Bridgeless PFC
Bridgeless PFC
Bridgeless PFC
APPLICATION NOTE
A "BRIDGELESS P.F.C. CONFIGURATION"
BASED ON L4981 P.F.C. CONTROLLER.
by Ugo Moriconi
This technical document describes an innovative topology dedicated to a medium to high power PFC
stage. The originality of this topology is the absence of the bridge that usually is placed between the
EMC filter and the PFC stage. The advantages of this topology can be found in terms of increased efficiency and improved thermal management.
This application features the L4981 PFC controller. It is a high performance device operating in average current
mode with many on-chip functions. The driver output stage can deliver 1.5A, which is very important for this type
of application.
A detailed device description can be found in AN628. A functional block diagram is shown in Figure 1.
Figure 1. Functional Diagram
November 2002
1/18
The conventional boost topology is the most efficient for PFC applications. It uses a dedicated diode bridge to
rectify the AC input voltage to DC, which is then followed by the boost section. See Figure 2.
This approach is good for a low to medium power range. As the power level increases, the diode bridge begins
to become an important part of the application and it is necessary for the designer to deal with the problem of
how to dissipate the heat in limited surface area. The dissipated power is important from an efficiency point of
view.
Figure 2.
Inductor
D
L
A
Load
M
Mains
V_Rs
Rs
controller
The bridgeless configuration topology presented in this paper avoids the need for the rectifier input bridge yet
maintains the classic boost topology.
This is easily done by making use of the intrinsic body diode connected between drain and source of PowerMOS
switches.
A simplified schematic of the bridgeless PFC configuration is shown in Figure 3.
Figure 3.
D1
D2
Inductor
Mains
L
O
A
D
M2
M1
Controller
2/18
Load
Fig4a
D1
D2
L
O
A
D
E
R
M1
C
H
O
P
v 0 v in
P
E
M1
M2
return
controller
D2
v in
D1
Fig4b
M2
L
O
A
D
v0
return
controller
Current Sensing.
The PFC function requires controlling the current drawn from the mains and shaping it like the input voltage
waveform. To accomplish this it is necessary to sense the current and feed its signal to the control circuit.
In average current conventional boost topology, we sense the rectified current rather than the AC input current.
This can be achieved by a simple sensing resistor in the return of the current to the bridge, as shown in Figure5a.
3/18
Fig.5a
Standard Sensing
Inductor
IL
Rs
vs
Iret.
Fig.5b
Magnetic sensing for high power
IL
Inductor
1:n
Rs
vs
Lm_p
Rs
Iret.
This type of sense transformer cannot operate at low frequency and for this reason it must be connected where
the current is switched at high frequency. The magnetic core must be allowed reset.
This is normally accomplished by using a diode. In order to reproduce the inductor's current in boost topology,
two of magnetic sense sections are needed and the simplified schematic is shown in figure 5b.
When the sense transformer solution is applied in the bridgeless topology, the simple sense as in fig5b, is no
longer valid.
4/18
iin
D1
D2
L1
L
O
A
D
Rs
Q1
M1
controller
M2
vs
L1
For the PowerMOSFET portion of the circuit, the complexity increases because during the half cycle when one
of the PowerMOSFETs is chopping, the other one has to handle the current flowing back to the mains.
Using the configuration of sensors as shown in Figure 6 it is possible to solve the problem without undue complexity. The unnecessary high frequency portion of the current signal is cancelled because of the method M1 is
connected to L1A as shown in Figure 6b. The problem due to the change of polarity during each half cycle is
solved by using a center tapped secondary and two rectifiers.
Since the coupling of the two windings must not permit the demagnetization of L1, an auxiliary transistor Q1 is
used that opens the circuit during the off-time. For the L4981 controller, the off-time is guaranteed not to be less
than 5% of the period. Q1 can be a small signal transistor because its switched current is low due to the fact
that the transformer secondary will have a large number of turns.
To realize the current sensing transformer, a high permeability toroidal core (ur=>5000) has been used. The
secondary has 50 turns as a compromise to reduce secondary current yet not require a large number of turns.
5/18
L2
L1a
L1b
D1
iin
D2
Dc
Inductor
Da
Rs
L
O
A
D
Db
Q1
controller
M1
L1a
L1b
L1
vs
L1=L1a+L1b
M2
L1a
vs
L1b
L1=L1a+L1b
L2
TOT
L1
L2
TOT
vs
vRs
vsa
vsb
vRs
6/18
B
R1
Coupled
Inductor
R1
v L(t)
vL(t)
iac(t)
iac(t)
R1
R1
C1
R2
R2
C1
CURRENT
MIRROR
It is based on the following consideration: the frequency of the signal of interest (tens of Hz), is much lower than
the switching frequency (tens of kHz). The boost inductor, for the low frequency, behaves like a short circuit.
