M190447EE-IEEE Conference Paper
M190447EE-IEEE Conference Paper
M190447EE-IEEE Conference Paper
FLYBACK CONVERTER
Chelimela Dinesh Dr. Mukti Barai
MTech Power Electronics Associate professor
dept. of Electrical Engineering dept. of Electrical Engineering
National Institute of Technology National Institute of Technology
Calicut Calicut
Kozhikode, India Kozhikode, India
[email protected] [email protected]
Abstract— Flyback converter is simple in construction with states and power conversion can be accomplished with
a smaller number of components like MOSFET, Mutual minimal power loss, which equates to high efficiency.
inductor, capacitor, diode and low cost. Hence it is popularly Flyback Converter is the Buck-Boost derived and isolated
used in the low power applications like cell mobiles, laptop
converter. Flyback converter is the most popular used
batteries, CRT, etc. Flyback converter has isolation therefore,
multiple output is also possible. So, it popularly used in
converter for low power applications <150w because it is
multivariable outputs. Now-a-days research on keeping the the converter which has less number of components and no
constant voltage cell, which makes battery constant voltage. In inductor at the output side which reduces the size of the
this converter is operated in the buck mode with the converter and also the cost of the converter. Flyback
Continuous conduction mode and the output specification of converter has an isolation from the input side and output
12V, 24W.The design calculation for the given specification is side which increases the safety of the load and supply. Due
done in the open loop. The Mutual inductor is designed with to the presence of the isolation multiple output is possible.
ferrite core for switching frequency of 20KHZ which decreases In the Electrical vehicle the battery cells have to kept
the size of the inductor with increase of the switching
constant for constant battery voltage without any leakage.
frequency. As, the system in the open loop condition becoming
unstable which is shown from the bode plot of the plant. So, to This constant can be made using the flyback converter to
make the system stable compensator is designed. Compensator each cell because size, cost and multiple output for low
design requires plant transfer function which has been derived power application is possible using it.Flyback converters are
by applying the small signal model and made system stable. mostly used in the cell phone battery, laptop battery, CRT
The opamp realization of the compensator is derived. Closed tubes, televisions ,etc..
loop flyback converter with the compensator design is In Isolated DC-DC Converters Flyback Converter is the
simulated in Ltspice software. Output voltage, output current simple converter to design since no Inductor is required at
and transformer parameters waveforms are shown. the load side and it has only a high switching frequency
Keywords—component, flyback, transfer function, Ferrite Core transformer, switch, diode and a capacitor.
inductor, OpAmp. Designing the mutual Inductance of the transformer is
different when compared to the other isolated converters
I. INTRODUCTION since it stores the energy when the switch is ON and the
stored energy is transferred to the secondary when the
The regulated power supply technology can be divided into
switch is OFF. Here the transformer is working as a coupled
two distinct technologies. Firstly, the linear or the series
inductor but not as a transformer. Hence the Flyback
regulator, secondly the switched mode conversion
transformer is also called as the Coupled inductor. Hence the
technique. The linear regulator behaves as a variable
magnetizing Inductance value is found to be very high in
resistance between the input and the output as it provides the
Flyback converter.
precise output voltage. But the major drawback of this linear
regulator is its efficiency which is very less when compared
to the switched mode conversion technique because output
of the linear regulator is taken out from the resistors due to
which efficiency reduced and only one output is possible
that to precise output. Due to resistive circuit heat
dissipation is more so heat sink is required which increases
the size of the circuit. So, in application where space and
efficiency is crucial linear regulators are not used. Though
they have desirable characteristics, such as simplicity, low
output ripple, good load and line regulation, fast response
time to load or line changes and low EMI. Fig 1 Circuit diagram
A switched mode power supply is a power supply that uses
II. PRINICPAL OF OPERATION
low loss components such as capacitors, inductors,
transformers and the use of switches that are in one of the Flyback converter is operates in two modes. Mode1 is when
two possible states ON or OFF. The advantage is that the the switch is on, Mode 2 is when the switch is off [1].
switch dissipates very little power in either of these two Mode1: When the switch is on the positive polarity of the
supply is connected to the dot terminal of transformer so
non-dotted terminal becomes the negative. The same is
reflected to the secondary hence the diode in the secondary
becomes the reverse bias and secondary winding do not
conducts. During period the load is supplied by the capacitor
which is initially charged. As the current is flowing into the
primary winding, primary inductor gets storing the energy in
it. The circuit diagram is show below fig.
