ECG trace (Fig. 5.6(c)), without wandering of the baseline
occurs when the patient is not relaxed or is cold. It is generally found in the case of older patients.Muscle tremor signals are especially bothersome on limb leads when a patient moves or themuscles are stretched. Therefore, for long-term monitoring, the electrodes are applied on the chestand not on the limbs. For normal routine ECG recordings, the patient must be advised to get warmand to relax so that muscle tremor from shivering or tension is eliminated.
The most critical component of the ECG recorder is the patient cable. The conventional PVC
insulation gets degraded and becomes rigid and breakable because of the arification of the softener.
Some manufacturers supply a patient-cable made of silicon-rubber, which provides better elasticity
over long periods.
/G35/G2E/G31/G2E/G34 /G4D/G69/G63/G72/G6F/G70/G72/G6F/G63/G65/G73/G73/G6F/G72/G20/G42/G61/G73/G65/G64/G20/G45/G43/G47/G20/G4D/G61/G63/G68/G69/G6E/G65/G73
Microprocessor technology has been employed in the electrocardiographs to attain certaindesirable features like removal of artefacts, baseline wander, etc. using software techniques.Automatic centering of the tracing is another feature which can be similarly achieved.
Microprocessor-based ECG machines can perform self-testing each time they are switched on.
These machines are programmed to check lead continuity and polarity and also indicate lead fall-off or reversal. Use of digital filters considerably improves signal quality during recording and
problems like baseline drift and excessive mains hum are thus automatically corrected. For this,
the programs are stored in EPROM to obtain good quality tracings. Minimising baseline driftwithout distorting the signal helps in monitoring the ECG of exercising or ambulatory subjects.

164 Handbook of Biomedical Instrumentation
The frequency components of ECG signals are low enough for the microprocessor to perform
reliable data acquisition, processing and display. For example, the highest frequency componentsin the waveform are in the QRS complex which typically last 60 to 100 ms. In this case, a sampling
rate of 200 samples per second, which will yield 12–20 points in the QRS complex, is considered
adequate (Hsue and Graham, 1976). Since the microprocessor instruction times are of the order of5ms, this permits approximately 1000 instructions between samples, which is sufficient for data
acquisition, processing and storage. In these cardiographs, the ECG is converted to digital form forwaveform preprocessing and then reconverted into analog form for display or telephonictransmission to the central computer. Using microprocessors, the need for the technician to switchfrom one lead to another during the recording process is eliminated Microprocessors have alsobeen used for simultaneous acquisition of multiple leads. Stored programs in the ROM direct theoperator about entering of patient data and display error codes. An important advantage ofmicroprocessor-based ECG machines is that it offers the potential for reducing the complexity ofanalysis algorithms by preprocessing the data. Chapter 7 details out methods of ECG analysis forautomatic determination of various kinds of arrhythmias.
/G35/G2E/G31/G2E/G35 /G4D/G75/G6C/G74/G69/G2D/G63/G68/G61/G6E/G6E/G65/G6C/G20/G45/G43/G47/G20/G4D/G61/G63/G68/G69/G6E/G65
Most of the electrocardiographs used for clinical purposes are single channel machines, i.e. themachine contains one amplifier channel and one recording system. Such machines usually carrya multi-position switch, by means of which the desired lead connection can be selected. Only onelead at a time can be recorded with such type of instruments.
Multi-channel ECG machines are also available. They carry several amplifier channels and a
corresponding number of recording pens. This facilitates recording of several ECG leadssimultaneously and thus considerably reduces the time required to complete a set of recordings.Another advantage of multi-channel recording is that the waveforms are recorded simultaneously
and they can be shown in their proper time relationship with respect to each other.
Modern multi-channel ECG machines use microprocessor to capture the heart signals from a
standard 12-lead configuration, sequencing the lead selector to capture four groups of three lead
signals and switching groups every few seconds. Figure 5.7 shows the block diagram of a three-channel microprocessor based ECG machine. The operating program controlling the lead selectionand other operations is stored in a ROM. The ECG signals selected by the microprocessor areamplified, filtered and sent to a three-channel multiplexer. The multiplexed analog signals arethen given to an analog-to-digital converter. For a 10 mV resolution referred to as the input (Fostik
et al, 1980), it is necessary to use a 10-bit A-D converter. Ten bits provide resolution of one part in
1024 (2
10 = 1024), which for a 10 mV peak-to-peak input range equals 10 mV. A suitable 10-bit A-D
converter is Analog Devices 7570. The maximum conversion rate of the device for 10-bit words is25000 per second. The sampling rate is usually 200–1000 samples/s. The microprocessor storesthe digitized signals in a RAM. The contents of the RAM are sent to a digital-to-analog converterfor reconstructing the analog signals (Shackil, 1981). The analog signals are demultiplexed and
passed to the video display or chart recorder.
We have seen that older versions of ECG machines recorded one lead at a time and then evolved
to three simultaneous leads. This necessitated the use of switching circuits to generate the various

Biomedical Recorders 165
12 leads. This is now eliminated in modern digital ECG machines by the use of an individual
single-ended amplifier for each electrode on the body. Each potential signal is then digitallyconverted and all the ECG leads can be calculated mathematically in software. This wouldnecessitate a 9- amplifier system. The machines make use of a 16-bit A–D converter, all within asmall amplifier lead stage. The digital signals are optically isolated and sent via a high speedserial link to the main ECG instrument. Here the 32-bit CPU and DSP (Digital Signal Processor)chip perform all the calculation and hard copy report is generated on a standard A-4 size paper.
These machines are therefore called page writers.
The recording system also operates digitally. The machine output is produced by an X-Y drive
mechanism which uses drive wheels to move the ECG chart paper in the paper axis direction
while moving a carriage mounted pen in the carriage axis direction. Each direction of movementis caused by identical low inertia dc servo motors with attached encoders to provide positionfeed back for the respective servo chips (integrated circuit). Drive from the servo motors is providedby toothed belts. The microprocessor controls the plotting process by sending plot commands tothe X- and Y-axis servo chips. These commands are in the form of a pulse train. The servo chipscontrol the motors and keep track of the position error between the current and the desired positionof the motor. Each of the motor drives is activated in the direction which reduces this positionerror. Motor position is fed back to the respective servo chip by a shaft encoder mounted on themotor. This position feedback decreases the servo chips position error as the motor approaches thedesired position. The responsible servo chip communicates to the microprocessor should a servoerror overflow occur, i.e. of a servo position has been lost. This is an irrecoverable error, whichA-D
converterD-A
converter
3-channel
preamp.
multiplexer
RAM
ROM3-channel
demultiplexer
3-channel
recorder
KeyboardAlpha
numeric
displayMICROPROCESSOR
Fig.5.7 Block diagram of a microprocessor based three channel ECG machine

166 Handbook of Biomedical Instrumentation
causes the microprocessor to shut down the whole recorder assembly and send back an error
message.
Modern ECG machines also incorporate embedded software for an automatic interpretation of
the ECG. These programs are quite sophisticated involving waveform recognition, calculation ofamplitude, duration, area and shape of every Pwave, QRS complex, Twave and STsegment in
every lead and rhythm analysis. The basis of these analysis are covered in Chapter 7.
/G20/G35/G2E/G32 /G56/G45/G43/G54/G4F/G52/G43/G41/G52/G44/G49/G4F/G47/G52/G41/G50/G48/G20/G28/G56/G43/G47/G29
Vectorcardiography is the technique of analyzing the electrical activity of the heart by obtaining
ECG’s along three axes at right angles to one another and displaying any two of these ECGs as avector display on an X-Y oscilloscope. The display is known as a vectorcardiogram (VCG). Incontrast, the electrocardiogram which displays the electrical potential in any one single axis, thevectorcardiogram displays the same electrical events simultaneously in two perpendicular axes.This gives a vectorial representation of the distribution of electrical potentials generated by theheart, and produces loop type patterns (Fig. 5.8) on the CRT screen. Usually a photograph is takenof each cardiac cycle. From such pictures, the magnitude and orientation of the P,Q,R,S and T
vector loops are determined.
Horizontal Frontal LT. sagittal204060
4020 20 60
40Vectorcardiogram [QRS T]
Fig.5.8 Typical normal loop patterns recorded in three planes on a direct writing
vectorcardiograph
VCG illustrates the phase differences between the voltages and also the various leads from
which it is derived. The major information that it provides is the direction of depolarization andrepolarization of the atria and the ventricles. Each vectorcardiogram exhibits three loops, showingthe vector orientation of the P wave, the QRS axis and the T wave. Because of the high amplitude
associated with QRS, loops from the QRS complex predominate. An increase in horizontal and
vertical deflection sensitivities is normally required to adequately display the loops resulting
from the P wave and Twave. Bourne (1974) describes circuit details of an automated vector ECG
recording system.
The VCG has been demonstrated to be superior to the standard 12-lead scalar electrocardiogram
in the recognition of undetected atrial and ventricular hypertrophy, sensitivity in identification of

Biomedical Recorders 167
myocardial infarction and capability for diagnosis of multiple infarctions in the presence of
fascicular and bundle branch blocks (Benchimol and Desser, 1975).
/G20/G35/G2E/G33 /G50/G48/G4F/G4E/G4F/G43/G41/G52/G44/G49/G4F/G47/G52/G41/G50/G48/G20/G28/G50/G43/G47/G29
The phonocardiograph is an instrument used for recording the sounds connected with thepumping action of the heart. These sounds provide an indication of the heart rate and itsrhythmicity. They also give useful information regarding effectiveness of blood pumping andvalve action.
Heart sounds are diagnostically useful. Sounds produced by healthy hearts are remarkably
identical and abnormal sounds always corelate to specific physical abnormalities. From thebeginning till today, the principal instrument used for the clinical detection of heart sounds is theacoustical stethoscope. An improvement over the acoustal stethoscope, which usually has lowfidelity, is the electronic stethoscope consisting of a microphone, an amplifier and a head set.Electronic stethoscopes can detect heart sounds which are too low in intensity or too high infrequency to be heard in a purely acoustal instrument. The phonocardiographs provide a recordingof the waveforms of the heart sounds. These waveforms are diagnostically more important andrevealing than the sounds themselves.
/G35/G2E/G33/G2E/G31 /G4F/G72/G69/G67/G69/G6E/G20/G6F/G66/G20/G48/G65/G61/G72/G74/G20/G53/G6F/G75/G6E/G64/G73
The sounds are produced by the mechanical events that occur during the heart cycle. These soundscan be from the movement of the heart wall, closure of walls and turbulence and leakage of bloodflow. A typical recording of these sounds is illustrated in Fig. 5.9. The first sound, whichcorresponds to the R wave of the ECG, is longer in duration, lower in frequency, and greater in
intensity than the second sound. The sound is produced principally by closure of the valvesbetween the upper and lower chambers of the heart, i.e. it occurs at the termination of the atrial
Mitral valve
closure
Ist soundTricuspad
valve closure3rd sound
termination of
ventricular filling
2nd sound and
aortic & pulmonary
valves closure4th sound
atria contract
Fig.5.9 Basic heart sounds in a typical phonocardiogram recording

168 Handbook of Biomedical Instrumentation
contraction and at the onset of the ventricular contraction. The closure of the mitral and tricuspid
valve contributes largely to the first sound. The frequencies of these sounds are generally in therange of 30 to 100 Hz and the duration is between 50 to 100 ms. The second sound is higher in
pitch than the first, with frequencies above 100 Hz and the duration between 25 to 50 ms. This
sound is produced by the slight back flow of blood into the heart before the valves close and thenby the closure of the valves in the arteries leading out of the ventricles. This means that it occurs atthe closure of aortic and the pulmonic valves.
The heart also produces third and fourth sounds but they are much lower in intensity and are
normally inaudible. The third sound is produced by the inflow of blood to the ventricles and thefourth sound is produced by the contraction of the atria. These sounds are called diastolic soundsand are generally inaudible in the normal adult but are commonly heard among children.
/G35/G2E/G33/G2E/G32 /G4D/G69/G63/G72/G6F/G70/G68/G6F/G6E/G65/G73/G20/G66/G6F/G72/G20/G50/G68/G6F/G6E/G6F/G63/G61/G72/G64/G69/G6F/G67/G72/G61/G70/G68/G79
Two types of microphones are commonly in use for recording phonocardiograms. They are thecontact microphone and the air coupled microphone. They are further categorized into crystaltype or dynamic type based on their principle of operation.
The crystal microphone contains a wafer of piezo-electric material, which generates potentials
when subjected to mechanical stresses due to heart sounds. They are smaller in size and moresensitive than the dynamic microphone.
The dynamic type microphone consists of a moving coil having a fixed magnetic core inside it.
The coil moves with the heart sounds and produces a voltage because of its interaction with themagnetic flux.
The phonocardiogram depends extensively on the technical design of the microphone, since it
does not transform the acoustic oscillations into electrical potential uniformly for all frequencies.Therefore, the heart sound recordings made with a microphone are valid only for that particulartype of microphone. As a consequence, microphones of various types cannot, as a rule, beinterchanged.
A new acoustic sensor, which enhances the audibility of heart sounds and enables recording of
quantitative acoustic spectral data is described by Kassal et al, 1994. This device is a polymer-
based adherent differential-output sensor, which is only 1.0 mm thick. The device is compliantand can be applied to the skin with gel and two-sided adhesive material, and can conform to thecontour of the patient’s body. The device can be used for phonocardiography, lung sounds andthe detection of Korotkoff sounds. The device is not a microphone and does not detect acoustic
pressure, rather it actually discriminates against it. Instead, the sensor detects the motion of the
skin that results from acoustic energy incident upon it from within the soft tissue. Its principlesensing components is PVDF (poly-vinylidene fluoride), which is a piezo-electric polymer. Itproduces charges of equal magnitude and opposite polarity on opposite surfaces when amechanical strain is imposed on the material. The voltage generated in the sensor due to theflexing motion forms the basis of electronic stethoscopes, and real time digital acoustic spectralanalysis of heart sounds.

Biomedical Recorders 169
/G35/G2E/G33/G2E/G33 /G41/G6D/G70/G6C/G69/G66/G69/G65/G72/G73/G20/G66/G6F/G72/G20/G50/G68/G6F/G6E/G6F/G63/G61/G72/G64/G69/G6F/G67/G72/G61/G70/G68/G79
The amplifier used for a phonocardiograph has wide bandwidth with a frequency range of about
20 to 2000 Hz. Filters permit selection of suitable frequency bands, so that particular heart soundfrequencies can be recorded. In general, the high frequency components of cardiovascular sound
have a much smaller intensity than the low frequency components and that much information of
medical interest is contained in the relatively high frequency part of this spectrum (Bekkering andVollenhoven, 1967). Therefore, high-pass filters are used to separate the louder low frequencycomponents from the soft and interesting high frequency murmurs. Experiments have shown thatthe choice of different filters does not have to be very critical but in general, sets of four or five high-pass filters with different cut-off frequencies and slopes are used in the commercially availableinstruments. PCG amplifiers usually have gain compensation circuits to increase the amplificationof high frequency signals, which are usually of low intensity. The frequencies at the higher end ofthe range are of particular significance in research applications.
The appropriate filter characteristics may be selected to attenuate the unwanted frequencies at
filter slopes of 12 dB/octave or 24 dB/octave. This is based on the fact that cardiac vibrationsfollow the inverse square law which is 12 dB/octave, i.e. as the frequency of the sound is increased,the intensity decreases approximately 12 dB/octave over a portion of the sound spectrum. The 12dB/octave approximation is valid from 50–200 Hz and the 24 dB/octave from 200–800 Hz. Below
50 Hz, 6 dB/octave is found to be the best choice. A filter with a 12 dB/octave slope causes the
intensity of the unwanted sounds to decrease to 0.25 times the original, when the frequency of thesound is one half its original value. Figure 5.10 shows heart sound amplifier characteristics.
0
–10
–20+10+20+30+40+50
2 4 100 2 4 1 K 2 K 6 K 10 K800 Hz
400 Hz
200 Hz
100 Hz
50 Hz
25 Hz6 db12 db
Frequency of heart sound (Hz)Relative amplitude db
Fig.5.10 Characteristics of amplifiers with commonly employed filters in
phonocardiography (Courtesy: Hewlett Packard, USA)

170 Handbook of Biomedical Instrumentation
/G35/G2E/G33/G2E/G34 /G57/G72/G69/G74/G69/G6E/G67/G20/G4D/G65/G74/G68/G6F/G64/G73/G20/G66/G6F/G72/G20/G50/G68/G6F/G6E/G6F/G63/G61/G72/G64/G69/G6F/G67/G72/G61/G70/G68/G79
In order to obtain a faithful reproduction of all the frequency components, the phonocardiogram
requires a recording system capable of responding to 2000 Hz. Light beam galvanometers have notrouble in meeting this demand. But these galvanometers are expensive and require more power
from the amplifiers when used for recording high frequencies, of the order specified above. Direct
writing recorders with an upper frequency response of about 150 Hz cannot be used to writefrequencies that lie beyond their working range. This drawback is overcome by using the“Envelope Detection” technique. The technique consists in using an artificial frequency, say100 Hz, in the heart sound amplifier. This is employed to oscillate the stylus so that the highfrequency sounds are modulated by 100 Hz. The shape of the sound wave is thus retained andreproduced with a response of 100 Hz. The heart sounds can now be recorded on a direct writingrecorder. This procedure, however, can only record heart sound intensity picked up every 10 ms,and events taking place within this time interval of 10 ms are not recorded.
Recording of phonocardiograms by direct recording methods has been made possible by the
ink jet recorder. Also, digital recorders such as the electrostatic recorder or the thermal recorder arealso suitable for phonographic recordings.
Many phonocardiograms also have a provision for recording the patient’s electrocardiogram
on the same chart. Simultaneous recording of the phonocardiogram and the electrocardiogramdisplays both the sounds and the electrical activity of the heart in their proper time relationship.This facilitates the work of the clinician to correlate sounds with the phase of the heart cycle
during which they occur. From such time correlations it is easier to identify, specifically, the
nature of the defect. The phonocardiogram is more informative than an electrocardiogram formonitoring heart valve actions.
While recording a phonocardiogram, the microphone picks up not only the sounds and
murmurs at the body surface, but also all extraneous noises in the immediate vicinity of thepatient. Therefore, common utilities like fans, air-conditioners and other noise producing gadgetsworking nearby create vibrations within the same frequency range as the heart sounds andmurmurs and will result in artefacts on the recording. This fact emphasizes the importance ofhaving a relatively quiet area for phonocardiography. The walls and ceiling of the room should beacoustic-tiled and the recording system placed on some cushioning material, to minimize externaland internal noise. It is advisable to secure the microphone to the patient’s chest with a strapbecause if it is hand held, its output may vary in accordance with the pressure applied. It is onlywith practice that one learns to apply a constant pressure to the microphone and the patient.
/G20/G35/G2E/G34 /G45/G4C/G45/G43/G54/G52/G4F/G45/G4E/G43/G45/G50/G48/G41/G4C/G4F/G47/G52/G41/G50/G48/G20/G28/G45/G45/G47/G29
Electroencephalograph is an instrument for recording the electrical activity of the brain, by suitablyplacing surface electrodes on the scalp. EEG, describing the general function of the brain activity,is the superimposed wave of neuron potentials operating in a non-synchronized manner in thephysical sense. Its stochastic nature originates just from this, and the prominent signal groups canbe empirically connected to diagnostic conclusions.

Biomedical Recorders 171
Monitoring the electroencephalogram has proven to be an effective method of diagnosing many
neurological illnesses and diseases, such as epilepsy, tumour, cerebrovascular lesions, ischemiaand problems associated with trauma. It is also effectively used in the operating room to facilitate
anaesthetics and to establish the integrity of the anaesthetized patient’s nervous system. This has
become possible with the advent of small, computer-based EEG analyzers. Consequently, routineEEG monitoring in the operating room and intensive care units is becoming popular.
Several types of electrodes may be used to record EEG. These include: Peel and Stick electrodes,
Silver plated cup electrodes and Needle electrodes.
EEG electrodes are smaller in size than ECG electrodes. They may be applied separately to the
scalp or may be mounted in special bands, which can be placed on the patient’s head. In eithercase, electrode jelly or paste is used to improve the electrical contact. If the electrodes are intendedto be used under the skin of the scalp, needle electrodes are used. They offer the advantage ofreducing movement artefacts. EEG electrodes give high skin contact impedance as compared toECG electrodes. Good electrode impedance should be generally below 5 kilohms. Impedancebetween a pair of electrodes must also be balanced or the difference between them should be lessthan 2 kilohms. EEG preamplifiers are generally designed to have a very high value of inputimpedance to take care of high electrode impedance.
EEG may be recorded by picking up the voltage difference between an active electrode on the
scalp with respect to a reference electrode on the ear lobe or any other part of the body. This type ofrecording is called ‘monopolar’ recording. However, ‘bipolar’ recording is more popular whereinthe voltage difference between two scalp electrodes is recorded. Such recordings are done with
multi-channel electroencephalographs.
EEG signals picked up by the surface electrodes are usually small as compared with the ECG
signals. They may be several hundred microvolts, but 50 microvolts peak-to-peak is the most
typical. The brain waves, unlike the electrical activity of the heart, do not represent the samepattern over and over again. Therefore, brain recordings are made over a much longer interval oftime in order to be able to detect any kind of abnormalities.
/G35/G2E/G34/G2E/G31 /G42/G6C/G6F/G63/G6B/G20/G64/G69/G61/G67/G72/G61/G6D/G20/G64/G65/G73/G63/G72/G69/G70/G74/G69/G6F/G6E/G20/G6F/G66/G20/G45/G6C/G65/G63/G74/G72/G6F/G65/G6E/G63/G65/G70/G68/G61/G6C/G6F/G67/G72/G61/G70/G68
The basic block diagram of an EEG machine with both analog and digital components is shown inFig. 5.11.
Montages: A pattern of electrodes on the head and the channels they are connected to is called a
montage. Montages are always symmetrical. The reference electrode is generally placed on a non-active site such as the forehead or earlobe. EEG electrodes are arranged on the scalp according toa standard known as the 10/20 system, adopted by the American EEG Society (Barlow et al, 1974).
Traditionally, there are 21 electrode locations in the 10/20 system. This system involves placement
of electrodes at distances of 10% and 20% of measured coronal, sagittal and circumferential arcsbetween landmarks on the cranium (Fig. 5.12). Electrodes are identified according to their positionon the head: Fpfor frontal-polar, F for frontal, C for central, P for parietal, Tfor temporal and Ofor
occipital. Odd numbers refer to electrodes on the left side of the head and even numbers representthose on the right while Z denotes midline electrodes. One electrode is labelled isoground and

172 Handbook of Biomedical Instrumentation
Electrodes
SubjectElectrode
montage
selector
Electrode
test/calibrate
JackboxAmplifiers
Filters
Hi-pass Low-pass Notch Sensitivity
Analog to
digital
converterWriter unit
Chart
driveInk-writing
oscillograph
Oscilloscope Computer
EEG
Fig.5.11 Schematic diagram of an EEG machine (after Isley et al, 1998)
C3
P3
O1O2
Inion
Top of Head Left side of HeadFront
Nassion
F3F7Fp1Pg1Pg2
Fp2
PzFz
P4F4F8
CzC4T4
T5T6T3A2A1
A1OrbitNasion
Inion Nasopharyngeal
leadP3P2C3F3
10%20%20%
20%
20%10%F2C2
F7
T3T5O1Fp1
Pg1Right Left
Ear lobeVertex
Fig.5.12 10-20 system of placement of electrodes
placed at a relatively neutral site on the head, usually the midline forehead. A new montage
convention has recently been introduced in which electrodes are spaced at 5% distances along thecranium. These electrodes are called closely spaced electrodes and have their own namingconvention ( Fig. 5.13).

Biomedical Recorders 173
FT7
TP7
P7P5
PO3PO1
O1Oz
I2O2POzPO2PO4P3P1PzP2P4P6P8F7
T7FT9
TP9F9
T9FT11Ch1LO1AF7AF3FP3SO1SP1IO1
AF1AFzFPzNz
AF2AF4FP2SO2SP2IO2
AF8
TP11F11
A1FC5
CP5F5
C5FC3
CP3F3
C3FC1
CP1F1
C1FCz
CPzFz
CzFC2
CP2F2
C2FC4
CP4F4
C4FC6
CP6F6
C6FT8
TP8F8
T8FT10
TP10F10F10
T10FT12
TP12F12F12
A2
Fig.5.13 Pictorial representation of closely spaced electrodes
Electrode Montage Selector: EEG signals are transmitted from the electrodes to the head box,
which is labelled according to the 10–20 system, and then to the montage selector. The montageselector on analog EEG machine is a large panel containing switches that allow the user to selectwhich electrode pair will have signals subtracted from each other to create an array of channels ofoutput called a montage. Each channel is created in the form of the input from one electrode minus
the input from a second electrode.
Montages are either bipolar (made by the subtraction of signals from adjacent electrode pairs)
or referential (made by subtracting the potential of a common reference electrode from each
electrode on the head). In order to minimize noise, a separate reference is often chosen for eachside of the head e.g. the ipsilateral ear. Bipolar and referential montages contain the same basicinformation that is transformable into either format by simple substration as long as all theelectrodes, including reference, are included in both montages and linked to one common reference.Many modern digital EEG machines record information referentially, allowing easy conversion toseveral different bipolar montages. The advantage of recording EEG in several montages is thateach montage displays different spatial characteristics of the same data.
Preamplifier: Every channel has an individual, multistage, ac coupled, very sensitive amplifier
with differential input and adjustable gain in a wide range. Its frequency response can be selectedby single-stage passive filters. A calibrating signal is used for controlling and documenting thesensitivity of the amplifier channels. This supplies a voltage step of adequate amplitude to theinput of the channels. A typical value of the calibration signal is 50 uV/cm.
The preamplifier used in electroencephalographs must have high gain and low noise
characteristics because the EEG potentials are small in amplitude. In addition, the amplifier musthave very high common-mode rejection to minimise stray interference signals from power lines

174 Handbook of Biomedical Instrumentation
and other electrical equipments. The amplifier must be free from drift so as to prevent the slow
movement of the recording pen from its centre position as a result of changes in temperature, etc.
EEG amplifiers must have high gain in the presence of unbalanced source resistances and dc
skin potentials at least up to 100 mV. Noise performance is crucial in EEG work because skinelectrodes couple brain waves of only a few microvolts to the amplifier. Each individual EEG
signal should be preferably amplified at the bedside. Therefore, a specially designed connector
box, which can be mounted near the patient, is generally employed with EEG machines. Thisensures the avoidance of cable or switching artefacts. The use of electrode amplifiers at the sitealso eliminates undesirable cross-talk effects of the individual electrode potentials. The connectorbox also carries a circuit arrangement for measuring the skin contact impedance of electrodes withac. Thus, poor electrode-to-skin contacts above a predetermined level can be easily spotted out.
Sensitivity Control: The overall sensitivity of an EEG machine is the gain of the amplifier multiplied
by the sensitivity of the writer. Thus, if the writer sensitivity is 1 cm/V, the amplifier must have anoverall gain of 20,000 for a 50 mV signal. The various stages are capacitor coupled. An EEG
machine has two types of gain controls. One is continuously variable and it is used to equalize thesensitivities of all channels. The other control operates in steps and is meant to increase or reducethe sensitivity of a channel by known amounts. This control is usually calibrated in decibels. Thegain of amplifiers is normally set so that signals of about 200 mV deflect the pens over their full
linear range. Artefacts, several times greater than this, can cause excessive deflections of the penby charging the coupling capacitors to large voltages. This will make the system unusable over aperiod depending upon the value of the coupling capacitors. To overcome this problem, most
modern EEG machines have de-blocking circuits similar to those used in ECG machines.
Filters: Just like in an ECG when recorded by surface electrodes, an EEG may also contain muscle
artefacts due to contraction of the scalp and neck muscles, which overlie the brain and skull. Theartefacts are large and sharp, in contrast to the ECG, causing great difficulty in both clinical and
automated EEG interpretation. The most effective way to eliminate muscle artefact is to advise the
subject to relax, but it is not always successful. These artefacts are generally removed using low-pass filters. This filter on an EEG machine has several selectable positions, which are usuallylabelled in terms of a time constant. A typical set of time constant values for the low-frequencycontrol are 0.03, 0.1, 0.3 and 1.0 s. These time constants correspond to 3 dB points at frequencies of5.3, 1.6, 0.53 and 0.16 Hz.
The upper cut-off frequency can be controlled by the high frequency filter. Several values can be
selected, typical of them being 15, 30, 70 and 300 Hz.
Some EEG machines have a notch filter sharply tuned at 50 Hz so as to eliminate mains
frequency interference. These however have the undesirable property of ‘ringing’ i.e. they producea damped oscillatory response to a square wave calibration waveform or a muscle potential. Theuse of notch filters should preferably be restricted to exceptional circumstances when all othermethods of eliminating interference have been found to be ineffective.
The high frequency response of an EEG machine will be the resultant of the response of the
amplifier and the writing part. However, the figure mentioned on the high frequency filter controlof most EEG machines generally refers to the amplifier. The typical frequency range of standardEEG machines is from 0.1 Hz to 70 Hz, though newer machines allow the detection and filtering

Biomedical Recorders 175
of frequencies up to several hundred Hertz. This may be of importance in some intracranial
recordings.
Noise: EEG amplifiers are selected for minimum noise level, which is expressed in terms of an
equivalent input voltage. Two microvolts is often stated as the acceptable figure for EEG recording.Noise contains components at all frequencies and because of this, the recorded noise increaseswith the bandwidth of the system. It is therefore important to restrict the bandwidth to that requiredfor faithful reproduction of the signal. Noise level should be specified as peak-to-peak value as it
is seen on the record rather than rms value, which could be misleading.
Writing Part: The writing part of an EEG machine is usually of the ink type direct writing recorder.
The best types of pen motors used in EEG machines have a frequency response of about 90 Hz.Most of the machines have a response lower than this, and some of them have it even as low as
45 Hz. The ink jet recording system, which gives a response up to 1000 Hz, is useful for some
special applications.
Paper Drive: This is provided by a synchronous motor. An accurate and stable paper drive
mechanism is necessary and it is normal practice to have several paper speeds available for
selection. Speeds of 15, 30 and 60 mm/s are essential. Some machines also provide speed values
outside this range. A time scale is usually registered on the record by one or two time marker pens,which make a mark once per second. Timing pulses are preferably generated independently of thepaper drive mechanism in order to avoid difference in timing marks due to changes in paperspeed.
Channels: An electroencephalogram is recorded simultaneously from an array of many electrodes.
The record can be made from bipolar or monopolar leads. The electrodes are connected to separateamplifiers and writing systems. Commercial EEG machines have up to 32 channels, although 8 or16 channels are more common.
Microprocessors are now employed in most of the commercially available EEG machines. These
machines permit customer programmable montage selection; for example, up to eight electrodecombinations can be selected with a keyboard switch. In fact, any desired combination of electrodescan be selected with push buttons and can be memorized. These machines also include a videomonitor screen to display the selected pattern (montage) as well as the position of scalp sites withelectrode-to-skin contact. Individual channel control settings for gain and filter positions can bedisplayed on the video monitor for immediate review. Therefore, a setting can be changed by asimple push button operation while looking at the display.
Modern EEG machines are mostly PC based, with a pentium processor, 16-MB RAM, atleast a
2 GB hard disk, cache memory and a 4 GB DAT tape drive. The system can store up to 40 hours ofEEG. The EEG is displayed on a 43 cm colour monitor with a resolution of 1280 ¥ 1024 pixels. The
user interface is through an ASCII keyboard and the output is available in the hard copy form
through a laser printer.
/G35/G2E/G34/G2E/G32 /G52/G65/G63/G6F/G72/G64/G69/G6E/G67/G20/G6F/G66/G20/G45/G76/G6F/G6B/G65/G64/G20/G50/G6F/G74/G65/G6E/G74/G69/G61/G6C/G73
If an external stimulus is applied to the a sensory area of brain, it responds by producing anelectrical potential known as the ‘evoked potential’. The most frequently used evoked potentials

176 Handbook of Biomedical Instrumentation
for clinical testing include brainstem auditory evoked responses, visual evoked responses and
somatosensory evoked potentials.
Evoked potential, recorded at the surface of the brain, is the integrated response of the action of
many cells. The amplitude of the evoked potential is of the order of 10 microvolts. The evokedpotentials are generally superimposed with electroencephalograms. Therefore, it is necessary to
remove the EEG by an averaging technique while making evoked potential measurements. Since
the background EEG and other unwanted signals often appear irregular, or do not synchronize tostimuli, they are markedly reduced by averaging across multiple trials. In general, averagingreduces noise proportional to the square route of the number of trials. Most of the improvements insignal-to-noise ratio occur within 40 to 100 trials.
Since many evoked potential components are of short duration, about 2 ms to 1 sec., rapid
sampling rates are needed to digitally record such low level potentials. Usually, the sampling rateis 1000/second. The amplitude of evoked potentials are normally measured on a vertical scalewith sample points measured as bits on a logarthmic resolution scale. Resolution of voltage isusually sufficient with a 8-bit recording, although 10-to 16-bit A/D systems are becoming available.
/G35/G2E/G34/G2E/G33 /G43/G6F/G6D/G70/G75/G74/G65/G72/G69/G7A/G65/G64/G20/G41/G6E/G61/G6C/G79/G73/G69/G73/G20/G6F/G66/G20/G45/G45/G47
Assessment of the frequency and amplitude of the EEG is crucial for rapid and accurateinterpretation. This involves the need for constant analysis of the EEG signal by a skilledtechnician and the acquisition of volumes of recording paper. Therefore, modern machines makeuse of computerized EEG signal processing to extract and present the frequency and amplitudeinformation in simple, visually enhanced formats that are directly useful to the clinician (Isleyet al, 1998).
Frequency Analysis: It takes the raw EEG waves, mathematically analyzes them and breaks them
into their component frequencies. The most popular method of doing this is called the Fast-FourierTransform or FFT.
Fast-Fourier Transformation of the digitized EEG waveform is a mathematical transformation
of a complex waveform (having varying frequency and amplitude content) into simpler, moreuniform waveforms (such as different sine waves of varying amplitudes). In this method, the EEGsignal is converted into a simplified waveform called a spectrum. It is then separated intofrequency bands at intervals of 0.5 Hz over a range of 1 to 32 Hz. The re-distribution of electricalactivity in the brain among certain frequency bands or the predominance of one band over theothers correlates with specific physiologic and pathologic conditions (Fig. 5.14(a)). The spectral
analysis transforms the analog EEG signal recorded on the time axis into a signal displayed on the
frequency axis.
Amplitude Analysis: Changes in the EEG amplitude can indicate clinical changes. The amplitude
changes result in changes in the power of the resulting frequency spectrum. As the amplitude
increases, so does the power. The most common number reflecting EEG amplitude is the total
power of the EEG spectrum. Due to the microvolt amplitude of the EEG, power is either innanowatts or picowatts. The power spectrum is calculated by squaring the amplitudes of the
individual frequency components. The powers of the individual frequency bands are also

Biomedical Recorders 177
commonly used and expressed as an absolute or a percentage of the total power. For example, 25%
Alpha would indicate that 25% of the total power is derived from the amplitudes of the Alphawaves.
Several different display formats have been developed for visually enhancing the computer-
processed information. These are
Compressed Spectral Array (CSA): In this format, a series of computer-smoothed spectral arrays
are stacked vertically, usually at two second intervals, with the most recent EEG event at thebottom and the oldest at the top. Peaks appear at frequencies, which contain more power or makelarger contributions to the total power spectrum. Since the origin of the plots shifts vertically with
time, this produces a pseudo three-dimensional graph (Fig. 5.14(b)). With this method, it is easy to
pick up changes in frequency and amplitude of each sample over a longer period of time as itcompresses a large amount of data into a compact, easy to read trend.
4
4theta delta alpha 8
812
1216
16Hz
Hz
HzEEG
signal
Signal
analyzed
(spectra)
Smoothened
waveform
(a)
(b)424681012
0 8 12 16 20 24 28 325¢[]t
Fig.5.14 (a) Typical ECG waveform broken down into frequency components
(b)Mathematical and display techniques used to generate the compressed
spectral array format

178 Handbook of Biomedical Instrumentation
Dot-density modulated Spectral Array (DSA): It is another method for displaying the power spectra.
This format displays a power spectrum as a line of variable intensities and/or densities withsuccessive epochs again stacked vertically as in the CSA plots. Areas of greatest density represent
frequencies, which make the greatest contribution to the EEG power spectrum. An advantage of
the DSA format is that no data is hidden by the peaks as in the CSA display. DSA displays couldbe in the form of gray or colour-scaled densities.
/G20/G35/G2E/G35 /G45/G4C/G45/G43/G54/G52/G4F/G4D/G59/G4F/G47/G52/G41/G50/G48/G20/G28/G45/G4D/G47/G29
Electromyograph is an instrument used for recording the electrical activity of the muscles todetermine whether the muscle is contracting or not; or for displaying on the CRO and loudspeakerthe action potentials spontaneously present in a muscle or those induced by voluntary contrac-tions as a means of detecting the nature and location of motor unit lesions; or for recording theelectrical activity evoked in a muscle by the stimulation of its nerve. The instrument is useful formaking a study of several aspects of neuromuscular function, neuromuscular condition, extent ofnerve lesion, reflex responses, etc.
EMG measurements are also important for the myoelectric control of prosthetic devices (artificial
limbs). This use involves picking up EMG signals from the muscles at the terminated nerve endings
of the remaining limb and using the signals to activate a mechanical arm. This is the most
demanding requirement from an EMG since on it depends the working of the prosthetic device.
EMG is usually recorded by using surface electrodes or more often by using needle electrodes,
which are inserted directly into the muscle. The surface electrodes may be disposable, adhesivetypes or the ones which can be used repeatedly. A ground electrode is necessary for providing acommon reference for measurement. These electrodes pick up the potentials produced by thecontracting muscle fibres. The signal can then be amplified and displayed on the screen of acathode ray tube. It is also applied to an audio-amplifier connected to a loudspeaker. A trainedEMG interpreter can diagnose various musculardisorders by listening to the sounds producedwhen the muscle potentials are fed to the loud-speaker. The block diagram (Fig. 5.15) shows atypical set-up for EMG recordings. The oscillo-
scope displays EMG waveforms. The tape
recorder is included in the system to facilitateplayback and study of the EMG sound waveformsat a later convenient time. The waveform canalso be photographed from the CRT screen byusing a synchronized camera.
The amplitude of the EMG signals depends upon various factors, e.g. the type and placement of
electrodes used and the degree of muscular exertions. The needle electrode in contact with a singlemuscle fibre will pick up spike type voltages whereas a surface electrode picks up manyoverlapping spikes and therefore produces an average voltage effect. A typical EMG signal rangesfrom 0.1 to 0.5 mV. They may contain frequency components extending up to 10 kHz. Such high
Oscilloscope
Tape
recorder
A.F.
amp.SpeakerEMG
amp.Input
Fig.5.15 Block diagram of a typical set-up
for EMG recording

Biomedical Recorders 179
frequency signals cannot be recorded on the conventional pen recorders and therefore, they are
usually displayed on the CRT screen.
Modern EMG machines are PC based (Fig. 5.16) available both in console as well as laptop
models. They provide full colour waveform display, automatic cursors for marking and making
Fig.5.16 PC based digital EMG recording and reviewing system for 2 to 4 channels. It
includes a pentium processor, hard disk, DAT-tape or optical disk
storage, laser printer, high resolution colour monitor (Courtesy: M/s Cadwell,USA)

180 Handbook of Biomedical Instrumentation
measurements and a keyboard for access to convenient and important test controls. The system
usually incorporates facilities for recording of the EMG and evoked potentials. The stimulators aresoftware controlled. For report generation in the hard copy form, popular laser printers can be
used.
Preamplifier: The preamplifiers used for EMG are generally of differential type with a good
bandwidth. The input impedance of the amplifier must be greater than 2 ¥ 50 M W. Present day
electronic devices easily provide input impedances of the order of 10
12 ohms in parallel with
5 picofarads. It is preferable to mount the preamplifiers very near the subject using very smallelectrode leads, in order to avoid the undesirable effects of stray capacitance between connectingcables and the earth. Also, any movement of the cable from the output of the electrode will notgenerate significant noise signals in the cable, which feeds into the subsequent amplifier. Thepreamplifier provides an output with low impedance and, therefore, the high frequencies do notget attenuated even if long cables are used to connect the preamplifier and the rest of the machine.The common-mode rejection should be greater than 90 dB up to 5 kHz. A calibrating square wavesignal of 100 mV (peak-to-peak) at a frequency of 100 Hz is usually available. The main amplifier
has controls for gain adjustment from 5 mV/div to 10 mV/div for selecting the sensitivity mostappropriate to the incoming signal from the patient.
Basmajian and Hudson (1974) suggested the use of a preamplifier to amplify the EMG signals
picked up by needle electrodes at the electrode site before transmitting them along wires. The effect
of electrical interference is substantially reduced and the microphonic artefacts generated in the
wires due to movement of the subject are virtually eliminated. When surface EMGs are to bemeasured, it is convenient to combine the electrode pair and a differential amplifier within a singlemodule. Johnson et al (1977) designed a miniature amplifier circuit fully encapsulated in epoxy
resin with two small silver electrodes of 6 mm diameter, exposed flush with the base of the module.The electrode is attached to the skin using adhesive tape. Figure 5.17 shows a circuit diagram of
IC2IC1
IC3R3R6
R7R4
R2R1
R10R5
C1R8
R9D1
D2+V
0V
–VV+V–
OutputElectrode
Electrode
Fig.5.17 Pre-amplifier circuit for an EMG machine (redrawn after Johnson et al, 1977;
by permission of Med. and Bio. Eng. and Comp.)

Biomedical Recorders 181
the preamplifier. The amplifier design provides for a flat frequency response between 10 Hz and
1 kHz, with a CMRR of 100 dB at the mains frequency. The noise level was found to be 2 mV rmsand the input impedance greater than 10 M W.
The two ICs in the input stage act as voltage followers, which present the desired high input
impedance to the electrodes. They are coupled via C
1 and R5to provide a high differential signal
gain. Capacitor C1 determines the low frequency performance of the circuit. It also eliminates the
effects, at the output, of any dc offsets due to IC1,and IC2 or any imbalance in electrode potentials.
The second stage IC3 provides further differential signal gain, while rejecting common-mode
signals. The overall gain of the amplifier is 1000.
Input impedance of the amplifier must be higher by several orders than the electrode impedance.
Also, selection of the electrode type without the knowledge of the amplifier’s input resistance
results in distorted records and considerable error. The larger the surface of the electrode, the lessinput resistance is allowable. For example, a needle electrode with a surface of 15,000 mm
2may
need an amplifier with input impedance of 5 M W, while a needle electrode with a surface of
500mm2 will ensure a record with acceptable distortion by means of an amplifier with minimum
input impedance of 100 M W.
The capacitance present parallel with the input resistance of the amplifier reduces the frequency
response of the amplifier as well as lowers the common-mode rejection at higher frequencies.Owing to these, the electrode cable, the extension cable and the input stage of the amplifier requirecareful designing. Generally, shielded cables are used which reduce the disturbing signals but atthe same time, the parasite capacitance will increase. By careful design, a capacitance value of
50 pF or less can be achieved for the input stage. McRobbie (1990) illustrate a rapid recovery EMG
preamplifier without AC coupling capacitors.
To ensure patient safety, the subject should be electrically isolated from any electrical connection
to the power line or ground. This isolation is achieved either through the use of optical isolators orthrough the use of isolation transformers.
Low Frequency and High Frequency Filters: These are used to select the passband of the incoming
signal and to modify the progressive reduction in voltage output which occurs at either end of the
frequency spectrum roll-off. The low frequency 3 dB point may be selected over the range of 0.016to 32 Hz while the high frequency 6 dB point can be selected over the range 16 Hz to 32 kHz. Thus,the passband may be varied over a very wide range but is normally made as narrow as possible,subject to the requirements of the particular application in order to restrict displayed noise.
Signal Delay and Trigger Unit: Sometimes, it is necessary to examine the signals from individual
fibres of muscle tissue. For this purpose, special needles are available with a 25 micron diameterelectrode surface and up to 14 pick-up surfaces down the side of one needle. These 14 points arescanned sequentially to determine which point is acquiring the largest signal. This point is thenconsidered as the reference and its signal is used to trigger the sweep. Signals from the remaining13 points are then scanned sequentially and recorded with respect to the reference signal. Toexamine these signals, it is necessary to trigger the sweep from the signal and to delay the signalso that the whole of its leading edge is displayed. The delay is achieved by passing the digitizedsignal through the shift register or random access memory into the recirculate mode to obtain anon-fade display of a transient phenomenon.

182 Handbook of Biomedical Instrumentation
EMG machines have a provision for the selection of sweep speeds from 0.05–500 ms per division
on the CRT.
Integrator: The integrator is used for quantifying the activity of a muscle. Lippold (1952)
established that a linear relationship exists between the integrated EMG and the tension producedby a muscle. The integrator operates by rectifying an incoming EMG signal, i.e. by converting allnegative potentials to identical positive deflections so that the EMG pattern consists of positivedeflections only. The area under the rectified potentials is accumulated using a low-pass filter so
that the module output, at any time, represents the total area summed from a selected starting time.
The integrator indicates the EMG activity either as a variable frequency saw tooth waveform or asa steady deflection. In the former case, the output curve is a measure of the total electrical activityper second, recorded from a muscle during voluntary contraction within the analysis time. Theslope of this curve, measured as the number of resets per second, can be used to detect changes inthe number of motor units firing over a period of time. The steady deflection or mean voltage modeis used in plotting mean voltage v. isometric tension curves of muscle interference patterns duringvoluntary contraction, to show changes in muscle activity due to neuro-muscular disease such asmuscular dystrophy, poliomyelitis, etc. Different time constants determine the amount ofsmoothing applied to the output signal. When rapid changes in the signal have to be followed, theshortest time constant provides maximum smoothing of the signal and the most easily read meanvalue.
Stimulators: The stimulators incorporated in the EMG machines are used for providing a single
or double pulse or a train of pulses. Stimulus amplitude, duration, repetition and delay are alladjustable and facilities are provided for external triggering. The output is either of the constantvoltage type or of the constant current type. The constant voltage type stimulator provides square
wave pulses with amplitudes in the range of 0–500 V, a pulse duration of 0.1–3 ms and frequency
between 0-100 Hz. Output of the constant current generator can be adjusted between 0 to 100 mA.
Usually, the electromyographic changes in an advanced diseased state are readily recognized
on an oscilloscope display and by the sound from a loudspeaker. However, since the loss ofmuscle fibres, and therefore, the action potential changes are relatively small in early or milddisease states, changes in the EMG signals may be obscured by the usual variability of actionpotentials. Quantitative analysis is thus necessary to determine when the waveforms havechanged beyond the normal range. Quantities measured for such analysis include zero-crossingrate, peak rate, negative wave duration and wave rise time. These time-domain techniques aresomewhat different from the classical frequency spectrum and correlation function methods, butare much simpler to implement with electronic techniques. Fusfeld (1978) details out circuits usedto implement quantitative analysis of the electromyogram.
/G20/G35/G2E/G36 /G4F/G54/G48/G45/G52/G20/G42/G49/G4F/G4D/G45/G44/G49/G43/G41/G4C/G20/G52/G45/G43/G4F/G52/G44/G45/G52/G53
/G35/G2E/G36/G2E/G31 /G41/G70/G65/G78/G63/G61/G72/G64/G69/G6F/G67/G72/G61/G70/G68
An apexcardiograph records the chest-wall movements over the apex of the heart. These
movements are in the form of vibrations having a frequency range of 0.1 to about 20 Hz. Thetransducer required for recording these movements is similar to that employed for a phonocard-

Biomedical Recorders 183
diagraph (PCG) but which has a frequency response much below the audio range. It can be an air-
coupled microphone or a contact microphone. The apexcardiograph has limited applications. Itis, however, useful in the diagnosis of the enlargement of the heart chambers and some type of
valvular disorders.
/G35/G2E/G36/G2E/G32 /G42/G61/G6C/G6C/G69/G73/G74/G6F/G63/G61/G72/G64/G69/G6F/G67/G72/G61/G70/G68/G20/G28/G42/G43/G47/G29
A ballistocardiograph is a machine that records the movement imparted to the body with each
beat of the heart cycle. These movements occur during the ventricular contraction of the heartmuscle when the blood is ejected with sufficient force.
In BCG, the patient is made to lie on a table top which is spring suspended or otherwise
mounted to respond to very slight movements along the head axis. Sensing devices are mountedon the table to convert these movements into corresponding electrical signals. The sensors usuallyare piezo-electric crystals, resistive elements or permanent magnets, moving with respect to fixedcoils. In all such cases, the output of the sensor is amplified and fed to an oscilloscope or to a chartrecorder. BCG has so far been used mainly for research purpose only. It is rarely used in routineclinical applications.
/G35/G2E/G36/G2E/G33 /G45/G6C/G65/G63/G74/G72/G6F/G2D/G6F/G63/G75/G6C/G6F/G67/G72/G61/G70/G68/G20/G28/G45/G4F/G47/G29
Electro-oculography is the recording of the bio-potentials generated by the movement of the eyeball. The EOG potentials are picked up by small surface electrodes placed on the skin near the eye.One pair of electrodes is placed above and below the eye to pick up voltages corresponding tovertical movements of the eye ball. Another pair of electrodes is positioned to the left and right ofthe eye to measure horizontal movement. The recording pen is centred on the recording paper,corresponding to the voltage changes accompanying it. EOG has applications mostly for researchand is not widely used for clinical purposes.
/G35/G2E/G36/G2E/G34 /G45/G6C/G65/G63/G74/G72/G6F/G72/G65/G74/G69/G6E/G6F/G67/G72/G61/G70/G68/G20/G28/G45/G52/G47/G29
It is found that an electrical potential exists between the cornea and the back of the eye. Thispotential changes when the eye is illuminated. The process of recording the change in potential
when light falls on the eye is called electroretinography. ERG potentials can be recorded with a
pair of electrodes. One of the electrodes is mounted on a contact lens and is in direct contact withthe cornea. The other electrode is placed on the skin adjacent to the outer corner of the eye. Areference electrode may be placed on the forehead. A general purpose direct writing recorder maybe used for recording electroretinograms.
The magnitude of the ERG voltage depends upon the intensity and duration of the light falling
on the eye. It may be typically about 500 mV.
/G20/G35/G2E/G37 /G42/G49/G4F/G46/G45/G45/G44/G42/G41/G43/G4B/G20/G49/G4E/G53/G54/G52/G55/G4D/G45/G4E/G54/G41/G54/G49/G4F/G4E
Feedback is a common engineering term and refers to its function to control a process. When this
concept is applied to biological processes within the body, it is known as biofeedback. Here again,

184 Handbook of Biomedical Instrumentation
biofeedback is a means for gaining control of the body processes to create a specially required
psychological state so as to increase relaxation, relieve pain and develop healthier and morecomfortable life patterns. The technique involves the measurement of a variable produced by the
body process and compares it with a reference value. Based on the difference between the measured
and reference value, action is taken to bring the variable to the reference value.
It may be noted that biofeedback is not a treatment. Rather, biofeedback training is an
educational process for learning specialized mind/body skills. Through practice, one learns torecognize physiological responses and to control them rather than having them control us. Theobjective of biofeedback training is to gain self-regulatory skills which help to adjust the activityin various systems to optimal levels.
Many different physiological processes have been evaluated for possible control by biofeedback
methods. However, the following four neural functions are commonly employed:
• Emotions or Electrodermal Activity (Galvanic skin response measurements)• Muscle tension or EMG (Electromyograph measurements)• Temperature/sympathetic pattern (Thermistor readings)• Pulse (Heart rate monitoring)
Electrodermal activity is measured in two ways: BSR (basal skin response) and GSR (galvanic
skin response) is a measure of the average activity of the sweat glands and is a measure of thephasic activity (the high and low points) of these glands. BSR gives the baseline value of the skin
resistance where as GSR is due to the activity of the sweat glands. The GSR is measured most
conveniently at the palms of the hand, where the body has the highest concentration of sweatglands. The measurement is made using a dc current source. Silver-silver electrodes are used tomeasure and record the BSR and GSR. Figure 5.18 shows the arrangement for measuring theseparameters. The BSR output is connected to an RC network with a time constant of 3 to 5 secondswhich enables the measurement of GSR as a change of the skin resistance.
dc current
source
Neutral
electrodeActive
electrodeA1A2
BSR
meterGSR
meterRC
Fig.5.18 Block diagram for measurement and record of Basal Skin Resistance (BSR)
and Galvanic Skin Response (GSR)

Biomedical Recorders 185
Biofeedback instrumentation for the measurement of EMG, temperature and pulse/heart rate
is not different from other instruments used for the measurement of physiological variables.Transducers and amplifiers are employed to measure the variable that is to be controlled by
the feedback process. The magnitude of the measured variable or changes in the magnitude are
converted into a suitable visual or auditory stimulus that is presented to the subject. Based on thestimulus, the subject learns to control the abnormal conditions. Reports have appeared in literatureregarding applications of biofeedback to control migraine headaches, to slow down heart rate, etc.Biofeedback techniques have been greatly refined and computerized biofeedback training andpsychological computer-assisted guidance programs in the privacy of one’s home are now areality.

186 Handbook of Biomedical Instrumentation
/G50/G61/G74/G69/G65/G6E/G74/G20/G4D/G6F/G6E/G69/G74/G6F/G72/G69/G6E/G67/G20/G53/G79/G73/G74/G65/G6D/G73
/G20/G36/G2E/G31 /G53/G59/G53/G54/G45/G4D/G20/G43/G4F/G4E/G43/G45/G50/G54/G53
The objective of patient monitoring is to have a quantitative assessment of the important physiological
variables of the patients during critical periods of their biological functions. For diagnostic andresearch purposes, it is necessary to know their actual value or trend of change. Patient monitoringsystems are used for measuring continuously or at regular intervals, automatically, the values ofthe patient’s important physiological parameters. There are several categories of patients who
may need continuous monitoring or intensive care. Critically ill patients recovering from surgery,
heart attack or serious illness, are often placed in special units, generally known as intensive care
units , where their vital signs can be watched constantly by the use of electronic instruments. The
long-term objective of patient monitoring is generally to decrease mortality and morbidity by:(i) organizing and displaying information in a form meaningful for improved patient care,(ii) correlating multiple parameters for clear demonstration of clinical problems, (iii) processingthe data to set alarms on the development of abnormal conditions, (iv) providing information,based on automated data, regarding therapy and (v) ensuring better care with fewer staff members.
During a surgical operation, the patient is deprived of several natural reaction mechanisms,
which normally restore abnormalities in his physical condition or alert other people. Indicationsor alarms that cannot be given by the patient himself can be presented by patient monitoringequipment. Besides this, in special cases, it is not uncommon for surgical procedures to last forseveral hours. During these lengthy operative procedures, it is difficult for the anaesthesiologistand the surgeon to maintain intimate contact with the patient’s vital signs and at the same time
attend to anaesthesia, surgery, fluid therapy and many other details that are required under such
circumstances. Also, when a patient is connected to a life-support apparatus, e.g. heart-lungmachine or ventilator, correct functioning of these has to be monitored as well. A patientmonitoring system thus better informs the surgeon and the anaesthesiologist of the patient’scondition. With patient monitoring systems, the risk that surgery involves has been considerablyreduced since it is possible to detect the complications before they prove dangerous as suitableremedial measures can be taken well in time.HAPTER
66

Patient Monitoring Systems 187
The choice of proper parameters, which have a high information content, is an important issue
in patient monitoring. It is, however, generally agreed that monitoring of the following biological
functions is often needed. Electrocardiogram (ECG), heart rate (instantaneous or average), pulse
rate, blood pressure (indirect arterial blood pressure, direct arterial blood pressure or venousblood pressure), body temperature and respiratory rate. In addition to these primary parameters,
electroencephalogram (EEG), oxygen tension (pO
2) and respiratory volume also become part of
monitoring in special cases. In addition to these, equipment such as defibrillators and cardiacpacemakers are routinely needed in the intensive care wards.
The general requirements for patient monitoring equipment have not changed much over the
past few decades. However, today’s equipment monitors more parameters and processes more
information. Trends in monitoring include software control, arrhythmia monitoring, haemodynamics
monitoring, monitoring during transportation of the patient and increased user friendliness. Withmore than 10 parameters to be monitored and scores of calculations to be made, the requirement
for an easy-to-use user interface has assumed great significance.
Monitoring is generally carried out at the bedside, central station and bedside with a central
display. The choice amongst these is dependent upon medical requirements, available space andcost considerations.
/G20/G36/G2E/G32 /G43/G41/G52/G44/G49/G41/G43/G20/G4D/G4F/G4E/G49/G54/G4F/G52
The most important physiological parameters monitored in the intensive care unit are the heartrate and the morphology or shape of the electrical waveform produced by the heart. This is done to
observe the presence of arrhythmias or to detect changes in the heart rate that might be indicativeof a serious condition. Thus, a cardiac monitor is specifically useful for monitoring patients withcardiac problems and the special areas in the hospitals where they are generally used are known
as cardiac care units or coronary care units (CCU). These instruments are also called ‘Cardioscopes’
and comprise of:
• Disposable type pregelled electrodes to pick up the ECG signal.
• Amplifier and a cathode ray tube (CRT) for the amplification and display the ECG which
enable direct observation of the ECG waveform.
• A heart rate meter to indicate average heart rate with audible beep or flashing light or both
with each beat.
• An alarm system to produce signal in the event of abnormalities occurring in the heart rate.
The cardioscope is basically similar to the conventional oscilloscope used for the display of
waveforms in electronic laboratories. They have the usual circuit blocks like vertical and horizontalamplifiers, the time base and the EHT (extra high tension) for the cathode ray tube. However, theydiffer in two important aspects as compared to the conventional instrument. These are slower
sweep speeds and a long persistence screen. The slow sweep is an outcome of the low frequency
character of the ECG signal. The slow sweep speed necessitates the use of a long persistencescreen so as to enable a convenient observation of the waveform. Without a long persistencescreen, one can only see a moving dot of light instead of a continuous trace. Typically for a 13-cm
screen, total sweep time is usually kept as 2.5 or 5 s. In this way, one can observe at least four heartbeats in a single sweep period.

188 Handbook of Biomedical Instrumentation
Most of the present day cardioscopes are designed to be used at the bedside. Some of them are
even portable and can work on storage batteries. A large screen with about 50 cm screen sizeinstruments are usually mounted in one corner of the operating room at a height at which it is
possible to conveniently observe the waveforms being displayed. Small cardioscopes using 3 ≤
diameter cathode ray tubes are mounted on anaesthesia trollies. These are called “Anaesthesiamonitors”. These monitors are use by the anaesthetist for continuous monitoring of the ECG ofanaesthetized patients.
Low frequency waveforms, which are available from various physiological parameters do not
show up well on a conventional oscilloscope CRT screen because the scan is so low that most ofthe wave-track is dark during the scan. This is particularly true in large screen displays. Non-fademonitors using digital memories have been developed to overcome the problem of the fading ofslow scanning CRT displays. By this technique, it is possible to generate a rolling waveformdisplay that produces an effect comparable to a pen write-out. The display is thus continuous,bright and flicker-free on a normal non-storage CRT. If required, the image can be held indefinitely,selectively erased, allowed to roll across the screen or made to simulate a normal non-memoryCRT according to the sequence of operations in the read-out circuit. These type of displays areespecially suitable for patient monitoring applications because the stored information can be
readily transmitted to a remote viewing station or a chart recorder without losing track of incoming
signals.
With conventional oscilloscopes, there is an unavoidable break in the chain of information
between the end of sweep and the beginning of the next, because of sweep circuit retrace and hold-off requirements. Furthermore, this lost display usually extends while the oscilloscope waits forthe next trigger. If the train of events being observed is sine wave, the loss of information during thehold-off period may be of no consequence. But if, for example, we are looking for arrhythmias in atape recorded ECG display with a conventional oscilloscope, a complete, QRS complex couldeasily be missed. By using non-fading type roll-mode display, waveforms can be paraded acrossthe CRT screen in a continuous stream with no disruptions of any kind. The viewing windowdepends upon the sweep rate. Figure 6.1 shows the same waveform train displayed by aconventional oscilloscope and by the memory monitor in the roll-mode. Moreover, the ability to‘freeze’ the display in the roll mode at any point makes it easy to capture important data, whichcan be outputted to a chart recorder.
Figure 6.2 shows the basic system for incorporating digital storage in the oscilloscopic displays.
Essentially, the system carries out high-speed real-time sampling of an incoming analog waveform,
followed by the digital measurement of each successive sample and the subsequent storage of the
stream of data. Once the data is stored in the digital form, it can be recalled for conversion back toanalog form or for other processing operations. This ‘replay’ process can be continuous and itsspeed can be chosen to provide a non-fading, flicker-free trace on the CRT, irrespective of the speedof the original recording, or to provide a low-speed output to drive a conventional chart recorder.
Any digital method of waveform recording will have an analog–digital converter, which feeds
data corresponding to the input signal into a digital store in a controlled ‘write’ cycle. The data isretrieved via a similar controlled ‘read’ cycle and is reformed via a digital/analog converter fordisplay. As the regenerated signal is based on a finite number of measurements of the input signal,it is inevitably degraded as compared to the original. Two important factors governing the final

Patient Monitoring Systems 189
resolution are the sample rate and word length. The former must be high enough to provide
sufficient resolution on the time axis, while the latter depends on the number of bits provided bythe analog/digital converter or store which determines the number of levels between zero and fullscale on the vertical axis (Y-axis). In actual operation, the selected trigger signal initiates a scanLost data
Conventional displaySweep rate 0.1 sec/div; equivalent to 50 mm/sec chart speed
Moving display
Time window at 0.2 sec/div
(in SAVE)Time window
0.1 sec./div.
0.2 sec./div.
0.5 sec./div.
Roll mode no lost dataConventional
sweep
Fig.6.1 This figure illustrates, how the roll mode in digital storage can capture signals
without any loss of data. The time window is selectable (Courtesy: Tektronix,USA)
Input
amp.A-D
converterD-A
converterDigital
storeY
output
amp.
CRT.X output
amp.Sweep
gen.Control
logicTrigger
amp.Y
input
Fig.6.2 Block schematic of an oscilloscope display system incorporating digital
storage

190 Handbook of Biomedical Instrumentation
and writes in sequence into each address of the 1024 word length store. The writing sequence
depends upon the instructions from the 10-stage write address counter, which is controlled by the
time base speed control. The write-cycle control logic is designed to update or refresh the storewhenever a trigger pulse is received and hence the display follows changes in the input waveform
as they occur. A separate read address counter continuously scans all addresses at a fixed rate
and drives the Y-deflection system of the CRT via a digital/analog converter. At the same time, thetime base ramp generator is initiated, with the required speed, by the address counter at the startof the scan. The address input to the store is alternated, if necessary, by a data selector between the
write and read address counters. Read out generally takes place on alternate clock pulses which
are otherwise unused for writing; but with a fast sweep rate, where all clock pulses are requiredwhile writing, read out takes place between successive writing scans.
Two basic types of storage devices are used to store digital information in memory monitors:
shift registers and random access memories. Both of them are equally good for this application.
The other important component of memory monitors is the analog-to-digital converter. Any
book on digital techniques would describe techniques for converting analog information into
digital form. Out of the methods available for A to D conversion, the counter or dual ramp methods
are very effective for slow conversion rates. For higher conversion rates, the successive approxima-tion, tracking, parallel or flash techniques are preferred.
/G36/G2E/G32/G2E/G31 /G53/G65/G6C/G65/G63/G74/G69/G6F/G6E/G20/G6F/G66/G20/G53/G79/G73/G74/G65/G6D/G20/G50/G61/G72/G61/G6D/G65/G74/G65/G72/G73
In a digital memory display system, operating parameters should be carefully selected to maintain
signal accuracy, bandwidth and fidelity. The following are important parameters:
Sampling Rate: Signals with high bandwidth require sampling to be carried out at a high rate to
faithfully obtain all its features. A high sampling rate, however, necessitates a large memory to
store all the data. Therefore, a compromise must be worked out. For ECG, AHA (American HeartAssociation) recommends a sampling rate of 500 samples per second for computer analysis.However, for routine monitoring purposes, such a high sampling rate is not needed. A sampling
rate, which is three times the highest frequency component to be displayed, has been found to be
generally adequate. For monitoring purposes the bandwidth is usually limited to 50 Hz. Therefore,a sampling rate of 150 or above will be satisfactory. Most of the single channel memory displaysystems have a sampling rate of 250 samples/s and multi-channel displays work on 180 samples/s.
Word Length: The ECG signals, before they are coded in the digital form, are amplified in a
preamplifier and brought to a level of 0–1 V. The accuracy of conversion of analog signals depends
on the number of bits used in the conversion. The greater the number of bits per word, the greater
will be the resolution, as a greater number of levels will be available to accurately define the valueof the sampled signal. In practice, however, the ultimate resolution of a given design is limited bythe noise in the various analog and switching circuits and by the linearity and monotonicity of the
converter. Usually, an 8-bit word is chosen, which defines 28 or 256 different words or levels that
can be resolved.
Memory Capacity: The contents of each channel memory are displayed with each sweep across
the CRT screen. Therefore, the stored signal must roll on the screen with a time interval, whichmust be convenient for viewing. For example, a display of 5 s of data (on a 13 cm CRT screen), with

Patient Monitoring Systems 191
he spot moving at a rate of 25 mm/s and at a sampling rate of 200 samples/s, would require a 1000
word memory. Increasing either the sampling rate or the display time interval will increase thememory size proportionately.
/G36/G2E/G32/G2E/G32 /G43/G61/G72/G64/G69/G61/G63/G20/G4D/G6F/G6E/G69/G74/G6F/G72/G20/G55/G73/G69/G6E/G67/G20/G44/G69/G67/G69/G74/G61/G6C/G20/G4D/G65/G6D/G6F/G72/G79
Modern ECG monitors not only include the non-fade display facility but also display heart ratealong with the ECG trace. Figure 6.3 shows a block diagram of a single channel cardioscope withdigital memory. The ECG signal is sensed differentially by the RA (right arm) and LA (left arm)electrodes and is amplified by an isolation ECG amplifier. The patient circuit is isolated by usinga transformer and by modulating the 102 kHz carrier signal with amplified and filtered ECG. Themodulated carrier is demodulated, amplified and applied to an analog-to-digital converter. The
converter samples the waveform at a rate of 250 samples/s, converts each sample to an 8-bit
parallel word, and enters the word into the recirculating memory where it replaces the oldest wordstored.
The recirculating memory, usually employed, consists of eight 1024-stage shift registers operating
at a clock rate of 250 kHz. The output of each shift register is fed back to its input, so that thecontents of the shift registers recirculate continuously. Hence, a waveform acquired at a rate of 250samples/s is available at a rate of 250,000 samples/s. The samples are reconverted into an analogsignal for presentation on the CRT display. The sweep time is so arranged that it matches the timeto read out 1024 samples. The most recent four seconds of the original ECG waveform are traced infour milliseconds. The relatively fast repetition rate of the stored information causes the displayedwaveform to appear bright with no fading.
At any time, the freeze control can be activated to obtain a fixed display on the CRT screen. In
this position, feeding of new signals into the memory is discontinued and all values in the memorythen return to the same positions after each recirculation through the memory. The display in thefrozen position helps to observe a particular event more conveniently. For the display of heart rateon the CRT screen, the original ECG waveform is shaped and supplied to the tachometer circuit to
compute and display the heart rate.
Adjustable heart rate alarm limits are indicated visually and in the audible form. Operation of
these limits is carried out by using two comparators to sense when the heart rate goes beyond set
limits.
Both types of CRTs, viz. electrostatic deflection type or electromagnetic deflection type, are used
to construct non-fade display monitors. However, it is more convenient to use tubes withelectromagnat deflection because of their small size. The deflection coils of CRTs having electro-magnetic deflection have to be specially designed to give a frequency response of about 50 kHz inthe vertical section and a low frequency response in the horizontal section for slow sweepoperations.
As compared to the directed beam display, it will be preferable to use the raster scan deflection
technique because of cost, brightness and power considerations, which make this technique highlydesirable for electromagnetic deflection type tubes. However, the normal scanning rates wouldresult in a display of waveforms with widely spaced dots. The separation of these dots provesunacceptable to physicians and nurses who are used to studying smooth waveforms. This problem

192 Handbook of Biomedical Instrumentation
Fig.6.3 Block diagram of cardioscope using digital memory (redrawn after Grobstein and Gatzke, 1977; by
permission of Hewlett Packard, USA)
Filter andprotectionnetwork
RA LA LLLeadselectorAmp.Amp.BalancedmodulatorBalanceddemodulator
Leadsoff
detecterLeads-offdetector
GuardAlarmAlarms
A-DSample
SamplingpulseTachometerTone burstgate
Beepercircuit
Filter
0.5-1.5 Hz
Amp.Speaker
CRTdisplay
ECG
D-A
Sweep trigger
Timingcircuits
102 KHz
Powersupply
+6 V–6 VCommon-modesignalRecirculatingmemory
Clock

Patient Monitoring Systems 193
can be solved by using a high raster frequency (70 kHz instead of the usual 15 kHz) combined with
beam width shaping on individual lines.
There is at present an intense competition between the long-established cathode ray tube (CRT)
and the newly emerging flat panel technologies for the display of alphanumeric and graphicinformation. Flat panels offer advantages of wide-angle visibility and brightness, resistance to
physical shock and immunity to electromagnetic disturbances. Compared to CRTs, flat panels are
thin, save space and have less weight and bulk. They employ line-addressed, dot matrics imaging,rather than the complex, pixel-by-pixel raster pattern of CRTs. They are in production in formatsranging from 278 ¥ 128 pixels to 1024 ¥ 864 pixels, with screen diagonals up to 19 inches. Flat
panels have found particular favour in medical instruments for high information rate applica-tions, such as patient monitoring systems and ventilators.
The three most common flat panel technologies are: liquid crystal displays, plasma displays
and electro-luminescent displays.
It will be of interest to diagramatically illustrate some of the special features of cardiac monitors,
particularly in respect of the frequency response, input circuitry and other special features, whichdistinguish them from electrocardiographs.
Frequency Response of Cardioscopes: Some monitors have two selectable frequency response
modes, namely Monitor and Diagnostic. In the ‘Monitor’ mode or ‘Filter-in’ mode, both the lowand high frequency components of an electrocardiogram are attenuated. It is used to reducebaseline wander and high frequency noise. The Monitor mode bandwidth is generally 0.4 to 50 Hz(3 dB points). In the ‘Diagnostic’ mode, the instrument offers expanded bandwidth capability of0.05 to 100 Hz. Some instruments include a 50 Hz notch filter to improve the common moderejection ratio and this factor should be kept in mind while checking the frequency response of theinstrument.
lnput Circuit: Figure 6.4 shows a complete input circuit used in present day cardiac monitors.
There are three prominent circuit blocks: (i) low-pass filter circuit, to suppress RF interference,
Patient
electrode 3.3 mH 10 K
Cable
capacitance
(100 pF)C1
C2(Shield)
100 pF27 K27 K 10 K 10 K
10 K
330 pF
High-voltage
protectionOver-voltage
protectionElectrosurgery
filterPatient cable
(shielded)
Q1
C4C3 330 pF
330 pF
+6 VGuard
Fig.6.4 Details of the input network in a cardiac monitor. The features include low
pass filter to suppress RF interference and voltage clamps to prevent defibril-lator pulses from damaging the sensitive input amplifier (redrawn afterGrobstein and Gatzke, 1977)

194 Handbook of Biomedical Instrumentation
(ii) high voltage protection circuit, similar to electrocardiographs, to provide voltage clamp in the
presence of defibrillator pulses, and (iii) over voltage protection circuit.
Electrosurgery Interference: RF interference occurs due to any one of two possible modes. The first
and usually the most severe is due to conduction, i.e. the RF energy is actually carried via thepatient into the monitor. The second is radiation, by which the RF energy is transmitted throughthe air and is induced into the circuits of the monitoring instrument, its leads and cables.
Electrosurgery machines generate RF signals within a range of 0.4 to 5 MHz with peak-to-peak
amplitudes of 100 to 1000 V, pulse modulated at rates from 1.5 to 25 kHz for coagulating or 120 Hz
for cutting. Cardiac monitors are often used in operation theatres, where the RF is applied througha pointed scalpel at the point of incision and the return path for the current is through a wide areaelectrode on the opposite side of the patient’s body. The ECG signal, on the other hand, is of theorder of 1 mV with frequency components below 100 Hz. The ECG input amplifier, the ECGelectrode-skin interface and the scalpel-tissue interface are the main sites where rectificationoccurs. Moreover, the common mode RF signal gets converted into a normal mode signal by animbalance in the capacitance between input to ground at the differential amplifier input.
In order to reduce interference caused by electrosurgery and radiofrequency emissions, it is
very essential to use filters at the input of leads of the monitoring instruments. Grobstein andGatzke (1977) give the design criteria for constructing such a filter. To reduce the interference toacceptable levels, the ratio of amplifier sensitivities to the ECG signal (1 mV at less than 100 Hz)and to the electrosurgery machine (100 V at 1 MHz or so), should be at least 10
5. If we use a filter
having a cut off frequency above 10 kHz, a three pole filter is needed to achieve this level of
reduction. Five poles are used in the monitors, three provided by the RC filters within the
instrument and two provided by the electrosurgery filters and the conductor-to-shield capacitanceof the cable (usually 1000 pF). The equivalent network for the ECG machine input cable is asshown in Fig. 6.5. The imbalance in capacitance from each input of the differential amplifier to theground also injects a differential RF signal that can be rectified and can produce interference. Theimbalance necessary to provide a 100 mV signal is surprisingly very small, of the order of 0.001 pF
at 1 MHz. To minimize this problem, a guard is used, which besides shielding the input circuitsfrom electromagnetic radiation also helps in equalizing the capacitance from input to ground ofeach amplifier input.
3.3 mH3.3 mH
3.3 mH
3.3 mH10 K10 K
10 K
10 KCapacitance of
trunk, cable
1000 pF each
Capacitance of
lead set
400 pF eachElectrosurgery
machinePatient
Fig.6.5 Equivalent network for the ECG machine input cable (after Grobstein and
Gatzke, 1977)

Patient Monitoring Systems 195
The following precautions should be taken to achieve good ECG display in the presence of
electrosurgery interference (Benders, 1976):
• The electrosurgery return plate should be directly under the surgical site, as far as pos-
sible.
• The ECG electrodes should be placed at the maximum possible distance away from the
surgical site.
• The electrodes should be equidistant from the surgical site.• All monitoring electrodes must be placed either on the frontal surface or on the posterior
surface.
• Only shielded ECG patient cables and electrode leads must be used.
Leads Off Detector: The “leads off’ detector circuit usually works on the principle that loss of body
contact of either the RA or LA electrode causes a rather high impedance change at the electrode/body contact surface, consequently causing a loss of bias at the appropriate amplifier input. Thissudden change makes the amplifier to saturate, producing maximum amplitude waveform. Thiswaveform is rectified and applied to a comparator that switches on an alarm circuit (leads off)when the waveform exceeds a certain amplitude.
Quick Recovery Circuit: In order to avoid problems due to dc drift and supply voltage variations,
ECG amplifiers are usually constructed with at least one stage of ac coupling. The circuit employsa high-value series capacitor in between two amplifying stages, a high value is used to maintaina good low frequency response. The presence of this capacitor, however, creates the problem of avery long recovery time after an over voltage appears at the input of the amplifier due to conditionslike leads off, excessive patient movement, defibrillator operation, etc. This problem is overcome byusing a quick recovery circuit. This circuitbasically provides a fast discharge of thecoupling capacitor once it has been ex-
cessively charged. Figure 6.6 shows the
working principle of this circuit, in whichthe over range voltage for A
2 is determined
by the voltage dividers ( R1,R2,R3,R4, res-
pectively)—the voltage may be plus or
minus. In either case, the plus or minusover range input to the differential ampli-fier A
3 causes the output to go positive.
This turns ‘on’ the FET Q, which rapidly
discharges C, causing the output baseline
to return to zero level.
Cardiac monitors are also available
which display more than one channel ofinformation such as ECG and delayed ECG
with digital heart rate display and alarm
facility. The display is for 4 second. The twochannels can also be cascaded to have aA1
A2
A3ECG
input
Output C
Q
R9(150 K)
R8
R1
R2
R3
R4R6D2D1
R5R7+VECG amplifier
Fig.6.6 Principle of operation of quick recovery
circuit

196 Handbook of Biomedical Instrumentation
8 second display (Fig. 6.7). They use time-multiplexing of complete sweeps to avoid the excessive
band-width introduced by electronic switching or the chopper technique. Each channel, in theseinstruments, can be independently controlled for updating or freezing the information.
Fig.6.7 Two channel digital storage ECG waveform display system, with digital
display of heart rate (Courtesty: M/s Schiller Health Care India Pvt. Ltd.)
/G20/G36/G2E/G33 /G42/G45/G44/G53/G49/G44/G45/G20/G50/G41/G54/G49/G45/G4E/G54/G20/G4D/G4F/G4E/G49/G54/G4F/G52/G49/G4E/G47/G20/G53/G59/G53/G54/G45/G4D/G53
Bedside monitors are available in a variety of configurations from different manufacturers. They
are designed to monitor different parameters but the common feature amongst all is the facility tocontinuously monitor and provide non-fade display of ECG waveform and heart rate. Someinstruments also include pulse, pressure, temperature and respiration rate monitoring facilities.
The advent of microcomputers has marked the beginning of a fundamentally new direction in
patient monitoring systems. Such systems are intended to replace the traditional monitoring
devices with a single general purpose unit capable of recognizing the nature of the signal source
and processing them appropriately. The hardware responsible for physiological signal analysis,information display and user interaction is actually a set of firmware modules implemented interms of a microcomputer program. The firmware gives the system its functional personality andthe usual switches, knobs, dials and meters can be replaced by a touch-sensitive character display.
A typical example of a microprocessor-based bedside patient monitoring instrument is shown
in Fig. 6.8. The system is designed to display an electrocardiogram, heart rate with high and low

Patient Monitoring Systems 197
alarms, pulse rate, dynamic pressure or other waveforms received from external preamplifiers.
It also gives immediate and historical data on the patient for trend information on heart rate,temperature, and systolic and diastolic blood pressures for periods up to eight hours. The system
basically consists of three circuit blocks: Preamplifier section ,Logic boards and Display part .
The preamplifiers incorporate patient isolation circuits based on optical couplers. The ECG wave-
form has facilities for lead-off detection, ‘pacer’ detection and quick recovery circuit for overload
signals. Various amplified signals are carried to a multiplexer and then to an analog-to-digitalconverter, included in the logic board. The central processing unit along with memory gives X andY output for the CRT display. The character generator output is mixed with the Y output for
numeric display on the CRT. The alarm settings, selection switches for different parameters and
the defibrillator synchronization system communicate with the CPU. The alarm signals are alsoinitiated under its control. The memory comprises 5 K bytes ROM and 3.25 K bytes RAM with 256samples of ECG. Eight seconds delayed ECG is available for recording purposes.
Several important trends in the design and function of bedside monitors have emerged in the
past few years. More bedside units are now software based, a feature that facilitates changes and
updates in function by the simple replacement of computer memory chips. Wider use of on-boardmicroprocessors also permits bedside monitors to perform increasingly sophisticated signal-processing tasks. Advances in monitoring the haemodynamic parameters are particularly note-
worthy. New smart algorithms help to carry out automatic calculations of indices of cardiovascular
functions and artefact removal tasks. The trend in ECG monitoring is towards display and analysisof data from multiple leads. Several manufacturers now include arrhythmia monitoring, including
the monitoring of the ST segment of the ECG, as a standard feature in bedside monitors.Central
processing
unit
Loud speakerAlarm outputAlarm switchesRate displayDefib. sync.ECG
Temp
PulseAmp.
Amp.
Amp. Amp.
EHT
supply
BlankingY
amp. X
amp.
CRT display
tubeY output
X outputIsolation
Diagnostic ECG
Rwave
detectorRate
QRS
Temp
PulseRate
ECG
Temp
Pulse
BPMultiplexerA-D
converter
Memory
Character
generatorX-position
generator
Fig.6.8 Block diagram of the bedside patient monitor (Courtesy: Albury Instruments, U.K.)

198 Handbook of Biomedical Instrumentation
While increasingly sophisticated monitoring capabilities have been added to bedside monitors,
many monitors today are much easier to use than their predecessors were. Improvements insoftware and features such as touch screen make today’s bedside monitors a user-friendly
equipment.
Patient monitors are also known as vital sign monitors as they are primarily designed to
measure and display vital physiological parameters. Figure 6.9 shows the system block diagramof a patient monitoring system. It basically consists of the modular parts for measurement of thefollowing:
• ECG and respiration measuring electronics
• Blood pressure (non-invasive) measuring electronics, pump and tubing• Blood pressure (invasive)• Temperature measuring electronics• Pulse probe and SpO
2 (pulse oximetry)
• Microprocessor board including analog signal multiplexer, A–D converter and real time
clock
• Video control board to convert the CPU commands into video signal• Video display module• Transformer and power supply board to generate necessary voltages• Mother board including signal buses and analog input signal buffers• Keyboard
The whole system works under the control of the processor 80C32, which works at 16 MHz. The
processor boards for each of the above parameters are described in the subsequent sections. Figure6.10 shows a typical vital signs monitor.
/G20/G36/G2E/G34 /G43/G45/G4E/G54/G52/G41/G4C/G20/G4D/G4F/G4E/G49/G54/G4F/G52/G53
With central monitoring, the measured values are displayed and recorded at a central station.Usually, the signal conditioners are mounted at the bedside and the display and alarms, etc. arelocated in a central station.
The central station monitoring equipment may incorporate a multi-microprocessor architecture
to display a flexible mixture of smooth waveforms, alphanumerics and graphics on a singlecathode ray tube. This presents all the information at a glance and thus assists the hospital staff inseveral ways. First, it generates audible and visual alarms if preset vital sign limits are exceeded.It is important that the central station announces these emergencies without generating too manyfalse alarms, arising due to patient movements, etc. Secondly, it displays the patient’s vital signdata. By watching this data, the attending staff can detect problems before they reach the alarm
stage. Trend plots of vital signs aid in guiding the patient’s therapy. Thirdly, it provides a recording
of the ECG and sometimes of other parameters, especially of the few seconds just before an alarm,which shows what kind of irregularity led to the alarm.
Central stations are primarily designed for coronary care patients to display ECG waveforms
and heart-rate information for eight patients. The display shows (Fig. 6.11) four seconds of real

Patient Monitoring Systems 199 Fig.6.9 Block diagram of bed side monitor
Trans-former AC inCuff
Pump
Non-invasive BPmeasurementNIBP
processorboardPressuretransducerMain processor board (I/O, MUX, ADC, CPU, ROM, RAM)
Power supplyRESPECG
InoperateInoperateECG electrodes(from patient)LeadsetVideocontrolboardVideodisplaymoduleCRTvideoscreenKey-board
SpO
measuringboard2
SpO
processorboard2Pulse oximetrytransducer
Multiplexer
Temperature
P1
P2Pressurechannels(invasive)

200 Handbook of Biomedical Instrumentation
time ECG waveform and alarm information. A long trend (for either 9 or 24 h) or short trends (90
min) may be selected for display for observation and/or documentation. The information for the
central monitor is collected from the bedside. Each bedside cable contains as many as fifteenanalog signals representing physiological parameters, which may include several bloodpressures, ECG, heart rate, respiration, end-tidal CO
2 and temperature. Status information such as
alarm signals is also carried by the same cable. The 80 or so incoming physiological values arethen sampled and digitized at appropriate rates by a 10-bit analog-to-digital converter. ECGwaveforms are sampled every two milliseconds to maintain the 0–100 Hz bandwidth. Slowlyvarying variables such as temperature are sampled every four seconds.
The display part has two subsections-raster type display for waveforms and a conventional
300¥ 260 picture-element bit map for alphanumerics and graphics. To make the waveforms look
smooth, a 1200 line vertical raster is used. The display subsection also manages a 16 K-wordmemory, which is used as temporary storage for waveform and hard copy data. ECG waveformdata for each patient is continually stored temporarily to provide a delayed (typically 8 s)waveform output for recording. The delay is used to capture a snapshot of any ECG abnormality
Fig.6.10 Vital S igns Monitor (Courtesy: Protocol Systems Inc., USA)

Patient Monitoring Systems 201
leading to an alarm. If the recorder happens to be busy, the display processor stores it until a
recorder becomes available.
The trend memory can hold patient data for 24 h. It contains 6 K words of CMOS RAM, which
requires very little power in the standby mode and, therefore, can be connected to a battery back-uppower supply. The patient data is thus held for at least 24 h after a power failure.
There has been an increasing realization with regard to assuring minimum breakdown of
patient monitoring equipment. Microcomputer based self-test provision in the system helps theequipment to automatically test, analyze and diagnose itself for failures. When some parts fail,others take over their functions to minimize the impact of the failure on the system operation.
Alarm notifications are typically audible and/or visual. Audible alarms can be distinguished
by varying the pitch, volume, duration and sequencing of the tones. Visual alarms can be indicatedby varying the colour of the display on the monitor screen. Alarm conditions can also be capturedin hard copy documentation via automatic generation of a recording at the time of the event (Slye,1995). Traditional limit alarms need to be set up by the nursing staff, usually process each input
independently and have unacceptably frequent false alarms. Dodd (1993) explains the use of
neural network techniques to minimize the false alarm indications frequently encountered in apatient monitoring environment.
The increasing use of workstations in central monitoring installations have made it possible to
monitor a large number of patients on a single monitor. Patient and equipment status are indicatedby simple, colour graphic symbols. The ease of networking the monitors on the existing computernetwork wiring has resulted in flexible systems. The patient’s bedside monitors can be viewed
and operated from the central station. Due to the large storage capacity of the workstations, all thewaveforms and numerics can be automatically captured, stored and retained to allow access to876544 3
6 52 1
8 7321
8292
52
Check paper/door-recorder # 1 17 Jul 79 145382(50 – 150)(50 – 150)(30 – 170)(HR-limits) (HR-limits)
49HR HR
(40 – 150) (35 – 150)*** HR < 40*** HR < 85
*** Leads of INOP
ECG invalidControlBED BED
Fig.6.11 Typical displays on the patient information centre. The display depicts
information on the heart rate, alarm limits, display of ECG waveformsfrom four to eight patients (Courtesy: Hewlett Packard, USA)

202 Handbook of Biomedical Instrumentation
data for up to three days after the patient is discharged. Figure 6.12 shows the view of a modern
central patient monitoring station.
Fig.6.12 PC based central station (Courtesy: M/s Protocol Systems Inc. USA)
/G20/G36/G2E/G35 /G4D/G45/G41/G53/G55/G52/G45/G4D/G45/G4E/G54/G20/G4F/G46/G20/G48/G45/G41/G52/G54/G20/G52/G41/G54/G45
Heart rate is derived by the amplification of the ECG signal and by measuring either the average or
instantaneous time intervals between two successive R peaks. Techniques used to calculate heart
rate include:
•Average calculation This is the oldest and most popular technique. An average rate (beats/
min) is calculated by counting the number of pulses in a given time. The average method ofcalculation does not show changes in the time between beats and thus does not representthe true picture of the heart’s response to exercise, stress and environment.
•Beat-to-beat calculation This is done by measuring the time ( T), in seconds, between two
consecutive pulses, and converting this time into beats/min., using the formula beats/min. = 60/ T. This technique accurately represents the true picture of the heart rate.
•Combination of beat-to-beat calculation with averaging This is based on a four or six beats
average. The advantage of this technique over the averaging techniques is its similaritywith the beat-to-beat monitoring system.