Since the Powermos's drains are, in turns, close to ground (via the body diode), the resulting equivalent circuit
is shown in fig7b.
The relation between the voltage (from the inductor) and the current that flows in to Iac pin is:
a)
1
1
Y ( s ) = ----------- -------------Req 1 + st
R1
2
The pole must be located at a frequency high enough not to distort the input waveform and at the same time,
low enough to filter the switching frequency.
In this application the equivalent resistance has been choosen
Req.=324-k that fits well with the current amplifier design.
The resulting R1 is 300k and R2 is 12k
The pole has been placed a decade before the switching frequency:
Fp = 5kHz
b)
that gives:
1
C1 = -------------------------------------------------- = 2.87nF
R1
2 fp -------- //R2
2
7/18
B
Ra
Ra
Ra
Rb
+
Ra
Rb
Ca
Cb
Rc
vLP(t)
Ca
Cb
Rc
vLP(t)
Defining HLP(s) the transfer functions between the voltage from the inductor and the voltage at the output of the
filter vLP (Fig.8B), we have the following relation:
c)
1
HLP = KL P ------------------------------------------------( 1 + st1 ) ( 1 + st2 )
Rc
KL P = ------------------------------------------------( Ra + 2Rb + 2Rc )
The time constants cannot be expressed in simple way and so that the position of poles can be numerically calculated.
The constant KLP is defined taking in to account the wide-range that is, V_mains is between 88V and 264V:
d)
2 2
V LP = VRMS ----------- KLP
e)
8/18
2 2 88 + 264
1.5 + 5.5
----------- ----------------------- K LP = ----------------------2
2
R b = 150k
R c = 30k
For the capacitors, we set 80 dB of attenuation on the fundamental frequency using the commercial values: Ca = 390nF
Cb = 470nF
The design places two poles at 3Hz and 14Hz and 80 dB of attenuation at 100Hz.
Practical examples.
The preceeding points of this note have described the topology peculiarity. Remainder of the topics, for PFC
design, are similar to standard P.F.C. boost applications based on L4981A/B (see the related references and
application notes).
Starting from now, we can refer to real design examples.
In fact, in order to verify the efficacy of the described configuration, it have been checked a pair of application's
size. For evaluation porpoise, it has been realized a printed circuit.
Let us beginnes with and 800W P.F.C application.
800W Target:
1 - Wide range input voltage variation 110Vrms to 220Vrms.
2 - Output power 800W.
2 - Output voltage 400Vdc.
A switching frequency of 50 kHz has been chosen as a good compromise between the coil-size and the powerMOS switching losses.
9/18
iL
Inductor
Mains
vs(t)
is(t)
Zin
Leq.
is(t)
Zin
vs(t)
Mains
Equivalent circuit.
Realizing the inductor in this manner improves common mode rejection and avoids the effect of the difference
between drain capacitance of the PowerMOSFETs. In order to simplify the model, assume a near unity coupling
factor and the equivalent circuit is shown in Figure 9b.
The inductance is proportional to the square of the number of turns. For the two windings it will be:
f)
N
N = ---- and
2
N N
N total = ---- + ---2 2
The required number of turns for a given inductance on the same core is the same as it is for one winding or
two windings. The only difference is that the two windings are separated into two sections. For simplicity we can
design the coupled inductor using the same criteria as for a standard inductor - core size, number of turns, and
size of copper wire.
For the core, the preferred design is a gapped ferrite core set.
The size of the core can be chosen considering the maximum current Ipk. that, for the 800W target's parameters
can exceed 14A (placing Ipk. = 15A).
g)
Where:
(mm )
4 lcore
K = 1.4 10 -------------lgap
For the 800W application, the nominal current ripple has been chosen around 25%. This fixes the boost inductance value L=450H.
10/18
Ae = 550mm^3;
lcore = 146mm;
m_core = >1600;
Vcore = 80.4*10^3mm^3
The air gap needed to avoid saturation and optimize the coil size is equal to lap = 3mm.
Vcore>67.5mm^3
This result confirms the core is well above the minimum size.