During this mode when the switch is OFF, the current III. DESIGNING OF FLYBACK CONVERTER:
flowing through the switch when the switch was ON will
now have no path so it flows through the switch capacitance In the Flyback converter the switch rating is depend upon
for a small duration of time. During this interval the primary the turns ratio(n) of transformer. As the n value is increased
current will charges the switch capacitance and the voltage the voltage stress on the switch increases so the rating of the
across the switch will become more than the input voltage. switch has to be increased which increases the cost of the
When the input voltage is less than the switch capacitor switch and also the heat losses. S to avoid this we select the
voltage, the capacitor will discharge through the clamped switch rating such that minimum rating of the switch is
circuit described later in the design specification. The maximum input voltage. Then accordingly we design the n
voltage clamp circuit consist of an ultrafast diode and also a value. So in designing process the first step is designing the
parallel connection of a capacitor and resistor. When the n value. Hence mosfet rating is taken as 200v/18A [2].
switch capacitor voltage is more than the input voltage the Relation between input voltage and output voltage:
polarity of the voltage across the primary winding of the Hence the ripple current becomes
transformer will be of opposite polarity and during this time I pmax−I pmin=(V ¿ ¿ ¿ d T s)/(L p)¿
with this polarity secondary voltage will be induced which
makes the diode to be forward biased condition and makes n ( V 0 +V r ) ( 1−d ) T s
I pmax−I pmin=
the diode to conduct when the switch is turned off. Lp ¿
¿
Lp 2
Where =n
Ls
d
V ❑0= (V −V r )
( 1−d ) n ¿
Calculation of turns ratio of transformer(n):
The role of the turns ratio (n) is in controlling the voltage 2 2
T s∗V inmax∗d min
reflected into the primary when the secondary diode is L pmin =
conducting. This reflected voltage along with the Vin will 2∗P 0 min
decide the voltage appearing across the switch. Another role the value of the magnetizing inductance to get 30% ripple in
of n is to calculate the secondary side current level. A large primary side is calculated as shown below.
value of n will result in high voltage stress in switch and
lower voltage stress across the diode, but high current stress I pmax−I pmin=(V ¿ ¿ ¿ d T s)/(L p)¿
in diode. Usually n is chosen such that n𝑉0 = Vin in certain
designs. However this will lead to the requirement that Primary and secondary currents:
MOSFET should have 2.3 Vin max voltage stress on it and Primary and secondary currents are found by solving the
to leave 20% margin then its voltage requirement must be equations of
2.3/(0.8) Vin max which is approximately equal to 3 Vin n ( V 0 +V r ) ( 1−d ) T s
max. this will lead to 1200V MOSFET requirement if Vin is I pmax−I pmin=
obtained by straight rectification of 230V AC 1200 V Lp
MOSFET’s are though available are costly and have large 2∗I 0
on state resistance. Further their on state resistance varies I pmax + I pmin=
significantly with temperature. A better strategy will be to n(1−d)
choose a MOSFET and then adjust n. But primary conducts the most under Vin min condition
hence mosfet
V inmax +n ( V 0 +V r ) +0.3 V inmax =0.8∗( voltage rating of chosen by using) dmax
Calculations of Duty Ratio(d):
2∗I 0 max
I +I =
pmax pmin
n(V 0 +V r) n(1−d max ❑)
d max = n ( V 0 +V r ) ( 1−d m ) T s
n ( V 0 +V r ) +V inmin
I pmax−I pmin=
n(V 0+V r) L pmin
d min = RATING OF THE MOSFET AND DIODE:
n ( V 0 +V r ) + V inmax MOSFET:
Primary Inductance of Transformer: Peak Current = 1.3A Average current =0.7A RMS current
Designing the primary inductance of the flyback transformer =0.813A
is different when compared to the other isolated converters
Peak voltage = V inmax +n ( V 0 +V r )=151 V
since it stores the energy when the switch is ON and delivers
the energy to the secondary when the switch is in OFF Hence the switch rating is selected as Vds =200V and
condition. For this reason the value of magnetizing Current is 18A
inductance value is high. Diode:
Minimum Primary inductance of the transformer has to be Peak Current = 7.4A Average current =2A RMS current
designed such that at maximum ripple in the inductor with =3.2A
the minimum loading condition the converter has to be in V¿
Peak voltage =V 0 + =20.57 V
CCM mode. So, Maximum ripple occurs with the dmax and n
corresponding input voltage is Vinmin, but in the below Fast recovery diodes are used to reduce the recovery time
equation 5 We don’t know the Lpmin to get Lpmin that has and hence to ensure quick flux reset during the off time of
to make the CCM it is known by using the dmin value and the switch. So, MUR860 is selected with V = 600 V and I =
corresponding dmin value input voltage is vinmax. Hence 8 A.
we get the Lpmin and maximum ripple in the circuit. If
primary inductance below the Lpmin then the circuit
becomes the DCM. IV. CLOSED LOOP DESIGN AND SIMULATION
n ( V 0 +V r ) ( 1−d ) T s Closed loop control is required to make system stable output
I pmax−I pmin= voltage with the variations in the load or variations in the
Lp input voltage. The closed loop control requires the feedback
For CCM mode I pmin>0 with P0 min hence at the boundary of from the output voltage, reference voltage, compensator or
CCM mode I pmin=0 hence we get above eqn. controller part, and pwm modulator. The heart of the closed
loop control is compensator design which compares the
feedback output voltage and reference voltage. By
n ( V 0 +V r ) ( 1−d ) T s comparing, if any error is present then generates the error.
