Lithium Charger1 [620309]

1

TECHNICAL UNIVERSITY OF CLUJ -NAPOCA

LITHIUM -ION CHARGER BASED ON BUCK CONVETER

Abstract: A procedure to charge a single cell lithium -ion battery using a Buck converter topology. This aims
to develop a strategy for the implementation by using external voltage control loops and charging current, thus
enabling to maintain the battery in the parameters specified and reduce the risks. Also, the frequency
compensation is carried out, keeping the circuit in the stability zone. The usefulness of this work comes in the
context of which lithium -ion batteries currently represent one of the most important renewable energy sources
on the market.

I. Introduction:

The first lithium -based batteries
appeared before 1970, they had the ano de
made of metallic lithium, were
unchargeable and unstable. The problem
resolved ten years later when lithium -ion-
based batteries appeared with the anode
made of graphite and the cathode from
active materials. Lithium is a very
lightweight metal with an electrochemical
potential and a high energy density on the
table which has made it possible to carry out
batteries whose properties are perfectly
folded on the current requirements of the
electronic market.
The cell consists of:

• A positive electrode (anode);
• A negative electrode (cathode)
• A separator
• electrolyte

Figure 1. Cell structure The positive electrode is usually formed
from transition metals, which have the property
to increase energy density, reduce the cost of
manufacturing, enhance safety, being most
common in batteries related to electric cars.
One of the most used materials at the present
time is LiCoO2. The negative electrode is
made of carbon, the separator is a very thin,
microperforated plastic separating the two
electrodes, and the electrolyte is a material with
which it allows very high conductivity of the
ions and only of them, being made of Salts,
solvents and additives. In terms of
functionality when the battery is loaded, the
cathode breaks out of li thium ions that are
transferred by electrolyte to the anode. When
the battery is discharged, the ions move
backwards by producing energy. Lithium -ion
batteries represent at the current stage one of
the most important energy storage devices on
the market. T heir use is included in a wide
range of devices such as laptops, telephones
and electric cars. It is expected that electric and
hybrid cars will represent 50 -60% of the total
existing cars before 2050.[1]
Currently, battery Engineering focuses on
solvin g the main deficits they face by trying to
reduce the problems caused by the rapid
degradation of batteries in order to be able to
extend the life and implicitly obtain

2

performance in both financial and ecological
terms. The phenomenon of accumulations of
batteries depends directly on the composition
of the electrodes which causes the electrolyte
dissolution and the modification of the battery
properties. The way the electrodes are made is
thus the main challenge of engineers. [2]
Another big challenge of these batteries is
the way they are charged and the
implementation of solutions as efficient as
possible to reduce the time of charging without
accelerating the ageing process. It is
understood that the circuits specialised in the
charging of batteries are found in great
diversity, given that the batteries are created for
a multitude of applications, they have different
properties depending on the necessary needs,
so different charger topologies are
individualised by each other both by way of
implementation and by their ability to fully
satisfy the requirements of certain batteries.
The current work aims to implement a
specialised circuit for the loading of lithium –
ion batteries using a Buck -converter topology
in order to identify the yield of this method and
fully understand its functionality. This
deployment solution can be categorized into
two large categories depending on the
switching device chosen. We can speak of such
charger based on synchronous and
asynchronous topology. Asynchronous
topology is di stinguished by simplicity of
implementation and reduction of
manufacturing costs. For a circuit that does not
work at large voltages the control circuit of the
transistor can be made simple, it is often
carried out of a regulator with a weaker
precision ne edles uses an external capacitor to
remove the current peaks occur during
switching. The disadvantage of this method is
due to the large power that dissipate on the
switching device which limits the current of the
battery charge to the values above to four
amps. The synchronous topology appears as a
solution for the large power dissipated by the
charger in asynchronous configuration. This
entails a higher cost because additional
components are needed in controlling the
switching element. In this configurati on a
transistor NMOS is used to achieve the lower
switching stage, and the upper transistor can be both NMOS and PMOS. A logical gateway is
added to the command circuit level to ensure
that never transistors will lead simultaneously.
Usually a dead time is introduced to ensure that
the transistors will not be in conduction at the
same time.
II. Implementation method