The used formula for the number of turnes, needed to design the total required inductance L, is:
h)
N =
1gap
l, core
L
------ ----------------------------- + --------------------------------------------2
0 r, core A
A+
--- 1gap
The resulting N=38, in our solution, has been realized with 19 turns +19 turns.
In order to minimize the high frequency losses, the winding has been made using the "multiple wire" approach.
It is possible to estimate the losses for a low frequency current.
Imposing a maximum power value to be dissipated in the copper (Pcu = 5W)
i)
P wire
R DC < ------------------------ = 60m
2
I RMS , max
l)
ltu rn N
R DC = Cu ----------------------- < 60m
2
--- d M
4
Were:
In practice 20 wires were used, each having a diameter d=0.4 mm.
11/18
m)
Po
C o = --------------------------------------------------2 2 f fo Vo
Power Devices.
The selection of the power devices is dependent upon the topology and the size of the application.
Operating in continuous current mode, fast reverse recovery diodes are needed.
The TURBOSWITCH "STM family", in the 600V voltage range, offers a very good solution for the two boost
diodes, the STTH8R06FP has been chosen.
The insulated TO-220 package makes it easy to assemble the parts on a heat sinke.
Concerning the Powermos requirements, a 500V blocking voltage (Bvdss) is needed, for this application.
The chip selection is more complex. To find the best solution, it must be considered all the parameters that affect
the power dissipation and to compare the results in terms of a cost to benefit ratio. The devices used in the 800W
application (2+2), are the type STY34NB50F.
The four Powermos are efficiently driven without any additional buffer, thanks to the smart characteristics of the
integrated driver.
Figure 10. 800W SCHEMATHIC DIAGRAM.
R1
Vcc
D1+D2+D3+D4
Dz1
NTC
L3
C2
D6
D5
D13
L 2c
L 2a
F1
Input
C3
C1
L 2b
Q1
R2
R27
R28
R29
R32
R33
R30
R31
D7
Q2
Q3
D8
R14
D12
R8
R24
R25
R26
D10
R5 Q5
R4
R9
R7
R10
C16
R23
C5
20
Vcc
C15
C14
8 2
R17
4
11
6
L4981A/B
7
C13
18
17
R22
C6
C7
R18
12
10
C8
3
1
DETAIL for L2
19
15 16
Vcc
9
C12
13
14
R19
R21
L 2a
C10
L 2c
C11
L 2b
C4
R12
R15
R13
R16
Q4
D9
R3
D11
R6
12/18
R11
L1
R20
Value
Name
Value
Name
Value
Name
Value
Name
Value
R1
68
R2,3,4,5
10
R6
1.8
R7
100
R8
R9
2.7 k
R10
1.5 k
R11,12,14,15
1 M
R13
22 k
R16
25.5 k
R17
3.9k
R18
27 k
R19
220 k
R20
2.7 k
R21
5.6 k
R22,24,27
30 k
R25,26,28,29
499 k
R34
12 k
RSN
12
C1,3
1 F
C2
220 F
C4
330 F
C5
10 nF
C6
1 F
C7
1.8 nF
C8
1 F
C10
120 nF
C11
1 nF
C12
5.6 nF
C13
470 nF
C14
2.7 nF
C15
100 nF
C16
390 nF
NTC
2.5
B57364
F1
20A
R23,30,31,32,33 150 k
Dz1 = 1N4746;
Q5 = BS170
D5,6 = STTH8R06FP;
Q1,2,3,4=STY34NB50F
Note: For the evaluation circuit and external coupled inductor EMC where utilized.
The filter has been achieved as follows: Coupled inductor (30 + 30) turns; wire diameter = 0.8mm on a toroidal (40x17x9 mm): Magnetizing inductance (each half inductor) Lm=8mH and a leakage inductance, Ld=50H.
600W Target:
1- Wide range input mains 110Vrms to 220Vrms.
2- Output power = 600W.
2- Output voltage = 400Vdc.
The switching frequency has been set at 75kHz to use a reduced size and high performance PowerMOS.
13/18
A = 357mm^2;
lcore = 123mm;
m_core =>1600;
Vcore = 43.7mm^3
For the 600W, the coil has been realized with 21 turns +21 turns.
For minimize the high frequency losses, the "multiple wire" solution has been used.
Imposing the copper losses (Pcu = 3.8W), it has been used 14 wires having a diameter d = 0.4 mm each.
Output Capacitor.
For the selection of CO, the relation as been described in (m). The commercial value = 330F/450V used for
the 800W application is still good for the 600W application.
Power Devices.