Patient Monitoring Systems 203
The normal heart rate measuring range is 0–250 beats/min. Limb or chest ECG electrodes are
used as sensors.
/G36/G2E/G35/G2E/G31 /G41/G76/G65/G72/G61/G67/G65/G20/G48/G65/G61/G72/G74/G20/G52/G61/G74/G65/G20/G4D/G65/G74/G65/G72/G73
The heart rate meters, which are a part of the patient monitoring systems, are usually of theaverage reading type. They work on the basis of converting each R wave of the ECG into a pulse of
fixed amplitude and duration and then determining the average current from these pulses. Theyincorporate specially designed frequency to a voltage converter circuit to display the average heartrate in terms of beats per minute.
/G36/G2E/G35/G2E/G32 /G49/G6E/G73/G74/G61/G6E/G74/G61/G6E/G65/G6F/G75/G73/G20/G48/G65/G61/G72/G74/G20/G52/G61/G74/G65/G20/G4D/G65/G74/G65/G72/G73
Instantaneous heart rate facilitates detection of arrhythmias and permits the timely observation ofincipient cardiac emergencies. Calculation of heart rate from a patient’s ECG is based upon thereliable detection of the QRS complex (Thakor, et al 1983). Most of the instruments are, however,
quite sensitive to the muscle noise (artefact) generated by patient movement. This noise oftencauses a false high rate that may exceed the high rate alarm. A method to reduce false alarm is byusing a QRS matched filter, as suggested by Hanna (1980). This filter is a fifteen sample finite-
impulse-response-filter whose impulse response shape approximates the shape of a normal QRS
complex. The filter, therefore, would have maximum absolute output when similarly shaped
waveforms are input. The output from other parts of the ECG waveform, like a T wave, will
produce reduced output.
Figure 6.13 is a block diagram of the scheme. The ECG is sampled every 2 ms. Fast transition
and high amplitude components are attenuated by a slew rate limiter which reduces the amplitudeof pacemaker artefacts and the probability of counting these artefacts as beats. Two adjacent 2 mssamples are averaged and the result is a train of 4 ms samples. In order to remove unnecessaryhigh frequency components of the signal, a 30 Hz, infinite-impulse-response, Butterworth filter isemployed. This produces 8 ms samples in the process. Any dc offset with the signal is removed bya 1.25 Hz high-pass filter. The clamped and filtered ECG waveform is finally passed through a
Slew rate
limit
averageECG
2 ms
samples30 Hz
low-pass
filter1.25 Hz
high-pass
filterQRS
matched
filterFibrillation
detector
Alarm
comp.Heart rate
detectorBeat
detectorLow limit
High limit
Fig.6.13 Block diagram of the cardiotachometer based on matched QRS filter (redrawn
after Hanna, 1980; by permission of Hewlett Packard, USA)

204 Handbook of Biomedical Instrumentation
QRS matched filter. The beat detector recognizes QRS complexes in the processed ECG waveform
value that has occurred since the last heart beat. If this value exceeds a threshold value, a heartbeat is counted. The beat interval averaged over several beats is used to calculate the heart rate for
display, alarm limit comparison, trending and recorder annotation. The threshold in this
arrangement gets automatically adjusted depending upon the value of the QRS wave amplitude
and the interval between the QRS complexes. Following each beat, an inhibitory period of 200 ms
is introduced during which no heart beat is detected. This reduces the possibility of the T wavefrom getting counted. The inhibitory period is also kept varied as an inverse function of the highrate limit, with lower high rate limits giving longer inhibitory periods.
Based on the power spectra estimation of the QRS complex, Thakor et al (1984 b) have suggested
that a bandpass filter with a centre frequency of 17 Hz and a Qof five, yields the best signal to noise
ratio. Such a simple filter should be useful in the design of heart rate meters, arrhythmia monitorsand implantable pacemakers.
The subject of reliable detection of R-wave continues to be of great interest for the researchers.
Besides the hardware approach, a number of software based approaches have been reported inliterature. Since the ultimate aim of detecting the R-wave is to automate the interpretation of ECGand detect arrhythmias, they are best covered in the succeeding chapter.
/G20/G36/G2E/G36 /G4D/G45/G41/G53/G55/G52/G45/G4D/G45/G4E/G54/G20/G4F/G46/G20/G50/G55/G4C/G53/G45/G20/G52/G41/G54/G45
Each time the heart muscle contracts, blood is ejected from the ventricles and a pulse of pressure istransmitted through the circulatory system. This pressure pulse when travelling through thevessels, causes vessel-wall displacement, which is measurable at various points of the peripheralcirculatory system. The pulse can be felt by placing the finger tip over the radial artery in the wristor some other location where an artery seems just below the skin. The timing and wave shape ofthe pressure pulse are diagnostically important as they provide valuable information.
The pulse pressure and waveform are indicators for blood pressure and flow. Instruments
used to detect the arterial pulse and pulse pressure waveforms in the extremities are calledplethysmographs. Most plethysmograph techniques respond to a change in the volume of bloodas a measure of blood pressure.
The pulse gives a measure of pulse wave velocity and can be recorded and compared with the
ECG signal (Fig. 6.14). The pulse wave travels at 5 to 15 m/s, depending on the size and rigidity of
the arterial walls. The larger and more rigid the artery walls, the greater the velocity. The velocity
is 10–15 times faster than blood flow, and is relatively independent of it.
The methods used for the detection of volume (pulse) changes due to blood flow are:
• Electrical impedance changes
• Strain gauge or microphone (mechanical)• Optical changes (changes in density)
An electric impedance method measures the impedance change between two electrodes caused
by the change in blood volume between them. The change in impedance (0.1 ohm) may be small ascompared to the total impedance (several hundred ohms). The impedance is measured by applyingan alternating current between electrodes attached to the body. An alternating signal (10–100kHz) is used (rather than dc) in order to prevent polarization of the electrodes.

Patient Monitoring Systems 205
The mechanical method involves the use of a strain gauge connected to a rubber-band placed
around a limb or finger. Expansion in the band due to change in blood volume causes a change inresistance of the strain gauge. In another technique, a sensitive crystal microphone is placed on
the skin’s surface to pick up the pulsation.
The most commonly used method to measure pulsatile blood volume changes is by the
photoelectric method . Two methods are common: Reflectance method and transmittance method.
In the transmittance method (Fig. 6.15(a)) a light-emitting diode (LED) and photoresistor are
mounted in an enclosure that fits over the tip of the patient’s finger. Light is transmitted through
the finger tip of the subject’s finger and the resistance of the photoresistor is determined by theamount of light reaching it. With each contraction of the heart, blood is forced to the extremities1
2
3Chart no Judson bigelow inc USA
Fig.6.14 Pulse pick up, showing time relationship with electrocardiogram (i) ECG
(ii) crystal microphone pulse pick-up (iii) photoelectric pulse pick-up

206 Handbook of Biomedical Instrumentation
and the amount of blood in the finger increases. It alters the optical density with the result that the
light transmission through the finger reduces and the resistance of the photoresistor increasesaccordingly. The photoresistor is connected as part of a voltage divider circuit and produces a
voltage that varies with the amount of blood in the finger. This voltage that closely follows the
pressure pulse and its waveshape can be displayed on an oscilloscope or recorded on a strip-chartrecorder.
The arrangement used in the reflectance method of photoelectric plethysmography is shown in
Fig. 6.15(b). The photoresistor, in this case, is placed adjacent to the exciter lamp. Part of the lightrays emitted by the LED is reflected and scattered from the skin and the tissues and falls on thephotoresistor. The quantity of light reflected is determined by the blood saturation of the capillariesand, therefore, the voltage drop across the photoresistor, connected as a voltage divider, will varyin proportion to the volume changes of the blood vessels.
Photoresistor
PhotoresistorLamp
Lamp
(a) (b)
Fig.6.15 Arrangement of photoresistor and lamp in a finger probe for pulse pick-up:
(a) transmission method, (b) reflectance method
The LED phototransistor-photoplethysmograph transducer (Lee et al, 1975) consists of a Ga-As
infrared emitting diode and a phototransistor in a compact package measuring 6.25 ¥ 4.5 ¥ 4.75
mm. The peak spectral emission of the LED is at 0.94 mm with a 0.707 peak bandwidth of 0.04 mm.
The phototransistor is sensitive to radiation between 0.4 and 1.1 mm (Fig. 6.16).
For pulse rate measurement, a photoelectric transducer suitable for use on the finger or ear lobe
is used. The signal from the photocell is amplified and filtered (0.5 to 5 Hz passband) and the timeinterval between two successive pulses is measured. The measuring range is 0–250 bpm. Careful
placement and application of the device is essential in order to prevent movement artefacts due to
mechanical distortion of the skin.
Figure 6.17 shows the block diagram for processing the plethysmographic signal detected from
a photoelectric transducer. The circuit consists of two parts, a LED oscillator and driver, whichproduce 300 Hz, 50 ms infrared light pulses to the finger probe attached to the patient, and a
phototransistor that picks up the attenuated light. The electrical signal obtained from thephototransistor is amplified and its peak value is sampled and filtered. An automatic gain control

Patient Monitoring Systems 207
circuit adjusts the amplifier gain to yield a constant average pulse height at the output. The ac
component with a frequency in the heart rate range (0.8–5 Hz), is further amplified to outputthe plethysmographic pulse rate form. This signal is transmitted across the isolation barrier,demodulated, low-pass filtered and transmitted to the analog multiplexer resident on the CPUboard.
APiezo-electric crystal can also be used to detect the pulse wave at certain places of the
peripheral system where considerable displacement of the tissue layer above the artery is involved.The arrangement consists of a piezo-electric crystal clamped in a hermetically sealed capsule
subject to displacement stresses. The displacement can be transmitted to the crystal through a soft
rubber diaphragm. The crystal can be connected to an ECG recorder for recording the pressurepulse waveform.0.3 0.5 0.8 1.1 1.220406080100UV Visible radiation
Gas Tungsten lampIR
Silicon photo
transistor
Wavelength m mRelative response
Fig.6.16 Relative spectral response for silicon phototransistor and the radiant
spectral distribution of a tungsten lamp and a gallium-aresenide lamp
(after Lee et al. 1975; reproducd by permission of IEEE Trans. Biomed. Eng.)
LED
Driver300 Hz
oscillator
BIAS
adjust
Automatic
gain controlLED
Sample &
holdPlethysmographtransducer
Photo
transistordc
amp.ac
amp.
IsolationPulse
output
Fig.6.17 Block diagram for processing plythysmographic signal

208 Handbook of Biomedical Instrumentation
There is another variation of the finger plethysmograph in which an air-coupled piezo-electric
transducer is employed. As the volume of blood in the finger varies during the cardiac cycle, slightchanges occur in the size of the finger. These changes can be transmitted as pressure variations in
the air column inside the plastic tubing. A piezo-electric transducer at the end of the tube converts
the pressure changes to a corresponding electrical signal. This signal can then be amplified anddisplayed. Similarly, a semiconductor strain gauge can be used to detect the displacement of thevessel wall due to a pulse wave.
Monitoring the peripheral pulse is more useful and dependable than monitoring the heart rate
derived from ECG in the case of a heart block because it can immediately indicate the cessation ofblood circulation in the limb terminals. Moreover, a photoelectric pick-up transducer is mucheasier to apply than the three ECG electrodes. The amplitude of the plethysmographic signalobtained is also quite large as compared to the ECG signal and therefore, gives better signal-to-noise ratio. However, the technique is severely subject to motion artefacts.
/G20/G36/G2E/G37 /G42/G4C/G4F/G4F/G44/G20/G50/G52/G45/G53/G53/G55/G52/G45/G20/G4D/G45/G41/G53/G55/G52/G45/G4D/G45/G4E/G54
Blood pressure is the most often measured and the most intensively studied parameter in medicaland physiological practice. The determination of only its maximum and minimum levels during
each cardiac cycle supplemented by information about other physiological parameters is an
invaluable diagnostic aid to assess the vascular condition and certain other aspects of cardiacperformance. Pressure measurements are a vital indication in the successful treatment andmanagement of critically ill patients in an intensive cardiac care or of patients undergoing cardiaccatheterization. The tremendous research and development for an automatic blood pressuremonitor has resulted in several methods but only very few have been commercialized due tocertain practical difficulties.
Blood is pumped by the left side of the heart into the aorta, which supplies it to the arterial
circuit. Due to the load resistance of the arterioles and precapillaries, it loses most of its pressureand returns to the heart at a low pressure via highly distensible veins. The right side of the heartpumps it to the pulmonary circuit, which operates at a lower pressure. The heart supplies blood toboth circuits as simultaneous intermittent flow pulses of variable rate and volume. The maximumpressure reached during cardiac ejection is called systolic pressure and the minimum pressureoccurring at the end of a ventricular relaxation is termed as diastolic pressure. The mean arterial
pressure over one cardiac cycle is approximated by adding one-third of the pulse pressure
(difference between systolic and diastolic values) to the diastolic pressure. All blood pressuremeasurements are made with reference to the atmospheric pressure.
Typical haemodynamic pressure values are shown in Fig. 6.18. The nominal values in the basic
circulatory system are as follows:
Arterial system 30 –300 mmHg
Venous system 5–15 mmHg
Pulmonary system 6–25 mmHg
The most frequently monitored pressures, which have clinical usefulness in medium and long-
term patient monitoring, are the arterial pressure and the venous pressure. There are two basic
methods for measuring blood pressure—direct and indirect.

Patient Monitoring Systems 209
LungSuperior
vena cava
AortaPulmonary
arteryHead and arms
Left atrium
(2-11 mmHg)
Lung
Pulmonary
vein
Mitral valve
Left ventricle 120/2-11
Endocardium
Myocardium
Epicardium
Aortic valve
ApexInferior
vena cavaRight
ventricle(14/0.32/6)
Lower
extremetiesTricuspid
valvePulmonary
valveRight
atrium
(0.6
mmHg)
Fig.6.18 Typical haemodynamic pressure values present in the basic circulatory
system (Courtesy: Hewlett Packard, USA)
The indirect methods consist of simple equipment and cause very little discomfort to the subject
but they are intermittent and less informative. They are based on the adjustment of a knownexternal pressure equal to the vascular pressure so that the vessel just collapses. On the otherhand, the direct methods provide continuous and much more reliable information about theabsolute vascular pressure from probes or transducers inserted directly into the blood stream. Butthe additional information is obtained at the cost of increased disturbance to the patient andcomplexity of the equipment.
Blood pressure readings vary with subjects and, among other variables, with the location of the
transducer. If manometric blood pressure readings are not taken at heart level, they should becompensated to correspond to the readings at heart level. For example, if a mercury-manometerreading is taken at h mm below heart level, the reading is high due to the weight of a column ofblood h mm high (this weight is r gh). The compensation factor is simply the ratio of densities:
For Mercury, r = 13.6 g/cm
3
For blood, r = 1.055 g/cm3
Ratio = 13 6
1 055.
. = 12.9
The equivalent reading at heart level is thus:
mmHg reading = (mm above or below heart level)
12.9
If the manometer is above heart level, add the correction; if below, subtract the correction.

210 Handbook of Biomedical Instrumentation
/G36/G2E/G37/G2E/G31 /G44/G69/G72/G65/G63/G74/G20/G4D/G65/G74/G68/G6F/G64/G73/G20/G6F/G66/G20/G4D/G6F/G6E/G69/G74/G6F/G72/G69/G6E/G67/G20/G42/G6C/G6F/G6F/G64/G20/G50/G72/G65/G73/G73/G75/G72/G65
The direct method of pressure measurement is used when the highest degree of absolute accuracy,
dynamic response and continuous monitoring is required. The method is also used to measure thepressure in deep regions inaccessible by indirect means. For direct measurement, a catheter or a
needle type probe is inserted through a vein or artery to the area of interest. Two types of probes
can be used. One type is the catheter tip probe in which the sensor is mounted on the tip of theprobe and the pressures exerted on it are converted to the proportional electrical signals. The otheris the fluid-filled catheter type, which transmits the pressure exerted on its fluid-filled column toan external transducer. This transducer converts the exerted pressure to electrical signals. Theelectrical signals can then be amplified and displayed or recorded. Catheter tip probes provide themaximum dynamic response and avoid acceleration artefacts whereas the fluid-filled cathetertype systems require careful adjustment of the catheter dimensions to obtain an optimum dynamicresponse.
Measurement of blood pressure by the direct method, though an invasive technique, gives not
only the systolic, diastolic and mean pressures, but also a visualization of the pulse contour andsuch information as stroke volume, duration of systole, ejection time and other variables. Once anarterial catheter is in place, it is also convenient for drawing blood samples to determine thecardiac output (by dye dilution curve method), blood gases and other chemistries. Problems of
catheter insertion have largely been eliminated and complications have been minimized. This has
been due to the development of a simple percutaneous cannulation technique; a continuous flushsystem that causes minimal signal distortion and simple, stable electronics which the paramedicalstaff can easily operate.
A typical set-up of a fluid-filled system for measuring blood pressure shown in Fig. 6.19. Before
inserting the catheters into the blood vessel it is important that the fluid-filled system should bethoroughly flushed. In practice a steady flow of sterile saline is passed through the catheter toprevent blood clotting in it. As air bubbles dampen the frequency response of the system, it shouldbe ensured that the system is free from them.
Figure 6.20 shows a simplified circuit diagram commonly used for processing the electrical
signals received from the pressure transducer for the measurement of arterial pressure. Thetransducer is excited with a 5 V dc excitation. The electrical signals corresponding to the arterialpressure are amplified in an operational amplifier or a carrier amplifier. The modern preamplifierfor processing pressure signals are of the isolated type and therefore comprise of floating andgrounded circuits similar to ECG amplifiers. The excitation for the transducer comes from anamplitude controlled bridge oscillator through an isolating transformer, which provides an
interconnection between the floating and grounded circuits. An additional secondary winding in
the transformer is used to obtain isolated power supply for the floating circuits. The input stage isa differential circuit, which amplifies pressure change, which is sensed in the patient connectedcircuit. The gain of the amplifier can be adjusted depending upon the sensitivity of the transducer.After RF filtering, the signal is transformer-coupled to a synchronized demodulator for removingthe carrier frequency from the pressure signal.
For the measurement of systolic pressure, a conventional peak reading type voltmeter is used.
When a positive going pressure pulse appears at A, diode D
3 conducts and charges C3 to the peak

Patient Monitoring Systems 211
value of the input signal, which corresponds to the systolic value. Time constant R3C3 is chosen in
such a way that it gives a steady output to the indicating meter.
The value of diastolic pressure is derived in an indirect way. A clamping circuit consisting of C1
and D1 is used to develop a voltage equal to the peak-to-peak value of the pulse pressure. This
voltage appears across R1. Diode D2would then conduct and charge capacitor C2to the peak value
of the pulse signal. The diastolic pressure is indicated by a second meter M2which shows thePressure
monitor3–4 ff
Intraflo
flush
valveSample
stop-cockPressure
transducerSolution
under
pressure
Fig.6.19 Typical set up of a pressure measuring system by direct method
+V
–V
Strain gauge
pressure
transducerAmplifierAmp.
A
P2 D3
C3C1D1
D2P1C1
R3R1 R2
M1M2Diastolic
indicator
Systolic
indicator
Fig.6.20 Circuit diagram for measurement of systolic and diastolic blood pressure

212 Handbook of Biomedical Instrumentation
difference between the peak systolic minus the peak-to-peak pulse pressure signal. The mean
arterial pressure can also be read by using a smoothing circuit when required.
Central venous pressure (CVP) measurements made with needle cannulation techniques prove
extremely useful in the management of acute circulatory failure and in the maintenance of bloodvolume in difficult fluid balance problems. Simple water manometers are still the most common
measuring device in use, although highly sensitive pressure transducers are preferred when
accurate measurements are required. However, the transducers cannot be conveniently mountedat the catheter tip and small positional changes cause large errors in venous pressure. Infusingintravenous fluids while measuring pressure through the same catheter is another problemencountered in these measurements. Central venous pressure is usually measured from a catheterlocated in the superior vena cava. The CVP reflects the pressure of the right atrium and is sometimesreferred to as right atrial pressure. The catheter can even be located in the right atrium. Major peri-pheral veins used as entry sites for CVP monitoring are the brachial, subclavian and jugular veins.
Catheters used for CVP monitoring are usually 25 to 30 cm long. Long catheters, is they remain
in place over extended periods of time are susceptible to the formation of fibrin sheaths along theirouter surfaces. Besides this, air can be aspirated into a catheter that is situated in an area of lowpressure (as compared to the atmospheric pressure), resulting in thrombo-embolic complications.A continuous infusion of heparin solution will reduce this tendency. Also, it should be ensuredthat there is no possibility of air intake. Development of the Swan-Ganz catheter- a balloon tipped,
flexible catheter that can be flow-directed from a peripheral vein into the pulmonary artery, has
made routine clinical monitoring of pulmonary artery pressure possible (Swan and Ganz, 1970).Information about pulmonary artery wedge pressure or end diastolic pressure in the pulmonaryartery gives a good indication of the left atrial pressure. This is a very valuable parameter inpredicting and treating left ventricular failure in myocardial infarction in patients undergoingcardiac surgery.
Clinical experience has demonstrated the difficulty in maintaining a high-quality arterial pulse
waveform during direct measurements of blood pressure. Minute leaks in the stopcocks permit asmall quantity of blood to enter the catheter where it clots. Even with a highly leak proof system,clots still form at the catheter tip due to the small volume of blood which may enter as a result ofgauge volume displacement (0.04 mm
3 per 100 mmHg) and any volume displacement of minute
entrapped air bubbles. This type of clotting at the catheter tip can be avoided by using thecontinuous flush system. Pressure transducers presently available incorporate a continuous flusharrangement. The source of fluid for the flushing system (Fig. 6.21) is a plastic bag (600 ml), which
is filled with normal saline and kept at a pressure of 300 mmHg. The high pressure fluid then
flows through a Millipore filter (0.22 m) which is essential to prevent clogging of the fine bore
resistance element and which also serves to filter any bacteria found in the solution. Continuousflush is achieved by using a large resistive element to convert the pressure source to a flow source.With a 0.05 mm diameter glass tubing 1 cm long, flow across the element with 300 mmHg pressureis about 3 ml/h. It is found that large flow rates can cause significant error when using the smalldiameter catheter. Flow rates of 3 ml/h for adults and 0.5 ml/h for children have been found to beadequate. To initially fill the transducer and catheter, a fast flush feature is needed. This is done byusing a rubber valve in the system which when operated permits a fast flush, fills the transducer,and purges the air bubbles from the flush system.

Patient Monitoring Systems 213
Venous pressure measurement can be made by using a strain gauge transducer and a similar
electronic signal processing circuitry. The transducers should be of higher sensitivity to give moreaccurate results at lower pressures. Since the blood pressure is always referred to as theatmospheric pressure at the height of the heart, a correction must be applied while making venouspressure measurements to compensate for the difference of level between the heart and the site ofmeasurement. A correction of 7.8 mmHg is applied for every 10 cm. The site of measurement is
below the height of the heart.
Frequency Response and Damping Adjustment of the Fluid-filled Catheters The frequency com-
ponents of a normal pressure pulse consist of a zero frequency (dc) component, a fundamentalcomponent at the heart rate and harmonics of the fundamental rate. To record a pressure pulse
without any distortion, the measuring system should be capable of recording all the frequency
components with equal amplification and phase shift. The range of uniform frequency responsestarting from dc is determined from an estimation of the highest heart rate expected and thenumber of harmonics to be taken into account. Various authors have suggested that blood pressurepulses contain from 6 to 20 significant harmonics, but generally it is accepted that frequenciesup to 10th harmonics produce significant components in the pressure pulse. If we assume theheart rate to be 90 beats per minute, the same shall be the rate of the arterial blood pressure waves.This means that the upper frequency response should be at least 15 Hz for a heart rate of 90 perminute.
An important factor which requires special consideration in relation to a fluid column pressure
measuring system is the natural frequency or the resonant frequency of the system. The measuringsystem can respond accurately only for frequencies well below the natural frequency. A simplifiedequation which defines the natural frequency of the system is given by
F=
D
LP
V 41
p¥D
D(i)IV set
valvePressure
sleeve
Flush
bag
Drip
chamberIntraflo
Continuous slow flow flushing
Fig.6.21 Continuous slow-flow flushing arrangement for pressure measurements
(Courtesy: Hewlett Packard, USA)

214 Handbook of Biomedical Instrumentation
where D= diameter of the fluid column
L= length of the fluid column
DP= pressure change
DV= volume change for a given DP
This equation assumes that the mass of the moving element of the transducer is very small as
compared to the mass of the fluid. It also ignores the specific gravity of the fluid. DPis the change
in pressure corresponding to a total volume change of DVby the application of the assumed
pressure.
Fluid column systems usually have a very low natural frequency to be commensurate with the
requirements, due to their large inertia and compliance. Therefore, to accurately record the pressurepulses, some form of compensation is necessary to improve the frequency response of the system.This compensation is called damping. In most pressure measuring systems, damping is providedby the viscous resistance of the liquid in the catheter and is given by
D=
4
31
h
r g p ()/2
ccl
EL
NMMO
QPPL
NMMO
QPP(ii)
where D= damping coefficient
r= liquid density in g/cm3
h= viscosity of the liquid in poises
gc= radius of the catheter bore in cm
lc= length of the catheter bore in cm
E= volume elasticity of the sensing element in dynes/cm5
This equation shows that the damping increases inversely as the cube of the catheter diameter
decreases. Equations (i) and (ii) become contradictory because natural frequency decreases directlywith the diameter of the fluid column while the damping ratio increases. Therefore, a compromisemust be reached to obtain a maximum flat frequency response.
Figure 6.22 shows the frequency response curve of a catheter manometer system if filled by the
usual method with cold water. To obtain a uniform flat response, the system has to be suitably
damped. As in the case of recorders, for optimum frequency response, damping coefficient should
be set at 0.7. The damping can be varied by introducing an adjustable constriction into the flowline. Two methods are common. A series damper makes use of a capillary introduced between thecatheter and the manometer. With series damping, it is possible to effect a considerable decrease inthe height of the resonance peak. However, the resonance peak is shifted to a lower value, whichmakes the frequency response worse than without damping. The other method is by using aparallel damper consisting of a variable needle resistance parallel to the manometer and in serieswith a distensible plastic tube connected to a syringe. The parallel damper is able to flatten theresonance peak and a flat response curve is achieved almost up to the original peak.
Pressure waveform distortion due to transmission in fluid-filled catheters can be compensated
by electronic means. The compensator is based on the fact that fluid-filled catheters can becharacterized as second order systems and the distortions introduced by catheters are typical

Patient Monitoring Systems 215
of the output of such systems. It is obvious that a compensator whose transfer function is the
exact inverse of such a second order system would facilitate recovery of the original waveformshape.
In an attempt to evaluate the pressure distortion due to clinical catheter-manometer systems, it
becomes necessary to establish their linearity. A system is considered linear if its transfer functioncoefficients are independent of pressure and time in the applicable zone of pressure and frequency.It may be noted that linearity is to be individually established for different catheters havingdiffering cross-sectional structures.
Special Considerations for the Design of Pressure Transducers for Medical Applications: Physio-
logical pressure transducers are usually linked directly to the patient’s heart and hence they mustensure complete safety of the patient. It is for this reason that the construction of the transducersshould be such that they provide patient-safe isolation. Figure 6.23 shows the construction of onesuch transducer from M/s American Optical, USA. This transducer has three modes of isolation:(i) External isolation of the case with a plastic sheath, which provides protection from extraneousvoltages. (ii) Standard internal isolation of the sensing (bridge) elements from the inside of thetransducer case and from the frame. (iii) Additional internal isolation of the frame from the caseand the diaphragm, in case of wire breakage.
Thus, isolation of the patient/fluid column from electrical excitation voltage is assured, even in
the event of failure of the standard internal isolation. Typically, transducers have maximum
leakage of 2 microamperes at 120 V ac 60 Hz.
Pressure transducers are commonly used in intensive care units, cardiac catheterization and
other situations in which there is a possibility of a defibrillator being used on the patient. If the
transducer breaks down with the application of a defibrillating shock, the current to earth throughthe transducer would be a few miliamperes, assuming that the catheter impedance is 1 M W. This
is probably not significant as far as the patient is concerned, in view of the very high current flowin between the defibrillator paddles. However, there is a possibility of irreversible damage, whichmay be done to the transducer. Transducers, therefore, should be able to withstand high voltages51 0 2 05 0
Frequency100 %
Relative amplitude1. System undamped
2. Parallel damped
3. Series damped1
2
3
Fig.6.22 Frequency response of a fluid-filled catheter system

216 Handbook of Biomedical Instrumentation
arising from electrocautery or defibrillator procedures. The minimum breakdown voltage for this
purpose should be 10,000 V dc.
Physiological pressure transducers need to be sterilized before every use. This is a costly and
time consuming exercise and results in transducer wear and tear due to repeated cleanings andsterilization. This problem has been overcome by designing a presterilized disposable dome witha built-in membrane that provides a sterile barrier between the patient and transducer.
Pressure transducers are sensitive devices. They are adversely affected by steam autoclaving
and Gamma irradiation methods and therefore these methods should not be used. Ethylene oxidegas sterilization followed by aeration is a preferred method. Chemical sterilization is alsoacceptable provided the procedure is based on a thorough knowledge of the agent and the process.Ultrasonic cleaning is damaging to the transducer and should not be used. Sharp and hard objectsshould not be used particularly on or near the diaphragm. Gross soilage and particulate materialand residue should be removed by soaking the transducer only for a few minutes in an acceptablecleaning solution or disinfectant. For best operational performance, the transducers should not beexposed to a temperature higher than 65°C (maximum for short duration) or 50°C for extended
periods of time. The transducer should be thoroughly cleaned as soon as practicable after each
application and maintained in a clean condition for optimum long-term service.
One of the commonest type of abuses of physiological transducers is applying a pressure in
excess of the working pressure either by standing on the pressure line or by flushing out thetransducer with the output closed. A weight of 2 kg applied to a 1 ml disposable syringe is knownto produce a pressure of about 10,000 mmHg. The calibration of the transducers should notsubstantially change (more than 1%) after the maximum rated pressure is applied. The diaphragmrupture pressure values are different for pressure transducers of different makes, and are usuallyin the value range of 3,000 to 10,000 mmHg.
Catheter Tip Pressure Transducers : Fluid-filled catheter and external pressure transducer
arrangements for the measurement of intravascular pressures have limited dynamic response.Metal inner caseStandard isolationIsolated outer case with integral coupling threads
Electrical standoff for secondary connection
Insulated
electrical cable Non-conductive
dome
Liquid
column
Metal sensing
diaphragm
Liquid column Redundant isolation*Internal frame
Unbonded strain gaugeArmatureLiquid column
Conductive material
Non-conductive material0.870
(22 mm)
DIA¢¢
0.710
(18 mm)
DIA¢¢
Fig.6.23 Cross-section of a pressure transducer showing complete isolation of the
patient (Courtesy: American Optical, USA)

Patient Monitoring Systems 217
The problem has been solved to some extent, by the use of miniature catheter tip pressure trans-
ducers. One such transducer which makes use of a semiconductor strain gauge pressure sensor isdescribed by Miller and Baker (1973). The transducer is 12 mm long and has a diameter of 1.65
mm. It is mounted at the tip of a No. 5 French Teflon catheter, 1.5 m long. The other end of the
catheter carries an electrical connector for connecting the transducer to a source of excitation anda signal conditioner. A hole in the connector provides an opening to the atmosphere for the rear ofthe diaphragm. The active portion of the transducer consists of a silicon-rubber diaphragm withan effective area of 0.75 mm
2. Pressure applied to the diaphragm is transmitted by a linkage to two
silicon strain gauges, which form a half bridge, while the other half is constituted by the electricalcircuit.
The transducer gives a high output (0.1 V per 300 mmHg) for 3.5 V excitation and shows an
excellent thermal stability of ±0.15 mmHg/°C over an ambient temperature range of 25–40°C. The
transducer has a working range of ±300 mmHg with a volume displacement of 2 ¥ 10
–3 mm3per
100 mmHg. The dynamic response of the transducer permits high fidelity recording of pressuretransients, anywhere in the vascular system. With a natural resonant frequency of 15 kHz, it isideally suitable for studies requiring accurate analyses of pressures and pressure waveforms,particularly the derivatives of waveforms. The transducer has an epoxy resin insulation over the
strain gauge element within the tip and gives a leakage current of less than 0.5 mA for an applied
voltage of 150 V. Since the transducer system is made to work on 5.4 V excitation, leakage currentfrom this source will be of the order of 0.0216 mA, which is much less than any ac level considered
dangerous for catheters placed directly within the heart chambers.
Nichols and Walker (1974) used a Miller PC-350 catheter tip transducer, a commercial version
of the above described transducer (manufactured by Miller Instruments Inc., Houston, Texas,USA) in experimental and clinical work and found it working satisfactorily as compared tofluid-filled systems (Fig. 6.24). Derived variables such as the maximum rate of change of the leftventricular pressures ( dp/dt ) could be accurately recorded, but were found to be consistently lower
20
cmH O2Mikro-tip pressure transducer “
0
Fluid filled catheter–external transducer
Fig.6.24 A high fidelity pressure tracing from mikro-tip pressure sensor in the pulmo-
nary artery and a simultaneous tracing with artefact from a fluid-filledcatheter (Courtesy: Miller Instruments Inc. USA)

218 Handbook of Biomedical Instrumentation
than those measured with fluid-filled systems. Catheter tip transducers with a fluid-velocity sensor
can be used for making high fidelity pressure and velocity measurements simultaneously, at oneor more locations in the heart (Fig. 6.25).
Matsumoto et al (1978) point out that the catheter tip pressure transducers do not remain stable
over variations in temperature and with long-term clinical use. Safety problems may also be
experienced due to the direct connection between external electronic devices and the heart.
To overcome these problems, they proposed the application of fibre optics to the catheter tippressure transducer for the measurement of intracardiac pressures. The pressure is detected by thephoto-electric transducer element whose output voltage is proportional to the applied pressure.Comparison of waveforms and frequency analysis of waveforms reveal that the fibre optic catheterhas characteristics by no means inferior to the tip transducer type catheter.
/G36/G2E/G37/G2E/G32 /G49/G6E/G64/G69/G72/G65/G63/G74/G20/G4D/G65/G74/G68/G6F/G64/G73/G20/G6F/G66/G20/G42/G6C/G6F/G6F/G64/G20/G50/G72/G65/G73/G73/G75/G72/G65/G20/G4D/G65/G61/G73/G75/G72/G65/G6D/G65/G6E/G74
The classical method of making an indirect measurement of blood pressure is by the use of a cuffover the limb containing the artery. This technique was introduced by Riva-Rocci for the deter-mination of systolic and diastolic pressures. Initially, the pressure in the cuff is raised to a levelwell above the systolic pressure so that the flow of blood is completely terminated. Pressure inthe cuff is then released at a particular rate. When it reaches a level, which is below the systolicpressure, a brief flow occurs. If the cuff pressure is allowed to fall further, just below the diastolicpressure value, the flow becomes normal and uninterrupted.
The problem here finally reduces to determining the exact instant at which the artery just opens
and when it is fully opened. The method given by Korotkoff and based on the sounds producedby flow changes is the one normally used in the conventional sphygmomanometers. The soundsfirst appear (Fig. 6.26) when the cuff pressure falls to just below the systolic pressure. They areproduced by the brief turbulent flow terminated by a sharp collapse of the vessel and persist as thecuff pressure continues to fall. The sounds disappear or change in character at just below diastolicpressure when the flow is no longer interrupted. These sounds are picked up by using a microphone
placed over an artery distal to the cuff. The sphygmomanometric technique is an ausculatory
method; it depends upon the operator recognizing the occurrence and disappearance of theKorotkoff sounds with variations in cuff pressure.
A number of automated blood pressure measuring instruments have been designed which
make use of the Riva-Rocci method. They operate in a manner analogous to that employed by ahuman operator, but differ in the method of detecting the pulsations of blood flow at the systolicand diastolic levels. Frequency bands that best discriminate the Korotkoff sounds at systole anddiastole from the sounds immediately preceding these events must be defined for achieving a highdegree of reliability in the automatic electronic blood pressure instruments. Golden et al (1974)
carried out a special analysis of seven Korotkoff sounds centred about the systolic and diastolicausculatory events and found that a maximum increase in amplitude at the systolic transitionoccurred in the 18–26 Hz band. Similarly, a maximum decrease in spectral energy of diastolicKorotkoff sounds, at ausculatory cessation, was observed within a 40–60 Hz passband.

Patient Monitoring Systems 219
EMF velocity probe
Pulmonaryartery sensor
Right ventricularsensor
Left ventricularsensorRight atrialsensor
AorticsensorRight heart
Left heartLVAO
RV
RAPAECG PA flowvelocity AO flowvelocity
1 second1000 0 50 20 50
cm/sec cm/sec mmHg mmHg
Fig.6.25 Simultaneous recording of pulmonary artery and aortic flow velocity signals with high fidelity
pressure waveforms using mikro-tip transducers (Courtesy: Miller Instruments Inc. USA)

220 Handbook of Biomedical Instrumentation
/G36/G2E/G37/G2E/G32/G2E/G31/G41/G75/G74/G6F/G6D/G61/G74/G69/G63/G20/G42/G6C/G6F/G6F/G64/G20/G50/G72/G65/G73/G73/G75/G72/G65/G20/G4D/G65/G61/G73/G75/G72/G69/G6E/G67/G20/G41/G70/G70/G61/G72/G61/G74/G75/G73/G20/G75/G73/G69/G6E/G67/G20/G4B/G6F/G72/G6F/G74/G6B/G6F/G66/G66/G92/G73/G20/G4D/G65/G74/G68/G6F/G64
The method consists in putting a cuff around the upper part of the patient’s arm and applying a
microphone over the brachial artery. The compressed air required for inflating the cuff is providedby a pumping system incorporated in the apparatus. Usually the inflating is done to a presetpressure level, well beyond the systolic value at the rate of approximately 30 mmHg/s. Thepressure in the cuff is then decreased at a relatively slow pace at the rate of 3–5 mmHg/s. The cuffis to be applied in such a way that the veins are not occluded.
While air is allowed to leak from the cuff, the Korotkoff sounds are picked up by a special piezo-
electric microphone. The corresponding electrical signals are fed to a preamplifier. The amplifiedsignals are then passed on to a bandpass filter having a bandwidth of 25 to 125 Hz. With thispassband, a good signaI-to-noise ratio is achieved when recording Korotkoff sounds from thebrachial artery beneath the lower edge of the cuff. The system is so designed that the appearance
of the first Korotkoff sound switches in the systolic manometer and locks the reading on the
indicating meter. In a similar way, the diastolic value is fixed by the last Korotkoff sound. The cuffis completely deflated, automatically, after an interval of 2–5 s after the determination of thediastolic value.
Instruments operating on this principle are subject to serious errors, particularly in restless
patients, unless steps are taken to ensure protection against artefacts. One method of doing this isPressure
RestdurationInflatingdurationMeasuring duration
One cycle of measurementTime
Deflatingduration
Korotkoff
sounds
Blood flow
pulsationsDiastolic pressure
valueSystolic pressure
value
Fig.6.26 Principle of blood pressure measurement based on Korotkoff sounds

Patient Monitoring Systems 221
to design the control system in such a way that when pressure is registered, the first sound must
be followed by a second one within the preset interval. If this is not the case, the recorded value isautomatically cancelled and the measurement starts again with the subsequent sounds. The
measuring accuracy of such type of instruments is not very high and the error is usually of the
same order (±5 mmHg) as is obtained in clinical sphygmomanometers.
A complete cycle of measurement consists of cuff pumping, controlled deflation, picking up
and evaluation of the Korotkoff sounds, fixing of the systolic and diastolic pressure and then acomplete deflation of the cuff. The cycle is initiated by a time delay and the operation is controlledby a command pulse. Manual operation is, however, always possible.
A number of shortcomings limit the application of the Riva-Rocci method. Probably the most
serious among them is that the measurement is not continuous. Even for a particular singlemeasurement, a number of heart cycles intervene between the determination of the systolic anddiastolic pressures. Moreover, large errors are common since the pressure applied to the exteriorvessel wall is not necessarily identical to that in the cuff, but is attenuated by the intervening tissueand an exact state of flow cannot be precisely determined. This problem is so severe that in most ofthe measurements made with instruments based on this principle, the diastolic pressure value isless reliable than the systolic. The diastolic pressure can be determined with greater accuracy andmore reliability if the microphone output is amplified and fed to a chart recorder. The recorder canbe calibrated in terms of pressure by feeding simulated signals corresponding to 60, 120, 180, 240
and 300 mmHg.
The automatic built-in pump system of inflating and deflating the cuff must be provided with
safety devices so that the patient does not experience any discomfort in the case of system failure.
Provision should be made for immediate switching off of the pump when the pressure in the cuffreaches the preset maximum over-systolic value and in no case should the pressure in the cuff beallowed to exceed 300 mmHg. An additional arrangement must switch off the pump at anypressure after 20 s from starting and deflate the cuff at a constant rate. These devices shall ensurethat the pressure in the cuff will not reach too high a value and that no pressure is kept longer thanapproximately 20 s.
/G36/G2E/G37/G2E/G32/G2E/G32/G54/G68/G65/G20/G52/G68/G65/G6F/G67/G72/G61/G70/G68/G69/G63/G20/G4D/G65/G74/G68/G6F/G64
A fully automatic apparatus for measuring systolic and diastolic blood pressures has been
developed using the ordinary Riva-Rocci cuff and the principle of rheographic detection of anarterial pulse. Here, the change in impedance at two points under the occluding cuff forms the
basis of detection of the diastolic pressure.
In this method, a set of three electrodes (Fig. 6.27), which are attached to the cuff, are placed in
contact with the skin. A good contact is essential to reduce the skin electrode contact impedance.
Electrode Bwhich acts as a common electrode is positioned slightly distal from the mid-line of the
cuff. Electrodes A and C are placed at a certain distance from the electrode B, one distally and the
other proximally. A high frequency current source operating at 100 kHz is connected to theelectrodes Aand C. When we measure the impedance between any two electrodes before
pressurizing the cuffs, it shows modulation in accordance with the blood flow pulsations in theartery. Therefore, arterial pulses can be detected by the demodulation and amplification of thismodulation.

222 Handbook of Biomedical Instrumentation
Distal
amp.
Diff.
amp.
BCA
Diastolic pressure
indicator PumpSystolic
pressure
indicator
Fig.6.27 Rheographic method of indirect blood pressure measurement
When the cuff is inflated above the systolic value, no pulse is detected by the electrode A. The
pulse appears when the cuff pressure is just below the systolic level. The appearance of the firstdistal arterial pulse results in an electrical signal, which operates a valve to fix a manometerpointer on the systolic value. As long as the pressure in the cuff is between the systolic anddiastolic values, differential signal exists between the electrodes A and C. This is because the
blood flow is impeded underneath the occluding cuff and the pulse appearing at the electrode A
is time delayed from the pulse appearing at C. When the cuff pressure reaches diastolic pressure,
the arterial blood flow is no longer impeded and the differential signal disappears. A commandsignal is then initiated and the diastolic pressure is indicated on the manometer.
In the rheographic method of measuring blood pressure, the cuff need not be precisely
positioned as in the case with the Korotkoff microphone, which is to be fixed exactly above an
artery. Also the readings are not affected by ambient sounds.
Similar to the rheographic method of measuring arterial blood pressure non-invasively, is the
photo-electric plethysmograph, which detects cardiac volume pulse and transduces it to an
electrical signal. The rate of pulsatile blood-volume change in a peripheral site, such as the earlobeor finger tip is used for indicating the magnitude of the arterial systolic blood pressure aftercalibration against a standard blood pressure measuring means. The correlation between the rateof the peripheral pulsatile blood-volume change and the arterial systolic blood pressure is moreclose if these quantities are measured during a portion of each cardiac cycle, i.e. period of diastoleor systole and for just one instant during this period.
/G36/G2E/G37/G2E/G32/G2E/G33/G44/G69/G66/G66/G65/G72/G65/G6E/G74/G69/G61/G6C/G20/G41/G75/G73/G63/G75/G6C/G74/G61/G74/G6F/G72/G79/G20/G54/G65/G63/G68/G6E/G69/G71/G75/G65
The “differential auscultatory technique” is a non-invasive method for accurately measuring
blood pressure. A special cuff-mounted sensor consisting of a pair of pressure sensitive elements,isolates the signal created each time the artery is forced open.
Figure 6.28 illustrates how high frequency pulses are created each time, the intra-arterial
pressure exceeds the cuff pressure. As long as the cuff pressure exceeds the pressure in the artery,
the artery is held closed, and no pulse is generated. However, as soon as the intra-arterial pressure
rises to a value, which momentarily exceeds the cuff pressure, the artery “snaps” open; and a pulseis created. Once the artery is open, blood flows through it giving rise to the low frequency pressure

Patient Monitoring Systems 223
wave signal, which lasts until the arterial pressure again drops below the cuff pressure. This
process is repeated until the cuff pressure drops to a value below the diastolic.
Figure 6.29 is a cut away view of an arm with a cuff partially occluding the brachial artery. Each
time the artery opens, the signal shown in Fig. 6.30 is created. Note that this signal consists of aslowly rising, low frequency component (in the frequency range of 0.5–5 Hz) with a fast “pulse”(frequencies approximately 10–80 Hz) superimposed on it. This signal is denoted by the arrowsmarked A in Fig. 6.29 transmitted from the artery to both the sensor and the air bag in the cuff.
Due to the air bag characteristics, the high frequency component is highly attenuated, leaving
only the low-frequency signal, as shown in Fig. 6.30(b). Therefore, only the low frequency signalis transmitted to the side of the sensor facing the air bag, as denoted by the arrows marked B in
Fig. 6.29. Since most artefact signals (unwanted signals due to motion, etc.) fall in a frequency
range below 10 Hz, they are also transmitted to both sides of the sensor.
The systolic pressure is determined as the pressure at which the first opening of the artery
occurs, as shown by the first pulse in Fig. 6.30(c) , because this pulse is created the first time theartery is forced open by intra-arterial pressure. Similarly, diastolic value is determined as thepressure at which the differential signal essentially disappears, because this corresponds to thelast time the artery is forced open. The differential sensor subtracts the side “B” signal from theside “A” signal, thereby cancelling out the pressure wave component and the motion artefactsignals, and the higher frequency Korotkoff signals are isolated.
/G36/G2E/G37/G2E/G32/G2E/G34/G4F/G73/G63/G69/G6C/G6C/G6F/G6D/G65/G74/G72/G69/G63/G20/G4D/G65/G61/G73/G75/G72/G65/G6D/G65/G6E/G74/G20/G4D/G65/G74/G68/G6F/G64
The automated oscillometric method of non-invasive blood pressure measurement has distinct
advantages over the auscultatory method. Since sound is not used to measure blood pressure inCuff pressure Intra-arterial
pressure
PulsePressure
waveTime
Time(b)
Signal
detected
by
sensor A(a)
Systolic
pressure
(mmHg)
diastolic
Fig.6.28 (a) Diagram showing the relationship between cuff pressure and
intra-arterial pressure
(b) Signal created by the relative pressure changes

224 Handbook of Biomedical Instrumentation
Artery
Wall expansion
boundaries
Cuff with
air bladder
Differential
sensor
AB
Air tube to
machine
Sensor signal
leads
Blood
flowArmDistal
endAir
bagProximal
endA
A
AB
B
Fig.6.29 Cut-away view showing signal detection
(a)
(b)
(c)Systolic
A-BDiastolicSignal A
Signal B
Fig.6.30 (a) Signal generated by artery as air is bled from cuff
(b) Signal ‘A’ after filtering by air bag (equivalent of oscillometric signal
created by expansion of arterial walls; frequencies are in 0.5 – 5 Hz range)
(c) Differential signal due to opening and closing of artery (frequency range
10-80 Hz)

Patient Monitoring Systems 225
the oscillometric technique, high environmental noise levels such as those found in a busy clinical
or emergency room do not hamper the measurement. In addition, because this technique does not
require a microphone or transducer in the cuff, placement of the cuff is not as critical as it is with
the auscultatory or Doppler methods. The oscillometric method works without a significant lossin accuracy even when the cuff is placed over a light shirt sleeve. The appropriate size cuff can beused on the forearm, thigh, or calf as well as in the traditional location of the upper arm. Adisadvantage of the oscillometric method, as well as the auscultatory method, is that excessivemovement or vibration during the measurement can cause inaccurate readings or failure to obtainany reading at all.
The oscillometric technique operates on the principle that as an occluding cuff deflates from a
level above the systolic pressure, the artery walls begin to vibrate or oscillate as the blood flowsturbulently through the partially occluded artery and these vibrations will be sensed in thetransducer system monitoring cuff pressure. As the pressure in the cuff further decrease, theoscillations increase to a maximum amplitude and then decrease until the cuff fully deflates andblood flow returns to normal.
The cuff pressure at the point of maximum oscillations usually corresponds to the mean arterial
pressure. The point above the mean pressure at which the oscillations begin to rapidly increase inamplitude correlates with the diastolic pressure (Fig. 6.31). These correlations have been derived
and proven empirically but are not yet well explained by any physiologic theory. The actual
determination of blood pressure by an oscillometric device is performed by a proprietary algorithmdeveloped by the manufacturer of the device.
The oscillometric method is based on oscillometric pulses (pressure pulses) generated in the
cuff during inflation or deflation. Blood pressure values are usually determined by the applicationof mathematical criteria to the locus or envelope formed by plotting a certain characteristic, calledthe oscillometric pulse index, of the oscillometric pulses against the baseline cuff pressure(Fig. 6.32). The baseline-to-peak amplitude, peak-to-peak amplitude, or a quantity based on thepartial or full time-integral of the oscillometric pulse can be used as the oscillometric pulse index.The baseline cuff pressure at which the envelope peaks (maximum height) is generally regardedas the MAP (mean arterial pressure). Height-based and slope-based criteria have been used todetermine systolic and diastolic pressures.
An envelope that has been normalized with respect to the peak index can also be used for the
determination of the oscillometric blood pressure. The ECG-gating technique has been used toassist in the identification of oscillometric pulse signals. Measurement sites for oscillometric bloodpressure measurement include the upper arm, forearm, wrist, finger and thigh.
Most of the patient monitoring systems are based on the oscillometric measuring principle.
Figure 6.33 shows the major functional parts of a NIBP (non-invasive blood pressure) measuring
system.
An air pump is used to automatically inflate the patient cuff. The pump is of a membrane type
and is enclosed in a foam rubber filled casing to attenuate noise. The pneumatic unit includes
damping chambers to (i) prevent a rapid increase of pressure caused by the pump, (ii) slow downthe pressure change in the measurement of infant and (iii) smooth down rapid pressure pulsescaused by the bleed valve. A safety valve prevents accidental cuff over-pressurization and operatesnominally at 330 mmHg. A bleed valve is incorporated to release the cuff pressure. The opening of

226 Handbook of Biomedical Instrumentation
04080120160200
Mean
DiastolicSystolic
Cuff pressure mmHgPressure mmHg
Oscillations in cuff pressureDiastolic Mean Systolic
Oscillometric methodCuff pressure oscillationsCuff pressure
110
100
90
80
70
60
50
40
30
20
10
0
(b)(a)
Fig.6.31 Illusration of oscillometric method of blood pressure measurement

Patient Monitoring Systems 227
Diastolic
pressure
(slope-based
criterion)
Diastolic
pressure
(height-based
criterion)Systolic
pressure
(height-based
criterion)Systolic
pressure
(slope-based
criterion)Mean arterial
pressure
(MAP)Point of
maximum slope
Point of
minimum slopeHHSM/
= systolic ratioHHDM/
= diastolic
ratioPoint of maximum height
Envelope/locus
HMHD
HS
Baseline cuff
pressure00Oscillometric pulse index
Fig.6.32 Criteria for oscillometric blood pressure determination
the valve is pulse–width controlled between 100% (valve fully open) and 0% (valve fully closed).
The driving signal frequency is 40 Hz. For quickly deflating the cuff, an exhaust valve is provided.The solenoid (magnetic) valves and the air pump are controlled by an open collector darlingtondriver circuit. A watch dog timer prevents a prolonged inflation. The operation of the pump and
the operation of the valves are all under microprocessor control.
A piezo-resistive pressure transducer is connected to the cuff. It measures the absolute pressure
of the blood pressure cuff and the pressure fluctuations caused by the arterial wall movement. The
pressure transducer is excited by a 4 mA constant current source. The output of the pressuretransducer is a differential signal, which is amplified, in a differential amplifier with a gain of 30.This is followed by zero-control and gain control circuits. A dc channel is used to measure thestatic or non-oscillating pressure of the blood pressure cuff. The ac component of the pressure datais amplified to allow the processor to analyze the small cuff pressure fluctuations, which are usedas a basis for blood pressure determination. This is followed by a second order high-pass filter toeffectively block out the dc component of the pressure signal. The next stage is a low-pass filter,which blocks the offset voltages.

228 Handbook of Biomedical Instrumentation
Fig.6.33 Fundamental parts of NIBP measuring system
AdultcuffInfantcufftwinhoseOuter/upperchannelInner/lowerchannel
DampingchamberIIDampingchamberIII
DampingchamberI
Safety valveExhaust valveBleed valveSwitch
Pressure transducerPressure dc channelAMP
Highpass
Highpass
Gaincontrol
Zerocontrol
NIBP boardB1
X1X3X2Y1P1P3P0P2
Y2
Y4
Y3Pressure AC channelGainlowpassA–Dconv-erter
AC-reset
Multiplexer
Darlingtondriver
PWMEPROM64 K byteRAM8 K byte
Data busAddress bus
AddresslatchAddressdecoder
Choke
MG
Pump moduleWatchdogcircuit Serialinterface
CPU

Patient Monitoring Systems 229
The analog multiplexer circuit is used to select either the dc or ac channel to the A-D convertor,
whose output goes to the processor circuit. The 8051 microcontroller alongwith its associatedmemory circuits allow the control of the whole measuring procedure.
/G36/G2E/G37/G2E/G32/G2E/G35/G55/G6C/G74/G72/G61/G73/G6F/G6E/G69/G63/G20/G44/G6F/G70/G70/G6C/G65/G72/G20/G53/G68/G69/G66/G74/G20/G4D/G65/G74/G68/G6F/G64
Automatic blood pressure monitors have also been designed based on the ultrasonic detection of
arterial wall motion. The control logic incorporated in the instrument analyzes the wall motionsignals to detect the systolic and diastolic pressures and displays the corresponding values.
As explained in Chapter 11, the observed Doppler frequency can be expressed as:
Df=2Vt
cl(i)
where D f= Doppler frequency (Hz)
Vt= velocity of the object (m/s)
lc= carrier wavelength (m)
For blood pressure measurement, the brachial artery is the object from where the ultrasound
gets reflected. Arterial movement produces the Doppler frequency shift.
lc=V
fc
c(ii)
where lc= wavelength (in metres) of the carrier frequency in the medium
Vc= velocity of the carrier frequency in the medium (1480 m/s in water)
fc= carrier frequency in the medium (2 MHz)
Substituting these values in equation (ii),
lc=1480
21 06¥ = 0.74 ¥ 10–3m
The Doppler frequency is expressed as
Df= 2
07 4 1 03Vt
.¥- = 2.7¥ 103Vt (Hz) (iii)
Thus, D fvaries directly with the target velocity, i.e. the motion of the brachial artery.
To measure blood pressure, the Doppler frequency shift due to the snapping action of the artery
must be known. The arterial movement with the opening and closing of the artery is 5 ¥ 10–3 m,
assuming that the snapping occurs in 0.1 s ( Dt), the arterial wall velocity is
Vt=D
Dd
t=51 0
013¥-
. = 50 ¥ 10–3 m/s
By substituting values for Vt, and lc into the Doppler equation (iii), the artery motion Doppler
frequency is
2.7¥ 103Vt= 2.7 ¥ 103¥ 50 ¥ 10–3 = 135 Hz
Instruments making use of ultrasonic Doppler-shift principle for the measurement of blood
flow are based on the detection of the frequency shift ascribed to back scattering from moving

230 Handbook of Biomedical Instrumentation
blood particles. On the other hand, the blood pressure instrument filters out these higher frequency
reflections and senses the lower frequency refractions originating from the movement of therelatively slow moving arterial wall.
In principle, the instrument consists of four major subsystems (Fig. 6.34). The power supply
block converts incoming ac line voltage to several filtered and regulated dc voltages required for
the pneumatic subsystem in order to inflate the occlusive cuff around the patient’s arm.
Pneumatic
subsystemControl
subsystemPower
supply
subsystemRF and
audio
subsystem
Patient’s
arm
ACDC
DC
DC Cuff
AC pressureControl
signalStart AudioTransmit 2 MHzReceive
Fig.6.34 Major subsystems in ultrasonic blood pressure monitor (Courtesy: Roche
Medical Electronics Division, USA)
At the same time, control subsystem signals gate-on the transmitter in the RF and audio
subsystem, thereby generating a 2 MHz carrier, which is given to the transducer located in the cuff.
The transducer converts the RF energy into ultrasonic vibrations, which pass into the patient’sarm. The cuff pressure is monitored by the control subsystem and when the pressure reaches thepreset level, further cuff inflation stops. At this time, audio circuits in the RF and audio subsystemsare enabled by control subsystem signals, and the audio signals representative of any Dopplerfrequency shift are thus able to enter the control subsystem logic.
The control subsystem signals the pneumatic subsystem to bleed off the cuff pressure at a rate
determined by the preset bleed rate. As air bleeds from the cuff, the frequency of the returned RF isnot appreciably different from the transmitted frequency as long as the brachial artery remainsoccluded. Till then, there are no audio signals entering the control subsystem.
At the systolic pressure, the occluded artery snaps open and the arterial blood flow starts. This
artery motion results in a Doppler shift in the returning ultrasonic vibrations. The converted audio

Patient Monitoring Systems 231
frequency signal is recognized as tentative systolic by the control subsystem logic. Four valid
artery returns must be recognized in order to register the tentative systole and for it to become fixedas true systole. This reduces the possibility of artefacts from recording a false systole reading. As
a further check, the audio returns are examined for width and rate of occurrence to prevent artefacts
from being accepted as true artery returns. Pulses more than 125 milliseconds wide or occurringmore frequently than every 250 milliseconds are rejected by the control subsystem logic. This setsthe upper limit (240 bpm) on the patient’s heart rate to which this instrument can function. Thelower limit of response is 24 bpm. At diastole, cuff pressure equals or slightly exceeds the arterialwall pressure. As a result, wall snapping ceases and the RF and audio subsystem no longerreceive the Doppler shifted returns. The reading is registered, and the cuff pressure is allowed todeflate rapidly to atmospheric pressure. The readings are held fixed until a new measurementcycle is initiated.
An occlusive cuff is placed on the arm (Fig. 6.35) in the usual manner, with an ultrasonic
transducer on the arm over the brachial artery. The cuff is inflated first to above systolic pressureand then deflated at a specified rate. A low energy ultrasonic beam (less than 50 mw/cm
2) at a
frequency of 2 MHz is transmitted into the arm. The portion of the ultrasound that is reflected bythe arterial wall shifts in frequency when the wall of the artery moves. Above systolic, the vessel
remains closed due to the pressure of the occluding cuff, and the monitor signals are not received.
As the cuff pressure falls to the point where it is just overcome by the brachial artery pressure, theartery wall snaps open. This opening wall movement, corresponding to the occurrence of the firstKorotkoff sound, produces a Doppler-shift which is interpreted by logic in the instrument assystolic and displayed accordingly. With each subsequent pulse wave, a similar frequency shift is
GeneratorRF unit Indicator
Receiver
Logic
controlPumpReceiver
transducerTransmit
transducerArm
Artery
Pressure
cuff
Fig.6.35 Measurement of blood pressure using ultrasonic Doppler-shift principle

232 Handbook of Biomedical Instrumentation
produced until at the diastolic pressure the artery is no longer occluded. Its rapid motion suddenly
disappears and the Doppler-shift becomes relatively small. The instrument notes the suddendiminution in the amplitude of the Doppler shift and cuff pressure at this point is displayed as
diastolic pressure. Special electronic circuits used in the instrument help to discriminate against
extraneous motion artefacts. A coupling medium is essential between the transducer and thepatients’ skin for the efficient transmission of ultrasonic energy.
Unlike the Korotkoff method, the instruments based on the ultrasonic Doppler-shift principle
often provide reliable blood pressure measurements in severe hypotensive states, at unfavourablesites such as the popliteal artery, in neonates where no other indirect method of measurement isfeasible, in patients too obese for successful ausculation, under unfavourable conditions such ashigh ambient noise, and in many species of laboratory research animals.
/G20/G36/G2E/G38 /G4D/G45/G41/G53/G55/G52/G45/G4D/G45/G4E/G54/G20/G4F/G46/G20/G54/G45/G4D/G50/G45/G52/G41/G54/G55/G52/G45
The transducer normally used for temperature measurement in a patient monitoring system is athermistor. Changes in resistance of the thermistor with changes in temperature are measured ina bridge circuit and indicated on a calibrated meter. The measuring range is 30–42°C.
In a patient monitoring system, provision for two channel temperature measurements are
usually made. Similar to ECG monitoring, the output circuits are isolated through opto-couplers.Provision for inoperate conditions are also included in such type of monitoring equipment.
/G20/G36/G2E/G39 /G4D/G45/G41/G53/G55/G52/G45/G4D/G45/G4E/G54/G20/G4F/G46/G20/G52/G45/G53/G50/G49/G52/G41/G54/G49/G4F/G4E/G20/G52/G41/G54/G45
The primary functions of the respiratory system are to supply oxygen and remove carbon dioxidefrom the tissues. The action of breathing is controlled by a muscular action causing the volume ofthe lung to increase and decrease to effect a precise and sensitive control of the tension of carbondioxide in the arterial blood. Under normal circumstances, this is rhythmic action with the resultthat the respiration rate provides a fairly good idea about the relative respiratory activity. Severaltechniques have been developed for the measurement of the respiration rate. The choice of aparticular method depends mostly upon the ease of application of the transducer and theiracceptance by the subject under test. Some of the commonly used methods for the measurement ofrespiration rate are explained below.
/G36/G2E/G39/G2E/G31 /G44/G69/G73/G70/G6C/G61/G63/G65/G6D/G65/G6E/G74/G20/G4D/G65/G74/G68/G6F/G64
The respiratory cycle is accompanied by changes in the thoracic volume. These changes can besensed by means of a displacement transducer incorporating a strain gauge or a variable resistanceelement. The transducer is held by an elastic band, which goes around the chest. The respiratorymovements result in resistance changes of the strain gauge element connected as one arm ofa Wheatstone bridge circuit. Bridge output varies with chest expansion and yields signalscorresponding to respiratory activity.
Changes in the chest circumference can also be detected by a rubber tube filled with mercury.
The tube is fastened firmly around the chest. With the expansion of the chest during an inspiratory

Patient Monitoring Systems 233
phase, the rubber tube increases in length and thus the resistance of the mercury from one end of
this tube to the other changes. Resistance changes can be measured by sending a constant currentthrough it and by measuring the changes in voltage developed with the respiratory cycle.
/G36/G2E/G39/G2E/G32 /G54/G68/G65/G72/G6D/G69/G73/G74/G6F/G72/G20/G4D/G65/G74/G68/G6F/G64
Since air is warmed during its passage through the lungs and the respiratory tract, there is adetectable difference of temperature between inspired and expired air. This difference of tempera-ture can be best sensed by using a thermistor placed in front of the nostrils by means of a suitableholding device. In case the difference in temperature of the outside air and that of the expired air issmall, the thermistor can even be initially heated to an appropriate temperature and the variationof its resistance in synchronism with the respiration rate, as a result of the cooling effect of the air
stream, can be detected. This can be achieved with thermistor dissipations of about 5 to 25 mW.
Excessive thermistor heating may cause discomfort to the subject. The thermistor is placed as partof a voltage dividing circuit or in a bridge circuit whose unbalance signal can be amplified toobtain the respiratory activity. The method is simple and works well except in the case of somepatients who object to having anything attached to their nose or face. This method is found tosatisfy the majority of clinical needs including for operative and post-operative subjects.
Occasionally, unconscious patients display a tendency for the uncontrolled tongue to block the
breathing system. Under such systems, we are often faced with the situation that not a singlemillilitre of air is inhaled but the patient’s thorax is carrying out large, even though frustralbreathing motions. In this condition, the impedance pneumograph and switch methods will showthe correct state. Putting the thermistor in a tracheal cannula is not simple. There it is very sooncovered with excretions. In the case of suffocated patients with no spontaneous respirationmotions, those few millilitres that pass through the cannula are sufficient to drive the breath ratemeter. This is a drawback in the technique of using thermistors for the detection of respiration rate.
/G36/G2E/G39/G2E/G33 /G49/G6D/G70/G65/G64/G61/G6E/G63/G65/G20/G50/G6E/G65/G75/G6D/G6F/G67/G72/G61/G70/G68/G79
This is an indirect technique for the measurement of respiration rate. Using externally appliedelectrodes on the thorax, the impedance pneumograph measures rate through the relationshipbetween respiratory depth and thoracic impedance change. The technique avoids encumberingthe subject with masks, tubes, flowmeters or spirometers, does not impede respiration and hasminimal effect on the psychological state of the subject. Impedance method for measuringrespiration rate consists in passing a high frequency current through the appropriately placed
electrodes on the surface of the body (Fig. 6.36) and detecting the modulated signal. The signal is
modulated by changes in the body impedance, accompanying the respiratory cycle. The electrodeused for impedance pneumograph are of the self-adhesive type. Contact with the skin is madethrough the electrode cream layer for minimizing motion artefacts. The electrodes, when the skinis properly prepared, offer an impedance of 150 to 200 W. The change in impedance corresponding
to each respiratory cycle is of the order of 1% of the base impedance.
The two electrode impedance pneumograph is convenient for use with quiet subjects. Movement
artefacts are produced due to changes in the electrode contact impedance, in case the subject is

234 Handbook of Biomedical Instrumentation
moving. These artefacts can be significantly reduced by using a four electrode impedance
pneumograph. In this case, the output from the oscillator is applied to the two outer electrodes. Bydoing so, the main oscillator current does not flow through the contact impedance of the measuringelectrodes. This system is useful for monitoring restless subjects such as babies.
To avoid the stimulation of sensory receptors, nerves and muscle, currents higher in frequency
than 5 kHz must be used for the measurement of physiological events by impedance. Frequencieslower than 5 kHz are particularly hazardous since ventricular fibrillation may be produced withsubstantial current flow. The use of higher frequencies not only provides the protection sought inthe avoidance of tissue stimulation, but also provides the safe use of currents of magnitude, whichcould be lethal if the frequencies were lower.
Electrical impedance changes associated with physiological activity have been studied exten-
sively. Some of the physiological quantities which have been measured and recorded by theimpedance method include respiration, blood flow, stroke volume, autonomic nervous systemactivity, muscle contraction, eye movement, endocrine activity and activity of the brain cells.
The impedance-based method of measuring respiration rate is commonly employed in patient
monitoring systems. The electrodes used for this purpose are the same as those used for ECGmeasurement. The dynamic measuring range of the amplifier is 0.1 to 3.0 W with a frequency
response of 0.2 to 3.0 Hz corresponding to respiratory rate of 12 to 180 per minute. The amplifier
operates within an impedance window established by the static impedance level (approx.3 k ohms) and its output produces a respiratory waveform from which respiratory rate is derived.
/G36/G2E/G39/G2E/G34 /G43/G4F/G32/G20/G4D/G65/G74/G68/G6F/G64/G20/G6F/G66/G20/G52/G65/G73/G70/G69/G72/G61/G74/G69/G6F/G6E/G20/G52/G61/G74/G65/G20/G4D/G65/G61/G73/G75/G72/G65/G6D/G65/G6E/G74
Respiration rate can also be derived by continuously monitoring the CO2 contained in the subject’s
alveolar air. Measurement of CO2 in expired air is otherwise useful in several ways; for example,
for originally setting up the respirator and in making adjustments to it afterwards, supervising
patients suffering from respiratory paralysis, and other cases where there is respiratory involve-
ment.
The measurement is based on the absorption property of infrared rays by certain gases. Suitable
filters are required to determine the concentration of specific gases (like CO2, CO, and NO2)
constituting the expired air. Rare gases and diatomic gases do not absorb infrared rays.Demodulator
and filterRecorderDC backing
voltage
50 kHz
amplifier50 kHz
oscillator
R
AZ DZ
BR
Fig.6.36 Principle of impedance pneumograph (two electrode method)

Patient Monitoring Systems 235
When infrared rays are passed through the expired air containing a certain amount of CO2,
some of the radiations are absorbed by it. There is a proportional loss of heat energy associatedwith the rays. The detector changes the loss in heating effect of the rays into an electrical signal.
This signal is used to obtain the average respiration rate.
Figure 6.37 shows the arrangement for the detection of CO
2 in expired air. Two beams of equal
intensity of infrared radiations from the hot-wire spirals fall on one half of each of the condenser
microphone assembly. The detector has two identical portions separated by a thin, flexible metaldiaphragm. The detector is filled with a sample of pure CO
2. Because of the absorption of CO2 in
the analysis cell, the beam falling on the test side of the detector is weaker than that falling on thereference side. The gas in the reference side would, therefore, be heated more than that on theanalysis side. As a result, the diaphragm is pushed slightly to the analysis side of the detector. Thediaphragm forms one plate of a capacitor. The infrared beams are chopped at 25 Hz and thealternating signal which appears across the detector is amplified, shaped and suitably integratedto give the respiration rate.
Integrator
RecorderCuvette for referenceInfrared sourceChopper motor
Chopper
Test
gas
Detector
Condenser
AmplifierCalibrated
meter
Fig.6.37 Schematic diagram for detection of CO2 in the expired air for continuous
monitoring of respiration rate
/G36/G2E/G39/G2E/G35 /G41/G70/G6E/G6F/G65/G61/G20/G44/G65/G74/G65/G63/G74/G6F/G72/G73
Apnoea is the cessation of breathing which may precede the arrest of the heart and circulation in
several clinical situations such as head injury, drug overdose, anaesthetic complications andobstructive respiratory diseases. Apnoea may also occur in premature babies during the firstweeks of life because of their immature nervous system. If apnoea persists for a prolonged period,brain function can be severely damaged. Therefore, apnoeic patients require close and constantobservation of their respiratory activity. Apnoea monitors are particularly useful for monitoringthe respiratory activity of premature infants.

236 Handbook of Biomedical Instrumentation
Several contactless methods are available for monitoring the respiration of infants. The most
successful apnoea monitors to-date have been the mattress monitors. These instruments rely fortheir operation on the fact that the process of breathing redistributes an infant’s weight and this is
detected by some form of a pressure sensitive pad or mattress on which the infant is nursed. The
mattress, in its simplest form, is a multi-compartment air bed, and in this case the weight redistri-bution forces air to flow from one compartment to another. The air flow, is detected by the coolingeffect it produces on a heated thermistor bead. Though the technique is simple, the main disad-vantage with the air mattress is the short-term sensitivity variation and the double peaking effectwhen inspiration or expiration produce separate cooling of the thermistor.
Alternatively, a capacitance type pressure sensor in the form of a thin square pad is usually
placed under or slightly above the infant’s head. Respiratory movements produce regular pressurechanges on the pad and these alter the capacitance between the electrode plates incorporated inthe pad. This capacitance change is measured by applying a 200 kHz signal across the electrodesand by detecting the current flow with a phase-sensitive amplifier. Two types of electrodes can beused: (i) 70 mm plates, 350 mm apart in a plastic tube which is placed alongside the body; (ii) 250mm long, 60 mm diameter cylinders placed one on either side of the body. This system is much toosensitive to people moving nearby and thus an electrically screened incubator is essential for the
infant.
Impedance pneumography is another practical method to monitor the breathing of the patient.
The technique also enables the simultaneous monitoring of the heart rate and respiration. The
heart rate is known to drop during apnoea. Monitoring the heart rate and respiration thus gives anextra measure of security. Electrodes measure the effort to breath and not the actual ventilation(Kulkarni, 1991). Impedance pneumography has certain inherent disadvantages. One is that theplacement of the electrodes is very critical and the other is cardiovascular artefact. This resultsfrom the detection of movement between the electrodes because of the cardiovascular system,rather than due to respiration. Apnoea monitors need to be designed to reject this artefact. Silvola(1989) describes a new non-invasive piezo-electric transducer for the recording of respiration,heart rate and body movements using the PVDF (polyvinylidenefluoride) polymer film. Thetransducer consists of an area of about 1000 cm
2 PVDF film (length 40–50 cm, width 20–30 cm)
with a thickness of 40 mm. The PVDF elements are placed directly on the bed mattress without
being fixed on the skin. The recordings can be performed when the subject is lying on their back,
stomach or on their side.
Apnoea monitors are generally designed to give audio-visual signals under apnoeic conditions
when no respiration occurs within a selectable period of 10, 20 or 30 s. The apnoea monitors are
basically motion detectors and are thus subject to other motion artefacts also which could givefalse readings. The instruments must, therefore, provide means of elimination of these errorsources. Figure 6.38 shows a block diagram of an apnoea monitor. It basically consists of an inputamplifier circuit, motion and respiration channels, a motion/respiration discrimination circuit,and an alarm circuit. The input circuit consists of a high input impedance amplifier which couplesthe input signal from the sensor pad to the logic circuits. The sensor may be a strain gaugetransducer embedded in the mattress. The output of the amplifier is adjusted to zero volts withoffset adjustment provided in the amplifier. The amplified signal goes to motion and respirationchannels connected in parallel. The motion channel discriminates between motion and respiration

Patient Monitoring Systems 237
as a function of frequency. In the case of motion signals, high level signals above a fixed threshold
are detected from the sensor. In the respiration channel, a low-pass filter is incorporated. Lowfrequency signals below 1.5 Hz (respiration) cause the output of the Schmidt trigger circuit topulse at the respiration rate. Higher frequency signals, above 1.5 Hz (motion), cause the output ofthe trigger to go positive. Absence of the signal (apnoea) causes the output of the Schmidt trigger to
go negative.
The outputs of the motion and the respiration signals are combined in a comparator circuit,
which compares the polarities of the motion and respiration channel signals to indicate respira-tion. The presence of respiration is indicated by a flashing lamp. The output of the discriminationdetector also goes to an apnoea period selector circuit, a low frequency alarm oscillator and driver,a tone oscillator and audio amplifier connected to a speaker. Audible alarm is given at a frequencyof 800–1000 Hz, which is pulsed at 2 Hz.
An alternative method of detecting apnoea is based on electromagnetic induction. It consists in
passing a high frequency alternating current through a transmitting coil and creating an alternatingmagnetic field. The transmitting coil is placed at some distance from the infant. The receiving coilis applied on to the abdomen of the infant. The alternating magnetic field induces an emf inthe receiving coil. The movement of the abdominal wall with the infant’s respiration results ininducing an amplitude-modulated signal in the receiving coil. If this amplitude-modulated signalis demodulated, the modulation frequency corresponding to the respiration frequency can be
recovered.
Another contactless method for monitoring the breathing activity of premature babies is by the
use of microwave energy. It operates by directing low intensity microwave energy (10 GHz) at the
individual to be monitored and detecting this energy, after its reflection from the moving surface.The difference between the transmitted and received microwave frequencies (Doppler-shift)Motionresp.
discrimination
circuitNegative
detectorPositive
detectorMotion channel
Motion
lamp
Resp.
lamp
Apnoea period
selector
circuitSensor
amplifierFrom
sensor
Low-pass
filterSchmitt
triggerAlarm
circuitRespiration channel
Fig.6.38 Block diagram of apnoea monitor (Courtesy: B-D Electrodyne, USA)

238 Handbook of Biomedical Instrumentation
provide a signal voltage, which can be amplified, and used as an indicator of the continuance of
respiratory activity.
/G20/G36/G2E/G31/G30 /G43/G41/G54/G48/G45/G54/G45/G52/G49/G5A/G41/G54/G49/G4F/G4E/G20/G4C/G41/G42/G4F/G52/G41/G54/G4F/G52/G59/G20/G49/G4E/G53/G54/G52/G55/G4D/G45/G4E/G54/G41/G54/G49/G4F/G4E
An important field where continuous patient monitoring can aid in carrying out sophisticatedprocedures which otherwise would have been either impossible or difficult to carry out is thecardiac catheterization laboratory. These procedures result in huge amounts of data which needto be acquired processed, analyzed, correlated and stored. The calculations involved can be donemore accurately and rapidly by using computer-based systems.
A cardiac catheterization laboratory is a place for carrying out specialized catheterization
procedures, providing surgical facilities for pacemaker implantation and carrying out advancedresearch in biomedical engineering. Facilities in catheterization laboratories include systems forrecording of electrocardiograms, intra-cardiac pressures in the left ventricle (LV) and pulmonaryartery, atrial pressure, right atria pressure and dye densitometer and then processing to get thepulmonary artery systolic/end diastolic pressure, LV systolic/diastolic pressures cardiac output(CO) and cardiac index (CI), stroke volume, intra-cardiac potentials, heart sounds, oximetry, etc.Number of channels available in each case depends upon the procedure to be carried out. Usually
two channels for pressure monitoring are considered to be sufficient. Various signals from the
patient are recorded on a multi-channel graphic recorder and simultaneously displayed on alarge screen monitor mounted at a height to facilitate convenient viewing from a distance. A multi-channel analog FM tape recorder is added to the system for the storage of physiological data. Thedata can be replayed at a later stage for future studies and detailed analysis.
The computerized system provides the physician with an immediate pressure waveform analysis
to allow a step-by-step assessment of the progress of catheterization. It helps to make rapiddecisions regarding the course and duration of the procedure. The heart of the system is the onlineanalysis of pressure waveform for calculations of systolic pressure, diastolic pressure, meanpressure, pressure derivative (dp/dt), etc. in the chambers of the heart, arteries or the aorta.Additional calculations like systolic ejection period, diastolic filling period, stroke volume, cardiacoutput and valve areas, etc. are also calculated. Figure 6.39 shows a system configuration employedin a typical catheterization lab.
The information is obtained online during the catheterization procedure from a selected ECG
lead, up to three pressure transducers to measure oxygen saturation and a dye dilution sensor formeasuring the cardiac output. A summary of the analyzed pressure gradients, valve areas and
cardiac output is displayed on the video monitor.
/G36/G2E/G31/G30/G2E/G31 /G50/G72/G65/G73/G73/G75/G72/G65/G20/G4D/G65/G61/G73/G75/G72/G65/G6D/G65/G6E/G74/G73
To make pressure measurements, the operator identifies the pressure sites and starts pressure
sampling. Five or eleven successive beats of the pressure signal are sampled. Although the analysisis done for three or nine beats, five or eleven beats are sampled to allow for the possibility of anincomplete first beat and for the part of the last waveform necessary to define, among other things,LV end diastolic pressure. Nine beat analysis is used for automatically rejecting PVCs and other

Patient Monitoring Systems 239
arrhythmias and for averaging out respiratory variations. Three beat analysis increases the speed
of analysis and is generally used where the rhythm is regular. This concept of ‘statistically selectedbeats’ is employed to minimize the effect of signal distortion by arrhythmias and the occasional
premature ventricular contractions elicited by the catheter position. Figure 6.40 shows a flow
chart for pressure signal processing in a catheterization laboratory.
The computer smooths noise fluctuation from the raw pressure data by use of a low-pass
Fourier filter, which also permits accurate dp/dtcalculation. The cut-off frequencies are preset to
17 Hz for ventricular pressure and 6 Hz for non-ventricular pressures. The control unit of thecomputer generates a QRS timing pulse from the electrocardiogram and this is used by the
computer for reference timing of the pressure waveforms.
The search for peak systolic pressure begins at 50 ms after the QRS pulse (Fig. 6.41). To reject the
possibility of detecting an early erroneous peak in the presence of catheter fling, the search iscontinued for 100 ms after a peak is found. Any large value found within this 100 ms is acceptedas the new peak. The maximum dp/dt is found for each beat by searching forward from the QRS
onset +50 ms to peak systolic pressure.
Another important parameter calculated is V
max, which is the isovolumetric index of the cardiac
function or contractility (Mirsky et al, 1974). This is obtained by extrapolating the downslope of
thedp/kP dt versus P curve back to zero pressure. In this expression, P is the total left ventricular
pressure, and K a stiffness constant, with a value of 30.
Beginning diastolic pressure is detected in the interval between the maximum peak systolic
pressure and a point 100 ms before the next QRS onset pulse. Maximum ( Pmax) and minimum
(Pmin) pressure are measured in that interval and the computer searches for the first point on the
downstroke where the pressure curve slope has a magnitude less than ()max minPP –
2 mmHg/s. IfRecording system
(transducers, signal
conditioners, display)
Tape recorderControl unitKeyboardVideo monitorVideo monitorCatheterization laboratory
Auto cal.Computer area
System console
(report printer)
Computer
A–D converter
Paper
tape recorder
DiscPlotterLine printerDigitizer
controlProjection
system and
keyboardCine film viewing
area
Digitizer
pen
Fig.6.39 Typical instrumentation layout in a catheterization lab

240 Handbook of Biomedical Instrumentation
this slope is not found before the minimum pressure point is reached, beginning diastolic pressure
is taken as the minimum pressure.
End diastolic pressure is located between the QRS onset-20 ms and the minimum dp/dt on the
next pressure complex. Starting from the maximum dp/dt point, a backward search is made for
end diastole, which is the first point encountered whose slope is less than ( Pmax–Pmin) mmHg/s.
Pattern recognition of direct and indirect (pulmonary artery wedge) arterial pressure involves the
timing of pressure oscillations with the electrocardiogram and the setting of maximum local values
for waves in the appropriate halves of the R-R interval.
Among the valvular parameters measured and/or calculated are outflow gradients and inflow
gradients. Gradients can be determined from either simultaneous or non-simultaneous pressureson each side of the valve. Pressures that are not sampled simultaneously are matched beat-by-beat.The pressure sites used for valve gradients are those closest to the valve.
Systolic ejection period begins at the pressure on the ventricular pressure waveform equal to the
diastolic arterial pressure. If the two do not occur simultaneously, the computer shifts the arterialAnalyse pressures
and calculate gradientsSample smooth and
store 5 or 11 beatsRaw pressure signal
Rank values for
each parameter
Determine
representative values
for each parameterResult is mean of the
(middle) three valuesDiscard three highest and
three lowest values in
nine-beat analysisDiscard first beatQRS leading edge used
for timing only
Two adjacent
pressure
sites sampled
Store and display
resultsCalculate valve areasIs
representative
gradient
significiant
YesYes
No
No
Fig.6.40 Flow chart for pressure signal processing in cath. Lab (Courtesy:
Hewlett Packard, USA)

Patient Monitoring Systems 241
DEP
BDP EDP Min.SepA BMax.
Aortic
pressure
Left atrial
pressure
Left ventricular
pressure
ECG
QRS onset pulse
BDP : Beginning diastolic pressure
EDP : End diastolic pressureDEP : Diastolic ejection period
A =
B = maxMax-Min
2
dp
dtmmHg = mean pressure
mmHg/s = maxi. slope of pressure with respect to time
Fig.6.41 Determination of ventricular pressure measurement points, systolic ejection
period and diastolic filling period (Courtesy: Hewlett Packard, USA)
pressure waveform backward relative to the ventricular pressure waveform by the required time
delay. The computer then searches for a second crossing point between both pressures. The searchstarts at a point on the arterial pressure waveform corresponding to the beginning of the ventricularsystole. The crossing point marks the end of the systolic ejection period.
The search for the diastolic filling period starts at the point detected as beginning diastolic
pressure on the ventricular waveform and proceeds backward until a crossing between theventricular and atrial or shifted pulmonary wedge pressure is found. The search for the end ofdiastolic pressure on the ventricular beats moves forward until an intersection is found. If no
crossing is found, then the end of the diastolic filling period is set at end diastolic pressure.
Some other useful calculations made during catheterization procedures are as follow:
Valve area (cm
2)= Valve flow (ml/s)
Valve gradient (mmHg)C
Where C = 40 for mitral valve and 44.5 for aortic, pulmonic and tricuspid valves.
The formula is applicable only to stenotic valves (Grolin and Grolin, 1957). All valve areas
greater than 3.5 cm2are not reported.

242 Handbook of Biomedical Instrumentation
Valve flow (mitral and tricuspid valves) = Cardiac output (ml/min)
Diastolic filling period (s /min)
Valve flow (aortic, pulmonic valves) = Cardiac output (ml/min)
Systolic ejection period (s /min)
Computerized measurement of blood flow can be made either by the dye dilution or Fick method.
Oxygen saturation readings of the blood sample withdrawn through a cuvette oximeter are
accomplished by sampling directly the output of the red and infrared cell of the instrument fourtimes per second. Logarithmic conversion in the computer permits values to read as percentsaturation of the blood with oxygen. This instrument is used for measuring cardiac output by theindicator dilution method. Dye injected into the circulation is detected by continuously samplingfrom a downstream sampling site. The clinical findings are entered into the selective retrieval fileand preparation of the final report and patient assessment are essentially complete at the con-clusion of each procedure.