I pmax=
Lp This error will be passed through the pwm modulator which
During the switch ON period the primary inductor stores the modulates or modifies the modulating signal. In the pwm
energy, modulator pulses are generated by comparing modulating
signal (dc signal) with the carrier signal (triangular signal).
2 When the carrier signal is higher than the dc signal then
f s∗L p∗I pmax pulses are obtained. So, when error from the compensator
P0 min ¿
2 varies then the modulating signal varies accordingly hence
Where I pmax is known from equation 1 by making pulse width varies till error from the compensator is zero. In
I pmin=0 and substituting we get Lpmin as this way closed loop control is achieved.
Compensator design requires the transfer function of the
plant (converter). This transfer function of the plant is
[]
derived by applying the small signal analysis to the plant. d ip
This analysis involves the non-ideality conditions of the
plant. Small signal analysis is performed using state space dt
=
modelling technique which requires the differential dVc
equations. These differential equations are written in the
dt
[ ][ ] [
state space model and applying small perturbation to the
parameters. These equations are linearized as time invariant −( r l 1 +r s )
][ ]
system and then transfer function of the plant is derived. 0 1 V¿
Bode plot is plotted to the obtained transfer function and Lp ip 0 0
check for stability and bandwidth of the system. Based on + Lp V ❑r
the bode plot compensator transfer function is designed to
( r l 1+ r s ) V c 0 0 0 i0
0
increases the stability and bandwidth of the plant. Using the c ( R+r c )
[ ]
compensator transfer function opamp realization is
V¿
[ ][ ]
performed. Then the obtained opamp circuit is connected to R ip
the plant. Then the obtained circuit will be closed loop [ V 0 ]= 0 + [ 0 0 0 ] V ❑r
circuit.[3] R+r c V c
i0
Small Signal Modelling and State Space Averaging in
CCM:[4] Mode-2: When the switch –OFF condition:
Consider the inductor current (i p ) and capacitor voltage (V c
) as state variables for the state space eqautions.
2
d i p −n ( rl❑2 +r d +r c ) i p−nV r +n r c i 0−nV c
=
dt Lp
d V c n i p i0
= −
dt c c
V 0=ic r c +V c
V 0= ( n i p −i 0 ) r c +V c
Fig.5 Equivalent circuit for mode-1 operation
The state space matrix representation for mode-2 is shown
L p∗d i p
−V ¿ + ( r l1 +r s ) i p + =0----------- (4.1) below
[ ][ ][ ] [ ][
dt 2
On simplifying above equation we get,
d ip −n ( rl❑2 +r d + r c ) −n −n −nr c
0 V
d i p V ¿ ( r l 1+ r s ) i p dt Lp L p ip Lp Lp
= + V
= − dVc n Vc −1
dt L p Lp 0 0 0
d V c −i 0 dt c c
[ ]
= V¿
[ ]
dt c ip
V 0=−i 0 r c +V c [ V 0 ] =[ n r c 1 ] Vc
+ [ 0 0 −r c ] V ❑r
The state space matrix representation for mode-1 condition i0
is shown below
Apply the state space averaging model to mode 1 and 2[5].
A=( A¿ ¿1∗d )+( A 2∗(1−d ))¿
B=(B ¿¿ 1∗d )+(B2∗(1−d))¿
C=(C¿¿ 1∗d)+(C 2∗(1−d ))¿
D=(D¿¿ 1∗d)+(D2∗(1−d ))¿ The compensator function K(s) used must be in the form of
k
k ( s ) = x K 1 (s) (where K is a positive gain value) since
Now apply perturbation for the plant i.e. replacing D with s
the +d ᶺ . Then the resultant matrix is integral control action is a must.
Type I compensator is a simple integrator with gain.
