15th International Power Electronics and Motion Control Conference, EPE-PEMC 2012 ECCE Europe, Novi Sad, Serbia
Single-Phase Single-Stage Bidirectional Isolated
ZVS AC-DC Converter with PFC
F. Jauch, J. Biela
Laboratory for High Power Electronic Systems, ETH Zurich
Physikstrasse 3, CH-8092 Zurich, Switzerland
Email: jauchf@ethz.ch
Abstract—This paper presents a single-phase bidirectional
isolated AC-DC converter with Power Factor Correction (PFC)
consisting of a half-bridge on the AC side and a full-bridge on
the DC side to accomplish single-stage power conversion. The
converter applies a new control scheme combining phase-shift
and frequency modulation to achieve Zero-Voltage-Switching
(ZVS) over the full range of the AC line voltage. Compared to
the conventional boost PFC approach, the proposed converter
eliminates high frequency harmonic distortions on the mains
due to the inherently integrated LC input filter stage. The
operating principle in AC-to-DC and DC-to-AC under ZVS
conditions by means of analytical considerations are provided.
Simulation results and a detailed loss model of a 3.3 kW electric
vehicle battery charger to connect to the 230 Vrms / 50 Hz mains
considering a battery voltage range of 280 V to 430 V validate
the theoretical analysis. The converter can also be used as a
submodule in a Cascaded H-Bridge Converter (CHB) for medium
or high voltage applications.
Keywords—AC-DC Converter, Power Factor Correction (PFC),
Bidirectional, Isolated, Zero-Voltage-Switching (ZVS)
I. I NTRODUCTION
Single-phase isolated AC-DC converters are widely used for
applications like charging (hybrid-)electric vehicles, interfac-
ing storage batteries (e.g. for uninterruptible power supplies)
or supplying energy from photovoltaic systems to the grid.
Some of these applications demand bidirectional power flow
capability, e.g. for implementing Vehicle-2-Grid (V2G) con-
cepts or grid battery storage systems.
Besides the basic two-stage approach with a boost Power
Factor Correction (PFC) rectifier and a subsequent high-
frequency isolated DC-DC converter such as Dual Half-Bridge
(DHB), Dual Active (Full-)Bridge (DAB) or resonant DC-
DC converters [1], several single-stage isolated AC-DC PFC
converter topologies have been proposed. A review of state-of-
the-art single-phase power quality AC-DC converters is given
in [2] whereas [3] summarizes single-phase non-isolated PFC
topologies based on the boost converter approach.
Unidirectional single-stage AC-DC PFC converters operat-
ing in Discontinuous Conduction Mode (DCM) are proposed
in [4], [5] and [6]. However, high power applications ask for
PFC operation in Continuous Conduction Mode (CCM) due to
high current stresses of the switching devices and conducted
emissions in DCM [7]. Furthermore, the output capacitances of
the switching devices resonate with the boost inductor during
the non-conducting interval.
In [8], [9] a true bridgeless high-efficient AC-DC PFC
converter operating in CCM based on a resonant Cuk con-
verter with high power factor is presented. There, integrating
978-1-4673-1972-0/12/$31.00 ©2012 IEEE LS5d.1-1
L1 i1
L2 i 2
S1a
vC 1 + C1
S1b S3 S5
Lσ iLσ
V1 + C3 +
vp Lm vs vC 3 V2
-
n:1
S2a S4 S6
vC 2 + C2
S2b
Fig. 1. Proposed single-phase single-stage bidirectional isolated AC-DC
converter with PFC which can be used as a submodule in a Cascaded H-Bridge
Converter (CHB) in star connection for medium or high voltage applications
(see Fig. 12).
isolation into the topology is also possible [10]. Though, the
converter system only allows unidirectional power flow.
Bidirectional power flow and single-stage AC-DC operation
can be achieved using a Cycloconverter on the primary and
a DAB on the secondary side of a medium-frequency trans-
former [11]. Due to the bidirectional switching devices on
the primary side, a sequential multi-step commutation scheme
must be used which in turn limits the switching frequency.
Other approaches use phase-shift control as shown in [12],
where a half-bridge on the primary and secondary side of the
transformer and an input diode rectifier is utilized. The diode
rectifier can be replaced by a synchronous rectifier applying
MOSFETs to further reduce conduction losses. Furthermore,
for shaping the transformer leakage inductance current in order
to achieve soft-switching over the full range of the line voltage,
a DAB can be used on the transformer secondary side [13].
This paper proposes a single-phase bidirectional isolated
AC-DC converter with PFC as shown in Fig. 1 employing
combined phase-shift and frequency control to achieve Zero-
Voltage-Switching (ZVS) over the full range of the AC line
voltage and additionally eliminates the input rectifier stage by
use of bidirectional switching devices.
First, in section II the proposed converter topology is
explained. The operating modes in AC-to-DC and DC-to-
AC case including the converter control to achieve ZVS
are described and analyzed in section III. Then, section IV