/G41/G72/G72/G68/G79/G74/G68/G6D/G69/G61/G20/G61/G6E/G64/G20/G41/G6D/G62/G75/G6C/G61/G74/G6F/G72/G79
/G4D/G6F/G6E/G69/G74/G6F/G72/G69/G6E/G67/G20/G49/G6E/G73/G74/G72/G75/G6D/G65/G6E/G74/G73
/G20/G37/G2E/G31 /G43/G41/G52/G44/G49/G41/G43/G20/G41/G52/G52/G48/G59/G54/G48/G4D/G49/G41/G53
Any disturbance in the heart’s normal rhythmic contraction is called an arrhythmia. Patients
undergoing a seemingly uneventful recovery from myocardial infarction may develop cardiacarrest as a direct and immediate result of ventricular fibrillation. It is possible to treat and reversemany of these dangerous episodes if they could be detected early and an advance warning of theironset could be made available. The necessity for early detection of these harbingers of catastrophic
arrhythmias led to the establishment of coronary care units in hospitals in the early 1960s for the
intensive monitoring and treatment of patients with acute myocardial infarction. The attempt inthese units was to effectively carry out resuscitation techniques such as cardiac massage andtransthoracic defibrillation. During the last decade, attention was focussed on early detection andtreatment of arrhythmias and the emphasis of treatment switched from resuscitation to aggressiveprophylactic therapy.
Detailed and extensive examination of ECG records has shown that abnormalities in the
functioning of the heart invariably manifest themselves in ECG waveform. Diagnostic statementsas observed from the ECG records are classified into two classes: (i) morphological statements —
primarily based on ECG waveshapes that attempt to describe the state of the working musclemasses and (ii) rhythm statements —concerned with the site and rate of the cardiac pacemaker and
the propagation of impulses through the conduction system.
Sometimes, irritation occurs in the ventricles, the self-triggering impulse does not arrive through
the AV node and thus travels a different and slower path in spreading over the ventricles. The QRS
then becomes widened, and is classified as ventricular ectopic beat. Thus, an ectopic beat is a beat,which starts in an abnormal location in the heart and is often premature, therefore also called
premature ventricular contraction (PVC), i.e. it occurs sooner than the next expected beat.
Ventricular ectopic beats result in an abnormal depolarization sequence, the ECG displays anHAPTER
77

244 Handbook of Biomedical Instrumentation
abnormal QRS morphology, often with a pronounced increase in width and change in amplitude.
Figure 7.1 shows waveforms corresponding to some common types of arrhythmias.
(a)
(b)
Fig.7.1 (a) ECG waveform shows ventricular ectopic beat on second pulse
(b) Widened QRS-classified as a ventricular ectopic beat
/G20/G37/G2E/G32/G41/G52/G52/G48/G59/G54/G48/G4D/G49/G41/G20/G4D/G4F/G4E/G49/G54/G4F/G52
An arrhythmia monitor is basically a sophisticated alarm system. It is not an ECG interpretation
system. It constantly scans ECG rhythm patterns and issues alarms to events that may bepremonitory or life threatening. Arrhythmia monitors are available in various degrees of sophisti-cation, but all are ventricular oriented, detecting most of the significant ventricular arrhythmiasincluding ventricular premature beats which comprise the majority of such events. While thecomplex computerised systems are useful for multi-patient set-ups and can help detect arrhythmiasof a wide variety at graded alarm levels, the comparatively simpler instruments mostly look forwidened QRS waves and heart timing for premature beats. With the availability of low cost
PCs, desk-top arrhythmia monitors are now available at lower cost, these provide comparativemonitoring facilities which were earlier derived from large computerized systems.
The arrhythmias, which the instruments are designed to detect, are premature QRS complexes,
widened QRS complexes and runs of widened complexes. Because each patient’s ECG may differ,
the instruments generally base their determination of abnormal or ectopic beats upon a reference
obtained from the patient himself. Therefore, any arrhythmia monitoring instrument will operatein the following sequence:

Arrhythmia and Ambulatory Monitoring Instruments 245
• Stores a normal QRS for reference, particularly QRS width and R–R interval. An external
ECG recorder is automatically activated during the store normal mode so that the referenceheart beats may be visually examined and determined as to whether they are truly repre-
sentative.
• Initiates an alarm automatically, when ectopic beats are detected—either the ventricular
prematured or widened varieties.
• Gives alarm light signals whenever the prematured or widened ectopic beats exist up to
the rate of 6/min or 12/min.
• Detects and triggers alarm when artifacts are present at the source, e.g. muscle tremor due
to patient movement, base line shift and improperly connected electrodes.
Figure 7.2 shows the major signal processing tasks performed by most automated arrhythmia
monitoring and analysis systems. The function of each of these blocks is as follows:
Summary, statistics
documentation
trend plots, alarmsECG signal
Signal conditioning
QRS
detectionNoise
detectionVentricular
fibrillation
detection
Morphology
characterzationTiming
classification
Beat
labelingAtrial
fibrillation
detection
Rhythm definition
Fig.7.2 Block diagram of basic arrhythmia monitoring system

246 Handbook of Biomedical Instrumentation
Signal Conditioning: Single or multiple ECG leads may be used for arrhythmia monitoring.
Although the first generation arrhythmia systems were used on a single lead, the present daysystems analyze two or more leads simultaneously. ECG signal is amplified, filtered (0.05–100 Hz
for diagnostic purposes, 1–40 Hz for monitoring purposes) and digitized using an 8-or 12-bit
analog-to-digital convertor with a typical sampling rate of 250 Hz.
Noise Detection: Inspite of analog or digital filtering performed on the ECG signal, some unwanted
noise and artifact still remain. Baseline wander, motion artifact and muscle noise all have some
energy that overlaps the ECG signal spectrum. Using specialized signal processing techniques,
unwanted noise and artifact are minimized.
For example, electrode motion artifact is the most troublesome for arrhythmia analyzers, since
it contains considerable power in the ECG signal band. This can be detected by measuring theelectrode-skin contact impedance. Motion of the electrode with respect to the skin causes changesin the contact impedance generating electrical artifact. The impedance signal can thus provide anindependent measure of electrode motion, which can be helpful to prevent false alarm.
The ECG waveform is processed by two digital filters: a detection filter and a classification
filter. The detection filter removes low frequency noise (baseline wander) and muscle artifact. P
waves and T waves are diminished. This filter helps avoid an erroneous detection of tall T waves
as beats. Even though the shape of the QRS is distorted, the output from the detection filter is used
only for beat detection.
The classification filter removes signal irregularities, and preserves the important features of
theQRS . So, the resulting ECG output can be used for feature measurements and beat classification
(Hewlett Packard, 1999a).
QRS Detection: Arrhythmia monitors require reliable R wave detectors as a prerequisite for
subsequent analysis. The steep, large amplitude variation of the QRS complex is the obvious
characteristic to use and this is the function of the R wave detector. Any subsequent analysis is
entirely dependent upon the output from the R wave detector and, therefore, it is important that
this unit should function reliably. Most analog devices use various filtering methods to extract theQRS complex by attenuating P and T waves and artifacts. Since the maximum of the QRS energy
spectrum is in the vicinity of 10 Hz (Clynes et al., 1970), the filter is designed to have a bandwidth
of about 15 Hz with a centre frequency of 10–12 Hz. By using a bandpass filter rather than a low-
pass filter, the amplitude of low frequency noise as well as the low frequency components of theECG will be reduced without affecting the QRS .
QRS detection is now almost universally performed digitally in a two-step process. The ECG is
first preprocessed to enhance the QRS complex while suppressing noise, artifact and non- QRS
portions of the ECG. The output of the preprocessor stage is subjected to a decision rule thatconfirms detection of QRS if the processor output exceeds a threshold. The threshold may be fixed
or adaptive.
Morphology Characterization: This is based on analyzing the shape of the QRS complexes and
separating beats into groups or clusters of similar morphology. Most algorithms for real timearrhythmia analysis maintain no more than 10–20 clusters at a time, in order to limit the amountof computation needed to assign a QRS complex to a cluster.

Arrhythmia and Ambulatory Monitoring Instruments 247
Timing Classification: It involves categorization of the QRS complexes as on time, premature or
late. The observed R–R interval is compared to an estimate of the expected R-R interval. An R-R
interval will be declared premature if it is less than 85% of the predicted interval. Similarly, an
R–R interval is long if it is greater than 110% of the predicted value.
Beat Labelling: A physiologic label is assigned to each QRS complex. The possible beat labels that
can be attached by a beat classification module include the following: normal, supraventricularpremature beat, PVC, etc. This is the most complex form of the algorithm and is rarely disclosed by
the manufacturers.
Rhythm Labelling: This is the final stage in arrhythmia analysis. It is based on defined sequences
ofQRS complexes. The analysis systems are heavily oriented towards detecting ventricular
arrhythmias, particularly single PVCs. Special detectors are employed to identify rhythms such as
atrial fibrillation or ventricular fibrillation.
Atrial Fibrillation Detection: It is based on detecting abnormal rhythms from the timing sequence
ofQRS complexes.
Ventricular Fibrillation: This is usually detected by frequency domain analysis. The system is
characterized as a narrow-band, low frequency signal with energy concentrated in a band around5–6 Hz. It can be distinguished from noise (16–18 Hz) by appropriately designing bandpassfilters.
Summary Statistics: These characterize the cardiac rhythm over long time periods. These statistics
may be presented in the form of a table or graphically. Trend plots of heart rate and abnormal beatsare particularly useful to the clinician.
Alarms: These are necessary to bring to the attention of the nursing staff the serious arrhythmias
suffered by the patient with appropriate alarms.
Advanced software techniques and sophisticated algorithms are constantly being developed
for the automated analysis of arrhythmias. Recent efforts are focusing on the possible applicationof artificial intelligence methodology into the design of such systems particularly to ensure optimalalarm strategies.
/G20/G37/G2E/G33 /G51/G52/G53/G20/G44/G45/G54/G45/G43/G54/G49/G4F/G4E/G20/G54/G45/G43/G48/G4E/G49/G51/G55/G45/G53
There are several methods and computer programs in existence for the automatic detection of QRS
complexes. These include the use of digital filters, non-linear transformations, decision processorsand template matching. Generally, two or more of these techniques are used in combination in aQRS detector algorithm.
A popular approach in the detection of arrhythmias is based on template matching. A model of
the normal QRS complex, called a template, is derived from the ECG complex of a patient under
normal circumstances. This template is stored and compared with the subsequent incoming real-time ECG to look for a possible match, using a mathematical criterion. A close enough match to thetemplate represents a detected QRS complex. If a waveform does not match the available template
but is a suspected abnormal QRS complex, it is treated as a separate template, and future suspected
QRS complexes are compared with it. Obviously, the system requires considerable memory for

248 Handbook of Biomedical Instrumentation
storing the templates. Alternatively, algorithms have been developed based on digital filters to
separate out normal and abnormal QRS complexes.
/G37/G2E/G33/G2E/G31 /G53/G54/G2F/G41/G52/G20/G41/G72/G72/G68/G79/G74/G68/G6D/G69/G61/G20/G41/G6C/G67/G6F/G72/G69/G74/G68/G6D
The ST/AR (ST and Arrhythmia) algorithm from Hewlett Packard is a multi-lead ECG algorithm
designed for both arrhythmias and ST segment monitoring. The algorithm processes the ECGsignals for both paced and non-paced patients and performs several actions on the incoming ECGwaveform, including filtering the signal, detecting and classifying the QRS , generating heart rate,
identifying ectopic events and rhythms and generating alarms, when necessary.
In order to reliably detect the QRS , the detection threshold is kept as 0.15 mV to prevent the
detection of T waves or baseline noise as QRS complexes during a complete heart block or asystole.
For optimal performance and to prevent false alarms, the lead selected for monitoring should haveadequate amplitude. Figure 7.3 shows the arrangement for generating the QRS detection signal
using multiple leads. The contribution from each ECG lead to the QRS detection signal is
propotional to its measured quality based on the waveform amplitude, and the amount of muscleand baseline noise. The weighting factors are updated atleast every 200 ms to allow for quickadaptation to signal quality changes (Hewlett Packard, 1999(a)).
QRS detection
signal
generationQRS detection signal
Signal quality
measurement
Signal quality
measurementWeighting
factor
determinationChannel 2
ECG signalChannel 1
ECG signal
Fig.7.3 Generating the QRS detection signal in ST/AR system
The QRS detector checks the QRS detection signal for the presence of the peak of an R wave.
Search begins after an absolute refractory period from the previously identified QRS complex. The
value used for the refractory period is 192 ms. This helps to prevent a T wave from being identified
as an R wave.

Arrhythmia and Ambulatory Monitoring Instruments 249
After a QRS complex is identified, a search is
made on each lead independently in the area prior totheR wave to determine if there is an associated P
wave. This area is 200 ms wide and ends 120 msbefore the R wave peak (Fig. 7.4). To be accepted as a
P wave, it must be atleast 1/32 of the R wave height
and the P–R interval must be close to the average
P–R interval. P wave detection is used to differentiate
between a Sinus Rhythm and a Supraventricular (SV)
Rhythm.
After detection of the QRS , a number of features
which represent beat characteristics and whichcan be used to discriminate between different typeof beats are measured. The features measured are:
height, width, area and timing ( R–R interval).
Once the QRS is detected and measured, the beat is labelled. Labelling means that the algorithm
assigns the complex one of the following labels: normal ( N), supraventricular premature ( S),
ventricular ectopic ( V), paced ( P), questionable ( ?) and learning ( L). If the signal quality is not
good, the algorithm assigns the label “inoperative ( I)” and “artifact ( A)”.
Beat labelling is based on the use of template families to represent recurring morphologies. For
each patient, up to 16 different active template families can be created for each individual lead. Tokeep the template family information current, they are dynamically created and replaced as thepatient’s beat shape changes. When a beat is detected, it is matched against the stored waveform
templates for that patient. This process involves overlaying the beat on the template and using a
mathematical procedure to measure the differences between the two shapes. Figure 7.5 illustratesthe technique of template matching.
A separate detector continuously examines the ECG signal for ventricular fibrillation. If a flutter
or sinusoidal waveform persists for more than 4 seconds in any ECG channel, then the monitor
alarms for ventricular fibrillation.
Normally, the heart rate is computed by averaging the most recent 12 R–R intervals. This
average gives a stable estimate of the heart rate even when the rhythm is irregular. Alarms areactivated by the alarm generator wherein higher priority alarms such as asystole take precedenceand supersede lower priority alarms, such as low heart rate.
Any computerized arrhythmia algorithm would not make 100% accurate analysis of all patients.
Data bases are available for the testing of arrhythmia detection algorithms. These data basesconsist of records of patient ECG waveform, together with a set of annotation files in which eachbeat has been labelled by an expert cardiologist. One such data base is from the American Heart
Association (AHA), distributed by the Emergency Care Research Institute (ECRI), USA which
contains data from 80 patients, two of whom are paced. The performance results from the ST/ARarrhythmia algorithm (Hewlett Packard 1999b) are shown to be giving very high accuracy andreliability (above 95%), in both single lead and multi-lead arrangements.
Friesen et al (1990) compared noise sensitivity of nine types of QRS detection algorithms. He
established that an algorithm using a digital filter had the best performance for the compositenoise corrupted data.Pwave search window
Rwave
120 msec200
msec
Fig.7.4 P-wave detection technique

250 Handbook of Biomedical Instrumentation
New beat
Stored template 1 Stored template 2
Poor match Good matchECG to be analyzed
Existing
template families
Fig.7.5 Template matching technique for detection of arrhythmias

Arrhythmia and Ambulatory Monitoring Instruments 251
/G37/G2E/G33/G2E/G32/G44/G61/G74/G61/G20/G43/G6F/G6D/G70/G72/G65/G73/G73/G69/G6F/G6E/G20/G61/G6E/G64/G20/G50/G72/G6F/G63/G65/G73/G73/G69/G6E/G67/G20/G6F/G66/G20/G74/G68/G65/G20/G45/G43/G47/G20/G53/G69/G67/G6E/G61/G6C/G20/G62/G79/G20/G41/G5A/G54/G45/G43
To handle the relatively high data rate of ECG signals for real-time analysis of rhythm, it is necessary
to use some technique of data compression so that less memory is required to store the waveformsand the overall processing time is less. Cox and Nolle (1968) suggested AZTEC (Amplitude-
Zone-Time-Epoch-Coding)—a pre-processing program for real time ECG rhythm analysis at an
approxi-mate data reduction rate of 10 : 1.
Basically, AZTEC regards the ECG signal as a series of straight line segments, each segment
representing a sequence of consecutive points with approximately the same amplitude. Therefore,AZTEC’s first step is to convert the ECG signal into such segments, some of short and others oflong duration. For example, QRS complex is converted into a series of short line segments as the
signal amplitude changes rapidly. A series of lines, each containing four samples (with 500samples/s) or less, is considered to be adequately represented by a constant rate of voltage changeor slope as long as the voltage difference between adjacent lines does not change sign. The slope isterminated by a line longer than the four samples or a change in signs. The slope duration and thevoltage between the lines bounding the slope are then stored as the next pair of data words. ThusAZTEC also examines sequences of short segments moving in the same direction of amplitudeand replaces them with slope segments, thereby further compressing the data. Figure 7.6 illustratesan ECG signal and its resulting AZTEC representation, which is a caricature of the original ECG
signal. It is composed of sequences of straight line segments and sloping line segments. Further,
straight line segments are stored in the computer as voltage and time duration whereas the slopingline segments are stored as voltage variation (+ or –) and time duration. AZTEC also containsdigital logic, which is able to detect base line wander. Also, the sum of absolute values of all slopesis used for noise detection, but sudden bursts of muscle artifact or an excursion from baseline isdistinguishable from an ECG wave only if the condition persists.
ECG signal
AZTEC
representation
Fig.7.6 ECG signal and its resulting AZTEC representation (after Cox and Nolle, 1968;
by permission of IEEE Trans. Biomed. Eng.)

252 Handbook of Biomedical Instrumentation
Accurate description and delineation of ECG waveform by a complex series of scans through
the AZTEC data is the next step. The module for processing this function identifies QRS complex
andPandTwaves. It systematically determines each complex’s beginning and end, and performs
the measurement of four QRS parameters (height, duration, offset and R–R interval). This module
functions as follows:
• The module runs once every second, depending upon the quality of the signal. If the signal
is ‘chaotic’ (a signal of poor quality that cannot be processed) or ‘noisy’ (a signal of inter-mediate quality), no processing is carried out.
• The memory is then scanned to find the QRS complex until the operation is temporarily
suspended to give the artifact time to terminate in the case of a poor quality signal.
• Starting with the last QRS found and processed, R wave is looked for after a refractory
period of 120 or 200 ms. The scan looks for a slope greater than 1/3rd the height of the R
wave of the patient’s ‘normal’ QRS . If no such slope is found within the interval, the scan
is repeated with the slope criterion relaxed to 1/6th of the height. If no QRS is detected in
a distance of two times the normal ‘ R–R’ interval, less 1/4th the previous R–R interval,
then it is a case of ‘missed beat’.
• After finding the ‘ R’ wave, the QRS complex is determined by the following criteria: (a) it
must be V-shaped and must not contain (wiggles) lines and slopes < 3/8th the largestslope and have a duration longer than 32 ms, (b) the other slope of the complex must be atleast 3/8th the amplitude of the largest slope found. A scan is then performed on each sideof the complex until a `wiggle’ with a duration greater than 32 ms is found. These two
points define the limits of QRS .
• After establishing the QRS and its extent, the QRS performs the four measurements
required to compare it with the patient’s ‘normal’ QRS . These measurements refer to QRS
height, duration, offset and R–R interval. The basis for making these measurements is
shown in Fig. 7.7.
• In order to confirm that the QRS detected as per the above mentioned procedure is the right
one, the processor goes back and scans for a distance of 300 ms following this assumedQRS . It is quite probable that an abnormally big sized P wave might have been mistaken as
aR wave. If more than one R wave possibilities are observed, the one with the highest
amplitude is taken as the true R wave and the process repeated to determine the complex
limits.
•A Twave following a QRS is also searched. A wave is considered as a T wave if (i) it occurs
within a 200 ms interval, (ii) its peak occurred within 1/3rd the R–R interval +80 ms or
within 240 ms whichever is greater, following the preceding R wave, and (iii) its height is
half the height of a normal R wave.
•P waves are detected by scanning for 300 ms, preceding the QRS , and measuring the
minimum and maximum amplitudes encountered. If these amplitudes are consistent withthe patient’s ‘normal’ P wave, the wave is taken as P wave. P waves are used to determine
the sinus rhythm—(more than 50% of the beats are preceded by P waves.)
The four measurements of QRS and a fifth measurement called ‘distance D’ based on height,
duration and offset to those of the patient’s ‘normal’ QRS are compared. Beats are classified on the

Arrhythmia and Ambulatory Monitoring Instruments 253
basis of their deviation from the normal and may be definitely normal, probably normal, probably
PVC or definitely PVC. Only those beats are used for updating the ‘stored’ normal which are
definitely normal. A table is prepared on the value of D calculated on the basis of ‘normal’ and
current values. The table provides 120 possibilities, which are isolated and the result in terms ofprobability of PVCs is presented.
/G37/G2E/G33/G2E/G33 /G44/G65/G74/G65/G63/G74/G69/G6F/G6E/G20/G6F/G66/G20/G56/G65/G6E/G74/G72/G69/G63/G75/G6C/G61/G72/G20/G46/G69/G62/G72/G69/G6C/G6C/G61/G74/G69/G6F/G6E
Most of the arrhythmia monitoring systems do not always generate an alarm during ventricularfibrillation, as a QRS complex with known and established characteristics is not found to be
present. Ventricular fibrillation is characterized by a sinusoidal ECG waveform resulting from
uncoordinated heart muscle activity. Therefore, the detector must be capable of using thesinusoidal characteristics of the waveform and generate alarm under such conditions. The blockdiagram of the detector is shown in Fig. 6.13.
The detection of ventricular fibrillation is based upon two criteria: the frequency of the sinusoid
and the filter leakage fraction. The average physiologically occurring frequency filter leakagefraction is computed as follows.
The average period of the ECG waveform for one second is first determined by the following
equation:
T= 2 p
V
VVn
nn s||–Â
1 1
where T= the number of sample points in an average period
Vn= the value of the nth sampleh
h
ddQRS
heightQRS
duration
hh
hh12
1 2–
+OFS = 64QRS offsetR–R
interval R–R
interval
h2h1
Fig.7.7 Detection criterion for QRS is based upon QRS height, QRS duration, QRS
offset, QRS polarity and R-to-R interval (reproduced with permission of
Hewlett Packard, USA)

254 Handbook of Biomedical Instrumentation
The period ( T) is then used to adjust a notch filter, the notch to be positioned at the average
frequency corresponding to the average period determined above. This filtering is achieved byadding two samples over the average half period.
(Filter output)
n=Vn + Vn–T/2
The unfiltered signal and filter output are summed over a one second period to calculate the
‘filter leakage fraction’ and is given by:
Filter leakage fraction = ||
|| | |/
/VV
VVnn T
nn T s+
+-
-Â2
2 1
For an ideal sinusoidal waveform, this fraction will be zero. As the ventricular fibrillation
waveform is not an ideal sinusoid, a higher leakage fraction is used as the threshold for detectionof this condition.
/G20/G37/G2E/G34 /G45/G58/G45/G52/G43/G49/G53/G45/G20/G53/G54/G52/G45/G53/G53/G20/G54/G45/G53/G54/G49/G4E/G47
Stress test or exercise electrocardiography is used when the diagnosis of coronary arterial diseaseis suspected or to determine the physical performance characteristics of a patient. The test involvesthe recording of the electrocardiogram during dynamic or occasionally isometric exercise. Thediagnostic value of exercise testing primarily concerns either depression or elevation of the STsegment present in myocardial ischemia.
Dynamic exercise is performed by the patient who walks on a treadmill on which the speed and
elevation can be adjusted, manually or automatically, to suit a variety of graded exercise protocols.Alternatively, the patient may be asked to pedal an electrically braked bicycle ergometer. Both thetreadmill and ergometer can be used as stand-alone devices for testing physical fitness. Advancedergometers and treadmills can store and display activity data, transfer it to a PC and downloadpatient data from the PC.
/G37/G2E/G34/G2E/G31 /G54/G72/G65/G61/G64/G6D/G69/G6C/G6C/G20/G54/G65/G73/G74
There are two basic kinds of exercise protocols used in treadmill tests:
•The Balke-Ware Protocol: It uses a constant speed of 3.3 miles/hour (5.3 km/hour), with
progressive increments in the load every 2 minutes. This is achieved by increasing thegrade or incline of the motor-driven treadmill. The exercising subject therefore walks “up-hill” and his own body weight serves as the load.
•The Bruce Protocol: It uses simultaneous increments in both speed and treadmill grade at
intervals of 3 minutes.
Both the protocols are satisfactory for most clinical purposes. However, they may have to be
modified to suit the condition of the individual being subjected to exercise testing.
/G37/G2E/G34/G2E/G32/G42/G69/G63/G79/G63/G6C/G65/G20/G54/G65/G73/G74
In this test, the speed is usually kept constant and incremental resistance is applied eithermechanically (friction belt or brake pads) or electrically. The work load can be precisely measured

Arrhythmia and Ambulatory Monitoring Instruments 255
in terms of kilograms per metre per minute. In bicycle ergometry, physiological measurements are
more easily obtained with the subject comfortably seated and the torso relatively immobile.
The stress test laboratory is equipped with an exercise device, a PC-based ECG display and
recording system, a blood pressure measuring instrument (sphygmomanometer) and a defibrillator.
Sophisticated algorithms (digital filters and programs for the morphologic recognition) have
been developed in order to obtain stabilized artifact-free ECG tracings. The system displays in realtime three user selectable leads, a reference ECG complex (template) and the current mean beat onwhich the system detects stress-induced changes. Automatic measurements of ST level and slope,current and maximum heart rate are displayed in real time and can be printed as alphanumericdata or as trends. The usefulness of the exercise stress test lies in assessing the extent and severity
of the cardiovascular disease and it remains a useful tool in providing the important prognostic
information.
The exercise test has become an established tool for the diagnosis of coronary artery disease due
to changes observed in the ECG induced by exercise. The principal ECG abnormalities that arerecognised as manifestations of ischemia are ST segment depression and ST segment elevation.
The current standard of determining the ST segment measurement is by measuring the voltage
difference between the value at a point 60–80 ms after the J point and the isoelectric baseline. Theisoelectric baseline is either between the P and Q waves (the P-R interval) or in front of the P-wave
(the T-P interval) (Fig. 7.8). Investigations have revealed that the slope of the ST segment
significantly influences the abnormal ST shift. Figure 7.9 shows normal and abnormal shiftsobserved in the ST segment. The algorithm used for ST segment monitoring is based on beat
detection and classification information provided by multi-lead arrhythmia algorithm and
performs several additional actions on the ECG waveforms including filtering the signals,
measuring ST values and generating ST alarms.
R
ST value
Isoelectric
pointST
measurement
pointQ
SPTJ-point
Fig.7.8 ST segment measurement
The ST segment is a signal of low amplitude and low frequency content. Therefore, a sampling
rate of 250 samples/s is adequate. To ensure that the ST segment can be measured accurately, the

256 Handbook of Biomedical Instrumentation
incoming ECG signals must have a low-end bandwidth of 0.05 Hz. This is to ensure that no signal
distortion is introduced in the ST segment. A special ST filter with a higher low-end bandwidthof 0.67 Hz is used to further remove unwanted baseline noise. Since ST segment values do notchange very rapidly, every single beat need not be measured. Instead, reliable ST measurementcan extracted using signal averaging techniques. Generally, all QRS complexes detected by the
arrhythmia algorithm within a discrete 15 second period are saved and used for ST segmentanalysis. Statistical analysis techniques are used and the largest deviation in the value of the STsegment from the reference is selected and displayed (Hewlett Packard, 1997). Alfonso et al (1996)
compare various algorithms employed for processing stress ECG signals.
It may be observed that the exercise test when applied in an appropriate manner and interpreted
by an experienced cardiologist is a vary useful tool in the functional assessment of normal and
abnormal cardiovascular physiology, particularly its capability to provide prognostic information.
/G20/G37/G2E/G35 /G41/G4D/G42/G55/G4C/G41/G54/G4F/G52/G59/G20/G4D/G4F/G4E/G49/G54/G4F/G52/G49/G4E/G47/G20/G49/G4E/G53/G54/G52/G55/G4D/G45/G4E/G54/G53
The traditional medical examination involves a number of chemical, physical and electro-physio-
logical measurements. These measurements are of very short duration and comprise no morethan a physiological snapshot of the patient’s condition. However, when one wants to performfunctional tests on a patient, which are expected to have some relationship to his behaviour innormal life, the measurements have to be made over a long period. Ambulatory monitoringconcerns itself with the extension of such measurements into the time domain on unrestrictedambulatory (mobile) patients during everyday stress and activity as well as during periods ofsleep. Therefore, the precise objective of ambulatory monitoring is to record one or more physio-logical variables continuously or repeatedly, without interference with the spontaneous activities
of the subject by the restraints of conventional laboratory instrumentation and without influencingAbnormal
Abnormal WorseNormalST elevation
ST depression
Fig.7.9 ST segment slope

Arrhythmia and Ambulatory Monitoring Instruments 257
the variable being measured. Ambulatory monitoring is not only an invaluable aid to the physician
in the differential diagnosis of many unexplained symptoms like dizziness, syncope and palpita-tion but it also provides accurate data for the evaluation of drug therapy, stress testing, artificial
pacemakers, status of myocardial infarction and several other problems in research programmes.
The technique is so well established now that it is predicted that within the next decade,ambulatory monitoring departments will become a common feature in the hospital service,accepted as a matter of course just like the X-ray or pathology department. Ambulatory monitoringof ECG is called ‘ Holter Cardiography ’, after Dr Norman Holter who introduced this concept
in 1962.
Currently, two types of systems are available for ambulatory ECG monitoring:
(i) Tape-based or solid-state systems that provide continuous recording of the complete
24–72 hours of ECG waveforms. All of the data is available to the system and screening
expert for retrieval, review and editing.
(ii)Event, time-activated or patient-demand recorders that are not conventional, continuous
tape-recorders, but systems that can provide limited monitoring for a specific type of car-diac event, such as ST segment analysis. These recorders can capture data both and duringpreceding the symptoms when a patient activates the unit by pressing a button. Eventrecorders document and record abnormal beats but provide no analysis. In order to docu-ment the abnormal rhythm that precedes the symptomatology, a memory loop can docu-ment the abnormal event prior to activation of the system at the time of the symptoms(Handelsman, 1990). An excellent review of ambulatory cardiac event recorders is pro-vided by Benz (1999).
/G37/G2E/G35/G2E/G31 /G44/G61/G74/G61/G20/G52/G65/G63/G6F/G72/G64/G69/G6E/G67
The core of the modern ambulatory monitoring system is a multi-channel sub-miniature taperecorder running normally at a speed of 2 mm/s. At this speed, a C-120 entertainment cassette willrun for 24 h. The recorders are designed to be fitted with a variety of plug-in circuit boards adaptedfor different signals or transducers. The main areas of interest in ambulatory monitoring arecentred on the cardiovascular system, with particular reference to the control of cardiac rhythmdisturbances, along with treatment of hypertension and the diagnosis and treatment of ischaemicheart disease. Another area of interest is EEG recording, with particular reference to epilepsy.Ambulatory monitoring of respiration has also considerably developed in the last few years intoa rapidly growing area of clinical research, which is beginning to make a substantial contribution
to respiratory medicine. The main non-clinical application of ambulatory monitoring is for studies
in work physiology and environmental health and it continues to make considerable impact onthese areas.
During replay, the tape is run at 120 mm/s (60 times the recording speed) to achieve rapid
manual or automatic scanning of ambulatory records. The tape recorders used for this purposehave some special features as compared to the usually available entertainment tape recorders. Ablock diagram of a two channel recorder is shown in Fig. 7.10. The tape recorder offers single-sided 24 h recording facility, with three recording channels—two for ECG and one for timing, forthe precise correlation of recorded events to patient activity. As in conventional circuits, the ECG

258 Handbook of Biomedical Instrumentation
channels feature high input impedance, differential inputs consisting of transient protection mode
switching and differential amplifier stages, followed by a single ended amplifier, with provisionfor fast transient recovery and reset.
Calibration is automatic, as it switches the ECG channels to receive the 1 mV pulse output of a
calibration circuit. The timing channel provides duty cycle coding for odd versus even minutes aswell as half hour transitions. Timing channel information is amplitude modulated at low frequencyso that the recorder user may mark significant episodes without disturbance to either the ECGor timing information. Time coding is a function of a crystal controlled counter, which controlsbiasing to the three recording channels. Other facilities incorporated are a cassette interlock thatprevents use of battery power without a cassette in place and the automatic shut down on lowbattery.
Conventional tape and solid-state systems commonly use data compression. This process
requires digitization at least at 250 samples/s. Digitization through tape or solid-state systemsresults in the loss of the original ECG data by compression since digitization only records certain
points along the waveform. However, for practical purposes, the loss in data may not be highly
significant so long as the device retains the necessary components for accurate reproduction of thevarious waveforms necessary for clinical evaluation.
/G37/G2E/G35/G2E/G32/G44/G61/G74/G61/G20/G52/G65/G70/G6C/G61/G79/G20/G61/G6E/G64/G20/G41/G6E/G61/G6C/G79/G73/G69/G73
One difficult problem, which is inevitably faced when examining long-term recordings, is thealmost overwhelming quantity of data which becomes available. Even a single channel 24 hourMotor drivePreamplifierECG 1
ECG 2Diff. to
single
end amp.
Cal.
std 1 mVCrystal
control
counterChannel 1
output
Display
boardHR set
Min. set
Amplitude
modulator
Time
Bias
inject BiasCal./
reset
controlDeck
Motor
Diff. to
single
end amp. Channel 2
outputPreamp.
Fig.7.10 Block diagram of the recording unit for ambulatory monitoring (Courtesy:
American Optical, USA)

Arrhythmia and Ambulatory Monitoring Instruments 259
ECG recording will contain in excess of 100,000 beats. Early replay and analysis equipment relied
on visual inspection of the replayed signals in accelerated time. This can become exceedinglytedious and is subject to error. For example, many events of clinical significance are quite transitory
and may occupy less than one minute of real time, which means that they last perhaps one or two
seconds on playback and in the course of replaying, a tape which runs for about half an hour, caneasily be missed. More complex equipment now has the ability to work largely automatically andis able to recognise abnormalities and to write these out on a conventional pen recorder forsubsequent examination. The displays in the form of R–R interval histograms, have also been
found to give important diagnostic information. The analyser part in the automatic scanning ofambulatory records look for four arrhythmic conditions. These are bradycardia, tachycardia,dropped beat and premature beat. A threshold control is associated with each of these and whenthe appropriate threshold is exceeded, an alarm condition is generated.
A typical example of a modern two channel ECG analysis system incorporating precise tape
control and data processing is shown in Fig. 7.11. The most prominent operational feature of thissystem is its dual microprocessors: control and timing and acquisition and display CPUs.
Accquistion and
display
CPUArrhythmia
processor
Direct
writerinterfaceArrhythmia
algorithm
controls
Front
panel
Control and
timing
CPU
Motor
control
Tape
deckECG and
timing
preampTiming
track2 CH
audio
ampSpeakersKeyboard2
channel
A-D conv.A-D
clock
2 ECGBI-directional serial link
HI-SPD
DMA
Display
interface
Display ampXYZ
7220
scope display
Fig.7.11 Functional diagram of data replay and analysis system for ECG (Courtesy:
American Optical, USA)
The control and timing CPU has overall system control responsibility. It also handles individual
functions such as keyboard and direct writer interface, tape deck control, timing data processing,

260 Handbook of Biomedical Instrumentation
and arrhythmia count totalizing via a high speed interrupt system. The control and timing pro-
gram is primarily resident in the 8 K PROM with spill-over occupying additional RAM. The RAMresident program portion is loaded during system initialization at start-up from the data acquisi-
tion and display PROM via a bidirectional serial communications link. The acquisition and
display CPU is responsible for the scan-mode A-D conversion and storage of 72 s worth of ECGdata for each of the two channels. Digital data storage is in RAM in circular buffer fashion forready access to facilitate display generation. This CPU is also responsible for accessing the storeddata in order to produce a variety of system scan mode CRT displays, as well as ECG strip andtrend-mode report write-outs.
The arrhythmia processor provides arrhythmia detection sounds, ectopic count tabulation and
a trend report write-out. After adjustment of detection sensitivities, the arrhythmia detector outputcan be set to either sound distinctive tones for different types of detected abnormalities therebyalerting the operator to their occurrence or to automatically stop the scanner and display thedetected abnormality.
A convenient way of managing enormous ECG data would be to perform real-time rhythm
analysis and discard normal signals instead of recording the total information on magnetictape—the continuous 24-hour electrocardiogram of an ambulatory patient. Tompkins (1978&1980) describes the basics of such a system based on the use of a microcomputer. It is designedto detect an abnormal rhythm and store in its RAM the 16 s of ECG prior to the time it sensed
the abnormal event. By using a manual override switch, the patient can put the instrument into
alarm mode to capture data during symptomatic episodes. In the alarm mode, a communication isestablished with a remote host computer and the collected data is transferred before it cancontinue in the monitoring mode once again. The contact with the host computer is establishedby the patient himself by dialling a telephone. Thakor et al (1984 a) discussed the design andimplementation of a microprocessor-based portable arrhythmia monitor for ambulatory ECG,which does not store normal complexes but recognizes and triggers an alarm on significantarrhythmias.
Modern Holter systems are PC-based diagnostic instruments designed to scan 24 hour analog
Holter tapes at high speed and produce an analysis report on cardiac arrhythmia event activity.The arrhythmia detection and analysis is performed at high speed (as high as 240 ¥ real time) on
C-60 cassettes recorded on the Holter recorder. Analysis is performed for mean heart rate, minimumand maximum heart rates, premature ventricular arrhythmias, runs of three or more ventricularectopics, premature supraventricular ectopic beats, supraventricular tachycardia and ST segment
measurements. Colour coded beat identification greatly enhances the ability to discriminate
and validate the computer analysis of abnormal beats. The PC-based system offers multiplepatient data storage and retrieval capability. Holter reports can be generated on a high speed laserprinter.
In addition to recording and analysing ECG, blood pressure is another parameter which has
been studied extensively through ambulatory monitoring. The method consists of inserting asmall diameter plastic cannula into the radial or brachial artery, and a connecting tube about 75cm long, from the cannula to the transducer, which is worn in a chest harness by the patient. Therecordings are made on a miniature tape recorder. The unit uses semiconductor gauge pressure

Arrhythmia and Ambulatory Monitoring Instruments 261
transducer and the perfusion is maintained by a miniature peristaltic pump. The frequency
response of the recorder and replay unit reduce the bandwidth to 10 Hz. This is, however,satisfactory for accurate pressure measurements in peripheral arteries (Pickering and Stott, 1980).
The system has been used for recording ambulatory blood pressure for periods up to 72 hour, with
satisfactory results.
Figure 7.12 shows a non-invasive blood pressure measuring system for ambulatory subjects.
This allows the patient to pursue normal daily activities while blood pressure measurements areautomatically computed and recorded throughout the ambulatory period. Repetitive bloodpressure measurement is achieved through pre-selection of time intervals for an automatic patientcuff inflation cycle. The pre-selected time intervals may be over-ridden by means of manualswitching to activate, restart or terminate the cuff inflation cycle. The apparatus utilizes a standardpneumatic cuff with a piezo-electric microphone transducer held in place with an adhesive discfor reliable detection of Korotkoff sounds. Detected Korotkoff sounds are exhibited as systolic anddiastolic readings on a digital liquid crystal display with a resolution of 1 mmHg. Accuracy of
Fig.7.12 Attachment of electrodes for ECG and cuff for arterial blood pressure
measurement of ambulatory subjects (Courtesy: Del Mar Avionics, USA)

262 Handbook of Biomedical Instrumentation
blood pressure measurement is achieved through the use of a closed loop design for electro-
pneumatic bleed-down of the patients cuff and repetitive electronic sensing of Korotkoff sounds.Sometimes, the sounds are gated by ECG R wave signals, which are also simultaneously recorded.
This reduces artifact since Korotkoff sounds are only accepted within a limited time period
following the R wave. The instrument is powered by a rechargeable battery which will operate for
24 h or for 192 blood pressure readings with intervals of 7.5 min between readings. The instrumentensures patient safety as it automatically releases cuff pressure if inflation and deflation cyclesexceed a 2 min period. Additionally, the ECG signal controls cuff-bleed-down at a rate of 3 mm perheart beat to ensure accurate readings of systolic and diastolic pressure.

Foetal Monitoring Instruments 263
/G46/G6F/G65/G74/G61/G6C/G20/G4D/G6F/G6E/G69/G74/G6F/G72/G69/G6E/G67/G20/G49/G6E/G73/G74/G72/G75/G6D/G65/G6E/G74/G73
During the last decade, several techniques, devices and instruments have become available which
provide reasonably reliable information and data instantaneously about the foetus during itsintrauterine life or at the time of delivery. Nevertheless, the development of foetal instrumentation
has been a difficult task because of the complex nature of the problems involved. The foetus while
in the uterus is mechanically shielded from the outside world so that it can safely develop there.This means that only a limited amount of information can be obtained directly about the foetal
condition. The only information which is readily available about the foetus and which can be
picked up from the maternal abdominal wall are the electrical potential of the foetal heart activityand the foetal heart sound signals. These signals when picked up are mixed up with the corres-
ponding maternal signals. The isolation and interpretation of the mixed up foetal signals requires
expert handling. Consequently, the obstetrician is faced with the problem of having very fewparameters available on which to base a diagnosis of foetal well-being or distress. In most cases,
the condition of the foetus is assessed by studying the blood flow in the foetal heart and its heart
rate. The foetal heart rate (FHR) yields important information about the status of the foetus, andtherefore, has become a widely studied parameter in maternity cases.
Foetal heart rate monitoring in the labour ward has generally been carried out on an intermittent
basis. It has been traditional to listen to the foetal heart sounds at intervals of up to every 15
minutes. This is done by using the Pinard stethoscope. Nevertheless, this technique does notprovide a complete picture of the foetal heart beat during the majority of contractions, specially if
they are strong, nor is it feasible to listen continuously. Thus, perhaps the most valuable informa-
tion of the foetal heart rate occurring during a contraction goes unrecorded (Day et al., 1968).
Moreover, the subjective element of human variability in counting compromises the data further
and the absence of permanent foetal heart rate records renders comparison of different patients
unsatisfactory.
/G20/G38/G2E/G31 /G43/G41/G52/G44/G49/G4F/G54/G4F/G43/G4F/G47/G52/G41/G50/G48
An assessment of the condition of the foetus can be made during labour from the foetal heartaction. Simultaneously, recording beat-to-beat foetal heart rate and uterine activity provides basicHAPTER
88

264 Handbook of Biomedical Instrumentation
information for assessing the compensatory potential of the foetal circulatory system. The instru-
ment which carries out a continuous and simultaneous recording of the instantaneous foetal heart
rate and labour activity is called cardiotocograph. In addition to detecting long-term bradycardia or
tachycardia, this instrument helps in the evaluation of foetal heart rate response of the undisturbed
circulatory system and response stimulated by uterine contractions. In the undisturbed, healthyfoetus, oscillation of the FHR is normal whereas, absence of FHR oscillation is considered a signof potential foetal distress (Gentner and Winkler, 1973). Uterine contraction may or may not causea response in the FHR. To determine the prognostic significance of a response, the shape and timerelationship of the change in FHR, with respect to the contraction is usually studied.
Cardiotocographs are designed to measure and record foetal heart rate on a beat-to-beat basis
rather than on an average basis. Normally, an accuracy of measurement may be 2–3% for classi-fication of responses. Sensitivity of 20 bpm/cm of recording chart allows adequate reading of therecorded FHR. Labour activity and FHR traces are usually recorded simultaneously on the sametime scale. Chart speed of 1–2 cm/min is adequate to provide sufficient resolution of the stimulus-response relationship. In addition, foetal monitoring instrumentation should allow the user achoice of various clinically accepted monitoring methods, be simple to operate and result inminimum patient annoyance within the constraints of high quality data presentation.
The following methods are commonly employed in most of the cardiotocographic monitoring
during labour:
Method Foetal Heart Rate Uterine Contraction
Indirect 1. Abdominal foetal electrocardiogram Tocodynamometry (using
(external) 2. Foetal phonocardiogram tocotonometer to sense
3. Ultrasound techniques (narrow beam changes in uterine tension
and wide-angle transducer) transmitted to the abdominal
skin surface)
Direct Foetal ECG with scalp electrode (spiral, Intrauterine pressure measure-
(internal) clip or suction electrode attached to the ment (using a fluid-filled
presenting part of the foetus) intracervical catheter with
strain gauge transducer)
/G20/G38/G2E/G32 /G4D/G45/G54/G48/G4F/G44/G53/G20/G4F/G46/G20/G4D/G4F/G4E/G49/G54/G4F/G52/G49/G4E/G47/G20/G46/G4F/G45/G54/G41/G4C/G20/G48/G45/G41/G52/G54/G20/G52/G41/G54/G45
/G38/G2E/G32/G2E/G31 /G41/G62/G64/G6F/G6D/G69/G6E/G61/G6C/G20/G46/G6F/G65/G74/G61/G6C/G20/G45/G6C/G65/G63/G74/G72/G6F/G63/G61/G72/G64/G69/G6F/G67/G72/G61/G6D/G20/G28/G41/G46/G45/G43/G47/G29
Foetal electrocardiogram is recorded by suitably placing the electrodes on the mother’s abdomen
and recording the combined maternal and foetal ECG. Figure 8.1 shows a typical recording of
foetal ECG picked up from the abdomen. The maximum amplitude of FECG ( R wave) recorded
during pregnancy is about 100 to 300 mV. This magnitude is much smaller than in the typical adult
ECG which is about 1 mV in the standard lead connection. The amplitude is still lower in somestages of pregnancy and may not be even properly detected. Low signal amplitude places verystringent requirements on the recording of the FECG if the signal-to-noise ratio (SNR) is to be kepthigh. Hence, the usual precautions of obtaining good ECG records are more carefully observed.They include low electrode skin contact impedance, proper electrode material with low depolariza-

Foetal Monitoring Instruments 265
tion effects and placement of the electrodes at appropriate positions. The signals must be properly
shielded, the equipment properly grounded and the patient electrically isolated from the equipment.
Van Bemmel et al (1971) suggest that the best place for abdominal electrodes is when one electrode
is near the umbilicus and the other above the symphysis.
The foetal heart rate is computed from the foetal ECG by appropriately shaping the foetal QRS
wave. The foetus heart rate is approximately twice that of the normal adult ranging approximately
from 110 to 180 bpm. The main problem in processing the foetal heart signals is the poor SNR.
There are periods, particularly during birth, during which instantaneous computation of FHR isnot possible because of excessive noise. Therefore, specific signal properties are made use of toimprove SNR.
The major sources of noise in the foetal ECG signal recorded from the maternal abdomen are
(i) amplifier input noise, (ii) maternal muscle noise (EMG), (iii) fluctuations in electrode polariza-
tion potential, and (iv) maternal ECG. For practical purposes, the first three of these sources can beconsidered as random whereas the maternal ECG is a periodic noise source. The frequencyspectrum of each noise source partially overlaps that of the foetal ECG and therefore, filteringalone is not sufficient to achieve adequate noise reduction. Mains frequency noise pick-up, whichis normally a problem in physiologic recordings, is usually eliminated by the use of a notch filter.
A method for dealing with this type of noise problem in foetal ECG, first suggested by Hon and
Lee (1963) and later refined by several investigators (Van Bernmel et al. 1968), utilize the technique
of signal averaging to improve the ratio of signal to random noise. The maternal ECG componentis effectively removed from each lead recording by creating an average maternal waveform at eachof its occurrences in the recording. The maternal component can then either be subtracted directly
from the foetal ECG recording before averaging the foetal waveforms or corresponding portions of
the separated maternal component can be averaged in parallel with the selected foetal waveformsand the resulting residual maternal signal subtracted from the foetal waveform average at theend of the process. The latter technique is more efficient as fewer calculations are required. A
particularly difficult problem encountered when applying signal averaging to the foetal ECG isthe selection of a signal to trigger the averaging process. The signal must bear the same preciseMMMMMFFFFFF F
50 Vm
t 1s(a)
(b)
Fig.8.1 Abdominal recording of foetal electrocardiogram. Bandwidth (a) 0.2 – 200 Hz
(b) 15-40 Hz

266 Handbook of Biomedical Instrumentation
temporal relationship to each waveform if the average is to be coherent. Unlike the signal averaging
often employed in stimulus-response work, there is no external stimulus to trigger the start of theaveraging procedure. Thus, the ECG signal itself must be used to provide that trigger.
Figure 8.2 shows a block diagram of the abdominal FECG processing circuit for computing
foetal heart rate. After proper placement of the electrodes, the signals are amplified in a preamplifier
which provides a very high input impedance (100 M W) and a high sensitivity and good common
mode rejection ratio (up to 120 dB). The input stage should preferably be kept isolated so that anyearth leakage currents that may develop under fault conditions comply with the safety require-ments. The preamplifier is a low-noise differential amplifier that has a wide dynamic range. Asizable common-mode signal manages to pass through the input amplifier, a circumstance to beexpected whenever electrodes spaced a few centimetres apart are attached to the human body in ahospital environment. Power line hum is responsible for most of the common-mode interferingsignal. This is suppressed by a notch filter following the input amplifier.
Bandpass
filter
24–59 HzPreamp. LA
RLRA Notch
filter
60 HzLow-pass
filter
20 HzM-pluse
generator
F-pluse
generatorPulse
insertion
logicInhibit
Output to heart
rate circuit
Fig.8.2 Block diagram of the abdominal foetal electrocardiogram processing
circuit (after Courtin et al. 1977)
The signal path then splits into two channels: the maternal ECG channel or Mchannel and the
foetal or F channel. Since the frequency spectrum of the foetal ECG differs somewhat from the
maternal ECG, some initial signal separation is achieved by using the appropriate bandpassfiltering in each channel. Polarity recognition circuits in each channel accommodate signals ofeither polarity. After filtering, the M signal is assured of being the largest signal component in the
M channel, so it can be detected on the basis of peak amplitude. It is used to generate a blanking
pulse for use in the F channel and in the pulse-insertion logic circuits. The F channel has a 30 ms
pulse generator that is triggered by the foetal ECG. It is inhibited, however, by the blanking pulsefrom the M channel, so it will not generate a pulse in response to the maternal ECG signal feeding
through to the F channel.
Foetal ECG signal detected via electrodes placed on the mother’s abdomen is complex and
requires attenuation of maternal signals for obtaining FHR. Also, due to the overlapping of the
foetal ECG with the maternal ECG, about 20% to 50% of the expected pulses may be missing.
Therefore, the pulse train generated in the F channel is fed to logic circuits. These determine the
rate at which the F channel pulses occur and if the timing indicates that there should be an F pulse

Foetal Monitoring Instruments 267
at a time when one is blanked or missing, a pulse is inserted into the F channel output pulse stream.
However, the logic circuits will not insert two pulses in a row, so there is no danger that theinstrument will continue to output normal pulses when no foetal ECG is present. The logic circuits
also keep track of the maternal heart rate. If the M and F channels have exactly the same rate, they
inhibit the F channel output during the maternal P wave. This precaution is taken because
otherwise it could be possible that when no foetal ECG is detected, the F channel would respond
to the maternal P wave and generate a train of pseudo F pulses.
The substitution logic requires a delay time to establish a missing foetal trigger pulse. On the
one hand, this delay has to be longer than the maximum permissible change in heart period(14 bpm change from 50–64 bpm = 262 ms) and on the other hand, it has to be shorter than theshortest period duration (216 bpm = 285.7 ms). It is thus kept as 270 ms. The range of FHR measure-ment is limited to 40–240 bpm because of the substitution logic. Thereafter, the output of logiccircuits go to standard heart rate computing circuits.
Clinical trials have shown that the AECG technique is usually effective in the most cases except
in those rare cases where the amniotic fluid fails to provide adequate electrical coupling fromfoetus to mother. However, during labour, the uterine and abdominal wall electromyogram signalstend to obliterate the FECG signal, making FHR counting quite difficult. At present, the abdominalFECG, therefore, does not seem to offer a practical reliable means of FHR monitoring during labourand delivery. Microprocessor-based signal averager for analysis of foetal ECG in the presence of
noise has been reported by Wickham (1982).
/G38/G2E/G32/G2E/G32 /G46/G6F/G65/G74/G61/G6C/G20/G50/G68/G6F/G6E/G6F/G63/G61/G72/G64/G69/G6F/G67/G72/G61/G6D
Foetal heart sounds can be picked up from the maternal abdomen by a sensitive microphone. The
heart sounds in the form of mechanical vibrations have to pass through tissue structure and thesignals picked up are rather weak because of distance effects and the small size of foetal heartvalves. Moreover, the heart sounds are greatly disturbed by maternal movements and externalnoise. To pick up the heart sounds, it is essential that the transducer be properly placed and its
impedance carefully matched. A crystal microphone is used for picking up phono signals. The
phono transducer signals are amplified by a low noise preamplifier and fed to a bandpass filterwhich rejects all frequencies outside the 70 to 110 Hz range. The preamplifier is incorporated inthe transducer housing to minimize interference signals being picked up. Much of the randomnoise is eliminated during this process and the record on paper after this stage is called foetalphonocardiograph.
From the normal foetal heart action, generally two sounds (Fig. 8.3) are produced corresponding
to the contraction and relaxation of the heart muscles. These two bursts of heart sounds are mixedup with unwanted signals which may succeed in passing through the filters. Such situations arequite complicated and require elaborate electronic circuitry to get one pulse for each foetal heartbeat. This is achieved by using the repetitive properties of the FHR considering the highest FHRto be 210 and the lowest FHR as 50/min. A certain interval between the two heart beats, whencomputed, is stored into a memory for comparison with the following interval. The latter is onlyaccepted if it does not differ more than a certain number of beats (±7 bpm) from the stored interval.
By this method, a reasonably reliable heart rate measurement can be made.

268 Handbook of Biomedical Instrumentation
11
222 211(a)
(b)
Fig.8.3 Foetal phonocardiogram: (a) unfiltered heart sounds, (b) filtered heart sounds
To ensure this, a detected heart sound triggers a one-shot multi-vibrator that inhibits succeeding
heart sounds from reaching the following circuits for the duration of the one-shot. The circuit mustbe able to operate a 4-to-1 range (50 to 210 beats/min. or 1.2 to 0.285 s/period). This necessitatesdesigning in the ability to adjust the one-shot ontime to the heart rate. If the time between twotriggers is less than 400 ms, the duration of the blanking pulse produced is 273 ms. If it is morethan 400 ms, then the blanking pulse is extended to 346 ms.
Figure 8.4 shows a block diagram of the arrangement used for obtaining a variable pulse
duration to inhibit triggering by the second heart sound. After peak detection, the processedpulses operate a one-shot circuit which gives a fixed pulse width of 230 ms. The output of one-shot(2) triggers a variable pulse width multi-vibrator (3) which adds and gives either 43 or 116 ms timedepending on the heart rate. The pulse width at the output will be either 230 + 43 = 273 ms or 230+ 116 = 346 ms.
Adjustable
width monoPeak
detector12
45 6 73
Phono signal
input230 ms
one shotOutput
400 ms
one shot20 ms
one shotIntegrator Comparator
Fig.8.4 Block diagram of the circuit arrangement used for obtaining variable pulse
duration to inhibit triggering by second heart sound
To detect the heart frequency, the 400 ms one-shot (4) is used. If the period duration is greater
than 400 ms, the one-shot will deliver a pulse. The negative slope of this pulse is used to trigger the20 ms one-shot (5). These 20 ms pulses are integrated by the integrator (6) and the output of thisintegrator is compared with a fixed voltage -V. If the output of the integrator is more negative than-V, the output of comparator (7) will become positive. Consequently, the reference level of the one-

Foetal Monitoring Instruments 269
shot (3) is shifted to more a positive level and pulse width of this one-shot increases from 43 ms to
116 ms, giving the total pulse width at the output between 273 ms to 346 ms. The integrator (6) is
used to delay the change in the time constant and to make sure that a change of on-time takes place
only if several (3 to 4) heart beats with the longer period duration (below 150 bpm) are present.
No output pulse will occur, if the period between two pulses is less than 400 ms. The 20 ms
pulses are, therefore, not generated and the integrator discharges slowly from the negative output
voltage to a positive output voltage. If the output of the integrator (6) is less negative than -V,the
output of comparator (7) will become negative. Now the reference level of the one-shot changes in
such a way that the time varies from 116 to 43 ms, resulting in a pulse width of 273 ms.
Phonocardiography provides a basically cleaner signal than does ultrasound, thus allowing a
greater chance of detecting a smooth baseline FHR. Unfortunately, phonocardiography is more
susceptible to artefacts introduced from ambient noise, patient movement or other intra-abdominal
sounds. Thus, even with phonocardiography, the baseline FHR may have an apparent increase in
variability that may not be real. Although phonocardiography has some advantages, it has almost
become obsolete for clinical monitoring because of its tendency to pick up too much background
noise, signal loss during uterine contractions and the general difficulty of obtaining a good signal.
/G38/G2E/G32/G2E/G33 /G46/G48/G52/G20/G4D/G65/G61/G73/G75/G72/G65/G6D/G65/G6E/G74/G20/G66/G72/G6F/G6D/G20/G55/G6C/G74/G72/G61/G73/G6F/G75/G6E/G64/G20/G44/G6F/G70/G70/G6C/G65/G72/G20/G46/G6F/G65/G74/G61/G6C/G20/G53/G69/G67/G6E/G61/G6C
An important clinical instrument for obstetric applications which makes use of the Doppler shift
principle is the foetus blood flow detector. The technique is extended to derive an integrated rate of
the foetus heart from blood flow signals and to display it on a suitable display system. In obstetric
applications, the site of investigation varies from 5 to 20 cm below the surface of the abdomen
(Fielder, 1968). This depends upon the patient and the stage of pregnancy. For obstetric studies,
ultrasonic frequency of about 2 to 2.5 MHz is usually employed, whereas in the study of blood
flow in arteries and superficial blood vessels frequencies around 5–10 MHz are preferred. The
level of ultrasonic energy transmitted into the body is generally kept between 10–15 mW/cm2.
Assuming a maximum of 50% conversion efficiency, this would mean that the transducer should
be powered with an electrical energy below 30 mW/cm2.
The Doppler-shift based ultrasound foetal blood flow detectors use hand-held probes which
may be either pencil-shaped or flat and contain two piezo-electric crystals. The probe is coupled to
the patient’s skin by means of an acoustic gel. This is done to exclude any air from the interface.
The presence of air severely attenuates the ultrasound, the problem being more acute during early
pregnancy. The transmitting crystal emits ultrasound (2 – 2.5 MHz) and the back-scattered
ultrasound is detected by the receiving crystal. The back-scattered ultrasound frequency would be
unchanged if the reflecting object is stationary. If the reflecting object is moving, as would be the
foetal heart blood vessels, then the back-scattered frequency is higher as the blood cell is
approaching the probe, and lower if it is moving away from the probe. The magnitude of the
frequency shift ( Df) varies according to the following formula:
Df = (2 fou cos q )/c
where, fo is the transmitted frequency, u is the blood velocity, cos q is the cosine of the angle of the
sound beam and the object’s direction and c is the velocity of the sound wave in the tissue.

270 Handbook of Biomedical Instrumentation
Representing the frequency shift as sound is the simplest method of processing the Doppler
signal and displaying the same. Battery operated portable instruments are commercially available.Figure 8.5(a) shows a typical instrument of this type which also displays the foetal heart rate andFig. 8.5(b) shows it in use.
Blood flow detectors based on ultrasonic Doppler shift can detect foetal pulse as early as the
tenth week of pregnancy and in nearly all cases by the twelfth week. At about 20 weeks it ispossible to detect multiple pregnancies especially if two instruments are used together and thepulse rates compared. Intrauterine death of the foetus can also be diagnosed. Later after about25 weeks of pregnancy, a distinctive sound from the placenta helps to determine its locationand facilitates diagnosis of placenta praevia. Blood flow through the umbilical cord can also beheard at this stage. Such types of instruments are also useful during labour to assess the conditionof the foetus.
Fig.8.5(a) Ultrasonic blood flow detector (Courtesy: M/s Huntleigh Health Care, USA)

Foetal Monitoring Instruments 271
Ultrasonic Doppler foetal heart signal is easy to
obtain but it is difficult to process and to get consistenttrigger pulses required for instantaneous beat-to-beat rate
measurement. This is mainly because the signal usually
has a lot of fading and the level, spectra and envelopewaveform change rapidly. We can hear the signalthrough a loudspeaker with a scarce chance of failing torecognize any beat, but a simple electronic circuit mayfail to trigger from this signal. Still, the ultrasoundDoppler shift method is more practical and easy to useduring labour. It is currently the most reliable method
for detecting the FHR pattern that is interpretable.
Signal processing for FHR determination can be
based either on detecting the foetal heart valve motionor on detecting the heart wall motion. The heart valve motion detection technique is based on thedistinct ultrasound frequency shift produced by the fast opening and closing of the heart valves.The technique, however, requires that the ultrasound beam must be directed against the relatively
small heart valves involving a longer search period and frequent repositioning of the transducer.
Therefore, it is not preferred for continuous monitoring applications. Movements of the foetalheart wall are slower as compared to valve movements and, therefore, produce a smaller frequencyshift. This signal is less precise than the heart valve signal and tends to produce more jitter on theFHR trace. However, since it is much easier to obtain these signals and transducer repositioningis necessary less often, they are better suited for continuous monitoring. In order to reduce jitter onthe trace, the usual practice is to incorporate a signal smoothing circuit with an averaging time
constant over a window of approximately three heart periods. This will nevertheless result in lesserbeat-to-beat variability details than those obtained with scalp electrodes.
Improved artefact rejection is usually accomplished in the Doppler mode of operation by some
form of short-term averaging. The averaged measure may lack the variability of the beat-to-beat,but it does provide adequate detail of the baseline trend of the FHR. Tuck (1981) suggests a twosecond average rate from only those intervals which have been determined to be valid based on thecriterion that they should be within ±10% of the most previous valid interval. The interval testing
algorithm suggested by him does not accept intervals greater than 1000 ms (corresponding to rates
< 60 bpm) and less than 250 ms (rates > 240 bpm), nor interval changes greater than ±10% of thelast valid accepted interval. However, should this inacceptance continue to occur, then a newvalid interval is recognized after three successive intervals fall within ±10% of each other and sothe process continues. This averaging approach offers an improvement over the continuousaveraging techniques by allowing an optimized trade-off between FHR measurement errorand lack of response in tracking the true FHR. This approach can be easily implemented in amicroprocessor-based instrument.
Two types of ultrasonic transducers for FHR measurement are in common use. They are the
narrow beam and the wide-angle beam types. The narrow beam transducer uses a single ultrasoundtransmitter/receiver piezo-electric crystal pair. The maximum ultrasound intensity is generallykept below 25 mW/cm
2. The typical transducer diameter is 25 mm. The narrow beam transducer
Fig.8.5(b) Ultrasonic foetal
heart beat detectorin use

272 Handbook of Biomedical Instrumentation
is very sensitive and produces a good trigger signal for instantaneous heart rate determination.
However, it takes time to detect a good signal and, therefore, frequent transducer repositioning isnecessary.
The broad beam transducers are available in many configurations. The transducers comprise a
number of piezo-electric crystals mounted in such a way as to be able to detect foetal heart move-
ments over a wider area. In one arrangement, the ultrasonic transducer is arranged in the shape of
a clover-leaf so that it provides a large area of ultrasonic illumination which allows the monitoringconsiderable lateral and descending foetal motion before requiring repositioning. The transducerhousing is flexible to permit it to follow the contour of the abdomen regardless of shape changeswith contractions. The transducer has three crystals on the other side acting as transmitterswhereas the crystal placed at the centre acts as a receiver. An alternative arrangement is the arraytransducer which has one transmitter and six peripheral ceramic receiving crystals (Fig. 8.6). Thetransmitting crystal emits a 40° divergent beam so that at 10 cm from the skin surface the beamcovers an area of approximately 10 cm diameter. This construction ensures continuous recordingof the foetal heart activity without the need to reposition the transducer which is otherwisenecessitated due to normal foetal movement. The transducer has a diameter of 6 cm and can beheld in place either by a simple buckle or a stretch belt.
Fig.8.6 Multireceiver transducer
Analysis of ultrasonic Doppler signals using a speech spectrograph shows that frequency
components in the range of 100–1000 Hz tended to be more distinctly related to the foetal heartcycle than components lying outside this frequency range. The bandpass filter, therefore, enhancesthe signal/noise ratio—the noise in this context being, for example, foetal movements at lowfrequencies and maternal placental blood flow at high frequencies.

Foetal Monitoring Instruments 273
With ultrasonic Doppler signals, there remains the possibility of more than one burst in each
cardiac cycle being detected. In some instruments, this difficulty is overcome by the dead-timegenerator, which inactivates the detector for a period of 0.3 sec after an amplitude burst has been
detected. This dead time is chosen on a compromise basis: it defines a maximum heart rate (200
bpm) that can be detected, while at rates which are less than half this maximum, i.e. 100 bpm, it isconceivable to ‘double-count’ the signal. Although the total signal processing in many instrumentsgoes far in minimizing this frequency-doubling possibility, the effect remains a fundamentallimitation of using the foetal ultrasonic Doppler signal for recording heart rate.
The principle of ultrasonic Doppler-shift based FHR measuring circuit is shown in the block
diagram in Fig. 8.7. This arrangement can be used both with a wide angle beam as well as anarrow beam transducer. The transmitted signal that leaks into the receiving path serves as alocal-oscillator signal for the mixing diodes in the demodulator. The output of the demodulator isdc except in the presence of a Doppler-shift frequency. The reflected signal is some 90 to 130 dBlower in amplitude than the transmitted signal. The high overall gain in the receiving channel(+110 dB) requires special measures to minimize the effects of interference. One measure used is alow noise, low distortion oscillator for the transmitter. This reduces interference caused byoscillator harmonics beating with radio and TV signals. Other measures involve filters in the
transducer connected for attenuating high-intensity high frequency radiation that could drive the
amplifiers into a non-linear operating region. The high frequency section of the circuits issurrounded by both magnetic and electrical shields.
RF amplifier
and
demodulatorDoppler
filter
amplifierRectifierEnvelope
filter
124
Digital display
of heart rateOne shot
multivib.2 MHz
oscillatorTransmitter
crystal
Receiver
crystal
Fig.8.7 Ultrasonic Doppler-shift based FHR measuring circuit (after Courtin et al., 1977)
Depending upon the transducer used, i.e. array or narrow beam, the filter circuits can be selected
to match the Doppler-shifted frequency components. A bandpass filter centred on 265 Hz isolatesthe Doppler frequencies resulting from the movement of the heart walls. The array transducerused with this circuit gives a broad ultrasonic beam that does not require careful positioning toobtain a strong Doppler return from the relatively large heart walls.

274 Handbook of Biomedical Instrumentation
The availability of the foetal heart Doppler signal from the twelfth week of gestation onwards,
its usually good signal-to-noise ratio and the lack of maternal interference renders it practical formeasuring each and every heart beat interval.
Lauersen et al (1976) describe a system which enhances ultrasonic reflections from certain
distances from the transducer and reduce those from others. This allows the operator to ‘range in’
on the best sounding foetal heart Doppler to optimize signal clarity. The depth ranging capability
is accomplished by superimposing a digital code on the transmitted 2 MHz continuous ultrasonicwave and awaiting the return of the code. As the ultrasonic propagation time through tissue isknown, the expected time from various distances is known. By the use of a correlation techniquethe system accepts only ultrasonic reflections at the distances selected (Fig. 8.8) by the operator.The correlation technique (Tuck 1982) also enhances the selected reflection three fold (10 dB), andreduces reflections from other distances by 30 fold (30 dB). So, by selecting the depth whichproduces the loudest, clearest, foetal heart Doppler, all returns from this area are enhanced andother Doppler sources are largely excluded. Doppler signals contain a multiplicity of componentsrepresating atrial contractions, A-V valve closure, A-V valve opening and to some extent aorticand pulmonic valve motion. As the Doppler signal changes or becomes less distinct, greater
Fig.8.8 Doppler signal corresponding to different events in the cardiac cycle (after
Laursen et al., 1976)

Foetal Monitoring Instruments 275
artifactual jitter is produced and long-term variability may be obscured. Lauersen et al (1976)
report that the Doppler from the ranging auto-correlation system may be further improved byelectronically selecting the Doppler signal from the moving heart valves either going toward or
away from the transducer for processing. This provides an additional 20-fold signal clarification.
Studies have shown that ultrasonic power levels generally used in foetal monitoring produce
no chromosome damage and/or other apparent damage to the full term foetus, even after long
exposures (Abdulla et al 1971). The Doppler signal is far more complex than the foetal electro-
cardiogram. It presents no easy single point for counting as does the R wave of the FECG. Also ,the
signal components are inconsistent in amplitude, shape and presence or absence. Therefore, theelectronic circuits must accommodate rapid gross amplitude changes, appearance and disappearanceof components and must attempt to count on the same component with each heart beat. Takeuchi
and Hogaki (1977) developed a special auto-correlation processor for obtaining a reliable beat-to-
beat record from ultrasonic Doppler foetal heart signals. The basic concept of the development isto make a extremely quick real-time auto-correlation algorithm equivalent to the beat-to-beat heartrate meter, by adoptively controlling the algorithm itself according to the “present heart rate”.Figure 8.9 shows the system block diagram. The system consists of preprocessor, correlator, post-processor and system controller.
Post processorSystem
controller
RecorderD-ACorrelator A-D PreprocessorSignal input
ultrasonic
Doppler foetal
signal
Fig.8.9 Block diagram of the auto-correlation processor for FHR measurement
(adapted after Takeuchi and Hogaki, 1977)
The input signal comes from the ultrasound Doppler foetal signal detector. The signal goes
through the bandpass filter, automatic gain control circuit (AGC) and envelope detector. The
bandpass filter passes the heart valve signal and the higher frequency part of the heart wall signal.
The envelope signal is again passed through another set of bandpass filter and AGC circuit. Thisremoves dc and lower frequency components and only the useful components are put forcorreIation after A-D conversion. The correlator has a data storage of 4 bit ¥ 256 word with a sampling
rate of 200 samples/s and a 16 bit ¥ 256 word in correlation storage. The system has special
external logic control facility to its operating mode which can be used to optimize the system inreal time according to FHR, i.e. to control the data length effective for correlation computation justnear to one beat-to-beat interval. Hence, the correlation output responds actually to beat-to-beatchanges in the heart rate. The effective data length is controlled by scaler, by changing therefreshing rate of correlation storage. The correlation output is read out in cyclic mode. A simple

276 Handbook of Biomedical Instrumentation
clocked peak detector measures the location of the first fundamental peak location. The reciprocal
value of the peak distance (i.e. heart rate) is obtained as analog voltage and put to the recorder.
Most foetal monitors check for potential false FHR with a credence check. The credence check
detects beat-to-beat interval changes which exceed the likely physiologic rate of change of theFHR. In some instruments, the recorder pen may be held for a few seconds until several rhythmic
heart beats are found and, if none are found, it may lift off the paper until rhythmic beats are found.
In other systems the recorder pen may immediately lift. Because random noise or low signals mayappear rhythmic for short periods of time, they tend to produce erratic lift and splatter. Experiencehas shown that at least 90% of patients who are calm and have normal presentation can bemonitored throughout labour by a Dopper FHR system. When patients are too restless, too obeseor have an unusual presentation, foetal scalp electrocardiography is preferred over the ultrasoundDoppler method for detecting foetal heart signals.
Signal response is not always as good after the rupture of the membranes as before this stage.
On many patients, however, ultrasound monitoring can be continued after the rupture of themembranes with extremely satisfactory results. However, once the membranes have ruptured anddilatation of the cervix is 1.5–2 cms or more and if satisfactory and accurate performance with theexternal ultrasound system is not obtained, the ECG facility should then be used and monitoringcontinued with scalp electrodes.
The percentage of successful results of monitoring with the ultrasound has been found to be
very much dependent upon the initial correct positioning of the transducer and the adjustment ofthe transducer position from time to time throughout labour as and when the need arises. The
correct site is not necessarily that where the best sound is obtained, and it is necessary to try
various positions on the abdomen to obtain the most reliable ratemeter operation. If the umbilicalcord comes between the transducer and the foetal heart, it can lead to bad counting. The same canoccur if the transducer is sited over an anterior placenta due to the relatively long delay betweenthe heart action and the actual blood flow at the placental termination of the cord.
While monitoring FHR with the ultrasound method, the ratemeter and chart recorder may give
inaccurate results. Many of these are due to operating conditions like the lack of a couplingmedium between the skin and the transducer, the transducer being too loosely held in position,accidental moving of the transducer by the patient, patient sitting up or turning on the side andmovement of the foetus in the uterus. Inaccuracy due to these factors can be eliminated by thecareful placement of the transducer and by checking the response on the audio output.
/G38/G2E/G32/G2E/G34 /G46/G48/G52/G20/G4D/G65/G61/G73/G75/G72/G65/G6D/G65/G6E/G74/G20/G77/G69/G74/G68/G20/G44/G69/G72/G65/G63/G74/G20/G46/G45/G43/G47
When the membranes are not ruptured, the foetal heart activity (FECG) can be recorded by usingabdominal or trans-abdominal electrodes. Usually, a hooked electrode of nichrome wire 0.5 mm indiameter can be placed subcutaneously in the foetal buttock through a puncture of the maternalabdomen. This minimizes interferences from maternal ECG.
However, after the rupture of the membranes, it is possible to attach an electrode directly to the
foetal scalp. Such electrodes are called scalp electrodes. The use of scalp electrodes gives anexcellent record of the electrocardiographic signals (Fig. 8.10) as the signal-to-noise ratio is muchhigher than in the abdominal recordings. QRS amplitudes of several hundred (50 to 300) microvolts

Foetal Monitoring Instruments 277
are easily obtainable. Maternal ECG signals are rarely present in the direct lead connections. For
this reason, foetal monitoring with scalp electrodes is the most reliable during the birth process.
While this method has the advantage of producing very clean foetal heart frequency curves, its useis limited to only during delivery periods.
Three electrodes (Fig. 8.11) are required for the direct detection of the ECG signal. One of the
electrodes, called the scalp electrode, is attached to the presenting part of the foetus. The othersignal electrode is fixed contacting the vaginal wall. The indifferent electrode is strapped to thematernal leg. Conducting jelly is used with the leg electrode to ensure good electrical contact withthe skin. An applicator is provided for attaching the scalp electrode. Three types of electrodes havebeen used for direct FECG studies, clip, spiral and suction (Fig. 8.12). The clip is an adaptation ofa surgical skin clip applied to the foetal scalp. It causes little damage to the foetal skin or vaginaltissue. The spiral electrode consists of a double small metallic spiral wire and can be ‘screwed’into the upper layers of the skin of the presenting part. This electrode provides larger amplitudeFECG signals than the clip electrode. The suction electrode is held on the scalp by suction. It is notpreferred because it can get dislodged. Spiral electrodes are easy to apply and the problem of their
dislodgement is nearly absent. The foetal ECG signal measured in this way has similar significance
to that of the vertical leads in conventional electrocardiography, although the amplitude of thesignal is attenuated to that which would be evident post-natally. In a breech delivery, where thefoetal electrode is in fact attached to the foetal buttock, the foetal ECG is comparable with theconventional aVF lead of electrocardiography, while in the normal vertex delivery, the foetal ECGis inverted with respect to the conventional aVF. This is of practical importance, since althoughthe ECG is a biphasic signal, a knowledge of the dominant polarity of the QRS complex is useful in
achieving consistent triggering of the ratemeter. Because the FECG has severe baseline wandering,the display is often filtered to keep the FECG in view on the display. This Filtering distorts theST segment and the T wave, rendering it difficult to judge ST-T changes. These changes are late
changes and are almost always preceded by the more common signs of foetal stress. Most physicians1 sec.7124.20133-01 652
F FFF FF F F
250 V m(a)
(b)
t
Fig.8.10 Intrauterine recording with scalp electrodes of foetal ECG during delivery
with filter response (a) 12 Hz – 200 Hz, (b) 15-40 Hz

278 Handbook of Biomedical Instrumentation
feel that when the FECG is necessary, the minimal risk of using direct monitoring techniques is
well worth the benefits obtained from the monitoring information.
The circuit arrangement for processing FECG signals is standard except for having capabilities
for handling low level signals. The input signal is passed through an electrically isolated amplifierand then filtered by a mains frequency ‘notch filter’ to eliminate power line interference. Abandpass filter suppresses the maternal ECG signal, if any. Together with the notch filter, thebandpass filter has a lower cut-off frequency of 30 Hz and an upper cut-off at 45 Hz. The availablesignal from a scalp electrode is 20 mV to 3 mV. There is one disadvantage of direct ECG monitoring.
Internal monitoring may carry potential maternal risks of infection and perforation of the uterus,as well as scalp injury and infection for the foetus.
/G20/G38/G2E/G33 /G4D/G4F/G4E/G49/G54/G4F/G52/G49/G4E/G47/G20/G4C/G41/G42/G4F/G55/G52/G20/G41/G43/G54/G49/G56/G49/G54/G59
During labour, the uterus muscle starts contractions of increasing intensity in a bid to expel out
the child. The intrauterine pressure can reach values of 150 mmHg or more during the expulsion
period. However, a normal patient in spontaneous active labour will demonstrate uterine con-
Fig.8.11 Scalp electrodes and accessories (Courtesy: Sonicaid, U.K.)

Foetal Monitoring Instruments 279
tractions occurring at intervals of three to five minutes, with a duration of 30 to 70 s and peak
intensity of 50 to 75 mmHg. Each uterine contraction diminishes placental perfusion and acts asa transient stress to the foetus, which may be damaged by excessive contractility or by prolongedduration of labour. Some patients will spontaneously exhibit much lower uterine activity, in termsof intensity and frequency of contractions than others but will still show progressive cervicaldilatation and an otherwise normal progress of labour.
The labour activity can be recorded either in terms of the intra-uterine pressure measured
directly by means of a catheter or a relative indication of the labour intensity measured through anexternal transducer. A plot of the tension of the uterine wall is obtained by means of a spring
loaded displacement transducer. The transducer performs a quasi-isometric measurement of the
tension of the uterus. The transducer carries a protruding tip which is pressed to the mother’sabdomen with a light force to ensure an effective coupling. The protruding surface of the transduceris displaced as the tension in the uterus increases. This movement is converted into an electricalsignal by a strain gauge in the transducer housing. The abdominal transducer provides a reliableindication of the occurrence frequency, duration and relative intensity of the contraction.
The toco-transducers are location sensitive. They should be placed over the fundus where there
is maximum motion with the contractions. The toco-tonometer transducer cannot be used in thesame place as the foetal heart rate detector, thus the patient must have two transducers on herabdomen.Suction
electrodeSpiral
electrodeClip
electrode
Fig.8.12 Three types of electrodes for studying direct FECG

280 Handbook of Biomedical Instrumentation
To sense uterine contractions externally, it is necessary to press into the uterus through the
abdominal wall. Resistance to pressure is measured either by the motion of a spring or the forceneeded to prevent a button from moving. External strain gauges are used to measure and record
the bending of a spring. In some instruments, a crystal which changes electrical characteristics
with applied pressure is used to measure force against a plunger. This method is automatic andprovides pertinent information.
Figure 8.13 shows a block diagram of the circuit which measures labour activity externally.
The transducer output is amplified in an ac amplifier. The low frequency labour activity signal isobtained from the synchronous detector and is further amplified by a dc amplifier. The activity canbe either displayed on a meter or on a direct writing chart recorder.
SynchronousdetectorA.C.
Amp.D.C.
Amp.Oscillator
Labour activity
recorderLabour activity
indicatorTransducer
LV DT
Fig.8.13 Block diagram of of labour activity monitor (external method)
The labour-activity transducers are pressure transducers that drive circuits for obtaining an
electrical indication of pressure by conventional means. The pressure channel on the recorder
is provided with a positioning control. This is done because the baseline is affected by the staticpressure on the transducer that results from the tension on the belt holding the transducer inplace. The control permits the operator to position the baseline on the zero-level line of the recordingchart.
In external toco-tonometry, movement of the foetus may be superimposed on the labour activity
curve. Stress imposed on the foetal circulatory system by the uterine contractions, foetal movementsor other factors are seen in the response of the foetal heart to these stimuli and are studied in thecorrect time relationship.
The internal method measures intra-uterine pressure (IUP) via a fluid-filled catheter. The catheter
is inserted into the uterus through a guide after the rupture of the foetal membranes. After allowingfree flow of amniotic fluid to ensure correct placement, the distal end of the catheter is usuallyattached to a pressure transducer of the type used for cardiac studies. Changes in amniotic

Foetal Monitoring Instruments 281
pressure are easily transmitted to the gauge by the incompressible fluid in the catheter. The
pressure transducer converts the catheter pressure into an electrical signal which can be displayedon the strip chart recorder. Strain gauges, though very accurate, tend to drift up to several mmHg/h
or drift with temperature changes. Therefore, when continuous monitoring is employed, it is
necessary to set zero and calibrate the transducer frequently. The peak pressure may vary accordingto which catheter is placed in the uterus. It is necessary to flush the catheter system to avoid anyblockage and to maintain the frequency response. The major applications of IUP measurement areaccurate assessment of the pressure during contractions and measurement of tonus, both impossibleby indirect means.
Although the system is inherently capable of having great accuracy, catheter-obtained uterine
contraction data may be distorted or inaccurate. The IUP may be accurately recorded only as longas a fluid pool is sustained around the tip of the catheter and leakage is completely controlled bythe descending foetal head (Caspo, 1970). Since there is no real control of catheter placement, itmay slip into an isolated pocket and receive very high pressure, especially if there is little fluid.Also, the uterus only approximates a closed fluid chamber, pressures are not necessarily transmittedequally to all segments. Open segments tend to lose fluid and thus may generate lower pressures.One study showed that IUP varies by as much as 25% at different points in the uterus. Thus, the
physiological measurement does not approach the instrument in accuracy or reproducibility.
/G20/G38/G2E/G34 /G52/G45/G43/G4F/G52/G44/G49/G4E/G47/G20/G53/G59/G53/G54/G45/G4D
Instantaneous “beat-to-beat” rate is displayed on a calibrated linear scale or digitally displayed
with a range from 50 to 210 bpm. A two-channel chart recorder is incorporated in instrumentsused for monitoring labour activity. One channel records FHR on a calibrated chart in beats perminute (50–210 bpm) while the other channel is used for recording uterine contractions calibrated0-100 mmHg. The standard chart speed is usually 1 or 2 cm/min. Both the contraction transducerand the foetal heart transducer are held together in position using stretch belts or bandages. Therecorder usually uses thermal writing and thus avoids the possibility of running out of ink. In onesystem, the stylus is a thick film resistor. To make the operation quieter, contactless positionfeedback is provided by a capacitive transducer on the galvanometer shaft. This contactless
feedback also enhances reliability by eliminating mechanical parts that could wear out. The
galvanometer, which needs a frequency response of only 3 Hz, is positioned by a servo motorthrough a silent step-down belt drive. Recording sensitivity is 20 bpm/cm giving a basic resolutionof 1 bpm for seeing small changes in the heart rate.
The chart paper is advanced by a direct-drive stepper motor eliminating the usual gear train.
Paper speed is changed simply by switching to a different motor drive frequency, rather than byshifting gears. The paper magazine is designed to make loading the chart paper an extremely easytask.
The ability to record large amounts of relatively artefact-free data, especially after the onset of
labour in the form of foetal electrocardiogram (FECG), foetal heart rate (FHR), and uterine con-tractions has been an encouraging and a valuable step forward. However, new problems havebeen generated. It is indeed impossible to analyze even in a rudimentary manner, about 9000foetal ECG complexes and various FHR and uterine contraction patterns associated with even an

282 Handbook of Biomedical Instrumentation
hour of labour, with a large percentage being, by known criteria, normal. Despite such monitors,
intelligence has still to be provided by the clinical staff observing and interpreting the records.Transferring information in this way from a chart record to a clinically useful result relies on the
perception of patterns and summation of activity over long periods of monitoring.
The first task of recognition of patterns is most amenable to the human observer. The second,
long-term survey of several metres of chart records to deduce trends, is naturally difficult for
human interpreters. Time available for the study of the records, the problems of translating a longanalog recording into meaningful data about trends and recognition of ominous shifts in thefoetal condition are other factors contributing to difficulties in manually processing the hugevolumes of data. Advocates of some form of machine intelligence in foetal monitoring contendthat all the criticism against human unreliability of interpretation can be solved by computersurveillance of the FHR and IUP patterns. Many centres, therefore, have undertaken studiesfor examining the feasibility of data reduction techniques with the use of a digital computer.These systems collect data from foetal monitors, analyze the patterns and produce results andinterpretations. Some systems also produce contraction, dilatation and effacement time curves aswell as records of treatment and obstetrical history.
Cardiotocographs using microprocessors for working out several computations on FHR and
IUP data are commercially available. Such systems compute beat-to-beat foetal heart rate with aresolution of 1 bpm. The FHR variability data is plotted in a bar graph form on strip chart recorder
from every 256 usable heart beats. This is computed in direct FECG mode and calculated only
between contractions. Data is edited by omitting beat-to-beat changes which are greater than20 bpm. The system stores the heart rate for variability computations if a contraction is not presentand if the beat-to-beat change is not greater than 12 bpm. The uterine activity is plotted also in bargraph form every 10 minutes.
The microprocessor also computes conditions of bradycardia and gives an audio alert alarm
after 35 s. The processing circuit contains the usual elements like: PROM containing the programcontrolling the CPU, RAM for temporary storage of data by the CPU, input/output ports forcontrolling flow of information into and out of the CPU, digital-to-analog converters for generatingF signals for the recorder and mode logic for decoding signals determined by the type of trans-ducers connected to the monitor.

