A highefficiency method is proposed to suppress magnetic core imbalance in phaseshifted fullbridge (PSFB) converters. Compared with conventional solutions, such as controlling peak current mode (PCM) or adding DC blocking capacitance, the proposed method has several advantages, such as lower power loss and smaller size, because the additional current sensor or blocking capacitor is removed. A time domain model of the secondary side is built to analyze the relationship between transformer core imbalance and cathode voltage of secondary side rectifiers. An approximate control algorithm is designed to achieve asymmetric phase control, which reduces the effects of imbalance. A 60 V/15 A prototype is built to verify the proposed method. Experimental results show that the numerical difference of primary side peak currents between two adjacent cycles is suppressed from 2 A to approximately 0 A. Meanwhile, compared with the PCM solution, the efficiency of the PSFB converter is slightly improved from 93% to 93.2%.
I. INTRODUCTION
Phaseshifted fullbridge (PSFB) converter is extensively used in mediumpower and highpower applications because of its advantages of high efficiency and low electric stress on power devices at the same power level
[1]

[4]
. Operating in the first and third quadrants of the
B–H
curve is an outstanding characteristic of PSFB converters, but easily leads to transformer core imbalance. Transformer core imbalance is normally observed as the unilateral saturation of transformers or excessive electric stress on power devices
[5]

[7]
.
Transformer core imbalance in the PSFB converter is a result of static current offset on the primary side. Current offset is caused by unavoidable unequal onstate resistance of different diagonal power switches, delay of driver circuits, and nonideal PCB parameters
[8]
. The effect of this imbalance can be reduced by simply adding a DC blocking capacitor to the primary winding of the transformer in engineering applications. However, the blocking capacitor method cannot completely eliminate transformer core imbalance. Moreover, this method has several disadvantages, such as grievous ratio loss, high cost, and low efficiency. In recent years, several new methods are proposed to eliminate the imbalance issue.
A compensation circuit is proposed to reduce transformer core imbalance in an asymmetrically regulated PSFB converter
[9]
. However, the method is difficult to apply because many external devices are introduced. Moreover, an uncommon asymmetric fullbridge controller that separately controls the phases of diagonal power switches is necessary.
An accurate model and a digital compensator are proposed in
[10]
. The model is built to analyze the relationship between nonideal parameters and offset current on the primary side. However, the current sensor in the proposed solution leads to extra cost and power loss, which limit the application range.
Given the disadvantages of low efficiency and high material cost in common transformer core imbalance solutions, a high efficiency and simple method is proposed in this study.
In this study, a transient model is built to analyze the relationship between the imbalance of the magnetic core and cathode voltage of output rectifiers on the secondary side. An approximate control algorithm that eliminates the effect of transformer core imbalance is achieved by a lowcost microcontroller. This paper is organized as follows: In Section II, a conventional PSFB converter is analyzed and simulated to investigate the relationship among the cathode voltage of output rectifiers on the secondary side, primary side current, and magnetic core imbalance. A digital compensator is proposed in Section III, including an external sample circuit and an internal control algorithm. In Section IV, a 60 V/15 A prototype is built and tested to verify the effect of the proposed method compared with conventional PSFB converters.
II. THEORETICAL ANALYSIS OF TRANSFORMER CORE IMBALANCE
 A. Transformer Core Imbalance in PSFB Converters
A conventional PSFB converter regulated based on a microcontroller is shown in
Figure 1
. The primary side includes input capacitor
C
_{IN}
and power MOSFETs
M1–M4
.
L
_{R}
represents the sum of the resonant inductor and leakage inductance of the transformer;
N
_{P}
represents the primary winding of the transformer; and
N
_{1}
and
N
_{2}
represent the secondary windings of the transformer. The secondary side consists of output rectifiers
D1–D2
, filter inductor
L
_{OUT}
, and filter capacitor
C
_{OUT}
.
R
represents the load. Feedback and control components are composed of sample resistors
R
_{VS1}
–
R
_{VS2}
and a microcontroller. The internal ADC module samples the feedback voltage, which is divided by
R
_{VS1}
–
R
_{VS2}
from output voltage
V
_{OUT}
. The function of the logic module is calculating the phase of diagonal transistors. The internal timers output four pulse signals, Driver1–Driver4, based on the logic module. Driver1–Driver4 control
V
_{GS}
of power MOSFETs when amplified by the isolated circuits DriverA–DriverD.