For the two boost diodes, as for the 800W application, the STTH8R06FP has been used.
Concerning the Powermos, the devices used in the 600W version application are two STW26NM50F.
14/18
Vcc
D1+D2+D3+D4
Dz1
NTC
L3
C2
D6
D5
D13
L2c
L2a
F1
R11
R14
R12
R15
R13
R16
L1
Input
C3
C1
L2b
D11
C4
D12
R8
Q2
Q5
D9
R27
R28
R32
R29
R33
D8
R36
R3
R4
R35
R30
R24
R25
R26
C16
R23
R10
R31
C5
20
16
Vcc
C15
R9
R7
R6
C14
8 2
R17
4
11
6
L4981A/B
7
C13
R22
18
17
C6
C7
R18
12
10
C8
3
1
DETAIL for L2
19
15
9
C12
Vcc
13
14
R19
R21
C9
L2a
L2c
R20
C11
C10
L2b
Value
Name
Value
Name
Value
Name
Value
Name
Value
R1
68
R3,4
10
R6
6.8
R7
100
R8
6.8
R9
2.7 k
R10
1.5 k
R11,12,14,15
1 M
R13
22 k
R16
25.5 k
R17
3.9k
R18
33 k
R19
220 k
R20
2.7 k
R21
5.6 k
R22,24,27
30 k
R25,26,28,29
499 k
R34
12 k
C1,3
1 F
C2
220 F
C4
330 F
C5
10 nF
C6
1 F
C7
1.8 nF
C8
1 F
C10
120 nF
C11
1 nF
C12
5.6 nF
C13
470 nF
C14
2.7 nF
C15
100 nF
C16
390 nF
NTC
2.5
57364
F1
15 A
R23,30,31,32,33 150 k
Dz1=1N4746;
Q5 = BS170.
15/18
Q2,3 = STW26NM50F
Note: For the evaluation circuit and external coupled inductor EMC where utilized.
The filter has been achieved as follows: Coupled inductor (30 + 30) turns; wire diameter = 0.8mm on a toroidal (40x16x8.5 mm): Magnetizing inductance (each half inductor) Lm=8mH and a leakage inductance, Ld=50uH.
Conclusion:
The innovative bridgeless PFC configuration as described in this application note has been successfully tested.
Details have been presented how to implement the technology, which should prove interesting to designers.
Figure 12 shows the test results of efficiency and power dissipation for the application's 800W prototype.
Figure 12.
EFFICIENCY
[%]
98
B.Less
Standard PFC
97
96
95
94
P0=800 W
93
92
88
110
132
154
176
198
220
242
264
Vin
DISSIPATED POWER
[w]
80
70
60
50
40
30
20
10
0
B.Less
Standard PFC
P0=800 W
88
110
132
154
176
198
220
242
264
Vin
16/18
Pout
Pin
PF
TDH
Efficiency
395VDC
800W
860W
0.999
94%
824W
0.997
97%
800W
Pout
Pin
PF
TDH
Efficiency
395VDC
652W
700W
0.998
6.7
93%
624W
0.994
96.5%
652W
References: a) Parsad N. Enjeti, R. Martinez "A high performance single phase AC to DC rectifier with input power factor
correction" IEEE APEC'93
b) Alexandre Ferrari de Souza and Ivo Barbi "A new ZVS Semi resonant High Power Factor Rectifier with Reduced Conduction Losses"
IEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL.46, NO.1 FEBRUARY 1999.
c) STM Application Notes AN628; AN824.
STMicroelectronics @ www.st.com, http://ccd.sgp.st.com/stonline/books/index.htm
17/18
Information furnished is believed to be accurate and reliable. However, STMicroelectronics assumes no responsibility for the consequences
of use of such information nor for any infringement of patents or other rights of third parties which may result from its use. No license is granted
by implication or otherwise under any patent or patent rights of STMicroelectronics. Specifications mentioned in this publication are subject
to change without notice. This publication supersedes and replaces all information previously supplied. STMicroelectronics products are not
authorized for use as critical components in life support devices or systems without express written approval of STMicroelectronics.
The ST logo is a registered trademark of STMicroelectronics
2002 STMicroelectronics - All Rights Reserved
STMicroelectronics GROUP OF COMPANIES
Australia - Brazil - Canada - China - Finland - France - Germany - Hong Kong - India - Israel - Italy - Japan -Malaysia - Malta - Morocco Singapore - Spain - Sweden - Switzerland - United Kingdom - United States.
http://www.st.com
18/18