shows a prototype system of the proposed converter including
a detailed loss model. Finally, simulation results verify the
theoretical analysis. Assuming further a negligible small capacitor voltage ripple,
the secondary DC-link voltage is vC3 = V2 for the following
II. C ONVERTER T OPOLOGY mathematical considerations.
The topology of the AC-DC converter is shown in Fig. 1. During one half-cycle of the AC line voltage, only two of
Due to the alternating line voltage, bidirectional switching the primary switching devices are switched at high frequency.
devices on the AC side have to be employed. These can be These are S1a and S2a for the positive and S1b and S2b for
realized by an anti-serial connection of two MOSFETs as the negative half-wave. The devices switched at low frequency
shown in Fig. 1 or by reverse blocking IGBTs. The AC side are turned on/off at nearly zero voltage and zero current.
of the proposed converter has been presented in [14], [15] As indicated in Fig. 2b, on the primary side of the trans-
considering single-stage AC-AC power conversion. Moreover, former, a square-wave voltage vp with the time-dependent
a similiar primary circuit is used in [16] to supply high amplitude |v1 (t)| /2 is applied. The voltage vs on the trans-
pressure sodium lamps. former secondary side consists of positive and negative voltage
Compared to the conventional boost PFC approach, the pulses with an amplitude of V2 , which allows shaping the
proposed converter eliminates high frequency harmonic distor- transformer leakage inductance current iLσ to transfer the
tions on the mains due to the inherently integrated input filter desired instantaneous power and to achieve ZVS. The voltage
stage. Capacitors C1 and C2 fully absorb the high frequency amplitudes of vp and vs are assumed to be constant over one
switched currents. Nevertheless, the remaining AC voltage switching cycle because the switching frequency is chosen to
ripple on the capacitors leads to a certain degree of high be well above the mains frequency.
frequency input current distortions.
In either power flow direction, the converter can operate in A. AC-to-DC Operation
buck or boost mode. The transformer turns ratio n is chosen In AC-to-DC operation, power flows from the mains to the
such that the primary referred DC voltage V2 is higher than DC side. To describe the phase-shift operation mathematically,
half of the maximum peak voltage of the AC line voltage two control variables g(t) and w(t) relative to the switching
in every operating point. This guarantees, that at the lowest period Ts are introduced. Fig. 3 shows the applied voltages
battery voltage the primary and secondary applied half-cycle on the transformer primary and secondary side vp and vs over
voltage-second products with respect to n can be kept equal. one switching period Ts and the definition of the phase-shift
control variables g and w during the positive and negative
III. O PERATING P RINCIPLE
half-cycle of the mains in AC-to-DC operation (Modulation 1).
The converter is operated with a combined phase-shift and Moreover, the resulting transformer leakage inductance current
frequency control as shown later. A simple equivalent circuit is iLσ and the states of the switching devices are given.
drawn in Fig. 2a which consists of the LC input filter stage and The transformer leakage inductance current iLσ in AC-to-
a parallel connected variable impedance load Zl . The converter DC operation is given by
is operated in such a way, that the reactive power consumed ⎧ |v1 |
by the filter capacitors is fully compensated by the variable ⎪
⎪ 2 + nV2
⎪
⎪ (τ − τ0 ) + iLσ (τ0 ) τ0 ≤ τ ≤ τ1
impedance load in order to achieve PFC. ⎪
⎪ Lσ
⎪
⎪ 1|v |
Fig. 2b shows a simplified representation of the converter ⎪
⎪
for phase-shift control. The magnetizing inductance Lm of ⎪
⎪
2
(τ − τ1 ) + iLσ (τ1 ) τ1 ≤ τ ≤ τ2
⎪
⎪ L
⎪
⎪ σ
the transformer is assumed to be much larger than the leakage ⎪ |v1 |
inductance Lσ and is therefore neglected in the following. The ⎪ 2 − nV2 (τ − τ ) + i (τ )
⎪
⎨ τ2 ≤ τ ≤ τ3
2 Lσ 2
primary half-bridge is switched with a constant duty cycle of iLσ (τ ) = Lσ
|v1 |
⎪
⎪ − 2 − nV2
50 %, which leads to equally distributed capacitor voltages vC1 ⎪
⎪ (τ − τ3 ) + iLσ (τ3 ) τ3 ≤ τ ≤ τ4
and vC2 . Neglecting the input inductor L1 , the average value ⎪
⎪ Lσ
⎪
⎪
⎪
⎪ − 2 |v 1 |
over one switching cycle of these voltages follows the AC line ⎪
⎪ (τ − τ4 ) + iLσ (τ4 ) τ4 ≤ τ ≤ τ5
⎪
⎪
⎪ L|v
voltage according to ⎪ σ
⎪
⎪ |
⎩ − 2 + nV2 (τ − τ5 ) + iLσ (τ5 ) τ5 ≤ τ ≤ τ6
1
V̂1 ⎪
vC1 (t) = vC2 (t) = sin (ωt) . (1) Lσ
2 (2)
with the values at the switching instants
L1 i1 |v1 |
+ (g − w)nV2
4
iLσ (τ0 ) = − , (3)
2fs Lσ
vC 1 + C1 Lσ iLσ
v1 (g − 14 ) |v1 | + (g + w)nV2
Zl iLσ (τ1 ) = , (4)
vC 2 + C2 vp vs 2fs Lσ
( 14 − w) |v1 | + (g + w)nV2
iLσ (τ2 ) = , (5)
2fs Lσ
(a) (b)
|v1 |
Fig. 2. Equivalent circuits of the proposed converter: (a) with variable 4 + (g − w)nV2
impedance load Zl and (b) for representing phase-shift control. iLσ (τ3 ) = , (6)
2fs Lσ