Basic PSFB converter.
R
_{DS(on)}
of
M1–M4
, related parasitic parameters of PCB, and delay time caused by isolated drivers cannot be completely equal because of design and manufacturing limitations. The inequality causes a deviation between positive and negative peak currents on the primary side, which leads to inevitable imbalance of the transformer core in the PSFB converter.
Transformer core imbalance is directly decided by the DC offset of the primary current and is affected by several other phenomena. For example, unequal peak values of the primary current also illustrate the imbalance condition of the transformer core, which has been utilized by peak current mode (PCM) control. A current sensor is required to sample the primary current, which increases the material cost and power loss of the PSFB converter.
 B. Transformer Core Imbalance Reflexed by a Voltage Signal
Aside from the peak values of the primary current, another phenomenon also causes transformer core imbalance, that is, voltage oscillation ring of the anode and cathode of rectifiers
D1
and
D2
on the secondary side, as shown in
Figure 1
. Even if
RC
snubbers are paralleled to
D1–D2
, voltage peaks can be detected on the anode and cathode voltages of
D1
during overlapping of
M1
and
M3
or on the anode and cathode voltages of
D2
during overlapping of
M2
and
M4
, as illustrated in
Figure 2
. Given that the cathode of rectifiers is directly connected,
V
_{K}
is simulated with primary current
I
_{P}
instead of voltage of anodes in
Figure 2
.
Simulated waveform of V_{K} and I_{P}.
In
Figure 2
, the peak values of
I
_{P}
and
V
_{K}
are measured; the range of
V
_{K}
is adjusted to show inequality. Unequal peak values of
I
_{P}
and
V
_{K}
in a switching cycle are illustrated, which indicate that the voltage signal includes information of the imbalance condition of the magnetic core. Moreover, the relationship between oscillation ring
V
_{K(D2)}
(
V
_{D2}
) and primary current
i
_{p}
(
t
) in the PSFB converter without
RC
snubbers is verified, which proves that
V
_{K}
is infected by
I
_{P}
.
Confirming whether the relationship remains correct when
RC
snubbers are added is necessary.
RC
snubbers are extensively used to prevent overvoltage stress on
D1
and
D2
. The relationship between
V
_{K}
and
I
_{P}
in PSFB converter with
RC
snubbers should be investigated, such that analysis based on
V
_{K}
can be proven effective.
We let
t
_{1}
denote the finish time of the ratio loss procedure, which transfers no energy from the primary side to the secondary side.
Figure 3
illustrates the schematic and equivalent circuit of the PSFB converter with
RC
snubbers during the energy transfer procedure from the primary side to the secondary side (defined as
t
_{1}
–
t
_{2}
). During
t
_{1}
–
t
_{2}
,
M1
and
M3
are on, and the onstate resistances are
R
_{DS1}
and
R
_{DS3}
. Inductors
L
_{R}
and
L
_{P}
are charged, and the primary side current
i
_{P}
(
t
) can be expressed as follows:
Equivalent circuit to reflex anode/cathode voltage of secondary side rectifier.
Secondary current has two parts (
Figure 3
(a)). The green line denotes that current flows through
L
_{OUT}
to provide output power.
L
_{OUT}
is sufficiently large that
L
_{OUT}
,
C
_{OUT}
, and R can be replaced by constant current source
I
_{LO}
during
t
_{1}
–
t
_{2}
. Meanwhile, the red line denotes that current flows through junction capacitance
C
_{j}
of
D1
and
RC
snubber, which causes resonance of
v
_{K}
(
t
) when
i
_{P}
(
t
) increases to its peak value. If
L
_{R}
is converted from primary side to secondary side, then the circuit can be simplified based on the operation principle of the transformer (
Figure 3
(b)).
