Family Controller PFC
Family Controller PFC
Family Controller PFC
power-factorcorrection circuits that operate at continuousconduction-mode. Both trailing-edge and leadingedge pulsewidth modulation are used. These controllers do not require
the muitiplier and the rectifid-linevoltage sensor, which are
needed by traditional control methods, and can be
implemented with a unified control circuit to achieve
simplicity. Controller examples are analyzed and verified
experimentally.
L Introduction
A single-phase diode bridge followed by a dc-dc converter
can form a rectifier with active power-factor correction
(PFC). The dc-dc converter controller forces the average
input current i to have the same shape as the input voltage
g
I'
Such a rectlfier is called a "resister emulator."
g.
The multiplier approach and voltage-follower approach
are two traditional control strategies for the resistor
emulator[ 11. Examples of the multiplier approach include
the average-current control[21 and the peak-current
control13 ) . etc. Examples of voltage-follower approach
include buck-boost converters operating at dtscontinuousconduction mode(DCM)[1]. Cuk converters at DCM (51, and
boost converters at the boundary of DCM and continuousconduction mode (CCM)[6], etc. Rectifiers under multiplierapproach control usually operate at CCM while rectifiers
control generally operate at
undcr \,oltage-follower-approach
DCM or at the boundary of DCM and CCM. Converters
operate at DCM are usually used for low-power-level
applications. The CCM becomes necessary when power level
goes higher. because the current stress on a switch and the
current ripple in the inductor are too large for a single DCM
converter to operate efficiently. The multiplier approach
requires a multiplier in its current loop and to sense the
rectified line voltage, hence, the control circuit is more
complicated.
A number of papers have been dedicated to control
methods for the CCM operation without a multiplier and the
input-voltage sensor [7-11). Some methods can be
implemented with a very simple control circuit under the
penalty of higher current dlstortion. The non-linear-carrier
control, proposed in [8, 91 for the boost converter and other
topologies, has the simplicity of the voltage-follower
approach and the performance of the multiplier approach.
Fig. 1 shows the block dlagram for PFC circuits with the
proposed controller family. Notice that no rectified-linevoltage sensor and multiplier appear in the diagram. The
proposed PFC controllers are similar to the current-mode
controllers for dcdc converters.
The PWM modulator and the current sensor in Fig. 1 are
distinct from those used in typical multiplier-approach
controllers. The PWM modulator in Fig. 1 is the general
PWM modulator which can perform leading-edge
modulation as well as trailing-edge modulation. The sensed
current can be valley inductor current or valley switch
current besides peak and average inductor or switch current.
One can see shortly that for trailing-edge modulation, the
sensed current need to be the peak current while for leadtngedge modulation, the sensed current is the valley current.
66
rHT-;&&
Sensor
Modulator
Compensator
Vnf
-1
-1
-1
@)instant-switch-cumnt sensing
(1)
where v l . v2. ... are control inputs to the modulator and are
linear combinations of the sensed current and the
modulation voltage vm in the PFC controller case, shown in
Fig. 1. These control inputs are slow signals compared to the
switching frequency. To find the d solution for (l), the
general modulator can be applied with Q of the flip-flop as
its trailing-edge modulation duty-ratio output. Equivalent
leading-edge modulation for the same control function can
67
R
I Inlegrator
w/ reset
R
I Mqwtur
R
I integrator
W i Imet
Wi
reset
VO
@)
(10) gives the trailingedge average-witch-currentcontrol modulation equation. This equation is the basis of
(5)
(7)
(8)
hence
R, iT d = v,/M(d)
(12)
Re-arranging ( 12) yields the trailing-edge instant-switchcurrent-control modulation equation. One can find out in
Table 1 that the difference between instant-switch-current
and average-switch-current control methods is to integrate
I . IIc-LJc
converters with Fig. 4(a) conjiguration
a) Trailing-edge modulation
For the hpe of converters shown in Fig. 4(a). average
currents
<ig>=<IT>.
= vm/M(d) = v,Q(d)/P(d).