LS5d.1-2
 Modulation 1
V2 transferrable instantaneous power
 
iLσ |v1 (t)| nV2 v12 (t) + 4 |v1 (t)| nV2 + 4n2 V22
pt,max (t) =
32fs Lσ (v12 (t) + 4n2 V22 )
|v1 | vp (13)
2
dependent on the AC line voltage v1 (t).
Ts
2 vs The transfer of active power takes place in the intervals
0 [τ1 , τ3 ] and [τ4 , τ6 ] which is also shown by the shaded area in
Fig. 3. During the intervals [τ0 , τ1 ] and [τ3 , τ4 ], only reactive
− |v21 | power occurs that is required to achieve ZVS. The control
gTs wTs gTs wTs
function g(t) will be given in section III-C.
B. DC-to-AC Operation
−V2 In DC-to-AC operation, power flows from the secondary
S1a
to the primary side of the transformer which demands a

v 1 ( t) < 0 v 1 ( t) > 0
S1b different sign of the applied transformer voltages vp , vs and
S2a the transformer leakage inductance current iLσ . Fig. 4 shows
S2b the transformer voltages vp , vs , the leakage inductance current
S1a iLσ and the gating signals of the switching devices for the
S1b
S2a
positive and negative half-wave of the mains (Modulation 1).
S2b The analytical expression of the transformer leakage induc-
S3 tance current iLσ in DC-to-AC operation can be determined
S4 the same way as in AC-to-DC operation considering the
S5 switching instants from Fig. 4. Setting iLσ (τ3 ) = iLσ (τ5 )
S6
leads to (11) for the control variable w(t).
τ0 τ1 τ2 τ3 τ4 τ5 τ6 = T s τ The instantaneous power transferred from DC to AC side
is given by (12) through −pt (t). The same holds for the
Fig. 3. Primary and secondary voltages vp , vs applied to the transformer
and resulting transformer leakage inductance current iLσ over one switching
period Ts in AC-to-DC operation. The gate signals of the primary switching
devices differ during the positive and negative half-wave of the AC line
Modulation 1
V2
voltage.
iLσ

(g − 14 ) |v1 | + (g + w)nV2 |v1 | vp

iLσ (τ4 ) = − , (7) 2
2fs Lσ
Ts
2 vs
( 1 − w) |v1 | + (g + w)nV2 0
iLσ (τ5 ) = − 4 . (8)
2fs Lσ
− |v21 |
The ZVS conditions considering a minimum commutation
current Is for the resonant transition are given by wTs gTs wTs gTs

iLσ (τ1 ) > Is , (9)

−V2
iLσ (τ3 ) > Is . (10) S1a
v 1 ( t) < 0 v 1 ( t ) > 0

S1b
In order to guarantee a trapezoidal current shape as shown in S2a
S2b
Fig. 3, iLσ (τ1 ) = iLσ (τ3 ) is demanded, which leads to one
S1a
single ZVS condition for the whole switching cycle inducing S1b
S2a
|v1 (t)| (1 − 2g(t)) S2b
w(t) = . (11)
4nV2 S3
S4
The instantaneous power over one switching cycle Ts trans- S5
ferred from the AC to the DC side can then be calculated S6
by
τ0 τ1 τ2 τ3 τ4 τ5 τ6 = T s τ
 
|v1 (t)| nV2 g(t) − 2g 2 (t) + w(t) − 2w2 (t)
pt (t) = (12) Fig. 4. Primary and secondary voltages vp , vs applied to the transformer
4fs Lσ and resulting transformer leakage inductance current iLσ over one switching
period Ts in DC-to-AC operation. The gate signals of the primary switching
taking (11) into account. Inserting (11) into (12) and setting devices differ during the positive and negative half-wave of the AC line
the derivative with respect to g to zero, leads to the maximum voltage.

LS5d.1-3
 10 pt (gmax (t))
maximum transferrable power, which is defined by (13) by
means of −pt,max (t).