Biomedical Telemetry and Telemedicine 283
/G42/G69/G6F/G6D/G65/G64/G69/G63/G61/G6C/G20/G54/G65/G6C/G65/G6D/G65/G74/G72/G79/G20/G61/G6E/G64/G20/G54/G65/G6C/G65/G6D/G65/G64/G69/G63/G69/G6E/G65
/G20/G39/G2E/G31 /G57/G49/G52/G45/G4C/G45/G53/G53/G20/G54/G45/G4C/G45/G4D/G45/G54/G52/G59
Wireless telemetry permits examination of the physiological data of man or animal under normal
conditions and in natural surroundings without any discomfort or obstruction to the person oranimal under investigation. Factors influencing healthy and sick persons during the performanceof their daily tasks may thus be easily recognized and evaluated. Wireless bio-telemetry has madepossible the study of active subjects under conditions that so far prohibited measurements. It
is, therefore, an indispensable technique in situations where no cable connection is feasible.
(Gandikola, 2000).
Using wireless telemetry, physiological signals can be obtained from swimmers, riders, athletes,
pilots or manual labourers. Telemetric surveillance is most convenient during transportationwithin the hospital area as well for the continuous monitoring of patients sent to other wards orclinics for check-up or therapy.
/G39/G2E/G31/G2E/G31 /G4D/G6F/G64/G75/G6C/G61/G74/G69/G6F/G6E/G20/G53/G79/G73/G74/G65/G6D/G73
The modulation systems used in wireless telemetry for transmitting biomedical signals makes useof two modulators. This means that a comparatively lower frequency sub-carrier is employed in
addition to the VHF, which finally transmits the signal from the transmitter. The principle of
double modulation gives better interference free performance in transmission and enables thereception of low frequency biological signals. The sub-modulator can be a FM (frequencymodulation) system or a PWM (Pulse Width Modulation) system, whereas the final modulator ispractically always an FM system.
Frequency Modulation: In frequency modulation, intelligence is transmitted by varying the
instantaneous frequency in accordance with the signal to be modulated on the wave, while keepingthe amplitude of the carrier wave constant. The rate at which the instantaneous frequency variesis the modulating frequency. The magnitude to which the carrier frequency varies away from thecentre frequency is called “Frequency Deviation” and is proportional to the amplitude of theHAPTER
99

284 Handbook of Biomedical Instrumentation
modulating signal. Usually, an FM signal is
produced by controlling the frequency of anoscillator by the amplitude of the modulating
voltage. For example, the frequency of oscillation
for most oscillators depends on a particular valueof capacitance. If the modulation signal can beapplied in such a way that it changes the value ofcapacitance, then the frequency of oscillation willchange in accordance with the amplitude of themodulating signal.
Figure 9.1 shows a tuned oscillator that serves
as a frequency modulator. The diode used is avaractor diode operating in the reverse-biasedmode and, therefore, presents a depletion layercapacitance to the tank circuit. This capacitanceis a function of the reverse-biased voltage acrossthe diode and, therefore, produces an FM wave
with the modulating signal applied as shown in
the circuit diagram. This type of circuit can allowfrequency deviations of 2–5% of the carrierfrequency without serious distortion.
Pulse Width Modulation: Pulse width modula-
tion method has the advantage of being lessperceptive to distortion and noise. Figure 9.2shows a typical pulse width modulator. TransistorsQ
1 and Q2 form a free-running multi-vibrator.
Transistors Q3 and Q4provide constant current
sources for charging the timing capacitors Cl and
C2 and driving transistors Q1 and Q2. When Q1is
‘off’ and Q2 is ‘on’, capacitor C2 charges through
R1 to the amplitude of the modulating voltage em.
The other side of this capacitor is connected to thebase of transistor Q
2 and is at zero volt. When Q1
turns ‘on’ switching the circuit to the other stage,
the base voltage of Q2drops from approximately
zero to -em.Transistor Q2 will remain ‘off’ until the base voltage charges to zero volt. Since the
charging current is constant at I, the time required to charge C2 and restore the circuit to the initial
stage is:
T2 = C
Iem2◊
Similarly, the time that the circuit remains in the original stage is:
T1 = C
Iem1◊R1
R2R3R4em
C3C2C1L
D
–V1:n
Fig.9.1 Circuit diagram of a frequency
modulator using varactor diode
R1R1R5R2R2R3R3R4
Q1Q3Q4
C1C2
Q2+V
Fig.9.2 Pulse width modulator

Biomedical Telemetry and Telemedicine 285
This shows that both portions of the astable period are directly proportional to the modulating
voltage.
When a balanced differential output from an amplifier such as the ECG amplifier is applied to
the input points 1 and 2, the frequency of the astable multi-vibrator would remain constant, but thewidth of the pulse available at the collector of transistor Q
2 shall vary in accordance with the
amplitude of the input signal.
In practice, the negative edge of the square wave is varied in rhythm with the ECG signal.
Therefore, only this edge contains information of interest. The ratio P:Q(Fig. 9.3) represents the
momentary amplitude of the ECG. The amplitude or even the frequency variation of the squarewave does not have an influence on the P:Q ratio and consequently on the ECG signal. The signal
output from this modulator is fed to a normal speech transmitter, usually via an attenuator, tomake it suitable to the input level of the transmitter.
WW
TQQ P PPulse width
modulated signalInput signalPulse generated by astable multi-vibrator (symmetrical 1000 Hz)
Fig.9.3 Variation of pulse width with amplitude of the input signal
W = Pulse width as generated by the multi-vibrator.
P = Variable pulse width; variation in accordance with input signal.
Q = Off-period, which also gets varied as the pulse width P varies with the
amplitude of the input signal.
/G39/G2E/G31/G2E/G32 /G43/G68/G6F/G69/G63/G65/G20/G6F/G66/G20/G52/G61/G64/G69/G6F/G20/G43/G61/G72/G72/G69/G65/G72/G20/G46/G72/G65/G71/G75/G65/G6E/G63/G79
In every country there are regulations governing the use of only certain frequency and bandwidth
for medical telemetry. Therefore, the permission to operate a particular telemetry system needs tobe obtained from the postal department of the country concerned. The radio frequencies normallyused for medical telemetry purposes are of the order of 37, 102, 153, 159, 220 and 450 MHz. Thetransmitter is typically of 50 mW at 50 W, which can give a transmission range of about 1.5 km in
the open flat country. The range will be much less in built-up areas. In USA, two frequency bandshave been designated for short range medical telemetry work by the FCC (Federal Communications

286 Handbook of Biomedical Instrumentation
Commission). The lower frequency band of 174–216 MHz, coincides with the VHF television
broadcast band (Channels 7–13). Therefore, the output of the telemetry transmitter must be limitedto avoid interference with TV sets. Operation of telemetry units in this band does not normally
require any licence. In the higher frequency band of 450–470 MHz, greater transmitter power is
allowed, but an FCC licence has to be obtained for operating the system.
Radiowaves can travel through most non-conducting material such as air, wood, and plaster
with relative ease. However, they are hindered, blocked or reflected by most conductive materialand by concrete because of the presence of reinforced steel. Therefore, transmission may be lost orbe of poor quality when a patient with a telemetry transmitter moves in an environment with aconcrete wall or behind a structural column. Reception may also get affected by radio frequencywave effects that may result in areas of poor reception or null spots, under some conditions ofpatient location and carrier frequency. Another serious problem that is sometimes present in thetelemetry systems is the cross-talk or interference between telemetry channels. It can be minimizedby the careful selection of transmitter frequencies, by the use of a suitable antenna system and bythe equipment design.
The range of any radio system is primarily determined by transmitter output power and frequency.
However, in medical telemetry systems, factors such as receiver and antenna design may makethe power and frequency characteristics less significant. The use of a higher-powered transmitterthan is required for adequate range is preferable as it may eliminate or reduce some noise effects
due to interference from other sources.
/G39/G2E/G31/G2E/G33 /G54/G72/G61/G6E/G73/G6D/G69/G74/G74/G65/G72
Figure 9.4 shows a circuit diagram of the FM transmitter stage commonly used in medical tele-
metry. The transistor T acts in a grounded base Colpitts R.F oscillator with L1 and C1 and C2as the
tank circuit. The positive feedback to the emitter is provided from a capacitive divider in thecollector circuit formed by C
1andC2. Inductor L1 functions both as a tuning coil and a transmitting
antenna. Trim capacitor C2 is adjusted to precisely set the transmission frequency at the desired
220 K
1.2 K330 pf47 pf
5–20 pf
0.001
fm0.001 f m
T1.35 VL1C1
C2
Fig.9.4 Typical circuit diagram of a FM telemetry transmitter

Biomedical Telemetry and Telemedicine 287
point. In this case, it is within the standard FM broadcast band from 88 to 108 MHz. Frequency
modulation is achieved by variation in the operating point of the transistor, which in turn variesits collector capacitance, thus changing the resonant frequency of the tank circuit. The operating
point is changed by the sub-carrier input. Thus, the transmitter’s output consists of an RF signal,
tuned in the FM broadcast band and frequency modulated by the sub-carrier oscillator (SCO),which in turn is frequency modulated by the physiological signals of interest. It is better to use aseparate power source for the RF oscillator from other parts of the circuit to achieve stability andprevent interference between circuit functions (Beerwinkle and Burch, 1976).
/G39/G2E/G31/G2E/G34 /G54/G68/G65/G20/G52/G65/G63/G65/G69/G76/G65/G72
In most cases, the receiver can be a common broadcast receiver with a sensitivity of 1 mV. The
output of the HF unit of the receiver is fed to the sub-demodulator to extract the modulating signal.
In a FM/FM system, the sub-demodulator first converts the FM signal into an AM signal. This isfollowed by an AM detector which demodulates the newly created AM waveform. With thisarrangement, the output is linear with frequency deviation only for small frequency deviations.
Other types of detectors can be used to improve the linearity.
In the PWM/FM system, a square wave is obtained at the output of the RF unit. This square
wave is clipped to cut off all amplitude variations of the incoming square wave and the averagevalue of the normalized square wave is determined. The value thus obtained is directly propor-tional to the area which in turn is directly proportional to the pulse duration. Since the pulse
duration is directly proportional to the modulating frequency, the output signal is directly
proportional to the output voltage of the demodulator. The output voltage of the demodulator isadjusted such that it can be directly fed to a chart recorder. The receiver unit also provides signaloutputs where a magnetic tape recorder may be directly connected to store the demodulated signal.
Successful utilization of biological telemetry systems is usually dependent upon the user’s
systematic understanding of the limits of the system, both biological and electrical. The two major
areas of difficulty arising in biotelemetry occur at the system interfaces. The first is the interfacebetween the biological system and the electrical system. No amount of engineering can correct ashoddy, hasty job of instrumenting the subject. Therefore, electrodes and transducers must be put
on with great care. The other major area of difficulty is the interface between the transmitter and
receiver. It must be kept in mind that the range of operation should be limited to the primary servicearea, otherwise the fringe area reception is likely to be noisy and unacceptable. Besides this, thereare problems caused by the patient’s movements, by widely varying signal strength and because
of interference from electrical equipment and other radio systems, which need careful equipmentdesign and operating procedures.
/G20/G39/G2E/G32 /G53/G49/G4E/G47/G4C/G45/G20/G43/G48/G41/G4E/G4E/G45/G4C/G20/G54/G45/G4C/G45/G4D/G45/G54/G52/G59/G20/G53/G59/G53/G54/G45/G4D/G53
In a majority of the situations requiring monitoring of the patients by wireless telemetry, theparameter which is most commonly studied is the electrocardiogram. It is known that the display
of the ECG and cardiac rate gives sufficient information on the loading of the cardiovascular
system of the active subjects. Therefore, we shall first deal with a single channel telemetry systemsuitable for the transmission of an electrocardiogram.

288 Handbook of Biomedical Instrumentation
/G39/G2E/G32/G2E/G31 /G45/G43/G47/G20/G54/G65/G6C/G65/G6D/G65/G74/G72/G79/G20/G53/G79/G73/G74/G65/G6D
Figure 9.5 shows the block diagram of a single channel telemetry system suitable for the transmission
of an electrocardiogram. There are two main parts:
• The Telemetry Transmitter which consists of an ECG amplifier, a sub-carrier oscillator and
a UHF transmitter along with dry cell batteries.
•Telemetry Receiver consists of a high frequency unit and a demodulator, to which an elec-
trocardiograph can be connected to record, a cardioscope to display and a magnetic taperecorder to store the ECG. A heart rate meter with an alarm facility can be provided tocontinuously monitor the beat-to-beat heart rate of the subject.
Transmitter
Subcarrier
modulator
ECG
amplifierECG
electrodes
BatteryR.F. amplifier
Demodulator
Heart rate
monitor
CardioscopeGraphic
recorderTape
recorderTransmitter
antennaReceiver
antenna
Fig.9.5 Block diagram of a single channel telemetry system
For distortion-free transmission of ECG, the following requirements must be met: (Kurper et al, 1996).
• The subject should be able to carry on with his normal activities whilst carrying the instru-
ments without the slightest discomfort. He should be able to forget their presence aftersome minutes of application.
• Motion artefacts and muscle potential interference should be kept minimum.• The battery life should be long enough so that a complete experimental procedure may be
carried out.
• While monitoring paced patients for ECG through telemetry, it is necessary to reduce
pacemaker pulses. The amplitude of pacemaker pulses can be as large as 80 mV comparedto 1–2 mV, which is typical of the ECG. The ECG amplifiers in the transmitter are slew rate
(rate of change of output) limited so that the relatively narrow pacemaker pulses are re-
duced in amplitude substantially.

Biomedical Telemetry and Telemedicine 289
Some ECG telemetry systems operate in the 450–470 MHz band, which is well-suited for
transmission within a hospital and has the added advantage of having a large number of channelsavailable. The circuit details of an ECG telemetry system are described below:
Transmitter: A block diagram of the transmitter is shown in Fig. 9.6. The ECG signal, picked up by
three pre-gelled electrodes attached to the patient’s chest, is amplified and used to frequencymodulate a 1 kHz sub-carrier that in turn frequency-modulates the UHF carrier. The resultingsignal is radiated by one of the electrode leads (RL), which serves as the antenna. The input
circuitry is protected against large amplitude pulses that may result during defibrillation.
Currentcontrolledsub-carriermulti-vibrator
X 2
frequency
multiplierRegulatorOsc. bias
and
shut downElectrode
off
ECG
input
amp.
Sub-carrierfilterVoltagecontrolledcrystal osc.X 2
multiplyingamplifier
230 MHz
460 MHz6.2 volts1 KHzBattery
Patient
electrodes
RA
RLLA
Fig.9.6 Block diagram of ECG telemetry transmitter (redrawn after Larsen et al.,
1972; by permission of Hewlett Packard, USA)
The ECG input amplifier is ac coupled to the succeeding stages. The coupling capacitor not
only eliminates dc voltage that results from the contact potentials at the patient-electrode interface,
but also determines the low-frequency cut-off of the system which is usually 0.4 Hz. The
sub-carrier oscillator is a current-controlled multi-vibrator which provides ±320 Hz deviation
from the 1 kHz centre frequency for a full range (± 5 mV) ECG signal. The sub-carrier filter removesthe square-wave harmonic and results in a sinusoid for modulating the RF carrier. In the event ofone of the electrodes failing off, the frequency of the multi-vibrator shifts by about 400 Hz. This
condition when sensed in the receiver turns on an ‘Electrode inoperative’ alarm.
The carrier is generated in a crystal-controlled oscillator operating at 115 MHz. The crystal is a
fifth overtone device and is connected and operated in the series resonant mode. This is followedby two frequency doubler stages. The first stage is a class-C transistor doubler and the second is a
series connected step recovery diode doubler. With the output power around 2 mW, the system
has an operating range of 60 m within a hospital.
Receiver: The receiver uses an omnidirectional receiving antenna which is a quarter-wave
monopole, mounted vertically over the ground plane of the receiver top cover. This arrangementworks well to pick up the randomly polarized signals transmitted by moving patients.
The receiver (Fig. 9.7) comprises an RF amplifier, which provides a low noise figure, RF filtering
and image-frequency rejection. In addition to this, the RF amplifier also suppresses local oscillator

290 Handbook of Biomedical Instrumentation
radiation to –60 dBm to minimize the possibility of cross-coupling where several receivers are
used in one central station. The local oscillator employs a crystal (115 MHz) similar to theone in the transmitter and ¥4 multiplier and a tuned amplifier. The mixer uses the square law
characteristics of a FET to avoid interference problems due to third-order intermodulation. Themixer is followed by an 8-pole crystal filter that determines the receiver selectivity. This filter witha 10 kHz bandwidth provides 60 dB of rejection for signals 13 kHz from the IF centre frequency(21.82 MHz). The IF amplifier provides the requisite gain stages and operates an AGC amplifierwhich reduces the mixer gain under strong signal conditions to avoid overloading at the IF stages.The IF amplifier is followed by a discriminator, a quadrature detector. The output of the discrimi-nator is the 1 kHz sub-carrier. This output is averaged and fed back to the local oscillator forautomatic frequency control. The 1 kHz sub-carrier is demodulated to convert frequency-to-voltageto recover the original ECG waveform. The ECG is passed through a low-pass filter (Fig. 9.8)
having a cut-off frequency of 50 Hz and then given to a monitoring instrument. The 1 kHz sub-
carrier is examined to determine whether or not a satisfactory signal is being received. This is doneby establishing a window of acceptability for the sub-carrier amplitude. If the amplitude is withinthe window, then the received signal is considered valid. In the case of AM or FM interference, an’inoperative’ alarm lamp lights up.
Different manufacturers use different carrier frequencies in their telemetry equipment. Use of
the FM television band covering 174 to 185.7 MHz (VHF TV channels 7 and 8) is quite common.However, the output is limited to a maximum of 150 mV/m at a distance of 30 m to eliminate
interference with commercial television channels.
Some transmitters are also provided with special arrangements like low transmitter battery and
nurse call facility. In both these situations, a fixed frequency signal is generated, which causes adeviation of the sub-carrier and when received at the receiver, actuate appropriate circuitry forvisual indications.
Also, some telemetry systems include an out-of-range indication facility. This condition is
caused by a patient lying on the leads or the patient getting out of range from the receiver capability.For this, the RF carrier signal from the transmitter is continuously monitored. When this signallevel falls below the limit set, the alarm will turn on.QuadraturedetectorIF
amplifierRF
amplifierCrystalfilter MixerLocal
oscillator438.18 MHz
21.82 MHz
AGC460 MHz1 KHzAFCAntenna
Fig.9.7 Block diagram of high frequency section of ECG telemetry receiver (adapted
from Larsen et al , 1972; by permission Hewlett Packard, USA)

Biomedical Telemetry and Telemedicine 291
For the satisfactory operation of a radiotelemetry system, it is important to have proper orienta-
tion between the transmitting and receiving antennas. There can be orientations in which none ofthe signals radiated from the transmitting antenna are picked up by the receiving antenna. It istherefore important to have some means for indicating when signal interference or signal dropoutis occurring. Such a signal makes it possible to take steps to rectify this problem and informs theclinical staff that the information being received is noise and should be disregarded.
/G39/G2E/G32/G2E/G32 /G54/G65/G6D/G70/G65/G72/G61/G74/G75/G72/G65/G20/G54/G65/G6C/G65/G6D/G65/G74/G72/G79/G20/G53/G79/G73/G74/G65/G6D
Systems for the transmission of alternating potentials representing such parameters as ECG, EEGand EMG are relatively easy to construct. Telemetry systems which are sufficiently stable to tele-meter direct current outputs from temperature, pressure or other similar transducers continuouslyfor long periods present greater design problems. In such cases, the information is conveyed as amodulation of the mark/space ratio of a square wave. A temperature telemetry system based onthis principle is illustrated in the circuit shown in Fig. 9.9.
Temperature is sensed by a thermistor having a resistance of 100 W (at 20°C) placed in the
emitter of transistor T
1. Transistors T1 and T2form a multi-vibrator circuit timed by the thermistor,
RI+ R2and C1. R1is adjusted to give 1:1 mark/space ratio at midscale temperature (35–41°C). The
multi-vibrator produces a square wave output at about 200 Hz. Its frequency is chosen keeping inview the available bandwidth, required response time, the physical size of the multi-vibrator,timing capacitors and the characteristics of the automatic frequency control circuit of the receiver.Filter
Peak-to-peak
detector
Peak-to-peak
detectorAmp.
Comp.
Comp.+
+–
–+
–Monostable
multi-vibratorECG
filterTo ECG output
amplifier
Electrode
inoperative
light
Period
comparator
Turn-off
delayBisable
multi-vibrator
Output
hold off
System
indicationORBRange/battery
inoperative light1 kHz
Fig.9.8 Schematic diagram of ECG demodulation and ‘inoperate’ circuits in ECG
telemetry receiver (after Larsen et al. 1972; by permission of Hewlett Packard,USA)

292 Handbook of Biomedical Instrumentation
This is fed to the variable capacitance diode D2via potentiometer R3.D2is placed in the tuned
circuit of a RF oscillator constituted by T3. Transistor T3forms a conventional 102 MHz oscillator
circuit, whose frequency is stabilized against supply voltage variations by the Zener diode
D3 between its base and the collector supply potential. T4is an untuned buffer stage between
the oscillator and the aerial. The aerial is normally taped to the collar or harness carrying thetransmitter.
On the receiver side, a vertical dipole aerial is used which feeds a FM tuner, and whose output,
a 200 Hz square wave, drives the demodulator. In the demodulator, the square wave is amplified,
positive dc restored and fed to a meter where it is integrated by the mechanical inertia of the meter
movement. Alternatively, it is filtered with a simple RC filter to eliminate high ripple content andobtain a smooth record on a paper. A domestic FM tuner can be used for this purpose. Temperature
measurements in this scheme were made with a thermistor probe having a temperature coefficient
of approximately – 4% per degree centigrade. It produces a change of mark/space ratio of about20% over a temperature range of 5°C. With a span of ±3°C, the system is found to be linear. The
circuit is designed to operate on 5.4 V, 350 mAh battery which gives a continuous operation for
100 hours.
/G20/G39/G2E/G33 /G4D/G55/G4C/G54/G49/G2D/G43/G48/G41/G4E/G4E/G45/G4C/G20/G57/G49/G52/G45/G4C/G45/G53/G53/G20/G54/G45/G4C/G45/G4D/G45/G54/G52/G59/G20/G53/G59/G53/G54/G45/G4D/G53
Medical measuring problems often involve the simultaneous transmission of several parameters.For this type of application, a multi-channel telemetry system is employed. Multi-channel telemetryis particularly useful in athletic training programs as it offers the possibility of simultaneouslysurveying several physiological parameters of the person being monitored.
With appropriate preamplifiers, the multi-channel systems permit the transmission of the
following parameters simultaneously depending upon the number of channels required, ECGand heart rate, respiration rate, temperature, intravascular and intra-cardiac blood pressure.R4R6
R3 R7
R5
R1R9R10R14R13R11
R2R8D1
D2 D3
D4C2C7
C4
C5
C6C8C9
T4T3T2T1C3L1
30 cm flex
Thermistor
Fig.9.9 Circuit diagram of a temperature telemetry system (after Heal, 1974; by per-
mission of Med. & Biol. Eng.)

Biomedical Telemetry and Telemedicine 293
In multi-channel telemetry, the number of sub-carriers used are the same as the number of
signals to be transmitted. Each channel therefore has its own modulator. The RF unit—the samefor all channels—converts the mixed frequencies into the transmission band. Similarly, the receiver
unit contains the RF unit and one demodulator for each channel.
Pulse width modulation is better suited for multi-channel biotelemetry systems. Such systems
are insensitive to carrier frequency shifts and have high noise immunity. FM-FM systems for
similar use may have low power consumption and high baseline stability, but they are morecomplicated and turn out to be more expensive. They can be troubled by interference betweendifferent channels. Techniques for separation usually require expensive and complex filters andeven with these, cross-talk can still be a problem. Since the FM-FM system employs a separate sub-carrier frequency for each data channel, it generally involves a high cost. Similarly, pulse-positionamplitude modulation easily gets into synchronization difficulties caused by noise and thusresults in a loss of the information transmitted. On the other hand, advantages of pulse-durationmodulation include lower sensitivity to temperature and battery voltage changes and its adaptabilityto miniaturization due to availability of suitable integrated circuits.
For multi-channel radiotelemetry, various channels of information are combined into a single
signal. This technique is called multiplexing . There are two basic methods of multiplexing. These
are:
• Frequency–division multiplexing: The method makes use of continuous-wave sub-carrier
frequencies. The signals frequency–modulate multiple subcarrier oscillators, each being atsuch a frequency that its modulated signal does not overlap the frequency spectra of theother modulated signals. The frequency modulated signals from all channels are addedtogether through a summing amplifier to give a composite signal in which none of theparts overlap in frequency. This signal then modulates the RF carrier of the transmitter andis broadcast.
• Time–division multiplexing : In this technique, multiple signals are applied to a commuta-
tor circuit. This circuit is an electronic switch that rapidly scans the signals from different
channels. An oscillator drives the commutator circuit so that it samples each signal for an
instant of time, thereby giving a pulse train sequence corresponding to input signals. Aframe reference signal is also provided as an additional channel to make it easy to recog-nize the sequency and value of the input channels.
/G39/G2E/G33/G2E/G31 /G54/G65/G6C/G65/G6D/G65/G74/G72/G79/G20/G6F/G66/G20/G45/G43/G47/G20/G61/G6E/G64/G20/G52/G65/G73/G70/G69/G72/G61/G74/G69/G6F/G6E
An FM-FM modulated radiotelemetry transmitter (Fig. 9.10) for detecting and transmitting ECGand respiration activity simultaneously on a single carrier frequency in the FM broadcast band is
described by Beerwinkle and Burch (1976). Respiration is detected by the impedance pneumo-
graphic principle by using the same pair of electrodes that are used for the ECG. A 10 kHzsinusoidal constant current is injected through electrodes E
1 and E2 attached across the subject’s
thoracic cavity. The carrier signal is generated by a phase shift oscillator. The varying thoracicimpedance associated with respiration produces an ac voltage whose amplitude varies with achange in impedance. The amplitude varying carrier is amplified by an amplifier A
1. An amplifier
filter A3 recovers the respiration signal by using rectifiers and a double pole filter. The ECG signal,

294 Handbook of Biomedical Instrumentation
detected by electrodes E1 and E2 is amplified in A1 along with the respiratory signal. It is passed
through a low-pass Butterworth filter stage A2 which passes the ECG signal but blocks respiratory
signal. The amplified ECG signal is then summed up with the preprocessed respiration signalinA
4.
The output of A4is a composite signal which is supplied to an astable multi-vibrator which acts
as a voltage-controlled sub-carrier oscillator operating at 7350 + 550 Hz. The sensitivity of the sub-carrier modulation system is 650 Hz/mV for the ECG signal and a 40 Hz/ W change in the case of
the respiration signal when the total thoracic impedance is between 600 and 800 W. The output of
the sub-carrier oscillator is then fed to a RF oscillator for transmission. The circuit requires lessthan 185 mA from a 1.35 V mercury battery. Signals can be transmitted over distances up to 15 m for
about four weeks before replacing the battery.
/G39/G2E/G33/G2E/G32 /G4F /G62/G73/G74/G65/G74/G72/G69/G63/G61/G6C/G20/G54/G65/G6C/G65/G6D/G65/G74/G72/G79/G20/G53/G79/G73/G74/G65/G6D
There has been a great deal of interest to provide greater freedom of movement to patients during
labour while the patient is continuously monitored through a wireless link. Thus, from a centrallocation, it is possible to maintain a continuous surveillance of cardiotocogram records for severalambulatory patients. In the delivery room, telemetry reduces the encumbering instrumentationcables at the bedside. Moreover, when an emergency occurs, there is no loss of monitoring in thevital minutes needed for patient transfer.
The patient carries a small pocket-sized transmitter which is designed to pick up signals for
foetal heart rate and uterine activity. The foetal heart rate is derived from foetal ECG which is
obtained via a scalp electrode attached to the foetus after the mother’s membranes are ruptured.TransmitterSub-carrier
oscillatorA4
A2A1A3Phase shift
oscillator
10 kHz
Fig.9.10 Schematic diagram of FM-FM modulated radiotelemetry transmitter for
ECG and respiration activity simultaneously (adapted from Beerwinkle andBurch, 1976; by permission of IEEE Trans. Biomed. Eng.)

Biomedical Telemetry and Telemedicine 295
Uterine activity is measured via an intra-uterine pressure transducer. If only foetal ECG is
measured, the patient herself can indicate uterine activity or foetal movement by using a hand-held push button.
The receiver (Fig. 9.11) located away from the patient, is connected to a conventional cardio-
tocograph. If the patient exceeds the effective transmission range or the electrode has a poor
contact, it is appropriately transmitted for corrective action.
Fig.9.11 Telemetry receiving system for monitoring foetal heart rate and uterine
contractions in use (Courtesy: Hewlett Packard, USA)
The telemetry system uses FM/FM modulation, with a carrier of 450 to 470 MHz and an RF
power output of 2 mW into 50 W load measured from RL electrode to a ground plane under
transmitter. The input signal range in the input for the ECG channel is 100 mV to 1 mV with a
frequency band 1 to 40 Hz. The toco channel has a sensitivity of 40 mV/ V/mmHg, and by using a
high sensitivity transducer, it can be 5 mV/ V/mmHg. The strain gauge transducer is excited with
0.25 Vrms for 40 mV/V/mmHg at 2.4 kHz. The frequency response of this channel is 3 Hz ± 0.5 Hz.
/G39/G2E/G33/G2E/G33 /G54/G65/G6C/G65/G6D/G65/G74/G72/G79/G20/G69/G6E/G20/G4F/G70/G65/G72/G61/G74/G69/G6E/G67/G20/G52/G6F/G6F/G6D/G73
The use of telemetry in operating rooms seems to be particularly attractive as it offers a means of
achieving a high degree of patient safety from electric shock as well as elimination of the hanginginter-connecting patient leads which are necessary in direct-wired equipment. Normally, there

296 Handbook of Biomedical Instrumentation
are several parameters which are of interest while monitoring surgical patients, the most common
being ECG, blood pressure, peripheral pulse and EEG.
Basically, in a four channel system, the signal encoding is based upon frequency modulation of
the four sub-carriers centred at 2.2, 3.5, 5.0 and 7.5 kHz, respectively. The system is designed to
give a bandwidth of dc to 100 Hz at the 3 dB point and the discriminator provides a 1.0 V dc output
for a 10% shift in the sub-carrier associated with each channel. This is obtained from 2 mV peak-to-peak of ECG signal (0.05 to 100 Hz), 100 mV of EEG signal (1 to 40 Hz), 100 mmHg of arterial
blood pressure (dc to 40 Hz) and 400 mV peak-to-peak of peripheral pulse (0.1 to 40 Hz). The four
sub-carriers are summed and used to frequency-modulate a radio frequency carrier oscillator
which is tuned to a spot frequency within the commercial FM band. The transmitted signals aretuned by a FM tuner whose output is fed into a fourth-channel discriminator which separates thesub-carriers through filtering and demodulates each using a phase-locked loop. The demodulated
signals are displayed on an oscilloscope.
/G39/G2E/G33/G2E/G34 /G53/G70/G6F/G72/G74/G73/G20/G50/G68/G79/G73/G69/G6F/G6C/G6F/G67/G79/G20/G53/G74/G75/G64/G69/G65/G73/G20/G74/G68/G72/G6F/G75/G67/G68/G20/G54/G65/G6C/G65/G6D/G65/G74/G72/G79
Monitoring of pulmonary ventilation, heart rate and respiration rate is necessary for a study ofenergy expenditure during physical work, particularly for sports such as squash, handball, tennis,track, etc. For this purpose, the transmitter uses pulse duration modulation, i.e. each channel is
sampled sequentially and a pulse is generated, the width of which is proportional to the amplitudeof the corresponding signal. At the end of a frame, a synchronization gap is inserted to ensure thatthe receiving system locks correctly onto the signal.
Each channel is sampled 200 times a second. With each clock pulse, the counter advances one
step, making the gates to open sequentially. At the opening of a particular gate, the corresponding
physiological signal gets through to a comparator where it is compared with the ramp. As soon asthe ramp voltage exceeds the signal voltage, the comparator changes state. Thus, the time required
for the comparator to change state would depend upon the amplitude of the signal. The counterand gates serve as a multiplexer.
The pulse train at the output of the comparator is used to frequency-modulate the RF oscillator
in the 88–108 MHz band. The transmitter is designed to work in a range of 100 m, which can beextended by using a whip antenna. For recording ECG, the electrodes are placed at the sternum.The pulmonary ventilation and respiration rate are derived from a mass flow transducer. At thereceiving end, the system contains an FM tuner and circuitry to convert the pulse width codedsignals back to analog signals and a multi-channel pen recorder to display the physiologicalsignals. Figure 9.12 shows a three channel telemetry system for monitoring the physiological dataof a sprinter.
/G20/G39/G2E/G34 /G4D/G55/G4C/G54/G49/G2D/G50/G41/G54/G49/G45/G4E/G54/G20/G54/G45/G4C/G45/G4D/G45/G54/G52/G59
The establishment of instrumented coronary care units have resulted in substantial reduction in
the mortality rates of hospitalized patients. When a patient’s condition has stabilized within a few
days, it is necessary that he is monitored during the early stages of increased activity and exertionto determine if his heart has sufficiently recovered. This can be conveniently done by the use of

Biomedical Telemetry and Telemedicine 297
telemetry which provides a sort of intermediate stage of care that smoothens the patient’s transition
back to a normal life. It thus permits surveillance of suspected coronaries without the unnaturalconstraints of confining the patient to bed. The main advantage of a multi-patient single parametertelemetry system is that patients making satisfactory recovery can vacate the hard wiredinstrument beds in the ICU/CCU units, which provides a positive psychological effect. Thepatients regain mobility after an extended period of confinement thereby improving their muscletone and circulation. Transmitters as small as 8 ¥ 6.25 ¥ 2.25 cm in size and weighing less than
115 g, including battery are commercially available. Data from different patients is received at thenurses central station. The station may have the facility of non-fade display of received waveforms,
an ECG recorder which gets activated when the patient goes into alarm, loose lead/loss of signal
alarm. The heart rate of each patient is derived and displayed simultaneously with a digitaldisplay. Multi-patient telemetry is usually done using crystal controlled circuits, which providefrequency stability to within ±0.0015%. Codes are necessarily provided on both the transmitter aswell as on the receiver units to indicate their calibrated frequencies.
The multi-patient telemetry systems, having utility mostly with cardiac patients, should have
transmitters provided with defibrillator protection to 5000 V, 400 Watt sec. pulse. ECG waveformsshould not be seriously affected in the presence of pacemaker pulses of 2.5 ms in width and at ratesup to 150 bpm.
Fig.9.12 A three channel telemetry system to monitor the physiological data of a
sprinter

298 Handbook of Biomedical Instrumentation
/G20/G39/G2E/G35 /G49/G4D/G50/G4C/G41/G4E/G54/G41/G42/G4C/G45/G20/G54/G45/G4C/G45/G4D/G45/G54/G52/G59/G20/G53/G59/G53/G54/G45/G4D/G53
Implantable telemetry systems allow the measurement of multiple physiological variables over
long periods of time without any attachment of wires, restraint or anaesthesia to the monitoredsubjects. Above all, no sensors need to be attached even to the body surface. Most of the work in
implantable telemetry has been used exclusively in animal research. Single or multi-channel
systems have been used successfully to monitor ECG, EEG, blood pressure, blood flow, tempera-ture, etc. For a multi-channel operation, a time-multiplex system is used to handle from 3 to 10channels.
The telemetry transmitters most often have to be made as small as possible so that they do not
cause any disturbance to the subject under investigation. They are made with subminiature,passive and active components which are to be kept minimum in number to avoid difficultiesencountered in interconnection and to minimize electrical interaction between them—which maymake the circuit unstable. These difficulties are greatly reduced by using thin film hybrid circuittechnology. This can permit greater complexity in design and the designer is not constrained bythe need to reduce the component numbers at the expense of circuit reliability. Transmitters in thinfilm circuits have been made by several researchers particularly for monitoring blood flow andtemperature.
Another difficult problem faced while designing implantable transmitters is that of energy
supply. The weight and size of the batteries must be minimal and the operating life must bemaximal. The energy consumption in the circuit can be minimized by using micropower
operational amplifiers and CMOS components for multiplexing, tuning and switching operations
in multi-channel telemetry systems. Besides this, the circuit design has to be such that thetransmitter energy consumption should be minimized and in the case of implantable transmitters,it should be particularly possible to turn the transmitter ‘on’ only when required. This is usuallydone by having a magnetically operated switch which allows the transmitter to be turned ‘off’externally during idle periods.
/G39/G2E/G35/G2E/G31 /G49/G6D/G70/G6C/G61/G6E/G74/G61 /G62/G6C/G65/G20/G54/G65/G6C/G65/G6D/G65/G74/G72/G79/G20/G53/G79/G73/G74/G65/G6D/G20/G66/G6F/G72/G20/G42/G6C/G6F/G6F/G64/G20/G50/G72/G65/G73/G73/G75/G72/G65/G20/G61/G6E/G64/G20/G42/G6C/G6F/G6F/G64/G20/G46/G6C/G6F/G77
In animal research, it is often necessary to obtain information about the blood flow over a period of
several months. This requirement is best met by the use of implantable flowmeters. Electromagneticflowmeters are not suitable for implant purposes, since they consume a lot of power and give riseto baseline shift due to a variety of reasons. Ultrasonic Doppler shift principle is the most widelyused technique for implantable blood flowmeter.
In this method, blood velocity information is converted to an electrical signal by means of two
ultrasonic transducers which are mounted in a rigid cuff surrounding the vessel. One of thetransducers is driven by a high frequency power source and the second receives the scatteredenergy with a shifted frequency. Figure 9.13 shows the block diagram of the flowmeter in whichthe implantable portion is shown within the dashed box. High frequency power for the flowtransducer is generated by the 6 MHz oscillator. The 6 MHz AM receiver converts the incomingultrasonic signal to an audio frequency signal by synchronous detection. Data recovery isaccomplished by an internal 100 MHz, FM transmitter and an external commercial FM receiver. A

Biomedical Telemetry and Telemedicine 299
demodulator, external to the body, converts the Doppler shift frequency to a flow estimate.
Basically, the demodulator measures the zero-crossing rate of the Doppler signal which isproportional to the blood velocity.
Rader et al (1973) describe a miniature totally implantable FM/FM telemetry system to
simultaneously measure blood pressure and blood flow. Pressure is detected by a miniatureintravascular transducer by placing it directly in the blood stream. It measures 6.5 mm in diameterand is 1 mm thick. Four semiconductor strain gauges connected in a conventional four-arm bridgeare bonded to the inner surface of the small pressure sensing diaphragm. The sensor produces
approximately 30 mV/300 mmhg. The flow is sensed and measured by an extravascular inter-
ferometric ultrasonic technique.
Barbaro and Macellari (1979) explain the construction of a radiosonde for the measurement of
intracranial pressures. The main sources of error, consisting of the thermal drift of the electroniccomponents and particularly of the pressure transducer are eliminated by simultaneouslytransmitting information regarding temperature, so that the data can be corrected accordingly. Astrain gauge pressure transducer is implanted epidurally and is connected in a flexible manner tothe body of the radiosonde. A thermistor is placed next to the pressure transducer. The sub-carrieroscillator is essentially an astable multi-vibrator operating at 4 kHz, and whose time constant isdetermined alternately by the two transducers. The switching of the transducers is done at 100kHz. The sub-carrier oscillator then amplitude-modulates a radio frequency carrier of 1050 kHz,which can be conveniently received by the common commercial receivers. The tuning coil of thecarrier oscillator acts as the transmitting aerial. The radiosonde is powered from outside throughelectromagnetic coupling and therefore contains circuits for converting the RF power into dc
voltage. The total circuit is enclosed in a case of polypropylene, which is tolerated well by the6 MHz AM
receiver6 MHz
oscillator100 MHz FM
transmitterPower supply RF switch
Recorder Demodulator100 MHz FM
receiverInternal
External
Fig.9.13 Block diagram of an implantable blood flowmeter based on ultrasonic
Doppler shift principle (after Dipietro and Meindl, 1973)

300 Handbook of Biomedical Instrumentation
human tissue. A film of silicone rubber covering the case further improves this capability. Cheng
et al (1975) used a Pitran transducer for telemetering intracranial pressures.
/G20/G39/G2E/G36 /G54/G52/G41/G4E/G53/G4D/G49/G53/G53/G49/G4F/G4E/G20/G4F/G46/G20/G41/G4E/G41/G4C/G4F/G47/G20/G50/G48/G59/G53/G49/G4F/G4C/G4F/G47/G49/G43/G41/G4C/G20/G53/G49/G47/G4E/G41/G4C/G53
/G4F/G56/G45/G52/G20/G54/G45/G4C/G45/G50/G48/G4F/G4E/G45
Telephony provides another convenient method of sending physiological signals over telephone
lines for remote processing. The method has the advantage that individual patients can bemanaged in remote areas. By sending ECG and other signals over the telephone lines, a patientcan communicate with the doctor or specialist from his home while lying in bed. Another necessityfor such a transmission is to use telephone lines for the collection of data for a central computer,from anaesthetized patients undergoing surgery in operating theatres and from conscious patientsin intensive care or recovery rooms, for maintenance of records for future reference.
Telephony deals, normally, with the transmission of human speech from one place to a second
distant place. Human speech consists of a large number of frequency components of differentvalues from about 100 to 4000 Hz having different amplitude and phase relations between them.On the other hand, most of the bioelectric signals and other physiological signals like ECG,
respiration, temperature and blood pressure consist of frequency components, which are much
below the audio band width permitted by telephone-line systems. Some system of frequency orpulse code modulation has to be employed for the transmission of such signals. However, most ofthe work reported in literature is based on the use of the frequency modulation technique.
A telephone telemetry technique for transmitting and receiving medical signals is shown in
Fig. 9.14. For frequency modulation, a modulator is used with a centre frequency of 1500 Hz. Thisfrequency is modulated ±200 Hz for a 1 V peak-to-peak signal. This deviation is linear within 1%range. The demodulator consists of an audio amplifier, a carrier rejection filter and a low-passintegrator output circuit to recover the input signal. Both at the transmitting as well as receivingends, coupling or isolation transformers are used to match the standard telephone line impedance.
DemodulatorAudio
amplifierLow-pass
integratorFrom telephone
linesPreamp.Frequency
modulationTo telephone
lines
Fig.9.14 Arrangement for transmission of analog signals over telephone lines
This technique makes use of a wired electrical connection of the amplified signal to the
telephone transmission system. This is largely due to the technical difficulties involved with
carbon microphones normally used as transmitters in telephone handsets.

Biomedical Telemetry and Telemedicine 301
Acoustic coupling both at the transmitting as well as receiving ends is desirable in case the
telephone transmission system is to be more widely employed. This however, necessitates the useof a superior type of carbon microphones. Transmission based on acoustic coupling has been
successfully demonstrated with the development of an ECG telephone transmitter for the remote
monitoring of potential cardiac patients, pacemaker studies and routine but short-term rhythmsampling. After proper amplification and filtration of the ECG signal, it is given to a voltagecontrolled oscillator (VCO). The centre frequency of the oscillator is set at 2000 Hz with a deviationof ±250 Hz. The output of the VCO is given to a dynamic earpiece, which provides sufficient audiosignal for transmission. Coupling of the earpiece to a standard telephone handset is accomplishedby clamping the two using a foam rubber gasket to ward off extraneous noise. The receiver is astandard data set, which reconstructs the waveform using zero-crossing detection technique.Even though the oscillator output is a square wave, it is satisfactory for single channel FMtelephony, since the upper harmonics are attenuated by the telephone line bandpass.
Real-time trans-telephonic ECG transmitters do not store, but can transmit in real time a
patient’s ECG to a remote receiver. These are called ‘event recorders’. They provide physicianswith ECGs from patients where such information facilitates timely therapeutic decision making.ECG transmissions are typically made using continuous frequency-modulated (FM) audible tones
emitted from a speaker built into the event recorder. This tone passes through the microphone on
a telephone handset, over telephone lines, to the remote receiver. For example, when a patient callsa physician’s office to transmit the ECG, he only needs to press a SEND button and hold thespeaker on the event recorder close to the mouthpiece on the telephone handset. This type ofacoustic transmission is compatible with telephone systems worldwide. The transmissionnormally takes place at a speed equivalent to real time, i.e. an ECG recorded over five minutes takesfive minutes to transmit. Receivers are located at a physician’s office, hospital or monitoringservice and are used to receive and print-out the ECG.
Modern event recorders (Benz, 1999) use modems, which transmit the ECG digitally, requiring
computerized receiving capability and a direct connection to the telephone system. However, thisdigital transmission is not necessarily compatible with all telephone systems throughout theworld.
/G39/G2E/G36/G2E/G31 /G4D/G75/G6C/G74/G69/G2D/G63/G68/G61/G6E/G6E/G65/G6C/G20/G50/G61/G74/G69/G65/G6E/G74/G20/G4D/G6F/G6E/G69/G74/G6F/G72/G69/G6E/G67/G20/G54/G65/G6C/G65/G70/G68/G6F/G6E/G65/G20/G54/G65/G6C/G65/G6D/G65/G74/G72/G79/G20/G53/G79/G73/G74/G65/G6D
Single and multi-channel systems for the transmission of electrocardiograms have been widelyemployed as remote diagnostic aids for cardiac patients recovering at home and for pacemakerperformance follow up. There is however, an increasing need for multi-channel parametermonitoring, especially the simultaneous transmission of ECG, blood pressure, respiration andalso temperature. Rezazadeh and Evans (1988), developed a remote vital signs monitor using adial-up telephone line. A frequency modulation system using 750 Hz, 1750 Hz and 2750 Hz wasemployed. Figure 9.15 shows the block diagram of the transmitter.
The physiological signals after amplification to a nominal 1 volt peak-to-peak amplitude are
input to an adder circuit. This adds an appropriate dc level to the signals prior to their connectionto the corresponding channel of the voltage controlled oscillator (VCO). After multiplexing, the signals
generated by the VCO, a low-pass filter cutting off at 3500 Hz is used to prevent out of band signals

302 Handbook of Biomedical Instrumentation
entering the line. A 1:1 600 W isolating transformer is required to link the system with the telephone
line. The transmitter uses the frequency division multiplexing (FDM) technique to accommodate
the 3-channels of data within the 3100 Hz available bandwidths of the telephone lines.
The three physiological signals frequency modulate three different VCOs. The VCO centre fre-
quencies are 750, 1750 and 2750 Hz. This frequency spacing ensures that the spectral components
of one channel do not directly overlap signals on either of the other modulated channels. The VCOused was the Intersil 8038, which can generate three output waveforms: square, triangular andsinewave. The sinewave is preferred for use in a multi-channel system where cross-talk is to beminimized. The bandwidths occupied by each channel was calculated as 600 Hz, leaving 400 Hzas the guardband between the channels to prevent inter-channel cross-talk.
On the receiver side (Fig. 9.16), the multiplexed signals are filtered after being terminated by the
isolating transformer. A second order low-pass filter with a roll-off frequency set at 3 kHz toDC
DC
DCChannel 1
Channel 2
Channel 3Multi-
plexerVCO 1
750 Hz
VCO 2
1750 Hz
VCO 3
2750 HzNotch
filter
To the telephone
line
Power = 5 mW
(7 dBm)
Fig.9.15 Block diagram of the three channel telephone transmitter (after Rezazadeh
and Evans, 1988)
N/filter
1750 and
2750 Hz
N/filter
750 and
2750 Hz
N/filter
750 and
1750 HzPLL 1
750 Hz
PLL 2
1750 Hz
PLL 3
2750 HzLPF 1
LPF 2
LPF 3Channel 1
Channel 2
Channel 3LPF
To the telephone
line
Isolating
transformer
Fig.9.16 Block diagram of the three channel telephone receiver (after Rezazadeh
and Evans, 1988)

Biomedical Telemetry and Telemedicine 303
prevent high frequency line noise is used. The main component of each receiver channel is a
phase-locked loop (PLL) used as a frequency demodulator. The PLLs are set to lock at frequenciesof 750, 1750 and 2750 Hz for the first, second and third channels respectively. The Signetics NE
565, was selected as the PLL. After low-pass filtering the multiplexed signals, filters are used as
band-reject notch filters, followed by PLL detectors in each channel. Finally, low-pass filtering ofeach output channel is needed to remove carrier ripple noise.
The channels were found to be identical in response to within ± 1 dB, cross-talk better than
–45 dB and total harmonic distortion. The system was used over a 40 kms distance on normaltelephone lines.
/G20/G39/G2E/G37 /G54/G45/G4C/G45/G4D/G45/G44/G49/G43/G49/G4E/G45
Telemedicine is the application of telecommunications and computer technology to deliverhealth care from one location to another. In other words, telemedicine involves the use of moderninformation technology to deliver timely health services to those in need by the electronictransmission of the necessary expertize and information among geographically dispersed parties,including physicians and patients, to result in improved patient care and management, resourcedistribution efficiency and potentially cost effectiveness (Bashshur, 1995).
Advanced information technology and improved information infrastructure the world over
have made telemedicine an increasingly viable health care service delivery alternative, measured
in clinical, technical and economic terms. However, most existing telemedicine programs, at
present, are operating in an investigational settings. Issues such as telemedicine technologymanagement and other barriers such as professional, legal and financial are still under debate.
From a technology stand point, the telemedicine technology includes hardware, software,
medical equipment and communications link. The technology infrastructure is a telecommunicationnetwork with input and output devices at each connected location.
/G39/G2E/G37/G2E/G31 /G54/G65/G6C/G65/G6D/G65/G64/G69/G63/G69/G6E/G65/G20/G41/G70/G70/G6C/G69/G63/G61/G74/G69/G6F/G6E/G73
Although telemedicine can potentially affect all medical specialities, the greatest currentapplications are found in radiology, pathology, cardiology and medical education.
Teleradiology: Radiological images such as X-ray, CT or MRI images can be transferred from one
location to another location for expert interpretation and consultation. The process involves imageacquisition and digitization.
Telepathology: To obtain an expert opinion on the microscopic images of pathology slides and
biopsy reports from specialists.
Telecardiology: Telecardiology relates to the transmission of ECG, echocardiography, colour
Doppler, etc.
In addition, telemedicine is being advantageously used for:
Tele-education: Delivery of medical education programmes to the physicians and the paramedics
located at smaller towns who are professionally isolated from major medical centres.

304 Handbook of Biomedical Instrumentation
Teleconsultation: Specialist doctors can be consulted either by a patient directly or by the local
medical staff through telemedicine technology. In the latter case, the patient is substituted by his/her electronic patient record (EPR) which has complete information on the physical and clinical
aspects of the patient.
Depending on the level of interaction required, the telecommunication infrastructure requirement
also varies: from a normal telephone, low-bit rate image transmission, real-time video transfer to
video conferencing.
/G39/G2E/G37/G2E/G32 /G54/G65/G6C/G65/G6D/G65/G64/G69/G63/G69/G6E/G65/G20/G43/G6F/G6E/G63/G65/G70/G74/G73
Store and Forward concept involves compilation and storing of information relating to audio,
video images and clips, ECG, etc. The stored information in the digital form is sent to the expert
for review, interpretation and advice at his/her convenience. The expert’s opinion can be trans-
mitted back without any immediate compulsion on the consultant’s time during his/her busyprofessional schedule.
Real Time telemedicine involves real-time exchange of information between the two centres
simultaneously and communicating interactively. It may include video conferencing, interviewingand examining the patients, transmission of images of various anatomic sites, auscultation of the
heart and lung sounds and a continuous review of various images.
/G39/G2E/G37/G2E/G33 /G45/G73/G73/G65/G6E/G74/G69/G61/G6C/G20/G50/G61/G72/G61/G6D/G65/G74/G65/G72/G73/G20/G66/G6F/G72/G20/G54/G65/G6C/G65/G6D/G65/G64/G69/G63/G69/G6E/G65
Telemedicine systems are based on multi-media computing, which not only support live multi-
way conversations between physicians, patients and specialists but can also facilitate off-lineconsultations among health-care team members. It is however, advisable to create a detailedelectronic patient record so that necessary information can be accessed, when desired. Thefollowing components relating to a patient are considered essential from the point of view of
telemedicine:
•Primary Patient Data: Name, age, occupation, sex, address, telephone number, registration
number, etc.
•Patient History : Personal and family history and diagnostic reports.
•Clinical information : Signs and symptoms are interpretations of data obtained from direct
and indirect patient observations. Direct observations include data obtained from the
senses (sight, sound, touch, smell, etc.) and through mental and physical interaction with
the patient, while indirect observations include data obtained from diagnostic instru-ments such as temperature, pulse rate, blood pressure.
•Investigations: Complete analysis reports of haemotology and biochemistry tests, stool and
urine examination.
•Data and Reports : Radiographs, MRI, CT, ultrasound and nuclear medicine images and
reports; pathology slides, electrocardiogram, spirogram.
In addition, there is a need to have video-conferencing facility for online consultations.
Figure 9.17 shows the principle and various sub-systems used in a telemedicine set-up.

Biomedical Telemetry and Telemedicine 305 Fig.9.17 Block diagram of a typical telemedicine system
DocumentsPatient
Online ECG X’RAY StoredECGECG
Pathological reportCT scanMRI PatientDATA12 Lead ECGmechine
Printer
Scanner(Transmitting centre)
[ECG and image display]Videocamera
DATAbasevideo clips moving imagesvideoconferencingaudiomessageSatellite
dish/pot/internet/ISDNDish/ISDN/pot/internet
Satellite
ArchivingAnalysistool
Applicationsoftware
Printer Report
generationServer at
receiving centre
[ECG medical imageand DATA display]
Transmitting centreReceiving centre

306 Handbook of Biomedical Instrumentation
/G39/G2E/G37/G2E/G34 /G54/G65/G6C/G65/G6D/G65/G64/G69/G63/G69/G6E/G65/G20/G54/G65/G63/G68/G6E/G6F/G6C/G6F/G67/G79
Transmission of Medical Images: One of the most important aspect of telemedicine is the
acquisition and transmission of medical images such as X-rays, CT, MRI, histopathology slides,
etc. These images are first required to be converted into digital form. The usual types of diagnosticimages used in telemedicine include:
• Images stored on traditional film or print media (e.g. X-ray film) and converted into digital
format by direct imaging or scanning in a raster sequence under controlled lighting condi-tions. CCD (charge coupled devices) and laser-based scanners are commercially available
for digitizing the film recorded X-ray images. A typical 11-by-17 inch chest film requires
atleast 2000 by 2000 pixels and an optical dynamic range of at least 4000 to 1 (12-bits) torepresent the image adequately.
• Computer-generated images (e.g. ultrasound, CT) available in standard video format
(NTSC), computer format (SVGA), or computer-file format (TIF). In modern digital radiog-raphy systems, the X-ray image is stored in the computer in the digital format. Being afilmless system, it does not require any further digitization.
The American College of Radiologists and the National Electrical Manufacturers Association
jointly developed a 13-part Digital Imaging and Communications in Medicine (DICOM) standard(ACR-NEMA, 1993) for the interconnection of medical imaging devices, particularly for
radiological imaging equipment such as digital radiography, CT (Computed Tomography), MRI
(Magnetic Resonance Imaging) and PACS (Picture Archiving and Communication Systems). Inaddition, ACR, 1994 is another standard, which deals with the transmission of radiological imagesfrom one location to another for the purposes of interpretation and consultation. The standardincludes equipment guidelines for digitization of matrix images digitized in arrarys of 0.5 k ¥ 0.5
k¥ 8-bits and large matrix images digitized in arrarys of 2 k ¥ 2 k ¥ 12-bits.
At the transmitting end, there is usually a requirement for the local storage of image data,
particularly when the storage and forward concept is adopted in the telemedicine system. Thenumber of images that may be stored depends upon the size of the storage facility (Hard disk) andthe amount of data compression applied to the images before storage. The storage requirements forvarious imaging modalities are given in Table 9.1.
Transmission of Video Images :Telemedicine applications generally require video and individual
still-frame images for interactive visual communication and medical diagnosis. NationalTelevision System Committee (NTSC) adopted, in 1953, the analog signal format used in USA forthe broadcast and cable transmission of television. The NTSC format consists of 30 image frames/s. Each frame is an interlaced raster scan of 512 horizontal lines sized to obtain a 4:3 aspect ratio.Vertical resolution is limited by the number of scan lines and horizontal resolution is limited bythe specified bandwidth of the signal—4.2 MHz. An alternative expression of resolution is the
term television lines (TVL), the number of alternating light and dark bars an observer can resolve
in the vertical dimension or along 75% of the horizontal dimension. NTSC luminance resolutionis 336 TVL.
At the resolution and frame rate associated with NTSC video, digital data are produced at a rate
of over 100 Mbps (Mega bits per seconds). Obviously, communication and storage limitations and

Biomedical Telemetry and Telemedicine 307
costs necessitate that some form of compression be used to reduce this data rate. The compression
and decompression of the signal is usually carried out by a device known as codec (coder/decoder).This device identifies and removes both spatial and temporal redundancies and reconstructs thevideo in such a way that the missing data are not readily perceived. Codec normally employed intelemedicine applications operate at data rates between 336 kbps (kilo bits per second) and 1536kbps.
Images are obtained from direct visualization by a video camera and a lens system for direct
observation (e.g. a skin lesion) or an optical adapter to a conventional scope (e.g. laparoscope,
microscope, otoscope) that provides magnification or remote access using fiber optics. The most
commonly used digital camera is based on the use of charge coupled device (CCD). Incident lightexposes an arrary of discrete light-sensitive regions called pixels, which accumulate electric chargein proportion to light intensity. The CCD sensor in a camera with a single sensor (single-chipcamera) contains a microscopic array of red, green and blue filters covering the pixels, while acamera with three CCD sensors (3-chip camera) contains large prisms and filters to form separatered, green and blue images. After a specific integration time, the accumulated charges in the pixelsare sequentially converted into conventional analog intensity and colour signals. Single CCDcameras offer 450 TVL (Television Lines) or higher resolution while high quality 3-chip CCDcameras offer 700 TVL or higher resolution.
Video is captured one frame at a time typically in a 640 by 480 pixel format, with an intensity
scale typically consisting of eight bits for monochrome and 24 bits for colour (8-bit each for red,green and blue). Other formats exist for high-resolution cameras; a 1024 by 1024 pixel format at8,10 or 12 bits per pixel can be achieved with a capture time, which depends in part on the time
needed to integrate a sufficient number of photons from dark areas of the image.
Transmission of Digital Audio: Audio channels are usually provided for diagnostic instruments
such as an electronic stethoscope or Doppler ultrasound. To reproduce heart and lung soundsaccurately, an electronic stethoscope must have a uniform frequency response from 20 Hz to/G95/G20/G54/G61/G62/G6C/G65/G20/G39/G2E/G31 Storage Requirement for Imaging Techniques (adapted from K.G. Baxter et al ,
1991)
Modality Image Size Dynamic Range Average Average storage
(pixels) (bits) number requirement
of images per exam
per exam (M Bytes)
Computed tomography 512 ¥ 512 12 30 15.0
Magnetic resonance imaging 256 ¥ 256 12 50 6.5
Digital subtraction angiography 1,000 ¥ 1,000 8 20 20.0
Digital Fluorography 1,000 ¥ 1,000 8 15 15.0
Ultrasound imaging 512 ¥ 512 6 36 9.0
Nuclear medicine 128 ¥ 128 8 26 0.4
Computed radiography 2,000 ¥ 2,000 10 4 32.0
Digitized film 4,000 ¥ 4,000 12 4 128.0

308 Handbook of Biomedical Instrumentation
2 kHz, while Doppler ultrasound requires a uniform frequency response from 100 Hz to 10 kHz.
Audio used for conversation and medical diagnosis in a telemedicine system must be digitizedand compressed before it can be combined with digital video and other information. Typical
audio compression algorithms, operate at data rates from 16 Kbps to 64 Kbps. Medical diagnostic
applications which require fidelity at higher audio frequencies will require higher data rates; 120Kbps is sufficient to reproduce the full auditory frequency spectrum from 20 Hz to 20 kHz over adynamic range of 90 dB which is adequate for the normal physiologic hearing range.
/G39/G2E/G37/G2E/G35 /G56/G69/G64/G65/G6F/G20/G43/G6F/G6E/G66/G65/G72/G65/G6E/G63/G69/G6E/G67
One of the essential components in a telemedicine system is the video conferencing facility, whichpermits real-time transmission of both audio and video information. A number of internationally
recognized standards have been established by TSS (Telecommunication Standardization Sector,
formerly Consultative Committee on International Telephone and Telegraph, CCITT) to assure ahigh degree of functional compatibility between like equipment supplied by different manufac-turers and also to standardize the interface protocols which access and control the communi-cations network.
Telemedicine is facing a number of key issues that must be addressed before it can realize its full
potential. First, standards must be set, ranging from a common language to the standardization ofcomputer architecture, so that health care providers can share data. Another key issue is licensing.Many states or countries may not allow a consultation via telemedicine unless the consultingphysician is licensed in that state or country. Medical reimbursement of expenses for telemedicineis also an important factor, which would have a major influence on the growth of telemedicine.
/G39/G2E/G37/G2E/G36 /G44/G69/G67/G69/G74/G61/G6C/G20/G43/G6F/G6D/G6D/G75/G6E/G69/G63/G61/G74/G69/G6F/G6E/G20/G53/G79/G73/G74/G65/G6D/G73
Telemedicine primarily demands a continuous and reliable communication link for the exchangeof information. There are various digital communication services available today for this purpose.
POTS: Using a modem (modulator/demodulator), the analog telephone systems (Plain Old
Telephone Service-POTS) digital signals at data rates up to about 30 kbps can be transmitted.However, depending upon the quality of the circuit, the maximum reliable data transfer rate maybe less than half of this rate. This service is adequate for data file transfers of still images.
DDS: Telephone traffic between major exchanges is today carried in digital format through the
digital data system (DDS). Switched-56 is the most common type of DDS service which provides adata rate of 56 kbps. To obtain increased data communication capacity a number of Switched-56lines can be combined in a single channel. Six Switched-56 lines give 336 kbps, which providerealistic reproduction of detail and motion when used for applications such as video conferencingand distance learning.
The digital data system of AT&T is based on the conversion of an analog voice signal into a
digital equivalent at a data rate of 64 kbps. This basic channel rate represents the bandwidthsneeded to encode telephone quality audio using the pulse code modulation encoding method.Channels are combined into larger units by time division multiplexing of digital voice data. If

Biomedical Telemetry and Telemedicine 309
carried on conventional copper wire cables, the digital signal is denoted by T representing a
transmission rate of 1536 kbps. T-1 and fractional T-1 service is available from service providers.
ISDN : Today, the all digital integrated services digital network (ISDN) is available for the
transmission of voice and data. The common form of ISDN, BRI (Basic Rate Interface) consists oftwo 64 kbps data (bearer)channels and 16 kbps data control channel (2B+D) multiplexed on twowire pairs. The data channels can be combined into 128 kbps channels; for example, a codec usedfor desktop video conferencing. ISDN dialling and other control functions are handled in the
D-channel. In some locations, ISDN, PRI (Primary Rate Interface) is available and provides
23 B-channels at 1472 kpbs and one D-channel at 64 kbps. Not all telephone ISDN services are thesame and depend upon the user’s equipment and the type of ISDN switch used by the serviceprovider.
ATM: ATM (Asynchronous Transfer Mode) is a high capacity communication link between
widely dispersed sites, usually connected by fibre communication channels such as OC-3(155 Mbps) or OC-12 (622 Mbps). This network service is well suited for transmitting digitalvideo and audio.
Table 9.2 shows the important features for various types of digital communication services.
/G95/G20/G54/G61/G62/G6C/G65/G20/G39/G2E/G32 Digital Communication Services
Type of Service Media/Carrier Data Rate
Switched-56 2 or 4 wire/T–0 56 Kbps
Dedicated line 4 wire/T–1 56 Kbps–1.54 MbpsFrame Relay
Data only 2 wire/ISDN-BRI 64 Kbps-128 Kbps
Data and Video 4 wire/ISDN-PRI 384 Kbps –1.54 Mbps
2B+D 2 wire/ISDN-BRI 128 Kbps
Broadband Fibre/ATM 1.5 Mbps and higher
The primary transmission media for communication systems are copper wires, fiber optics,
microwave links and/or satellite transponders. A hybrid network may rely on satellite trans-ponders for remote facilities, fiber optics for video images and copper wires for data, signallingand control. The communication requirement is predicted on real-time/store and forward modesof operation. The real-time interactive mode requires the transmission of a large amount ofinformation in a short time. The major emphasis is on transmission speed and bandwidth. On theother hand, store and forward tends to move blocks of data at a time and is less demanding in
speed and bandwidth requirements (Lin, 1999).
/G39/G2E/G37/G2E/G37 /G54/G65/G6C/G65/G6D/G65/G64/G69/G63/G69/G6E/G65/G20/G55/G73/G69/G6E/G67/G20/G4D/G6F/G62/G69/G6C/G65/G20/G43/G6F/G6D/G6D/G75/G6E/G69/G63/G61/G74/G69/G6F/G6E
Mobile communication and satellite communication have opened up new possibilities for mobile
telemedicine in emergency situations. A typical set-up is shown in Fig. 9.18. In a moving vehicle,

310 Handbook of Biomedical Instrumentation
Fig.9.18 Principle of telemedicine using mobile satellite communication
MicMic
SPSP
ECGinECGout
CCDCRT
VideoProc.VideoProc.
MUXTxTx
RxRxDmux
AudioProc.AudioProc.
Mobile StationFixed Station
SatelliteComp.
Comp.
MicVideoProcDMUX
SPComp
CCD AudioProc.RxCRT
:Microphone: Video processor: Demultiplexer
: Speaker: Computer
: Charge coupleddevice cameraMUXTx: Multiplexer: Transmitter
: Audio Processor: Receiver: Cathode ray tube

Biomedical Telemetry and Telemedicine 311
colour images, audio signals and physiological signals such as ECG and blood pressure are
obtained from the patient. These images and signals are multiplexed and transmitted to a fixedstation. In the fixed station, the signals received are demultiplexed and presented to a medicalspecialist. Instructions from the specialist are then transmitted back to the mobile station throughthe communication link.
In mobile communication, the capacity of transmission link is generally limited and is typically
10 kbps–100 kbps. These capacities are far below those required for the transmission of medicallysignificant information. For example, the transmission of a colour video signal requires a trans-mission capacity of 1–10 Mbps with data compression for a moving image (Shimizu, 1999). Byadopting data compression on video, audio, ECG and blood pressure signals, the total capacityrequired is about 19 kbps, which is well within the practical capacity of mobile communicationlink. Table 9.3 gives transmitting information on various parameters and data reduction ratiosused in mobile communication.
/G95/G20/G54/G61/G62/G6C/G65/G20/G39/G2E/G33 Parameters and data compression ratios (After Shimizu, 1999)
Data Sampling Compression Ratio Bit Rate
Video 256 ¥ 256 pixels/plane 10:1 8 kbit/sec
8-bit RGB/pixels1 plane/20 sec
Audio 8 bit/sample 4.8:1 10 kbit/sec
6000 sample/sec
ECG 3 channel 8:1 600 bit/sec
8 bit/sample (3 channel)
200 sample/sec
Blood Pressure 1 sample/min 1:1 0.3 bit/sec
16 bit/sample
/G39/G2E/G37/G2E/G38 /G55/G73/G65/G20/G6F/G66/G20/G49/G6E/G74/G65/G72/G6E/G65/G74/G20/G52/G65/G73/G6F/G75/G72/G63/G65/G20/G66/G6F/G72/G20/G54/G65/G6C/G65/G6D/G65/G64/G69/G63/G69/G6E/G65
The world wide web (WWW) is an internet resource through which information-producing sites
offer hyper-linked multi-media information to the general public or in some cases restricted accessto a certain group of people. Graphical browser programs such as Netscape are specially designedto access WWW resources and view their contents in text, graphics images and video. Suggestionshave been made to make use of the WWW for telemedicine applications. However, it is advan-tageous and widely recommended to have a dedicated link as it offers higher security to the data,is more reliable due to lesser links in the channel, is better tuned for real-time applications and isan ideal solution for an integrated application to the requirements of both the transmitting as well
as the receiving ends; e.g. image processing applications, video conferencing, video-microscopy,
scanning, etc.

312 Handbook of Biomedical Instrumentation
/G4F/G78/G69/G6D/G65/G74/G65/G72/G73
/G20/G31/G30/G2E/G31/G4F/G58 /G49 /G4D/G45/G54/G52/G59
Oximetry refers to the determination of the percentage of oxygen saturation of the circulating
arterial blood. By definition:
Oxygen saturation = []
[] [HbO
HbO Hb]2
2+
where [HbO2] is the concentration of oxygenated haemoglobin
[Hb] is the concentration of deoxygenated haemoglobin
In clinical practice, percentage of oxygen saturation in the blood is of great importance. This
saturation being a bio-constant, is an indications of the performance of the most important
cardio-respiratory functions. It is maintained at a fairly constant value to within a few percent inan healthy organism. The main application areas of oximetry are the diagnosis of cardiac andvascular anomalies; the treatment of post-operative anoxia and the treatment of anoxia resultingfrom pulmonary affections. Also, a major concern during anaesthesia is the prevention of tissuehypoxia, necessitating immediate and direct information about the level of tissue oxygenation.Oximetry is now considered a standard of care in anaesthesiology and has significantly reducedanaesthesia-related cardiac deaths.
The plasma (liquid part of the blood) is a very poor carrier of oxygen. At the pressures available,
only 0.3 ml of oxygen can dissolve in 100 ml of plasma, which is quite insufficient for the needs ofthe body. The red blood cells contain haemoglobin which can combine with a large volume ofoxygen so quickly that in the lungs it may become 97% saturated forming a compound calledoxyhaemoglobin. The actual amount of oxygen with which the haemoglobin combines dependsupon the partial pressure of oxygen. The total quantity of oxygen bound with haemoglobin in the
normal arterial blood is approximately 19.4 ml percent at a pO
2of 95 mmHg. On passing through
the tissue capillaries this amount is reduced to 14.4 ml per cent at a pO2of 40 mmHg. Thus, under
normal conditions, during each cycle through the tissues, about 5 ml of oxygen is consumed by thetissues from each 100 ml of blood which passes through the tissue capillaries. Then, when theHAPTER
1010

Oximeters 313
blood returns to the lungs, about 5 ml of oxygen diffuses from the alveoli into each 100 ml of blood
and the haemoglobin is again 97% saturated.
/G31/G30/G2E/G31/G2E/G31 /G49/G6E/G20/G56/G69/G74/G72/G6F/G20/G4F/G78/G69/G6D/G65/G74/G72/G79
When blood is withdrawn from the subject under anaerobic conditions and measurement foroxygen saturation is made at a later time in the laboratory, the procedure is referred to as in vitro
oximetry. For discrete blood samples, a spectrophotometric measurement of oxygen saturationcan be made by either a transmission method or a reflection method.
Transmission Oximetry: Measurement of the
degree of oxygen saturation of the blood can bemade by spectrophotometric method. In spectro-photometry, the concentrations of substancesheld in solution are measured by determiningthe relative light attenuations that the lightabsorbing substances cause at each of several
wavelengths. Applying Beer-Lambart’s model
(Fig. 10.1) of a light absorbing medium ofconcentration C and thickness b, the intensity of transmitted light I is related to the incident light
I
O, as follows:
I = IO–kCb
where K is known as the extinction coefficient and varies as a function of the substance and the
wavelength of light. The quantity KCb is called the absorbance A.
The spectral transmission characteristics of oxyhaemoglobin and reduced haemoglobin in the
visible and infrared regions of the spectrum are shown in Fig. 10.2(a). It is obvious that thewavelength best suited for the measurement of oxygen saturation of blood is between 600 and 700b
CI I0
Fig.10.1 Principle of transmission
oximetry
600 500 700 800 900
Wavelength (nm)
(a)20
0406080100
Precent transmissionHbHb02
Fig.10.2 (a) Spectral characteristics of blood

314 Handbook of Biomedical Instrumentation
nm where the difference between the extinction
coefficients of oxidized and reduced blood is thegreatest. However, if the measurement is based ona given wavelength the result does not dependonly on the extinction coefficient but also upon the
total haemoglobin content and the thickness of the
blood quantity equivalent to be found in the tissues.Generally, the total haemoglobin varies with theindividual and the thickness of the equivalentchanges in time with a periodicity determined bycardiac function. It is thus necessary to attempt theelimination of these factors. This is usually achievedby making measurements at two wavelengths, sothat the variations in the haemoglobin concentra-tion gets nullified.
The extinction coefficient for haemoglobin and
oxyhaemoglobin, when plotted against wavelengthin the visible and near infrared regions of the spec-trum, are shown in Fig. 10.2(b). It is observed that
at 805 nm, the molecular extinction coefficient for fully saturated and fully reduced blood, are
equal. This point is vital to the design and principle of operation of oximeters. One measurementis made at a wavelength of 650 nm (red) and the second at 805 nm (infrared). The red channelprovides a signal which depends upon the amount of oxygen in the blood and the amount ofblood and tissues in the optical path. The infrared channel signal is independent of oxygensaturation and carries information on the amount of blood and tissue in the optical path.
The Lambert-Beer law or the spectrophotometric technique applies only to haemolyzed blood,
i.e. to blood in which the individual red cells have been destroyed and the pigments contained inthe cells are homogeneously distributed in the whole solution. This is necessary to eliminateartefacts associated with multiple scattering of the measuring light from erythrocytes. Because ofthe high density of haemolyzed blood, very short light paths have to be used. Several designs ofmicro and ultramicro cells have been made for the purpose.
The calibration curve of the oximeters departs from linearity at 90 to 100% saturation. This is
due to the fact that the light absorbing characteristics of oxygenated and reduced haemoglobin donot exactly follow Lambert-Beer’s law in this region (Matthes and Gross, 1939). Also, the relationshipof concentration to light absorption applies only in the case of monochromatic light. Oximeters
which make use of filters fall far short of giving monochromatic light. Therefore, the calibration
curve some what deviates from theoretical predictions.
Reflection Oximetery: Reflection oximetry is based on the scattering of light by the erythrocytes.
For the light scattered from the unhaemolyzed blood sample, oxygen saturation is given by:
Oxygen saturation =
I
Ir
r()
()l
l2
1 + br (i)l2l1e1e2
l3Hb
Hb02
Wavelength (nm)
(b)Extinction coefficient
Fig.10.2 (b) Molecular extinction coeffi-
cient for haemoglobin andoxyhaemoglobin plottedagainst wavelength

Oximeters 315
Polanyi and Hehir (1960) showed experimentally that a linear relationship exists between
Ir(l2)/Ir(l1) and oxygen saturation.
They computed the relationship as follows:
Oxygen saturation = 1.13 – 0.28 ¥I
Ir
r()
()805
650(ii)
Figure 10.3 shows the schematic arrangement of a reflection oximeter. Light from a tungsten
filament lamp (E) is condensed on the plane bottom of a cylindrical cuvette (F). The cuvette has a15 mm internal diameter and contains about 2 ml of whole blood. A portion of the light scatteredby the sample at an angle of about 135
0 with respect to the impinging light is condensed on two
matched photoconductors (A and B). Two interference filters (C and C’) limit the light reachingeach cell to a narrow band centred at l
1 = 650 nm and l2 = 805 nm, respectively. The ratio of the
resistance of the photocells is measured by means of a conventional Wheatstone bridge. The ratio
of the light intensity scattered by the sample is obtained as the inverse of this ratio.
FG
C C¢
B A
EG
E
F
A B,
C C,¢= Blood sample
= Tungsten lamp
= Glass cuvette= Photocells= Interference filters
Fig.10.3 Schematic arrangement of reflection oximeter (after Polanyi and Hehir,1960)
The optical path of a typical instrument which measures the oxygen saturation of unhaemolyzed
blood on samples as small as 0.2 ml by using a micro-cuvette is shown in Fig. 10.4. Operating onthe principle of reflection, the instrument employs two wavelengths (660 and 805 nm) andheparinized whole blood to measure the percentage of oxygen independent of wide haematocritchanges. The light reflected at 660 and 805 nm (isobestic point) is measured and the ratio of thetwo light intensities is, therefore, a direct function of the blood oxygen saturation of blood.
/G31/G30/G2E/G31/G2E/G32 /G49/G6E/G20/G56/G69/G76/G6F/G20/G4F/G78/G69/G6D/G65/G74/G72/G79
In vivo oximetry measures the oxygen saturation of blood while the blood is flowing through thevascular system or it may be flowing through a cuvette directly connected with the circulatory

316 Handbook of Biomedical Instrumentation
system by means of a catheter. The blood in this case is unhaemolyzed. Both techniques, reflection
and transmission, are utilized for in vivo oximetry.
/G20/G31/G30/G2E/G32 /G45/G41/G52/G20/G4F/G58/G49/G4D/G45/G54/G45/G52
Ear oximeters usually make use of the transmission principle to measure the arterial oxygensaturation. In this case, the pinna of the ear acts as a cuvette. Blood in the ear must be made similar
to arterial blood in composition. This is done by increasing the flow through the ear without
appreciably increasing the metabolism. Maximum vasodilatation is achieved by keeping the earwarm. It takes about 5 or 10 min for the ear to become fully dilated after the ear unit has been putup in place and the lamp turned on.
Merrick and Hayes (1976) describe details of an ear oximeter which enables the measurement of
oxygen saturation of blood. This measurement is independent of a wide range of encounteredvariables and is made without involving patients in any calibration or standardization procedure.In brief, the technique involves measuring the optical transmittance of the ear at 8 wavelengths inthe 650 to 1050 nm range. A 2.5 m long flexible fibre ear probe connects the patient to the instrument.The ear probe can be either held in position for discrete measurements or can be convenientlymounted to a headband for continuous display. The resulting light transmissions are processeddigitally according to a set of empirically determined constants and the resulting oxygen satura-tion results are displayed in the digital form.
The instrument is based on the Beer-Lambert law. However, it is assumed that the optical
absorbers act independently and additively and that the effects of light scattering by the ear tissuecan be minimized by a proper source and detector geometry. The mathematical statement of this
law for wavelength can be written as:
Aj = E
1jC1D1 + E2jC2D2+…+ EijCiDi+…..+ ENjCNDNOptical standard
Heat eliminating
mirror
Condenser
lens
Lamp
Light shutter
Photocell
Interference
filterCuvette
Collimating
lens
Heat
dissipator
Beam splitter
Light
attenuator
Interference
fillter
Collimating
lens
Photocell
Fig.10.4 Optical path of dual beam oximeter (Courtesy: M/s American Optical, USA)

Oximeters 317
where Ajis the total absorbance due to N layers of absorbers of concentration Ci, thickness Di and
with extinction coefficients Eij.If the measurements are made at eight wavelengths, the absorbance
Aj at wavelength j can be related to the transmittance Tj in the following way:
Aj = –log Tjwhere Tj = [I/IO]j
where IO is the intensity of the light falling on the ear and Iis the transmitted light intensity, all at
wavelength j. The instrument is designed to measure the percentage of functional haemoglobin
combined with oxygen. This can be expressed as:
SO2 = C
CCO
OR+¥ 100
where CO is the concentration of oxyhaemoglobin and CRis the concentration of deoxyhae-
moglobin. These are two of the eight absorbers thought to be present. A working expression forsaturation can be found by manipulating the set of eight simultaneous absorbance equationsinvolving eight concentrations. This results in an expression with the following form:
SO
2 = AA T A T
BB T B TO
O++ +
++ +11 88
11 88log ….. log
log ….. log
The coefficient AO through A8 and BO through B8 are constants which were empirically
determined by making a large number of measurements on a select group of volunteers. Havingfound the coefficient set, the instrument measures ear transmittance at eight wavelengths 20 timesper second, performs the indicated calculations, and displays the results.
Figure 10.5 explains the basic operation of the instrument. The light source is a tungsten-iodine
lamp that has a high output in the spectrum of interest. A lens system collimates the light beamand directs it through thin-film interference filters that provide wavelength selection. These filters
Analog-to-
digital
converterDigital
averaging
and storageCentral
processorDigital
display
Analog
output
Alarms/
inopsSync.
circuitsDetector Reference
path Lamp
Filter wheel
Optical sync.
pin holesFibre optic
ear probe
Fig.10.5 Block diagram of ear oximeter Model 47201A H.P. (Courtesy: Hewlett
Packard, USA)

318 Handbook of Biomedical Instrumentation
are mounted in the periphery of a wheel rotating at 1300 rpm and thus cut the light beam
sequentially. The filtered light beam then enters a fibre optic bundle that carries it to the ear.Another fibre optic bundle carries the light passing through the ear back to a detector in the
instrument. A second light path is developed with a beam splitter in the path of the collimated
light beam near the source. This path also passes through the filter wheel and then through a fibreoptic bundle directly to the photodetector. So, the detector receives two light pulses for eachwavelength. The processor takes the ratio of two pulses as the measured value; so readings arecompensated for any changes in the spectral characteristics of the light source and optical system.
The current developed at the photodetector is only 0.5 nA or less during a light pulse. This is
amplified in a high gain amplifier and then converted to a 16-bit digital form by an A-D convertersynchronized with the wheel rotation. The 16-bit words are given to a digital signal averager thatperforms two functions. First, it averages out the noise content of the signal with a time constant of1.6 s and secondly it serves as a buffer to hold information till it is required for computation.
Computation of percent oxygen saturation is accomplished by a 24-bit algorithmic-state
machine. It uses serial processing with the program stored in ROM and the necessary coefficientsof the equations stored on a field programmable ROM. The computation circuits also derive thequantity of total haemoglobin seen within the field of view of the earpiece. If this quantity is low,the instrument displays an ‘Off Ear’ indication. From the computational section, data is transferredin pulse-decimal form to the output circuit board where it is converted to BCD for the front panel
digital display.
The patient related part consists of arterializing blood flow in the pinna by a brisk 15 s rub .
Application of the probe to the ear results in a suitable display in about 30s. A built-in heater
regulated to 41
oC maintains arterialization. Restandardization is not required when the instrument
is to be used on other patients.
Measurements at eight wavelengths provide a great deal of information, which makes it
possible to account for eight unknowns. This is sufficient to take into consideration the patient topatient variables and account for the various forms of haemoglobin. The procedure is simple,requiring only the storage of initial light intensities at each of the eight wavelengths. However, itis still necessary to arterialize blood flow by warming the ear, and a large ear probe incorporatingfibre optics is necessary to make the system work.
/G20/G31/G30/G2E/G33 /G50/G55/G4C/G53/G45/G20/G4F/G58/G49/G4D/G45/G54/G45/G52
Pulse oximetry is based on the concept that arterial oxygen saturation determinations can be madeusing two wavelengths, provided the measurements are made on the pulsatile part of the wave-form. The two wavelengths assume that only two absorbers are present; namely oxyhaemoglobin
(HbO
2) and reduced haemoglobin (Hb). These observations, proven by clinical experience, are
based on the following:
(i) Light passing through the ear or finger will be absorbed by skin pigments, tissue, cartilage,
bone, arterial blood, venous blood.
(ii)The absorbances are additive and obey the Beer-Lambert law:
A = –log T = log lo/I = eD C

Oximeters 319
where Io and I are incident and transmitted light intensities, e is the extinction coefficient,
D is the depth of the absorbing layer and C is concentration.
(iii) Most of the absorbances are fixed and do not change with time. Even blood in the capillar-
ies and veins under steady state metabolic circumstances is constant in composition andflow, at least over short periods of time.
(iv) Only the blood flow in the arteries and arterioles is pulsatile.
Therefore, only measuring the changing signal, measures only the absorbance due to arterial
blood and makes possible the determination of arterial oxygen saturation (SaO
2). This is
uninfluenced by all the other absorbers which are simply part of the constant background signal.
Figure 10.6 (a) shows a typical finger tip oximeter probe in use whereas Fig. 10.6(b) shows the
construction of a typical pulse oximeter probe. This has two LEDs (light emitting diodes), one thattransmits infrared light at a wavelength of approximately 940 nm and the other transmitting lightat approximately 660 nm. The absorption of these select wavelengths of light through livingtissues is significantly different for oxygenated haemoglobin (HbO
2) and reduced haemoglobin
(Hb). The absorption of these selected wavelengths of light passing through living tissue ismeasured with a photosensor.
940 nm infrared LED660 nm red LED
Photosensor(a)
(b)
Fig.10.6 (a) A typical finger tip pulse oximeter probe in use
(b) Components of a pulse oximeter probe
The red and infrared LEDs within the probe are driven in different ways, depending on the
manufacturer. Most probes have a single photodetector (PIN-diode), so the light sources aregenerally sequenced on and off. A typical pulsing scheme of the LEDs is shown in Fig. 10.7. To

320 Handbook of Biomedical Instrumentation
compensate for ambient light during the time when both LEDs are off, the light level is measured
and then subtracted from each light channel between cycles. This minimizes the effects due toambient conditions which may vary during monitoring. Depending on the make and model ofpulse oximeters, the drive currents of LEDs, pulse widths, off and on cycles between pulses andcycle times can all vary (Ackerman and Weith, 1995).
The output of the photodiode has a raw signal represented in Fig. 10.8. There will be one signal
that represents the absorption of red light (660 nm) and one that represents infrared light (940 nm).The ac signal is due to the pulsing of arterial blood while the dc signal is due to all the non-pulsing
absorbers in the tissue. Oxygen saturation is estimated from the ratio ( R) of pulse-added red
absorbance at 660 nm to the pulse-added infrared absorbances at 940 nm.
R =
ac dc
ac dc660 660
940 940/
/
Figure 10.9 shows the analog signal processing technique used in pulse oximeters. To simplify
the diagram, the circuitry required to drive the LEDs in the sensor are omitted, and only the analog
signal processing blocks between the sensor and the digital processing circuitry are shown. The
signal from the sensor is a current. The first amplifier stage is a current to voltage converter. Thevoltage signal then goes through the following circuits: amplifiers to further amplify the signal;noise filters to remove different kinds of interference, a demultiplexer to separate the interleaved
red and infrared signals; bandpass filters to separate the low frequency (dc) component from the
pulsatile, higher frequency (ac) component; and an analog–digital converter to convert thecontinuously varying signal to a digital representation.
An advancement over the analog signal processing arrangement has been described by Reuss
(2000) which eliminates analog circuitry for signal processing and replaces it with a digital signal
processing the microprocessor. The output from the sensor is directly given to a high dynamic
range analog-to-digital convertor followed by a microprocessor which supports the requireddigital signal processing. This technique offers the advantages of less circuitry, higher reliability,smaller size and lower cost.Photodiode current
TimeOff
timeRed
Infrared InfraredRed
Off
timeOff
time1000 HzCycle time
Off
time
Fig.10.7 Typical pulsing of red and infrared light emitting diodes by a pulse oximeter

Oximeters 321
An accuracy of 1% or better has been reported for the saturation range of above 80% for most
transmission type pulse oximeters. Usually, the accuracy is less at lower saturation because of
non-linear effects of absorption.
The pulse oximeter offers the following advantages:
• It removed the requirement of arterializing blood flow. No heating or rubbing is necessary.
The measurement requires that pulsatile activity should be present, but the level is notcritical.Incident lightNo pulsation Pulsatile bloodTimeTimeVenous bloodArterial bloodVariable absorption
due to pulse added
volume of arterialbloodImax
IminAC- componentImax(DC-component)Intensity of
transimitted
light
Transmitted
light
Fig.10.8 Infrared light absorption in the finger

322 Handbook of Biomedical Instrumentation
• Since a change in signal is measured, it is not necessary to store any initial light intensity
values, simplifying operational procedures.
• The instrument can be empirically calibrated. Subject variability (skin pigmentation, thick-
ness, tissue, sensor location, etc.) has no significant influence on the measurement.
• True arterial saturation is measured because the pulsatile signal comes from the arterial
blood.
A limitation of the pulse oximeter is that ambient lights have been shown to interfere with the
measurement. Therefore, covering the cuff with an opaque material is necessary to prevent suchinterference. Motion artifact is also a potential problem. This is because the information containingpulse activity is in the same frequency range as motion artifact.
/G20/G31/G30/G2E/G34 /G53/G4B/G49/G4E/G20/G52/G45/G46/G4C/G45/G43/G54/G41/G4E/G43/G45/G20/G4F/G58/G49/G4D/G45/G54/G45/G52/G53
For the measurement of oxygen saturation level of blood in localized areas of oxygen deprived
tissues on the limbs, head and torso, a skin reflectance oximeter can be employed. The instrument
basically depends on monitoring backscattered light from living tissue in two wavelengths. Thebackscattered light data is then used for the in vivo determination of the blood’s relative oxygensaturation.
Cohen and Wadsworth (1972) bring out the difficulties in the extraction of useful information
from backscattered light intensity from human tissue. There are vast variations of tissueconstruction and optical properties among various subjects and in different locations on the samesubject. Moreover, blood constitutes only a small fraction of the tissue under investigation. Thefraction of volume occupied by blood is unknown, as are the exact optical properties andconstruction of the skin and lower tissues. Cohen and Longini (1971) suggested a theoreticalsolution to some of these problems. They considered human tissues to be composed of parallelsemi-infinite layers of homogeneous materials. By using normalization techniques based on dataexperimentally collected from the living tissue under varying conditions of oxygenation and bloodvolume from various subjects, and processed by a computer, it was found possible to design an
oximeter which was less sensitive to variations in parameters such as blood volume, skin
pigmentation and construction, age, etc.Sensor
Modulator
red and IRAmp. DemultiplexerBandpass filter
Bandpass filterred
IRDC
DCAC
ACAmp.
Amp.Analog
digital
converter
Digital
MicroprocessorSource signal
Fig.10.9 Analog signal processing by pulse oximetry

Oximeters 323
However, in comparison to transmission, the reflection pulse oximeter has poorer signal-to-
noise ratio. Mendelson et al (1988) utilized multiple photodiodes around the light source to
enhance signal level. This arrangement is shown in Fig. 10.10. A pair of red and infrared light
emitting diodes are used for the light source, with peak emission wavelengths of 665 nm (red) and
935 nm (infrared). The reflected light from the skin at these two wavelengths is detected by asilicon diode. These detected signals are processed in the form of photo-plethysmographs todetermine So
2.
Red and infrared LEDs
Delrin holder
Airpax package
Optical shield
(a)
(b)Conductive
epoxyCeramic
substratePhotodiode
5 m
PhotodiodeLEDs
Ceramic
substrateOptically clear epoxy
Optical shield
Airpax Package
Delrin holder
Fig.10.10 Schematic diagram of the reflectance oximeter sensor
(redrawn after Mendelson et al, 1988)
A heater was incorporated in the sensor to warm the tissue so as to increase local blood flow.
Excellent correlation in comparison with the transmission oximeter has been shown from the calfand thigh.
/G20/G31/G30/G2E/G35 /G49/G4E/G54/G52/G41/G56/G41/G53/G43/G55/G4C/G41/G52/G20/G4F/G58/G49/G4D/G45/G54/G45/G52
For intravascular oximetry, modern instruments make use of optical fibres to guide the light signalinside the vessel and the reflected light from the red blood cells back to the light detector. Forestimating SO
2, usually the reflectance at two wavelengths, one in the red and the other in the near
infrared regions, are used. The relationship is given by:
SO2 = A + B (Rl1/Rl2)
where A and B are the constants that depend upon the fiber geometry and physiological parameters
of blood (Polany and Hehir, 1960). The equation forms the basis of reflection oximetry.
Although the computation of reflectance ratios at two wavelengths has been suggested to cancel
the effects of non-linearity due to scattering, effects of haemotacrit change may become important.