By converting inductance
L
_{R}
and voltage across
L
_{P}
to the secondary side, the turns ratio can be expressed as
K
= (
L
_{P}
/
L
_{S1}
)
^{0.5}
= (
L
_{P}
/
L
_{S2}
)
^{0.5}
,
L
_{R}
becomes
L
_{RS}
=
K
^{2}
L
_{R}
/4, and voltage across the secondary windings is
V
_{S}
= 2
V
_{IN}
/
K
. By transferring the circuit from the time domain to the complex frequency domain, as shown in
Figures 3
(b) and
3
(c), the voltage across
C
_{j}
can be calculated as follows:

(a)Current loop duringt1–t2

(b)Equivalent circuit in the time domain duringt1–t2

(c)Equivalent circuit in the complex frequency domain duringt1–t2
In Equation (2),
I
_{P}
(0) represents the primary current at
t
_{1}
. Moreover, if the forward drop of the rectifier is neglected, then
v
_{K}
(
s
) in the complex frequency domain can be expressed as follows:
v
_{K}
(
s
) is a threeorder function of variable s according to Equation (3). The initial value of the primary current infects the voltage, which is simulated in the time domain by
Matlab
(
Figure 4
). Two values of
I
_{P}
(0) are compared in the same coordinate system, which shows that a higher
I
_{P}
(0) is equivalent to a higher
V
_{K}
. A comparison between the simulation results and tested results proves that the shape of the simulated waveform is similar to the tested waveform when the DC value is neglected. This comparison proves that
v
_{K}
(
t
) reflexes
i
_{P}
(
t
), particularly with the difference between the two directions during one cycle.
Simulated time domain waveform of VK compared with the tested waveform.
According to
Figure 4
, the theory that
V
_{K}
is related to
I
_{P}
is proven, which indicates that the difference in the peak values of
V
_{K}
in one cycle reflexes the balance condition of the transformer core in the PSFB converter. Magnetization processes in the first and third quadrants of the
B–H
curve are asymmetric; thus, the offset value of the primary side current can be observed, which should be zero in a balanced condition.
Given that the peak value of
V
_{K}
is directly affected by the primary side current,
V
_{K}
can be used to acquire imbalance state information of the transformer core. If the transformer core is balanced, then the peak value
V
_{KPEAK(D1)}
should be equal to peak value
V
_{KPEAK(D2)}
and
D1
and
D2
represent different rectifiers in different phases. According to the simulations, tests, and theoretical analysis described in this section, the difference between
V
_{KPEAK(D1)}
and
V
_{KPEAK(D2)}
affects the imbalance of the transformer core. The conclusion represents the theoretical foundation of the proposed suppression method.
III. DESIGN OF A DIGITAL PHASE COMPENSATOR
 A. Control Algorithm and Program Loop Design
Based on the analysis presented in Section II, a new method that samples
V
_{K}
as another feedback variable can be proposed. Transformer core imbalance is currently suppressed accurately cyclebycycle using PCM. However, in practical applications, serious consequences, such as unilateral transformer saturation or power device failure, are caused by a continuous singledirection imbalance. A current offset in one singleswitching cycle causes no harm. Thus, the main idea of suppression in this study is to limit the imbalance range and swing imbalance direction.
If the static offset value of primary side current
I
_{P_OS}
is controlled to alternate between positive and negative, then the average value of
I
_{P_OS}
would be approximately zero during most periods:
In Equation (4),
n
represents the amount of switching cycles.
Based on the equation, regulating
I
_{P_OS}
with an approximate method is considered to be effective. When the sign of
I
_{P_OS}
changes, the magnetic offset of the transformer can also be controlled to alternate between positive and negative, and the amplitude of imbalance in one switching cycle can be suppressed as well. Although transformer core imbalance occurs during the entire operation procedure, the imbalance is bidirectional and significantly limited; thus, the average value of magnetic flux
is also approximately zero in most switching cycles.
According to the analysis, an asymmetric phase regulation method should be applied to maintain a low
I
_{P_OS}
in each cycle. Moreover, the positive or negative sign of
I
_{P}
cannot be maintained for a long period of time. Thus, the phase difference between
M1–M3
and
M2–M4
should be regulated based on the imbalance state of the transformer core, which is represented by the numerical relationship between the maximum value of
V
_{KPEAK(D1)}
and
V
_{KPEAK(D2)}
in the proposed solution (
Figure 5
).
Control diagram of the proposed solution.
Compared with the PSFB converter with a conventional controller, an external voltage sampler that is used to sense cathode voltage
V
_{K}
from output rectifiers is inserted to the control loop (
Figure 5
). Moreover, a few extra algorithm modules are added to the program cycle of the microcontroller in the proposed solution.
In a normal switching cycle, output voltage
V
_{OUT}
of the power topology is sampled by output sampler
R
_{VS1}
–
R
_{VS2}
(
Figure 1
). The sampled voltage
V
_{FB}
turns into the error signal
V
_{E}
after being amplified.