Instantaneous power (kW)

8 pt (gp (t))
Active power is transferred in the intervals [τ0 , τ2 ] and
Δpt (t)
[τ3 , τ5 ] which is also shown by the shaded area in Fig. 4. 6 pt (gmin (t))
Only reactive power occurs during the intervals [τ2 , τ3 ] and
[τ5 , τ6 ] that is required to achieve ZVS. 4

C. Converter Control under ZVS Conditions (Modulation 1) 2

To achieve ZVS for all switching devices during one half- 0

cycle of the mains, the turn-off currents have to be large 0 2 4 6 8 10
Time t (ms)
enough to charge/discharge the drain-source capacitances of
(a)
the switching devices in a bridge-leg. Looking at the reso-
nant transition and introducing energy equivalent capacitances
0.25
Ceq,p , Ceq,s for the parallel connection of the drain-source Border to upper limit of switching frequency

Instantaneous power (kW)

capacitances in a bridge-leg both for AC and DC side, the 0.2
conditions for ZVS in both AC-to-DC and DC-to-AC opera-
tion are 0.15 pt (gp (t))
pt (gmax (t))
v1 V2
iLσ (τ1 ) >  and iLσ (τ1 ) >  , (14) 0.1
Lσ Lσ p∗t (t) pt (gmin (t))
Ceq,p Ceq,s
0.05
v1 V2
iLσ (τ3 ) >  and iLσ (τ3 ) >  . (15) 0
Lσ Lσ 0 0.1 0.2 0.3 0.4 0.5 0.6
Ceq,p Ceq,s Time t (ms)
(b)
Conditions (14) and (15) result in a minimum for the control
variable g(t) given by Fig. 5. Instantaneous power transferred from AC to DC side over a half-
⎧ ⎫ cycle of the mains (a) and in the range of the zero-crossing (b) in AC-to-DC
⎨ v operation for controlling with gmin (t), gp (t) and gmax (t). Additionally, the
4fs Lσ Is 1 V2 ⎬ reference value of the transfer power p∗t (t) according to (20) is drawn in (b).
gmin (t) = Is = max  , . The vertical line indicates the border to the switching frequency saturation
|v1 (t)| + 2nV2 ⎩ Lσ Lσ ⎭
region.
Ceq,p Ceq,s
(16)
Furthermore, the maximum transferrable instantaneous power
according to (13) occurs at which has to be transferred to the transformer secondary side
for a given input current amplitude Iˆ1∗ can be expressed as
v12 (t) − |v1 (t)| nV2 + 2n 2
V22
gmax (t) = , (17) ω (C1 + C2 ) 2
2 (v12 (t) + 4n2 V22 ) p∗t (t) = V̂1 Iˆ1∗ sin2 (ωt) − V̂1 sin(ωt) cos(ωt),
4
which is set to be the upper limit for the control variable (20)
g(t). To keep a certain power control reserve without losing which takes the reactive power consumed by the capacitors
ZVS, the control variable gp (t) is introduced, which provides C1 , C2 into account to achieve PFC. Inserting (18) into (12)
the average power between controlling with (16) and (17) as while considering (11) and setting
shown in Fig. 5a. The analytical expression is found to be
pt (t) = p∗t (t) (21)
3
α |v1 (t)| + (αnV2 + β) v12 (t) + 4αn3 V23 + 4βn2 V22
gp (t) = gives the solution of the additional control variable fs (t). The
4 (v12 (t) + 4n2 V22 ) (|v1 (t)| + 2nV2 ) upper limit of the switching frequency is set to 120 kHz, which
(18) is reached near the zero-crossing of the AC line voltage. There,
with α, β given by the reference value of the transfer power p∗t (t) according to
√ √
α=2− 2, β = 8 2fs Lσ Is . (19) (20) falls below the transferrable power obtained by control-
ling with gmin (t) as can be seen in Fig. 5b. In order to avoid
In Fig. 5a the instantaneous power transferred from the AC to high switching losses, ZVS is maintained by controlling with
the DC side during a half-cycle of the mains for controlling gmin (t) what in turn leads to a small input current distortion.
with gmin (t), gp (t) and gmax (t) is drawn. At every point in Fig. 6 shows the three control variables gp (t), wp (t) and
time, a power control reserve of ±Δpt (t) is obtained. fs (t) over the half-wave of the mains in AC-to-DC operation
To keep the power control reserve and guarantee the transfer for an input current of 16 Arms and an output voltage of V2 =
of the required instantaneous power, an additional degree of 355 V.
freedom, the switching frequency fs is considered as time-
dependent control variable. To maintain ZVS at low powers D. Partial Load Operation (Modulation 2)
around the zero-crossing of the mains voltage, g is kept greater Decreasing the load, especially below 50 % of the maximum
than zero. Therefore, the switching frequency has to rise, in output power, the saturation region of the switching frequency
order to lower the transferred power. The instantaneous power around the zero-crossing of the AC line voltage is extended

LS5d.1-4
 0.32 140 Modulation 2
V2

Switching frequency fs (kHz)

wp (t)
0.24 110
Ts
τ

iLσ
Relative time

|v1 | vp
0.16 80 2

Ts
0.08 gp (t) 50 2 vs
0
fs (t)
0 20 − |v21 |
0 2 4 6 8 10
Time t (ms)
wTs gTs wTs gTs
Fig. 6. Control variables gp (t), wp (t) and fs (t) over a half-cycle of the
mains period in AC-to-DC operation for a mains voltage of 230 Vrms , an
input current of 16 Arms and an output voltage of V2 = 355 V. The controller −V2
adjusts gp (t) in the bounds [gmin (t), gmax (t)], the solution of gp (t) is used
as feed-forward variable. τ0 τ1 τ2 τ3 τ4 τ5 τ6 = T s τ

Fig. 8. Primary and secondary voltages vp , vs applied to the transformer

25 Modulation 2 Modulation 1 1 and resulting transformer leakage inductance current iLσ over one switching
PF
Total Harmonic Distortion (%)

period Ts in partial load AC-to-DC operation.