[]
ᶺ Type II has one pole and one zero in addition to integrator
d iL with gain.
dt Type III has two coincident poles and two coincident zeros
ᶺ in addition to integrator with gain.
dVc k
dt Type 1 :k ( s )=
s
[ ][ ]
2
−( r s+ r l 1 ) D−−n ( r l 2 +r d + r c ) ( 1−D ) −n ( 1−D ) k (1+ s τ z )
Type 2 :k ( s )=
Lp Lp iᶺL s(1+ s τ p )
¿ 2
−n ( 1−D ) −D V ᶺc k (1+s τ z)
Type 3 :k ( s )= 2
c c ( r c+ R ) s(1+ s τ p)
The specifications required for the design of compensator
+¿ are
[ ][ ]
1) Cross over frequency- Consider the cross frequency
DR+ ( 1−D ) ( r c + R ) i ᶺL
[ V 0 ]= n r c (1−D )
rc + R
as less than 0.1 times of switching frequency=2khz.
V ᶺc 2) Phase Margin – Consider Phase Margin as 45deg.
[]
The compensator has to be designed such that the loop gain
ᶺ magnitude at the specified cross over frequency (fco) will be
d
[ ]
1 and the angle of loop gain at that frequency will be (360o
rc V ᶺ¿ – PM) where PM is the phase margin specified. Loop gain in
+ −n r c I L −( ) 0 0 −r c (1−D) ᶺ
r c+ R Vr the closed loop feedback system is ( s )=−G ( s ) K (s) . The
i ᶺ0 requirement stated above translates to
1
|K ( j wc 0 )|=
The above mathematical equations are written in the |G( j wc 0)|
MATLAB software and bode plot is obtained. Phase Delay Angle
The transfer function of the plant obtained from the
MATLAB code is
|K ( jw c 0 )|=180 − phase Delay Angle G ( j wc 0 ) −PM
0
−1 0
The Transfer function of the type-3 compensator in analog
Lead Angle K 1 ( j wco ) =4 tan k −180 circuit form
[ 1
]
0 −1
Phase Delay Angle of K 1 ( j wco )=180 −4 tan k −
This must be equal to [ V o (s) R1 ( C 1+C 2 )
0
90 − phase Delay Angle G ( j wc 0 )−PM ¿ = ∗(1+ s R2 C 2)¿ ¿
V ¿ (s) s
Therefore, Where,
[ ]
−1 0
4 tan k =90 + PhaseDelay Angle of G ( j w c0 ) + PM 1
K= ; τ =R2 C2= ( R1+ R3 ) C3 ; τ p =R 2 (C ¿ ¿ 1 ⊕C
0
90 + PhaseDelay Angle of G ( j wc 0 ) + PM R 1 ( C 1+ C 2 ) z
k =tan ( )
4
¿
2∗π∗f co
K= 2
k ∗|G( j wc 0)| TABLE I
k
τ z= PARAMETERS VALUES
w co
rl1(Primary inductance 0.01Ω
1 resistance)
τ p=
k∗wco rl2 (Secondary inductance 0.01Ω
Therefore, the transfer function of the compensator is given resistance)
by Rc(Capacitance Resistance) 0.036Ω
( )
36.744 1+s∗(1.054∗10 )
−3 2 R(load resistance) 6Ω
K ( s )= ∗ n(Turns ratio of transformer) 7
s 1+ s(9.43∗10−5 ) D(Duty ratio) 0.626
C(capacitance) 1200uF
The Bode Plot obtained with combination of plant Lpmin(Primary Inductance 2.72mH
(converter) and type-3 compensator minimum to be ccm)
From the above obtained plot the stability parameters like Lp 5.21mH
Gain Margin and Phase Margin are positive with finite Rs (Source resistance) 0.18Ω
frequency. Hence, the system becomes stable.
Rd(Diode resistance) 0.01Ω
Gain Margin = 18.5 @ 1.33*104 rad/sec and Phase
3
k 3.343
Margin = 34.5deg @ 3.61*10 rad/sec K 36.744
Further, compensator has to be designed in analog domain. Tz 1.054/1000sec
For type III compensator, the analog circuit is shown in Tp 9.43/100000sec
below fig. It consists of three capacitors and three inductors. R1 247.862kΩ
R2 10.54kΩ
R3 24.366kΩ
C1 9.8nF
C2 0.1uF
C3 3.87nF
Results:
REFERENCES
[1] L. Umanand, “Lecture on Fly-Back Type Switched
Mode Power Supply” NPTEL Module-3, Lesson 22
[2] Abraham I. Pressman; Keith, “Switching Power
Supply Design” Textbook- McGraw Hill, Third
Edition.
[3] L. Umanand, “Modelling of DC-DC Converters”
NPTEL Module-8, lecture-19.
[4] L. Umanand, “State Space Modelling” NPTEL
Module-8, lecture-20 and 21.
[5] L. Umanand, “Circuit Averaging” NPTEL Module-
8, lecture-22 and 23.
[6] K.S. Suresh, “Modelling Control of DC-DC
Converters” lecture notes.