V
_{E}
is the input of the phase modulator, which is used to calculate phase
D
. In a conventional PSFB converter, the phases of
M1–M3
and
M2–M4
output by the microcontroller are equal to
D
. However, the phases of power MOSFETs are not the same as previously mentioned. This issue is one of the main factors of transformer core imbalance; thus, imbalance problems can be solved by deliberate unequal phases of
M1–M3
and
M2–M4
.
Figure 6
shows the procedure of the phase controller. To achieve asymmetric regulation in the proposed solution, the peak values of voltage signal
V
_{K}
are captured by the internal ADC of the microcontroller and then converted to digital variables
voltageKnode1
and
voltageKnode2
in the sample function. The difference between
voltageKnode1
and
voltageKnode2
reflects the imbalance state of the transformer core in the current cycle. Extra transformer imbalance modulator outputs time variable
t
_{D}
, which is the other input of the logic unit, except for phase
D
. The function of the logic unit is to generate phases
D
_{13(n)}
and
D
_{24(n)}
based on the current phases
D
_{13(n − 1)}
and
D
_{24(n − 1)}
, calculated phase
D
, and time variable
t
_{D}
.
D
_{13(n)}
and
D
_{24(n)}
would be assigned to related registers of internal timers to modify the phases of
M1–M3
and
M2–M4
in the subsequent cycle.
Procedure of the phase controller.
Compared with the conventional digital phase controller of the PSFB converter, the proposed digital phase controller inserts different phase calculation functions to the main program loop.
The driver signals of
M1
and
M2
are generated by one of the advanced timers in the microcontroller. These two signals would never be regulated, which indicates that the period, ratio, and dead time of
M1
and
M2
remain constant. The driver signals of
M3
and
M4
are generated by another advanced timer in the microcontroller. They are regulated in program interruptions based on the values of
voltageFB
,
voltageKnode1
, and
voltageKnode2
. Phase
D
of the subsequent cycle is set with an initial value, which is calculated based on the phases and sampled
voltageFB
in the current cycle. In the current cycle,
voltageKnode1
and
voltageKnode2
are also sampled and compared with the obtained magnetic offset condition.
The numerical value of
t
_{D}
is set to be constant to simplify the control loop; however, the sign of
t
_{D}
would be positive or negative according to the relationship between
voltageKnode1
and
voltageKnode2
. The phases of
M1–M3
and
M2–M4
in cycle
n
are calculated as follows:
As Equation (5) shows, the phases in the subsequent cycle are related to the phases in the last cycle, the calculated phase in the current cycle, and the imbalance state in the current cycle. If the direction of transformer core imbalance is reflected as
V
_{KPEAK(D1)}
>
V
_{KPEAK(D2)}
, then
I
_{PEAK(M1–M3)}
>
I
_{PEAK(M2–M4)}
in the current cycle
T
_{n}
, and
t
_{D}
would be negative to decrease
D
_{13}
and increase
D
_{24}
in the subsequent cycle
T
_{n}
_{+1}
. If the direction of transformer core imbalance in
T
_{n}
_{+1}
retains the state of
T
_{n}
, then
t
_{D}
would again have a negative value and
D
_{13}
in
T
_{n}
_{+2}
becomes smaller than
D
_{13}
in
T
_{n}
_{+1}
.
Moreover, the sign of
t_{D}
would not change until condition
V
_{KPEAK(D1)}
>
V
_{KPEAK(D2)}
becomes
V
_{KPEAK(D1)}
<
V
_{KPEAK(D2)}
. When
V
_{KPEAK(D1)}
<
V
_{KPEAK(D2)}
is built in one cycle, the phase controller would change the sign of
t
_{D}
to build
V
_{KPEAK(D1)}
>
V
_{KPEAK(D2)}
. Thus, a limited transformer core imbalance is achieved, which no longer causes unilateral saturation of the transformer and power device failures.