20 0.99
Power Factor

15 0.98
Border of modulation
IV. P ROTOTYPE S YSTEM
10 0.97 As a prototype system, a 3.3 kW electric vehicle battery
THD charger to connect to the single-phase 230 Vrms / 50 Hz mains
5 0.96 with an output voltage range of 280 Vdc to 430 Vdc of a
Lithium-ion battery is considered. The primary and secondary
0 0.95 switching devices are chosen to be 650 V MOSFETs of type
0 0.5 1 1.5 2 2.5 3 3.5 4
Input Power (kW) STY139N65M5 with an on-resistance of 14 mΩ at a junction
temperature of 25 ◦C [17]. The system parameters are listed
Fig. 7. THD of the input current and PF over the input power range from 10 % in detail in Table I. A 3D drawing of the prototype system is
to 100 % of maximum input power for a battery voltage of 355 V applying shown in Fig. 13.
the proposed modulation scheme (Modulation 1) according to Fig. 3 and
Fig. 6 (solid lines). The dashed lines result from the change of the modulation
scheme (Modulation 2) according to Fig. 8 at 45 % of the input power (vertical
A. Converter Components and Loss Model
line). In the following, the converter components are described
with their loss models to calculate the efficiency of the
converter applying the proposed control scheme. Table II
with the proposed modulation strategy which results in a summarizes the components of the prototype system whereas
remarkable input current distortion. Fig. 7 shows the Total Fig. 9 shows the calculated efficiencies over the input power
Harmonic Distortion (THD) of the input current and the Power range from 10 % to 100 % for the battery voltages 280 V,
Factor (PF) of the converter over the input power range up 355 V and 430 V. For the proposed modulation strategy
to the maximum input power of 3.68 kW. From there, it can during high load conditions (Modulation 1), a peak efficiency
be concluded that the proposed control strategy is applicable of 97.8 % is reached at an input power of 2.58 kW and a
almost down to 40 % of the maximum input power, keeping battery voltage of 280 V. The estimated power density is
the THD below 5 % and the PF above 0.985. The limit is set around 3.9 kW/L. Fig. 10 shows the calculated distribution
to 45 % of the input power where the vertical line is drawn in of the power losses among the converter components at the
Fig. 7.
In case of an input power below 45 % of the maximum TABLE I
input power, the modulation is slightly changed (Modulation PARAMETERS OF THE PROTOTYPE SYSTEM .
2). Phase-shift control as shown in Fig. 8 is applied, where the
positive pulse of the secondary voltage waveform lies totally Mains voltage V1 230 Vrms ± 10 %
within the positive pulse of the primary voltage waveform. Mains frequency fg = 1/Tg 50 Hz
Battery voltage V2 280 Vdc . . . 430 Vdc
Demanding iLσ (τ0 ) = iLσ (τ2 ) (again in order to get one
Output power P2 3.3 kW
single ZVS condition) leads in this case also to (11) for the Switching frequency fs 20 kHz. . . 120 kHz
solution of w(t). To maintain ZVS also at low input powers, Transformer turns ratio n 10/13
g(t) > 0 has to be guaranteed for a minimum commutation Transformer leakage inductance Lσ 20 μH
current of zero. Practically, ZVS can not be maintained be- Transformer magnetizing inductance Lm 11 mH
cause the resonant transition after turn-off to charge/discharge Inductors L1 , L2 100 μH
Capacitors C1 , C 2 10 μF
the output capacitances of the switching devices requires a
Capacitor C3 20 μF
certain turn-off current.

LS5d.1-5
 TABLE II
C OMPONENTS OF THE PROTOTYPE SYSTEM . Rds,on of the chosen MOSFET is found to be 18.2 mΩ.