324 Handbook of Biomedical Instrumentation
Therefore, currently available fiber-optic oximeters utilize more than two wavelengths to adjust
for haematocrit variation (Takatani and Ling, 1994).
One of the problems in fiber optic oximeters is that damage to optical fibers result in severe
measurement error. In order to overcome this problem, catheter tip type oximeters using hybridtype miniature sensors such as shown in Fig. 10.11 have been developed.
0.2 inchRed LEDNear infrared
LEDPhotodiode Resistor
0.04 inch
Fig.10.11 A catheter – tip hybrid circuit oximeter for intravascular
measurement (after Takatani, 1994)
Intravasacular oximeters are normally used to measure mixed venous saturation, from which
the status of the circulatory system can be deduced. Mixed venous saturation varies in reflectingthe changes of oxygen saturation, cardiac output, haematocrit or haemoglobin content and oxygenconsumption.

Blood Flowmeters 325
/G42/G6C/G6F/G6F/G64/G20/G46/G6C/G6F/G77/G6D/G65/G74/G65/G72/G73
Blood flow is one of the most important physiological parameters and also one of the most difficult
to measure accurately. This is because instruments for measuring the flow through blood vesselswithin the body have to meet certain stringent specifications; e.g. sensitivity and stabilityrequirements depend upon the magnitude of flow, location and the diameter of the individualvessels. The average velocities of blood flow vary over a wide range in vessels with diametersranging from 2 cm to a few millimetres. Besides meeting the technical requirements, the measuringsystem must meet the specific clinical requirement of being the least traumatic. Blood flow
measurement is thus a difficult engineering and clinical problem. It is only natural that a variety
of techniques have since been developed in an effort to meet the requirements of an ideal flowmetering system.
There are many widely used techniques for measuring the blood flow and velocity. They are
categorized into invasive (surgical) and non-invasive (through the skin). The most accuratemethod, obviously, is to simply sever the vessel and time the blood flow into a calibrated beaker.But this procedure is too radical for most protocols. The most commonly used blood flow metersare described in this chapter.
/G20/G31/G31/G2E/G31 /G45/G4C/G45/G43/G54/G52/G4F/G4D/G41/G47/G4E/G45/G54/G49/G43/G20/G42/G4C/G4F/G4F/G44/G20/G46/G4C/G4F/G57/G4D/G45/G54/G45/G52
The most commonly used instrument for the measurement of blood flow is of the electromagnetictype. With this type of instrument, blood flow can be measured in intact blood vessels withoutcannulation and under conditions which would otherwise be impossible. However, this methodrequires that the blood vessel be exposed so that the flow head or the measuring probe can be put
across it.
The operating principle underlying all electromagnetic type flowmeters is based upon
Faraday’s law of electromagnetic induction which states that when a conductor is moved at right
angles through a magnetic field in a direction at right angles both to the magnetic field and itslength, an emf is induced in the conductor. In the flowmeter, an electromagnetic assembly providesHAPTER
1111

326 Handbook of Biomedical Instrumentation
the magnetic field placed at right angles to the blood
vessel (Fig. 11.1) in which the flow is to be measured. Theblood stream, which is a conductor, cuts the magnetic
field and voltage is induced in the blood stream.
This induced voltage is picked up by two electrodesincorporated in the magnetic assembly. The magnitudeof the voltage picked up is directly proportional to thestrength of the magnetic field, the diameter of the bloodvessel and the velocity of blood flow, i.e.
e = CHVd
where e= induced voltage
H= strength of the magnetic field
V= velocity of blood flow
d= diameter of the blood vessel
C= constant of proportionality
If the strength of the magnetic field and the diameter of the blood vessel remain unchanged,
then the induced voltage will be a linear function of the blood flow velocity. Therefore, e= C
1 V
where C1 is a constant and equal to CHd.
Further, the flow rate Q through a tube is given by
Q=VA
therefore,
V= Q /A
where A is the area of cross-section of the tube, hence
e= C1¥ Q/A = C2¥ Q
where C2 is a general constant and is given by C1/A.This equation shows that the induced voltage
is directly proportional to the flow rate through the blood vessel.
The induced voltage picked up by the electrodes is amplified and displayed/recorded on a
suitable system. The system is calibrated in terms of volume flow as a function of the inducedvoltage. The diameter of the blood vessel is held constant by the circumference of the hole in theprobe that surrounds it.
The above relationship is true only if there exist conditions of axial symmetry and the blood
velocity is considered independent of the velocity profile. Only then, the induced voltage is directlyrelated to the blood volume flow.
Design of the Flow Transducer: In actual practice, the electromagnetic flowmeter transducer
(Wyatt, 1984) is a tube of non-magnetic material to ensure that the magnetic flux does not bypassthe flowing liquid and go into the walls of the tube. The tube is made of a conducting material andgenerally has an insulating lining to prevent short circuiting of the induced emf. The induced emfis picked up by point electrodes made from stainless steel or platinum. The flow head (Fig. 11.2)contains a slot through which the intact blood vessel can be inserted to make a snug fit. Severalprobes of different sizes must therefore accompany the flowmeter to match the full range of sizes
of the blood vessels which have various diameters. It is naturally more difficult to construct flowBlood
vesselSignal
sensing
electrodesElectro-
magnet
Magnetic
fieldMagnet
currentSignal
voltage
Flow
Fig.11.1 Principle of electromag-
netic flowmeter

Blood Flowmeters 327
heads suitable for use with very small blood vessels. However, flow heads having as small as
1 mm external diameter have been reported in literature.
The flow-induced voltage of an electromagnetic flowmeter is, within certain limitations,
proportional to the velocity of the flow. This velocity is the average across the flow stream with anaxis symmetric velocity profile. The average flow velocity appears to be 20 to 25 cm/s in arteriesand 10 to 12 cm/s in veins. For designing the probe, velocity for the cardiovascular system is takenas 15 cm/s. For non-cannulated probes, a uniform magnetic field over the measuring area is soselected that it has a convenient shape and the smallest size (Cunningham et al. 1983). Ironcored electromagnets are used in probes having a diameter between 1 to 8.2 mm, and air coredelectromagnets are use in diameters above 8.2 mm. Cannulated probes for extracorporeal use canhave greater field strengths and magnet size as the constraint of small size is no longer present.
To obtain a reliable recording of the flow, a certain constriction of the vessel within the probe is
necessary to maintain good contact. In order to limit the amount of constriction, a 20% incrementalrange of sizes is chosen. To protect the probe from chemical attack, it must be encapsulated in abiologically inert material having a high electrical and chemical resistance, e.g. silicone rubber.
The probes can generally be sterilized by chemical means. Probe calibration is carried out in 0.9%
saline during manufacture and each probe is given a calibration factor that is engraved on theconnector. This factor is set on a multi-turn potentiometer to adjust the amplification to give a fullscale output on the display meter.
Fig.11.2 Different types of flow heads (Courtesy: Cardiovascular Instruments
Ltd., UK)

328 Handbook of Biomedical Instrumentation
The transducers are usually equipped with an internal ground electrode so that no external
grounding is necessary. The cable from the transducer to the instrument should comprise of ateflon insulated wire completely shielded with a tinned copper braid. The entire cable is sleeved
with medical grade silicone rubber tubing and impregnated with silicone rubber to minimize
leakage and electrical noise. The transducers should be tested for 1,000 megaohms minimumresistance between the coil and electrodes, after prolonged immersion in saline.
Common-mode rejection is influenced by the difference between the impedances of the pick-up
electrodes of a transducer and is maximum only when these impedances are identical. Electrodeimpedances in saline generally lie between 1 k W and about 10 k W, values of 1.5–2 k W being typical
of platinized platinum and dull gold electrodes. The common-mode impedance of the measuringcircuit should be at least 100 M W to ensure that variations of several hundred ohms between
the impedances of a pair of transducer electrodes do not significantly affect the common moderejection.
The early models of electromagnetic blood flowmeters employed permanent magnets which
were subsequently replaced by electromagnets powered at mains frequencies. However, thetremendous interference from mains voltages at the transducer electrodes resulted in baselinedrifts and poor signal-to-noise ratio. Various types of waveforms at different frequencies havesince been tried to overcome these difficulties. Today, we have electromagnetic flowmeters whosemagnetic coils work on sine, square or trapezium current waveforms.
The probes have an open slot on one side which makes it possible to slip it over a blood vessel
without cutting the vessel. The probe must fit the vessel during diastole so that the electrodes make
good contact. Probes are made in 1 mm increments in the range of 1 to 24 mm to ensure that they
can be used on a variety of sizes of arteries. The probes generally do not operate satisfactorily onveins because the electrodes do not make good contact when the vein collapses. Attempts havebeen made to miniaturize the flow head to an extent that it can be mounted on the tip of a catheter.Such devices are needed in experimental as well as in clinical cardiology for mapping the velocityconditions of the entire circulatory system.
/G20/G31/G31/G2E/G32 /G54/G59/G50/G45/G53/G20/G4F/G46/G20/G45/G4C/G45/G43/G54/G52/G4F/G4D/G41/G47/G4E/G45/G54/G49/G43/G20/G46/G4C/G4F/G57/G4D/G45/G54/G45/G52/G53
Basically all modern flowmeters consist of a generator of alternating current, a probe assembly, aseries of capacitance coupled amplifiers, a demodulator, a dc amplifier and a suitable recordingdevice. The shape of the energizing current waveform for the electromagnet may be sinusoidal orsquare.
/G31/G31/G2E/G32/G2E/G31 /G53/G69/G6E/G65/G20/G57/G61/G76/G65/G20/G46/G6C/G6F/G77/G6D/G65/G74/G65/G72/G73
In a sine wave flowmeter, the probe magnet is energized with a sine wave and consequently theinduced voltage will also be sinusoidal in nature. The major problem encountered with thesinusoidal type of magnetic field is that the blood vessel and the fluid contained in it act as thesecondary coil of a transformer when the probe magnet is excited. As a result, in addition to theinduced flow voltage, there is an induced artefact voltage generally referred to as ‘transformervoltage’.

Blood Flowmeters 329
The ‘transformer voltage’ is much larger than the signal or flow induced voltage and is 90o out
of phase with it. This also causes baseline drift which necessitates high phase stability in theamplifier and demodulator circuits.
In earlier versions of sine wave flowmeters, this unwanted voltage was eliminated by injecting
into the signal a voltage of equal strength, but having an opposite phase. The artefact signal is thus
cancelled and only the flow induced voltage is left behind for display.
An alternative method to eliminate the transformer induced voltage in sinewave flowmeters is
by using a gated amplifier. The function of this amplifier is to permit the amplification of the
signals only during the portion of the cycle where flow induced voltages are maximum and thetransformer induced voltages are minimum. By this method, the artefact voltage is prevented fromgetting amplified. This type of instrument is known as a ‘gated sine wave flowmeter’.
The sine wave flowmeters require complicated electronic circuitry for the removal of transformer
induced voltage from the flow induced voltage. As both the waveforms are of the same type,complete elimination of the artefact voltage becomes extremely difficult. The sine wave system, nodoubt yields good signal-to-noise ratio, but imposes stringent phase stability requirements withincreasing frequency.
/G31/G31/G2E/G32/G2E/G32 /G53/G71/G75/G61/G72/G65/G20/G57/G61/G76/G65/G20/G45/G6C/G65/G63/G74/G72/G6F/G6D/G61/G67/G6E/G65/G74/G69/G63/G20/G46/G6C/G6F/G77/G6D/G65/G74/G65/G72/G73
This differs from a sine wave flowmeter in that the energizing voltage given to the magnet is asquare wave and therefore, the induced voltage is also a square wave. The square wave flowmeterhas less stringent requirements of phase stability than the sine wave type as it can suppress thequadrature voltages relatively easily. Also, it is easier to control the magnitude and wave shape ofthe energizing current in the case of a square wave system.
The transformer induced voltage in this case is only a spike, superimposed on the beginning
of the square wave flow induced voltage. Separation of these two voltages becomes easier asthe amplifier can be gated only for a very short period. In a square wave flowmeter, blanking isrequired only during the portion when the current in the magnet is reversing and the amplifierworks during the flat portion of the square wave. The square wave is amplitude modulated by thevariation in blood flow and requires to be demodulated before it can be fed to a recorder.
Figure 11.3 shows the block diagram of a square wave electromagnetic blood flowmeter.
Transducer: The flow transducer consists of an electromagnet, which provides a magnetic field
perpendicular to the direction of flow and lying within the field are a pair of pick-up electrodeswhose axis is perpendicular to both the field and the flow axis. The electrodes may be in contactwith either the flowing blood or the outer surface of the blood vessel carrying the flowing blood.The former is called ‘Cannulating flowmeter’ and the latter ‘Cuff flowmeter’.
Preamplifier: The induced voltage picked up by the electrodes is given to a low noise differential
amplifier through a capacitive coupling. The preamplifier must have a very high common-moderejection ratio and input impedance. The preamplifier used by Goodman (1969(a)) has a CMRR of106 dB (200,000 : 1) with a common mode input impedance of 150 M W. The preamplifier gain is of
the order of 1000. The preamplifier also must incorporate the facility for ‘probe balance’ by whichsignals in phase with the magnet current can be selected to balance background voltages in phase

330 Handbook of Biomedical Instrumentation
with flow voltages. A calibrating signal of 30 mV amplitude can be connected to the preamplifier
with an input selector switch.
Noise voltage generated in the preamplifier is an important factor in the performance of an
electromagnetic flowmeter. The noise voltage is seen as a random movement of the baseline of therecorded flow. When expressed in terms of flow, it is typically 1–2% of the full scale output of thechosen probe. For example, a 2.7 mm probe gives full scale deflection for 500 ml/min, so the noise
is equivalent to 10 ml/min.
Gating Circuit: A gating amplifier helps to remove spurious voltages generated during magnet
current reversal. For the flowmeter to exhibit a satisfactory base line stability, it is essential thatthe spurious signals produced during the magnet current reversal and those in phase with flow
voltages are made negligible. The gating action is controlled by the circuit which provides an
excitation current to the electromagnet.
Bandpass Amplifier: Following the gating amplifier is an active RC bandpass amplifier, which
selectively passes through it the amplified square wave signal. The peak response is kept for 400
Hz. The 3 dB points are at 300 and 500 Hz. The gain of this amplifier is typically 50. The shape of
the wave after this amplifier is a distorted sinusoid.Preamplifier
filterDetectorFilter
and
outputFlow
Quad
Flow meterLogicPower
amplifierMagnet
current
driveVessel
Magnet
Lumen
earth
Fig.11.3 Block diagram of a square wave electromagnetic flowmeter

Blood Flowmeters 331
Detector: A phase sensitive detector is used to recover the signal, which is an analogue of the flow
rate being measured. This type of demodulator not only offers maximum signal-to-noise ratio butalso helps in the rejection of interfering voltages at frequencies well below the carrier frequency.
Low-Pass Filter and Output Stage: The demodulated signal is given to an active RC low-pass
filter, which provides a uniform frequency response and a linear phase shift from 0–30 Hz. This isfollowed by an integrator circuit to provide an output corresponding to the mean flow. The outputsignal thus obtained can be put to a recorder to read the blood flow rate from the calibrated scale.
Magnet Current Drive: The excitation current supplied to the electromagnet is a one ampere peak
square wave current. It is given from a source of high impedance to ensure that it remains constantfor variations in magnet winding resistance of up to 5 W. The square wave input to the power
amplifier stage which supplies current to the electromagnet is fed from a free running multi-
vibrator working at 400 Hz.
Zero-Flow Reference Line: Before measurement can be made for blood flow with electromagnetic
flowmeters, it is essential to accurately establish the signal corresponding to zero-flow. Althoughde-energizing the magnet should produce a zero reference line, unfortunately this line does not
always coincide with the physiological zero-flow line. This is owing to some effects at the electrode
vessel interface. An alternative method could be to occlude the blood vessel in which flow is to bemeasured. Several arrangements have been used to act as occluders (Jacobson and Swan, 1966).However, there is a serious objection to their use because the necessity for occlusion of the vessel,in order to obtain a zero flow reference, introduces the possibility of producing a spasm and hencea change in the blood flow.
/G20/G31/G31/G2E/G33 /G55/G4C/G54/G52/G41/G53/G4F/G4E/G49/G43/G20/G42/G4C/G4F/G4F/G44/G20/G46/G4C/G4F/G57/G4D/G45/G54/G45/G52/G53
There are basically two types of ultrasonic blood flow-velocity meters. The first type is the transit-time velocity meter and the second is the Doppler-shift type. For routine clinical measurements,the transcutaneous Doppler instrument has, by far, superseded the transit-time type. Therefore,most of the recent efforts have been concentrated on the development of Doppler-shift instruments,which are now available for the measurement of blood velocity, volume flow, flow direction, flowprofile and to visualize the internal lumen of a blood vessel.
/G31/G31/G2E/G33/G2E/G31 /G44/G6F/G70/G70/G6C/G65/G72/G2D/G73/G68/G69/G66/G74/G20/G46/G6C/G6F/G77/G2D/G76/G65/G6C/G6F/G63/G69/G74/G79/G20/G4D/G65/G74/G65/G72/G73
It is a non-invasive technique to measure blood velocity in a particular vessel from the surface ofthe body. It is based on the analysis of echo signals from the erythrocytes in the vascular structures.Because of the Doppler effect, the frequency of these echo signals changes relative to the frequencywhich the probe transmits. The Doppler frequency shift is a measure of the size and direction ofthe flow velocity. The principle is illustrated in Fig. 11.4. The incident ultrasound is scattered bythe blood cells and the scattered wave is received by the second transducer. The frequency shiftdue to the moving scatterers is proportional to the velocity of the scatterers. Alteration in frequencyoccurs first as the ultrasound arrives at the ‘scatterer’ and second as it leaves the scatterer. If theblood is moving towards the transmitter, the apparent frequency f
1 is given by

332 Handbook of Biomedical Instrumentation
f1=(c o s )Cv
C- q(i)
where f= transmitted frequency
C= velocity of sound in blood
q= angle of inclination of the incident wave to the direction of blood flow
v= velocity of blood cells
Assuming that the incident and scattered radiations are both inclined at q to the direction of
flow as shown in Fig. 11.4,
f2=f1C
Cv+L
NMO
QPcosq(ii)
The resultant Doppler shift
Df=f–f2 = f–f1C
Cv+L
NMO
QPcosq(iii)
From equation (i), substituting f1, we get
=f–fC v
C(c o s )- LNMOQPq C
Cv+L
NMO
QPcosq
=f1–
+L
NMO
QP(c o s )
(c o s )Cv
Cvq
q
Since C >> v,
Df=2fv
Ccosq
v=D◊fC
f2c o s q
This relationship forms the basis of measuring blood velocity. Depending on the application,
either a signal proportional to the average instantaneous velocity or a signal proportional to thepeak instantaneous velocity may be required. The peak signal is easier to obtain and has beenshown to be of particular importance in localizing and quantifying the severity of peripheralvascular diseases (Johnston et al, 1978).FrequencydifferencedetectorTransmit
Receive
ff v+ cosD qqfVessel
Blood
cells
Fig.11.4 Principle of ultrasonic Doppler-shift flow velocity meter

Blood Flowmeters 333
In order to measure absolute velocity, the angle of inclination of the ultrasonic beam to the
direction of flow must be known. Several methods are available for doing so. The first uses theprinciple that the Doppler shift signal is zero when the ultrasonic beam is at right angles to the
direction of flow. So, by finding the position of the zero Doppler signal with the probe over the
vessel and then by moving it through a known angle of inclination, the angle becomes known.However, due to the separation of the transmitter and receiver, the Doppler shift frequencies arenot zero. In such cases, the position at which the minimum Doppler shift frequencies are presentis taken for the probe to be at right angles.
The early instruments used a continuous wave (CW) ultrasonic beam. Basically, a CW
ultrasonic Doppler technique instrument works by transmitting a beam of high frequencyultrasound 3–10 MHz towards the vessel of interest. A highly loaded lead zirconate titanatetransducer is usually used for this purpose. The transducer size may range from 1 or 2 mm to aslarge as 2 cm or more. A separate element is used to detect the ultrasound back scattered from themoving blood. The back-scattered signal is Doppler shifted by an amount determined by thevelocity of the scatterers moving through the sound field. Since the velocity varies with the vesseldiameter to form a velocity profile, the returned signal will produce a spectrum corresponding tothese velocities.
Flax et al (1973) discuss the circuit blocks (Fig. 11.5) of the Doppler ultrasonic blood flowmeter.
The piezo-electric crystal
AAAAA is electrically excited to generate ultrasonic waves, which enter the
blood. Ultrasound scattered from the moving blood cells excites the receiver crystal. The electrical
signal received at B consists of a large amplitude excitation frequency component, which is
directly coupled from the transmitter to the receiver, plus a very small amplitude Doppler-shiftedcomponent scattered from the blood cells. The detector produces a sum of the difference of thefrequencies at D. The low-pass filter selects the difference frequency, resulting in audio frequencies
atE. Each time the audio wave crosses the zero axis, a pulse appears at G. The filtered output
level at H will be proportional to the blood velocity. The following two pitfalls are encountered in
Doppler ultrasonic blood flowmeters. High frequency response is usually inadequate whichintroduces a non-linearity into the input-output calibration curve. Also, the low frequency gain is
R.F. amplifier Detector
Low-pass
filterLow-pass
filterAudio
amplifierZero
crossing
detectorOutputReceiver
crystal5 MHz
exciterBloodTransmitter
crystal
A
H G F EDB
Fig.11.5 Block diagram of Doppler shift blood flowmeter (after Flax et al, 1973; by
permission of IEEE Trans. Biomed. Eng.)

334 Handbook of Biomedical Instrumentation
normally too high, resulting in wall motion artefacts. The maximum Doppler shift has been
calculated as about 15 kHz. The wall motion signal can be significantly reduced by filtering outfrequencies below 100 Hz.
/G31/G31/G2E/G33/G2E/G32 /G52/G61/G6E/G67/G65/G20/G67/G61/G74/G65/G64/G20/G50/G75/G6C/G73/G65/G64/G20/G44/G6F/G70/G70/G6C/G65/G72/G20/G46/G6C/G6F/G77/G6D/G65/G74/G65/G72/G73
Instruments based on a CW Doppler can be better characterized as flow detectors rather thanflowmeters. Even a simple measurement of the mean flow velocity requires that the angle betweenthe sound beam and the velocity vector should be known. Baker (1970) states that recordings inliterature that are calibrated in terms of Doppler frequency shift are both misleading and frequentlyinaccurate. The investigators did not or could not make the angle measurement required to quantifytheir data. Many of the difficulties associated with the CW system can be overcome if the ultrasonic
source is pulsed and the Doppler shift of the returning echo is determined. If the return signal is
range gated, then the distance to the moving interface (blood vessel diameter) as well as its velocitywith respect to the beam can be measured.
Figure 11.6 shows the block diagram of a pulsed ultrasonic Doppler flow detector. The system
consists of:
Receiver
limiterPhase
detectorLow-
pass
filterSample
and
amplifier
Power
amplifierPulse
amplifierSample
pulse
gen.Low-pass
filterAudible
doppler
output
Master
oscillator
4.5–5.5 MHzFreq.
division
20, 10, 5, 2.5,
1.25 KHzTransmitter
elementReceiver
element
Fig.11.6 Block diagram of a pulsed Doppler flowmeter (after Baker, 1970)
Transmitter: It comprises a master oscillator whose frequency may vary from 2 to 10 MHz; however
5 MHz is a good compromise for detecting the blood flow in vessels 4-6 cms in depth. The choiceof frequency is a compromise between the transmission attenuation losses over a given path
length. The master oscillator also drives the pulse repetition frequency (PRF) ripple counter and
provides a continuous reference signal to the receiver phase detector for the detection of theDoppler-shifted echoes.
The selection of PRF depends on the depth of the vessel, expected Doppler shift and attenuation
characteristics of the tissue. The following values of PRFs and range depth combinations foroptimum results have been suggested:

Blood Flowmeters 335
PRF (kHz) Approximate Depth Range (cm) Max. Df Detectable (kHz)
25 3.0 12.518 4.3 9.012.5 6.0 6.25
The PRF pulse triggers a 1 ms one shot multi-vibrator that generates the gating signal. When the
gate opens, 5 cycles of the master oscillator pass through to a power amplifier. The peak poweroutput applied to the transmitting transducer is in the range of 10 to 30 W during the 1 ms excitation
burst. With the PRF of 25 kHz, the average power applied to the transducer ranges from 0.25 to0.75 W depending on the transducer impedance, which may be 50 to 200 W. The radiating area of
the transducer ranges from 0.5 to 2 cm
2.
The transducers for pulsed Doppler applications should ideally have the sensitivity of the
lightly backed narrow-band CW Doppler units and the wide bandwidth of the pulse echounits. The Qof transducers designed for CW applications ranges from 10 to 30 or more. These
transducers ring at their resonant frequency long after the electrical signal stops. Pulse echo-
transducers use tungsten epoxy backing to have a Qas low as 1.5–2.5. However, the Q is not
lowered to a desirable value of about 1.5 to 5 because this would greatly decrease both the efficiencyof the transmission and the sensitivity of the reception. The Q of pulsed Doppler transducers is
generally kept at 5 to 15 and some ringing is thus tolerated. These transducers have a backing ofaluminium powder in epoxy.
Receiver: The back-scattered Doppler-shifted signals from a blood vessel range in intensity from
50 dB to more than 120 dB down from the transmitted signal. The receivers designed for thispurpose have a bandwidth of 3 MHz and gain in excess of 80 dB. This is followed by a single sideband type quadrature phase detector which separates the upper and lower Doppler side bands forsensing flow direction. The detector consists of a phase-shift network, which splits the carrier intotwo components that are in quadrature, which means they are 90°. These reference cosine and sinewaves must be several times larger than the RF amplifier output. The Doppler shift of the receivedecho, back scattered by the moving blood is detected by sensing the instantaneous phase differencebetween the echo and a reference signal from the master oscillator. If the echo at a particular rangeor depth contains a Doppler-shift, then the amplitude of a sample of the instantaneous phase
difference will vary in amplitude exactly with the Doppler difference frequency. The envelope
frequency from the phase detector is the Doppler difference frequency.
If the flow of blood is in the same direction as the ultrasonic beam, then it is considered that the
blood is flowing away from the transducer. In this case, the Doppler-shift frequency is lower thanthat of the carrier and the phase of the Doppler wave lags behind that of the reference carrier. If theflow of blood is towards the transducer, then the Doppler frequency is higher than the carrierfrequency and the phase of the Doppler wave leads the reference carrier. Thus, by examining thesign of the phase, the direction of flow can be established.
The depth at which the Doppler signal is sensed within the blood vessel and tissue depends on
the delay interval between the transmitted burst and the sample gate. For a velocity of sound, intissue of approximately 1500 m/s, the range factor is 13.3 ms delay/cm of depth. A one-shot multi-
vibrator is used to develop the adjustable range delay calibrated in millimetres of depth. This isfollowed by a ‘sample gate’. The detected Doppler shift frequency will correspond to the mean

336 Handbook of Biomedical Instrumentation
velocity averaged over the sample volume. However, the Doppler signals at the output of the sample
and hold circuit are low in amplitude and contain frequency components from the transmitterpulse repetition rate and the low frequency large amplitude signals produced by the motion of
interfaces such as vessel walls and the myocardium. Therefore, both high-pass and low-pass
filters are used to extract the signals of interest. Once the signals are filtered with a passband of 100Hz to 5 kHz, it is further amplified to drive an external audio amplifier or a frequency meter.
Zero-crossing Detectors: In order to measure blood velocity, a frequency meter is needed to analyze
the frequency components of the Doppler signal. Nearly all Doppler velocimeters use zero-
crossing detectors for this purpose. The function of the zero-crossing detector is to convert theaudio frequency amplifier output to a proportional analog output signal. It does this by emitting aconstant area pulse for each crossing of the zero axis from negative to positive. The output of thezero-crossing detector is a series of pulses. These pulses are passed through a low-pass filter toremove the high frequency component. The filter pass frequencies from 0 to 25 Hz in order toreproduce the frequencies of interest in the flow pulse. The zero-crossing detector is capable ofmeasuring blood velocity to within 20% and can detect changes in velocity of about 5%. A majorlimitation of the zero-crossing detector used in simple flowmeters is that it cannot detect thedirection of flow. A better approach is to use the quadrature-phase detector approach, which iscommonly used in radar technology. This detector gives not only the velocity with which theblood flows but also its direction.
Spectrum Analyzers: Because of the limitations of the zero-crossing detector, alternative methods
have been employed for the analysis of Doppler signals for velocity measurements. For example,spectrum analyzers are used to derive blood flow velocity information from Doppler signals. Aspectrum analyzer processes short lengths of audio signal to produce spectral displays, which
have frequency as the abscissa, time as the ordinate and spectral intensity represented by record
darkening.
Spectral analysis of Doppler signals with a spectrum analzser is an off-line technique. The
process of obtaining a good record involves aural recognition of the relevant Doppler wave forms,the recording of these on tape and the subsequent analysis of short sections at a time. In order tostudy effects extending over several cardiac cycles and also to allow the operator to optimize therecording during clinical investigation, a real-time continuous form of analysis is needed.Macpherson et al (1980) describe a real-time spectrum analyzer for ultrasonic Doppler signals.
The analyser calculates spectral components using a type of Discrete Fourier Transform (DFT)known as the Chirp-Z-Transform.
Development of a real-time system for the three dimensional display of Doppler-shifted
frequencies within an ultrasound flowmeter signal is described by Rittgers et al (1980). The device
is based on a Fast Fourier Transform (FFT) spectrum analyzer, which provides both high frequencyand time resolution over a variable range of frequencies up to 40 kHz. Signal pre-processing filtersout unwanted low frequency artefacts due to arterial wall motion and automatically controlsinput gain to limit peak amplitudes while matched phase shift networks translate forward and
reverse components for a directional output.
Ultrasound directional Doppler blood-velocity detection may be achieved using any of the
three basic detection systems. These are single sideband, heterodyne and phase-quadrature

Blood Flowmeters 337
detection systems. Out of these, phase-quadrature detection system is the most commonly used.
Phase-quadrature detection systems operate by detecting both the amplitude and phase of theDoppler-shifted frequency.
Schlindwein et al (1988) used a digital signal processor chip to perform the online spectrum
analysis of the Doppler ultrasound signals. It is a flexible and a programmable method, which has
every aspect of the system such as window function, frequency ranges and the transform rate
under software control.
/G31/G31/G2E/G33/G2E/G33 /G42/G6C/G6F/G6F/G64/G20/G46/G6C/G6F/G77/G20/G4D/G65/G61/G73/G75/G72/G65/G6D/G65/G6E/G74/G20/G74/G68/G72/G6F/G75/G67/G68/G20/G44/G6F/G70/G70/G6C/G65/G72/G20/G49/G6D/G61/G67/G69/G6E/G67
Doppler ultrasound is not only used for the measurement of the absolute value of blood velocityand volume flow, but it also helps to have a direct visualization of the blood vessels and to studythe blood-velocity/time wave form shapes over the cardiac cycle. The imaging facility helps tomeasure beam/vessel angle to detect the location of the required site of measurement of velocity
and flow. This technique is particularly attractive in studying blood flow in carotid and femoral
arteries and to visualize the bifurcation of the former.
The block diagram of an imaging instrument (Fish, 1975) using a pulsed flow detector is shown
in Fig. 11.7. The probe is mechanically coupled to the position resolvers, which gives electricaloutputs proportional to the movements of the probe. The position of the spot on the CRT screencorresponds to that of the sampling volume. When the sampling volume is within a blood vessel,the spot is intensified and stored when Doppler signals are received. A vessel is imaged by movingthe probe over the skin surface and by adjusting the probe/sampling-volume distance untilthe sampling volume has been swept through the section of the vessel of interest. The three-dimensional information about the geometry of the vessel is displayed by selecting, in turn, two ofthe three dimensions available for display on the monitor. Thus, it is possible to construct anterior-posterior, lateral and cross-sectional scans of blood vessels.
Position
computerCoordinate
selector
Ultrasonic
flow detector
Storage monitorVessel
image
VesselSkin
BeamProbeProbe
position
resolver
Arm
movements
Fig.11.7 Doppler imaging system-block diagram (redrawn after Fish, 1975)
Obviously, the time needed to complete a single cross-sectional scan of a blood vessel is long
(about 3 min) even after the vessel is located. An additional difficulty, if the flow detector is not
direction-resolving, is that of separating the images of contiguous veins and arteries. In order to

338 Handbook of Biomedical Instrumentation
overcome these difficulties, Fish (1978) developed a multi-channel direction-resolving flow
detector. Although the instrument has the same basic form as the single channel instrument, it isa 30 channel device and can determine the direction of the component of flow along the beam. The
instrument can map flow towards and away from the probe and, therefore, permits separation of
the images of the vein and artery when these are contiguous. The depth of the vessel can bemeasured by writing the position of the skin surface on the screen by intensifying a spot on thescreen corresponding to the end of the probe.
The imaging facility can aid the fixing and maintenance of sample-volume/vessel geometry.
Figure 11.8 shows lateral, anterior-posterior and cross-sectional views of a length of the femoralartery, together with recordings of relative velocity from three points on a diameter. The position ofthe 30 sampling volumes is clear from the lateral and cross-sectional views. From the images, it isquite convenient to see where the measurement is being made and to maintain the samplingvolumes in position during the measurement. As it is possible to record velocity profiles over thecardiac cycle, it is possible to compute mean blood flow, peak flow, peak reverse flow by samplingmeasurements at intervals, say 40 ms. The system computes flow information from the velocityprofiles and cross-sectional diameter of the blood vessel under investigation. The transducer
A-P
Cross-section VelocityLateral
Fig.11.8 Lateral, anterior-posterior and cross-section views of a length of femoral
artery together with recordings of relative velocity from three points on a
diameter of blood vessel

Blood Flowmeters 339
frequency is usually 5 MHz and it transmits ultrasound beam focussed at 2–4 cm, which is the
approximate depth of the carotid or femoral artery beneath the skin surface. The signal processingcircuit generates simultaneous forward and reverse flow-velocity channels by the use of a phase-
quadrature detection system and the processor scans the instantaneous blood-velocity spectrum
as it is formed. The circuit rejects low frequency signals due to arterial wall movement andspurious signals due to noise or transducer movement artefacts by adjusting the threshold forDoppler signals.
Colour Doppler sonography is based on the measurement of the local flow velocity in real-time
and the display the surrounding structures in colour coded form, together with the section scan ofthe vessel walls. The colour, usually red or blue, indicates the direction of blood flow. The colourintensity indicates the local flow velocity. The brighter the colour intensity, the higher the velocity.The image gives an immediate overview of flow direction and flow dynamics, as well as ahaemodynamic effect of alterations to the vascular structures.
One of the interesting applications of using pulsed Doppler ultrasound is to measure intra-
cranial blood flow-velocities and their related pathologies. This highly sensitive techniqueprovides a window to the brain for access to important clinical information previously unavailablewith non-invasive techniques. The ultrasound probes used for blood flow-velocity range from 2MHz to 8 MHz, and 20 MHz for micro-vascular applications. Figure 11.9 shows a transcranialDoppler system from M/s Nicolet, USA.
Fig.11.9 Transcranial Doppler blood flow measuring machine (Courtesy: M/s
Nicolet, USA)

340 Handbook of Biomedical Instrumentation
/G20/G31/G31/G2E/G34 /G4E/G4D/G52/G20/G42/G4C/G4F/G4F/G44/G20/G46/G4C/G4F/G57/G4D/G45/G54/G45/G52
Nuclear magnetic resonance principle offers yet another non-invasive method for the measurement
of peripheral blood flow or blood flow in various organs. The method pertains to a quantummechanical phenomenon related to the magnetic energy levels of the nucleus of some elements
and their isotopes. For blood flow measurement work, behaviour of the two hydrogen atoms of
water is studied, since blood is approximately 83% water. Due to the magnetic moment of thehydrogen atom, the nucleus behaves as a microminiature magnet which can be affected byexternally applied magnetic fields. The hydrogen nuclei orient themselves to produce a net parallelalignment to a steady magnetic field. The nuclear magnets precess around the magnetic fieldlines until they become aligned. The angular frequency (Larmor frequency) of this precession isgiven by
W = 2 pv= rB
o
where r is the ratio of the magnetic moment to the angular momentum (the magnetic gyro ratio) and
Bo is the density of the steady magnetic field and v is the frequency of radiation.
When the frequency of the magnetic field is near the Larmor frequency, the nuclear magnets can
be rotated and their presence detected secondary to an externally applied radio frequency magneticfield. Also, a region of demagnetization can be made by disorienting the nuclear magnets with aradio frequency field very close to the Larmor frequency.
An arrangement for the measurement of blood flow using the NMR technique is shown in
Fig. 11.10. The blood vessel of interest is positioned in a uniform steady magnetic field B
o.The
nuclear magnets of the hydrogen atoms, which before insertion were randomly oriented, nowbegin to align themselves with B
o. Some begin to align parallel, whereas others commence to
align anti-parallel. The alignment occurs exponentially with a time constant T1 known as the
longitudinal relaxation time. There are more nuclear magnets aligned parallel to Boin contrast
to those aligned anti-parallel. For blood at 37°C, there are 0.553 ¥ 1023 hydrogen nuclei per
N
S
Magnetizer ( )
permanent magnetB0Detector DC field
()
electromagnetBDLumen
Receiver coil
Modulation coilTransmitter
coil
Fig.11.10 Arrangement for measurement of arterial blood flow using NMR principle.
Hydrogen nuclei in the blood are magnetised by the magnetiser Bo. The nulcei
are detected downstream by the receiver coil

Blood Flowmeters 341
millilitre. For Bo = 1000 gauss, there are 1.82 ¥ 1016per millilitre net nuclei aligned. The maximum
magnetization Mo due to the magnetic field Bois given by Mo = Xo Bo, where Xo is the static nuclear
magnetic susceptibility equal to 3.23 ¥ 10–9for blood at 37°C. Thus for Bo = 1000 gauss, Mo = 3.23
microgauss.
The magnitude of the magnetization can be related to either flow velocity or flow rate. Since the
magnetization can be changed at some point in a magnetic field, and the change detected at some
distal point, it is possible to determine the transport time between the two points. A crossed coilconfiguration is used to detect the level of magnetization in the limb. Two magnets are used, astrong permanent magnet B
o for premagnetization and a weaker, homogeneous electromagnet BD
for detection. A graph plotted between the magnetization at the centre of the receiver coil for
typical distances as a function of velocity shows that the magnetization changes proportionally tovelocity and thus a measurement of the magnetization can yield velocity information. If a receivercoil is used, the voltage induced in the coil by the magnetization is proportional to the cross-sectional area of the vessel carrying the blood. The NMR signal voltage proportional to velocity V,
and multiplied by area A, will give a volumetric flow rate Q. With a homogeneous detector field B
D,
the magnetization in all the blood vessels in the region is detected. In a non-homogeneous field, BD
can be tuned to detect the magnetization of selected vessels. The major shortcoming of NMR blood
flowmeters is the physical size of the magnets and the sensitivity of the system needed to detect the
magnetically tagged bolus of blood.
For the measurement of arterial blood flow, the blood in the upper arm is magnetized by the
niagnetizer BO (Fig. 11.10) and the magnetized blood is detected in the forearm by the receiving
coil. The NMR signal waveform will be pulsatile because the arterial flow is pulsatile. For venousflow measurement, the arm is inserted in an opposite direction to that shown in the figure. Sincevenous flow is relatively steady, the magnetization at the receiver coil is also steady. For cerebralflow studies, a 78 cm wide, 45 cm high, 44 cm deep permanent magnet system of 540 gauss field isused. The size is large enough to accommodate a human head. NMR based flowmeters are limitedin their applications to the measurement of blood flow in limbs, since the part of the body to bemeasured must be located inside the lumen of a cylinder.
/G20/G31/G31/G2E/G35 /G4C/G41/G53/G45/G52/G20/G44/G4F/G50/G50/G4C/G45/G52/G20/G42/G4C/G4F/G4F/G44/G20/G46/G4C/G4F/G57/G4D/G45/G54/G45/G52
A system utilizing the Doppler-shift of monochromatic laser light to measure blood flow in skin is
described by Watkins and Holloway (1978). When a laser beam is directed towards the tissue
under study, absorption and scattering takes place. Radiation scattered in movable structures,such as red cells, is shifted in frequency due to the Doppler effect, while radiation scattered in non-moving soft tissue is unshifted in frequency. A part of the total scattered radiation is brought to fallon the surface of a photodetector. The effective radiation penetration depth is approximately 1 mmin soft tissue and scattering and absorption take place mostly in the papilla region and theunderlying corium—two dermal layers containing the capillary network of the skin.
In principle, light from a low power (5 mW) He–Ne laser is coupled into a quartz fibre and
transmitted to the skin. The light is reflected from both the non-moving tissues (reference beam)and moving red blood cells (Doppler-shifted beam). The two beams are received by a plastic fibre

342 Handbook of Biomedical Instrumentation
and transmitted back to a photo-diode where optical heterodyning takes place. The heterodyned
output signal which is proportional to the Doppler shift frequency is amplified and both RMS anddc values are calculated. The RMS value is weighted against the back-scattered light intensity
using the measured dc value as an index of total received power. This gives the output flow
velocity.
Figure 11.11 shows a block diagram of the laser Doppler system. The He–Ne laser operating at
632.8 nm wavelength is used. The laser output is coupled into the fibre using a converging lens,which results in an increased power density at the skin surface and thus enables the detection offlow in the more deeply seated veins and arteries. The receiving fibre is coupled to the photodiodethrough a laser line filter. The photodetector functions as a square law device and gives outcurrent, which is proportional to the intensity of the incident light and, therefore, to the frequencyof beating of the shifted and unshifted signals. The light falling on the photodetector is an opticallymixed signal involving a Doppler-shifted signal back scattered from the moving red blood cellswith the ‘reference’ signal reflected from the non-moving skin surface. The diode is connected in aconfiguration so that it gives wideband performance (dc–100 kHz). The amplifier is constructedusing a standard operational amplifier with a low noise preamplifier. System output is obtainedby taking the RMS value of the total signal, separating it from the total zero light noise, and
normalizing it for total back scattered light.
Helium
neon laser
5 mW
632.8 mm
Wideband
amplifierPreamp.
40 Hz–40 kHzPhoto
diode
High-pass
filter (300 Hz)
Low-pass
filter
25 kHzRMS
converterBuffer
BufferDigital
read out
Audio outputSkinPlastic
receiving
fiberQuartz laser
transmitting
fibre
Fig.11.11 Block diagram of a laser Doppler system for blood flow measurement
in skin (after Watkins and Holloway, 1978 by permission of IEEE Trans.
Biomed. Eng.)
An audio output of the signal before RMS conversion is also produced to ‘hear’ the flow pattern.
It may be noted that the instrument measures an averaged red blood cell velocity and not true flow,as both the cross-sectional area of flow and the angle the incident light beam makes with each
capillary are not known.

Blood Flowmeters 343
Laser Doppler flowmetry is a non-invasive technique and seems to offer several advantages
like light reproducibility and sensitivity. However, its disadvantages like poor selectivity, baseline instability and restriction in site of measurement are still limiting factors in its successful
clinical utilization. The blood cells move through the capillaries at about 1 mm/s, which in tissue
is approximately 2.10
11 mm/s. The Doppler-shift will be the same fraction of the light frequency.
To make adequate measuring possible under such conditions, the noise level of the entire system,particularly of the light source and the photo-elements, must be kept extremely low, lower thanwhat can normally be achieved with ordinary photodetectors and low power lasers.
The laser probe is easily adapted to suit different applications. Its main part, the core, is a thin
piece of stainless steel tubing (diameter 1.5 mm and length 40 mm), in which the terminating endsof the three fibres—one efferent and two afferent—are moulded together. For easier handling, thecore can be inserted into a tight fitting plastic sleeve. The core can also be inserted into a thermostathead where the core-end forms the centre of a 15 mm diameter contact surface, the temperature ofwhich can be set from 25 to 40°C.
Evaluation of the laser Doppler flowmeter has shown that a linear relationship exists between
flowmeter response and the flux of red cells, with red cell velocities and volume fractions withinthe normal physiologic range of the microcirculatory network of the skin. It has also been foundthat different degrees of oxygenation influenced the Doppler signal only to a minor extent.

344 Handbook of Biomedical Instrumentation
/G43/G61/G72/G64/G69/G61/G63/G20/G4F/G75/G74/G70/G75/G74/G20/G4D/G65/G61/G73/G75/G72/G65/G6D/G65/G6E/G74
Cardiac output is the quantity of blood delivered by the heart to the aorta per minute. It is a major
determinant of oxygen delivery to the tissues. Obviously, problems occur when the supply ofblood from the heart is unable to meet the demand. A fall in cardiac output may result in low bloodpressure, reduced tissue oxygenation, acidosis, poor renal function and shock. It is a reflection ofthe myocardial function and when taken with other measurements such as blood pressure andcentral venous pressure, the rational treatment of cardiac disorders becomes clearer. Stroke volumeof blood pumped from the heart with each beat at rest varies among adults between 70 and 100 ml,
while the cardiac output is 4 to 6 l/min.
The direct method of estimating the cardiac output consists in measuring the stroke volume by
the use of an electromagnetic flow probe placed on the aorta and multiplying it by the heart rate.
The method involves surgery and, therefore, is not preferred in routine applications. Another wellknown method for measuring cardiac output is the Fick’s Method, which consists in determiningthe cardiac output by the analysis of the gas-keeping of the organism. Even this method is rathercomplicated, difficult to repeat, necessitates catheterization and, therefore, cannot be consideredas a solution to the problem, though it is practised at many places even now. The most popularmethod group is the one applying the principle of indicator dilution.
/G20/G31/G32/G2E/G31 /G49/G4E/G44/G49/G43/G41/G54/G4F/G52/G20/G44/G49/G4C/G55/G54/G49/G4F/G4E/G20/G4D/G45/G54/G48/G4F/G44
Indicator dilution principle states that if we introduce into or remove from a stream of fluid aknown amount of indicator and measure the concentration difference upstream and downstreamof the injection (or withdrawal) site, we can estimate the volume flow of the fluid. The method
employs several different types of indicators. Two methods are generally employed for introducing
the indicator in the blood stream, viz: it may be injected at a constant rate or as a bolus. The methodof continuous infusion suffers from the disadvantage that most indicators recirculate, and thisprevents a maxima from being achieved. In the bolus injection method, a small but known quantityof an indicator such as a dye or radioisotope is administered into the circulation. It is injected intoHAPTER
1212

Cardiac Output Measurement 345
a large vein or preferably into the right heart itself. After passing through the right heart, lungs and
the left heart, the indicator appears in the arterial circulation. The presence of an indicator in theperipheral artery is detected by a suitable (photoelectric) transducer and is displayed on a chart
recorder. This way we get the cardiac output curve shown in Fig. 12.1. This is also called the
dilution curve.
TimeConcentration
Fig.12.1 The run of the dilution curve
The run of the dilution curve is self-explanatory. During the first circulation period, the
indicator would mix up with the blood and will dilute just a bit. When passing before thetransducer, it would reveal a big and rapid change of concentration. This is shown by the risingportion of the dilution curve. Had the circulation system been an open one, the maximum con-centration would have been followed by an exponentially decreasing portion so as to cut the timeaxis as shown by the dotted line. The circulation system being a closed one, a fraction of theinjected indicator would once again pass through the heart and enter the arterial circulation. Asecond peak would then appear. When the indicator is completely mixed up with blood, the curvebecomes parallel with the time axis. The amplitude of this portion depends upon the quantity ofthe injected indicator and on the total quantity of the circulating blood.
For calculating the cardiac output from the dilution curve, assume that
M= quantity of the injected indicator in mg
Q= cardiac output
then Q=
M
average concentration of indicator per curve duration
litre of blood for duration of curve in secondsl/s¥◊
=M¥60
area under the curve l/min
Suppose that 10 mg of the indicator was injected and the average concentration as calculated
from the curve was 5 mg/l for a curve duration of 20 s; then Q =6 I/min.

346 Handbook of Biomedical Instrumentation
The area under the primary curve obtained by the prolongation of the down slope exponential
curve to cut the time axis, encloses an area showing the time concentration relationship of theindicator on its first passage round the circulation and does not include any of the subsequent
recirculations. It demands a considerable time to perform the exponential extrapolation for
calculating the area. The evaluation of the dilution curve is simplified by replotting the curve on asemilogarithmic scale paper. The indicator concentration (Y-axis) is plotted on a logarithmic scaleand the time (X-axis) on a linear scale. The decreasing exponential portion of the curve appears asa straight line, which is projected downwards to cut the time axis. The area under the replottedprimary dilution curve is then measured either with a mechanical planimeter or by counting the
square units under the curve. It can be approximated by summing the indicator concentration
occurring at one second intervals from the start to the end of the curve.
/G20/G31/G32/G2E/G32 /G44/G59/G45/G20/G44/G49/G4C/G55/G54/G49/G4F/G4E/G20/G4D/G45/G54/G48/G4F/G44
The most commonly used indicator substance is a dye. Fox and Wood (1957) suggested the use ofIndocyanine green (cardiogreen) dye which is usually employed for recording the dilution curve.
This dye is preferred because of its property of absorbing light in the 800 nm region of the spectrum
where both reduced and oxygenated haemoglobin have the same optical absorption. While usingsome of a the blue dyes, it was necessary to have the patient breathe oxygen. The concentration ofcardiogreen can be measured with the help of infra-red photocell transducer. Dye cuvettes of assmall volume as 0.01 ml are available.
The procedure consists in injecting the dye into the right atrium by means of a venous catheter.
Usually 5 mg of cardiogreen dye is injected in a 1 ml volume. The quantity used may be 2.5 mg inthe case of children. A motor driven syringe constantly draws blood from the radial or femoralartery through a cuvette. The curve is traced by a recorder attached to the densitometer. After thecurve is drawn, an injection of saline is given to flush out the dye from the circulating blood.
There are problems relating to the use of the indicator indocyanine green. It has been experi-
mentally determined that above a dye concentration of approximately 20 mg/ml of blood, the
optical density rises less with an increase in dye concentration than below this level (Chamberlain,1975). Thus for optimum accuracy, the amount of dye chosen for injection should result in dyecurves whose peak concentration is less than 20 mg/ml.
Figure 12.2 shows a diagrammatic representation of a densitometer which can be used for the
quantitative measurement of dye concentration. The photometric part consists of a source ofradiation and a photocell and an arrangement for holding the disposable polyethylene tubeconstituting the cuvette. An interference filter with a peak transmission of 805 nm is used to permitonly infrared radiation to be transmitted. This wavelength is the isobestic wavelength for haemo-globin (Jarlov and Holmkjer, 1972) at various levels of oxygen saturation. In order to avoid the
formation of bubbles, the cuvette tubing should be flushed with a solution of silicone in ether. A
flow rate of 40 ml/min is preferred in order to get as short a response time as possible for thesampling catheter. The sampling syringe has a volume of 50 mi/min. The output of the photocellis connected to a low drift amplifier. It has a high input impedance and low output impedance.The amplification is directly proportional to the resistance value of the potentiometer R. Apotentiometric recorder records the amplifier signal on a 200 mm wide recording paper and apaper speed of 10 mm/s.

Cardiac Output Measurement 347
In the recording of dye dilution curves, it is generally necessary that the densitometer be at some
point removed from the site of interest. A catheter is used to transport the blood containing dyefrom the sampling site, inside the cardiovascular system, to the densitometer located outside thebody. Sampling through the catheter densitometer system distorts the concentration time curve.First, the velocity of flow within the catheter is not uniform, which causes the dye to mix within thetube as it travels downstream. The mixing is a function of the flow rate and the volume of thesampling system, the viscosity of the sampled fluid and the shape of the configuration of thesampling tube. The second source of distortion is the measuring instrument itself, which may nothave response characteristics fast enough to record instantaneous dye concentration as it actuallyoccurs in the lumen. Distortion is very important when the indicator dilution method is used tomeasure volume since it is the measurement of the mean transit time of an indicator from the point
of injection to the point of sampling, which is of interest. To reduce distortion, computer software-
based corrections have been devised.
/G20/G31/G32/G2E/G33/G54/G48/G45/G52/G4D/G41/G4C/G20/G44/G49/G4C/G55/G54/G49/G4F/G4E/G20/G54/G45/G43/G48/G4E/G49/G51/G55/G45/G53
A thermal indicator of known volume introduced into either the right or left atrium will produce aresultant temperature change in the pulmonary artery or in the aorta respectively, the integral ofwhich is inversely proportional to the cardiac output.
Cardiac output =
"a constant" (blood temp. injectate temp.)
area under dilution curve¥-
Although first reported by Fegler (1954), thermal dilution as a technique did not gain clinical
acceptance until Branthwaite and Bradley (1968) published their work showing a good correlation
between Fick and thermal measurement of cardiac output in man. However, the technique ofA
RecorderPhotocell
50 ml
syringeFilter6 V
lamp
+–Cuvette4 ml/min
Fig.12.2 Diagrammatic representation of a densitometer for quantitative measurement
of dye concentration (redrawn after Jarlov and Holmkjer, 1972; bypermission of Med.& Biol. Eng.)

348 Handbook of Biomedical Instrumentation
cannulation of the internal jugular vein and the difficulty of floating small catheters into the
pulmonary artery prevented a rapid clinical acceptance of the technique.
In 1972, a report appeared in the American Heart Journal describing a multi-lumen thermistor
catheter, known today as the Swan-Ganz triple lumen balloon catheter (Ganz and Swan, 1972).The balloon, located at or near the tip, is inflated during catheter insertion to carry the tip through
the heart and into the pulmonary artery. One lumen terminates at the tip and is used to measure
the pressure during catheter insertion. Later, it measures pulmonary artery pressure and inter-mittently, pulmonary–capillary wedge pressure. A second lumen typically terminates in the rightatrium and is used to the monitor right atrial pressure (central venous pressure) and to inject thecold solutions for thermal dilution. A third lumen is used to inflate the balloon. For use withthermal dilution, the pulmonary-artery catheter carries a thermistor proximal to (before) the balloon.The thermistor is encapsulated in glass and coated with epoxy to insulate it electrically from theblood. The wires connecting the thermistor are contained in a fourth lumen (Fig. 12.3). This cathetersimplified the technique of cardiac cannulation making it feasible to do measurements not only in
Proximal
(right atrial)
side hole
Distal
end holeDistal
port
Balloon
inflatedThermistor
Proximal
port Balloon
inflation port
Thermistor
connector
Distal
end holeThermistor
Fig.12.3 Swan-Ganz Catheter – A 4-lumen catheter. Distal Lumen – connects
to transducer system to monitor (i) pulmonary artery pressure (ii)Wedge pressure with balloon inflated. Inflation Lumen – connects toballoon located approximately 1 mm from catheter-tip. Ballooninflates with 1 to 1.5 ml of air. Proximal lumen – to monitorcentral venous pressure or right atrium pressure. ThermistorLumen – connects with cable to cardiac output computer.

Cardiac Output Measurement 349
the catheterization laboratory but also in the coronary care unit. The acceptance of the thermal
dilution technique over the past few years can only be attributed to the development of thiscatheter.
Figure 12.4 shows a typical cardiac output thermal dilution set up. A solution of 5% Dextrose in
water at room temperature is injected as a thermal indicator into the right atrium. It mixes in the
right ventricle, and is detected in the pulmonary artery by means of a thermistor mounted at the tip
of a miniature catheter probe. The injectate temperature is also sensed by a thermistor and thetemperature difference between the injectate and the blood circulating in the pulmonary artery ismeasured. The reduction in temperature in the pulmonary artery (due to the passage of theDextrose) is integrated with respect to time and the blood flow in the pulmonary artery is thencomputed electronically by an analog computer which also applies correction factors. A meterprovides a direct reading of cardiac output after being muted until integration is complete so as toavoid spurious indications during a determination.
Sup. vena cava
Thermistor
LV
RVIndicator
injection
siteRA
Fig.12.4 Cardiac output thermal-dilution set-up
The electronic computation is relatively simple, because there is no significant recirculation of
the indicator in man. The calculation rests upon the integral of the inscribed curve, the restingtemperature in the pulmonary artery, the temperature of the injectate, and a number of constants.Absence of the need to subtract that part of the area under the curve due to recirculation, and theease with which an unsatisfactory curve can be detected by failure to return to the original baselinevalue of temperature, contribute to the internal consistency of the results.
The system calibration is based upon the use of an injection of 10 ml of 5% Dextrose solution at
a temperature in the range of 18–28°C. Within this range, the injectate temperature is measured toan accuracy of ± 0.2°C , and is also displayed on a meter.
Blood temperature is measured over a range of 30 to 40°C to an accuracy of ±0.2°C. During a
determination the incremental temperature is automatically derived, relative to a baseline valueequal to the blood temperature, immediately before starting the determination. The incrementaltemperature is measured and displayed in the range 0-1°C full scale to an accuracy of ± 0.02°C.

350 Handbook of Biomedical Instrumentation
The sensitivity of the measuring system is adjusted in such a way that a temperature variation
of 0.3–0.5°C (normally encountered temperature fluctuation at the measuring point), measured bythe catheter tip, gives a full scale deflection on the recorder. The total quantity injected is normally
10 ml which gives a full scale deflection with the aforesaid sensitivity of the recorder. The
thermodilution curve can be recorded on an electrocardiograph machine. The normal paper speedis 10 mm/s and the normal curve amplitude is 30–50 mm.
/G31/G32/G2E/G33/G2E/G31 /G43/G6F/G6D/G70/G75/G74/G69/G6E/G67/G20/G53/G79/G73/G74/G65/G6D
Assuming that the injectate mixes thoroughly with the blood and that negligible net heat flowoccurs through the vessel wall during the passage of the blood injectate mixture between thepoints of injection and measurement, the heat (or coolth) injected can be equated to the heat
(coolth) detected. If the volume of injectate is small as compared with the volume of the blood it
mixes with, it can be shown that within a close approximation,
V D
i Si (Ti – Tb) = Q Db Sb DTd tz
which, when rearranged gives
Q=VT T
Td tib()-zD◊DS
DSii
bb
where Q= volumetric flow
V= volume of injectate
T= temperature
DT= incremental temperature of the blood injectate mixture
D= density
S= specific heat
The suffixed i and b refer to the injectate and blood respectively. The thermal dilution curve can
be validated to ensure that the downslope is exponential and that the area under an acceptablecurve is determined by integration. The equation for determining the cardiac output by this methodthus reduces to:
Cardiac output =
( . )( )( )( )( )10 8 6 0 CV T T
Td tib-zD
where: 1.08 is the ratio of the products of specific heats and specific gravities of 5% dextrose in
water and blood
C= 0.827 for 10 ml injectate at ice temperature (0 to 2°C)
= 0.747 for 5 ml injectate at ice temperature= 0.908 for 10 ml injectate at room temperature (22 to 26°C)
= 0.884 for 5 ml injectate at room temperature
V= volume of injectate (ml)
T
b= initial temperature of blood (°C)

Cardiac Output Measurement 351
Ti= initial temperature of injectate (°C)
DzTd t = integral of blood temperature change (°C.s).
Figure 12.5 shows a block diagram of the thermal dilution system. The linearizing amplifier
works on the principle that when a fixed resistance of suitable value is placed in parallel with athermistor, a virtually linear relation between the temperature and resistance of the parallelcombination can be obtained over a limited range of operation. The integrator responds to a DT
signal corresponding to a maximum temperature change of typically 0.3°C with reference to a T
b
baseline, which itself may suffer medium term variations of a similar amount and may additionally
vary from one patient to another over a range of 30-40°C. Also, the circuit must hold the integralobtained so that the value of the cardiac output computed from it may be displayed.
Timer/
control
Integ.
Preset
adjust
constantComputerCardiac
outputLinearizing
amplifierBlood
temp.
thermistor
Injectate
temp.
thermistorLinearizing
amplifier
Fig.12.5 Block diagram of the processing and computing circuit of thermal dilution
method (redrawn after Cowell and Bray, 1970)
The summing and multiplication/division operations required for the evaluation of cardiac
output are performed by a simple analog computing circuit. The timer control unit generates theswitching signals necessary for the proper integration of the DTsignal and for the display of the
computed value of the cardiac output. The cardiac output is displayed in two ranges: 0–10 l/minand 10–20 l/min.
Philip et al (1984) used a resistive element in a modified Swan-Ganz catheter and energized
it with a periodic electrical waveform. The resulting thermal signal is diluted by the bloodflow and sensed by a thermistor in the pulmonary artery. The thermal signal is processed by a

352 Handbook of Biomedical Instrumentation
microprocessor-based instrument. This technique avoids the introduction of fluids and thereby
reduces the risks of fluid overload and sepsis (Bowdle et al, 1993).
Neame et al (1977) illustrated the construction of thermistor probes intended for use in cardiac
output estimations by the thermodilution method. Ideally, the probes should have infinitesimalthermal capacity and hence an infinite frequency response in order to reproduce thermal transients
accurately; minimal size so as to cause no flow obstruction, produce negligible heat dissipation, a
linear temperature response and an infinite working life. Bead type of thermistors are available in
the smallest form and are thus suited to applications for cardiac output measurements.
Although the excellent linearity characteristics of thermocouples and their fast dynamic
response, as opposed to the slower non-linear thermistor beads, make them more desirable, the
inherent low sensitivity of thermocouples, however, provides a severe limitation. Compared tothermistors excited by a current below the effective self-heating range, thermocouples are twoorders of magnitude less sensitive than thermistors. For example, temperature coefficient is 0.04mV/°C for chromel-constantan as compared to 3 mV/°C for a 2000 W thermistor with a dissipation
constant of 0.2 mW/°C excited by 0.1 mA. Flow directed precalibrated thermistor catheters are
about 100 cm long and contain a thermistor located 4 cm from the distal tip. The thermistor resis-
tance, typically, may be 7000 Wat 36.6°C, with a temperature coefficient of resistance being 250 W
per °C ±5%. The injectate orifice is 26 cm proximal to the thermistor.
Basic source of inaccuracies in thermo-dilution cardiac output measurements arise with imperfect
indicator mixing. Pneumatically controlled injectors are used for rapid and consistent injections.
These injectors are powered by compressed air or CO
2 from a pressurized cylinder or a hospital air
system. A two-way automatic valve between the injection syringe and the catheter make possiblea fully automatic operation with injection and automatic refill. Injection speed is 7 ml/s viastandard 7F triple luman thermistor catheter. Both injection and refill speed are usually keptadjustable. Slow return speed minimizes chances of air bubbles formation during refilling.Injection volume is also adjustable in the range 1 to 10 ml.
As blood flow is pulsatile, injection can be done in any part of the heart cycle. Therefore, time
allowed for indicator mixing can vary as much as a full heart cycle. In addition, dilution curveobtained in the pulmonary artery will look different if injection was done in diastole or systole ofthe heart. Optimum injection moment is during diastole and, therefore, injectors are equippedwith ECG synchronizer to ensure injection at the right moment.
ECG synchronizer consists of an ECG amplifier with adjustable gain. A threshold detector
senses appearance of R wave and the interval between two adjacent R waves is measured and
stored in a the memory. Delay of injection is set manually in a percentage of the cardiac cycle, andnot in milliseconds as in the majority of angiographic injectors. The method of presetting delay ofinjection in a percentage of the cardiac cycle makes the synchronizer insensitive to variations in
heart rate.
An electrosurgical unit emits radio frequency energy that can distrupt or scramble the thermo-
dilution temperature curve and result in wildly inaccurate results. Such a disruption can causemajor problems for the anaesthesiologist measuring cardiac output intra-operatively. The effectof high frequency interference on the measurements becomes readily apparent by an inspection
of the thermodilution temperature. A visual display of the thermodilution curve is therefore
necessary.

Cardiac Output Measurement 353
/G20/G31/G32/G2E/G34 /G4D/G45/G41/G53/G55/G52/G45/G4D/G45/G4E/G54/G20/G4F/G46/G20/G43/G4F/G4E/G54/G49/G4E/G55/G4F/G55/G53/G20/G43/G41/G52/G44/G49/G41/G43/G20/G4F/G55/G54/G50/G55/G54/G20/G44/G45/G52/G49/G56/G45/G44
/G46/G52/G4F/G4D/G20/G54/G48/G45/G20/G41/G4F/G52/G54/G49/G43/G20/G50/G52/G45/G53/G53/G55/G52/G45/G20/G57/G41/G56/G45/G46/G4F/G52/G4D
It is not always possible, particularly in critically
ill patients, to estimate cardiac output on a beat-
to-beat basis. Thermal dilution measurementscan be repeated more frequently than the othermethods, which has made this method popular,but even then the measurements can only be madeat certain intervals. The technique also requiresthe presence of well-trained operators for obtain-ing reliable measurements. These problems arecircumvented in an instrument developed byWessling et al. (1976a), which is based on theanalysis of the pressure pulse contour illustratedin Fig. 12.6.
The method is based on the analysis of the aortic
pressure wave, and estimates the left ventricular
stroke volume from the pulse contour. The analysis
of the pressure wave depends on simple hydraulic relationships between flow, pressure and time.During the ejection phase flow is ejected into the aorta, the total amount depending on the drivingpressure, on the duration of the ejection period and on the impedance to flow in the aorta(Wesseling et al, 1974). Expressed mathematically:
Stroke volume =
1
ZPP d t
aoao edToTE() -z
This equation shows that the area ‘ A’ under the systolic portion of the aortic pressure wave is
integrated. This area ‘ A’ is divided by the patient calibrator (Zao: aortic characteristic impedance)
to obtain a quantity approximating the left venticular stroke volume. Thus, changes in this area ‘ A’
reflect corresponding changes in stroke volume. The instantaneous heart rate is derived as theinverse of the time lapse between the onset of ejection for every two consecutive pressure pulses.Cardiac output is computed for each beat as the instantaneous product of stroke volume and heartrate:
Stroke volume (SV) =
A
Zao (cm3)
Heart rate = 60
T(bpm)
Cardiac output = () ( )SV HR ¥
1000=60
1000◊
◊A
ZTao (l/min)
The instrument uses pattern recognition techniques to detect the onset ( To) and the end ( TE) of
ejection, and integrates the difference between the actual aortic pressure and the end diastolic
pressure over that time period. The average value of Zao is found to be 0.140 for adults.Pressure
TimePedTA
SystolePtao()
Fig.12.6 Pressure pulse contour method
of measuring beat-to-beat car-diac output

354 Handbook of Biomedical Instrumentation
In this technique, all that is required is one pressure measurement at any site in the aorta. To
measure the pressure signal, both catheter tip manometers as well as ordinary catheter manometersystems can be used. It should be ensured that pressure is measured in the aorta or near the aorta
proximally in a major side branch such as the subclavian artery. Other pressure signals are much
distorted and reduce the accuracy of the method. For reliable operation of the computing circuit,the resonant frequency of the catheter manometer system must be 25 Hz or higher. If no artefactdistorts the pressure waveform, the resonance frequency may be as low as 15 Hz.
The cardiac output computer consists of three modules: (i) the pressure transducer amplifier,
(ii) the cardiac output computer, (iii) dual numerical display and alarm limits. Quantities, whichcan be displayed, are mean systolic and diastolic pressure, cardiac output, stroke volume andpulse rate. Two parameters must be set initially for each patient: (i) patient age and (ii) the patientcalibrator Z
ao. In addition, waveform display is also provided for visually checking the pressure
curve shape and to ensure the correct triggering of the processing circuit. The instrument can beused to monitor trends in cardiac output or for displaying absolute values of cardiac output aftercalibration.
Deloskey et al (1978) showed a comparison of stroke volumes calculated using the Wesseling
pulse-contour formula and those obtained with the electromagnetic flowmeter with severalinterventions introduced to vary stroke volume in one patient. The agreement between the twomethods was very good (0.91) over a wide range of values for stroke volume, heart rate, mean
pulmonary artery pressure and pulmonary-artery resistance.
/G20/G31/G32/G2E/G35 /G49/G4D/G50/G45/G44/G41/G4E/G43/G45/G20/G54/G45/G43/G48/G4E/G49/G51/G55/G45
The technique used for the measurement of cardiac output by the impedance method is illustrated
in Fig. 12.7. If pis the resistivity, the resistance of the thorax between two sensing electrodes (2 and
3) is given by
R0=rL
A
where L is the separation between the electrodes and Ais the cross-sectional area of the thorax.
Assuming that with each ejection of stroke volume dV, the resistance decreases by dR,it can be
derived that
dV=–rL
R2
02L
NMO
QPdR
R can be replaced by Z if an ac signal is used for transthoracic impedance measurement, thus
giving dV = –r (L2/Z2
o)dz.In this relationship, dV is the stroke volume in ml, r is the resistivity of
the patient’s blood in W.cm and dzis the decrease in Zo during a particular systolic ejection.
The stroke volume is given by the product of the initial rate of change of impedance and the
time the aortic and pulmonic valves open, i.e., dz = T(dz /dt)maxwhere ( dz/dt) max corresponds to
the peak negative value of dz/dtfound during systole and T is the interval between dzldt = 0 and
the second heart sound.

Cardiac Output Measurement 355
The equation used by Kubicek et al (1966) for stroke volume can be expressed as
dV=–rL
Z2
02L
NMO
QP◊T◊dz
dtLNMOQPmax
The value of pis related to the patient’s haematocrit and for normal values, a blood resistivity of
150W◊cm can be assumed.
For experimentally calculating the stroke volume, a constant current at 100 kHz is applied
between electrodes 1 and 4. The resulting voltage fluctuations occurring across the thorax
coincident with cardiac activity are detected at the inner pair of electrodes 2 and 3.
The basal impedance between these electrodes is found to be about 25 W and this diminishes by
about 0.1 W with each systole. The voltage signal due to changes in impedance is amplified and
demodulated to obtain Z. The dzldt is calculated using a differentiator. A two-channel recorder is
used to record dz/dtand the phonocardiogram. For each beat the maximum value of the dz/dtat
systole is noted as is the ejection time from the moment the dz/dttracing crosses the dz/dt= 0 line
at the commencement of systole until the onset of the second heart sound.
The method of measuring cardiac output from transthoracic impedance plethysmograms has
several advantages in clinical use, especially in monitoring each stroke volume non-invasively.For this reason, there have been many correlation studies of cardiac-output values betweenthose measured by this method and those by other methods such as indicator dilution, Fick andpressure-gradient methods. Application of the thoracic bio-impedance technique has not beenvery encouraging because the correlation to thermodilution appears generally unacceptable(Bowdle et al, 1993). A completely non-invasive and continuous bio-impedance cardiac output
meter would be a preferred device over thermodilution, but only if the problems of accuracy can be
further solved.S432 1
A1A2
Differentiator
z(output to indicate
changes in impedance)d dz dt(output as / )
Amplifier and
demodulatorPotential
electrodes
Current
electrodesSignal
source
oscillator
= 100 kHzf
Fig.12.7 Technique of measuring cardiac output by impedance changes (after Hill and
Thompson, 1975; by permission of Med. & Biol. Eng)

356 Handbook of Biomedical Instrumentation
/G20/G31/G32/G2E/G36 /G55/G4C/G54/G52/G41/G53/G4F/G55/G4E/G44/G20/G4D/G45/G54/G48/G4F/G44
Ultrasound can be used to measure the velocity of blood flow in the ascending aorta by the
application of the Doppler principle (Huntsman et al, 1983). If the area of cross-section of the aorta
is known or can be measured, the blood flow rate can be calculated as follows:
Blood flow = velocity (cm/sec) ¥ area (cm2)
= cm3/sec = ml/sec ¥ L/100 ml ¥ 60 sec/min
= L/min
The blood flow in the ascending aorta would be identical to the cardiac output minus the
quantitatively negligible flow to the coronary arteries. However, the equation above is an oversimplification because blood flow in the aorta is not constant but pulsatile and blood velocity inthe aorta varies during the cardiac cycle.
Cardiac output measurement devices based on ultrasound actually measure the stroke volume
during the cardiac cycle as per the relationship given below:
SV=CSA ¥
VETzV(t)dt
where SV= stroke volume
CSA = cross-sectional area of the aorta
VET = ventricular ejection time
V= blood velocity
Stroke volume is then multiplied by the heart rate to give cardiac output.The blood velocity V calculated from the Doppler equation is as follows:
V=CF
Fd
o¥
2¥cosq
where: C= speed of sound
Fd= frequency shift
Fo= frequency of the emitted sound
q= angle between the emitted sound and the moving object
Blood velocity in the ascending aorta can be measured by a probe that is held in the operator’s
hand and positioned just above the sternal notch. Blood velocity in the descending aorta can bemeasured (Mark et al, 1986) by a probe attached to an esophageal stethoscope. When compared to
standard thermodilution measurement of cardiac output, the correlation coefficients range from0.63 to 0.91 (Wong et al, 1990). These variations account for the relatively poor accuracy of this
technique.
The area of cross-section of the aorta can be measured by an echo technique, but this is seldom
done in practice. Average human dimensions for the aorta are available from nomograms thatconsider sex, age, height and weight. However, the individual patients may vary considerably
from the average. Therefore, lack of a precise area of cross-section of the particular blood vessel
poses a significant potential source of error. Due to lack of accuracy, determination of cardiacoutput by the ultrasound method has not become popular in clinical practice.
An esophagial ultrasound Doppler instrument for quick and continuous cardiac output
and stroke volume has been introduced by M/s Deltex Medical, UK. This is shown in Fig. 12.8.

Cardiac Output Measurement 357
A single use, small diameter (6mm) probe is inserted into the esophagus and a continuous
Doppler signal is obtained. The equipment utilizes descending aortic flow to provide a real-time assessment of the left ventricular performance. The system works on a continuous 4 MHzultrasound frequency.
Fig.12.8 Cardio Q – Doppler shift based cardiac output and aortic volume
measurement and display system (Courtesy: Deltex Medical, USA)

358 Handbook of Biomedical Instrumentation
/G50/G75/G6C/G6D/G6F/G6E/G61/G72/G79/G20/G46/G75/G6E/G63/G74/G69/G6F/G6E/G20/G41/G6E/G61/G6C/G79/G7A/G65/G72/G73
/G20/G31/G33/G2E/G31 /G50/G55/G4C/G4D/G4F/G4E/G41/G52/G59/G20/G46/G55/G4E/G43/G54/G49/G4F/G4E/G20/G4D/G45/G41/G53/G55/G52/G45/G4D/G45/G4E/G54/G53
Three basic types of measurements are made in the pulmonary clinic: ventilation, distribution and
diffusion. Ventilation deals with the measurement of the body as an air pump, determining itsability to move volumes of air and the speed with which it moves the air. Distribution measure-ments provide an indication of where gas flows in the lungs and whether or not disease has closedsome sections to air flow. Diffusion measurements test the lung’s ability to exchange gas with the
circulatory system.
The most widely performed measurement is ventilation . This is performed using devices called
spirometers that measure volume displacement and the amount of gas moved in a specific time.
Usually this requires the patient to take a deep breath and then exhale as rapidly and completelyas possible. Called the forced vital capacity, this gives an indication of how much air can be movedby the lungs and how freely this air flows.
Distribution measurements quantify degrees of lung obstructions and also determine the residual
volume, which is the amount of air that cannot be removed from the lungs by the patients effort.The residual volume is measured indirectly, such as with the nitrogen washout procedure.
Diffusion measurements identify the rate at which gas is exchanged with the blood stream. This is
difficult to do with oxygen since it requires a sample of pulmonary capillary blood, so it is usually
done by measuring the diminishment of a small quantity of carbon monoxide mixed with theinhaled air.
Pulmonary function analyzers provide the means for automated clinical procedures and
analysis techniques for carrying out a complete evaluation of the lung function or the respiratoryprocess. The respiratory activity ensures supply of oxygen to and removal of carbon dioxide fromthe tissues. These gases are carried in the blood—oxygen from the lungs to the tissues and carbondioxide from the tissues to the lungs. During quiet breathing, the ordinary intake of air or tidalvolume is about 0.5 l. However, only part of this volume takes part actually in oxygenating theblood, because no exchange of gases between air and blood takes place in the mouth, trachea andHAPTER
1313

Pulmonary Function Analyzers 359
bronchi. The air filling these parts is called ‘Dead Space’ air and in adults it typically amounts to
about 0.15 l. The rest of the inspired air ventilates the alveoli and takes part in the exchange ofgases. In each minute, under normal conditions, about 250 ml of oxygen is taken up and 250 ml of
CO
2 is given out by the body. The average composition of atmospheric air and alveolar air is given
in Table 13. 1.
/G95/G20/G54/G61/G62/G6C/G65/G20/G31/G33/G2E/G31 Atmospheric and Alveolar Air Composition
Gas Atmospheric air (%) Alveolar air (%)
O2 20.9 14
CO2 0.1 5.5
N2 79.0 80.5
It might be thought that as the ultimate function of the lungs is to exchange gas with the
environment, measurement of the arterial blood gases would be sufficient to assess lung function.However, the respiratory reserve of the lungs is so considerable that in many cases, symptoms orsigns do not develop until an advanced degree of functional impairment is present. Many of theavailable tests are very sensitive and will detect abnormalities in lung function well before lungdisease has become clinically apparent.
The pulmonary function can be assessed by means of two major classes of tests. These are:
(i) Evaluation of the mechanical aspects of pulmonary function, which affects the bulk gas
transport into and out of the lungs.
(ii)Evaluation of gas exchange or diffusion at the alveoli.
The ability of the pulmonary system to move air and exchange oxygen and carbon dioxide is
affected by the various components of the air passages, the diaphragm, the rib cage and its asso-ciated muscles and by the characteristics of the lung tissue itself. Among the basic tests performed
are those to determine the volumes and capacities of the respiratory system. These are defined as
follows in Fig. 13.1.
/G31/G33/G2E/G31/G2E/G31 /G52/G65/G73 /G70/G69/G72/G61/G74/G6F/G72/G79/G20 /G56/G6F/G6C/G75/G6D/G65/G73
Tidal Volume (TV): The volume of gas inspired or expired (exchanged with each breath) during
normal quiet breathing, is known as tidal volume.
Minute Volume (MV): The volume of gas exchanged per minute during quiet breathing. It is equal
to the tidal volume multiplied by the breathing rate.
Alveolar Ventilation (AV): The volume of fresh air entering the alveoli with each breath.
Alveolar Ventilation = (Breathing rate) ¥ (Tidal volume – Dead space).
Inspiratory Reserve Volume (IRV): The volume of gas, which can be inspired from a normal end-
tidal volume.
IRV = VC – (TV + FRC)

360 Handbook of Biomedical Instrumentation
Expiratory Reserve Volume (ERV): The volume of gas remaining after a normal expiration less the
volume remaining after a forced expiration.
ERV = FRC – RV
Residual Volume (RV): The volume of gas remaining in the lungs after a forced expiration.
/G31/G33/G2E/G31/G2E/G32 /G52/G65/G73 /G70/G69/G72/G61/G74/G6F/G72/G79/G20/G43/G61 /G70/G61/G63/G69/G74/G69/G65/G73
Functional Residual Capacity (FRC): The volume of gas remaining in the lungs after normal
expiration.
Total Lung Capacity (TLC): The volume of gas in the lungs at the point of maximal inspiration.
TLC = VC + RV
Vital Capacity (VC): The greatest volume of gas that can be inspired by voluntary effort after
maximum expiration, irrespective of time.
Inspiratory Capacity (IC): The maximum volume that can be inspired from the resting end
expiratory position.
Dead Space: Dead Space is the functional volume of the lung that does not participate in gas
exchange.
/G31/G33/G2E/G31/G2E/G33 /G43/G6F/G6D /G70/G6C/G69/G61/G6E/G63/G65/G20/G61/G6E/G64/G20/G52/G65/G6C/G61/G74/G65/G64/G20/G50/G72/G65/G73/G73/G75/G72/G65/G73
Compliance (C): Change in volume resulting from unit change in pressure. Units are l/cmH2O.
Lung Compliance (CL): Change in lung volume resulting from unit change in transpulmonary
pressure ( PL)TLCVCIC
FRCIRV
TV
ERV
RV RVMax. expiratory levelResting expiratory
levelMax. expiratory level
Fig.13.1 Volume and capacities of the lungs-standardization of definitions and
symbols in respiratory physiology

Pulmonary Function Analyzers 361
Chest-Wall Compliance (Ccw): Change in volume across the chest wall resulting from unit change
in transchest-wall pressure.
Static Compliance (CST): Compliance measured at point-of-zero airflow by interruption or breach-
hold technique.
Elastance (E): Reciprocal of compliance. Units are cmH2O/litre.
Transpulonary Pressure (PL): Pressure gradient developed across mouth ( Pao) and pleural surface
at lung ( PPL).
Transalveolar Pressure (PEL): Pressure gradient developed between alveolar wall ( PALV–PAL).
Transairway Pressure (PRES): Pressure gradient developed between alveoli and mouth.
Static Elastic Recoil Pressure (RST(L)): Pressure developed in elastic fibers of the lung by expansion.
/G31/G33/G2E/G31/G2E/G34 /G44/G79/G6E/G61/G6D/G69/G63/G20/G52/G65/G73 /G70/G69/G72/G61/G74/G6F/G72/G79/G20/G50/G61/G72/G61/G6D/G65/G74/G65/G72/G73
A number of forced breathing tests are carried out to assess the muscle power associated with
breathing and the resistance of the airway. Among these are:
Forced Vital Capacity (FVC): This is the total amount of air that can be forcibly expired as quickly
as possible after taking the deepest possible breath.
Forced Expiratory Volume (FEV): The percentage of the VC that can be forced out of the lungs in a
given period with ‘maximal exertion’. This is written as FEVT where T is usually in seconds.
Maximum Mid-Expiratory Flow (MMEF or MMF) or Maximum Mid-Flow Rate (MMFR): The
maximum rate of flow of air during the middle half of the FEV spirogram. One half VC is obtainedfrom the volume indicated by the curve between 25 and 75% VC. This is illustrated in Fig. 13.2.
Mid-Expiratory Time (MET): It is the time in seconds over which this volume is forcibly exhaled.
The MMEF is calculated from MMEF = (1/2 VC) ¥ (1/MET)
0255075100
01234Forced
expired vital capacity
FEV1.0
FEV3.0
Vital capacity (%)
Fig.13.2 The calculation of MMEF from the FEV spirogram

362 Handbook of Biomedical Instrumentation
Normal values for each of these volumes and capacities have been calculated. They have been
found to vary with sex, height and age. All pulmonary volumes and capacities are about 20 to 25% less in females than in males.
A particular pattern of abnormal lung volume may occur in a particular form of lung disease
and such a pattern is useful confirmatory evidence of a diagnosis made on clinical grounds.
Further, serial lung function testing is of use in demonstrating progressive deterioration in
function or in confirming a satisfactory response to therapy. For example, if the ratio (FEV
1)/(FVC)
of the volume of gas that can be exhaled forcibly in one second from maximum inspiration (FEV1)
to the forced vital capacity (FVC) is less than 70%, airway obstruction as in chronic bronchitis islikely to be present. If the FEV
1/FVC is greater than 85%, a so called ‘restrictive’ defect may be
present. This is seen in cases of diffuse pulmonary fibrosis.
Pulmonary function tests are performed for the assessment of the lung’s ability to act as a
mechanical pump for air and the ability of the air to flow with minimum impedance through theconducting airways. These tests are classified into two groups: single-breath tests and multiple-breath tests.
There are three types of tests under the single-breath category. These are
• Tests that measure expired volume only.
• Tests that measure expired volume in a unit time.
• Tests that measure expired volume/time.
In the class of multiple-breath test measurements is the Maximal Voluntary Ventilation (MVV)
which is defined as the maximum amount of air that can be moved in a given time period. Here, thepatient breathes in and out for 15 s as hard and as fast as he or she can do. The total volume of thegas moved by the lungs is recorded. The value is mulitiplied by 4 to produce the maximum volumethat the patient breathed per minute by voluntary effort.
A resting person inspires about 0.5 litre of air with each breath, with the normal breathing
rate of 12 to 20 breaths per minute. With exercise, the volume may increase 8 to 10 times and thebreathing rate may reach 40 to 45 breaths per minute. A respiratory disease may be suspected ifthese volumes, capacities or rates are not in the normal range.
/G20/G31/G33/G2E/G32 /G53/G50/G49/G52/G4F/G4D/G45/G54/G52/G59
The instrument used to measure lung capacity and volume is called a spirometer. Basically, therecord obtained from this device is called a spirogram. Spirometers are calibrated containers thatcollect gas and make measurements of lung volume or capacity that can be expired. By adding atime base, flow–dependent quantities can be measured. The addition of gas analzsers makes thespirometer a complete pulmonary function testing laboratory.
/G31/G33/G2E/G32/G2E/G31 /G42/G61/G73/G69/G63/G20/G53 /G70/G69/G72/G6F/G6D/G65/G74/G65/G72
Most of the respiratory measurements can be adequately carried out by the classic water-sealed
spirometer (Fig. 13.3). This consists of an upright, water filled cylinder containing an invertedcounter weighted bell. Breathing into the bell changes the volume of gases trapped inside, and the

Pulmonary Function Analyzers 363
change in volume is translated into vertical motion, which is recorded on the moving drum of a
Kymograph. The excursion of the bell will be proportional to the tidal volume. For most purposes,the bell has a capacity of the order of 6–8 l. Unless a special light weight bell is provided, thenormal spirometer is only capable of responding fully to slow respiratory rates and not to rapidbreathing, sometimes encountered after anaesthesia. Also, the frequency response of a spirometer
must be adequate for the measurement of the forced expiratory volume. The instrument should
have no hysteresis, i.e. the same volume should be reached whether the spirometer is being filledor being emptied to that volume.
As the water-sealed spirometer includes moving masses in the form of the bell and counter-
weights, this leads to the usual problems of inertia and possible oscillation of the bell. This canlead to an over-estimation of the expiratory volume. A suggested compensation is by the use of aspirometer bell having a large diameter and which fits closely over the central core of the spiro-meter, so that the area of water covered by the bell is small in relation to that of the water tank. If thespirometer is used for time-dependent parameters, then it must also have a fast response time,with a flat frequency response up to 12 Hz. This requirement applies not only to the spirometer,but also to the recorder used in conjunction with the recording device.
The spirometer is a mechanical integrator, since the input is air flow and the output is volume
displacement. An electrical signal proportional to volume displacement can be obtained byusing a linear potentiometer connected to the pulley portion of the spirometer. The spirometeris a heavily damped device so that small changes in inspired and expired air volumes are notrecorded. The spirometers can be fitted with a linear motion potentiometer, which directly converts
spirometer volume changes into an electrical signal. The signal may be used to feed a flow-volume
differentiator for the evaluation and recording of data. The response usually is ± 1% to 2 Hz and± 10% to 10 Hz.Counter-
balance
weightOxygen
chamber
WaterFloating drumKymograph
Fig.13.3 Basic water sealed spirometer

364 Handbook of Biomedical Instrumentation
Tests made using the spirometer are not analytical. Also, they are not completely objective
because the results are dependent on the cooperation of the patient and the coaching efforts of agood respiratory technician.
There have been efforts to develop electronic spirometers which could provide greater informa-
tion-delivering and time-saving capabilities. Also, there have been efforts to obtain more definitive
diagnostic information than spirometry alone can provide.
Calculating results manually from the graph of the mechanical volume spirometer requires
considerable time. Transducers have been designed to transform the movement of the bell, bellows
or piston of volume spirometers into electrical signals. These are then used to compute the numeri-cal results electronically. The popularity and low cost of personal computers have made them anattractive method of automating both volume and flow spirometers. An accurate spirometerconnected to a personal computer with a good software programme has the potential of allowinguntrained personnel to obtain accurate result.
/G31/G33/G2E/G32/G2E/G32 /G57/G65/G64/G67/G65/G20/G53 /G70/G69/G72/G6F/G6D/G65/G74/G65/G72
A wedge spirometer (Fig. 13.4) consists of two square pans, parallel to each other and hinged
along one edge. The first pan is permanently attached to the wedge casting stand and contains apair of 5 cm inlet tubes. The other pan swings freely along its hinge with respect to the fixed pan.A space existing between the two pans is sealed airtight with vinyl bellows. The bellows isextremely flexible in the direction of pan motion but it offers high resistance to ‘ballooning’ orinward and outward expansion from the spirometer. As a result, when a pressure gradient existsbetween the interior of the wedge and the atmosphere, there will only be a negligible distortion ofthe bellows.
Fig.13.4 Wedge spirometer (Courtesy: Med. Science, USA)

Pulmonary Function Analyzers 365
As gas enters or leaves the wedge, the moving pan will change position in compensation
for this change in volume. The construction of the wedge is such that the moving pan willrespond to very slight changes in volume. Under normal conditions, the pressure gradient
that exists between the wedge and the atmosphere amounts to only a fraction of a millimetre
of water.
Volume and flow signals for the wedge are obtained independently from two linear transducers.
The transducers are attached to the fixed frame and are coupled to the edge of the moving pan. Onetransducer produces a dc signal proportional to displacement (volume), while the other has a dcoutput proportional to velocity (flow).
The transducer outputs are connected to an electronics unit, which contains the power supply,
an amplifier, and the built-in calibration networks.
A pointer attached to the moving pan and a scale affixed to the frame, combine to provide a
mechanical read out for determining the approximate volume position of the spirometer. Whenopen to the atmosphere and standing upright, the wedge will empty itself due to the force ofgravity acting on the moving member. An adjustable tilt mechanism provides the means forchanging the resting point of the moving pan to any desired volume point. An adjustable magneticstop insures a more highly defined resting position.
Neither the tilt nor the magnetic stop has any noticeable effect on the moving pan position once
it is connected to a closed system. This is primarily due to the large surface area of the pans, whichserves to convert small pressures into large forces.
Thus, the relatively small forces due to gravity and the magnetic stop are overcome by a negligi-
ble rise in pressure in the patient’s lungs. When gravitational return of the moving pan to theresting position is deemed undesirable, the wedge may be turned on its side so that at any point,the pan will be in a state of equilibrium. The wedge may be calibrated with a selector switch, whichdetermines the magnitude of the calibration signal. The volume may be calibrated with a signal
corresponding to either 0.5 ml or 5 l. The flow calibration signals for each particular wedge are
adjusted, using special fixtures. A volume of one litre is introduced at a certain point and a flowrate of 1 l/s is introduced at another point, with the calibration signals then being adjusted toproduce equal signals.
As on conventional spirometers, all standard pulmonary function tests may be performed on
the wedge. X-Y recorders featuring high acceleration slew rates may be used in recording flow/volume loops.
/G31/G33/G2E/G32/G2E/G33 /G55/G6C/G74/G72/G61/G73/G6F/G6E/G69/G63/G20/G53 /G70/G69/G72/G6F/G6D/G65/G74/G65/G72
Ultrasonic spirometers depend, for their action on transmitting ultrasound between a pair of
trans-ducers and measuring changes in transit time caused by the velocity of the intervening
fluid medium (McShane, 1974). They employ piezo-electric transducers and are operated at theircharacteristic resonant frequency for their highest efficiency. Gas flowmeters generally operatein the range from about 40 to 200 kHz. At frequencies higher than 200 kHz, absorption losses inthe gas are very high whereas sounds below 40 kHz are audible and can be irritating.