 B. Design of the RC Snubber of Output Rectifiers
The
RC
snubber is used to decrease voltage stress across the rectifier
[11]
. The purpose of designing the RC snubber is based on power loss and stress margin. Quality factor
Q
is defined as the reflex oscillation in the
RLC
circuit:
In PSFB applications, the oscillation ring should be suppressed to an underdamping state to make our method effective while ensuring that voltage stress is safe for the rectifiers. The value of
R
_{1}
can be determined as follows:
In Equation (7),
C
_{j}
can be derived from the datasheet and
L
_{i}
can be calculated as follows:
C
_{1}
is selected in the range of:
Parameter
I
_{P}
is the initial value of the primary side current, and
T
_{S}
represents the period of the PSFB converter. Based on these design considerations, the RC snubber ensures a single ring while the amplitude of the ring is suppressed.
 C. Special design of the voltage sensor to sample transformer core imbalance
Program achievement of asymmetric regulation has been proposed; however, the ADC sampler needs special consideration because of difficulty in sampling the peak value of
V
_{K}
.
To sample a reliable
V
_{K}
, a digital signal process function is designed in the microcontroller to capture voltage at a specified time. Sampler design has two rules, as follows:
1) According to the errors of the sampler, several different points should be sampled to filter spikes.
2) Several points unavoidably lead to error of the desired point; thus, the start and end times of the sampler should be set properly based on the relationship between PWM driver and
V
_{K}
waveform.
Based on the rules, the special design of the sampler should be considered. Period
T
_{S}
of the PSFB converter is set by the designer and executed by the MCU, and phase
t
_{P}
of two diagonal MOSFETs is decided by the MCU. Moreover, duty cycle loss time can be calculated as follows:
I
_{peak}
is the peak value of the primary side current,
L
_{R}
is the resonant inductance,
L
_{leak}
is the leakage inductance of the transformer, and
V
_{in}
is the input voltage.
According to these settings and calculations, the specific time of the oscillation ring peak is easy to locate; however, sampling the exact value during software execution is difficult particularly because no external sample/hold circuit is added.
Thus, sample consideration in the software should be based on multiple samples; in this case, average value of sampled points is used to reflex true
V
_{K}
. The proposed mechanism is to sample several values right after
t
_{loss}
and
T
_{S}
/2 +
t
_{loss}
, as shown in
Fig. 7
. A higher peak value of
V
_{K}
indicates a higher sampled value at the corresponding time during the first oscillation ring, such that the accurate average value can be calculated if the sample time is ensured to be sufficiently accurate. Then, these values are averaged to be
V
_{K}
in the half period, as illustrated in
Fig. 8
.
Waveform of VGS and primary side current.
Digital sample design consideration.
When data
V
_{K}
[1, …,
n
] are stored as variables in each phase, an extra process is necessary to generate a reliable variable, which is called the calculating function of
t
_{D}
:
The average method is the easiest way to process data if point number
x
in Equation (11) is set as 2
^{M}
, such as 4, 8, and 16. The point number should be set according to the cost of the sampler because a large amount of data in the sampling module can easily cause poor loop response. Finally, the number of
M
is set as 3, such that the average value of
V
_{K}
can be rapidly calculated.
Based on these considerations, an asymmetric phase controller is developed. Two external resistors are connected between the cathode of the output rectifiers and output ground to divide the large voltage. The divided voltage is one of the analog inputs of the microcontroller. The designed program generates four channels of PWM signal with asymmetric phases to control power MOSFETs. The test results are given in the subsequent section to show the effect of the proposed method.
IV. EXPERIMENTAL RESULTS
A 220 VAC input, 60 V/15 A output prototype is built to verify the effect of the proposed method (
Fig. 9
). The PSFB converter consists of two PCBs vertically connected via plugin parts. The PSFB topology board includes basic PSFB converter, isolated drivers of four power MOSFETs, and power supply for chips, fan, and driver circuits. The control board is made up of voltage followers, a 48 MHz microcontroller (STM32F051xx from STMicroelectronics), and driver ICs. The main parameters that correspond to the circuit in
Figure 1
are shown in
Table I
.
Photograph of the 60 V/15 A prototype.
PARAMETERS OF THE PSFB CONVERTER
PARAMETERS OF THE PSFB CONVERTER
In a regular PSFB converter without asymmetric control under 310 VDC input and 59 V/15 A output condition, waveforms of the primary side current and cathode voltage of output rectifiers are tested, as shown in
Fig. 10
(a).
Test waveforms of VK and IP (test condition: 310 V input, 60 V/15 A output): khaki, VK (100 V/div); brown, abs(IP) (10 A/div); green, IP (10 A/div); time, 10 μs/div: (a) Basic PSFB converter without asymmetric control, (b) SFB converter with the proposed control method.