MOSFETs S1a , S1b , S2a , S2b 2x STY139N65M5, 650 V, 14 mΩ
MOSFETs S3 , S4 , S5 , S6 2x STY139N65M5, 650 V, 14 mΩ
Transformer 2x 2x AMCC-4 VITROPERM 500
10 primary turns, 120 μm copper foil
13 secondary turns, 120 μm copper foil
Inductor L1 2x Kool Mu E 4317 26u
27 turns, round wire, diameter 2 mm
Inductor L2 2x Kool Mu E 4317 26u
28 turns, round wire, diameter 2 mm
Capacitors C1 , C2 18x Syfer 1825J500564KX, 560 nF
Capacitor C3 36x Syfer 1825J500564KX, 560 nF
98 V2 = 280 V V = 355 V
2
2) Transformer: The turns ratio of the transformer n is
chosen such that vC3 
> V̂1 /2 (with vC3 referred to the primary
side of the transformer) is always satisfied, also at the lowest
battery voltage of 280 V. The maximum transformer leakage
inductance Lσ is determined in such a way, that the peak of the
instantaneous power P̂1 (neglecting the reactive power term in
(20)) at full input power of 3.68 kW can be transferred at the
lowest switching frequency of 20 kHz and the lowest battery
voltage of 280 V by controlling the converter with gp (t). Lσ
can be obtained by inserting (18) into (12) while considering
(11) and solving

97.5 P̂1 = V̂1 Iˆ1 = pt (t = Tg /4). (23)

Efficiency (%)

97
The calculated Lσ guarantees the above mentioned power
96.5 V2 = 430 V control reserve ±Δpt (t) also at the peak of the instantaneous
power.
96
Border of modulation For the transformer, two C-cores of size AMCC-4 [18] with
95.5 VITROPERM 500 [19] material form an E-core where two of
Modulation 2 Modulation 1 them are stacked. The primary winding is wound around the
95
0.5 1 1.5 2 2.5 3 3.5 4 center leg and the secondary winding around the center and an
Input Power (kW) outer stray leg as shown in [20]. In the stray leg, an air gap
is inserted with a gap length of 0.85 mm to get the desired
Fig. 9. Calculated efficiencies over the input power range from 10 % to 100 % leakage inductance. In the loss model, the core losses per
of maximum input power for battery voltages 280 V, 355 V and 430 V. The
vertical line defines 45 % of the input power where the modulation scheme volume are calculated by applying the improved Generalized
is changed to limit input current distortions. Steinmetz Equation (iGSE) [21].
For the primary and secondary windings, the optimal foil
Capacitors: 5 %
Auxiliary: 9 % thickness is determined as 132 μm and 116 μm according to
Transformer: 27 % [22], which gives a minimum value of effective AC resistance.
120 μm copper foil is used for the prototype system, with
10 primary and 13 secondary turns. The skin and proximity
effect losses per unit length in foil conductors for each current
harmonic are then calculated according to [23]. The external
Secondary Inductor: 2 %
Primary Inductor: 4 %
magnetic field strength for calculating proximity losses is
Power MOSFETs: 52 % derived by a 1D approximation using the Dowell method [24].
Fig. 10. Calculated distribution of the power losses among the converter
components at the maximum input power of 3.68 kW and a battery voltage 3) Inductors: The primary and secondary inductors are
of 355 V. built with two stacked E-cores of type Kool Mu 4317 with
material 26u from Magnetics [25]. Powder cores are ideally
suited for the prototype system because they offer a distributed
maximum input power of 3.68 kW and a battery voltage of air gap and a high saturation flux density of around 1 T and are
355 V. therefore advantageous over a ferrite core with a large air gap
exhibiting considerable fringing magnetic field. Both inductors
1) Power MOSFETs: The losses of the power MOSFETs are wound with round wire of 2 mm diameter since the high
are mainly determined by conduction losses. Switching losses frequency components in the input/output current are relatively
PSi,sw per MOSFET are approximated by measurement data. small. The number of turns for primary inductor L1 are 27,
To reduce conduction losses, Ns number of MOSFETs are for secondary inductor L2 28, so that a minimum inductance
paralleled, so that the power loss per switching device is then value of 100 μH is guaranteed at the highest peak current.
approximated by The core losses per volume are calculated by using the
Rds,on 2 iGSE, the Steinmetz parameters are obtained from [25]. The
PSi = I + Ns PSi,sw (22) skin and proximity effect losses per unit length in round
Ns Si,rms
conductors for each current harmonic are calculated according
with i = {1a, 1b, 2a, 2b, 3, 4, 5, 6}. The prototype system to [23]. Also for the inductors, the external magnetic field
applies two MOSFETs in parallel for all switching devices strength is derived by a 1D approximation using the Dowell
(Ns = 2). For conduction loss calculations, a junction method [24].
temperature of 55 ◦C is assumed where the on-resistance

LS5d.1-6
  V2 
4) Capacitors: For primary and secondary capacitors
v1
C1 , C2 , C3 , paralleled 560 nF ceramic capacitors with di-

Voltage (V)

Current (A)
 
i2
electric X7R from Syfer [26] are used. Multilayer ceramic  
i1
capacitors exhibit high energy density and are therefore ideally  
suited to achieve high power densities. Since the voltage
 
ripple on the capacitors is relatively small at high switching
frequencies, dielectric losses are not accounted for in the loss 
model. Therefore, only the thermal losses according to vs

Voltage (V)
 vp
Resr 2 0
PCi = I i = {1, 2, 3} (24)
Nc Ci,rms 

where Resr denotes the equivalent series resistance obtained
from datasheet and Nc the number of capacitors paralleled, 
iLσ
are considered.