366 Handbook of Biomedical Instrumentation
Ultrasonic spirometers utilize a pair of ultra-
sonic transducers mounted on opposite sides of aflow tube (Fig. 13.5a). The transducers are capable
of both transmitting and receiving ultrasonic pulses.
In conventional ultrasonic flowmeters, pulses
are transmitted through the liquid or gas in the flow
tube, against and then with the direction of flow.The pulse transit time upstream, t
1, and down-
stream, t2, can be expressed:
t1 =D
Cv-¢and t2 = D
Cv+¢
where D is the distance between the transducers, C
is the velocity of sound propagation in the fluid
and v‘ is the fluid velocity vector along the path of
the pulses. The average gas velocity v through the
flow tube is a vector of v’ so that
v=v cos q
A frequency ( f) is usually measured, which is the reciprocal of the transit times:
11
21tt- =f2–f1 = Cv
D+¢–Cv
D-¢=2¢v
D=2v
Dcosq
The flow velocity is:
v=D
2c o s q11
21tt-LNMOQP=D
2c o s q[f2–f1]
The velocity of sound, C, does not appear in the final equation. Thus, the output accuracy
is unaffected by fluid density, temperature, or viscosity. In gas flow measurements, pulmonaryfunction tubes larger than 3 cm in diameter must be used; the single frequency systems thatmeasure time delay directly must be able to resolve nanoseconds since the total transit delay, t, is
usually measured in microseconds. This technique is not easily implemented because of thedifficulty in measuring these small time differences.
In these flowmeters, disc type flat transducers are mounted in recessed wells on opposite
sides of the flow tube at an angle to the flow as shown in Fig. 13.5(a). This arrangement avoidsflow disruption and provides relative immunity to transducers from contaminations by flowing
substances. Acoustically transparent windows are employed in liquid flowmeters to protect
transducer surfaces and to improve impedance matching to the medium. For measurement ofrespiratory gas flows, however, the medium must be in close contact with the transducer to achievegood acoustic transmission. Consequently, the recess has to be kept open which can lead tounwanted turbulence and moisture collection.
The open tubular wells in which the crystals are mounted, by their geometry, are subject to a
fluid accumulation, which interferes with or obliterates coherent ultrasonic transmission. TheV D
V¢qqX1
X2
(a)
Fig.13.5(a) Diagram of flow tube and
the position of ultrasonictransducers used in tran-sit-time based ultrasonicspirogram

Pulmonary Function Analyzers 367
dead space associated with them is also undesirable. In addition to the condensation problem, the
zero flow base line signal shows drift and is found to show oscillations with changes in temp-erature. The situation is worse when the transducers are applied to patients undergoing ventilatory
assistance as highly inaccurate signals are likely to be obtained due to fluid accumulation.
Blumenfeld et al (1975) modified the construction and designed a coaxial ultrasonic pneumotacho-
meter. In their design (Fig. 13.5(b)), the crystals are mounted midstream in the line of flow, withtheir transmission axis on the centre line of the tubular housing. The principle of measurement offlow is that of measuring transit times, which is a function of the linear gas velocity and hence offlow.
X1X2S
(b)
Fig.13.5(b) Coaxial ultrasound pneumotachometer (after Blumenfeld et al, 1975;
by Permission of Med. & Biol. Eng.)
Although, in theory, the technique seems simple, it is fraught with several sources of error.
For example, a reflected ultrasonic transmission when combined with the primary transmission,may produce artefacts of significant magnitude. Similarly, there could be alterations in the effectivejoint path length of flow and ultrasound transmission as a function of gas velocity and composi-tion. The method requires further experimental work to minimize such problems. Furthermore,use of this transducer for the estimation of flow in gases of varying composition, temperature andhumidity requires a correction for the corresponding variation of C (ultrasound velocity).
Plaut and Webster (1980 a) have attempted to
construct an ultrasonic pneumotachometer whichpossesses the desirable characteristics of open tubegeometry but avoids problems of flow disruption,fluid collection, reduced sensitivity and dead
space which are present in the cross-the-stream
and coaxial transducer configurations. They usedcylindrical shell transducers with their innersurfaces flushed with the walls of the flow tube toprevent flow disruption and fluid accumulation.The axis of transmission of the ultrasound beingparallel to the flow axis enhances the sensitivity ofthe device. Figure 13.5(c) shows the configurationof the cylindrical shell transducers and the airway.A circuit for measuring phase shift as a measure of
(c)
Fig.13.5(c) The configuration of ultra-
sound pneumotachometershowing the cylindrical shelltransducer and the airway(after Plaut and Webster,1980; by Permission of IEEETrans. Biomed. Eng.)

368 Handbook of Biomedical Instrumentation
transit time change due to flow is illustrated by Plaut and Webster (1980 b). They conclude that the
design described by them can give accurate quantitative information when gas composition andtemperature are limited to a narrow range of conditions. It can also measure qualitative parameters
associated with respiration while presenting little obstruction to breathing. They however caution
that poor understanding of the nature of the ultrasonic field and how it inter-acts with moving gasremains the most troublesome problem for the successful development of ultrasonic pneumotacho-meters. The other problems are the poor acoustic efficiency of ultrasonic trans-mission throughgases, the wide variation in gas composition, temperature and humidity and the need for highaccuracy and wide dynamic range.
/G20/G31/G33/G2E/G33 /G50/G4E/G45/G55/G4D/G4F/G54/G41/G43/G48/G4F/G4D/G45/G54/G45/G52/G53
Pneumotachometers are devices that measure the instantaneous rate of volume flow of respiredgases. Basically, there are two types of pneumotachometers, which are:
(i)Differential manometer— It has a small resistance, which allows flow but causes a pressure
drop. This change is measured by a differential pressure transducer, which outputs asignal proportional to the flow according to the Poiseuille law, assuming that the flow islaminar. The unit is heated to maintain it at 37°C to prevent condensation of water vapourfrom the expired breath.
(ii)Hot–wire anemometer —It uses a small heated element in the pathway of the gas flow. The
current needed to maintain the element at a constant temperature is measured and itincreases proportionally to the gas flow that cools the element.
Pneumotachometer is commonly used to measure parameters pertaining to pulmonary function
such as forced expiratory volume (FEV), maximum mid-expiratory volume, peak flow and togenerate flow-volume loops. Although these devices directly measure only volume flow, they canbe employed to derive absolute volume changes of the lung (spirometry) by electronicallyintegrating the flow signal. Conventional mechanical spirometers, though more accurate thanpneumotachometers, have limitations due to their mechanical inertia, hysteresis and CO
2 build-
up. Pneumotachometers, on the other hand, are relatively non-obstructive to the patient and thismakes them suitable for long-term monitoring of patients with respiratory difficulties. A basicrequirement of pneumotachometers (PTM) is that they should present a minimum resistance tobreathing. An acceptable resistance would be between 0.5 and 1.0 cm H
2O s/l. The pressure drop
across the flow head at peak flow is also indicative of PTM resistance. Fleisch PTMs normally
have a peak flow pressure drop of around 1.5 cm H2O. Normal respiratory phenomenon has
significant frequency components up to only 10 Hz and devices with this response should bequite suitable for most applications. More often, it is not the frequency response but the responsetime, which is generally specified. The response time of a typical ultrasonic spirometer is 25 ms.The dead space volume of the flow head should be as small as possible. A bias flow into the flowhead is sometimes introduced to prevent rebreathing of expired air. A good zero stability is aprerequisite of PTMs to prevent false integration during volume measurements. The popular typesof pneumotachometers are explained below.

Pulmonary Function Analyzers 369
/G31/G33/G2E/G33/G2E/G31 /G46/G6C/G65/G69/G73/G63/G68/G20/G50/G6E/G65/G75/G6D/G6F/G74/G61/G63/G68/G6F/G6D/G65/G74/G65/G72
Flow transducers generally used in respiratory
studies are the Fleisch-type pneumotachometers.These transducers are made by rolling a sheet
of thin, corrugated metal with a plain strip of
metal and inserting this core within a metal cover(Fig. 13.6). Thus, these transducers are resistanceelements consisting of small, parallel metalchannels. This construction helps to maintain alaminar flow at much higher flow rates thanwould be possible for a gauze of similar area. Incase of laminar flow, the pressure drop acrossthe element is directly proportional to the flowrate of a gas passing through it.
The output of the flow transducer appears as
a differential pressure. To convert this pressureinto an electrical signal, a second transducer isrequired. A capacitance type pressure transducer
is used in such applications. They are more stable
and less vibration-sensitive than resistive orinductive type transducers.
At high flow rates, turbulence develops in the hose leading to the pneumotach and its response
tends to become non-linear. This limits the usable range of the transducer. The relationship betweenpressure drop and flow is given by DP = AV + BV
2, where the term BV2introduces the non-linearity
effect. This non-linearity is generally corrected electronically.
According to Poisseuille’s law, the pressure developed across a pneumotach by laminar gas
flow is directly proportional to the gas viscosity. The viscosity and temperature coefficient ofviscosity for a variety of respiratory gases (Blais and Fanton, 1979) are given in Table 13.2.
The viscosity ( h) of a mixture of gases is approximated by the equation
1
h=X1
1h = X2
2h = X3
3h+… … .
where X1is the fraction of gas having the viscosity h1. This necessitates the application of an
automatic correction factor to the flow rate for changes in viscosity.
Hobbes (1967) studied the effect of temperature on the performance of a Fleisch head and found
that the output increased by 1% for each degree C rise. He also noted that the effect of saturating airat 37°C with water vapour was to reduce the output from the head by 1.2% as compared with dryair at the same temperature. The calibration of a pneumotachograph head in terms of volume flowrate can be done by passing known gas flows through it. The flow can be produced by a compressorand measured with a rotameter type gas flowmeter.
Most respiratory parameters are reported in BTPS conditions (body temperature, ambient
pressure, saturated with water vapour). This is the condition of air in the lungs and the mouth. ToTo heater
To pressure
transducer
HeaterBodyFlowFlow
channels
Fig.13.6 Construction of pneumotach
transducer (Courtesy: HewlettPackard, USA)

370 Handbook of Biomedical Instrumentation
prevent condensation and maintain the gas under these conditions, the temperature of the
pneumotach is maintained at 37°C. The heater that warms the pneumotach is electrically isolatedfrom the metal case for patient safety, and is encapsulated so that the entire unit may be immersedin liquid for sterilization. The thermistor that senses the temperature and controls the heaterthrough a proportional controller, is buried in the metal case.
/G31/G33/G2E/G33/G2E/G32 /G56/G65/G6E/G74/G75/G72/G69/G2D/G74/G79 /G70/G65/G20/G50/G6E/G65/G75/G6D/G6F/G74/G61/G63/G68/G6F/G6D/G65/G74/G65/G72
This type works similarly to the Fleisch pneumotachometer, but have a venturi-throat for the
linear resistance element. The resulting pressure drop is proportional to the square of volumeflow. They have open geometry and therefore are less prone to problems of liquid collection. Theirmain disadvantages are the non-linearity of calibration and the requirement for laminar flow.
/G31/G33/G2E/G33/G2E/G33 /G54/G75/G72/G62/G69/G6E/G65/G2D/G74/G79 /G70/G65/G20/G50/G6E/G65/G75/G6D/G6F/G74/G61/G63/G68/G6F/G6D/G65/G74/G65/G72
In this design, air flowing through the transducer rotates a very low mass (0.02 g) turbine blade
mounted on jewel bearings. Rotation of the turbine blade interrupts the light beam of a light-emitting diode (LED). The interrupted light beam falls on a phototransistor, which produces atrain of pulses, which are processed and accumulated to correspond to an accumulated volume inlitres.
A special feature of this transducer is a bias air flow, applied to the turbine blades from a pump.
This flow keeps the blades in constant motion even without the sample flow through it. Thisallows measurement of sample air flow in the range of 3 to 600 l/min in the most linear range of thevolume transducer, by overcoming much of the rotational inertia of the turbine. The ‘ZERO’ controlof the volume transducer adjusts the bias air flow to produce a train of clock pulses of exactlythe same frequency as those generated by the crystal oscillator. Figure 13.7 is a diagram of the
transducer.
/G20/G31/G33/G2E/G34 /G4D/G45/G41/G53/G55/G52/G45/G4D/G45/G4E/G54/G20/G4F/G46/G20 /G56/G4F/G4C/G55/G4D/G45
The volume of gas flowing into and out of the lungs is a factor of considerable importance in
investigations of lung function. Whilst the volume of a single breath, or the total volume expired in/G95/G20/G54/G61/G62/G6C/G65/G20/G31/G33/G2E/G32 Viscosity and Temperature Coefficient of Viscosity for Respiratory Gases
Viscosity (Micropoises) Temperature coefficient
Gas at 25°C (Micropoises/°C)
Air 183 0.47
Nitrogen 177.3 0.43
Oxygen 205 0.52
Carbon Dioxide 148.35 0.47
Helium 200.15 0.43
Water Vapour 162 –

Pulmonary Function Analyzers 371
a given time, can be measured by continuously acting spirometers, continuous breath-by-breath
measurements are often difficult. One method is to integrate the flow rate electronically and record
the resulting signals. The flow rate is measured as the pressure change across a pneumotachographhead with a micromanometer whose output is a voltage proportional to the pressure difference atthe manometer input, i.e.
V
1=K(P1–P2)
where K is a constant.
The output from the integrator is a voltage V0 such that
V0=1
12
RCVd tittz
A simplified integrator set up is shown in Fig. 13.8 for flow and volume measurement. It consists
of an ‘autozero’ flowmeter together with a threshold detector and an integrator. The threshold
detector selects which portion of the flow signal is to be integrated and this is normally set to
switch on when the flow signal moves past zero in a positive direction, and off again when the
flow signal returns to zero. This means either inspiration or expiration can be measured dependingon how the flowhead is connected.
When it is intended to measure tidal volume, the flow signal moves positive and continues
until the flow output returns to zero, when the integrator output is reset. The display shows thesize of each breath which is referred to as constant baseline. In case cumulative volume is to bemeasured the volume displayed after each breath is held up. The next breath integrated is addedVolume circuit
100 counts = 1 litreDigital gainOne-way valveOne-way valve
Turbine chamberExpired breath
outExpired breath
Bias flow outTurbineTemperature sensorTemperature
circuitTurbine bias flow
(~ 12 l/min steady)
(adjust for zero)
Mixing
chamber
B B (breath-by-breath)
sample¥
Gas exchange in lungsNon-rebreathing
valves
Non-rebreathing
valve aeratorInspired
breath
pulses
Mouthpiece
Fig.13.7 Turbine type volume transducer (Courtesy: Beckman Instruments Inc., USA)

372 Handbook of Biomedical Instrumentation
to its predecessor, thus producing a staircase pattern. The pattern can be recorded on a chart
paper.
Unless extremely high quality amplifiers are used, the integrator circuit will drift and give false
readings. Drift can be minimized by starting each volume from a fixed baseline on the record.
/G33/G2E/G34/G2E/G31 /G46/G6C/G6F/G77/G2D /G56/G6F/G6C/G75/G6D/G65/G20/G43/G75/G72/G76/G65
This is a plot of instantaneous maximum expiratory flow rate versus volume. The shape of the
flow-volume curve does not vary much between normal subjects of different age, size and sex,although absolute values of flow rate and volume may vary considerably. In patients with obstruc-tive airway disease, the shape of this curve is drastically altered. For this reason, the flow-volumecurve is a good early indication of abnormality. Typical MEFV curves are shown in Fig. 13.9.
There are various methods of producing the flow-volume curve. The method which has been
very common in the past was to record it on the storage oscilloscope and then make a permanentrecord by photographing it with a polaroid camera. This procedure, obviously, is time consumingand expensive.
General purpose X-Y recorders are not fast enough to follow the rapid changes encountered in
the signals while recording flow-volume curves. Therefore, special recorders have been designedto meet this requirement. For example, in the H.P. Pulmonary Function Analyzer, the recorderused has an acceleration of 76.2 m/s
2 and a slewing speed of over 0.762 m/s will result in
approximately a 7.5 cm deflection. The recorder will thus be able to accurately plot a MEFV curvein which the subject reaches a peak flow of 10 l/s in less than one-tenth of a second.SelectTransducer
with
auto zeroThreshold
detectorIntegratorReset
O/P O/PSet level
Manual
zero
Pneumotachograph
head
A-Pneumotachograph (air flow) B-Tidal volume C- Cumulative volume
Fig.13.8 Schematic diagram of electrospirometer for measurement of flow and vol-
ume (Courtesy: Mercury Electronics, Scotland)

Pulmonary Function Analyzers 373
A plot of the inspiratory flow-volume curve is also found to be useful in the detection of certain
lung abnormalities, though it does not yield as much information about the lung mechanics as
does the expiratory flow-volume curve. A useful indicator of the relative degrees of inspiratoryand expiratory obstruction is the MEF
50%/MIF50% ratio (Jordanoglou and Pride, 1968) MIF50% is
maximum inspiratory flow at 50% of vital capacity.
A microcomputer is incorporated in the modern equipment to calculate the maximum
spirometer value stored; FVC, the FEVI and the ratio FEVI/FVC. Besides these, some other indiceswere also evaluated. For example, the average flow over the middle portion of the spirogram hasbeen the most widely accepted parameter for the early detection of increased airway resistance.A microcomputer based system facilitates automating many of such indices which are underinvestigation.Flow volume loopMMFPFFEV /FVC3FEV /FVC1FEVFEV2FEV1FVCOne forced manoeuver produces
Flow volume loop
MMF point
FEV1
FEV2
FEV3PF
MMF
VCExpiratory
phase
Inspiratory
phase
Fig.13.9 Typical flow-volume loops

374 Handbook of Biomedical Instrumentation
/G31/G33/G2E/G34/G2E/G32 /G41/G72/G65/G61/G20/G6F/G66/G20/G74/G68/G65/G20/G46/G6C/G6F/G77/G2D /G56/G6F/G6C/G75/G6D/G65
The area under the maximum expiratory flow-volume curve (AFV) is a sensitive indicator of lung
function impairment. This is because within the scope of a single measurement, maximum flow atall lung volumes, the vital capacity and FEV
1 are all accessible. It has been found that AFV is the
only variable capable of detecting significant inter and intra-individual differences in respect of
both the immediate and the delayed asthmatic responses.
The area under the flow-volume curve can be easily computed by using a square-and-integrat-
ing circuit. In the derivation of area, the following equation is used:
A= Fd V
oVTz
where dV = F dt and therefore,
A= Fd t
oT2z
where VT= total volume exhaled during the time. T is a time greater than the period taken to
complete the respiratory manoeuvre.
A multiplier module like integrated circuit FM 1551 can be used to square the flow rate signal
and then integrate it with respect to time to obtain the area.
/G31/G33/G2E/G34/G2E/G33 /G4E/G69/G74/G72/G6F/G67/G65/G6E/G20/G57/G61/G73/G68/G6F/G75/G74/G20/G54/G65/G63/G68/G6E/G69/G71/G75/G65
Nitrogen washout technique is employed for the indirect determination of RV, FRC and TLC. Inthis technique, the subject breathes 100% oxygen. A nitrogen analyzer is placed near the mouth-piece to continuously monitor the nitrogen content. Some nitrogen is eliminated on everyexpiration, but none is inhaled. The analyzer records nitrogen content which decreases with eachsuccessive expiration since it is progressively replaced with oxygen. The alveolar nitrogenconcentration eventually decreases to 1% when an almost steady state is reached. Nitrogenwashout curves are plotted with time on the X-axis and % N
2in the expired air on the Y-axis.
Theoretically, a plot of % N2versus expired volume will be a straight line on a semi-log paper.
However, with the presence of anatomical dead space, the relationship between nitrogenconcentration and expired volume is no longer exponential but is dependent upon the deadspace/tidal volume ratio. To compensate for this, the dead space is automatically measured in theanalyzers at the start of the test and subtracted from each breath during the test, thus yielding awashout recording unaffected by the patients breathing pattern. A typical complete multi-breathnitrogen washout test would take about 10 min. with modern instruments.
The single breath nitrogen washout test is another index of alveolar ventilation in addition to
providing closing volume information. The test is performed with the subject exhaling to residualvolume, making a maximal inspiration of 100% oxygen and exhaling his vital capacity slowly.Nitrogen concentration is plotted versus volume during the expiration to yield a curve as shownin Fig. 13.10. Information on different parameters as obtained in this test is shown in the diagram.For example, closing volume is an important parameter which gives an early indication of smallairway disease. Closing volume increases with age since the lung loses its elastic recoil and more

Pulmonary Function Analyzers 375
and more airways are closed at higher lung volumes. Closing volume is that volume exhaled from
the onset of closure to the end of vital capacity. It is the first permanent upsloping departure of thecurve from the straight line.
Computational techniques have been developed to eliminate the main drawbacks of traditional
multiple breath nitrogen washout methods. It reduces the time required to deliver an accuratenitrogen washout test and the results are available at the same instant the subject completes thisbreathing manoeuvre.
/G20/G31/G33/G2E/G35 /G50/G55/G4C/G4D/G4F/G4E/G41/G52/G59/G20/G46/G55/G4E/G43/G54/G49/G4F/G4E/G20/G41/G4E/G41/G4C/G59/G5A/G45/G52/G53
A complete pulmonary function analyzer contains all the equipment necessary for testing variousparameters. It comprises a nitrogen analyzer, a vacuum pump, an X-Y recorder, pneumotachs, adigital display, plumbing and valves and other electronic circuits. A simplified block diagram ofthe system is shown in Fig. 13.11. Modern instruments are designed to completely automatethe measurements of ventilation, distribution and diffusion. The systems are designed aroundcomputers which control the procedures by opening and closing appropriate valves, measuringflow rates and the concentrations of various gases, and calculating and printing the results. Ananalog-to-digital converter supplies the measurement data to the computer. Inputs to the A-DVtVt+ ERV
%N2
Closing
volume
Onset of airway
closure
Expired volume VC
Fig.13.10 Single-breath N2 washout curve (Courtesy: Hewlett Packard, USA)

376 Handbook of Biomedical Instrumentation
converter are from various measurement devices, which include a pneumotach that provides a
signal proportional to the air flow for various measurements and carbon monoxide and heliumanalyzers for diffusion measurements. The software controlled pulmonary measurement procedureallows new programs to be added or existing ones to be modified.
Figure 13.12 is a typical example of a microprocessor based pulmonary function testing system.
It makes use of a modified ‘Fleisch-tube’ and replaces the capillary system with a sieve whichserves as a flow transducer. This enables it to achieve a large range of linearity from 0-15 l/s. Thisoffers better precision for the recording of the flow volume loop in the range of high flows as wellas of low flows. The flow proportional pressure drop over a calibrated screen resistor is led to apressure transducer via two silicon hoses. The pneumatic value is converted into an electrical
signal and fed into an amplifier whose output represents ‘Flow’. The signal is integrated to volume
by a voltage to frequency converter. Both signals, ‘flow’ and ‘volume’ are converted into digitalform in an A-D converter and given to the processing unit.
Processing, evaluation and representation of the data is carried out by an 8-bit microprocessor.
The program memory capacity amounts to a maximum of 46 k byte resident in the computer; out ofthis 16 k bytes are used for the software system and 30 k bytes for the measuring and organizationprogrammes. For the intermediate storage of data, the 12 k bytes RAM is utilized. The inputof characters and numbers, mainly of patient data, is done through a keyboard. The systemcommunication is enabled by a 20-digit alphanumerical display line. A 20-digit alphanumericalprinter-plotter produces a hard copy of output values. Forced expiratory curves can be presentedin the form of histograms also.BTPS compen
sation
linear AMPN
analyzer2 N
linearizer2%N2
ProcessorDigital
displays
Auto gain
circuit
Pneumotach
flow
transducer
Pneumotach
flow
transducerPressure
transducerAnalog-to-
digital
converters40 s
delaym
Feedback for autozeroingViscosity
compensationDigital-to-
analog
converterX-Y
recorder
Fig.13.11 Block diagram of pulmonary function analyzer (Courtesy: Hewlett
Packard, USA)

Pulmonary Function Analyzers 377
Fig.13.12 Microprocessor based pulmonary function testing system (Courtesy:
Erich Jaeger, W. Germany)
/G31/G33/G2E/G35/G2E/G31 /G49/G6D /G70/G65/G64/G61/G6E/G63/G65/G20/G50/G6E/G65/G75/G6D/G6F/G67/G72/G61 /G70/G68
Impedance pneumography is an indirect technique for the measurement of respiration. It measures
the respiratory volume and rate through the relationship between respiratory depth and thoracic
impedance change. Impedance pneumography avoids encumbering the subject with masks,
tubes, flowmeters or spirometers; does not impede respiration, and has a minimal a effect onthe psychological state of the subject. Also, it can provide a continuous volumetric record ofrespiration. During measurement of the thoracic impedance change signal, an ac excitation isapplied to the subject. Choice of the optimum frequency for recording transthoracic impedancechanges accompanying respiration is dictated by two important considerations. The first is that ofexcitability of the various tissues between the electrodes and the second is the nature of otherrecordings made at the same time, notably those of bioelectric origin like the ECG. The excitationfrequency employed is usually in the range of 50–100 kHz, with an amplitude of the order of onemilliampere peak-to-peak and in terms of power, it is less than one milliwatt. The excitation is toohigh in frequency and too low in amplitude to stimulate the tissues. The signals of this frequencyget attenuated in almost all biological amplifiers whose frequency response is limited to well
below 10 kHz. Also, a natural rejection of bioelectric events occurs when frequencies in the tens of
kilocycles range are used for the detection of impedance changes with respiration. Therefore, atuned amplifier can be used which has practically zero response for the spectrum of bioelectricevents while at the same time provides high amplification for the carrier used for impedancepneumography.

378 Handbook of Biomedical Instrumentation
Changes in transthoracic impedance during respiration should be independent of impedance
resulting from the sum of the resting thoracic impedance and the contact impedance betweenthe electrodes and the skin. This is possible if a constant current is maintained through the
subject over a large range of thoracic and electrode-skin impedance, likely to be encountered in
practice.
The thoracic impedance and impedance change signal may be viewed as either polar or
cartesian vectors as shown in Fig. 13.13. The respiratory signal DZis shown as the change between
an initial impedance Z
0 and a new impedance Z. The impedance change DZ is drawn much larger
on the diagram than it usually occurs. Several investigators have studied the changes in trans-thoracic electrical impedance associated with respiration after applying low-intensity sinusoidalcurrents from 20 to 600 kHz. DZhas been shown to be essentially a change in the resistive
component, virtually independent of frequency within the usually employed frequency range andcorrelates well with changes in the respired volume (DV)so that the impedance changes can be
used, within certain limits of accuracy, as a quantitative measure of respired volume.
q0Dqq¢DxcDz
z0z¢
DR
Rz()Initial subject
impedance–1 ( )z
Respiratory
signal
Fig.13.13 Change in impedance ( DDDDDZ) expressed in terms of polar or cartesian
vectors
The transthoracic impedance is a function of frequency and the type and size of electrodes.
Silver/silver chloride electrodes with a diameter of 9.5 mm have a typical impedance of 500–
800W at 50 kHz, whereas 4 mm diameter electrodes exhibit 1000–1500 W at the same frequency.
Typically, the impedance change signal is of the order of 3 W/l of respiratory volume change.

Pulmonary Function Analyzers 379
Baker et al (1966) report that the maximum change in transthoracic impedance per unit volume of
respired air (DZIDV)was recorded from electrodes secured to the eighth rib bilaterally on the
midaxillary lines. The magnitude of dZld V decreased rapidly with distance along the midaxillary
line in both directions from the eighth rib. Also, for a given electrode location, the ratio DZIDV was
found to be essentially the same whether the lungs were inflated by spontaneous respirationor inflated by a positive pressure respirator after relaxation of spontaneous efforts followinghyperventilation.
There are disadvantages in the use of the impedance pneumograph if absolute measurements
of respiration are required. This is because the conversion of impedance change to lung-volumechange is a function of electrode position, body size and posture.
Miyamoto et al (1981) studied the reliability of the impedance pneumogram as a method to
assess ventilation during exercise in comparison with the standard pneumatic method. It is foundthat a correlation coefficient of 0.92 to 0.99 exists between tidal volume changes and impedancevariation. However, it was observed that the accuracy of the impedance method is still inferiorto the standard pneumatic methods. Individual as well as regional variations of the impedancesensitivity also require time consuming calibration procedures for each subject.
/G20/G31/G33/G2E/G36 /G52/G45/G53/G50/G49/G52/G41/G54/G4F/G52/G59/G20/G47/G41/G53/G20/G41/G4E/G41/G4C/G59/G5A/G45/G52/G53
A knowledge of the qualitative and quantitative composition of inspired and expired gas and
vapour mixtures is of great importance in investigations connected with respiratory physiology,
lung function assessment and anaesthesia. A number of physical methods have been utilizedas the basis for a gas analyzer. Commercially available analyzers depend for their action on themeasurement of quantities such as infrared or ultraviolet absorption, paramagnetism, thermalconductivity or the ratio of charge to mass of ionized molecules. There are analyzers, designedprimarily for a analysis of a single component of the gas mixture whereas there are instrumentssuch as the mass spectrometer and gas chromatograph which provide a multi-component analysis.The most common gases of interest for measurement and analysis are carbon dioxide, carbonmonoxide, nitrous oxide and halothane in mixtures of respiratory or anaesthetic gas samples.
/G31/G33/G2E/G36/G2E/G31 /G49/G6E/G66/G72/G61/G72/G65/G64/G20/G47/G61/G73/G20/G41/G6E/G61/G6C/G79/G73/G65/G72/G73
Infrared gas analyzers depend for their operation upon the fact that some gases and vapoursabsorb specific wavelengths of infrared radiation. One of the most commonly measured gas usingthe infrared radiation absorption method is carbon dioxide. The technique used for this purposeis the conventional double-beam infrared spectrometer system having a pair of matched gas cellsin the two beams. One cell is filled with a reference gas which is a non-absorbing gas like nitrogenwhereas the measuring cell contains the sample. The difference in optical absorption detectedbetween the two cells is a measure of the absorption of the sample at a particular wavelength.Commercial gas analyzers used for measuring the amount of carbon dioxide (CO
2), carbon
monoxide (CO), nitrous oxide (N2O) or halothane in mixtures of respiratory and anaesthetic gases
make use of a non-dispersive infrared (NDIR) analysis technique. A major advantage of this

380 Handbook of Biomedical Instrumentation
technique is that it is highly specific to the gas being measured and, therefore, separate pick-up
heads selective to the wavelengths absorbed by the gas are necessary.
Infrared gas analyzers are particularly useful for measuring carbon dioxide in the respired air.
For the measurement of gases in the respired air two types of samplers are employed—a microcatheter-cell and a breathe-through cell. The micro-catheter cell is used with a vacuum pump to
draw off small volumes from the nasal cavity or trachea. Its typical volume is 0.1 ml and it is parti-
cularly useful when larger volumes could cause patient distress. The breathe-through cell acceptsthe entire tidal volume of breath with no vacuum assistance. It can be connected directly into thecircuit of an anaesthesia machine. These instruments have a typical response time of 0.1 s and asensitivity range of 0 to 10% CO
2.
The block diagram of an infrared gas analyzer is shown in Fig. 13.14. The solid state detector is
PbSe. The chopper has a high speed of 3000 rpm and provides a response time of up to 100 ms for90% reading. The infrared source operates at a temperature of about 815°C where it emits infraredenergy optimized for the spectral bands of interest. The infrared energy source is located at thefront focal plane of a parabolic reflector, so that the reflected energy from the reflector is effectivelycollimated. The collimated energy is chopped by the coaxial chopper, which allows the energy topass alternately through the reference and sample tubes. Since the energy is collimated, it passesthrough these tubes without internal reflections so that gold foil coatings on the inside of thesetubes are not necessary. The sample tube length can be selected according to the absorption
strength and concentration of the sample gas. At the output end of the two tubes, a second
parabolic reflector images the energy onto the detector filter assembly. The filter is a narrowbandpass interference filter with band pass characteristics matched to the absorption spectra ofthe gas of interest.
Reference
cellReference
cell
Diaphragm
distended
Capacitive outputInfrared
source
Chopper
Sample inSample in
Sample out Sample outSample
cellSample
cellMirror 1
Mirror 2Sig. sync.Ref. sync.
Chopper
Detectors filtersSource
Fig.13.14 Principle of infrared gas analyzer: conventional (two sources) and
improved design (with single source) (Courtesy: Infrared Industries,
USA)

Pulmonary Function Analyzers 381
/G31/G33/G2E/G36/G2E/G32 /G50/G61/G72/G61/G6D/G61/G67/G6E/G65/G74/G69/G63/G20/G4F/G78/G79/G67/G65/G6E/G20/G41/G6E/G61/G6C/G79/G73/G65/G72
Oxygen has the property of being paramagnetic in nature, i.e. it does not have as strong a mag-
netism as permanent magnets, but at the same time it is attracted into a magnetic field. Nitricoxide and nitrogen dioxide are two other gases which are paramagnetic in nature. Most gases
are, however, slightly diamagnetic, i.e. they are repelled out of a magnetic field. The magnetic
susceptibility of oxygen can be regarded as a measure of the tendency of an oxygen molecule tobecome temporarily magnetized when placed in a magnetic field. Such magnetization isanalogous to that of a piece of soft iron in a field of this type. Similarly, diamagnetic gases arecomparable to non-magnetic substances. The paramagnetic property of oxygen has been utilizedin constructing oxygen analyzers.
The paramagnetic oxygen analyzer was first described by Pauling et al (1946). Their simple
dumb-bell type of instrument has formed the basis of more modern instruments. The arrangementincorporates a small glass dumb-bell suspended from a quartz thread between the poles of apermanent magnet. The pole pieces are wedge-shaped in order to produce a non-uniform field.Referring to Fig. 13.15, when a small sphere is suspended in a strong non-uniform magnetic field,it is subject to a force proportional to the difference between the magnetic susceptibility of thissphere and that of the surrounding gas. The magnitude of this force can be expressed as:
F = C(K – K
o)
where C= a function of the magnetic field strength and gradient
Exciter lampSatationary mirror
Magnetic pole
pieces
High pot-
ential
vaneMirror
Test body
Light beam
Translucent
screen
ImageNull
lineLow
potential
vaneZero
controlSpan control
Null adjust
Range
resistors
To mains
supply
Fig.13.15 Functional diagram of Oxygen analyzer based on magnetic susceptibil-
ity (Courtesy: Beckman Instruments Inc., USA)

382 Handbook of Biomedical Instrumentation
K0= magnetic susceptibility of the sphere
K= magnetic susceptibility of the surrounding gas
The forces exerted on the two spheres of the test body are thus a measure of the magnetic
susceptibility of the sample and therefore, of its oxygen content.
The magnetic forces are measured by applying to one sphere an electrostatic force equal and
opposite to the magnetic forces. The electrostatic force is exerted by an electrostatic field establishedby two charged vanes mounted adjacent to the sphere. One vane is held at a higher potential thanthe test body, the other at a lower potential. Since the glass test body must be electrically conductive,it is sputtered with an inert metal.
The test body is connected electrically to the slider of the ‘Null Adjust’ potentiometer R
20. This
potentiometer is part of a voltage-dividing resistor network connected between ground and B+.
Potential to the test body can be adjusted over a large range. An exciter lamp directs a light beam
on to the small mirror attached to the test body. From the mirror, the beam is reflected to a stationarymirror and then on to a transluscent screen mounted on the front panel of the instrument. Thegeometry of the optical system is so arranged that a very small rotation of the test body causes anappreciable deflection of the image cast by the beam.
Zero control of the instrument is provided by ganged R
13–R15 setting, which changes the voltage
present on each vane with respect to the ground, but does not change the difference in potentialexisting between them. This adjustment alters the electrostatic field. Rheostat R
19 sets the upscale
standardization point, i.e. provides span or sensitivity control. When no oxygen is present, themagnetic forces exactly balance the torque of the fibre. However, if oxygen is present in the gassample drawn in the chamber surrounding the dumb-bell, it would displace the dumb-bell spheresand they would move away from the region of maximum magnetic flux density. The resultingrotation of the suspension turns the small mirror and deflects the beam of light over a scale of theinstrument. The scale is calibrated in percentages by volume of oxygen or partial pressure of
oxygen. Paramagnetic oxygen analyzers are capable of sampling static or flowing gas samples.
The null position is detected by a sensitive light, mirror and scale arrangement.
Oxygen analyzers are available with continuous readout 0–25% or 0–100% oxygen. The instru-
ments are calibrated with the reference gas specified. Standard cell volume is 10 ml and responsetime is about 10 s.
The recommended flow rate is 40 to 250 cc/min when the sample enters the analysis cell. Before
the gas enters the analyzer, it must be pressurized with the pump and passed through a suitablecleaning and drying system. In many cases, a small plug of glass wool is sufficient for cleaningand drying functions. The entry of moisture or particulate matter into the analyzer will changeinstrument response characteristics. Therefore, the use of a suitable filter in the sample inlet line isrecommended.
Any change in the temperature of a gas causes a corresponding change in its magnetic suscepti-
bility. To hold this temperature of the gas in the analysis cell constant, the analyzer incorporatesa thermostatically-controlled heating circuit. Once the instrument reaches temperature equilibrium,the temperature inside the analysis cell is approximately 60
oC. The sample should be admitted
to the instrument at a temperature between 23oC and 43oC. If the temperature is less than 10oC,
then the sample may not have time to reach the temperature equilibrium before entering theanalysis cell.

Pulmonary Function Analyzers 383
Calibration of the instrument consists of establishing two standardization points, i.e. a down
scale and an up scale standardization point. These two points can be set by passing standardgases through the instrument at a fixed pressure, normally atmospheric pressure. First a zero
standard gas is admitted and the zero control is adjusted. Then the span gas is admitted and the
span control is adjusted. Alternatively, the required practical pressures of oxygen are obtained byfilling the analysis cell to the appropriate pressures with non-flowing oxygen or air. If the highestpoint is not greater than 2l% oxygen, then dry air is used to set the span point. If the point is greaterthan 21% oxygen, the oxygen is used to set this point.
/G31/G33/G2E/G36/G2E/G33 /G50/G6F/G6C/G61/G72/G6F/G67/G72/G61/G70/G68/G69/G63/G20/G4F/G78/G79/G67/G65/G6E/G20/G41/G6E/G61/G6C/G79/G7A/G65/G72
Polarographic cells are generally used to meas-
ure the partial pressure or percentage of oxygen
from injected samples, continuous streams or instatic gas monitoring. They find maximum utilityin the respiratory and metabolic laboratories.Polarographic cells are based on the redoxreactions in a cell having both the electrodesof noble metals. When a potential is applied,oxygen is reduced at the cathode in the presenceof KCI as the electrolyte and a current will flow.The cathode is protected by an oxygen permeablemembrane, the rate at which oxygen reaches thecathode will be controlled by diffusion through
the membrane. The voltage current curve will be
a typical polarogram (Fig. 13.16). A residualcurrent flows in the cell at low voltages. Thecurrent rises with the increase in voltage until itreaches a plateau where it is limited by the diffusion rate of oxygen through the membrane. For agiven membrane and at a constant temperature, this would be proportional to the partial pressureof oxygen across the membrane. When the voltage is applied in the plateau region, the current inthe cell is proportional to oxygen concentration.
Polarographic cells are temperature sensitive as the diffusion coefficient changes with temp-
erature. The temperature coefficient is usually 2–4% per degree centigrade. Therefore, temperaturecompensation circuits are used to overcome this problem.
Polarographic oxygen cells are used mainly for portable gas detectors, where simplicity, low
cost and light weight are important. They are preferably used for measuring oxygen in liquids,especially in water pollution and medical work. The commercial oxygen analyzers use oxygensensors, which contain a gold cathode, a silver anode, a potassium chloride electrolyte gel and athin membrane. The membrane is precisely retained across the exposed face of the gold cathode,
compressing the electrolyte gel beneath into a thin film. The membrane, permeable to oxygen,
prevents airborne solid or liquid contaminants from reaching the electrolyte gel. The sensor isinsensitive to other common gases. A small electrical potential (750 mV) is applied across theanode and cathode.Voltage
Half wave
potentialCurrent
Residual
currentCurrent
plateau
Fig.13.16 Typical polarogram showing
relationship of current gener-
ated in the polarographic cell asa function of voltage

384 Handbook of Biomedical Instrumentation
Although the composition of the atmosphere is remarkably constant from sea level to the highest
mountain, i.e. oxygen, 21% and nitrogen 79%, there is a great difference in the partial pressure of
oxygen at different altitudes. The polarographic sensor, which actually senses partial pressure,
would, therefore, require some adjustment to read correctly the approximate percentage of oxygenat the altitude at which it is used. Humidity can also affect oxygen readings, but to a lesser degree.Water vapour in air creates a water vapour partial pressure that slightly lowers the oxygen partialpressure. Therefore, for precision work, it is often desirable to use a drying tube on the inlet sampleline. Also, care should be taken to calibrate and sample under the same flow conditions as requiredfor the gas to be analyzed. The range of the instrument is 0–1000 mmHg O
2 and the response time
is 10 s for 90%, 35s for 99% and 70s for 99.9%. The instrument can measure oxygen against abackground of nitrogen, helium, neon, argon etc. with no difficulty. The sensor is very slightlysensitive to carbon dioxide and nitrous oxide with typical error less than 0.1% oxygen for 10%carbon dioxide and 4% oxygen for 100% nitrous oxide.
/G31/G33/G2E/G36/G2E/G34 /G54/G68/G65/G72/G6D/G61/G6C/G20/G43/G6F/G6E/G64/G75/G63/G74/G69/G76/G69/G74/G79/G20/G41/G6E/G61/G6C/G79/G73/G65/G72/G73
Thermal conductivity of a gas is defined as the quantity of heat (in calories) transferred in unittime (seconds) to a gas between two surfaces 1 cm
2 in area and 1 cm apart, when the temperature
difference between the surfaces is 1°C. The ability to conduct heat is possessed by all gases but invarying degrees. This difference in thermal conductivity can be employed to determine quantita-tively the composition of complex gas mixtures. Changes in the composition of a gas streammay give rise to a significant alteration in the thermal conductivity of the stream. This can beconveniently detected from the rise or fall in temperature of a heated filament placed in the path of
the gas stream. The changes in temperature can be detected by using either a platinum filament
(hot wire) or thermistors. In a typical hot-wire cell thermal conductivity analyzer, four platinumfilaments (Fig. 13.17) are employed as heat-sensing elements. They are arranged in a constantcurrent bridge circuit and each of them is placed in a separate cavity in a brass or stainless steelblock. The block acts as a heat sink. The material used for the construction of filaments must havea high temperature coefficient of resistance. The material generally used for the purpose aretungsten, Kovar (alloy of Co, Ni and Fe) or platinum.
Sample gas
S1
S2 R2R1Ref.
gasDB+
B–A
FGE
MeterAdjust
heating
currentHeating
current
indicatorAdjust balance with
ref. gas in all cells
Fig.13.17 Arrangement of a thermal conductivity based gas analyzer

Pulmonary Function Analyzers 385
Two filaments ( R1 and R2) connected in opposite arms of the Wheatstone bridge act as the
reference arms, whereas the other two filaments ( S1 and S2) are connected in the gas stream, and act
as the measuring arms. The use of a four-cell arrangement serves to compensate for temperature
and power supply variations.
Initially, the reference gas is made to flow through all the cells and the bridge is balanced
precisely with the help of the potentiometer D. When the gas stream passes through the measuring
pair of filaments, the wires are cooled and there is a corresponding change in the resistance of thefilaments. The higher the thermal conductivity of the gas, the lower would be the resistance of thewire and vice-versa. Consequently, the greater the difference in thermal conductivities of thereference and sample gas, the greater the unbalance of the Wheatsone bridge. The unbalancecurrent can be measured on an indicating meter or on a strip chart recorder.
Thermistors can also be used as heat sensing elements arranged in a similar manner as hot wire
elements in a Wheatstone bridge configuration. Thermistors possess the advantage of beingextremely sensitive to relatively minute changes in temperature and have a high negativetemperature coefficient. When used in gas analyzers, they are encapsulated in glass. Thermistorsare available which are fairly fast in response.
Thermal conductivity gas analyzers are inherently non-specific. Therefore, the simplest analysis
occurs with binary gas mixtures. A thermal conductivity analyzer can be used in respiratoryphysiology studies to follow CO
2 concentration changes in the individual breaths of a patient. A
high speed of response necessary for this purpose can be obtained by reducing the pressure of thegas surrounding the filaments to a few millimetres of mercury. The variations in the proportions of
oxygen and nitrogen in the sample stream will have little effect, since they both have almost the
same thermal conductivity. The effect of changes in water vapour content can be minimized byarranging to saturate the gas fed to both the sample and reference filaments.
/G31/G33/G2E/G36/G2E/G35 /G4E/G32/G20/G41/G6E/G61/G6C/G79/G7A/G65/G72/G20/G42/G61/G73/G65/G64/G20/G6F/G6E/G20/G49/G6F/G6E/G69/G7A/G61/G74/G69/G6F/G6E/G20/G54/G65/G63/G68/G6E/G69/G71/G75/G65
With sufficient electrical excitation and at suitable pressures, certain gases emit radiation in diffe-rent ways like spark, arc, glow discharge in different parts of the radiation spectrum. Measurementof the emitted radiation can help in the determination of the unknown concentration of a gas in a
mixture. This technique has been utilized for the measurement of nitrogen gas, particularly in
respiratory gases. The measuring technique is essentially that of a photo-spectrometer, wherein agas sample is ionized, selectively filtered, and detected with a photocell, which provides anappropriate electrical output signal. The presence of nitrogen is detected by the emission of acharacteristic purple colour when discharge takes place in a low pressure chamber containing thegas sample.
The instrument consists of a sampling head and a discharge tube. The sampling head contains
the ionizing chamber, filter and the detector. The other part contains the power supply, amplifierand display system. The ionizing chamber or the discharge tube is maintained at an absolute press-ure of a few torr. A rotary oil vacuum pump draws a sample and feeds it to the discharge tube. Thevoltage required for striking the discharge in the presence of nitrogen is of the order of 1500 V dc. Thelight output from the discharge tube is interrupted by means of a rotating slotted disc (Fig. 13.18)so that a chopped output is obtained. This light is then passed through optical filters having a

386 Handbook of Biomedical Instrumentation
wavelength corresponding to the wavelength of the discharge. The intensity of light is measured
with a photocell and an amplifier specifically tuned to the chopping frequency. The light intensityis proportional to the nitrogen concentration. The sampling rate is adjusted with the help of a
needle valve, which is normally set at 3 ml/min. The vacuum system provides 600–1200 mHg. The
instrument is calibrated for water saturated mixtures of nitrogen and oxygen as a reading error upto 2% can be expected with dry gases. Compensation for this error can be simply made by adjustingthe sampling head needle valve, if it is desired to monitor dry gases.Flowing gas mixture
Inlet orfice
RecordFilter
Light
interrupterPhotocell
Amp.
Vacuum
pumpCathodeAnode
Power
supply
Fig.13.18 Schematic diagram showing principle of operation of nitrogen gas
analyzer

Clinical Laboratory Instruments 387
/G43/G6C/G69/G6E/G69/G63/G61/G6C/G20/G4C/G61 /G62/G6F/G72/G61/G74/G6F/G72/G79/G20/G49/G6E/G73/G74/G72/G75/G6D/G65/G6E/G74/G73
/G20/G31/G34/G2E/G31 /G4D/G45/G44/G49/G43/G41/G4C/G20/G44/G49/G41/G47/G4E/G4F/G53/G49/G53/G20/G57/G49/G54/G48/G20/G43/G48/G45/G4D/G49/G43/G41/G4C/G20/G54/G45/G53/G54/G53
Most of the pathological processes result in chemical changes in the internal environment of the
human body. These changes can generally be detected by the analysis of various samples takenfrom the body. The analysis not only helps in the diagnosis of various ailments but also in deter-mining the progress of treatment and for making a prognosis. Samples taken from the body areanalyzed in three different areas within the clinical laboratory set up. These are:
• Chemistry section deals with the analysis of blood, urine, cerebrospinal fluid (CSF) and
other fluids to determine the quantity of various important substances they contain. Mostof the electronic instruments in the clinical laboratory are available in this section.
• Haematology section deals with the determinations of the number and characteristics of the
constituents of the blood, particularly the blood cells.
• Microbiology section in which studies are performed on various body tissues and fluids to
determine the presence of pathological micro-organisms.
The most common substance for analysis from the body is blood. This is because the blood
carries out the most important function of transportation and many pathological processes mani-fest themselves as demonstrable changes in the blood. Deviations from the normal composition ofurine also reflect many pathological processes. The liquid part of the blood—the blood plasma,and the formed elements—the blood cells are analyzed during a chemical examination. The bloodplasma accounts for about 60% of the blood volume and the blood cells occupy the other 40%. Theplasma is obtained by centrifuging a blood sample. During centrifugation, the heavy blood cellsget packed at the bottom of the centrifuge tube and the plasma can thus be separated. The plasmais a viscous, light yellow liquid, i.e. almost clear in the fasting stage.
/G20/G31/G34/G2E/G32 /G53/G50/G45/G43/G54/G52/G4F/G50/G48/G4F/G54/G4F/G4D/G45/G54/G52/G59
Spectrophotometry is the most important of all the instrumental methods of analysis in clinicalchemistry. This method is based on the absorption of electromagnetic radiation in the visible,HAPTER
1414

388 Handbook of Biomedical Instrumentation
ultraviolet and infrared ranges. According to the quantum theory, the energy states of an atom or
molecule are defined and change from one state to another, would, therefore, require a definiteamount of energy. If this energy is supplied from an external source of radiation, the exact quantity
of energy required to bring about a change from one given state to another will be provided by
photons of one particular frequency, which may thus be selectively absorbed. The study of thefrequencies of the photons which are absorbed would thus provide information about the natureof the material. Also, the number of photons absorbed may provide information about the numberof atoms or molecules of the material present in a particular state. It thus provides us with amethod to have a qualitative and quantitative analysis of a substance.
Molecules possess three types of internal energy—electronic, vibrational and rotational. When
a molecule absorbs radiant energy, it can increase its internal energy in a variety of ways. Thevarious molecular energy states are quantized and the amount of energy necessary to cause anychange in any one of the above energy states would generally correspond to specific regions of theelectromagnetic spectrum. Electronic transitions correspond to the ultraviolet and visible regions,vibrational transitions to the near infrared and infrared regions and rotational transitions to theinfrared and far-infrared regions. The method based on the absorption of radiation of a substanceis known as Absorption Spectroscopy . The main advantages of spectrometric methods are speed,
sensitivity to very small amounts of change and a relatively simple operational methodology. The
time required for the actual measurement is very short and most of the analysis time, in fact, goesinto preparation of the samples.
The region in the electromagnetic spectrum which is normally used in spectroscopic work is
very limited. Visible light represents only a very small portion of the electromagnetic spectrum andgenerally covers a range from 380 to 780 m m. The ultraviolet region extends from 185 m mto the
visible. Shorter wavelengths lie in the far ultraviolet region, which overlaps the soft X-ray part ofthe spectrum. Infrared region covers wavelengths above the visible range.
/G31/G34/G2E/G32/G2E/G31 /G49/G6E/G74/G65/G72/G61/G63/G74/G69/G6F/G6E/G20/G6F/G66/G20/G52/G61/G64/G69/G61/G74/G69/G6F/G6E/G20/G77/G69/G74/G68/G20/G4D/G61/G74/G74/G65/G72
When a beam of radiant energy strikes the surface of a substance, the radiation interacts with theatoms and molecules of the substance. The radiation may be transmitted, absorbed, scattered orreflected, or it can excite fluorescence depending upon the properties of the substance. Theinteraction however does not involve a permanent transfer of energy.
The velocity at which radiation is propagated through a medium is less than its velocity in
vacuum. It depends upon the kind and concentration of atoms, ions or molecules present in themedium. Figure 14.1 shows various possibilities which might result when a beam of radiationstrikes a substance. These are:
(a) The radiation may be transmitted with little absorption taking place, and therefore, with-
out much energy loss.
(b) The direction of propagation of the beam may be altered by reflection, refraction, diffrac-
tion or scattering.
(c) The radiant energy may be absorbed in part or entirely by the substance.

Clinical Laboratory Instruments 389
In absorption spectrophotometry, we are usually concerned with absorption and transmission.
Generally, the conditions under which the sample is examined are such that they keep reflectionand scattering to a minimum.
Absorption spectrophotometry is based on the principle that the amount of absorption that
occurs is dependent on the number of molecules present in the absorbing material. Therefore, theintensity of the radiation leaving the substance may be used as an indication of the concentrationof the material.
Let us suppose , Pois the incident radiant energy and Pis the energy which is transmitted. The
ratio of the radiant power transmitted by a sample to the radiant power incident on the sample isknown as the transmittance.
Transmittance T= P / P
0
% Transmittance = ( P/P0)¥ 100
The logarithm to the base 10 of the reciprocal of the transmittance is known as absorbance.
Absorbance = log10 (1/T)
= log10(P0/ P)
Optical density = log10 (100/ T)
According to the Beer-Lambert Law,
P=P010–e b c
where c is the concentration, b the path length of light and e the extinction coefficient. The concent-
ration c of the substance is calculated from the absorbance, also called the extinction E, as follows:
Absorbance ( E) = log P0/P = ebc
In spectrophotometric measurements e and b are nearly constant so that for a particular
determination, absorbance ( A) varies only with concentration ( c). Therefore, if absorbance is plottedBAbsorbed
radiation
Incident radiation Transmitted radiation
ReflectionScatter
Fluorescence
Fig.14.1 Interaction of radiation with matter

390 Handbook of Biomedical Instrumentation
graphically against concentration, a straight line is obtained. A graph derived from the transmitt-
ance data will not be a straight line, unless transmittance (or per cent transmission) is plotted onthe log axis of a semi-log paper. The constant ain the Beer Lambert law must have units corres-
ponding to the units used for concentration and path length of the sample.
The most usually employed quantitative method consists of comparing the extent of absorption
or transmittance of radiant energy at a particular wavelength by a solution of the test material
and a series of standard solutions. It can be done with visual colour comparators, photometers orspectrophotometers.
/G20/G31/G34/G2E/G33 /G53/G50/G45/G43/G54/G52/G4F/G50/G48/G4F/G54/G4F/G4D/G45/G54/G45/G52/G20/G54/G59/G50/G45/G20/G49/G4E/G53/G54/G52/G55/G4D/G45/G4E/G54/G53
Figure 14.2 shows an arrangement of components of a typical spectrophotometer type instrument.The essential components are:
• A source of radiant energy, which may he a tungsten lamp, a xenon-mercury arc, hydrogen
or deuterium discharge lamp, etc.
• Filtering arrangement for the selection of a narrow band of radiant energy. It could be a
single wavelength absorption filter, an interference filter, a prism or a diffraction grating.
• An optical system for producing a parallel beam of filtered light for passage through an
absorption cell (cuvette). The system may include lenses, mirrors, slits, diaphragm, etc.
• A detecting system for the measurement of unabsorbed radiant energy, which could be the
human eye, a barrier-layer cell, phototube or photo-multiplier tube.
• A readout system or display, which may he an indicating meter or a numerical display.
Optical components
Sample holderDetector
IndicatorSource of
radiant
energy
Dispersing elements
Fig.14.2 Various components of a spectrophotometer type instrument
/G31/G34/G2E/G33/G2E/G31 /G52/G61/G64/G69/G61/G74/G69/G6F/G6E/G20/G53/G6F/G75/G72/G63/G65/G73
The function of the radiation source is to provide a sufficient intensity of light which is suitable for
making a measurement. The most common and convenient source of light is the tungsten lamp.This lamp consists of a tungsten filament enclosed in a glass envelope. It is cheap, intense andreliable. A major portion of the energy emitted by a tungsten lamp is in the visible region and onlyabout 15 to 20% is in the infrared region. When using a tungsten lamp, it is desirable to use a heat

Clinical Laboratory Instruments 391
absorbing filter between the lamp and the sample holder to absorb most of thc infrared radiation
without seriously diminishing energy at the desired wavelength. For work in the ultraviolet region,a hydrogen or deuterium discharge lamp is used.
In these lamps, the envelope material of the lamp puts a limit on the smallest wavelength which
can be transmitted. For example, quartz is suitable only up to 200 mm and fused silica up to
185 m m. The radiation from the discharge lamps is concentrated into narrow wavelength regions
of emission lines. Practically, there is no emission beyond 400 m m in these lamps. For this reason,
spectrophotometers for both the visible and ultraviolet regions always have two light sources,which can be manually selected for appropriate work.
For work in the infrared region, a tungsten lamp may be used. However, due to high absorption
of the glass envelope and the presence of unwanted emission in the visible range, tungsten lampsare not preferred. In such cases, nernst filaments or other sources of similar type are preferred.They are operated at lower temperatures and still radiate sufficient energy. For fluorescent work,an intense beam of ultraviolet light is required. This requirement is met by a xenon arc or a mercuryvapour lamp. Cooling arrangement is very necessary when these types of lamps are used.
Mercury lamps are usually run direct from the ac power line via a series ballast choke. This
method gives some inherent lamp power stabilization and automatically provides the necessaryionizing voltage. The ballast choke is physically small and a fast warm-up to the lamp operatingtemperature is obtained.
Modern instruments use a tungsten-halogen light source, which has a higher intensity output
than the normal tungsten lamp in the change over region of 320–380 nm used in colorimetryand spectrophotometry. It also has a larger life and does not suffer from blackening of the bulbglass envelope. In the ultraviolet region of the spectrum, the deuterium lamp has superseded the
hydrogen discharge lamp as a UV source. The radiation sources should be highly stable and
preferably emit out a continuous spectrum.
Deuterium arc lamp provides emission of high intensity and adequate continuity in the 190–
380 nm range. A quartz or silica envelope is necessary not only to provide a heat shield, but alsoto transmit the shorter wavelengths of the ultraviolet radiation. The limiting factor is normally thelower limit of atmospheric transmission at about 190 nm. Figure 14.3 shows the energy output asa function of wavelength in the case of a deuterium arc lamp and a tungsten-halogen lamp.
In the modern spectrophotometers, the power supply arrangements including any necessary
start-up sequences for arc lamps as well as change-over between sources at the appropriate wave-length are automatic mechanical sequences. Lamps are generally supplied on pre-set focus mountsor incorporate simple adjustment mechanisms for easy replacement.
/G31/G34/G2E/G33/G2E/G32 /G4F/G70/G74/G69/G63/G61/G6C/G20/G46/G69/G6C/G74/G65/G72/G73
A filter may be considered as any transparent medium which by its structure, composition orcolour enables the isolation of radiation of a particular wavelength. For this purpose, ideal filtersshould be monochromatic, i.e. they must isolate radiation of only one wavelength. A filter mustmeet the following two requirements:
(a) high transmittance at the desired wavelength and (b) low transmittance at other wave-
lengths. However, in practice, the filters transmit a broad region of the spectrum. Referring to

392 Handbook of Biomedical Instrumentation
Fig. 14.4, they are characterized by the relative light transmission at the maximum of the curve Tl,
the width of the spectral region transmitted (the half-width—the range of wavelength between thetwo points on the transmission curve at which the transmission value equals 1/2 T
l) and Tres(the
residual value of the transmission in the remaining part of the spectrum). The ideal filter wouldhave the highest value of T
l and the lowest values for the transmission half-width and Tres.Filters
can be broadly classified as absorption filters and interference filters.400 200 600 800
lnmEnergyDeuterium
arc lampTungsten-
halogen lamp
Fig.14.3 Energy output as a function of wavelength for deuterium arc lamp and
Tungsten-halogen lamp
Radiation available
from source
Radiation
transmittedRadiation energyRadiation energyRadiation absorbed
or reflected by
colour filter
Wavelength-wave numberTresHalt-peak
intensity
Spectral
half
bandwidthPeak intensity
Tl
1/2Tl
Wavelength
Fig.14.4 Optical properties of a light filter

Clinical Laboratory Instruments 393
Absorption Filters: The absorption type optical filter usually consists of colour media: colour
glasses, coloured films (gelatin, etc.), and solutions of the coloured substances. This type of filterhas a wide spectral bandwidth, which may be 40 to 50 m in width at one-half the maximum
transmittance. Their efficiency of transmission is very poor and is of the order of 5 to 25%.
Composite filters consisting of sets of unit filters are often used. One combination set consists of
a long wavelength and sharp cut-off filters and the other of short wavelength and cut-off filters
combinations are available from about 360 nm to 700 nm.
Interference Filters: These filters usually consist of two semi-transparent layers of silver, deposited
on glass by evaporation in vacuum and separated by a layer of dielectric (ZnS or MgF
2). In this
arrangement, the semi-transparent layers are held very close. The spacer layer is made of a
substance which has a low refractive index. The thickness of the diaelectric layer determines thewavelength transmitted. Figure 14.5 shows the path of light rays through an interference filter.Some part of light that is transmitted by the first film is reflected by the second film and againreflected on the inner face of the first film, as the thickness of the intermediate layer is one-half awavelength of a desired peak wavelength. Only light which is reflected twice will be in-phase andcome out of the filter, other wavelengths with phase differences would cause destructive inter-ference. Constructive interference between different pairs in superposed light rays occurs onlywhen the path difference is exactly one wavelength or some multiple thereof.
Semi-transparent silver filmSlitSlitTransparent spacer film
(one half wavelength/inch)
Fig.14.5 Path of light rays through an interference filter
Interference filters allow a much narrower band of wavelengths to pass and are similar to
monochromators in selectivity. They are simpler and less expensive. However, as the selectivityincreases, the transmittance decreases. The transmittance of these filters varies between 15 to 60
per cent with a spectral bandwidth of 10 to 15 nm.

394 Handbook of Biomedical Instrumentation
For efficient transmission, multilayer transmission filters are often used. They are characterized
by a bandpass width of 8 nm or less and a peak transmittance of 60-95%. Interference filters can beused with high intensity light sources, since they remove unwanted radiation by transmission
and reflection, rather than by absorption.
/G31/G34/G2E/G33/G2E/G33 /G4D/G6F/G6E/G6F/G63/G68/G72/G6F/G6D/G61/G74/G6F/G72/G73
Monochromators are optical systems, which provide better isolation of spectral energy than the
optical filters, and are therefore preferred where it is required to isolate narrow bands of radiantenergy. Monochromators usually incorporate a small glass of quartz prism or a diffraction gratingsystem as the dispersing media. The radiation from a light source is passed either directly or bymeans of a lens or mirror into the narrow slit of the monochromator and allowed to fall on the
dispersing medium, where it gets isolated. The efficiency of such monochromators is much better
than that of filters and spectral half-bandwidths of I nm or less are obtainable in the ultravioletand visible regions of the spectrum.
Prism Monochromators: Isolation of different wavelengths in a prism monochromator depends
upon the fact that the refractive index of materials is different for radiation of different wavelengths.
If a parallel beam of radiation falls on a prism, the radiation of two different wavelengths will bebent through different angles. The greater the difference between these angles, the easier it is toisolate the two wavelengths. This becomes an important consideration for the selection of materialfor the prisms, because only those materials are selected whose refractive index changes sharplywith wavelength.
Prism may be made of glass or quartz. The glass prisms are suitable for radiations essentially in
the visible range whereas the quartz prism can cover the ultraviolet spectrum also. It is found thatthe dispersion given by glass is about three times that of quartz. However, quartz shows theproperty of double refraction. Therefore, two pieces of quartz, one right-handed and one left-handed are taken and cemented back-to-back in the construction of 60° prism (Cornu mounting),or the energy must be reflected and returned through a single 30° prism, so that it passes throughthe prism in both directions (Littrow mounting). The two surfaces of the prism must be carefullypolished and optically flat. Prism spectrometers are usually expensive, because of exacting
requirements and difficulty in getting quartz of suitable dimensions.
Diffraction Gratings: Monochromators may also make use of diffraction gratings as a dispersing
medium. A diffraction grating consists of a series of parallel grooves ruled on a highly polishedreflecting surface. When the grating is put into a parallel radiation beam, so that one surface of the
grating is illuminated, this surface acts as a very narrow mirror. The reflected radiation from this
grooved mirror overlaps the radiation from neighbouring grooves (Fig. 14.6).
The waves would, therefore, interfere with each other. On the other hand, it could be that the
wavelength of radiation is such that the separation of the grooves in the direction of the radiationis a whole number of wavelengths. Then the waves would be in phase and the radiation would bereflected undisturbed. When this is not a whole number of wavelengths, there would be destructiveinterference and the waves would cancel out and no radiation would be reflected. By changing theangle at which the radiation strikes the grating, it is possible to alter the wavelength reflected.

Clinical Laboratory Instruments 395
The expression relating the wavelength of the radiation and the angle ( q) at which it is reflected
is given by
ml = 2dsinq
where dis the distance separating the grooves and is known as the grating constant and mis the
order of interference.
When m= 1, the spectrum is known as first order and with m= 2, the spectrum is known as
second order.
The resolving power of a grating is determined by the product mN, where Nis the total number
of grooves or lines on the grating. Higher dispersion in the first order is possible when there are alarger number of lines. When compared with prisms, gratings provide a much higher resolvingpower and can be used in all spectral regions. Gratings would reflect, at any given angle, radiationof wavelength l and also l/2,l/3, etc. This unwanted radiation must be removed with filters or
premonochromators, otherwise it will appear as stray light.
Most modern instruments now use a diffraction grating as a dispersing element in the monochro-
mator, as prisms in general have a poorer stray light performance and require complex precisioncams to give a linear wavelength scale. Replica gratings can even be produced more cheaply thanprisms and require only a simple sine bar mechanism for the wavelength scale.
A typical reflection grating may have 1200 grooves/mm, which means the grooves are spaced
at about 800 nm intervals. The grating may have a width of 20 mm or more, giving a total of at least24,000 grooves. To obtain constructive interference across this large number of grooves with littlelight scattering, the spacing and form of the grooves must be accurate to within a few nanometers
to give a high quality grating. Development of mechanical diamond ruling engines which give a
high quality grating near to this accuracy gives rise to difficult technological problems.
Holographic Gratings: Precision spectrophotometers use holographic or interference gratings,
which have superior performance in reducing stray light as compared to diffraction gratings.
Holographic gratings are made by first coating a glass substrate with a layer of photo-resist,
which is then exposed to interference fringes generated by the intersection of two collimatedbeams of laser light. When the photo-resist is developed, it gives a surface pattern of parallelgrooves. When coated with aluminum, these become the diffraction gratings.
Compared with ruled gratings, the grooves of holographic gratings are more uniformly spaced
and shaped smoother. These characteristics result in much lower stray light levels. Moreover, theOut of phase
Incident lightInphase
Fig.14.6 Dispersion phenomenon in diffraction gratings

396 Handbook of Biomedical Instrumentation
holographic gratings can be produced in much less time than the ruled gratings. Holographic
gratings used in commercial spectrophotometers are either original master gratings produceddirectly by an interferometer or are replica gratings. Replica gratings are reproduced from a master
holographic grating by moulding its grooves onto a resin surface on a glass or silica substrate.
Both types of gratings are coated with an aluminum reflecting surface and finally with a protectivelayer of silica or magnesium fluoride. Replica gratings give performance which is as good as themaster gratings. The holographic process is capable of producing gratings that almost reach thetheoretical stray-light minimum.
/G31/G34/G2E/G33/G2E/G34 /G4F/G70/G74/G69/G63/G61/G6C/G20/G43/G6F/G6D/G70/G6F/G6E/G65/G6E/G74/G73
Several different types of optical components are used in the construction of analytical instruments
based on the radiation absorption principle. They could be windows, mirrors and simple
condensers. The material used in the construction of these components is a critical factor anddepends largely on the range of wavelength of interest. Normally, the absorbance of any materialshould be less than 0.2 at the wavelength of use. Some of the materials used are:
• Ordinary silicate glasses which are satisfactory from 350 to 3000 nm.
• From 300 to 350 nm, special corex glass can be used.• Below 300 nm, quartz or fused silica is utilized. The limit for quartz is 210 nm.• From 180 to 210 nm, fused silica can be used, provided the monochromator is flushed with
nitrogen or argon to eliminate absorption by atmospheric oxygen.
Reflections from glass surfaces are reduced by coating these with magnesium fluoride, which is
one-quarter wavelength in optical thickness. With this, scattering effects are also greatly reduced.
With a view to reduce the beam size or render the beam parallel, condensers are used. These
condensers operate as simple microscopes. To minimize light losses, lenses are sometimesreplaced by front-surfaced mirrors to focus or collimate light beam in absorption instruments.Mirrors are aluminized on their front surfaces. With the use of mirrors, chromatic aberrations andother imperfections of the lenses are minimized.
Beam splitters are used in double-beam instruments. These are made by giving a suitable multi-
layer coating on an optical flat. The two beams must retain the spectral properties of the incidentbeam. Half-silvered mirrors are often used for splitting the beam. However, they absorb some of thelight in the thin metallic coating. Beam splitting can also be achieved by using a prismatic mirroror stack of thin horizontal glass plates, silvered on their edges and alternatively oriented to theincident beam.
/G31/G34/G2E/G33/G2E/G35 /G50/G68/G6F/G74/G6F/G73/G65/G6E/G73/G69/G74/G69/G76/G65/G20/G44/G65/G74/G65/G63/G74/G6F/G72/G73
After isolation of radiation of a particular wavelength in a filter or a monochromator, it is essentialto have a quantitative measure of its intensity. This is done by causing the radiation to fall on aphotosensitive element, in which the light energy is converted into electrical energy. The electriccurrent produced by this element can be measured with a sensitive galvanometer directly or aftersuitable amplification.

Clinical Laboratory Instruments 397
Any type of photosensitive detector may be used for the detection and measurement of radiant
energy, provided it has a linear response in the spectral band of interest and has a sensitivity goodenough for the particular application. There are two types of photo-electric cells; photo-voltaic
cells and photo-emissive cells. These cells have been described in Chapter 3.
/G31/G34/G2E/G33/G2E/G36 /G53/G61/G6D/G70/G6C/G65/G20/G48/G6F/G6C/G64/G65/G72/G73
Liquids may be contained in a cell or cuvette made of transparent material such as silica, glass or
perspex. The faces of these cells through which the radiation passes are highly polished to keepreflection and scatter losses to a minimum. Solid samples are generally unsuitable for directspectrophotometry. It is usual to dissolve the solid in a transparent liquid. Gases may be containedin cells which are sealed or stoppered to make them air-tight. The sample holder is generally
inserted somewhere in the interval between the light source and the detector. For the majority of
analyses, a 10 mm path-length rectangular cell is usually satisfactory.
In analyses where only minimal volumes of liquid samples are practical, microcells, which
have volumes as small as 50 ml, can be employed. Most of the rectangular liquid cells have caps
and, for the analyses of extremely-volatile liquids, some of the cells have ground-glass stoppers toprevent the escape of vapour. Studies of dilute or weakly-absorbing liquid samples, or of sampleswhere trace components must be detected, require a cell with a long path-length. For suchapplications, a 50 cm path-length with about a 300 ml volume cell is employed.
Cylindrical liquid cells offer higher volume to path-length ratios than do rectangular cells, being
available in path-lengths of 20, 50 and 100 mm and in volumes of 4, 8, 20 and 40 ml, respectively.
Similar to these cylindrical cells are the demountable cells that have silica windows which are
easily removable. This demountable feature is especially useful for the containment of samplesthat are difficult to remove and clean from conventional cylindrical cells. Demountable cells areequipped with ground-glass stoppers. Figure 14.7 shows a selection of typical sample cuvettes.
/G20/G31/G34/G2E/G34 /G43/G4F/G4C/G4F/G52/G49/G4D/G45/G54/G45/G52/G53
A colorimetric method in its simplest form uses only the human eye as a measuring instrumentThis involves the comparison by visual means of the colour of an unknown solution, with the
colour produced by a single standard or a series of standards. The comparison is made by obtaining
a match between the colour of the unknown and that of a particular standard by comparison witha series of standards prepared in a similar manner, as the unknown. Errors of 5 to 20% are notuncommon, because of the relative inability of the eye to compare light intensities.
In the earlier days, visual methods were commonly employed for all colorimetric measurements,
but now photo-electric methods have largely replaced them and are used almost exclusivelyfor quantitative colorimetric measurements. These methods are more precise and eliminate thenecessity of preparing a series of standards every time a series of unknowns is run.
Strictly speaking, a colorimetric determination is one that involves the visual measurement of
colour and a method employing photoelectric measurement is referred to as a photometric orspectro-photometric method. However, usually any method involving the measurement of colour

398 Handbook of Biomedical Instrumentation
in the visual region of the electromagnetic spectrum (400–700 nm) is referred to as the colorimetric
method.
In a colorimeter, the sample is normally a liquid. The sample compartment of a colorimeter is
provided with a holder to contain the cuvette, in which the liquid is examined. Usually this holderis mounted on a slide with positions for at least two cuvettes, so that sample and reference cuvettesare measured first and a shutter is moved into or out of the light beam until the microammeter givesa full-scale deflection (100% T-scale reading). The sample is then moved into the beam and the
light passing through it is measured as a percentage to the reference value.
Sample concentration = Standard concentration ¥
Sample reading
Reference reading
Colorimeters are extremely simple in construction and operation. They are used for a great deal
of analytical work, where high accuracy is not required. The disadvantage is that a range of filtersis required to cover different wavelength regions. Also the spectral bandwidth of these filters islarge in comparison with that of the absorption band being measured. The basic component of acolorimeter are those which are shown in Fig. 14.2. They are available as single beam or doublebeam configurations.
/G31/G34/G2E/G34/G2E/G31 /G4D/G75/G6C/G74/G69/G2D/G63/G68/G61/G6E/G6E/G65/G6C/G20/G43/G6F/G6C/G6F/G72/G69/G6D/G65/G74/G65/G72/G20/G28/G50/G68/G6F/G74/G6F/G6D/G65/G74/G65/G72/G29
An increasing number of chemical analyses is carried out in the laboratories of industry andhospitals, and in most of these the final measurement is performed by a photometer. Obviously, Fig.14.7 Selection of sample cuvettes

Clinical Laboratory Instruments 399
it is possible to increase the capacity of the laboratory by using photometers, which have a large
measuring capacity. One of the limitations for rapid analyses is the speed at which the samplescan be transferred in the light path.
In a multi-channel photometer, instead of introducing one sample at a time into a single light
path, a batch of samples is introduced. Measurements are carried out simultaneously, using a
multiplicity of fiber optic light paths (Fig. 14.8) and detectors, and then the samples are scanned
electronically instead of mechanically. The 24 sample cuvettes are arranged in a rack in a three keyeight matrix. The 25th channel serves as a reference beam and eliminates possible source anddetector drifts. The time required to place the cuvette rack into the measuring position correspondsto the amount of time necessary to put one sample into a sample changer.
DVM and
computer
interfaceDigital logicLamp power
supplySourceChopperFilter disk
Fibre optics
light pathsSample traysDetectors
Commutator
Amplifier
251
Fig.14.8 Schematic of a multi-channel photometer
/G20/G31/G34/G2E/G35 /G53/G50/G45/G43/G54/G52/G4F/G50/G48/G4F/G54/G4F/G4D/G45/G54/G45/G52/G53
A spectrophotometer is an instrument which isolates monochromatic radiation in a more efficient
and versatile manner than colour filters used in filter photometers. In these instruments, light fromthe source is made into a parallel beam and passed to a prism or diffraction grating, where light ofdifferent wavelengths is dispersed at different angles.
The amount of light reaching the detector of a spectrophotometer is generally much smaller
(Fig. 14.9) than that available for a colorimeter, because of the small spectral bandwidth. Therefore,a more sensitive detector is required. A photomultiplier or vacuum photocell is generally employed.The electrical signal from the photoelectric detector can be measured by using a sensitive microam-meter. However, it is difficult and expensive to manufacture a meter of the required range andaccuracy. To overcome this problem, either of the following two approaches are generally adopted:
(a) The detector signal may be measured by means of an accurate potentiometric bridge. A
reverse signal is controlled by a precision potentiometer, until a sensitive galvanometershows that it exactly balances the detector signal and no current flows through thegalvanometer. This principle is adopted in the Beckman Model DU single-beam spectro-photometer.

400 Handbook of Biomedical Instrumentation
(b) The detector current is amplified electronically and displayed directly on an indicating
meter or in digital form. These instruments have the advantage in speed of measurement.As in the case of colorimeters, the instrument is adjusted to give a 100% transmissionreading, with the reference sample in the path of the light beam. The sample is then movedinto the beam and the percentage transmission is observed.
Modern commercial instruments are usually double beam, digital reading and/or recording
instruments, which can provide absorbance, concentration, per cent transmission and differential
absorbance readings. It is also possible to make reaction rate studies. They can be used to include
specialized techniques, such as automatic sampling and batch sampling, with the addition ofcertain accessories. The measurements can be generally made with light at wavelengths from 340to 700 nm and from 190 to 700 nm with a deuterium source. In variable-slit type of instruments, theslit can be made to vary from 0.05 to 2.0 mm. The wavelength accuracy is ± 0.5 nm. The recordersare usually single channel, strip-chart potentiometric recorders. They are calibrated from 0.1 to 2.0A or 10 to 200% T full-scale. The recorder used with spectrophotometers have four wavelength
scanning speeds (100, 50, 20 and 5 nm/mm) and seven chart speeds (10, 5, 2, 1, 0.5, 0.2 and 0.1inch/mm). It has a sensitivity of 100 mV absorbance units or 100 mV/100% T.
When scanning a narrow wavelength range, it may be adequate to use a fixed slit-width. This
is usually kept at 0.8 mm. In adjustable slit-width instruments, it should be so selected that theresultant spectral slit-width is approximately 1/10th of the observed bandwidth of the sample, i.e.if the absorption band is 25 nm wide at half of its height, the spectral slit-width should be 2.5 nm.This means that the slit width set on the instrument should be 1.0 mm. This is calculated from the
dispersion data, as the actual dispersion in grating instruments is approximately 2.5 nm/mm
slit-width.Radiation energy
from monochromator
Radiation energy
transmitted
by colour filterRadiation available
from source
Radiation energy
Wavelength—wave number
Fig.14.9 Comparison of radiation energy from a monochromator and colour
filter

Clinical Laboratory Instruments 401
Spectrophotometers generally employ a 6 V tungsten lamp, which emits radiation in the wave-
length region of visible light. Typically, it is 32 candle power. These lamps should preferably beoperated at a potential of say 5.4 V, when its useful life is estimated at 1200 h. The life is markedly
decreased by an increase in the operating voltage. With time, the evaporation of tungsten produces
a deposit on the inner surface of the tungsten lamp and reduces emission of energy. Dark areas onthe bulb indicate this condition. It should then be replaced.
The useful operating life of the deuterium lamp normally exceeds 500 h under normal condi-
tions. The end of the useful life of this lamp is indicated by failure to start or by a rapidly decreasingenergy output. Ionization may occur inside the anode rather than in a concentrated path in frontof the window. Generally, this occurs when the lamp is turned on while it is still hot from previousoperation. If this occurs, the lamp must be turned off and allowed to cool before restarting.
Wavelength calibration of a spectrophotometer can be checked by using a holmium oxide filter
as a wavelength standard. Holmium oxide glass has a number of sharp absorption bands, whichoccur at precisely known wavelengths in the visible and ultraviolet regions of the spectrum.Holmium oxide filter wavelength peaks are given below:
Ultraviolet range with deuterium lamp : 279.3 and 287.6 nm
Visible range with tungsten lamp : 360.8, 418.5, 453.4,536.4 and 637.5 nm
The wavelength calibration can also be checked in the visible region by plotting the absorption
spectrum of a didymium glass, which has been, in turn, calibrated at the National Bureau ofStandards, USA.
/G31/G34/G2E/G35/G2E/G31 /G4D/G69/G63/G72/G6F/G70/G72/G6F/G63/G65/G73/G73/G6F/G72/G20/G42/G61/G73/G65/G64/G20/G53/G70/G65/G63/G74/G72/G6F/G70/G68/G6F/G74/G6F/G6D/G65/G74/G65/G72/G73
In spectrophotometry, computers have long been used, especially for on line or off line data
processing. Since the advent of the microprocessor, their application has not only been limited to
processing of data from analytical instruments, but has also been extended to control of instrumentfunctions and digital signal processing, which had been performed conventionally by analogcircuits. This has resulted in improved performance, operability and reliability over purely analoginstruments.
A microprocessor, in a spectrophotometer, could be used for the following functions:
•Control functions: Wavelength scanning, automatic light source selection, control of slit-
width, detector sensitivity, etc.
•Signal processing functions: Baseline correction, signal smoothing, calculation of % T,
absorbance and concentration, derivative, etc.
•Communication functions: Keyboard entry, menu-driven operations, data presentation,
warning display, communication with external systems, etc.
Figure 14.10 shows a block diagram of a microprocessor controlled spectrophotometer. The
diagram shows only the post-detector electronic handling and drive systems, all controlled via asingle microprocessor. Once the operator defines such parameters as wavelength, output modeand relevant computing factors, the system automatically ensures the correct and optimum combi-nation of all the system variables. Selection of source and detector are automatically determined;

402 Handbook of Biomedical Instrumentation
any filters introduced at appropriate points and sample and reference cells are correctly managed
in the sample area. Output in the desired form (transmittance, absorbance, concentration, etc.) is
presented along with the sample identification. Secondary routines such as wavelength calibra-tion and self-tests become available on demand.
For wavelength scanning, a stepper motor is used, which ensures accurate and fast scanning.
The automatic selection of samples is also made with a motor driven system under the control ofa microprocessor.
The signal from the photodetector is amplified in a preamplifier and converted into digital form
in an A-D converter. The signals are differentiated into sample signal S, reference signal Rand
zero signal Z and stored in the memory. From these values, the microprocessor calculates the
transmittance T = (S-Z/R-Z ) and absorbance = –Log T.In order to obtain RorS values within a
specified range, the microprocessor provides control signals for slit-width and high voltage for thephotomultiplier.
Baseline compensation due to solvent and optical unmatching of cells, which has been difficult
with conventional instruments is conveniently possible in microprocessor based systems.Improvements have also been achieved in such functions as auto-zero, expanding and contractingof the photometric scale, automatic setting of wavelength as well as in ensuring repeatable andmore accurate results.Display
Micro-
processorLamp
control
Converter
Detector
RS 232
and
local
networkMemory
Recorder
output
Motor
driveMotor
driveMotor
drive
Filter
disc
motorGrating
motorCell
holder
motor
Fig.14.10 Block diagram of a microprocessor controlled spectrophotometer

Clinical Laboratory Instruments 403
The digital output from the microprocessor is converted into analog form with a D-A converter
and given to an X-Y recorder as the Y-axis signal, whereas the wavelength forms the X-axis, to
obtain absorption or reflected spectra. Microprocessors have also enabled the making of such
measurements as higher order derivative spectra and high speed sampling and storage of fast
reaction processes and for presenting processed data during and after the completion of thereaction.
PC-based spectrophotometers are now commonly available. In these instruments, wavelength
drive, slit drive, filter wheel drive, source selector, mirror drive, detector change and grating driveare all carried out by stepper motors. Stepping control is effected by the computer, with the pulsefrequency depending upon the individual scan speed. The computer processes the data suppliedby the microprocessor and transmits it with the suitable format to the video display screen andprinter/plotter.
/G20/G31/G34/G2E/G36 /G41/G55/G54/G4F/G4D/G41/G54/G45/G44/G20/G42/G49/G4F/G43/G48/G45/G4D/G49/G43/G41/G4C/G20/G41/G4E/G41/G4C/G59/G53/G49/G53/G20/G53/G59/G53/G54/G45/G4D/G53
/G31/G34/G2E/G36/G2E/G31 /G54/G68/G65/G20/G53/G79/G73/G74/G65/G6D/G20/G43/G6F/G6E/G63/G65/G70/G74/G73
The development of new concepts and more advanced techniques in analytical methodology have
resulted in the estimation of blood constituents as a group, whose metabolic roles are related and
which collectively provide more meaningful information than the individual analyses. Forinstance, the group of important anions and cations of the blood plasma (electrolytes) like sodium,potassium, chloride and bicarbonate, which together with serum urea form a related set of tests—performed on patients with electrolyte disturbances. Another group consists of the analyses—protein, bilirubin, alkaline phosphatase and SGOT, which together assess liver function. Theeffect of this trend is in the replacement of single isolated analysis by groups of analyses, all ofwhich are carried out routinely on each sample with highly reproducible and accurate results.With this object in view, automatic analysis equipment have been designed and put to use.Automated analysis systems are available in multichannel versions and a full description of thedetailed working instructions and details of techniques for individual substances are given in theliterature supplied by the manufacturers to the purchaser of their equipment. The major benefit of
automation in the clinical laboratory is to get rid of the tasks that are repetitive and monotonous
for a human operator, which may lead to improper attention that may cause an error in analysis.However, it may be remembered that improvement in reproducibility does not necessarily enhancethe accuracy of test results, because accuracy is basically influenced by the analytical methodsused. The automated systems are usually considered more reliable than normal methods, due toindividual variations that may appear in handling various specimens.
The automated system is usually a continuous flow system, in which individual operations are
performed on the flowing stream as it moves through the system. The end-product passes throughthe colorimeter, where a balance ratio system is applied to measure concentrations of variousconstituents of interest. The final results are recorded on a strip-chart recorder along with acalibration curve, so that the concentration of the unknowns can be calculated. The output mayalso be connected to a digital computer to have a digital record along with the graphic record.