1. Buck converter design

Figure 2. Buck converter implemented in
Psim

The scheme of a stabilizer is relatively
simple, the alternative voltage taken from the
supply is transformed by the rectifying circuit
and properly filtered until the level of the
pulsed pulse is ensured for the other
components the continuous voltage applies.
The switching element usually represented by
a power transistor. The transistor is controlled
by a high frequency signal that can reach
hundreds of Hz. The rectangular -shaped
voltage resulting from the switching transistor
is applied to the power transformer and in the
secondary is obtained the same vo ltage, but
with the amplitude determined by the
transformation ratio, which is then recovered
and filtered becoming the output voltage of the
converter. The protection circuit ensures
overcurrent and overvoltage protection and
ensuring the correct operatio n.
If the input voltage changes its value
during the operation, the output of the
converter will vary depending on the variation
in the input, undesirable effect that can cause
the circuit to malfunction. To prevent the effect
of the network variation the voltage from the
output is taken and compared with a prescribed

3

value. The command and control circuit make
the comparison and generates a correction
signal for adjusting the output.[3]
For the sizing of the component elements
of the Buck converter it was considered that the
battery with Lithium -ion NCR 186500B, with
a saturation voltage of 4.2 V, a loading current
1.625 A and the output voltage ripple of 30 mV
maximum. Nominal input voltage of the
converter has been chosen 12 V, and the
converter will wor k at switching frequency of
100 khz. The diode will have to withstand a
greater sensitivity than the maximum input
voltage and which has a voltage breakdown in
the conduction of 0.6 V. The Schottky MBRD
360 diode with Vdiode_max = 60 V, UD = 0.6
V, RD = 20 mΩ is to be chosen. The power
dissipated on the diode will be 0.83 W. The
selection of the Capacitor imposes conditions
to meet the requirements related to the ripple,
the output resistance of the capacitor must
produce less than 75% of the total ripple. After
computing we need to use a minimum
capacitance of 22 μF and a maximum ESR of
0.16 Ω. We will choose an tantalum capacitor
B45197A 68 μF, ESR = 0.1 Ω. The converter
was implemented in PSIM.

For the control and frequency
compensation we have chosen th e voltage
regulation mode and as stated in Equation 18
the transfer function of the low pass filter from
the structure of the converter contains two
complex conjugated poles that will introduces
a gain attenuation feature with -40dB/decade,
as well as a -180 ° attenuation of the phase.
Besides the effect of the poles is added the
effect generated by the internal resistance of
the condenser that will introduce a zero, which
will generate an increase in the gain feature by
20 db/decade and will increase the p hase by +
90 °. The purpose of frequency compensation
feature is to obtain a phase margin greater than
45 ° to maintain the circuit in the stability zone
and achieve a gain equivalent to -20dB/decade
needed to increase the phase edge and reduce
the risk of having a positive gain at high
frequencies. The compensation chosen will be the type
3 implemented with the error amplifier. The
clearing of the complex conjugate poles and
the zeroes will follow the following sizing
strategy:
• A pole will be inserted in the origin for
a gain of -20 db/decade.
• Two zeroes will be inserted at the filter
cutting frequency.
• A pole will be inserted at the frequency
of the ESR.
• A third pole will be inserted at half the
switching frequency.
One of the most important aspects that we
need to take attention in the design is the
frequency where the gain of the control loop is
unitary, or the frequency to 0 dB. To have a
stable circuit it is very important that the phase
margine measured at this frequency is grea ter
than 45 °. The frequency setting to 0 dB was
achieved by controlling the gain of the control
loop, in this scope was dimensioned resistor
RC1 knowing the Rf1 and the gain of the Open
loop converter. The capacitor capacity value
will be determined CC1 k nowing that the first
zero frequency will be equal to the filter's
cutting frequency. The resulting values were
represented in the figure 3:

Figure 3. Resulting value of the compensator

2. Charging loop

When battery is charging the current
supplied by converter must be maintained at a
constant value specified by battery datasheet.