Current (A)


5) Auxiliary Losses: Besides the load dependent loss 
shares shown in the previous sections, a constant loss share
 
for pre-charging relay, gate drives, control, sensing and  5  15    35 
cooling of 8 W is considered. Additional losses caused by an Time (ms)
EMI filter are approximated by an equivalent resistance of
4 mΩ. Fig. 11. Simulation results of the proposed converter
√ with parameters given
in Table I for an input current reference Iˆ1∗ = 2 · 16 Arms and an output
voltage V2 = 355 V in AC-to-DC operation for a mains voltage of 230 Vrms .
6) Cooling System: The number of semiconductors ba-
sically defines the base plate size of the heat sink as Phase c
80 mm × 65 mm for primary and secondary side switching Phase b
Phase a Storage EV Battery EV Battery
devices, so that a double-sided heat sink can be used. Two Battery
DC + +
40 mm × 40 mm fans of type San Ace 40 are applied for forced Submodule
+
-
- -

convection cooling. After optimizing the cooling system as DC

described in [27], a thermal heat sink to ambient resistance of
Storage
Rth,s−a = 0.25 K/W results which in turn leads to a Cooling Battery
DC
System Performance Index (CSPI) of 12.6. +
Submodule -
DC
B. Simulation Results Storage
Battery
MV AC Grid DC
The proposed converter is simulated in GeckoCIRCUITS +
Submodule
[28] with a simulation model according to Fig. 1 and the La -

parameters
√ given in Table I for an input current reference
DC

Iˆ1∗ = 2 · 16 Arms and an output voltage V2 = 355 V in

AC-to-DC operation for a mains voltage of 230 Vrms as can Fig. 12. The proposed AC-DC converter applied as submodules connected in
be seen in Fig. 11. series in the three phase branches which form a star connection in a battery
energy storage system. Subsequent DC-DC converters can be connected to
the storage batteries and paralleled at the outputs to achieve high charging
V. F URTHER A PPLICATIONS powers for DC fast charging of (hybrid-)electric vehicles [30], [31].

The AC-DC converter depicted in Fig. 1 can also be used as

a submodule in a medium/high voltage battery energy storage
system shown in Fig. 12 where the submodules per phase are
connected in series sharing single filter inductances La , Lb , Lc
and the three phase branches form a star connection like in
the Cascaded Multilevel PWM Converter presented in [29]
also known as the Cascaded H-Bridge Converter (CHB). The
battery energy storage system drawn in Fig. 12 additionally
functions as a fast charging station for (hybrid-)electric vehi-
cles [30], [31]. Subsequent DC-DC converters are connected to
the storage batteries whose outputs are paralleled to achieve
high output powers for DC fast charging. Alternatively, the
outputs of the submodules can be directly connected to the
battery of an electric vehicle.
VI. C ONCLUSION
A single-phase single-stage bidirectional isolated AC-DC
converter with PFC is proposed. The converter achieves ZVS
for all switching devices over the full range of the AC
line voltage due to the combined phase-shift and frequency
modulation. The inherently integrated LC input filter stage
allows a substantial reduction of harmonic distortions on
the mains. For validating the theoretical analysis, a 3.3 kW
electric vehicle battery charger to connect to the single-phase
230 Vrms / 50 Hz mains with a battery voltage range of 280 V
to 430 V is investigated. During high load conditions, a peak
efficiency of 97.8 % is reached at a power density of around
3.9 kW/L. The converter can also be used as a submodule