Given that the peak value of the primary side current in waveforms has alternating signs, the green line is difficult to detect. Thus, a square–square root mathematical (absolute value,
abs
) calculation is executed to shift the negative to positive, shown as the brown line in
Fig. 10
(a). The difference between two adjacent cycles is measured as 2 A according to the red auxiliary lines, which indicates that magnetic offset is sufficiently significant to affect the reliability of power devices.
Fig. 10
(b) shows the experimental results when the proposed method is utilized by a conventional PSFB converter. Waveforms are tested under the condition of 310 V input and 59 V/15 A output. The test results are equal to the prototype under symmetrical control technique and higher than the converter with blocking capacitance.
In
Fig. 10
(b), the primary side current is also processed to a nonnegative range, shown as a brown line. A red auxiliary line is drawn in
Fig. 10
(b), where the peak value of
abs
(
I
_{P}
) is approximately equal to the adjacent cycle, which indicates that the imbalance of the magnetic core is effectively suppressed.
In addition, other indices should be tested to verify the proposed method. Efficiency under different load currents is tested.
Fig. 11
shows that efficiency is higher in the proposed method than that in the PCM in the load range of 30–100%. The peak value of efficiency in the proposed method is 93.2%, which is 0.2% higher than that in the PCM solution.
Efficiency comparison of PCM and the proposed method (test condition: 310 V input, 60 V/1–15 A output).
Moreover, the suppression effect of the two methods is compared to prove the effectiveness of our method. Notably, the proposed method does not reduce the imbalance of the transformer core in a single period. The purpose of the proposed method is to regulate the direction of the imbalance, such that the average peak current on the primary side is zero for a long period of time.
Table 2
shows the comparison between PCM and the proposed method. The experimental results prove that the effect of our method is not as good as that in PCM, although the proposed method achieves the objective of imbalance suppression for the entire operation procedure.
SUPPRESSION EFFECT COMPARISON BETWEEN PCM AND THE PROPOSED METHOD
SUPPRESSION EFFECT COMPARISON BETWEEN PCM AND THE PROPOSED METHOD
According to these experimental results, the proposed method is proven to be effective in suppressing the imbalance of the transformer core, and its efficiency is slightly higher than that in the PCM solution.
Light to heavy load variation is also tested; the waveforms of
V
_{K}
and primary side current are shown in
Fig. 12
. Transient cost is approximately 30 ms. The time cost is reasonable given the large output capacitance.
Load variation from 6 A to 15 A (test condition: 310 V input, 60 V/6–15 A output).
According to the lowclock frequency of the selected microcontroller, regulation accuracy is insufficient to reduce the imbalance of the transformer core completely; however, the proposed method can maintain a relatively limited imbalance state. The imbalance direction is controlled to invert every switching cycle in steady state, such that the range of imbalance is limited. The possibility of power device failure is reduced in the PSFB converter because of the limited range. Furthermore, only several external passive devices are introduced compared with the conventional solution, which indicates that the proposed method hardly affects the efficiency of the PSFB converter.
V. CONCLUSION
A highefficiency and lowcost solution is proposed to suppress transformer core imbalance in PSFB converters. Compared with the common methods, the proposed solution creates an asymmetric control algorithm, which does not accurately balance the transformer core. An approximate control consideration is achieved by a microcontroller, which generates unequal phases in the subsequent cycle according to the imbalance state in the current cycle. The proposed solution is verified by a 60 V/15 A prototype. Experimental results show that the goal of a balanced transformer core can be achieved. The efficiency of the proposed method is slightly higher than that of the PCM, and the performances of other methods are also tested and compared.
Acknowledgements
This work was supported by the National Nature Science Foundation of China (51277026), Qing Lan Project, and Suzhou Application Basic Research Project (SYG201450).
BIO
Juzheng Yu received his B.S. and M.S. degrees in electronics engineering from Southeast University, Nanjing in 2010 and 2013, respectively. He is currently working toward his Ph.D. degree in Southeast University. His research interests include digital control technique, EV/PEV charger, and highfrequency power converter.
Qinsong Qian received his Ph.D. degree in electronics engineering from Southeast University, Nanjing, China in 2012. He joined the School of Electronic Science and Engineering, Southeast University in 2012, where he is currently a lecturer. His research interests include power device design, simulations, and power converter.