404 Handbook of Biomedical Instrumentation
/G31/G34/G2E/G36/G2E/G32 /G54/G48/G45/G20/G53/G59/G53/G54/G45/G4D/G20/G43/G4F/G4D/G50/G4F/G4E/G45/G4E/G54/G53
The automated system consists of a group of modular instruments (Fig. 14.11) interconnected
together by a manifold system and electrical systems. The various sub-systems are:
• Sampling unit
• Proportioning pump• Manifold• Dialyzer• Heating bath or constant temperature bath• Colorimeter/flame photometer/Fluorometer
• Recorder
• Function monitor
Sampler Pump Mixer heater
dialyzerRecorder Colorimeter Digital printer
94-8
84-682-4
PhotocellLampFlow
cellDialysis
Mixing
ReagentAirAir
sampleDiluent
Fig.14.11 Schematic diagram of an automated continuous flow type analysis
system
The sample to be analyzed is introduced into a stream of diluting liquid flowing in the narrow
bore of a flexible plastic tube. The stages of the analytical reaction are completed by the successivecombination of other flowing streams of liquids with the sample stream, by means of suitablyshaped glass functions. Bubbles of air are injected into each stream, so that the liquid in the tubesis segmented into short lengths separated by air bubbles. This segmentation reduces the tendency
for a stationary liquid film to form on the inner walls of the tubes and decreases the interaction
between a sample and the one which follows it. The diluted samples and reagents are pumpedthrough a number of modules, in which the reaction takes place, giving a corresponding sequenceof coloured solutions, which then pass into a flow-through colorimeter. The corresponding extinc-tions are plotted on a graphic recorder, in the order of their arrival into the colorimeter cell. The airbubbles are removed before the liquid enters the colorimetric cell or flame emission. Details of theindividual units are described below:
Sampling Unit: The sampling unit enables an operator to introduce unmeasured samples and
standards into the auto-analyzer system. The unit in its earlier form consisted of a circular turntable(Fig. 14.12) carrying around its rim 40 disposable polystyrene cups of 2 ml capacity. The sampleplate carrying these cups rotates at a predetermined speed. The movement of the turntable issynchronized with the movements of a sampling crook. The hinged tubular crook is fitted at a

Clinical Laboratory Instruments 405
corner of the base. The crook carries a thin flexible polythene tube, which can dip into a cup and
allow the contents—water, standard or test solution—to be aspirated. At regular intervals, the
crook is raised, so that the end of the sample tube is lifted clear of the cup.
Between each sampling, the crook enters a receptacle of water or other suitable wash fluid, to
reduce cross-contamination of one sample with another. The ratio of sampling time to wash timeis normally 2:1. The plate then rotates a distance sufficient to allow the tube, when it next movesdown, to dip into the next cup. One complete rotation of the plate thus presents 40 samples. As thesample plate completes a cycle, a switch is operated, which stops the rotating action of the plate
and the sampling action of the sample probe. The sampling rate can be adjusted to 20, 40 or 60 per
hour. According to the above ratio, the time during which a sample is being drawn in, will be twominutes, one minute or 40s respectively. The volume of liquid taken up in most cases ranges fromabout 0.2 to 1.0 ml. This depends upon the rate at which the plate is run and the diameter of thepump tube.
The earlier version has been replaced by a more versatile form of the sampler, in which during
the time the sample tube is out of the specimen, the crook quickly comes down into the water, andthus successive samples are separated by a column of water instead of air. This provides a betterseparation between them. With this sampler, the sample size may range from 0.1 to 8.5 ml. Itutilizes cups of sizes 0.5, 2, 3 and 10 ml. The sample plate is kept covered to prevent evaporation,which may sometimes lead to errors up to 5%. Sampling and washing periods are controlled by aprogramming cam. The sample speed and sample wash cycles are selected by the markings on thecam, such as 40 and 2:1. This implies that the speed is 40 per hour at a sample wash ratio of 2:1.
Mechanical cams were used in the earlier modules of auto-analyzers to initiate and control
sample aspiration and wash cycles. Modern systems use electronic timers to do the same function.These timers provide a greater flexibility in the control of the sample-to-wash ratios, which in turn
allows flexibility in setting up parameters for analyses.Sample cupProbe
armTimer(A)
(B)Programmer
Timing
cam
Timing
motor
Sample
advance switchSampler
drive motorSample
tray
(A) For SMA system
(B) For auto-analyzer
Fig.14.12 Sampler controls (Courtesy: M/s Technicon Corp., USA)

406 Handbook of Biomedical Instrumentation
The Proportioning Pump: The function of the proportioning pump is to continuously and
simultaneously push fluids, air and gases through the analytical chain. In fact, it is the heart of theautomatic analysis system. Here, all the sample and reagent streams, in any particular analysis,
are driven by a single peristaltic pump, which consists of two parallel stainless steel roller chains
with finely spaced roller thwarts.
A series of flexible plastic tubes, one from the sampler, the others from reagent bottles or which
simply draw in air are placed lengthwise along the platen spring-loaded platform. The roller-head assembly is driven by a constant-speed gear motor. When the rollers are pressed down andthe motor switched on, they compress the tubes containing the liquid streams (sample, standardand reagents) against the platen. As the rollers advance across the platen, they drive the liquidbefore them.
The roller head rotates at a constant speed. The variable flow rates required in the different
streams (0.15 to 4 ml/min) are achieved by selecting tubes of appropriate internal diameter, but ofconstant wall thickness. Since the proportions of the various reagents are fixed by the tube sizes,no measurements are needed.
Proportioning pumps are available either for single-speed or for two-speed operations. The
single-speed pump has the capacitor synchronous gear head utilizing 10 rpm output shaft at50 Hz. The two-speed pump has a non-synchronous 45 rpm motor. The slow speed in this pumpis used for the ordinary working during a run and a much quicker one for filling the system withreagents before a run and for rapid washing to clear out reagents after the run. It is also utilized forrapid cleaning of the heating bath, or of the complete system, when fibrin (an insoluble protein)
problems are evident and are disturbing the run. High speed is not used for analysis. Heavy-duty
pumps are also available, which enable 23 pump tubes to be utilized simultaneously.
The plastic tubes are held taut between two plastic blocks having locating holes, which fit on to
pegs at each end of the platen. Before beginning a run, the tubes are stretched. With use and time,the tubes loose elasticity and pumping efficiency is reduced. Therefore, each block has three sets ofholes, so that the tubes can be increasingly stretched and the tension thus maintained. The tubesare replaced at the first sign of aging. In fact, they should be replaced at regular intervals to fore-stall failure. When not in use, one of the blocks is removed, so that the tubes are not kept in tension.
Actually, the sample or reaction stream is separated by air bubbles into a large number of dis-
tinct segments. The air bubbles completely fill the lumen of the tubing conducting the flow, therebymaintaining the integrity of each individual aliquot. In addition, the pressure of the air bubbleagainst the inner wall of the tubing wipes the surface free of droplets which might contaminate thesamples which follow. The proportioning includes an air bar device (Fig. 14.13), which adds airbubbles to the flowing streams in a precise and timed sequence. The air bar is actually a pinchvalve connected to the pump rollers that occludes or opens the air pump tubes at timed intervals.Every time a roller leaves the pump platen—and this occurs every two seconds—the air bar rises
and lets a measured quantity of air through. The release of air into the system is carefully controlled,
thereby insuring exactly reproducible proportioning by the peristaltic pump.
The continuous flow analyzers make use of liquid reagents. Large volumes of reagents are
stored in the systems and their quantity is adequate for the operation of the analyzer for severalhours or days. Some automated systems use reagents in a dry tablet form. When required, thetablet is dispensed into a one-test reaction vessel and dissolved. The sample is then added for the

Clinical Laboratory Instruments 407
reaction to take place. This is basically a unit-dose concept, which offers several advantages like
less storage space and operator time, long stability of reagents and lesser wastage.
Manifold: A manifold mainly consists of a platter, pump tubes, coils, transmission tubing, fittings
and connections. A separate manifold is required for each determination and the change can beeffected within a few minutes. The pump tubing and the connected coils are placed on a manifold
platter, which keeps them in proper order for each test. The pump tubing are specially made, are of
premeasured length and are meant to introduce all constituents of an analysis into the system. Thephysical and chemical properties of the tubing are extremely important in the correct functioningof the pump. It must not be so flexible as to expand beyond its normal internal dimensions onrelease of pressure, which may lead to variation in the flow, thereby affecting reproducibility andaccuracy of the system. The tubes should be chemically inert for the constituents which are expectedto flow through the tube. The constant and correct tension also provides the continual delivery ofa constant volume. The inside diameter of the pump tubing determines the flow rate per minute.
Several other tubes are required to introduce reagents and to transport the specimen from one
module to another. There are five types of such tubings. They are of varying sizes and are to beselected according to the requirements. These are: standard transmission tubing (Tygon), solvaflextubing, acidflex tubing, polyethylene tubing and glass tubing.
Two types of coils are employed in the system—mixing coils and delay coils. Coils are glass
spirals of critical dimensions, in which the mixing liquids are inverted several times, so thatcomplete mixing can result.
Mixing coils are used to mix the sample and/or reagents. As the mixture rotates through a coil,
the air bubble along with the rise and fall motion produces a completely homogeneous mixture.The mixing coils are placed in a horizontal position to permit proper mixing. Delay coils are
employed when a specimen must be delayed for the completion of a chemical reaction before
reaching the colorimeter. These coils are selected in length according to the requirements. Thestandard delay coil is 40 ft long, 1.6 mm I.D., and has a volume of approximately 28 ml. The timedelay can be calculated by dividing the volume of coil by the flow rate of specimen plus bubbles.
Phasing: With 12 tests to be recorded on each sample and a sampling rate of 60 samples per hour,
it follows that 5 s are allowed to record each steady state plateau. The reaction streams in the 12channels and up to four blank channels must, therefore, be phased to arrive at the colorimeter inAir bar
1 sec
2 sec
Proportionating pump
Fig.14.13 Principle of air segmentation in the continuous flow system

408 Handbook of Biomedical Instrumentation
waves 5 s apart. For example, if the cholesterol stream arrives at X time, calcium must arrive at
X + 5 s, total protein at X+ 10 s, albumin at X+ 15 s, etc. In order to ensure a proper sequencing
for the presentation of results, a number of devices have been provided to make this adjustment an
extremely simple operation. Phasing coils are used to permit the channels to enter the colorimeter
in the proper sequence.
Dialyzer: In analytical chemistry, it is often necessary to remove protein cells to obtain an
interference-free analysis. This is accomplished by dialysis in the auto-analyzer. The dialyzer
module (Fig. 14.14) consists of a pair of perspex plates, the mating surfaces of which are mirror
grooved in a continuous channel, which goes in towards the centre, on itself and returns to theoutside. A semi-permeable cellophane membrane is placed between the two plates and theassembly is clamped together, similar to the kidney dialyzer. The continuous groove channel thusgets divided into two halves and the dialysis occurs across the membrane. A solution containingthe substance to be analyzed passes along one-half, usually the upper one, of the channel, whilethe solvent that is receptive to the substance to be removed enters the other half. The substance tobe separated from the sample diluent stream, will diffuse through the semi-permeable membraneby osmotic pressure into the recipient stream and the non-diffusable particles will be left behind.
Membrane
Upper dialyzer
plate
Dialyzer
grooves
Membrane
Lower
dialyzer plateMicro
moleculesMacro moleculesAir
Sample
stream
(donor)
recipient
stream
Fig.14.14 Simplified block diagram of the dialysis process
The cellophane membrane usually used in the dialyzer has a pore size of 4–6 mm. The rate of
dialysis is stated to be dependent upon the temperature, area, and concentration gradient. For this
reason, the dialyzer unit is usually immersed in a water bath maintained at a constant temperature(37± 0.1°C). The temperature is kept constant with a thermostatically controlled heater and a

Clinical Laboratory Instruments 409
motorized stirrer. Both streams pass through preheating coils, before entering the dialyzer unit.
The channel path is 87 inch long, which provides a large surface presentation to the dialyzingmembrane. The plates of the dialyzer must be a matched set. If the plates are not a matched set, the
channels may be slightly off, causing leakage, poor bubble patterns and loss of dialyzing area,
which would ultimately result in the loss of sensitivity.
The quantity of solute that passes through the membrane in the dialyzer is determined by the
concentration gradient across the membrane, the duration of contact of the two solutions, the areaof contact, the temperature and by the thickness and porosity of the membrane. Other factorswhich affect the rate of transfer are the size and shape of the molecules, their electrical charge andthe composition of the fluids across the membrane.
A decrease in the flow rate of the liquid streams increases sensitivity in continuous flow
systems, since more concentrated samples and thinner membranes can be used. Modern dialyzers,therefore, have shallower and shorter grooves, resulting in reduced sample interaction and carryover. Membranes have been found to age with use and time due to protein deposition on theirsurface and therefore need periodical replacement. In the recent systems, the computer informs theoperator to investigate the need for membrane replacement.
Heating Bath: On leaving the dialyzer, the stream may be combined by one or more additional
reagents. It is then passed to a heating bath. This module is not used in all the tests performed bythe auto-analyzer. The heating bath is a double-walled insulated vessel, in which a glass heatingcoil or helix is immersed in mineral oil. A thermostatically controlled immersion heater maintainsa constant temperature within ±0.1°C, which can be read on a thermometer. Inside the bath, thestream passes along a helical glass coil about 40 ft long and 1.6 mm I.D., immersed in oil, which isconstantly stirred. The heating bath may have a fixed temperature, as 95° or 37°C or an adjustable
value. Passage through the heating coil takes about five minutes, but it would obviously vary with
the rate at which the liquid is moving, which, in turn, depends on the diameter of the tubes in themanifold.
Measurement Techniques : Although automated analyzers are mostly using absorption spectro-
photometry as the major measurement technique, several other alternative photometric approaches
have been utilized in the recent years. These are reflectance photometry, fluorometry, nephelometryand fluoresence polarization. Use of ion-selective electrodes and other electrochemical measurementtechniques are also becoming popular.
In the absorption photometers used in automated systems, the radiant energy sources employed
include tungsten, quartz, halogen, deuterium, mercury and xenon lamps and lasers. The spectrumcovered is usually from 300 to 700 nm. Spectral isolation is generally achieved with interference
filters in most automated systems. These filters have peak transmittance of 3~80% and bandwidthsof 5-15 nm. The filters are usually mounted on a rotating wheel and the required filter is brought inposition under the command of a microprocessor. Some automated systems also make use ofmonochromators, which obviously provide greater flexibility for the development and addition ofnew assays. The most popular detector used in the automated systems is the photomultiplier
although some of the recent systems also employ photodiodes.
Proper alignment of flow cells or cuvettes with the light path is as important in automated
systems as in manual methods. Stray energy and internal reflections are required to be kept as low

410 Handbook of Biomedical Instrumentation
as possible, to approximately less than 0.2%. This is usually done by careful design of the
wavelength isolation filters, or monochromator, or by the use of dual filters to increase rejection ofstray light.
The colorimeters used in the automated systems continuously monitor the amount of light
transmitted through the sample. They employ flow-through cuvettes. In the earlier designs of flow
cells, the arrangement was such that as the incoming stream entered the cell, the air bubbles
escaped upwards through an open vent, so that a continuous stream of liquid could fill the cellbefore going to waste. The flow cell size varied from 6 to 15 mm. The later designs of flow cells areall of tubular construction. This requires a much smaller volume of fluid, so that a smaller volumeof sample can be used. Being completely closed, it does not require separate cleaning. Before thestream enters the flow cell, it is pumped to a debubbler, where the air bubbles are removed. Thestream is then pulled through the flow cell under the action of another pump.
Some automatic continuous flow chemical analysis systems, incorporate a multi-channel
colorimeter The colorimeter employs a fiber optics system, using a single high-intensity quartziodine lamp, which is coupled to a highly stabilized power supply. This lamp passes its energythrough light guides, which are connected to up to five independent colorimeter modules. Eachcolorimeter has two detectors, one for the sample cell and the other for the reference beam. Aninsulating wall is placed between the light source and the colorimeter modules, so as to maintaingood temperature and electrical stability. The colorimeter contains a glass continuous-flow sample
cell with an integral debubbler.
Light guides from the main light source plug into one side of the cell block and the photo-
detectors are placed on the other side of the block. The filters are inserted into a slot between the
light guide and flow cell. The detectors are closely matched and their output is connected to ahighly stabilized bridge circuit. The output fed to the recorder is logarithmic, hence for solutionswhich obey Beer’s Law, the recorder output is linear in concentration over the normal range ofoptical densities observed in a continuous-flow automatic analysis system. The measurement ofsodium and potassium is carried out by a flame photometer. Fluorimetric analysis permitsmeasurements to be made at concentrations as low as 0.01 part per billion. The fluorimeters usedfor automated work, like colorimeters, have flow-through systems. The continuous flow cuvette ismade of Pyrex glass, which transmits light from the visible region to approximately 340 nm. Forthe ultraviolet region below 320 nm, quartz cuvettes are available.
Signal Processing and Data: The availability of low-cost microprocessors had a major impact on
the signal processing and data handling of analytical procedures in automated systems. Real-time acquisition and processing of data, by means of specific algorithms, so that the output isimmediately useful and meaningful, has become possible. Transformation of complex, non-linearstandard responses into linear calibration curves have allowed automation of procedures, suchas reflectance spectrometry. Specifically, microprocessors are now being used in automated
methods for the following functions:
1. Complete control of the electromechanical operation of the analyzer in relation to the
transfer of solutions, selection and placement of proper filters and continuous monitoringof the operation. This ensures that all functions are performed uniformly, repeatably andin the correct sequence.

Clinical Laboratory Instruments 411
2. Acquisition, assessment, processing and storing of operational data from the analyzer.
3. Providing effective communication between the analyzer and the operator through alpha-
numeric display on the CRT. Some systems even monitor the equipment function and giveout messages describing the site and type of problem in a malfunctioning equipment.
4. Facility to communicate to with the main-frame computer through the RS-232 interface for
an integration of the instrument with laboratory information.
5. Facility to communicate over the telephone lines, using a modem, with the manufacturer’s
central service department, thereby enhancing ability of the on-site operator to service andrepair the analyzer.
All 12 tests for each sample are reported in directly readable concentration units, on a single
strip of precalibrated chart paper. Since the normal ranges for each parameter are printed asshaded areas, the physician does not have to remember the normal values. Thus each abnormalitystands out clearly.
Actual measurement are made only after the
analytical curve (Fig. 14.15) reaches its steady-stateplateau (equilibrium condition in the system atwhich there are no changes in concentration withtime). At this steady-state plateau, all effects ofpossible sample interaction have been eliminated,
and the recorded signal gives a true reflection of
the concentration of the constituent being measured.Herein lies the importance of segmenting each ofthe sample and reagent streams with air bubbles.In effect, the air bubbles act as barriers to divideeach sample and reagent stream into a large numberof discrete liquid segments. Equally important, theair bubbles continually scrub the walls of the tub-ing. This sequential wiping of the walls diminishesthe possibility of contamination in succeeding segments of the same sample. Thus, should there beany interaction between two samples, it can easily be seen that the effects of this interaction willoccur only in the first few segments of the second sample. In the middle segments, the air bubbles
immediately preceding will have effectively cleansed the system and prevented further interaction.
It is these middle and final segments, free from interaction, which are recorded as the steady-stateplateau and appear as flat lines on the graph.
/G20/G31/G34/G2E/G37 /G43/G4C/G49/G4E/G49/G43/G41/G4C/G20/G46/G4C/G41/G4D/G45/G20/G50/G48/G4F/G54/G4F/G4D/G45/G54/G45/G52/G53
The flame photometer is one of the most useful instruments in clinical analyses. This is due to thesuitability of the flame photometer for determining sodium, potassium and calcium, which are ofimmense importance in the development of the living being and are indispensable for its physio-logical functions. In the clinical analysis of sodium and potassium, the flame photometer gives,rapidly and accurately, numerous differential data for normal and pathological values.Sample shown
at steady state
Fig.14.15 Typical curve showing the
steady state conditions whenmeasurements are made in acontinuous flow system

412 Handbook of Biomedical Instrumentation
The method of flame photometric determinations is simple. A solution of the sample to be
analyzed is prepared. A special sprayer operated by compressed air or oxygen is used to introducethis solution in the form of a fine spray (aerosol) into the flame of a burner operating on some fuel
gas, like acetylene or hydrogen. The radiation of the element produced in the flame is separated
from the emission of other elements by means of light filters or a monochromator. The intensity ofthe isolated radiation is measured from the current it produces when it falls on a photocell. Themeasurement of current is done with the help of a galvanometer, whose readings are proportionalto the concentration of the element. After carefully calibrating the galvanometer with solutions ofknown composition and concentration, it is possible to correlate the intensity of a given spectralline of the unknown sample, with the amount of the same element present in a standard solution.
A flame photometer has three essential parts (Fig 14.16). These are:
(a)Emission System: Consists of the following:
(i)Fuel gases: and their regulation: comprising the fuel reservoir, compressors, pressure
regulators and pressure gauges.
(ii)Atomizer: consisting, in turn, of the sprayer and the atomization chamber, where the
aerosol is produced and fed into the flame.
(iii) Burner: receives the mixture of the combustion gases.
(iv) Flame: the true source of emission.
(b)Optical System: It consists of the optical system for wavelength selection (filters or mono-
chromators), lenses, diaphragms, slits etc.
(c)Recording System: It includes detectors like photocells, photo-tubes, photomultipliers, etc.
and the electronic means of amplification, measuring and recording.
Measuring
recording
systemAmplification
systemDetection
systemSampleGases
AtomizerOptical
systemSelection
systemEmission system
Recording system
Fig.14.16 Essential parts of a flame photometer
Dedicated instruments for the simultaneous analysis of sodium, potassium and lithium are
available. In these instruments, sample handling is automatic, as the system has a turntable,which will hold up to 20 samples in cups and an automatic positive piston displacement dilutor,that dilutes the sample prior to entering the spray chamber.

Clinical Laboratory Instruments 413
Ignition and shutdown of the flame are automatic. When the calibrate button is depressed, the
flame is ignited and the circuits are energized. Passing a standby button extinguishes the flame,but maintains thermal equilibrium. Twenty-four microlitres of the sample are aspirated for the
simultaneous measurement of sodium and potassium, providing microsample analysis suitable
for paediatric or geriatric work. After analysis, the results are displayed directly in millimoles perlitre or milliequivalents per litre.
For precise and accurate determination of NaandKconcentrations, use is made of the fact that
lithium normally not present in significant concentrations in serum, exhibits about the same flameemission characteristics as Na and K. Lithium ions are added to the diluent used for samples,standards and controls. The lithium in the diluent is referred to as the internal standard.
The diluted sample containing the fixed known amount of lithium, in the form of a dissolved
salt, is nebulized and carried by the air supply into the first of two compartments in the spraychamber (Fig 14.17 ).Heavier droplets fall out of the stream on to the chamber walls or separate
from the stream upon striking a partition in the chamber and flow to a drain from the compartment.Propane enters the first chamber to mix with the air and sample stream, and carry it through atubular glass bridge into a second compartment. The aerosol and propane mixture travels up fromthe chamber to the burner head, where the mixture is burned. Exhaust gases are vented to room airfrom a cover located on top of the instrument.
BurnerFlame
Propane
Air
Diluted
sample
Drain
Fig.14.17 Flame spray chamber and burner of KliNa flame-photometer (Courtesy:
M/s Beckman Instruments, USA)
To provide internal standardization, the response of the sodium and potassium detector is a
ratio function of the response by the lithium detector. Thus, any change in air-flow rate or fuelpressure that may affect the sample would proportionately affect the lithium detector.
The flame photometer determines and provides digital display of sodium and potassium or
lithium concentrations in a sample, by responding optically and electronically to the intensity of

414 Handbook of Biomedical Instrumentation
the principal emission lines that characterize each ion as it is excited in the propane and air flame.
Figure 14.18 shows the schematic diagram of a flame photometer.
PMTPMT
PMT
Printer and
interfaceHV controlLKN 143
5.0
Displays Switch
Reference
voltageSystem
monitor
Changer DilutorGas, air
diluted sampleFlameFilters
Fig.14.18 Schematic diagram of a flame photometer
The flame is monitored continuously by three photomultiplier detectors. Each detector views
the flame through an optical filter that passes only that wavelength band which is of interest to theparticular detector. The sodium detector therefore, responds only to wavelengths in a narrowband centered at 589 nm; the potassium detector responds only to wavelengths in a narrow bandcentered at 766 nm; the lithium detector responds only to wavelengths in a narrow band centeredat 671 nm.
The system accuracy is ± 0.2 mmol/litre for potassium and lithium, and ±2.0 mmol/litre for
sodium. Potassium and lithium both show linearity to 20 mmol/litre, while sodium is linear to200. In addition to the 0–20 scale, potassium may be rescaled to readout to 200 mmol/litre forconvenient analysis of urine samples.
Modern flame photometers come with many useful accessories. For example, the diluter is a
motor driven cam-programmed system that functions through a cycle of operations. Theseoperations involve sample pick-up and transport to an internal mixing cup, the addition of ameasured volume of diluent, mixing to ensure a properly prepared sample aliquot, coupling of themixing cup to the spray chamber, so that sample aspiration can occur, and finally the washing
and draining of the mixing cup. The diluter uses positive displacement pumps to assure exact
sample dilutions in the operator selectable ratios of 50:1, 100:1 or 200:1.
The automatic changer enables automatic presentation to the diluter sample probe up to 20
successive samples. It is a turntable, which rotates stepwise to locate each sample cup under theextended and down position of the diluter sample probe. The probe is extended from the diluter,once for each sample determination. The probe tip enters the sample and a measured volume istaken for transport to the diluter mixing cup. Individual sample trays are placed on the changerturntable. Each tray can hold 20 sample cups. The sample cup could be of 0.25 ml, 0.5 ml or 2.00 mlsize. The 2.0 ml size is generally recommended.

Clinical Laboratory Instruments 415
The concentration of solutions is usually expressed in parts per million (ppm) in flame
photometry. This type of expression enables to make easy calculations on dilute solutions and theconcentrations can be expressed in weight/weight, weight/volume and volume/volume ratios.
/G20/G31/G34/G2E/G38 /G53/G45/G4C/G45/G43/G54/G49/G56/G45/G2D/G49/G4F/G4E/G20/G45/G4C/G45/G43/G54/G52/G4F/G44/G45/G53/G20/G42/G41/G53/G45/G44/G20/G45/G4C/G45/G43/G54/G52/G4F/G4C/G59/G54/G45/G53/G20/G41/G4E/G41/G4C/G59/G5A/G45/G52
Over the past decade, the pH meter has been at the centre of a most important change in the fieldof analytical measurements due to the introduction of selective-ion electrodes. As their name implies,these electrodes are sensitive to the activity of a particular ion in solution and quite insensitive tothe other ions present. As the electrode is sensitive to only one ion, a different electrode is needed foreach ion to be studied. Approximately 20 types of selective-ion electrodes are presently available.
Ion-selective electrodes are classified into four major groups:
Glass Electrodes: The first glass ion-selective electrode developed is the one sensitive to hydrogen
ions. Glasses containing less than 1% of Al
2O3 are sensitive to hydrogen ions (H+) but almost
insensitive to the other ions present. Glasses, of which the composition is Na2O–11%, Al2O3–18%,
Si O2–71% is highly selective towards sodium, even in the presence of other alkali metals. Glass
electrodes have been made that are selectively sensitive to sodium, potassium, ammonium andsilver.
Solid State Electrodes: These electrodes use single crystals of inorganic material doped with a rare
earth material. Such electrodes are particularly useful for fluoride, chloride, bromide and iodideion analysis.
Liquid-Liquid Membrane Electrodes: These electrodes are essentially liquid ion-exchangers,
separated from the liquid sample by means of a permeable membrane. This membrane allowsthe liquids to come in contact with each other, but prevents their mixing. Based on this principle,cells have been developed that are selective to calcium and magnesium. These cells are used formeasuring water hardeners.
Gas Sensing Electrodes: These electrodes respond to the partial pressure of the gases in the sample.
The most recent of these to be developed are the gas sensing electrodes for ammonia and sulphurdioxide. Ammonia or sulphur dioxide is transferred across a gas permeable membrane, until thepartial pressure in the thin film of filling solution between the glass electrode membrane and theprobe membrane equals that in the sample. The resultant pH change is measured by a combinationpH electrode. A potential is developed related to the partial pressure and hence the ammonia orsulphur dioxide concentration is measured.
Applications using ion-selective electrodes are many, most being time saving and simple to
use. The electrodes are now used for the continuous monitoring of important electrolytes in theblood such as sodium, potassium, calcium, chloride, etc.
/G31/G34/G2E/G38/G2E/G31 /G49/G6F/G6E/G20/G41/G6E/G61/G6C/G79/G7A/G65/G72/G73
Ion analyzers are basically pH/mV meters, which enable the operator to calculate the concentra-tion of specific ions from the potentials developed at the ion-sensitive electrode, when dipped insample solution. By measuring both the electrodes potential in a standard solution and in the

416 Handbook of Biomedical Instrumentation
sample solution, it is possible to calculate the unknown solution concentration by solving the
following equation:
Cx= Cs¥10DE/S
where, Cx =concentration of the unknown solution
Cs = concentration of the standard solution
E= difference between the observed potential in the sample solution and the observed
potential in the standard solution
S = electrode slope (change in electrode potential per ten-fold change in concentration)
Ion analyzers are mostly microprocessor-based instruments, which are programmed to calculate
sample concentration from a set of input data, such as electrode potentials, standard concentration,slope and blank correction. The instruments measure relative millivolts, pH and concentration ofspecific ions. The program for direct measurement concentration is based on the Nernstianelectrode response:
E
x = Eo + S log ( Cx + Cb)
where Ex = electrode potential
Eo = constant
Cb = blank concentration
The blank correction (Cb)accounts for the finite lower limit of detection of electrodes. If a solid or
liquid-membrane electrode is placed in pure water, the membrane dissolves slightly, producing
an equilibrium concentration of the measured ion. This concentration is a constant background
for all measurements and is represented by Cb. Typical electrode response curves are generally
given by the electrode manufacturers. If the sample concentration fall in the linear response region,a blank correction may not be necessary. But, if the sample concentrations are low, and fall in thenon-linear region of the response curve, blank correction must be applied.
The standard calibration procedure for a specific ion meter is similar to that used to calibrate a
pH meter with pH buffers. Two standard solutions are used, which are a decade apart in con-centration and approximately bracket the expected concentration range of the unknown samplesolution.
A block diagram of the ion analyzer is shown in Fig. 14.19. The first stage is the input buffer
amplifier, which provides a very high input impedance and less than 1 pA input bias current. Theelectrode potentials are individually buffered by unity gain amplifiers with FET front ends. Figure14.20 shows the input buffer stage. The two FETs are operated as source followers, each runningat a constant drain current determined by its associated op-amp. The voltage at the + input of eachop-amp is held constant, and therefore the drain current in the FETs will be constant. To do this,the op-amp output voltage must maintain a constant V
GSand must therefore follow the input
voltage. The op-amps effectively serve the dual purpose of controlling the operating current of the
FET and providing current gain. As with other similar circuits, the high input impedance of thebuffer amplifier gets degraded by the presence of dirt, moisture or solder flux. Also, the input FETis delicate and will get destroyed by static discharge. When the inputs are not being driven by asignal, they must be grounded with shorting straps. The input amplifier is followed by a differentialamplifier, before the signal is given to an A-D converter.

Clinical Laboratory Instruments 417
The results of the A-D converter are held in the A-D data latch by using shift-registers and the
loading function is controlled by the A-D converter. The output of the latch remains in a highimpedance state, until they are enabled by a signal from the control port decoder. Thus, the loadingand reading of data from the A-D are independent. The microprocessor may read data from the A-D converter, regardless of the timing of the analog-to-digital conversion cycle.
The microprocessor sends and receives information through the input-output (I/O) bus. The
bus is driven by only one source at a time and all other sources must be disabled, i.e. kept in a highimpedance state. The bus may be driven by the CPU, A-D converter, slope switches, standardvalue switches and mode switches. The CPU and display receive data from the bus.
The microcomputer consisting of a microprocessor, memory interface and read-only memory
(ROM) perform well-defined processing functions. Therefore, the program is stored in permanent
read-only memory. Under program control, the microprocessor generates signals on the control
port to select the path along which data will flow on the I/O bus. The CPU communicates with theAmplifier
multiplexerA-D
converterCPU Display
Slope and
standardFront panel
control
PreampInput
signalAnalog board
Fig.14.19 Block diagram of a microprocessor based ion-analyzer
+ +– – A2A1
VAoutVBoutVAinVBin
49.9 K 49.9 K
390 K150 K 49.9 K 49.9 K15 V
FET1FET2SSDDGVDG VDG
Fig.14.20 Input-buffer amplifier of an ion analyzer

418 Handbook of Biomedical Instrumentation
memory and the memory interface through the microprocessor data bus. Through this bus,
instructions and numerical constants flow from the memory outputs into the CPU. The memoryinterface performs the task of generating the address for each instruction stored in the memory. It
does this by maintaining a program counter according to commands from the CPU. The timing for
the microprocessor and for all signals on the bus is generated by the CPU clock.
Because of the low level of signal generated and high impedance of the ion-selective electrodes,
the grounding system is designed very carefully. Usually, the ion-analyzing instruments havethree grounds:
(i) The chasis and the electrostatic shield in the power transformer are connected to the earth
ground through the third wire of the AC line. This provides isolation from line noise
(ii)Digital ground provides the return path for all the logic signals, including the micropro-
cessor signals and the display current
(iii) Analog ground provides a reference point for the electrode input signals and a return path
for all analog current.
The analog and digital grounds are kept separate, so that digital signal return currents never
flow through the same conductor as the analog signal returns. The earth ground is not connectedto either digital or analog ground.
/G31/G34/G2E/G38/G2E/G32 /G43/G68/G65/G6D/G69/G63/G61/G6C/G6C/G79/G20/G53/G65/G6E/G73/G69/G74/G69/G76/G65/G20/G53/G65/G6D/G69/G63/G6F/G6E/G64/G75/G63/G74/G6F/G72/G20/G44/G65/G76/G69/G63/G65/G73
Considerable effort has recently been directed towards the development of ion-sensitive electrodesbased on a modification of the metal-oxide-semiconductor field transistor. In these devices,chemical sensitivity is obtained by fabricating the gate insulation of the FETs out of ion-sensitivematerials, usually a polymer or SiO
2. These devices are called ISFETs (ion-selective field-effect
transistors).
In these devices, the ion-sensitive material is bonded to the FET itself. This requires the material
and its method of fabrication to be compatible with the substrate (high purity silicon). This very
significant requirement puts a severe limitation on the use of some of the best characterized mem-brane materials, including ion sensitive glasses. The development of a pH-sensitive electrode bymeans of thick-film screening techniques has extensively been reported. This electrode retains theadvantages of ion-sensitive FET transducers, but eliminates the restrictions on membrane selectionand fabrication. Here, the ion-sensitive structure is physically separated from the FET. In this waythe ion-sensitive membrane can be fabricated on a compatible substrate and the FET can then beattached appropriately and placed in close proximity to the ion-sensitive membrane. A hybridelectrode structure permits the incorporation of a source follower FET amplifier, directly adjacentto the pH membrane, significantly reducing response time and noise pick-up (Janata, 1989).
Figure 14.21 presents the cross section of an ISFET, which is essentially a conventional insulated
gate field effect transistor, that has its metallic gate contact replaced by a chemically sensitivecoating and a reference electrode. In solution, the gate region can be coated with an ion-sensitivemembrane. Interaction of ions in solution with the membrane results in a change of the interfacial
potential and corresponding alteration of drain current. By this technique, numerous cations and
anions have been sensed (H
+, K+, Ca2+, Cl–, I– and CN–). The ISFET has advantages in its small size

Clinical Laboratory Instruments 419
(less than 1 mm2) and low output impedance, which makes it ideal for in vivo monitoring or
analysis of small sample volumes. However, problems like ion-selective coating adhesion anddevice encapsulaition have prevented large scale use of ISFETs.
ISFETs for up to eight different sensors have been fabricated on a single silicon chip. In addition,
probes 50 mm in diameter have been fabricated for on-chip circuitry that can measure pH, glucose,
oxygen saturation and pressure for biomedical applications (Webster, 1995). In addition to theISFET sensor, integrated circuits for signal processing can also be deposited on a single chip.4
123
(Substrate)(Source) (Drain)456789
(Reference
electrode)
(Selective
membrane)
(Anylate solution)
(Encapsulant)
(Insulator)
(Metalized
connections)
Fig.14.21 Constructional details of an ion-sensitive field-effect transistor

420 Handbook of Biomedical Instrumentation
/G42/G6C/G6F/G6F/G64/G20/G47/G61/G73/G20/G41/G6E/G61/G6C/G79/G7A/G65/G72/G73
Blood gas analyzers are used to measure the pH, partial pressure of carbon dioxide (pCO2) and
partial pressure of oxygen (pO2) of the body fluids with special reference to the human blood. The
measurements of these parameters are essential to determine the acid-base balance in the body. Asudden change in the pH and pCO
2 could result in cardiac arrhythmias, ventricular hypotension
and even death. This shows the importance of the maintenance of physiological neutrality inblood, and consequently the crucial role that the blood gas analyzers play in clinical medicine.
/G20/G31/G35/G2E/G31 /G41/G43/G49/G44/G2D/G42/G41/G53/G45/G20/G42/G41/G4C/G41/G4E/G43/G45
The normal pH of the extracellular fluid lies in the range of 7.35 to 7.45, indicating that the bodyfluid is slightly alkaline. When the pH exceeds 7.45, the body is considered to be in a state ofalkalosis. A body pH below 7.35 indicates acidosis. Both acidosis or alkalosis are disease condi-tions widely encountered in clinical medicine. Any tendency of the pH of blood to deviate towardsthese conditions is dealt with by the following three physiological mechanisms: (i) buffering bychemical means, (ii) respiration, (iii) excretion, into the urine by kidneys.
The blood and tissue fluids contain chemical buffers, which react with added acids and bases
and minimize the resultant change in hydrogen ions. They respond to changes in carbon dioxideconcentration in seconds. The respiratory system can adjust sudden changes in carbon dioxidetension back to normal levels in just a few minutes. Carbon dioxide can be removed by increasedbreathing and therefore, hydrogen concentration of the blood can be effectively modified. Thekidney requires many hours to readjust hydrogen ion concentration by excreting highly acidic oralkaline urine to enable body conditions to return towards normal.
Arterial blood has a pH of approximately 7.40. As venous blood acquires carbon dioxide, forms
carbonic acid and hydrogen ions, the venous blood pH falls to approximately 7.36. This pH drop
of 0.04 units occurs when the CO
2 enters the tissue capillaries. When CO2 diffuses from the pul-
monary capillaries into the alveoli, the blood pH rises 0.04 units to bring the normal arterial valueof 7.40. It is quite difficult to measure the pH of fluids inside the tissue cells, but from estimatesbased on CO
2and (HCO–
3) ion concentration, intracellular pH probably ranges from 7.0 to 7.2.HAPTER
1515

Blood Gas Analyzers 421
In order to maintain pO2, pCO2 and pH within normal limits, throughout the wide range of
body activity, the rate and depth of respiration vary automatically with changes in the metabolism.Control of alveolar ventilation takes place by means of chemical as well as nervous mechanisms.
The three important chemical factors regulating alveolar ventilation are the arterial concentrations
of CO
2, H+ and O2. Carbon dioxide tension in the blood stream and cerebrospinal fluid is the major
chemical factor regulating alveolar ventilation. The carotid and aortic chemoreceptors stimulaterespiration when oxygen tension is abnormally low. In fact, so many organs participate in thecontrol of respiration that it is difficult to include all aspects in this text. The readers may like toread any standard textbook on human physiology to appreciate the mechanism of respirationcontrol and the maintenance of physiological neutrality of the blood.
Table 15.1 Lists out the normal range for pH, pCO
2, pO2, total CO2, base excess and bicarbonate,
all measurements made at 37°C (Gambino, 1967).
/G95/G20/G54/G61/G62/G6C/G65/G20/G31/G35/G2E/G31 Typical Expected Values of Blood Gas Parameters
Arterial or arterialized Venous plasma
Parameter capillary blood (separated at 37 °C)
pH 7.37 to 7.44 7.35 to 7.45
pCO2 men 34 to 35 mmHg 36 to 50 mmHg
women 31 to 42 mmHg 34 to 50 mmHg
pO2 resting adult 80 to 90 mmHg 25 to 40 mmHgresting adult over 65 years 75 to 85 mmHg
Biocarbonate men 23 to 29 mmol/l 25 to 30 mmol/l
women 20 to 29 mrnol/l 23 to 28 mmol/1
Total CO
2 men 24 to 30 mmol/l 26 to 31 mmol/l
(plasma) women 21 to 30 mmol/l 24 to 29 mmol/l
Base Excess men – 2.4 to +2.3 mmol/l 0.0 to +5.0 mmol/l
women –3.3 to +1.2 mmol/l –1.0 to + 3.5 mmol/l
/G20/G31/G35/G2E/G32 /G42/G4C/G4F/G4F/G44/G20/G70/G48/G20/G4D/G45/G41/G53/G55/G52/G45/G4D/G45/G4E /G54
The acidity or alkalinity of a solution depends on its concentration of hydrogen ions. Increasing
the concentration of hydrogen ions makes a solution more acidic, decreasing the concentration ofhydrogen ions makes it more alkaline. The amount of hydrogen ions generally encountered insolutions of interest is extremely small and, therefore, the figure is usually represented in the moreconvenient system of pH notation. pH is thus a measure of hydrogen ion concentration, expressed
logarithmically. Specifically, it is the negative exponent (log) of the hydrogen ion concentration.
pH = –log (H
+)
If the number 10–7 represents the concentration of hydrogen ions in a certain solution, then its
pH would be 7. As hydrogen ion concentration rises, pH falls because the logarithm gets smaller,
and as hydrogen ion concentration falls, pH rises because the logarithm gets larger. As we deal in
logarithms to base 10, a pH of 7 represents 10 times the number of hydrogen ions as does a pH of 8.

422 Handbook of Biomedical Instrumentation
Since pure water dissociates into 10–7 mol/l of (H+) and 10–7 mol/l of (OH–), a pH of +7 is con-
sidered a neutral solution; a pH of +6 represents an acid, a pH of +8 an alkali. Since 10–6 is a larger
number than 10–8, the former solution has a larger hydrogenion concentration. Thus, a pH of 6 is
more acidic than a pH of 8. Concentration of ions, like the concentration of atoms or molecules, isexpressed in terms of mols/l (1 mol = 6.02 ¥ 10
–23 molecules, known as the Avogadro’s Number.
This is the number of molecules in one mole—an amount of material in grams equal to themolecular weight). Whole blood with a (H
+) of 4 ¥ 10–8 moles/1 would have a pH of 7.4; an
increase in the (H+) to 1 ¥ 10–7 moles/l would correspond to a decrease in pH to 7.0.
Electrochemical pH determination utilizes the difference in potential occurring between solutions
of different pH separated by a special glass membrane. If the pH of one of the solutions is kept con-stant, so that the potential varies in accordance with the pH of the other solution, then the systemcan be used to determine pH. The device used to effect this measurement is the glass electrode.
Glass Electrode: The potential ( E) of the glass electrode may be written by means of the Nernst
equation:
E = E
o–23 0 3 6. RT
F◊D pH
where Eo= standard potential
R= gas constant
T= absolute temperature
F= Faraday constant
DpH = pH value deviation from 7
The above relation shows that the emf developed in the electro-chemical pH cell is a linear
function of DpH.
Change of pH of one unit = 58.2 mV at 20°C = 62.2 mV at 40°C
The factor –2.3036 RTF is called the slope factor and is clearly dependent upon the solution
temperature. With a 1°C change in temperature, the emf changes by 0.2 mV. It is also obvious thatthe measurement of pH is essentially a measurement of millivolt signals by special methods.Figure 15.1 shows a relationship between the pH and emf at different temperatures.The reference
6 5 4 3 2 1 08 9 10 11 12 13 14
7
–379.3 mVpH+379.3 mV
–414.0 mV+414.0 mV
–448.8 mV+448.8 mV50°C
50°C25°C
25°C0°C
0°C+ mV
– mV
Fig.15.1 Relationship between pH and emf at different temperatures (Courtesy:
Beckman Instruments Inc., USA)

Blood Gas Analyzers 423
electrode provides a constant potential against which the potential of the indicator or glass
electrode is measured. An almost universally employed reference electrode is the saturated calomelelectrode.
pH Measurement For making pH measurements, the solution is taken in a beaker. A pair of
electrodes: one glass or indicating electrode and the other reference or calomel electrode, areimmersed in the solution. The voltage developed across the electrodes is applied to an electronicamplifier, which transmits the amplified signal to the display. The pH meter is usually equipped
with controls for calibration and temperature compensation.
The glass electrode exhibits a high electrical resistance, of the order of’ 100–1000 M W. The emf
measurement, therefore, necessitates the use of measuring circuits with high input impedance.
Further, the high resistance of glass electrodes render them highly susceptible to capacitive pick-up from ac mains. In order to minimize such effects, it is advisable to screen the electrode cable. Thescreen is usually grounded to the case of the measuring instrument.
The error caused in pH measurements due to temperature effect can be compensated either
manually or automatically. In manual adjustment the instrument is calibrated at 25°C. Then thecontrol is simply set to the actual measuring temperature. By this adjustment, the output current ofthe amplifier gets corrected to the desired temperature. In automatic adjustment, a variable resistorwhich is usually a thermistor or wire wound resistance that has an approximate desired resistancetemperature coefficient is inserted in the circuit. During measurement, it is placed in the testsolution. The use of an automatic temperature compensator will ensure that the pH meter isoperating with a correct mV/pH conversion ratio.
If it is desired to have the accuracy of a pH measurement as 0.001 pH, then the voltage must be
measured with an accuracy of 0.058 mV, assuming an ideal sensitivity of 58 mV per pH unit. Witha symmetrical scale on the measuring device of 6 pH units around pH 7, the maximum voltage willbe ±348 mV. Therefore, the accuracy requirement of the measurement will be 0.01%. This implies
that the internal resistance of the measuring device has to be 10
4 times the internal resistance of the
glass membrane (Bergveld and de Rooji, 1979).
Electrodes for Blood pH Measurement: Several types of electrodes have been described in literature
for the measurement of blood pH. They are all of the glass electrode type but made in different
shapes so that they may accept small quantities of blood and yield accurate results. The most
common type is the syringe electrode, which is preferred for the convenience of taking smallsamples of blood anaerobically. The small ‘dead space’ between the electrode bulb and the innersurface of the syringe barrel is usually filled with dilute heparin solution to prevent bloodcoagulation. Before making measurements, the syringe should be rolled between the hands toensure thorough mixing.
Microcapillary glass electrodes are preferred when it is required to monitor pH continuously:
for example during surgery. These types of electrodes are especially useful when a very smallvolume of the sample is to be analyzed.
Typically, a micro-electrode for clinical applications requires only 20–25 ml of capillary blood
for the determination of pH. The electrode is enclosed in a water jacket with circulating water at aconstant temperature of 38°C. The water contains 1% NACI for shielding against static inter-ference. The capillary is protected with a polyethylene tubing. The internal reference electrode is

424 Handbook of Biomedical Instrumentation
silver/silver chloride and the calomel reference electrode is connected to a small pool of saturated
KCI, through a porous pin. An accuracy of 0.001 pH can be obtained with this electrode against aconstant buffer. Figure 15.2 shows the constructional details of a typical blood pH electrode and
the measurement set-up used in practice.
Electrode
pH sensitive
glass
Capillary
tube
Sample
Sampling
catheter
BeakerKCE
sol.Ref sol.Ref
electrodeTo ref
inputTo pH
inputAspiration
Peristalticpump lever
Heating
resistorspH sensitive
glass
Heating
indicator
lamp
Thermometer
Sample
catheterCapillary
tubeAspiration
tube
Fig.15.2 Microcapillary electrode for measurement of blood pH (Courtesy:
Corning Scientific Instruments, USA)
Quite often, combination electrodes comprising both measuring and reference electrodes offer
single-probe convenience for all pH measurements. Several instruments offer the ability to measurepH in small containers with as little as 250 ml of the sample.
Ahn et al. (1975) bring out the drawbacks of the conventional macro-and micro-size pH
electrodes when used for biomedical applications. These are due to the relatively large size of themacro-electrode and the fragility of the micro-electrode. They constructed a miniature pH glasselectrode, using Corning 015 glass as the hydrogen-ion-sensitive glass. The dimensions of theelectrode were –1.0 mm outside diameter and 0.25 mm wall thickness. The inner electrolyte is asolution of 0.1 N hydrochloric acid and the inner reference electrode is a silver/silver chloride
electrode. The silver/silver chloride electrode is made from a silver wire (0.127 mm in diameter
and 99.9% purity) by electrolytic method. A FET input operational amplifier is integrated into thepH electrode. Temperature response of the electrode was –1.51 mV/K at a pH value of 7. Evaluationof the stability of the electrode showed a 1% drift over a 7–hour operational time. The pH temp-erature hysteresis effects showed a 0.5 and 1.0% deviation, respectively. The response time waswithin 4s for a 99% response.

Blood Gas Analyzers 425
Effect of Blood on Electrodes Glass electrodes deteriorate if allowed to remain in contact with
blood for a long time. This results in a change of the emf-pH slope. The poisoning effect appears tobe due to protein deposition. Therefore, as a precautionary measure, in an apparatus where blood
necessarily remains in contact around the electrode for long periods (more than 20 min.), the
response must be checked frequently against buffer solutions. The poisoning effect can be reducedby putting the electrode in pepsin and 0.1 N HCI, followed by careful wiping with a tissue paper.
The pH of blood is found to change linearly with temperature in the range of 18° to 38°C. The
temperature coefficient for the pH of blood is 0.0147 pH unit per degree centigrade. This neces-sitates the use of a highly accurate temperature controlled bath to keep the electrodes with theblood sample at 37°C ± 0.01°C. A circuit diagram for controlling temperature of the bath is shown
in Fig. 15.12.
Another important point to be kept in mind while making blood pH measurements is that
because of the possible individual variations in the temperature coefficient of blood pH, the methodof measuring at some temperature other than 37°C followed by correction is not recommended.It is advisable to keep both the glass as well as the reference electrode at the temperature ofmeasurement.
Buffer Solutions: Buffer solutions are primarily used for (i) creation and maintenance of a desired,
stabilized pH in a solution and (ii) standardization of electrode chains for pH measurements. Abuffer is, therefore, a substance which by its presence in a solution is capable of counteracting pHchanges in the solution as caused by the addition or the removal of hydrogen ions. Buffer solutionsare characterized by their pH value. They are available in tablets of pH value 4.7 and 9.2.Buffersolutions used in blood pH measurements are the following:
• 0.025 molar potassium dihydrogen phosphate with 0.025 molar disodium hydrogen phos-
phate. This solution has a pH value of 6.840 at 38°C and 6.881 at 20°C.
• 0.01 molar potassium dihydrogen phosphate with 0.04 molar disodium hydrogen phos-
phate. This buffer has a pH value of 7.416 at 38°C and 7.429 at 20°C.
These buffers should be stored at a temperature between 18° to 25°C. To maintain an accurate
pH, the bottles containing them should be tightly closed.
/G20/G31/G35/G2E/G33 /G4D/G45/G41/G53/G55/G52/G45/G4D/G45/G4E /G54/G20/G4F/G46/G20/G42/G4C/G4F/G4F/G44/G20/G50/G43/G4F/G32
The blood pCO2 is the partial pressure of carbon dioxide of blood taken anaerobically. It is
expressed in mmHg and is related to the percentage CO2 as follows:
pCO2 = Barometric pressure – water vapour pressure ¥%C O2
100
At 37°C, the water vapour pressure is 47 mmHg, so at 750 mm barometric pressure, 5.7% CO2
corresponds to a pCO2 of 40 mm.
All modern blood gas analyzers make use of a pCO2 electrode of the type described by Stow et al
(1957). It basically consists of a pH sensitive glass electrode having a rubber membrane stretchedover it, with a thin layer of water separating the membrane from the electrode surface. Thetechnique is based on the fact that the dissolved CO
2 changes the pH of an aqueous solution. The

426 Handbook of Biomedical Instrumentation
CO2 from the blood sample defuses through the membrane to form H2CO3, which dissociates into
(H+) and (HCO–
3) ions. The resultant change in pH is thus a function of the C02 concentration in
the sample. The emf generated was found to give a linear relationship between the pH and the
negative logarithm of pCO2. Although the electrode could not provide sensitivity and stability
required for clinical applications, it made way for realizing a direct method for the measurementof pCO
2.
The basic construction of the electrode was modified by Severinghaus and Bradley (1958) to a
degree that made it suitable for routine laboratory use. In the construction worked out by them, thewater layer was replaced by a thin film of an aqueous sodium bicarbonate (NaHCO
3) solution.
The rubber membrane was also replaced by a thin Teflon membrane, which is permeable to CO2
but not to any other ions, which might alter the pH of thebicarbonate solution. The CO
2from the blood diffuses into
the bicarbonate solution. There will be a drop in pH due toCO
2 reacting with water forming carbonic acid. The pH
falls by almost one pH unit for a ten-fold increase in theCO
2 tension of the sample. Hence, the pH change is a linear
function of the logarithm of the CO2 tension. The optimum
sensitivity in terms of pH change for a given change in CO2
tension is obtained by using a bicarbonate solution of
concentration of about 0.01 mole/l. The electrode iscalibrated with the known concentration of CO
2. The
response time of the CO2 electrode is of the order of 0.5 to
3 min. This electrode was twice as sensitive and driftedmuch less than the Stow’s electrode. Figure 15.3 shows theconstruction of a typical pCO
2 electrode.
Further improvements in stability and response time
were achieved by Hertz and Siesjo (1959). They used adilute solution of NaHCO
3 (0.0001 N), which helped in
reducing the response time but the drift introduced posedserious problems. The compromise between response timeand drift was achieved by using a 0.001 N solution of
NaHCO
3. Silver/silver chloride reference electrode was
replaced by a calomel cell which was made an integral partof the electrode.
Severinghaus (1962) made a further improvement upon
the earlier Severinghaus-Bradley electrode in the low pCO
2
range by replacing the cellophane spacer with a very thinnylon mesh. Glass fibres or powdered glass wool were alsofound to be good separators. He used a membrane of 3/8mil Teflon and glass wool for the separator. Electrodesconstructed in this way had a 95% response in 20 s.
Reyes and Neville (1967) constructed a pCO
2 electrode
using 0.5 mm polyethylene as a membrane and used noMembraneNylon
spacerpH sensitive
glassElectrolyte
reservoirReference
reservoirFill holeFill hole
capLead
Fig.15.3 Construction of pCO2
electrode (Courtesy:
Corning ScientificInstruments, USA)

Blood Gas Analyzers 427
separator between the glass surface and this membrane. They added carbonic anhydrase to the
electrolyte. The response time was found to be 6 seconds for 90% of a step change from 2% to 5%CO
2. Use of a pCO2 electrode for the measurment of blood or plasma pCO2 has been studied
repeatedly and has been found to be accurate, precise and expedient, (Hill and Tilsley, 1973). An
extension of the miniature pH electrode (Ahn et al, 1975) is the miniature pCO2 electrode described
by Lai et al (1975).
/G31/G35/G2E/G33/G2E/G31 /G50/G65/G72/G66/G6F/G72/G6D/G61/G6E/G63/G65/G20/G52/G65/G71/G75/G69/G72/G65/G6D/G65/G6E/G74/G73/G20/G6F/G66/G20/G70/G48/G20/G4D/G65/G74/G65/G72/G73/G20/G55/G73/G65/G64/G20/G66/G6F/G72/G20/G70/G43/G4F/G32/G20/G4D/G65/G61/G73/G75/G72/G65/G6D/G65/G6E/G74
The emf generated by a pCO2 electrode is a direct logarithmic function of pCO2. It is observed that
a ten-fold change in pCO2 causes the potential to change by 58 ± 2 mV. The pH versus log pCO2
relationship is linear within ±0.002 pH unit from 1 to 100% carbon dioxide. Since 0.01 unit pH
change corresponds to a 2.5% change in pCO2 or 1 mmHg in 40 mmHg, for achieving an accuracy
of 0.1 mmHg, it is desirable to read 0.001 pH unit, i.e. a resolution of 60 mv. This order of accuracy
can be read only on a digital readout type pH meter or on an analog meter with expanded scale.The instrument should have a very high degree of stability and a very low drift amplifier. Theinput impedance of the electronic circuit must be atleast 10
12W.
It is essential to maintain the temperature of the electrode assembly constant within close limits.
It is experimentally shown that variation in the temperature of ± 1°C produces an error of ±1.5mmHg or about ±3% at 5 mm pCO
2. The combined effects of temperature change upon the
sensitivity of the pH electrode and upon the pCO2 of the blood sample amount to a total variation
in sensitivity of 8% per degree centigrade.
Calculated Bicarbonate, Total CO2and Base Excess: Acid-base balance determinations are based
on several calculations, which are routinely used in conjunction with blood pH and gas analysis.An accurate picture of acid-base balance can be determined from the equilibrium.
CO
2 + H2OÆ H2CO3
H2CO3Æ H+ + HCO3–
which for bicarbonate has an equilibrium constant
KH2 CO3/HCO3–=[H ][HCO ]
HC O3
23+-
where (H+), (HCO–
3) and (H2CO3) refer to the concentration of these substances.
Since H2CO3= 0.03 pCO2
and since pH = –log [H+]
Therefore, pH = pK + log[HCO ]
0.03 pCO3
2-
where pK equals 6.11 for normal plasma at 37°C. This formula is used in blood gas analysers for
calculating actual bicarbonate.
Total CO2 is calculated from the relationship:
[HCO3–] + ( 0.03 ¥ pCO2) = total CO2 in millimoles/l

428 Handbook of Biomedical Instrumentation
Base excess is calculated from the formula described by Siggaard-Andersen (1963).
Base excess = (1 – 0.0143 ¥ Hb) [HCO3–] – (9.5 + 1.63 Hb) ¥ (7.4 – pH) – 24
where Hb represents the patients’ haemoglobin value.
Base excess is the number of milliequivalents of a strong acid or base which would be required
per litre of blood to restore it to a pH of 7.400 at 37°C with pCO2 held at 40 torr. This is usually
estimated from pH and pCO2 measurements done at 37°C in a sample of blood using Siggaard-
Andersen’s alignment monogram (Siggaard-Andersen, 1963).
/G20/G31/G35/G2E/G34 /G42/G4C/G4F/G4F/G44/G20/G70/G4F/G32/G20/G4D/G45/G41/G53/G55/G52/G45/G4D/G45/G4E /G54
The partial pressure of oxygen in the blood or plasma indicates the extent of oxygen exchange
between the lungs and the blood, and normally, the ability of the blood to adequately perfuse thebody tissues with oxygen. The partial pressure of oxygen is usually measured with a polarographicelectrode. There is a characteristic polarizing voltage at which any element in solution ispredominantly reduced and in the case of oxygen, it is 0.6 to 0.9 V. In this voltage range, it isobserved that the current flowing in the electrochemical cell is proportional to the oxygenconcentration in the solution.
Most of the modern blood gas analyzers utilize an
oxygen electrode first described by Clark (1956) formeasuring oxygen partial pressure. This type of elect-rode consists of a platinum cathode, a silver/silverchloride anode in an electrolyte filling solution and apolypropylene membrane. The electrode is of a singleunit construction and contains the reference electrode
also in its assembly. Figure 15.4 shows the construction
of a typical Clark-type oxygen electrode. The entire unitis separated from the solution under measurement bythe polypropylene membrane.
Oxygen from the blood diffuses across the membrane
into the electrolyte filling solution and is reduced at thecathode. The circuit is completed at the anode, wheresilver is oxidized, and the magnitude of the resultingcurrent indicates the partial pressure of oxygen. Thereactions occurring at the anode and cathode are:
Cathode reaction:
O
2 + 2H2O + 4e–Æ 4 OH–
Anode reaction:
4Ag Æ 4Ag++ 4e–
The Clark electrode for measuring pO2 has been
extensively studied and utilized. It is found to be ofparticular advantage for measuring blood samples. TheReservoir
fill hole
Ag-agcl
anodeElectrolyte
reservoir
MembranePlantium
cathodeLead
Fig.15.4 Construction of pO2
electrode (Courtesy:
Corning ScientificInstruments, USA)

Blood Gas Analyzers 429
principal advantages are: (i) sample size required for the measurement can be extremely small,
(ii) the current produced due to pO2 at the electrode is linearly related to the partial pressure of
oxygen, (iii) the electrode can be made small enough to measure oxygen concentration in highly
localized areas, (iv) the response time is very low, so the measurements can be made in seconds. As
compared to this, it takes a very long time if the measurements are made by chemical means.
McConn and Robinson (1963) observed that zero
electrode current was not given by a solution having zerooxygen tension, but occurred at a definite oxygen tension,which they called the ‘electrode constant’. So, for calib-rating the electrode it was necessary to know this con-stant for that particular electrode. They further showedthat when the straight line calibration curves (Fig. 15.5)were extended backwards, they did not pass throughthe origin, but intersected the oxygen tension axis at anegative value. To obtain a true zero-current (less than10 nA), the electrolyte of the electrode is deoxygenated bybubbling nitrogen through it for about half an hour and
then placing the electrode in water redistilled from
alkaline pyragallol.
The platinum cathode of the oxygen electrode tends to
become contaminated or dimensionally unstable withtime and use. The result is usually an inability to calib-rate and slope the electrode on any pO
2 range. The manufacturers usually recommend application
of ammonium hydroxide on the tip of the electrode (10% solution), with a gentle, rotary motionusing a swab. The silver chloride gets dissolved in ammonium hydroxide. It is then flushed withdistilled water.
The polarographic electrodes usually exhibit ageing effect by showing a slow reduction in
current over a period of time, even though the oxygen tension in the test solution is maintained ata constant level. Therefore, it needs frequent calibration. This is probably associated with thematerial depositing itself on to the electrode surface. The effect due to ageing can possibly beavoided by covering the electrode with a protective film of polyethylene, but it has the undesirableeffect of increasing the response time.
The measurement of current developed at the pO
2 electrode due to the partial pressure of oxygen
presents special problems. The difficulty arises because of the extremely small size of the electricalsignal. The sensitivity (current per torr of oxygen tension) is typically of the order of 20 pA per torr
for most commercial instruments. It is further subject to a constant drift and is also not independent
of the sample characteristics. Measurement of oxygen electrode current is made by using highinput impedance, low noise and low current amplifiers. Field effect transistors usually form theinput stage of the preamplifiers.
Hahn (1969) used a field-effect transistor operational amplifier to measure small polarographic
currents. The op-amp. is connected as a transresistance converter, the output of which can be readdirectly by a digital voltmeter. Figure 15.6 shows the circuit. The polarizing voltage is suppliedby the cell B (1.3 V) and variable resistance VR
1. The standing current from the electrochemical cell0.1 0.2 0.3100
0
–40200300
Current (micro amp.)Oxygen tension (mmHg)
Fig.15.5 Calibration curve of pO2
electrode (after McConn
and Robinson, 1963)

430 Handbook of Biomedical Instrumentation
is cancelled by means of VR2, Battery B and 1 G W resistance. Capacitor C (100 pF) is included to
limit the bandwidth of the amplifier to reduce noise and to ensure good dynamic stability.
+–
VR3
Output to
digital voltmeter300 M100 pF
Pt. Ag-AgCl1 G 2.5 K
2.5 KB2
B1VR2
VR1
1.3 V
Fig.15.6 Current amplifier for use with pO2 electrode
/G20/G31/G35/G2E/G35 /G49/G4E /G54/G52/G41/G2D/G41/G52 /G54/G45/G52/G49/G41/G4C/G20/G42/G4C/G4F/G4F/G44/G20/G47/G41/G53/G20/G4D/G4F/G4E/G49 /G54/G4F/G52/G49/G4E/G47
Arterial blood gas analysis is beneficial in the assessment and management of patients requiring
mechanical ventilation and for those suffering from cardiopulmonary and other difficulties.Arterial blood gas values provide vital information about the adequacy of oxygenation, ventilation,acid-base balance and gas exchange in the lungs. In vitro blood gas analyzers, though commonly
used, have several limitations. They require that blood be drawn and the sample analyzed, often
at a distant blood gas laboratory. In addition to problems associated with blood handling, the
need to send the sample to a laboratory delays results and treatment. Blood gas values can fluctuaterapidly in critically ill patients, and therefore, patient care decisions based on delayed informationmay be inappropriate. In vivo methods of blood gas monitoring have been developed to overcome
these drawbacks.
/G31/G35/G2E/G35/G2E/G31 /G43/G61/G74/G68/G65/G74/G65/G72/G20 /G54/G69/G70/G20/G45/G6C/G65/G63/G74/G72/G6F/G64/G65/G20/G66/G6F/G72/G20/G4D/G65/G61/G73/G75/G72/G65/G6D/G65/G6E/G74/G20/G4F/G66/G20/G70/G4F/G32/G20/G61/G6E/G64/G20/G70/G43/G4F/G32
Miniature electrodes are required for in vivo transcutaneous measurements of pO2 and pCO2. The
electrodes must be small enough to be mounted on the catheter tip and should preferably perform
measurements of more than one parameter. One such electrode capable of simultaneous measure-ment of both pO
2 and pCO2 is described by Parker et al (1978). The device is built into the tip of a SF
(1.65 mm) catheter, 40 cm in length. The electrode (Fig. 15.7) comprises a pH sensitive glass bulb atthe tip of the catheter for measuring changes in pH and hence pCO
2 according to the method
described by Severinghaus and Bradley (1958).
A 180 mm diameter silver cathode constitutes a pO2 measuring electrode. The common electrode
used is of silver/silver chloride. A semi-solid electrolyte is common for both the pO2 and pCO2
electrodes. The electrodes are dip-coated with a thin polystyrene diffusion membrane. When thedevice is placed in blood, water vapour diffuses through the membrane and together with theNaHCO
3 and NACI crystals deposited in the hydrogel film constitutes the electrolyte normally
used with a pCO2 electrode. Under these conditions, the output signal from both the pO2 and pCO2

Blood Gas Analyzers 431
electrodes is obtained. The response time for 90% response is found to be 2 min. However, the time
for 100% stabilization of the outputs after a step change in pCO2 and pO2 in solution is pretty long.
/G31/G35/G2E/G35/G2E/G32 /G70/G4F/G32/G20/G4D/G65/G61/G73/G75/G72/G65/G6D/G65/G6E/G74/G20/G77/G69/G74/G68/G20/G43/G75/G74/G61/G6E/G65/G6F/G75/G73/G20/G45/G6C/G65/G63/G74/G72/G6F/G64/G65/G73
Continuous intra-arterial monitoring of oxygen is unsatisfactory as a clinically reliable procedure.It is expensive and relatively traumatic. Oximetric methods of monitoring oxygen tension fromoxygen saturation are unreliable at the pO
2 level above 50 mmHg (Scacci et al, 1976). Relatively
simple and non-invasive methods are required to continuously monitor pO2 to detect changes or
establish trends. Eberhard et al (1973) concluded, after measuring oxygen tension using a Clark-
type electrode applied directly to the skin, that an excellent correlation (0.98) exists between skinpO
2 and arterial pO2 in infants and new borns (Fig. 15.8). This could provide more immediate
detection of hypoxia or hyperoxia than arterial sampling.
The principle underlying the skin sensor is that since oxygen is able to diffuse through body
tissue and skin, the measurement of pO2 can be obtained indirectly by applying a Clark-type
electrode sensor to the skin, heated to a constant temperature higher than the skin (44°C). Activevasodilation of the cutaneous vessels is achieved by warming the cathode and anode of the oxygensensor to a temperature, which is higher than the normal body surface temperature. Oxygen1.6 mmLead out
wiresCatheterSilver
cathodeSilver/silver chloride
internal referenceGelled
electrolyte
Diffusion
membrane
pH glass
Semi-solid
electrolyte
Silver/silver
chloride common
reference
electrode
Silicon rubber
seal
Fig.15.7 Catheter tip electrode for measurement of pO2 and pCO2 (after Parker et
al. 1978; reproduced by permission of Med. & Biol. Eng. and Comput.)

432 Handbook of Biomedical Instrumentation
diffuses from the arterialized capillary bed through the epidermis to the skin surface and is
measured there by an electrochemical reduction at the cathode of a Clark-type sensor. The electrodeis 14 mm in diameter, with a 4 mm gold cathode, silver/silver chloride anode, covered with a 6 m
thick Mylar membrane (Fig. 15.9). The electrolyte used is a solution of KCI buffered to pH 10 which
has a 3 to 45 days life if kept moist. A coil of resistance wire embedded in the Sensor heats the
electrode to 44°C, a temperature selected to provide good hyperemisation as well as safety forcontinuous application over several days. A thermistor is also imbedded in the sensor to providea control signal for monitoring constant temperature. The response time of the electrode is 60 s for95% response to a full-scale step change. The sensor is applied to the skin using adhesive tape.The polarogram of this sensor in the gas phase shows a stable plateau extending from –700 mV to100 50 0 100 150 300 250 350100
50
0150200250300350yCutaneously
measured
pO (mm Hg)2
Arterial blood p O (mm Hg)a2
Sensor temperature: 44°C
Sensor position: thorax
Number of measurements: 490
from 6 hospitals
Regression line: pO = 8 + 0.95 p O2a 2
Correlation coefficient: 0.93x
Fig.15.8 Relationship between central arterial pO2 and cutaneous pO2 of new-
borns. A significant correlation (0.93) was found to exist between the twovalues (after Eberhard et al, 1976)

Blood Gas Analyzers 433
–1.1 V with an operating voltage of –900 mV. The zero-current in 100% nitrogen is less than 0.5
nA/mm2. Eberhard et al (1975) explain the details of the electronic circuit used with cutaneous
pO2 sensors.
The cutaneous oxygen electrode is a satisfactory indicator of the changes in arterial oxygen
tension in only those patients who have a good peripheral circulation. Skin pO2is probably more
dependent on perfusion changes than on arterial tension. Consequently, this measurement may
be more sensitive to physiologic changes than is the arterial oxygen tension.
In spite of the fact that a significant correlation between cutaneous pO2 and arterial pO2 exists
if the sensor is heated to a temperature of 44°C, the cutaneous measurement technique cannot beregarded as a substitute for conventional blood-gas analysis via sampling of arterial blood(Eberhard and Mindt, 1976). In some situations, e.g. during marked peripheral vasoconstriction,the cutaneous pO
2 does not correlate with arterial blood pO2, and hyper-oxemic or hypoxemic
situations may not be reliably detected. Cutaneous pO2 measurement is a supplemental technique,
which helps to continuously monitor clinically significant changes in the oxygenation state of thesubject, which may otherwise go unnoticed in the time interval between blood sample analysis.
/G20/G31/G35/G2E/G36 /G41/G20/G43/G4F/G4D/G50/G4C/G45 /G54/G45/G20/G42/G4C/G4F/G4F/G44/G20/G47/G41/G53/G20/G41/G4E/G41/G4C/G59/G5A/G45/G52
Blood gas analyzers are designed to measure pH, pCO2 and pO2 from a single sample of whole
blood. The size of the sample may vary from 25 μl to a few hundred microlitres. The estimations
take about 1 minute. With built-in calculators, the instruments can also compute total CO2, HCO3
and Base Excess. A typical block diagram of a blood gas analyzer machine is shown in Figure
15.10. In this machine, separate sensors are used for pH, pCO2 and pO2. The instrument contains
three separate high input impedance amplifiers designed to operate in the specific range of eachmeasuring electrode. A separate module houses and thermostatically controls the three electrodes.It also provides thermostatic control for the humidification of the calibrating gases. A vacuum12 10 9 11
23 4 5 6 1 2 7 81
(1) Skin (5) Electrolyte (9) Heating element
(2) Adhesive ring (6) Cathode (10) Thermistors
(3) Contact fluid (7) Anode (11) Encapsulation
(4) Membrane (8) Membrane ratainer
ring(12) Encapsulation
Fig.15.9 Cross-section of cutaneous oxygen sensor (Courtesy: Roche Medical
Electronics Inc., USA)

434 Handbook of Biomedical Instrumentation
Fig.15.10 Block diagram of a complete blood gas analyser (Courtesy: Corning Scientific Instruments, USA)
CalCalCal
pHamp.InputInput
Rec.SlopeSlopeSlope
pCOamp.2pOamp.2Rec.
Range Sw
CO zero pot2
Adjust/operateswitch
Antilogconvert.Rec.
TCO2
HCO3Baseexcess Hb setA-D boardRead-out board
MeasureCKT. Timer

Blood Gas Analyzers 435
system provides aspiration and flushing service for all three electrodes. Calibrating gases are
selected by a special push button control and passed through the sample chamber when required.Two gases of accurately known O
2 and CO2 percentages are required for calibrating the analyzer
in the pO2 and pCO2 modes. The gases required are: O2 value of 12% Cal and 0% Slope and CO2
value of 5% Cal and 10% Slope. These gases are used with precision regulators for flow and press-
ure control. Two standard buffers of known pH are required for calibration of the analyzer in thepH mode. The buffers that are used are 6.838 (Cal) and 7.382 (Slope). It is generally recommendedthat the sample chamber should control 7.382 buffer when in the standby mode.
Input signal to the (HCO
3–) calculator (Fig. 15.11a) comes from the outputs of the pH and pCO2
amplifiers. The outputs are suitably adjusted by multiplying each signal by a constant and are
given to an adder. The next stage is an antilog-generator similar to the one used in a pCO2 amplifier.
The output of this circuit goes to an A–D converter for display. Resistance R is used to adjust zero
at the output.
Total CO2 is calculated (Fig. 15.11b) by summing the output signals of the (HCO3–) calculator
and the output of the pCO2 amplifier. Facilities for adjusting the slope and zero at the output are
available.
pH slope
adjust
From pH
amp.
From pCO
amp.2From pCO
amp.2
From HCO
output3TCO2
TCO2SlopepCO
slope adjust2 RAntilog
gen.
Antilog
gen.HCO3
To A-D
converter
(a)
(b)
(c)Base excess
+VHaemoglobin (gain) setpot
From
pH amp.
From
HCO output3
RR
Fig.15.11 Circuit diagrams for computation of
(a) Bicarbonate (HCO3–)(b) Total CO2(c) Base excess

436 Handbook of Biomedical Instrumentation
The base excess calculator (Fig. 15.11c ) consists of three stages. In the first stage, the output of
the pH amplifier is inverted in an operational amplifier whose gain is controlled with a potentio-meter (Haemoglobin value) placed on the front panel. The output of the HCO
3– calculator is
inverted in the second stage. The third stage is a summing amplifier A3 whose output is given to an
A–D converter.
The analog output of the selected parameter channel is given to the input of an A–D converter.
The output of the A–D converter goes to a digital readout circuit like LEDs.
The three electrodes (pH, pO2 and pCO2) are housed in a thermostatically controlled chamber.
It also provides thermostatic control for the humidification of the calibrating gases. The thermalblock and the humidifier block heat control circuits are of the same type (Fig. 15.12). Thetemperature is set with a potentiometer for exactly 37°C. The heater circuit is controlled by athermistor in the block, which acts as a sensor. As the heat increases, the resistance of the thermistordecreases. At 37°C, the thermistor is calibrated to have a resistance of 25 K.
Current supply for thermistor
Thermistor
Noise filter
0.01
Fm0.1 Fm75 K
7.5 KTV
3.16 K
2.61 KSet point
resistors
Feedback
resistors1 MEG
Voltage
feedback
divider10 K100 K T2T1Heater
resistorsOutput
transistors5.11 K
5.11 KCurrent
limiter
Amp.+30 V
Fig.15.12 Temperature control circuit for thermostated chamber (Courtesy:
Corning Scientific Instruments, USA)
Supposing the temperature of the block decreases, the resistance of the thermistor will increase.
The increase in resistance will cause the voltage at inverting input of op-amp to become morenegative. This results in the output voltage becoming more positive, increasing the base current oftransistors T
1 and T2.The increase in base current increases the collector current, which goes
directly to the heater resistor on the block. As the heater resistor heats up the block, the thermistorwill decrease until it returns to 25 K.
Many of the blood gas analyzers have a provision for checking the membrane of pO
2and pCO2
electrodes. In the check position, a potential is applied across the membrane. Any leak in the
membrane of sufficient magnitude will result in a considerable lowering of the resistance may be

Blood Gas Analyzers 437
from 100 M W to 500 K W. The change in resistance can be used to have a change of potential to
switch on a transistor, which would cause a lamp to light on the front panel of the instrument.This would indicate that a new membrane is needed.
/G31/G35/G2E/G36/G2E/G31 /G46/G69/G62/G65/G72/G2D/G6F/G70/G74/G69/G63/G20/G42/G61/G73/G65/G64/G20/G42/G6C/G6F/G6F/G64/G20/G47/G61/G73/G20/G53/G65/G6E/G73/G6F/G72/G73
Forin vivo measurements and for reliably analyzing blood gases, a small, stable, accurate and bio-
compatible sensor is required which could be inserted in the blood flow of an artery (Miller, 1993)through an arterial cannula and remain in place for several days. In addition, it has to be low costso that it could be used as a disposable item. Advances in fiber-optics and the development of pHand oxygen sensitive dyes have made such a sensor possible. Blood gas analyzers based on suchsensors are now commercially available.
Figure 15.13 shows the schematic diagram of a fiber-optic based blood gas analyzer (Soller,
1994). The sensors are interfaced with an electro-optic monitor. The monitor supplies the excitation
light, which may be from a monochromatic source such as a diode laser or a broadband source
MicroprocessorDisplay
SourceDetector
Filter Filter
Beam
splitter
Lens
Optical
connectorFibre optic
cable
Catheter
Selective
biocompatible
membrane
Sensing
chemistryProtective tubingThermocouple Fiber
Fig.15.13 Block diagram of fiber optic based blood gas sensor and monitor
(after Soller, 1994)

438 Handbook of Biomedical Instrumentation
like a xenon lamp whose light is filtered to provide a narrow bandwidth of excitation. Two
wavelengths of light are provided, one wavelength is sensitive to changes in the species to be
measured, while the other wavelength is unaffected by changes in the analyte concentration. Theunaffected wavelength serves as a reference and is used to compensate for fluctuations in the
source output and detector efficiency. The light output from the monitor is coupled into a fiberoptic
cable through appropriate lenses and optical connectors. The cable is sufficiently long to permiteasy patient access by allowing the monitor to be placed at a distance.
Within the sensor assembly (Fig. 15.14) are three optical fibers—one each for measuring blood
O
2, CO2 and pH. The optical-fiber is approximately 10 cm long and also has a thermocouple or
thermistor wire running along side the fiber to measure temperature near the sensor tip. Temp-erature correction is necessary for optical blood gas sensors. The solubility of the gases namely O
2
and CO2 in the sensing material, is a function of temperature and the optical properties of the
sensing chemistry also change as the temperature varies. The fibers and the temperature sensorare encased in a protective tubing to contain any fiber fragments in case of sensor breakage.
Thermocouple
CO fibre2O fibre2
pH fibre
Fig.15.14 The sensor assembly contains three fibres that measure pH, pCO2 and
pO2, bundled together with the thermocouple
Each fiber is as thin as human hair and coated at the tip with a specific chemical dye (Fig. 15.15).
When light of a known wavelength strikes the dye, the dye fluoresces, giving off light of a differentwavelength. The fluorescent emission changes in intensity as a function of the concentration ofthe analyte (O
2, CO2 or pH) in the blood. The emitted light travels back down the fiber to the monitor
where it is converted into an electrical signal by using a solid state detector or a photomultiplier.The signal is amplified before it is given to a digitizer. Signal processing to relate the light intensityto the analyte concentration is achieved using a microprocessor and is digitally displayed.
Because a detector produces noise or dark current when it is not illuminated, accurate signal
measurements require that ambient light be subtracted from the total signal. Thus, a signalmeasurement is made with the flash lamp off. This ambient light is subtracted from the total signalby correlated double integration circuits. Another factor that affects accurate blood gas calculationsis the background fluorescence of the materials in the optical block. This value is obtained bymeasuring the current developed by the detector when no sensor is connected. As with ambientlight, the background fluorescence is subtracted from the signal measurement.

Blood Gas Analyzers 439
Considerable effort has gone into identifying organic molecules, which would make suitable
sensors. These molecules must have a high fluorescent intensity at excitation and emissionwavelengths that match the available light sources and detectors. They must be photostable, i.e.their emission properties should not change as they are continually illuminated by the excitationsource. Sensors based on fluorescence quenching of organic dyes such as perylene dibutyratehave been reported for the measurement of pO
2. Oxygen sensors based on the phosphorescence
quenching of metal—loporphyrins and terbium complexes have also been successfully tried.
It has been found from experimental studies that as the partial pressure of oxygen increases, the
sensitivity decreases. The best sensitivity is achieved in the region of 30 to 150 mmHg, but dropsoff considerably with higher pO
2, making it difficult to resolve small changes in pO2, when the
oxygen partial pressure is greater than 200 mmHg. Further, at high pO2 as the quenching increases,
the light reaching the detector decreases. A compromise is thus required to be made in selecting a
sensing material that provides adequate sensitivity over the required pO2 measurement range and
simultaneously offers good signal-to-noise ratio at the detector. The performance range of thesensor is normally limited to under 300 mmHg in order to achieve good sensitivity and adequatelight detection.
pH sensor: Designs are based on dye molecules whose optical properties change as the pH is
varied between 6.8 and 7.8. At any pH in the range of interest, both the acid and base forms of thedye molecule are present and each form has distinct optical characteristics. pH sensors have beendeveloped which take advantage of the fact that the excitation wavelength for fluorescenceemission of some dyes is different for the acid and base forms and the ratio of emission excited atthese two wavelengths can be used to calculate pH. Additionally, sensors have been developedwhich utilize the difference in absorption maxima for both the acid and base forms of the dye.Excitation light signal
DyeReturn (re-emitted)
light signal
Fig.15.15 Within each fibre’s core, excitation light reflects along the fibre toward
the fluorescent dye at the fibre's tip. The dye at the tip reacts to theexcitation light and analyte concentration by fluorescing. The fluorescentsignal then returns in the same fibre to the monitor, which measures theintensity of the signal

440 Handbook of Biomedical Instrumentation
The commonly used pH-sensitive dye is phenol red whose absorption spectra is shown in
Fig. 15.16. The largest peak is observed from base form of phenol red at 560 nm and is used tomeasure pH because it is more sensitive to pH changes than the acid peak at 430 nm. A wavelength,
which is insensitive to pH changes, is used as a reference; either a wavelength greater than 600 nm
or the isobestic point at 480 nm. The relationship between pH and the base form of the dye is givenby the Henderson-Hasselbalch equation:
pH = pKa– log
[HA]
[A ]-
where pH is the negative logarithm of hydrogen and concentration and pKa is the negative
logarithm of the equilibrium constant Ka, which describes the dissociation of the acid, HA.
450 400 500 550 600 6506.07.08.0pH = 9.0
Wavelength (nm)Absorbance
Fig.15.16 Absorption spectra of pH sensitive dye (Phenol red)
One of the difficulties in designing a pH sensor is to achieve resolution of 0.01 pH units
over the range of 6.8 to 7.8. An effective way to achieve this is to optimize the pKa of the dye
material. This can be done through proper choice of a functional group attached to the dyemolecule or by immobilizing the pH-sensitive material on a polymer with the appropriate ioniccharacteristics.
Most fiberoptic sensors for measuring pCO
2 use the same approach as a pCO2 electrode. A
pCO2 sensor is fabricated by surrounding a pH sensor with a gas permeable membrane containing
a bicarbonate ion (HCO–
3 ) buffer. The membrane allows gaseous CO2 and water vapour to enter
the sensor, and they combine to form carbonic acid as per the following equations:
CO2 + H2OÆ H2CO3
H2CO3Æ H+ + HCO–
3

Blood Gas Analyzers 441
The partial pressure of CO2 can be related to the measured pH through the equilibrium constants
for the above reactions and the equation is
pH = log N + pK1– log ( Ks pCO2)
where N = concentration of bicarbonate ion in the sensor
pK1= negative log of the acid dissociation constant for H2CO3 times the hydration constant
for CO2.
Ks = solubility coefficient for CO2
This principle for the design of pCO2 sensors has been implemented using both fluorescence-
based and absorption-based pH sensors (Vurek et al., 1983).
The methods for measuring pH, CO2 and O2 are similar, except that the wavelength of light
used for different blood gas parameters vary. The optics is composed of three channels, each for
measuring one of the parameters. Provision for calibration is made in the measuring system to
compensate for individual physical variations between sensors and monitors. The calibrationtechnique involves placing the sensor in a calibration solution, then bubbling precision mixturesof O
2, CO2 and nitrogen (N2) through the fluid. When equilibrium is reached, there are known
partial pressure of O2 and CO2 in the solution. The pH is also known from the gas tensions and the
chemical composition of the solution. Bubbling is repeated with a second gas mixture to providea second calibration point. Using both calibration points, the monitor can calculate the appro-priate calibration factors for that sensor.
With the development of fiber-optic based blood gas sensors, routine electrode membraning
and maintenance have become history. Continuous self-monitoring provides clear and immediateinformation of instrument performance. The keyboard-based user interface provides advancedanalytical performance and data processing capabilities. Along with measurement of the bloodpH, pO
2 and pCO2, some instruments like the AVL OPTI Critical Care Analyzer (Fig. 15.17) also
include facilities for measuring other important ions such as Na+, Ka+, Ca++ and Cl– in the blood.
This is possible with the development of optical sensors based on fluorescence emission. All the
sensors are mounted on a cassette shown in Fig. 15.18. The syringe adapter shown at the rightside of the cassette allows the automatic aspiration of a sample directly from a syringe. Removingthe adapter allows for direct sample aspiration from a capillary or microsampler. The sensorcalibration is verified by the system automatically after insertion of the cassette. The cassette isremoved after sample analysis.
The optical sensors in the cassette are designed in a way that the analytes bind with the
fluorescent sensor molecule. The sensor molecule is selective for the specific analyte, i.e. the pHsensor molecule reacts only with “H
+”, the O2 sensor only with O2 molecules, etc. The intensity of
the emitted fluorescent light varies with the concentration of the ions (H+, Na+, K+) or the partial
pressure of the gas molecules (O2 and CO2) in the sample. The relationship is specific for each
analyte. The corresponding calibration information for each component is encrypted in the barcode. Before the analyte can bind to the fluorescent molecule, it is made to pass through an opticalisolator. The isolator prevents interference by unspecific light with the light detection system. The
pO
2 sensor also makes it possible to measure and compute the total haemoglobin and oxygen
saturation. The equipment works on a minimum sample size of 125 ml.