4

To achieve that we need a current loop to
regulate the output current. When the battery
reaches the saturation voltage the charging
current automatically decrease and the voltage
across at its terminal must be maintained stable
until the saturation is complete with an external
voltage loop who regulate the battery voltage.
This two loops are separate by two diode.

Figure 4. Current and voltage loops

3. Practical i mplementation

For the implementation of the adjusting
loops I chose the MAX745 regulator. It can
control the charging of a maximum of four
batteries of several types, tensions and
reference current can be controlled externally.
The reference current setti ng will be achieved
through the SETI pin, using an internal
reference of 4.2 V or by using an external
reference. With the help of CS and BATT pins,
the current on the Rsense resistance is
measured, thus being converted into tension,
amplified by 6 and the n compared to the
tension on the SETI pin. Such control shall be
carried out on the current loop. In the case of
this integrated the adjusting and control circuit
is implemented with transconductor.
In order to be able to monitor the current,
a series coil resistance of 0.1 Ω is used. To set
the charging current to 1,625, choose a control
voltage on the SETI pin of 3.68 V. The voltage
is obtained by using a resistance of 100 and
13.5 kΩ. Because the control circuit operates in
the current mode the output fi lter changes from
an RLC filter of order two with complex poly conjugates to a single -pole 1 RC Filter and thus
the compensation is made easier. By applying
on the VADJ pin of a voltage equivalent to half
of the reference voltage, the battery charge
voltag e is set to 4.2 V, so two resistors of 100
kΩ will be used to divide the tension and the
compensation is carried out on the CCV pin by
introducing a pole -zero pair. Selecting the
number of cells is carried out by order of
CELL0 and CELL1 pins.
For the ch oice of components we will take
into account that the control circuit works at a
frequency of 300 kHz which makes it possible
to reduce the dimensions of the converter
components. Using equation 36 A minimum
value of 8 mH is obtained for the coil used. We
will choose a coil of 22 mH. From Equation 40
we obtain a minimum value of 4 mF for the
capacitor and 0120 Ω for ESR. We will thus
choose a tantalum capacitor B45197A 68 mF,
ESR = 0.1 Ω.
The same transistor IRF7303 will be used
because it can support a m aximum current of
7.5 A and a voltage of 30 V. To achieve
protection when reversing the current polarity
we use the same Schottky diode MBRD360G.
In order to be able to follow the load current
value, the current sensor HV7800 that retrieves
the voltage fro m the current detection resistor
and transfer it to the output, and in order to be
able to obtain the battery charge voltage a
repeater stage is realised with AD711
operational.

Figure 5. Max 745 connexion diagram

III. Experimental results

5

In this project Buck converter works in
CCM mode which is proved by inductor
current who never reach zero.

Figure 6. Circuit currents

To maintain a stable voltage at battery
saturation we need a small output ripple. As we
can see the voltage output ripple is s maller than
50 mV.

Figure 7. Output voltage ripple

The stability of the circuit is showed
below:

Figure 8a. Bode plot: Gain

Figure 8b. Bode plot: Phase
As we can see the margin phase of the
circuit is above 60 ° and the stability is
acquired. The PWM of the practical
implementation using MAX745 is showed
below.

Figure 9a. Upper transistor command voltage

Figure 9b. Lower transistor command voltage

Because we use a synchronous configuration of
Buck converter we can see that the transist or
work in antiparallel configuration.

IV. Conclusion

In this project we successfully manage to
implement a lithium -ion charger using a DC –
DC converter with good performance and
results and finally we can affirm that DC -DC
converters are one of the best sui table ways to
implement a charger for one of the most used
source of energy.

V. References

[1](.t.B.U.website,„https://batteryuniversity.com/lear
n/article/charging_lithium_ion_batteries”.
[2]M. Brain, “How Lithium -ion Batteries Work",” 14
November 2006.
[3]D. Petreus, Electronica surselor de alimentare,
Cluj Napoca: Mediamira Cuj Napoca, 2002.

6

7