LS5d.1-7

Fig. 13. 3D drawing of the prototype system.
in a CHB in star connection for medium and high voltage
applications
ACKNOWLEDGMENT
The authors would like to thank Swisselectric Research and
the Competence Center Energy and Mobility (CCEM) very
much for their strong financial support of the research work
and VACUUMSCHMELZE for providing the core material for
the transformer.
R EFERENCES
[1] R. L. Steigerwald, “A comparison of half-bridge resonant converter
topologies,” IEEE Transactions on Power Electronics, vol. 3, no. 2, pp.
174–182, 1988.
[2] B. Singh, B. N. Singh, A. Chandra, K. Al-Haddad, A. Pandey, and
D. P. Kothari, “A review of single-phase improved power quality AC-
DC converters,” IEEE Transactions on Industrial Electronics, vol. 50,
no. 5, pp. 962–981, 2003.
[3] J. P. M. Figueiredo, F. L. Tofoli, and B. L. A. Silva, “A review of
single-phase PFC topologies based on the boost converter,” in Proc. 9th
IEEE/IAS International Conference on Industry Applications (INDUS-
CON), 2010, pp. 1–6.
[4] W. Choi and J.-S. Yoo, “A bridgeless single-stage half-bridge AC/DC
converter,” IEEE Transactions on Power Electronics, vol. 26, no. 12, pp.
3884–3895, 2011.
[5] W.-Y. Choi and J.-Y. Choi, “A novel single-stage AC-DC converter to
supply sustain power for plasma display panels,” Journal of Display
Technology, vol. 7, no. 9, pp. 494–502, 2011.
[6] A. J. Sabzali, E. H. Ismail, M. A. Al-Saffar, and A. A. Fardoun, “New
bridgeless DCM Sepic and Cuk PFC rectifiers with low conduction and
switching losses,” IEEE Transactions on Industry Applications, vol. 47,
no. 2, pp. 873–881, 2011.
[7] K. D. Gusseme, D. M. Van de Sype, A. P. M. Van den Bossche, and
J. A. Melkebeek, “Input-current distortion of CCM boost PFC converters
operated in DCM,” IEEE Transactions on Industrial Electronics, vol. 54,
no. 2, pp. 858–865, 2007.
[8] S. Cuk, “True bridgeless PFC converter achieves over 98% efficiency,
0.999 power factor,” Power Electronics Technology Magazine, vol. 36,
no. 7, pp. 10–18, July 2010.
[9] S. Cuk, “Bridgeless PFC converter achieves 98% efficiency, 0.999 power
factor,” Power Electronics Technology Magazine, vol. 37, pp. 34–40,
August 2010.
[10] S. Cuk, “Single-stage bridgeless isolated PFC converter achieves 98%
efficiency,” Power Electronics Technology Magazine, vol. 39, pp. 22–31,
October 2010.
[11] S. Norrga, “Experimental study of a soft-switched isolated bidirectional
AC-DC converter without auxiliary circuit,” IEEE Transactions on
Power Electronics, vol. 21, no. 6, pp. 1580–1587, 2006.
LS5d.1-8
[12] M. H. Kheraluwala and R. W. De Doncker, “Single phase unity power
factor control for dual active bridge converter,” in Proc. IEEE Industry
Applications Society Annual Meeting, 1993, pp. 909–916.
[13] J. Everts, J. Van den Keybus, and J. Driesen, “Switching control strategy
to extend the ZVS operating range of a dual active bridge AC/DC
converter,” in Proc. IEEE Energy Conversion Congress and Exposition
(ECCE), 2011, pp. 4107–4114.
[14] D. Chen and J. Liu, “The uni-polarity phase-shifted controlled voltage
mode AC-AC converters with high frequency AC link,” IEEE Transac-
tions on Power Electronics, vol. 21, no. 4, pp. 899–905, 2006.
[15] D. Chen, “Novel current-mode AC/AC converters with high-frequency
AC link,” IEEE Transactions on Industrial Electronics, vol. 55, no. 1,
pp. 30–37, 2008.
[16] G. C. R. Sincero and A. J. Perin, “High pressure sodium lamp high
power factor electronic ballasts using AC-AC converters,” IEEE Trans-
actions on Electronics, vol. 22, no. 3, pp. 804–814, 2007.
[17] [Online]. Available: http://www.st.com
[18] [Online]. Available: http://www.hitachimetals.com
[19] [Online]. Available: http://www.vacuumschmelze.de
[20] U. Badstuebner, J. Biela, and J. W. Kolar, “Power density and efficiency
optimization of resonant and phase-shift telecom DC-DC converters,”
in Proc. 23rd Annual IEEE Applied Power Electronics Conference and
Exposition (APEC), 2008, pp. 311–317.
[21] K. Venkatachalam, C. Sullivan, T. Abdallah, and H. Tacca, “Accurate
prediction of ferrite core loss with nonsinusoidal waveforms using only
Steinmetz parameters,” in Proc. 8th IEEE Workshop on Computers in
Power Electronics, June 2002, pp. 36–41.
[22] W. G. Hurley, E. Gath, and J. G. Breslin, “Optimizing the AC resistance
of multilayer transformer windings with arbitrary current waveforms,”
IEEE Transactions on Power Electronics, vol. 15, no. 2, pp. 369–376,
2000.
[23] J. Lammeraner and M. Stafl, Eddy Currents. Iliffe Books Ltd., 1966.
[24] P. Dowell, “Effects of eddy currents in transformer windings,” Proc.
Institution of Electrical Engineers, vol. 113, no. 8, pp. 1387–1394, 1966.
[25] [Online]. Available: http://www.mag-inc.com
[26] [Online]. Available: http://www.syfer.com
[27] U. Drofenik, A. Stupar, and J. W. Kolar, “Analysis of theoretical limits
of forced-air cooling using advanced composite materials with high
thermal conductivities,” IEEE Transactions on Components, Packaging
and Manufacturing Technology, vol. 1, no. 4, pp. 528–535, 2011.
[28] [Online]. Available: http://www.gecko-research.com
[29] L. Maharjan, S. Inoue, and H. Akagi, “A transformerless energy storage
system based on a cascade multilevel PWM converter with star configu-
ration,” IEEE Transactions on Industry Applications, vol. 44, no. 5, pp.
1621–1630, 2008.
[30] [Online]. Available: http://ufcev.epfl.ch
[31] M. Vasiladiotis, A. Rufer, and A. Beguin, “Modular converter architec-
ture for medium voltage ultra fast EV charging stations: Global system
considerations,” in Proc. IEEE International Electric Vehicle Conference
(IEVC), 2012, pp. 1–7.