442 Handbook of Biomedical Instrumentation
With the miniaturization of the direct reading electrodes, it is possible to combine them into a
single cuvette so that a complete blood gas determination could be made on a single small sample.
Fig.15.17 Critical care analyser for measuring blood gases and other
parameters (Courtesy: M/s AVL Medical Instruments)
Na sensor+
K sensor+
Ca or Cl sensor+–pCO sensor2
pO /tHb/SO sensor22
pH sensor
Fingergrips
Adapter (red) for syringe
samples (removable for
capillary samples/
AVL microsampler)
Fig.15.18 Sensor cassette (Courtesy: M/s AVL Medical Instruments)

Blood Gas Analyzers 443
The introduction of the microprocessor and its use in blood gas analyzers free the medical
personnel from monitoring the reaction in the electrode chamber and from the tedious chores ofcalculating and copying the results.
All commercial blood gas analyzers make use of the same basic electrodes and signal conditioner
circuitry. The main differences between instruments manufactured by various companies are not
the measurements of the parameters but the degree of automation and the technique by which the
sample is presented to the electrodes.

444 Handbook of Biomedical Instrumentation
/G42/G6C/G6F/G6F/G64/G20/G43/G65/G6C/G6C/G20/G43/G6F/G75/G6E/G74/G65/G72/G73
/G2016.1 TYPES OF BLOOD CELLS
Changes in the normal functioning of an organism are often accompanied by changes in the blood
cell count. Therefore, the determination of the number and size of blood cells per unit volume oftenprovides valuable information for accurate diagnosis. The blood constitutes 5–10% of the totalbody weight and in the average adult, it amounts to 5–6 l. Blood consists of corpuscles suspendedin a fluid called plasma in the proportion of 45 parts of corpuscles (cells) to 55 parts of plasma. Thepercentage of cells in the blood is called the haematocrit value or packed cell volume (PCV). Themajority of the corpuscles in blood are red blood cells (erythrocytes), others being white blood cells
(leucocytes) and platelets (thrombocytes).
Blood cells are divided into groups according to their form and function as shown in Table 16.1.
/G95/G20/G54/G61/G62/G6C/G65/G20/G31/G36/G2E/G31 Blood Cell Types
Blood cell Number of Mean cell Relative proportion
types cells in mm3volume (MCV) of different leucocyte
Inmm3count (differential)
1. Erythrocytes (4.8–5.5) ±1 ¥ 10490
2. Leucocytes 5000–10,000 100%
(a) Neutrophils 2000–7500 450 59 ± 18%
(b) Lymphocytes 1500–4000 250 34 ± 10%
(c) Eosinophils 40 –400 450 2.5%
(d) Basophils 10 –100 450 0.5%
(e) Monocytes 200 –800 600 4%
3. Thrombocytes 1.5 ¥ 105– 4 ¥ 1058–
Erythrocytes (Red Blood Cells): Red blood cells have the form of a bi-concave disc with a mean
diameter of about 7.5 m and thickness of about 1.7 m. The mean surface area of the cell is aboutHAPTER
1616

Blood Cell Counters 445
134mm2. There are about 5.5 million of them in every cubic millimetre of blood in men and nearly
5 million in women. In the whole body, there are about 25 billion erythrocytes and they areconstantly being destroyed and replaced at a rate of about 9000 million per hour. The normal red
cell lasts approximately 120 days before it is destroyed.
The erythrocytes have no nucleus. They are responsible for carrying oxygen from the lungs to
the tissues and carbon dioxide from the tissues to the lungs. Anaemia (reduction in the oxygen
carrying capacity of blood) can develop from a change in the number, volume or Hb concentrationof erythrocytes, caused by bone marrow dysfunction resulting in the poor production rate of RBCs.Since these changes are specific, the measurement of packed cell volume (PCV), the number ofRBCs and the haemoglobin (Hb) are very important.
Leucocytes (White Blood Cells): Leucocytes are spherical cells having a nucleus. There are normally
5000–10,000 white cells per cubic mm of blood but their number varies during the day. They livefor seven to fourteen days and there is a rapid turn over, with constant destruction and replacement.
Leucocytes form the defence mechanism of the body against infection. They are of two main
types: the neutrophils and the lymphocytes. Neutrophils ingest bacteria and lymphocytes areconcerned with immunological response. The number and proportion of these types of leucocytesmay vary widely in response to various disease conditions. For thus reason, it is important toknow the total leucocyte count. The change, however, is often so small that the WBC count remainswithin normal limits and only the differential count would indicate any abnormality.
Neutrophils are nearly twice as big as the red cells and contain both a nucleus divided into
several lobes and granules in their protoplasm. Lymphocytes are of the same size as the red cells butcontain a large density staining nucleus and no granules. Monocytes are another type of leucocytes,which are twice as big as the neutrophils. They have a single large nucleus and no granules.
Thrombocytes (Platelets): Platelets are usually tiny, round, oblong or irregularly shaped cells of
the blood with an average diameter of approximately 2 m. They play an important role in the blood
coagulation process. There are usually 250,000–750,000 platelets in every cubic mm of blood.
/G31/G36/G2E/G31/G2E/G31 /G43/G61/G6C/G63/G75/G6C/G61/G74/G69/G6F/G6E/G20/G6F/G66/G20/G53/G69/G7A/G65/G20/G6F/G66/G20/G43/G65/G6C/G6C/G73
Mean Cell Volume (MCV): It is calculated from the PCV and the number of red cells present per litre
of blood. For example, if the PCV is 0.45, i.e. 1 litre of blood contains 0.45 litres of red cells and ifthere are 5 ¥ 10
12 red cells per litre, then
Mean volume of one cell = 04 5
51 012.
¥ = 90 f/l f/l = Femolitres
1f/l = 10–15
Normal mean red cell volume is 86 ± 10 f/l. In diseased conditions, it may fall to 50 f/l or rise
upto 150 f/l.
Mean Cell Haemoglobin (MCH): It is calculated from the Hb and red cell count. For example, if
there are 15 g of Hb per decilitre (dl) of blood, there will be 150 gram Hb per litre of blood. Supposingthe number of red cells is 5 ¥ 10
12 per litre, then
MCH =150
51 012¥ = 30 picogram (pg)

446 Handbook of Biomedical Instrumentation
Normal mean cell haemoglobin is 29.5 ± 2.5 pg. In diseased conditions it may rise to 50 pg or fall
to 15 pg.
Mean Cell Haemoglobin Concentration (MCHC): It can be calculated if PCV and Hb per dl are
known. For example, if PCV is 0.45 and there are 15 g Hb per dl of blood, then
MCHC = 15
04 5. g/dl
= 33.3 g/dl
Mean Platelet Volume (MPV): It is the ratio of the integrated platelet volume to the platelet count
and is expressed in femolitres.
Plateletcrit (PCT): It is the percentage of the total specimen volume occupied by the platelets.
Information from the platelet count (PLT) and mean platelet volume is expressed by the followingequation:
PCT% = MPV(fl) (PLT) (10 l)
109¥¥ /
Red Cell Distribution Width (RDW): It is a numerical expression of the width of the size distribu-
tion of red cells. It is derived by analog computation. The total erythrocyte count is scanned by acontinuously variable thresholding circuit. The upper threshold is moved progressively lowerfrom a level equivalent to 360 femolitres until 20 per cent of all erythrocytes present have a sizeabove a certain value. This is recorded as the twentieth percentile value. The lower threshold is
moved downwards until it reaches a level, which 80% of all erythrocytes exceed. This is labelled
as the eightieth percentile value.
The RDW index is expressed by the following equation:
RDW =
(20th 80th) Percentile volume
(20th 80th) Percentile volume-
+¥ 100 ¥K
The constant K is the calibration factor to produce a result of 10 for a normal population.
Platelet Distribution Width (PDW): This index is related to the size range covered by those platelets
lying between the sixteenth and eighty fourth percentile. This is the conventional geometric
standard deviation of the mean platelet size and is derived from the distribution curve based on
the data in a 64-channel pulse height analyzer.
/G20/G31/G36/G2E/G32/G4D/G45/G54/G48/G4F/G44/G53/G20/G4F/G46/G20/G43/G45/G4C/G4C/G20/G43/G4F/G55/G4E/G54/G49/G4E/G47
/G31/G36/G2E/G32/G2E/G31 /G4D/G69/G63/G72/G6F/G73/G63/G6F/G70/G69/G63/G20/G4D/G65/G74/G68/G6F/G64
The most common and routinely applied method of counting blood cells even today, particularly
in small laboratories, is the microscopic method in which the diluted sample is visually examinedand the cells counted. Commonly known as the counting chamber technique, it suffers from severalcommon drawbacks. Apart from the inherent error of the system, which may be about 10%, there is
an additional subjective error of ±10% entailing poor reproducibility of the results. Furthermore,
the lengthy procedure involved results in the rapid tiring of the person making the examination.There is also poor time and labour utilisation.

Blood Cell Counters 447
Another problem with microscopic counting is that the data gathered by this measurement
is not directly suitable for storage or for further processing and evaluation. Furthermore, theincreasing number of examinations carried out in large series in busy laboratories necessitated the
development of automatic instruments for counting the blood particles, with the errors of counting
significantly reduced than a counting chamber. Agoston and Zillich (1971) compared the resultsof microscopic counting with those made by electronic counters (Fig. 16.1). It may be observed thatinstead of the ±20% measuring accuracy in microscopic counting, the electronic counters canprovide an accuracy of ±3%.
Microscopic measurement
Electronic counter
3579 1 1 1 3 1 5 1 7 1 9 2 1 2 3 195100105110
Percent deviation from nominal value (100%)
Number of measurements
Fig.16.1 Comparison of results obtained with microscopic counting and
electronic techniques (after Agoston and Zillich, 1971).
/G31/G36/G2E/G32/G2E/G32 /G41/G75/G74/G6F/G6D/G61/G74/G69/G63/G20/G4F/G70/G74/G69/G63/G61/G6C/G20/G4D/G65/G74/G68/G6F/G64
The method is based on collecting scattered light from the blood cells and converting it into
electrical pulses for counting. Figure 16.2 shows one type of arrangement for the rapid countingof red and white cells using the optical detection system. A sample of dilute blood (1:500 forwhite cells and 1:50,000 for red cells) is taken in a glass container. It is drawn through a countingchamber in which the blood stream is reduced in cross-section by a concentric high velocity liquidsheath. A sample optical system provides a dark field illuminated zone on the stream and the lightscattered in the forward direction is collected on the cathode of a photomultiplier tube. Pulses are

448 Handbook of Biomedical Instrumentation
produced in the photomultiplier tube corresponding to each cell. These signals are amplified in a
high input impedance amplifier and fed to an adjustable amplitude discriminator. The discriminatorprovides pulses of equal amplitude, which are used to drive a digital display.
Instruments based on this technique take about 30 s for completing the count. An accuracy of
2% is attainable. The instruments require about 1 ml of blood sample.
/G31/G36/G2E/G32/G2E/G33 /G45/G6C/G65/G63/G74/G72/G69/G63/G61/G6C/G20/G43/G6F/G6E/G64/G75/G63/G74/G69/G76/G69/G74/G79/G20/G4D/G65/G74/G68/G6F/G64
Blood cell counters, operating on the principle of conductivity change, which occurs each time acell passes through an orifice, are generally known as Coulter Counters. The method was patentedby Coulter in 1956 and it forms the basis of several particle counting instruments manufactured bya number of firms throughout the world. The technique is extremely useful for determining thenumber and size of the particles suspended in an electrically conductive liquid.
The underlying principle of the measurement is that blood is a poor conductor of electricity
whereas certain diluents are good conductors. For a cell count, therefore, blood is diluted and thesuspension is drawn through a small orifice. By means of a constant current source, a direct
current is maintained between two electrodes located on either side of the orifice. As a blood cell is
carried through the orifice, it displaces some of the conductive fluid and increases the electricalresistance between the electrodes. A voltage pulse of magnitude proportional to the particle volumeis thus produced. The resulting series of pulses are electronically amplified, scaled and displayedon a suitable display.Read out Discriminator
Amplifier
Photo
multiplier
Volume
relay
SampleSheath flowDark
ground
illuminator
Fig.16.2 Optical method of counting cells

Blood Cell Counters 449
To achieve optimum performance and to enable the relationship of change in resistance with
volume of the cell to hold good, it is recommended that the ratio of the aperture length to thediameter of the aperture should be 0.75:1, i.e. for an orifice of 100 m diameter the length should
be 75 m.
The instrument based on the Coulter principle works most satisfactorily when the average
diameter of the particles ranges between 2 to 40% of the diameter of the measuring hole. Therefore,
the following condition must be met for the measuring range:
D/50£d£D/2
where d =maximum particle size
D = diameter of the measuring aperture
The lower limit of measurement in the system is governed by the noise sources involved. The
noise sources include the thermal noise of the detector due to the resistance of the fluid flowingacross the orifice and the noise inherent in the electronic circuits.
Particles of sizes larger than the diameter of the measuring aperture can only pass through
the aperture if their longest dimension is parallel to the axis of the measuring aperture; otherwisethey cause the clogging of the aperture. The upper limit of measurement is thus imposed by theincreasing size of the particles. When the size of the particle approximates the diameter of theaperture, an amplitude linearity error is produced. Therefore, to count particles of different sizes,the diameter of the measuring aperture must be chosen in such a way as to meet the conditionsof measurement. Simple procedures enable the extention of the range as needed via the use ofsuccessive aperture size—the total range covered is from about 0.5 microns to upwards of 500
microns. The applicability of the Coulter principle for measuring particles of various sizes is
shown in Fig. 16.3.
0.0001 0.001 0.01 0.1 1 10 100 1000 10000
1 A° 1 cm
Atoms ColloidsLight waves limit of visibilityCoulter principle
Sieves
Microscope
Sedimentation CentrifugeElec
micr.
Ultra
cent
Fig.16.3 The applicability of Coulter principle in the range from 0.5 microns
to 500 microns
/G20/G31/G36/G2E/G33 /G43/G4F/G55/G4C/G54/G45/G52/G20/G43/G4F/G55/G4E/G54/G45/G52/G53
A wide range of particle counting instruments designed to meet a wide variety of needs in the
haemotology laboratory are being commercially produced. These instruments range from the small
counters used primarily for red and white cell counts in very small hospitals and clinics, to the

450 Handbook of Biomedical Instrumentation
multi-parameter microprocessor controlled instrument featuring fully automatic diluting of
samples and printing of results.
Figure 16.4 shows a block diagram showing the principle of a Coulter counter. A platinum
electrode is placed inside the orifice tube and a second electrode is submerged into the beakercontaining the cell dilution, creating an electrical circuit between the two electrodes. Current will
flow from one electrode to the other through the orifice. When the cell suspension is drawn through
the orifice, cells will displace their own volume of electrolyte and cause a resistance change, whichis converted to a voltage change, and is amplified and displayed.
ABMain
amplifier
Horizontal
sweepThreshold
circuitPulse
amplifier
Counter
driver
Digital
registerPulse brightening signal
Scope
Counter “Start-Stop”
Fig.16.4 Principle of Coulter counter
In practice, the cell suspension is drawn through the orifice by means of a mercury manometer.
This manometer includes two platinum wire contacts (A and B) set through the glass walls.Contact A will start the count and contact B will stop it when precisely 0.5 ml of the dilution haspassed through the orifice tube. Thus, it provides a count of the number of particles in a fixedvolume of suspension. Figure 16.5 shows the sequence of building up the pulse in terms of increasein resistance at different positions of the cell with respect to the orifice.
To enable the instrument to count only those pulses, which fall within certain preset size limits,
the threshold facility is required. The threshold is also necessary to enable the instrument to ignoreany electronic noise, which may be present in the system. The lower threshold sets an overallvoltage level, which must be exceeded by a pulse before it can be counted. The upper threshold willnot allow pulses to be counted which exceed its preset level.
The Coulter counters are usually provided with an oscilloscope monitor to display the pulse
information, which has passed through the amplifier, and acts as a visible check on the counting
process indicating instantaneously any malfunctions such as a blocked orifice. In particular, it
provides information regarding (i) relative cell size, (ii) relative cell size distribution, (iii) settingsof the threshold level control, and (iv) means to check the performance of the instrument forreliability of counts. The voltage pulses produced each time a cell passes through the orifice aredisplayed on the oscilloscope screen as a pattern of vertical spikes.

Blood Cell Counters 451
Coulter counters also help to give an idea of the size distribution of various types of cells. It has
been stated that the pulse-height is to a first approximation, proportional to the volume of theparticle. Converting the pulse height into a digital number, through an A-D converter, and storingit in memory can help to obtain a plot of the number of cells as a function of their size (McGannet al 1982).
Taylor (1970) suggests that an aperture diameter of 100 m would be generally useful. For such an
aperture, a length of about 200 μ and flow rate of 0.04 ml/s would be optimum. The aperture canbe made using ruby watch jewels bonded to a glass surface. Typically, an aperture of 100 m
diameter and 200 m length, separating two solutions of phosphate buffered saline, has a resistance
of about 25 k W and capacitance of 120 pF. The minimum rise time is about 5 ms. This means that the
electronic circuit must have upper frequency response greater than 70 kHz. The preamplifier usedin cell counters must be of very low noise preferably having noiset voltage less than 2 nA at therequired bandwidth of 70 kHz.
Calibration: The calibration factor is constant for a given aperture size, electrolyte resistivity and
amplifier gain setting. It is used for the conversion of threshold settings to particle volumes or theircube roots to equivalent spherical diameters. Calibration is done simply and quickly by observingthe threshold for monosized particles of known diameter (adjusting the threshold level to the peaksof the single-height pulses on the oscilloscope screen). Ragweed pollen (19 micron in diameter)
and polystyrene latex particles (6–14 micron in diameter) seem to meet these requirements. Of the
two, polystyrene latex is preferred for calibration purposes (Thom, 1972). The particles when usedseldom plug the orifice. These can be conveniently obtained in the range of 5 million particles percubic mm.
/G31/G36/G2E/G33/G2E/G31 /G4D/G75/G6C/G74/G69/G2D/G70/G61/G72/G61/G6D/G65/G74/G65/G72/G20/G43/G6F/G75/G6C/G74/G65/G72/G20/G43/G6F/G75/G6E/G74/G65/G72
Figure 16.6 shows the block diagram of a multi-parameter Coulter counter. It provides theuniversally accepted profile of white cell count, red cell count, mean cell volume, haemotocrit,
mean cell haemoglobin concentration, mean cell haemoglobin and haemoglobin. Besides this, the12 3 4 5 6Orifice resistance
Position of particle123456
Fig.16.5 The sequence of building up the pulse in terms of increase in resistance at
different positions the cell has with respect to the orifice

452 Handbook of Biomedical Instrumentation
following five parameters are presented: platelet count, red cell distribution width, mean platelet
volume, plateletcrit, and platelet distribution width. This instrument is microprocessor-controlledand provides the flexibility of expressing in various forms the available count and size data storedin memory.
All the directly measured parameters are measured in triplicate and the average results are
displayed. All the 14 parameters are obtained from 1 ml of whole blood in 34–50 s dependingon the number of platelets present. All diluting and mixing of samples is controlled by themicroprocessor, which at the same time monitors the instrument and checks all the electronic
circuits and pneumatics for correct functioning. A specially designed ticket is provided showing
results of the measurements and computations. All this information is presented simultaneously(Fig. 16.7) so that the data can be quickly reviewed just by scanning the display. Since eachdistribution is an average of the information from the three measurement apertures, data from anyone of the three apertures, can be examined, one after the other, or overlaid, one upon the other, fora comparative review.MCV 1
MCV 2
MCV 3MCV
comp.
module
RBC 1
RBC 2
RBC 3
PLT 1
PLT 2
PLT 3PLT
comp.
moduleRDWRBC coincidence
correction and
comparison moduleHCT
MCH
MCHC
comput.Printer
X-Y
recordPlatelet
size
dist.RBC
bath
WasteWaste
RinseRinseRinse
Hgb
cuvetteWBC
bathWBC 1
WBC 2WBC 3WBC
comp.
module
WBC
Dil. mod
RBC
Dil. mod
Fig.16.6 Block diagram of multi-parameter Coulter counter (Courtesy: Coulter
Electronics, USA)

Blood Cell Counters 453
Coincidence Compensation: Coincidence error occurs when two or more particles are present in
the sensing zone at the same time. This will result in the instrument detecting fewer particles than
are actually present. This will also result in the instrument adding the pulses together to producea single much longer pulse (Fig. 16.8). As the instrument detects fewer pulses than are actuallypresent, to render the total count accurate, it is necessary to add on the pulses that have been ‘lost’due to coincidence. The rate at which this happens has been mathematically determined. ModelS-plus Coulter counter automatically compensates for this loss. With some instruments, acorrection chart is available to allow the correct number to be determined. Under a total count of10,000 pulses, primary coincidence is negligible and can be ignored.
Coulter counters have a serious drawback linked with the mercury manometer arrangement.
The surface of the mercury gets dirty as a consequence of which the contact bordering the volumebecomes uncertain, which may make the measured values uncertain. The mercury also dirties theneighbouring glass wall, which further causes uncertainty in counting. Also, the instrument canonly be transported and stored after removing the mercury from it.
/G31/G36/G2E/G33/G2E/G32/G50/G69/G63/G6F/G73/G63/G61/G6C/G65
Based on the same principle of detecting change in conductivity in the presence of a cell in theorifice of a measuring tube, there is another cell counting instrument known by the name Picoscale.This instrument does not make use of a mercury manometer for fixing the volume thus eliminatingthe problems associated with its use.
This instrument is primarily meant for counting RBC, WBC and PBC and is manufactured by
MEDICOR, Budapest. In this instrument (Fig. 16.9), a glass measuring tube ‘C’ provided with an5050
2100100
10200200
20 Cubic micrometers PLTRBCWBCREL
No.
300300 400WBC15/01/81 026
OP CODE
RBC
HGB
HCT
MCVMCHMCHCRDWPLTPCT
MPV
PDWLY9.5
4.9615.243.888.330.634.612.8315..2608.215.4
34
Fig.16.7 Simultaneous display of normal white cell data, red cell distribution curve
on the data terminal of the model S-plus Coulter counter

454 Handbook of Biomedical Instrumentation
aperture ‘A’ is immersed into the suspension. The pressure difference created between the two
sides of the aperture draws the suspension to flow through the aperture. This pressure differenceis generated by a simple mechanical pump consisting of a syringe, a relay and other parts.
A constant current is normally passed between the electrodes E
1 and E2. Therefore, the electric
resistance of the liquid measured between these two electrodes changes rapidly when a particle
having electric conductance differing from the conductance of the electrolyte passes through the
aperture. This results in the generation of a voltage pulse, which is amplified in a preamplifier ofhigh gain and low noise level. The output signal of this stage goes to a discriminator, whichcompares the amplitude of the pulse arriving at its input with the preset triggering level. If theinput signal exceeds the triggering level, the discriminator gives out a pulse of constant shape andamplitude. These pulses go to a counting circuit for the display of the measured parameter.1
2122 112PassageR
R
R1
Time
Time
Time2
1 + 2
Coincident passage
Fig.16.8 Presence of more than one particle in the orifice resulting in loss of some
valid pulses

Blood Cell Counters 455
The measuring tube C is provided with a third electrode E3 which helps to monitor the suction
of a limited volume of the suspension. When the liquid level reaches E3, the pump is changed over
from the suction phase to the pressure phase. The counting process also occurs during the time theelectrolyte is forced out of the measuring tube. After the liquid looses contact with the electrodeE
1, the counting automatically stops and the unit becomes ready once again for the operation. The
manufacturers recommend that to increase the reliability of the results, it is preferable to repeat themeasurements several times, and calculate the mean value based on these measurements.
One great advantage of this instrument is that the clogging of the capillary is greatly eliminated
by applying a bi-directional flow during the measurement procedure. The number of particles N in
a unit volume is determined from the relation,
N=HLE
V
where
H= factor of dilution
L= scaling factor of the counter
V= measured volume
E= result displayed on the digital display.
For example, if the diluting factor is 63,000 (typical for red cell count in this instrument) and 520
appears on the display; L being 60, and V equal to 0.378 cm3, then
N= (. ) ( ) (. )
(. )6 3 10 60 5 20 10
37 8 1 042
2¥¥ ¥
¥
= 5.20 ¥ 106per mm3Suction
pressureVolume
limiter
Amplifier Counter
Digital
display
ACE2E3 E1
Fig.16.9 Block diagram of 'Picoscale' blood cell counter

456 Handbook of Biomedical Instrumentation
In other words, the solution contains 5.2 million blood cells per cubic millimetre.
The capillary diameter for red cell count is 72 mm and the dilution factor is 63,000. For white
cells, the diameter is 102 mm, and the dilution factor is 630. For platelet count, the diameter of the
capillary is 72 mm and a dilution of 6300 is used.
/G31/G36/G2E/G33/G2E/G33 /G45/G72/G72/G6F/G72/G73/G20/G69/G6E/G20/G45/G6C/G65/G63/G74/G72/G6F/G6E/G69/G63/G20/G43/G6F/G75/G6E/G74/G65/G72/G73
There are a number of errors that may occur in the electronic cell counting technique. Briefly, these
errors are categorized as follows:
Aperture Clogging: Partial clogging of the measuring aperture adversely affects the results of
counting. The aperture diameter becomes constricted and even small particles that were notintended for counting get included in the final count. The measuring time also becomes larger.Clogging of the aperture is cleared by cleaning the aperture and using a suction-pressure cycle bymeans of a pump.
Uncertainty of Discriminator Threshold: The uncertainty of triggering at threshold level and the
threshold hysteresis means that particles of the same size are sometimes counted and sometimesleft out. In addition to this, a threshold uncertainty error may also be introduced if the measuringcapillary has a diameter larger than that rated because the passage of cells in a capillary of largerdiameter will result in smaller changes in the resistance.
Coincidence Error: Coincidence error will be observed if more than one particle passes through the
measuring aperture simultaneously. For the calculation of this error, a Poisson distribution isassumed whereby it is possible to determine the probability of 0, 1, 2, . . . , n particles entering the
measuring capillary during the time only a single count is produced. The coincidence loss dependsupon the length and diameter of the aperture, the number of particles per unit volume and thedilution ratio. A correction for the coincidence loss is usually incorporated in the instruments.
Settling Error: This error arises due to the settling of the particles in the solution, with the result
that the measurements show a decreasing tendency with time. If the readings are taken within4–5 min., the settling error is less than 1%.
Statistical Error: Suppose that during a time t, n particles are detected. Then, during an additional
time t, the number of particles detected will be n±
n. Therefore, the mean statistical error = n
and the relative error = ±100/ n%. Assuming Gaussian distribution, the mean statistical error
means that 67% tests fall into the interval n± n with 33% of the measurements greater than that.
To obtain the statistical error, the instrument reading should be multiplied by the scaling factor
of the counter. This will yield the value of n.
Error in Sample Volume: The measuring volume should be as accurate as possible because the
magnitude of the measuring volume affects the accuracy of measurement. Greater the samplevolume, the more accurate will be the result. The diameter of the measuring tube should be veryaccurate as it appears on the second power in the volume calculations.
Error due to Temperature Variation: Conductivity of the electrolyte increases with rising temperature.
Hence, the specific resistance of the solution decreases, resulting in lowering the amplitude of the

Blood Cell Counters 457
pulse generated by the particle. For accurate results, the temperature of the solution should be
maintained constant.
Biological Factors: Factors like deformation of cells in the diluting solutions while passing through
the aperture, the haemolyzing effect under the influence of electric fields and the presence ofbacteria, etc. also introduce errors in the measurement process.
Dilution Errors: To obtain accurate measurements, the suspension containing the particles
should be prepared with utmost care and with minimal error possible. The dilutions should bemade accurately and the solution should be of a nature that the cells do not shrink or swell.
Error due to External Disturbances: A high intensity magnetic field in the vicinity of the instrument
may interfere with the measurement. It might be necessary to screen the instrument with somemagnetic material.
The sources of errors enumerated above originate from random factors and since the results too
would contain random errors, different results may be observed in repeated tests. The actual valueshould therefore be determined with due consideration of the fluctuations, measurement lossesand other circumstances.
/G20/G31/G36/G2E/G34 /G41/G55/G54/G4F/G4D/G41/G54/G49/G43/G20/G52/G45/G43/G4F/G47/G4E/G49/G54/G49/G4F/G4E/G20/G41/G4E/G44/G20/G44/G49/G46/G46/G45/G52/G45/G4E/G54/G49/G41/G4C
/G43/G4F/G55/G4E/G54/G49/G4E/G47/G20/G4F/G46/G20/G43/G45/G4C/G4C/G53
Along with the automated instruments for obtaining the erythrocyte, leucocyte and platelet counts,
there has been a considerable interest in developing automated techniques for identifying andcounting the different types of cells within a given class. Examples of this could be the immaturered cell count, the differential leucocyte count and the recognition of normal versus malignantcells in other cell types. Various diseases affect the mechanism of blood cell formation in differentways. In particular, fractional proportion of the five major leucocytes is a sensitive measure forindicating and assisting in the diagnosis of various diseases. Furthermore, some diseases causeimmature forms, which normally are present in the blood forming tissue to appear in the peripheralblood. Some diseases affect red cell morphology, and a qualitative report of red cells also forms apart of the differential count. There is, thus, a need for automation of acquisition and interpretation
of data in routine clinical differential cell count. Several approaches have been employed in the
pursuit of automating the techniques.
Miller (1976) describes a differential white blood cell classifier based upon a three-colour flying
spot-scanner approach. It utilizes recognition parameters based on the principle of geometricalprobability functions, which are generated at high speed in a dedicated computer. The systemuses a conventional microscope with automatic focus and stage motion. A block diagram of thecell identification steps of the system are shown in Fig. 16.10. The system is built around a Zeissmicroscope with two 15 ¥ eyepieces and a 40 ¥ oil immersion objective and with computer-
controlled focusing. A television monitor displays the data and shows the relative position of thecells in each field. The slide is positioned precisely on the transport to allow reloading, oil isapplied from a reservoir, and the objective is focused. A cathode ray tube generates a flying spot,which traverses the slide in 1 mm-wide sweeps. When a cell nucleus is located, the coordinates of
the beam are immediately stored. A 24 ¥ 32 mm window is established around the centre of the

458 Handbook of Biomedical Instrumentation
nucleated cell, and the spot scanner is reset to sweep the area in 0.25 mm wide passes. The light
passing through the cell is split and sent through three filters to break it into its red, green, and bluecomponents, with highly sensitive photomultiplier tubes detecting the resulting radiation. This
system improves cell classification and allows wedged as well as spun slides to be used. Based on
the number of grid positions and three colours, the system has over 3 million data values to use ineach cell identification. These are used to first locate and then characterize the nucleii based onsize and segmentation. The system is capable of determining segmented neutrophils, bands,eosinophils, basophils, lymphocytes, and monocytes, as well as abnormal cells such as atypicallymphocytes, blasts, nucleated red cells, and immature granulocytes. In addition, the systemcarries out the red cell morphology, evaluating size, shape, and colour, counts the reticulocytes,estimates the platelet count and plots a distribution of red cell diameters. The instrument therefore,performs all the functions that a human being sitting at a microscope would be expected to do.However, the system is much faster than a human technologist as it can process up to 100 slidesper hour.
If the cell is identified as one of the six normal cell types, it is counted otherwise it is shown as
a suspect cell to be identified by the operator. The same field is searched again for other nucleatedcells. If no other is found, then the stage drive is incremented and the electronic search resumes in
the new field presented to the system.
Video
scanner Colour
analyzer
Stage and Focus
Motor DrivesImage memory
Reference
memory
Morphological
analyzerRecognition
computerKeyboard
Normal
cell
counter
‘‘Suspect’’
cell
counterMicroscope
Video
displayTicket printer
Fig.16.10 Block diagram of cell identification system

Blood Cell Counters 459
The automatic focus mechanism operates from high frequency video information as an error
signal to position a special piezoelectric focus drive. The piezoelectric crystal adjusts the lensassembly so as to maximize the high frequency content of the video signal.
The pattern recognition program was developed from 21000 normal and abnormal cells and
artefacts, from 700 different people and was stored on magnetic tape. It was observed that the
degree of agreement between the instrument developed (Hematrak, registered trade mark of
Geometric Data Corporation, Wayne, PA. 19087) and the technologist was 98%. The time requiredto perform a differential count on the instrument was small enough to yield a throughput of 40slides per hour. An example of an automated cell recognition and counting system is that of the“diff-3” system which is illustrated below.
/G31/G36/G2E/G34/G2E/G31 /G93/G64/G69/G66/G66/G2D/G33/G94/G20/G53/G79/G73/G74/G65/G6D
System Concept: The “diff-3” System (Perkin-Elmer, USA) is a completely automated system for
the evaluation of red and white human cells and platelets. The system electronically examinesconventional microscope blood smear slides and employs optical pattern recognition techniquesto achieve the following:
• Counts and differentiates seven important categories of red blood cells (erythrocytes);
three based on size, two on colour, one on shape and one covering red cells with nucleii
(nucleated red cells).
• Enumerates white blood cells (leucocytes) and differentially classifies them into the 10
most significant medical categories and estimates their total number.
• Surveys platelets (thrombocytes) cell size and sufficiency.
The system is designed to analyze standard slides at a 35 to 40 slides per hour rate. The actual
analysis task takes only 90 s. In fact, the system largely duplicates mechanically and opto-electronically, the manual procedures followed while examining blood smears with a microscope.
System Operation: The blood sample is first made into a conventional slide by a slide-spinner that
ensures a monolayer of cells on the slide. About 200 m of the sample is aspirated via the built-in
pipetter-diluter and dispensed on to the standard 1 ¥ 3 unit (25 ¥ 75 mm) slide which has been
labeled previously and put into a slide spinner unit. The slide will spin under low RPM (3200)for 0.6 s and comes to a stop. While the slide is spinning, the pipetter is flushed automatically toprevent sample carry over. The low spin speed and short spin time helps create consistently highquality slides with excellent morphology and random cell distribution. Slides are then removed byhand and taken to the slide staining unit—use of Wright’s stain and an Ames Hema Tek automatic
stainer helps assure the consistency of stained smears. If necessary, the system automatically com-
pensates for normal variations in stain. After staining the slide, they are loaded into the magazine.The system keeps track of the magazine and position number to ensure correct patient identification.
Upon loading, the slides go to the blood smear analyzing part, which is an open L-shaped desk
console unit. Built into the system is a microscope with a binocular eyepiece angled for easyobservation by the operator; easy to read TV screens and the signal processing equipment.
Slides inserted into the smear analyzer are moved onto the stage of a binocular microscope
(400 ¥ 1.0 N.A. Zeiss Optics) receiving a drop of oil for the microscope’s immersion type objective

460 Handbook of Biomedical Instrumentation
lens as it moves from loading breech to the stage. Automatically, the slide is focussed for the
electronic analyzer system. Light from the illuminated slide is split between the electronic patternrecognition unit and the microscope eyepiece for viewing. The microscope’s motorized stage
moves the slide in a programmed, back and forth raster pattern over the slide. Its movement is
halted upon recognition of white cells. At each stop, the system studies isolated white and redcells and platelets.
Each white cell thus found is isolated from its surroundings via electronic image scanning and
digital techniques. From then onwards, an image processor (Golay Logic Processor) takes overand carries out a total of 50 measurements on the white cells. These measurements are thosecommonly used in manual practice, i.e. evaluation of cell size, colour of the cytoplasm, number ofnuclear lobes, etc. On the basis of these measurements, the cells of the blood are differentiated andcounted. White cells, while being counted, are categorized as one of the 10 following types:
(i) segmented neutrophils, (ii) bonded neutrophils, (iii) eosinophils, (iv) basophils, (v) immature
granulocytes, (vi) lymphocytes, (vii) monocytes, (viii) atypical lymphocytes, (ix) blasts, and(x) others (generally bizarre and usually less than 1% of the cells checked). In the final print-out, anestimate of the WBC count per cubic millimetre is also available.
The red cells are isolated and examined at the same time as the whites are being studied. Some
20 measurements are made on each individual red cell, so that RBCs can be classed by formand structure into five primary groups with several subdivisions (i) nucleated red blood cells:reported as NRBCs per 100 WBCs, (ii) hypochromic cells, (iii) polychromatophilic cells, (iv) sizedistribution—the printout indicates the relative percentage of red cells which are small, normal or
large in size and (v) mis-shaper cells. These along with the spherocytes and target cells are grouped
under the term ‘poikilocytes’.
Platelets are examined for size and sufficiency. Giant platelets are reported as count per 100
WBCs whereas platelet count is shown per cubic millimetre. Sufficiency of platelets is estimated aslow, normal or high.
Image Processor: The system uses two computers. A general purpose minicomputer operates and
controls the system and makes it inherently more flexible than systems using hard-wired logic.
The second computer is a special purpose pattern recognition computer, the Golay Logic Processor(Golay, 1969), which enables the system to transform cell pattern recognition information intodifferential results. Golay logic (Preston et al 1979) enables the system to ‘see’ a cell in much the
same way as a technologist does. For example, in classifying nucleated cells, it evaluates the cellarea, nuclear area, nuclear colour, nuclear clumps, cytoplasmic colour, granular colour, nuclearlobe count, and about 35 other measurements. Similarly, for red cell morphology, the systemevaluates the cell area, area of central pallor, optical density of the entire cell, concavities,irregularities of shape, and about 10 other measurements.
In performing classification, it positions a nucleated cell in a 50 mm¥ 50 mm field. This field is
scanned as 4096 discrete data points called picture elements. The separation between data pointsis 0.8 mm. This low resolution scan is used to automatically adjust the fine focus and confirms that
a suitable cell has been found rather than dirt or debris. Then, automatically switching to a higherresolution, a field of 25 mm¥ 25mm is scanned and another 4096 data points are examined at a
resolution of 0.5 mm. Each picture element, representing a 0.4 mm¥ 0.4mm point carries a digital

Blood Cell Counters 461
number representing its optical density at two different wavelengths. This picture is stored in the
computer’s memory (Fig. 16.11a).
Fig.16.11(a) Image of a band neutrophil and surrounding red cells as stored in
video picture memory
The contents of the video memories are processed by the special purpose pattern recognition
computer that examines these pictures. The process starts by generating separate, simplifiedpictures of separate parts of the cell such as the nucleus cytoplasm, granules, etc. The pictures aremade by applying the principle that picture elements with about the same transmission are mostlikely located in the same parts of the cell. For example, all the very dark picture elements havinglow transmission values at a certain wavelength are likely to be in the nucleus, and so on. Eachpicture element is then turned ‘off’ or ‘on’ when it falls within an optical density range underinvestigation. The resulting pictures resemble line drawings, or silhouettes of parts of the cell.
The logic processor examines the ‘on’ and ‘off’ status of each picture element and its six nearest
neighbours at many optical density levels. In this way, it reconstructs the image being scanned foridentification. For example, a point will be judged to be an ‘interior point’ if the point is determinedto be on and its six nearest neighbours are also determined to be on. For a background point, allseven points will be judged to be off. An edge point is recognized as a combination of on and off
points. In all, 14 combinations are possible to define elements such as convex edges, concave
edges, strings, branches and end points (Fig. 16.11b). The processor can examine the outline of acell by turning on only the edge points of the cytoplasm. The picture represented in Fig. 16.11(c)would result. By turning on only interior points of the nucleus, Fig. 16.11(d) would result. Havingisolated a single cell, the processor evaluates the nucleus and cytoplasm of the cell for shape,colour, texture and other elements.
In addition to a thorough examination of each cell, the processor must also ensure that only one
cell is evaluated at a time. Also, the measurements of the cell area, nuclear colour, etc. are used inmathematical equations to decide the type of cell present. Comparisons are conducted on each celland a required number of conditions are needed before classification is made into any category. Ifno category meets a sufficient number of conditions, then the cell will be classified as an ‘other’.

462 Handbook of Biomedical Instrumentation
Fig.16.11(b) Examples of points identified by the Golay logic processor
(c) (d)
Fig.16.11(c) Outline of whole cell made by turning on only edge points in
picture of the whole cell
(d) Picture of the nucleus generated by turning on only interior points of
the necleus
This may happen with some frequency on abnormal samples, but it happens only occasionally on
normal samples.
At every observation step, the instrument shows the cells it is studying on a CRT screen, exactly
as the technologist would view it in a microscope at the same time. The results of examinations arealso shown upon completion of the analysis on the CRT. End results are printed in triplicate on an
8¥ 22.5 cm ticket.

/G41/G75/G64/G69/G6F/G6D/G65/G74/G65/G72/G73/G20/G61/G6E/G64/G20/G48/G65/G61/G72/G69/G6E/G67/G20/G41/G69/G64/G73
/G20/G31/G37/G2E/G31 /G4D/G45/G43/G48/G41/G4E/G49/G53/G4D/G20/G4F/G46/G20/G48/G45/G41/G52/G49/G4E/G47
Sound waves are longitudinal waves in which the motion of each particle of the medium in which
the wave is travelling, moves backward and forward along a line in the direction in which thewave is propagated. The human aural system reacts to these oscillating pressure changes andtransmits them to the brain through a series of steps. Figure 17.1 shows the anatomy of the humanear. The outer ear consists of the pinna or auricle, together with the ear canal, the external auditory
meatus, which is a convoluted tube, about 1 cm
3in volume and terminates at its inner end in theHAPTER
1717
Tympanic
membraneMiddle ear
cavityEustachian
tubeCochleaAuditory
nerveStapes in
oval windowMiddle ear
bonesEar canal
Pinna
Fig.17.1 Anatomy of ear

464 Handbook of Biomedical Instrumentation
tympanic membrane. The pinna scatters acoustic waves so that some of the scattered energy enters
the auditory canal and pushes against the tympanic membrane during a wave of compression.The distance membrane moves is a function of the force and velocity with which the air molecules
strike it and is, therefore, related to the loudness of sound.
The tympanic membrane separates the ear canal from the middle ear cavity. The middle ear is
exposed to atmospheric pressure only through the eustachian tube, which connects it to the
pharynx and nose or mouth. The sound energy from the tympanic membrane is transmittedthrough the cavity of the middle ear, to the receptor cells in the inner ear, which are surrounded byfluid. Thus, the major function of the middle ear is to transfer movements of the air in the outer earto the fluid-filled chambers of the inner ear. A chain of three small, middle ear bones couple thetympanic membrane to a membrane covered opening, called the oval window. The total force onthe oval window is the same as that on the tympanic membrane. The size of the window being verysmall, it experiences much greater force per unit area. One of the bones, called the stapes, restsupon the lower end of the cochlea and passes the vibrations directly into the fluid within.
The inner ear or cochlea is a fluid-filled coiled passage in the temporal bone. It is almost
completely divided lengthwise by the basilar membrane. Most of the pressure wave received bythe cochlea is transmitted to this membrane, which is deflected into the scale tympani. Themembrane has different resonating properties along its length, responding to high frequencies atthe stapes end and to low frequencies at its upper end. The membrane contains the sensitive
receptor cells, which transform sound energy or pressure waves into action potentials. The nerve
impulses thus initiated are propagated along the acoustic nerve fibres to the brain with a speed of100 m/s. The pattern of nerve impulses arriving in the brain is associated with the subjectivelyexperienced sound, which has attributes of loudness, pitch and timbre (quality).
The appreciation of sound is mainly a cerebral function. However, the recognition of notes is
partly a function of the cochlea. Therefore, if it is defective, the individual may not hear certaintones.
Hearing is affected by anything which interferes with the conduction of sound waves to the
cochlea, such as a perforated tympanic membrane (ear drum), disease of the middle ear or diseaseof the cochlea itself or its connection in the central nervous system.
The sounds reaching the ear are characterized by loudness (intensity), which depends upon
the amplitude of the waves; by pitch, which depends upon their frequency; and by quality, whichresults from the combination and interaction of the waves. The human ear responds to vibrationsranging from 20 to 20,000 Hz. The waves of speech and many other common sounds are not ofsingle frequency but are complex waves made up of several frequencies of vibration. The numberof sound frequencies in addition to the fundamental tone, i.e. the degree of purity of the soundwave is related to the quality or timbre of the sound. The human ear can in fact, distinguish some400,000 different sounds.
/G31/G37/G2E/G31/G2E/G31 /G41/G69/G72/G20/G61/G6E/G64/G20/G42/G6F/G6E/G65/G20/G43/G6F/G6E/G64/G75/G63/G74/G69/G6F/G6E
Air conduction, by definition, is the transmission of sound through the external and middle ear tothe internal ear. Bone conduction, on the other hand, refers to transmission of sound to the internalear mediated by mechanical vibration of the cranial bones and soft tissues. The most important

Audiometers and Hearing Aids 465
diagnostic differential from the standpoint of the functional hearing tests is the relationship
between air and bone conduction acuity. Clinical observation has shown that hard-of-hearingpatients with middle ear disease usually have normal hearing by bone conduction, whereas
patients with inner ear involvement have decreased or diminished bone-conduction.
It has been concluded from clinical observations that an approximate 60 dB loss is the maximal
air conduction impairment to be anticipated with middle ear defect. Therefore, if the air conduction
loss in a patient with apparently typical middle ear pathology exceeds 60 dB, it is likely that innerear impairment is superimposed on the middle ear lesion.
Figure 17.2 shows that upto a frequency of about 14–16 kHz, the thresholds for air conduction
and bone conduction have the same shape. At this frequency, they both show an abrupt fall insensitivity with a slope of 50 dB/octave. The start of this slope defines the ‘end point’ of the ear. Forair conducted signals, the fall in sensitivity continues, so that for instance at 25 kHz, 5 W ofacoustic power (equivalent to about 500 W of electrical power) is needed to produce a hearingresponse. On the other hand, for bone conducted signals, there is a change in slope again, at about2 kHz above the end point. From then on up to 200 kHz, the threshold sensitivity falls at a rate of10 to 15 decibels per octave. So, in the ultrasonic region, a bone conducted signal of less than oneelectrical watt is audible.
2020406080100120
143 5 6 8 10 14 16 20 30 48 55 77 95dB
Air conduction
15 dB/octaveBone conduction
50 dB/octave
Frequency kHz
Fig.17.2 Air and bone conduction response curves of a normal ear

466 Handbook of Biomedical Instrumentation
There is a rapid drop in impedance of the middle ear at high frequencies and very little of the
acoustical energy fed to the ear by air conduction is transmitted to the cochlea. But bone-conductedsound by-passes the middle ear, as is shown in Fig. 17.3. This to some extent explains the different
threshold shapes at high frequencies.
Ear drum and
middle ear
Inner ear
Cochlea
Basilar
membrane
To brain
Auditory nerves
Bone
conducted sound
to cochlea through skullBone
vibratorEar canalAir conducted sound
to cochlea via
ear canal and
middle ear
Loudspeaker
Fig.17.3 Mechanism of air and bone conduction
/G31/G37/G2E/G31/G2E/G32 /G54/G68/G72/G65/G73/G68/G6F/G6C/G64/G20/G6F/G66/G20/G48/G65/G61/G72/G69/G6E/G67
The threshold pressure level of a sound is the lowest level at which an observer can discriminate
between the desired sound and the noise background always present in the auditory system. In
other words, it is minimal intensity on the stimulus scale, which is just barely adequate to elicit aresponse. The hearing threshold is not an invariable fixed intensity above which sound is alwaysheard and below which sound is never heard. In fact, the sensitivity of the auditory mechanism isfound to vary with interactions between certain physiological, psychological and physical factors.Therefore, the threshold may be regarded as an intensity range within which sound stimuli at ornear the statistically determined threshold may or may not be perceived.
The base of each audiological examination is the determination of the hearing threshold. A
complete audiogram, one which shows both air and bone conduction threshold bilaterally, is notonly a graphic representation of the dB loss at different frequency levels for both air and boneconduction but also indicates the type and location of the hearing impairment. Thus, on the basisof the audiograms, the conduction and perceptive hearing deficiencies can be easily separatedfrom each other.
After establishing the sensitivity of the ears to air bone sounds, it is necessary to know whether
the hearing loss is due to a disorder of the hearing organ (cochlea) and connecting nerves with thebrain (perceptive loss) or whether the loss arises because of reduced transmission of sound
vibrations through the middle ear mechanism itself (conductive loss).

Audiometers and Hearing Aids 467
/G20/G31/G37/G2E/G32 /G4D/G45/G41/G53/G55/G52/G45/G4D/G45/G4E/G54/G20/G4F/G46/G20/G53/G4F/G55/G4E/G44
Sound intensity may be defined as the amount of energy flow per unit time through a unit area
perpendicular to the direction of energy flow. It is expressed as watts per square centimetre.However, the common receivers of sound are microphones, which do not measure sound intensity
directly. They are sensitive to sound pressure and therefore, it is more pertinent to measure sound
in terms of sound pressure which is given in dynes per square centimetre or in microbars (onemicrobar equals 1 dyne per cm
2). Sound pressure, for a given sinusoidal event, is related linearly
to both amplitude and frequency. Sound intensity is proportional to the square of sound pressure.
In most acoustic considerations what is of importance is not the absolute magnitude of the
intensity but its relative magnitude as compared to some assumed reference standard. The con-venient unit for making such comparisons and to express the sound intensity and sound pressuredata for all practical purposes, is the decibel (dB). The dB is 1/10 of a larger unit, the bell, namedafter Alexander Graham Bell. Decibel expresses the logarithm of the ratio between two soundintensities, powers or sound pressures. Since dB is merely a ratio, it is a dimensionless entity.
IfI
1and I2are two intensities in watts per square centimetre, then the number of decibels with
which they are related can be expressed as:
N= 10 log I1/I2
= 20 log P1/P2
where P1andP2are sound pressures in dynes per square centimetre. Since intensity is proportional
to the square of sound pressure, the constant 20 has been used in this case.
Attenuation is commonly expressed in negative dB numbers whereas amplification is given in
positive dB numbers.
Use of decibels as units for comparison of intensities help to avoid all mathematical calculations
except algebraic addition or subtraction of small numbers. The transmission efficiency of anymedium like air, a hearing aid or an amplifier is usually expressed in dB as a gain when the outputis greater and as a loss, if less. When two media are connected in series, the net efficiency is the sumof the two. For example, if a patient has a hearing loss of 70 dB and is fitted with a hearing aid witha gain of 60 dB, the net efficiency of the air-to-bone transmission system comes out to be -70 + 60 ora 10 dB loss.
Level (Volume): This is sound power measured in dB from a chosen zero level of power. The zero
level is usually 10
–16 W per cm2. This is the sound power (intensity) in air flowing past an area of
1 sq cm at right angles to the direction of the sound, at the rate of 10–16 W. This is approximately the
just audible power of a 1000 Hz pure tone for young healthy adults. The corresponding referencevalue in terms of sound pressure is 0.0002 dyne per sq cm. The smallness of both the sound
pressure and sound intensity values show that the human hearing mechanism is very sensitive.
/G31/G37/G2E/G32/G2E/G31 /G54/G72/G61/G6E/G73/G64/G75/G63/G65/G72/G73
In audiometry, generally employed transducers are (i) earphone, (ii) microphones, (iii) bone
vibrators, and (iv) loud speakers.
Earphones: They are usually of the moving coil type. They give a reasonably flat frequency response
up to 6 kHz after which their sensitivity falls rapidly. They are not specially designed for audiometric

468 Handbook of Biomedical Instrumentation
applications but for communication purposes. In their miniature form, they are used in hearing
aids. When used via insert earphone and ear moulds, they provide greater acoustic power to betransferred to the small volume of the external ear. It may be noted that audiometer earphones
are not interchangeable and must remain identified with a specific instrument to preserve its
calibration.
In conditions of ambient noise being too high for unshielded earphones, specially designed
audio cups are used. They use a fully-articulated suspension system which leaves the standardear caps free to locate against the pinnate with normal pressure, and at the same time to enclosefully the external ears with noise-excluding shells, sealed with soft plastic cushions, to excludebackground noise which will otherwise result in elevated threshold measurements.
Microphones: These are used to translate wave motion in air into electrical signal. Usual types are:
(i) carbon button which changes resistance with air pressure, (ii) electrodynamic where a voltageis induced in a coil by its motion relative to a magnet, (iii) condenser where capacitance of acondenser is varied by the vibration of one of the condenser plates. High quality condensermicrophones of diameters 12.5, 6.25 and 3.125 mm are currently used, depending on the frequencyto be measured. For special purposes, microphones can be fitted to the ear caps and used inreciprocal arrangement to transmit sound to the ear.
Bone Vibrators: In the early days, bone vibrators were composed of a rod connected to an
electromagnetically excited driving system. They were often held by hand against the mastoidprocess. Such units were large and cumbersome and are not in clinical use today. In the presentform, bone vibrators are of the hearing-aid type in which the transduction mechanism changes thealternating current into a vibratory force through a diaphragm. The diaphragm and its basicmechanical parameters like mass, compliance and resistance are important in establishing itsresponse characteristics. Though convenient, it is a very inefficient means of transduction and hasa rather limited and peaky frequency response. The plane circular contact area of a bone vibrator
is recommended to be 175 ± 25 mm
2. It is held in position by a headband.
Loudspeakers: They are used to deliver auditory stimuli, when it is not possible to have close
coupling of the transducer to the ear. Obviously, the acoustic energy loss into the surroundingswill be much greater than when stimulation is applied directly via an earphone. The acoustics of
the test room and masking of the non-test ear are two important factors which merit special
considerations while using the sound field of a loudspeaker.
/G20/G31/G37/G2E/G33 /G42/G41/G53/G49/G43/G20/G41/G55/G44/G49/G4F/G4D/G45/G54/G45/G52
An audiometer is a specialized equipment, which is used for the identification of hearing loss inindividuals, and the quantitative determination of the degree and nature of such a loss. It isessentially an oscillator driving a pair of headphones and is calibrated in terms of frequency andacoustic output. Both frequency and output are adjustable over the audio range. The instrument isalso provided with a calibrated noise source and bone-conductor vibrator.
Audiometers may be divided into two main groups on the basis of the type of stimulus they
provide to elicit auditory response: pure-tone audiometers and speech audiometers. A pure-tone
audiometer is used primarily to obtain air-conduction and bone-conduction thresholds of hearing.

Audiometers and Hearing Aids 469
These thresholds are helpful in the diagnosis of hearing loss. Pure-tone screening tests are employed
extensively in industrial and school hearing conservation programmes. Speech audiometers arenormally used to determine speech reception thresholds for diagnostic purposes and to assess
and evaluate the performance of hearing aids.
Screening audiometers are used to separate two groups of people. One that can hear as well as
or better than a particular standard and the other that cannot hear so well. Applications of these
instruments are found in industry, schools and military service.
An important application of audiometers is in industry. They help to assess the hearing function
of personnel at different stages of their detection of changes in auditory acuity, identify noisesusceptible persons and evaluate the effectiveness of ear protectors and noise control measures.
In conventional pure-tone audiometry, head phones are worn by the subject and a set of
responses is obtained for air-conducted sounds directed to each ear in turn. A bone conductorvibrator can then be attached to the head at the centre forehead position to see whether the hearingthreshold improves. If it does, then the disorder is most likely wholly or partly conductive inorigin. To avoid stimulation of the ear not under test with the vibrator, it can be temporarily madedeaf by introducing a suitable masking noise in the non-test ear via an earphone. A narrow-bandnoise centred on the pure-tone test frequency or a wide-band white noise is used for this purpose.The problem of how to recognize the need for masking and then applying the correct intensityposes a considerable difficulty.
/G31/G37/G2E/G33/G2E/G31 /G47/G65/G6E/G65/G72/G61/G6C/G20/G52/G65/G71/G75/G69/G72/G65/G6D/G65/G6E/G74/G73/G20/G6F/G66/G20/G41/G75/G64/G69/G6F/G6D/G65/G74/G65/G72/G73
Modern audiometers are solid-state instruments covering a frequency range from approximately100 to 10,000 Hz. Some instruments produce this range in discrete octave or semi-octave steps orintervals, while others provide for continuously variable frequency over their designed range. Thefrequency must remain sensibly constant at a value within 1–3% of the indicated value. Whereautomatic recording facilities include a continuous sweep frequency, the rate of change is normallykept as one octave per minute. If an automatic recording audiometer provides fixed frequencies,then a minimum period of 30 s must be allowed at each frequency.
The test frequencies should have sufficient purity of tone or approximation to the ideal sine
wave form to ensure response only to the desired fundamental frequency. The maximum harmonicdistortion in pure-tone air conduction audiometry is specified as 2% for the second and third
harmonic and much less at higher order harmonics. The total harmonic distortion should not be
more than 3%.
The intensity range of most audiometers starts from approximately 15 dB above normal to 95 dB
below normal over a frequency range from approximately 500 to 4000 Hz. The intensity range issomewhat less for frequencies below 500 Hz and above 4000 Hz. This is partly because of certaininstrumental limitations imposed by the earphone or vibrator and partly due to the desire to avoidthe threshold of feeling from stimulation at the lower frequency levels. The threshold of feeling isthe sensation of pain or tickle in the ear, which results from sound pressures and limits themaximal sound intensity that can be tolerated by the ear. The intensity level at which the thresholdof feeling is stimulated varies with frequency. For example, the threshold of feeling is stimulated atan intensity level approximately 120 dB above the normal threshold of audibility from about 500

470 Handbook of Biomedical Instrumentation
to 4000 Hz, but at 64 Hz the threshold of feeling is stimulated by sound pressures approximately
65 dB above the normal threshold value.
The attenuation dials on the audiometers provide variable intensity or volume controls. They
are calibrated in decibels usually in discrete steps, which differ by 5 dB in intensity from step tostep. Auditory acuity for each frequency is thus measured in dB above or below the normal hearing-
zero dB reference level for that frequency. This level is the minimal intensity at which each given
frequency can be perceived by the normal ear in a noise free environment and is experimentallydetermined by averaging the results of measurement on a large number of normal individualsbetween 18 and 25 years of age.
Audiometers usually have two channels with single pure-tone generators. The first channel
has pure-tone or speech output while the second channel has nominal masking. The pure-toneand speech can be switched to both channels for special tests. Channel two can have either wideor narrow-band masking. Each channel has an accurate independent attenuator output and thetransducers are switched to each attenuator as required.
In the recent years, numerous audiometric products incorporating microprocessors have been
introduced in the market. Such equipment offers greater convenience in calibration, test signalpresentation and versatility. Automated data collection and storage are also useful featuresincluded in such equipment. It may however be noted that the audiometric measurement principlesas described in the following text are not altered with the use of microprocessors and digitaltechnology.
It is extremely important in audiometry to ensure that only the testing signal reaches the ear.
Therefore, all testing must be done in a noise-free environment. Since environmental noise isdifficult to control, the noise-free conditions are achieved by performing the audiometric testing in
a sound isolating enclosure. Such enclosures help to attenuate all frequencies within the sensory
range below the threshold of hearing of normal ears. By using double wall construction andappropriate sound absorbing material, it is common to achieve 25 dB attenuation at 125 Hz and60 dB attenuation at frequencies between 1000 and 8000 Hz.
/G31/G37/G2E/G33/G2E/G32 /G4D/G61/G73/G6B/G69/G6E/G67/G20/G69/G6E/G20/G41/G75/G64/G69/G6F/G6D/G65/G74/G72/G79
In the presence of monaural and asymmetrical binaural hearing losses, there is serious difficultyin obtaining accurate measures of hearing for the poorer ear. The answer to the problem is to
eliminate responses from the better ear by masking in order, to shift the threshold to a higher level,
permitting greater intensities to be presented to the poorer ear without any danger of cross-over. Ifthe difference in air conduction acuity between the two ears is 50 dB or more, then it is advisableto place a masking noise over the better hearing ear while determining the threshold in the other.Masking efficiency depends upon the nature of masking sound as well as its intensity. A pure tonecan be used to mask other pure tones. However, over a range of test frequencies, masking efficiencyof a pure tone is low as compared to a noise composed of many frequencies, as usually providedin commercial audiometers.
Saw-tooth noise and white noise have been most commonly used for masking in clinical
audiometry, but narrow band noise, i.e. a restricted frequency bandwidth of white noise is alsooften used. Saw-tooth noise is a noise in which the basic repetition rate (fundamental frequency)

Audiometers and Hearing Aids 471
is usually that of the mains voltage and contains only those frequencies that are multiples of the
fundamental. The intensity of these multiples decreases as their frequencies increase. Noisesreferred to as ‘complex’ or square waves are similar in that they are composed of a fundamental
frequency and components that are multiples of it.
White noise is a noise containing all frequencies in the audible spectrum at approximately
equal intensities. However, the spectrum is limited at the ear by the frequency response of the
earphone, which may essentially be flat to 6000 Hz and may drop rapidly beyond. An excellentcomplex masking noise can be obtained by using the thermal or random electronic emission froma semiconductor diode, since it generates all frequencies simultaneously and with equal amplitudeover a frequency range wider than the response of the ear.
Narrow-band noise has been used by a number of investigators in audiometric studies. It is
produced by selectively filtering white noise. It has been found that narrow band noise is the mostefficient masking noise in pure-tone audiometry. The masking audiograms for normal hearingsubjects and the clinical results for hearing-impaired subjects show that for equal intensity levels,narrow band noise produces greater threshold shifts than do either of the other two types andthereby provides greater protection from false responses due to cross-over of the test tone.
/G20/G31/G37/G2E/G34 /G50/G55/G52/G45/G20/G54/G4F/G4E/G45/G20/G41/G55/G44/G49/G4F/G4D/G45/G54/G45/G52
A wave in air, which involves only one frequency of vibration, is known as pure-tone. Pure-tone
audiometry is used in routine tests and, therefore, it is the most widely used technique for
determining hearing loss. Pure-tone audiometers usually generate test tones in octave steps from
125 to 8000 Hz, the signal intensity ranging from –10 dB to +100 dB.
Pure-tone audiometry has several advantages, which makes it specifically suitable for making
threshold sensitivity measurements. A pure-tone is the simplest type of auditory stimulus. It can
be specified accurately in terms of frequency and intensity. These parameters can be controlled
with a high degree of precision. Speech audiometry normally allows measurements to bemade within the frequency range of 300–3000 Hz. Some patients may have impaired highfrequency response due to high intensity level occupational noise at 4000 or 6000 Hz. Pure-tone
measurements at these frequencies prove to be a more sensitive indicator of the effect of such noise
on the ear than speech tests. Changes in threshold sensitivity associated with various middle earsurgical procedures can be monitored more accurately with pure-tone than speech tests.
A pure-tone audiometer basically consists of an LC oscillator in which the inductance and
tuning capacitance are of close tolerances for having a precise control on the frequency of
oscillations. The oscillator is coupled to an output current amplifier stage to produce the required
power levels. The attenuators used in these instruments are of the ladder type, of nominal 10 W
impedance. The signals are presented acoustically to the ear by an earphone or small loudspeaker.
The available sound pressure levels in a typical audiometer are given in Table 17.1.
/G20/G31/G37/G2E/G35 /G53/G50/G45/G45/G43/G48/G20/G41/G55/G44/G49/G4F/G4D/G45/G54/G45/G52
Besides tonal audiometry, it is sometimes necessary to carry out tests with spoken voices. Thesetests are particularly important before prescribing hearing-aids and in determining the deterioration

472 Handbook of Biomedical Instrumentation
of speech understanding of patients. Specially designed speech audiometers are used for this
purpose. They incorporate a good quality tape recorder, which can play recorded speech. A doubleband tape recorder is preferred to interface the two channel audiometer units. Masking noise issupplied by the noise generator. The two channels supply the two head-phones or the two loudspeakers which are of 25 W each.
The tape recorder has a capacity for recording a limitless variety of test material and a
consistency of speech input, which cannot be obtained for live-voice audiometry in relation to test-retest repeatability. Another advantage of the tape recorded material is that the test words andsentences can be selected to cater for the widely differing needs of age, intelligence, dialect and
language.
In speech audiometers, live-voice facilities are incorporated primarily for communication
purposes as the inherent unreliability of live-voice speech tests may lead to serious errors. The
microphone amplifier used for this purpose is a simple two stage amplifier. The frequency responsecharacteristics of a live-voice channel should be such that with the microphone in a free soundfield having a constant sound pressure level, the sound pressure level developed by the earphoneof the audiometer in the artificial ear at frequencies in the range 250 to 4000 Hz does not differ fromthat at 1000 Hz by more than 110 dB. Also, it shall not rise at any frequency outside this band bymore than 15 dB, relative to the level at 1000 Hz.
/G20/G31/G37/G2E/G36 /G41/G55/G44/G49/G4F/G4D/G45/G54/G45/G52/G20/G53/G59/G53/G54/G45/G4D/G20/G42/G45/G4B/G45/G53/G59
George Van Bekesy, a Hungarian scientist, designed an automatic audiometric testing method forplotting the hearing threshold based on the patient’s signal. A principal feature of the method,differentiating it from conventional pure-tone audiometric techniques, is the interdependence of
the patient’s response and stimulus intensity: responses govern intensity and are affected by/G95/G20/G54/G61/G62/G6C/G65/G20/G31/G37/G2E/G31 Test Tones and Signal Intensity in Audiometers
Pure-tone Pure-tone Narrow band Narrow band
Frequency (head- (bone Balance masking (head- masking (bone
phones) conduction) channel phones) conduction)
125 70 – – – –
250 90 45 90 80 50500 110 60 110 90 60
1000 110 60 110 90 60
1500 110 60 110 90 602000 110 60 110 90 603000 110 60 110 90 50
4000 110 60 110 90 50
6000 90 – – 80 –
8000 90 – – 80 –
Speech 110 – 110 – –

Audiometers and Hearing Aids 473
changes they introduce in it. An audiogram traced by the Bekesy method represents the absolute
threshold values at all frequencies in the range tested. In addition, it shows the difference, indecibels, between levels at which the patient just hears a signal of increasing intensity and those
at which he just ceases to hear the signal when its intensity is decreasing. This latter characteristicoften varies significantly with the type of hearing impairment, and can aid in establishing the site
of lesion within the auditory system. On the basis of the audiograms, one can easily separate the
conduction and perceptive hearing deficiencies from each other.
Audiometers Bekesy are relatively simple for the patient to operate. The instrument generates a
pure-tone signal, which is presented to him through an air-conduction earphone. The subject istold to press a switch when the tone is heard and to release the switch when it is not heard. This
switch controls the motor-driven attenuator of the audiometer: when it is pressed, signal intensity
decreases and when it is released, signal intensity increases. A pen connected to the attenuatortraces a continuous record of the patient’s intensity adjustments on an audiogram chart,
producing a graphic representation of the subject’s threshold. The test signal may be presented in
a variety of ways, each suited to the investigation of a particular problem.
A block diagram of the audiometer system Bekesy is shown in Fig. 17.4. It consists basically of
an electrical section and a mechanical section. The electrical section includes an oscillator andmodulator circuits for the generation of the desired test signal, an automatic attenuator linked to
the writing system, control circuits for the drive motors of the mechanical section and a master
clock generator for the control of all timing functions via a logic control circuit. The carriage drive
and the writing system with their separate drive motors constitute the mechanical section.
/G31/G37/G2E/G36/G2E/G31 /G45/G6C/G65/G63/G74/G72/G69/G63/G61/G6C/G20/G53/G65/G63/G74/G69/G6F/G6E
Sine Wave Oscillator: This oscillator generates test signals with frequencies of 125 ,250, 500, 1000,
1500, 2000, 3000, 4000, 6000 and 8000 Hz. This sequence is first presented to the left earautomatically, each tone for 30 s, and then to the right ear, the shift between the frequencies beingnoiseless. After both ears have been tested, a 1 kHz tone is presented to the right ear to provide auseful indication of test reliability.
Modulator: From the oscillator the test signal is fed to the modulator, where the mode of operation
is selected by the ‘Tone’ switch, via the logic control circuit. Two models, ‘Pulse’ or ‘Cont’, are
available. In the ‘Pulse’ mode the test signal is modulated giving a signal, which is easilyrecognized by the patient. In the ‘Cont’ mode no modulation is applied, giving a signal suitable foruse, when calibrating the audiometer.
Automatic Attenuator: The signal from the modulator feeds the automatic attenuator situated on
the carriage together with the writing system. The attenuator consists of a logarithmic potentiometerwhich has its wiper attached to the pen drive so that the attenuation of the potentiometercorresponds to the position ( y-axis) of the pen on the audiogram chart. The potentiometer has
infinite resolution. The attenuation range is 100 dB, thereby covering the range of hearing levels
from -10 to +90 dB. When the test is initiated, the attenuator starts at its top position of -10 dB andthen increases the level with a rate of 5 dB/s. Also, when the test signal switches between the earsand when retesting at 1 kHz, the attenuator decreases the signal level to –10 dB to ensure that theright ear does not receive a tone at the elevated level possibly required at 8 kHz.

474 Handbook of Biomedical Instrumentation
Hearinglevel
calibration
Sine waveoscillatorModulatorBufferamplifierSeparateadjustmentof each testfrequencyEarphonesHand-switchRemotecontrolsocket
Carriagedrive
Audiogramchart
Pendrive
Carriage with writing system(X-Y recorder)
Thick-film
potentiometerFibre pen
Masterclock
generator
LLRL+L–L
CarriagePen
Limit switches“Pulse” “Cont.”ToneReturnForwardStopTestHoldResponsecounter
Front panel switches760272
01020
0500500Right
Left
NameNameName
NameNameDateDateDateDateTime
10001000
20002000
30003000
40004000
50005000
600060007000
102010
1020
2030
3040
4050
5060
6070
7080
8090
90Logic control circuit
Fig.17.4 Block diagram of the audiometer system Bekesy

Audiometers and Hearing Aids 475
Hand Switch: The pen drive is controlled via the logic control circuit by means of the hand-switch
operated by the patient. Pressing the switch decreases the output from the potentiometer andthereby the level in the earphones, while releasing the switch increases the output both ways with
a speed of 5 dB/s.
Buffer Amplifier and Calibration Circuit: From the attenuator the signal is fed via a buffer amplifier
to the hearing level calibration circuit. The buffer amplifier isolates the attenuator from thecalibration circuit in order not to affect its output. The calibration circuit consists of seven
potentiometers, one for each test frequency. During calibration, the potentiometers are adjusted
one at a time until the correct level, measured in a coupler, is obtained in the earphones.
Earphones: The earphones are a matched pair with distortion, typically less than 1%.
Master Clock Generator: A stable clock generator supplies the necessary signals for the control of
motor speed, attenuator speed, frequency shift, modulation and other timing functions. This makesthe system independent of variations in line voltage and frequency.
/G31/G37/G2E/G36/G2E/G32 /G4D/G65/G63/G68/G61/G6E/G69/G63/G61/G6C/G20/G53/G65/G63/G74/G69/G6F/G6E
Carriage: The carriage with the writing system is driven by a stepping motor via a toothed belt. The
speed and direction of rotation of the motor are automatically controlled via the logic controlsystem. When the test is initiated and the patient indicates that he hears the signal by pressing thehand switch, the carriage moves along the X-axis (frequency axis) of the audiogram in agreementwith the frequency of the test signal. When the frequency shifts, the carriage stops until the patientagain, by pressing the hand switch, indicates that he hears the signal. This avoids wastage of
recording space on the audiogram if a patient’s hearing threshold varies from frequency to
frequency or from left to right ear. When the complete test is finished the carriage and writingsystem return to the start position. To prevent carriage over-run, two limit switches are included inthe carriage drive circuit.
Writing System: The writing system is operated by the pen drive, which is driven by a stepping
motor. The pen drive moves the pen, and with it the wiper of the automatic attenuator, along theY-axis (hearing level axis) with a constant speed corresponding to the change in attenuation of 5dB/s. The direction of movement of the pen is determined by the position of the hand switchoperated by the patient. Limit switches are also included with the pen drive.
Audiogram Chart: The audiogram is printed in standard A5 format (148 ¥ 210 mm). The recording
space is large, 0.8 dB/mm, to enable easy reading. Space is provided on the audiogram side forregistration of information on the patient, audiometer, operator, etc. while the other side has spacefor recording the patient’s medical and occupational history. Four holes in the chart give preciseand automatic location of the audiogram on the chart bed.
In order to establish a more exact diagnosis applying adaptation and hearing fatigue tests,
several other tests besides the pure-tone Bekesy audiometry, have been suggested and can beperformed using the basic Bekesy system. For example, for carrying out the Fowler loudnessbalance test, a second channel is provided. The second channel has a continuously variableintensity over the range 0 to 110 dB and is calibrated in 1 dB increments.

476 Handbook of Biomedical Instrumentation
/G20/G31/G37/G2E/G37 /G45/G56/G4F/G4B/G45/G44/G20/G52/G45/G53/G50/G4F/G4E/G53/G45/G20/G41/G55/G44/G49/G4F/G4D/G45/G54/G52/G59/G20/G53/G59/G53/G54/G45/G4D
Evoked response audiometry has been the subject of research for several years. This work has
established evoked response electroencephalography resulting from an auditory stimulus abovethe hearing threshold. Instruments based on this principle have been found particularly suitable
for determining auditory threshold in the absence of voluntary response in subjects such as
infants, uncooperative adults, or animals.
The system basically comprises a conventional wide range pure-tone audiometer, which
operates under the control of an automatic programmer and provides a series of auditory stimulito the subject via either a loudspeaker or standard earphones. The EEG signal is picked up bystandard electrodes placed in contact with the subjects scalp. One electrode is usually placed onthe vertex, one at the post auricular area, and a third (ground) on the earlobe or forehead. Theinstrument stores and evaluates that part of the EEG signal, which follows each individualstimulus presentation. At the end of the programmed series of stimuli, it writes out on a paperchart a waveform that is the average response to stimuli. The presence of characteristic amplitudesand latencies in this waveform give an indication that the test intensity exceeded the subject’sthreshold at the test frequency. Similar trials at other intensity levels and other frequencies establishthe threshold contour.
Figure 17.5 shows a block diagram of the evoked response audiometry system. It consists of the
following five major subsystems.
The Tone Generator: It is a wide range pure-tone audiometer whose frequency output can be
selected at 250, 500, 750, 1000, 1500, 2000, 3000, 4000, 6000 and 8000 Hz. The output power levels
Stimulus
generator
–Light
–Speech
Audio
meterProgrammerEEG
amp.Signal
average
Viewer RecorderEvoked
response
waveformIntercomPre amp.IntercomTone input to
earphones or
speaker
EEG
signalTest tone
Fig.17.5 Block diagram of the evoked response audiometer

Audiometers and Hearing Aids 477
are adjustable from –5 to + 110 dB in 5 dB steps. In addition to the pure tones, internally generated
broad-band noise may be used as the stimulus. Provision is also made for external input fromother types of stimulus. A special feature of the generator is the selectable rise/decay times for
1–100 ms. Outputs are provided for the left ear, right ear, or both. A variable intensity masking
noise source is included in the generator. A power amplifier is incorporated to drive a speaker ortactile transducer. Total harmonic distortion should not exceed 2%.
EEG Amplifier: It is a conventional high gain, high impedance, low noise amplifier. The first stage
of the EEG amplifier is preferably kept in a separate “preamp head” located near the subject. Its
design and location minimize power line frequency pick-up. An ohm meter provided in thepreamp head enables the measurement of the electrode contact impedance and thus indicateswhen satisfactory contact resistance of the electrodes is obtained. The EEG preamplifier gain isfixed at 200. The overall sensitivity is adjustable in steps of 10 to 1000 mV/div on the chart recorder.
The amplifier also provides selectable roll-offs at the high and low ends of the spectrum of interestand a 60 Hz sharp notch filter. The low frequency roll-off points are at 0.15, 0.30.0.60, 1.5, 3, 6,10,15,30, 60 Hz at 6dB (half amplitude), whereas high frequency points are at 1.5, 3, 6, 10, 15, 30, 60,100 Hz at 6 dB (half amplitude) yielding a 12 dB/octave roll-off.
The Programmer: A logic device that controls the system operation in correct time sequence. It
helps to have a selectable rate of stimulus presentation, stores the number of pulses that theoperator chooses to constitute a run, starts the recorder at the beginning of the run, turns theaudiometer tone generator on and off to provide the auditory stimuli at the proper time. It alsospeeds up the chart drive for the detailed signal samples, stops the recorder after providing forpaper clearance, erases the signal averaging computer and clears and resets itself for the next run.
Total count in the programmer is selectable from 1–109 stimuli with a selector switch. The pulse
interval is normally kept as 0.1, 0.2, 0.5, 1, 2, 5, 10 and pulse duration as 1–10,000 ms.
Signal Averaging Computer: This separates evoked responses from the normal EEG activity by
ignoring those components, which are not synchronized with stimuli. Because the waveform ofthe evoked potential will be essentially the same every time in response to the tone presentation,and the other electrical activity will vary randomly, the evoked response “grows” in the computermemory and the noise component tends to average to zero with repeated presentations. It may beprovided with either 50 or 100 averaging points depending upon the degree of resolution required.The computer includes a provision for selection of integrating time constant (5, 10, 20, 50, 100, 200,
500 s) and sweep duration or analysis time (0.1, 0.2.0.5, 1, 2, and 5 s). A delay circuit is incorporated
to select a delay between the onset of the stimulus and the start of analysis by the signal averager.The delay time is selectable as 0, 0.2, 0.5 and 1 s. The amplitude of the evoked response may benormalized to that of the on-going EEG monitoring signal using a gain control (gain variable from20 to 50). The computer provides outputs for display either on a conventional oscilloscope or onan X-Y plotter.
Chart R ecorder: It is a two channel recorder. One of the two channels is used to display the
averaged response after it has been processed by the computer and the other displays unprocessedEEG. There are two event markers, one of which is activated by each gating pulse from theprogrammer to show the beginning and duration of each stimulus and the other is available forregistering any mark at the desired instant. The chart can be driven at four different speeds (1, 5,

478 Handbook of Biomedical Instrumentation
25, 125 mm/s), which are automatically switched by the programmer. Translucent chart paper is
usually employed so that records may be compared by overlaying one on another on the
illuminated opal-glass viewer.
Evoked response audiometer systems also contain a provision for ‘External’ mode of operation
where any other type of stimulus generator can be connected into the system and controlled from
the programmer. This could include narrow band noise, speech, a high frequency auditory signal
for animal research or even a photic or tactile stimulator.
Modern evoked response audiometers are built around microcomputers. In these instruments,
the stimulators, preamplifiers and amplifiers are all digitally controlled via the central processingunit, which automatically avoids undesirable parameters. They have no push button or dials asthe parameters are varied by means of the keyboard. The parameters can be controlled to a verywide range, which would not be possible with conventional knobs and switches. Stimuli from 12to 16 kHz are included to facilitate investigation into high frequency hearing loss and ototoxicdrug effects. Texts and waveforms are displayed on a large size TV screen. A built-in chart recorderhelps to make recordings under the control of the keyboard.
/G20/G31/G37/G2E/G38 /G43/G41/G4C/G49/G42/G52/G41/G54/G49/G4F/G4E/G20/G4F/G46/G20/G41/G55/G44/G49/G4F/G4D/G45/G54/G45/G52/G53
The purpose of audiometric testing is to compare. The comparison may be between an individual’spresent audiogram and one taken previously, or between his audiogram and the audiograms of
others. Whatever the reason for recording an audiogram, clearly the test conditions and the
reference levels for the measurements must be identical for valid comparisons to be made betweenthem. Furthermore, for valid comparisons with audiograms from other instruments, the testconditions and reference levels must be universal. Various standard institutions have publishedaudiometric standard threshold values for several widely used combinations of earphones andartificial ears for carrying out audiometer calibrations.
Accurate calibration of audiometers is essential to ensure that the instrument produces a pure
tone at the specified level and frequency, and that the signal is present only in the transducer towhich it is directed. For pure tone and speech audiometers, the parameters, which are commonlychecked, are frequency and intensity.
The frequency output from the audiometer is checked using either a counter-timer or an
oscilloscope. A counter-timer is preferable because it gives a quick, direct reading. It is connectedacross one earphone, and the audiometer is set to give maximum output. The specifications foraudiometers allow a tolerance of ± 3% for frequency output from a fixed frequency pure-toneaudiometer.
Sound pressure levels are best checked by an ‘artificial ear’ or coupler together with a sound
level meter. The artificial ear consists of a condenser microphone and a 6 ml coupler. The coupleroriginally designed by the National Bureau of Standards, USA encloses a volume approximately
the same as the volume under a earphone for a human ear. The earphone is placed squarely on the
coupler with a 500 g weight. It should be ensured that it forms a good seal. The output on thesound level meter is read directly in dB and compared with the expected output per frequency. Itmay, however, be noted that volume displacement is only one component of acoustic impedance

Audiometers and Hearing Aids 479
and, therefore, it is recognized to be a rather crude acoustic approximation of the average human
ear. This is why the reference levels vary with the earphone type. The disadvantage of differentreference levels is overcome with a more elaborate design of the wide-band artificial ear. A typical
example of an artificial ear is that of type 4153 Artificial Ear from Bruel and Kjaer. It is designed in
accordance with IEC R 318. It has a three cavity coupler and provides acoustic impedance, whichclosely resembles that of the human ear. Different types of couplers are available for use withartificial ears for measurements on headphones, insert type hearing aids and other earphones.
While the artificial ear is used as a standardized substitute for the human ear when making
measurements using an air-conduction earphone, an artificial mastoid is used as a standardizedsubstitute for the human mastoid. The artificial mastoid consists of a mechanical simulation of thehuman one, incorporating a built-in force transducer to monitor the output of the device to becalibrated. The artificial mastoid must meet standard specifications such as the American NationalStandard ANSI S3 13-1972 or the IEC R 373.
Audiometric tests should be conducted in rooms which are reasonably quiet. A subject under
test should not be disturbed either by sounds created within the room or by those intruding fromthe outside. Such rooms are known as sound-treated rooms. In order to keep out extraneousnoises, the outside walls of such rooms consist of heavy hard-surfaced shell. The inside is linedwith absorptive material to keep reverberation low and to prevent practically any reflection.
/G20/G31/G37/G2E/G39 /G48/G45/G41/G52/G49/G4E/G47/G20/G41/G49/G44/G53
Hearing loss has many forms. The most common is related to the body aging process and to long-term cumulative exposure of the ear to sound energy. As one grows older, it becomes more difficultto hear. The ear becomes less sensitive to sound, less precise as a sound analyzer and less effectiveas a speech processor. Loss of hearing differs greatly in different individuals. Changes in the earoccur gradually over time. However, by the time the changes are manifested, it is estimated thatapproximately 30 to 50 percent or more of the sensory cells in the inner ear have suffered irreparablestructural damage or are missing (Engebretson, 1994). Under these conditions, the only choiceavailable for hearing-impaired individuals is to wear a hearing aid.
Hearing impairment is caused by either loss in sensitivity (loss in perceived loudness), or loss
in the ability to discriminate different speech sounds or both. Loss of loudness may be due to eitherincreased mechanical impedance between the outer ear and the inner ear or by the reducedsensitivity of the sensory organ of hearing. Loss of the discrimination ability is basically associatedwith damage to the sensory organ, although, other neural structures at higher levels may also be
involved.
The modern hearing aid became possible with the invention of the transistor, which has enabled
to develop small, power-efficient amplifier circuits that could be packed in a form that fits behind
or in the ear. Even though the primary function of an hearing aid is to compensate for the loss ofsensitivity of the impaired ear, in practice, it is not this simple. The ear behaves differently for softsounds near the hearing threshold than it does for loud sounds. Therefore, a frequency responsethat restores normal hearing thresholds for soft sounds will not, in general be appropriate forlouder sounds. Furthermore, even when speech sounds are made audible for the hearing-impaired

480 Handbook of Biomedical Instrumentation
listener, it does not follow that he/she will be able to understand speech. Hearing-impaired
listeners experience more difficulty in understanding speech in background noise than normal-hearing listeners.
/G31/G37/G2E/G39/G2E/G31 /G43/G6F/G6E/G76/G65/G6E/G74/G69/G6F/G6E/G61/G6C/G20/G48/G65/G61/G72/G69/G6E/G67/G20/G41/G69/G64
Modern hearing aids have evolved from single-transistor amplifiers to modern multi-channeldesigns containing hundreds and even thousands of transistors. A typical design is shown inFig. 17.6. The basic functional parts include a microphone and associated preamplifier, anautomatic gain control circuit (AGC), a set of active filters, a mixer and power amplifier, an outputtransducer or receiver. The total circuitry works on a battery. The use of multiple channels in thisdesign provides different compression characteristics for different frequency ranges. Typically,
the crossover frequencies of the channels and the compression characteristics can be adjusted
with potentiometers. Most of the latest hearing aids are electronically programmable. Theprogrammable parameters are downloaded from a computer-based system and stored in digitalregisters. The register outputs are used to switch resistor networks that control various analogcircuitry. The active filters are adjusted to generally provide for low-frequency attenuation of up to30–40 dB relative to the high-frequency response. This is because most hearing aid wearers requirehigh frequency gain.
MICAMP
AGCLPF
COMP
HPF
COMPMixer
AMPREC
Digital registers
Audiometer/programmerHearing aid
Fig.17.6 Conventional analog type hearing aid
The transducer in a hearing aid, which is a microphone, can be realized in an integrated form
with a field-effect transistor preamplifier (Fig. 17.7). The preamplifier is housed in the metallic,microphone case to shield its input from extraneous noise. On the other hand, the receiver is anelectromagnetic device, which drives a miniature diaphragm to produce acoustic output. Theacoustic output is routed to the ear-mould through a flexible tubing whose frequency response canbe altered to boost the high-frequency response. This is done by tapering its inside diameter fromthe ear mould back to the receiver port end.

Audiometers and Hearing Aids 481
All the electronics circuitry is packaged in a housing, which can be designed for fitting to the
ear in any one of the following ways:
1. Placing all the components in a pocket-sized enclosure or box which is connected to the
output transducer worn in the ear. The box can be carried in the shirt pocket or carriedwith a belt around the waist. With the availability of miniature-sized aids, this approachis no longer employed.
2. The components are packaged in a curved module, which is designed to fit comfortably
behind the ear.
3. The most popular design is in which the total package can be put inside the outer ear.
Much depends on the performance of the filters for further reduction of the size and improve-
ment in the working of the hearing aids. The dynamic range of an operational amplifier, which isthe basic building block of an electronic filter, decreases as the three halves power of feature size.
Since dynamic range of analog hearing aids is already marginally acceptable, it appears that
further reduction in size to achieve increased processing complexity is not practical. The potentialfor greater dynamic range, with less power consumption and greater complexity in hearing aiddesign is feasible only with digital processing technologies.
/G31/G37/G2E/G39/G2E/G32 /G44/G69/G67/G69/G74/G61/G6C/G20/G48/G65/G61/G72/G69/G6E/G67/G20/G41/G69/G64
A typical digital hearing aid is illustrated in Fig.17.8. The major parts are the microphone,an analog-to-digital converter (ADC), the digital signal processor (DSP), the digital-to-analog
converter (DAC), the receiver and a two port memory. Essentially, sound waves picked up by the
microphone and transformed into electrical signals are converted into digital form by an A-Dconverter. A typical microphone will have an internal noise of 20 dB SPL (sound pressure level)when referred to the input and maximum undistorted output corresponding to a signal of about90 dB SPL. Allowing some margin for peak performance, the total dynamic range required of theADC is 80 dB. This requirement can be achieved with a 14 bit A–D converter.
The DSP is a fixed (wired-program) digital processing device containing an array of adders,
multipliers and registers which provide the fundamental operations necessary for implementingFET pre-
amplifier+++++++++++++++++++++++++–––––––––––––––––––––––––Acoustic input
Electret
Back plateMetallized plastic
diaphragm
Support
Ground
Power
inputSignal
output
Fig.17.7 Integrated microphone and FET preamplifier (Electret microphone)