Weifeng Sun received his B.S., M.S., and Ph.D. degrees in electronic engineering from Southeast University, Nanjing, China in 2000, 2003, and 2007, respectively. Since 2006, he has been with the School of Electronic Science and Engineering, Southeast University, where he is currently the dean of the School of Electronic Science and Engineering. His research interests include new power device design, power IC, power device model, and power system.
Taizhi Zhang received his B.S. degree in electrical engineering from the Hangzhou Dianzi University, Hangzhou in 2010. He is currently working toward his Ph.D. degree in Southeast University. His research interests include AC–DC, DC–DC, and singlestage PFC converters applied in LED drivers and battery chargers.
Shengli Lu received his Ph.D. degree in information and physics from Nanjing University, Nanjing, China in 1994. Since 1994, he has been with the School of Electronic Science and Engineering, Southeast University, Nanjing, China, where he is currently a professor at the National ASIC System Engineering Research Center. His research interests include VLSI and applicationspecific integrated circuit.
Hsieh Y.C.
,
Huang C.S.
2011
“Liion battery charger based on digitally controlled phaseshifted fullbridge converter,”
IET Power Electron.
4
(2)
242 
247
DOI : 10.1049/ietpel.2009.0206
Bai H.
,
Zhang Y.
,
Semanson C.
,
Luo C.
,
Mi C. C.
2011
“Modelling, design and optimisation of a battery charger for plugin hybrid electric vehicles,”
IET Electrical Systems in Transportation
1
(1)
3 
10
DOI : 10.1049/ietest.2010.0038
Mishima T.
,
Akamatsu K.
,
Nakaoka M.
2013
“A high frequencylink secondaryside phaseshifted fullrange softswitching PWM DC–DC converter with ZCS active rectifier for EV battery chargers,”
IEEE Trans. Power Electron.
28
(12)
5758 
5773
DOI : 10.1109/TPEL.2013.2258040
Gu B.
,
Lai J.S.
,
Kees N.
,
Zheng C.
2013
“Hybridswitching fullbridge DC–DC converter with minimal voltage stress of bridge rectifier, reduced circulating losses, and filter requirement for electric vehicle battery chargers,”
IEEE Trans. Power Electron.
28
(3)
1132 
1144
DOI : 10.1109/TPEL.2012.2210565
Zhou X.
,
Chen D.
,
Jamerson C.
“Leadingedge modulation voltagemode control with flux unbalance correction for pushpull converter,”
in 15th Annual IEEE Applied Power Electronics Conference and Exposition (APEC)
2000
Vol. 1
327 
333
Eberle W.
,
Liu Y.F.
“A zero voltage switching asymmetrical halfbridge DC/DC converter with unbalanced secondary windings for improved bandwidth,”
in IEEE 33rd Annual Power Electronics Specialists Conference(PESC)
2002
Vol. 4
1829 
1834
Kim J.
,
Song H.S.
,
Nam K.
2011
“Asymmetric duty control of a dualhalfbridge DC/DC converter for singlephase distributed generators,”
IEEE Trans. Power Electron.
26
(3)
973 
982
DOI : 10.1109/TPEL.2010.2070877
Gusev B. A.
,
Meleshin V. I.
,
Ovchinnikov D. A.
“Transformer core unbalancing issue in a fullbridge DCDC converter with current doubler rectifier,”
in International Conference on Power Electronics, Drives and Energy Systems
Dec. 2006
1 
6
Claassens J. A.
,
Hofsajer I. W.
“A flux balancer for phase shift ZVS DCDC converters under transient conditions,”
in 21st Annual IEEE Applied Power Electronics Conference and Exposition
Mar. 2006
Cho J.H.
,
Park K.B.
,
Park J.S.
,
Moon G.W.
,
Youn M.J.
2012
“Design of a digital offset compensator eliminating transformer magnetizing current offset of a phaseshift fullbridge converter,”
IEEE Trans. Power Electron.
27
(1)
331 
341
DOI : 10.1109/TPEL.2011.2158115
Zhu G. R.
,
Zhang D. H.
,
Chen W.
,
Luo F.
“Modeling and analysis of a rectifier voltage stress mechanism in PSFB converter,”
in 27th Annual IEEE Applied Power Electronics Conference and Exposition (APEC)
Feb. 2012
857 
862