= vmQ(d)/P(d).
the instant-switch current one more time than the averageswitch current. These two methods are physically equivalent.
They both are listed for convenience of application.
Substituting (7) into (12) results in
(9)
vn, = %v&
(6)
and R, is the equivalent current-sensing resistance. For a dcdc con\.erter at CCM steady state, the instant active-switch
(MOSFET) current IT does not always equal the inductor
current iL uhen the active switch is on. instead
iD = N(d) iL
-n?on-
vg vom(d),
(4)
where M(d) = P(d)/Q(d) is the conversion ratio of the dcdc
converter. P(d) and Q(d) are polynomials in d. M(d) for five
commonly-used converters are listed in Table 1. Thus the
general PFC control function is
iT = N(d) iL .
Similarly the instant &ode current
@$T-Jqq+
<ig > = v g b .
(3)
where
is the emulated resistance. At quasi-steady state in
a line cycle, the duty ratio d needs to be controlled to satisfy
'r'lT
5 i~ d = vmW(a?N(d)l.
(13)
Re-arranging ( 13 ) yields the inductor-current-control
modulation equation.
t 10)
68
b) Leadmg-edge modulation
From (13) one can obtain a leading-edge inductor-currentcontrol modulation equation by applying (2). During t E [0,
d' T,], (8) holds. Substituting ( 8 ) into the inductor-currentcontrol equation yields the instant-switch-current-control
modulation equation. Finally, substituting
<iD> = d' iD
(14)
into the switch-current-control equation results in the
average-switch-current-control modulation equation for
leading-edge modulation.
2. Dc-dc converters with Fig. 4(b) conjguration
a) Trailing-edge modulation
For the converter type shown in Fig. 4@),
(15)
%<iL> = vm/M(d).
This equation gives the inductor-current control
(16)
>
R& = vmm(d)
(17)
under the small current ripple assumption. Substituting (7)
into (17) yields the instant-Awitch-current-control
modulation equation
E
[v"(d)l/M(d).
(18)
Finally, substituting (11) into (18) yields the averageswitch-current-control modulation equation
$ 1 ~=
&<iT>
[v, d N(d)]/M(d).
?*
(19)
b) Leading-edge modulation
Applying (2) to (17) leads to the leading-edge inductorcurrent-control modulation equation. Substituting (8) into
the inductor-current-control equation gives the instant.witch-current control. Finally. one can use (14) to get the
leading-edge average-switch-current-control modulation
equation.
Derivation results for the five commonly-used converters
are listed in Table 1. Same procedures can be used to obtain
control circuits for other topologies. Notice that the trailingedge a\.erage-switch-curn~control for boost converter is
the same control method as the non-linear-carrier control[8].
The instant-switch-current control for the trailing-edge boost
converter. which turned out to be the same method as the
Linear Peak Current Mode control proposed in [lo]. is an
approximation of the peak-current non-linear-carrier
control[81. Accurate instant-switch-current control can be
obtained through the same derivation as (13) in [SI which
can directly apply the general modulator.
For those controllers with input v4 = 0 in Table 1, two
stages of integrators are required whle those with inputs v3
= VJ = 0
.-el
69
.--1
s
r
x,
,I)'=
Yo
(26)
(2 1)
where f, = l/Ts is the switching frequency. Thus
dn' = I ; ( &
Rs
r;
-I
___ ) - I
(22)
Lfs
For constant V,
R, and Vo, the valley current is
proportional to the input voltage. The average inductor
current is then found to be
hence,
70
18
16
14
12
i? 10
B e
I-
4
2
0
0
20
40
60
80
powsr Level f%WI Loslb
100
Fig. 8. Experimental waveforms at different load power levels. Top: duty ratio ( 0 3 d i v ) ; Middle: line voltage (110 Vrms); Bottom:
l i e current, SA/div, ?,A/div, and O.SA/div respectively for (a), (b), and ( c ) . (a) at full load; (b) at 40%of full load; (c)at 10%of full
load.
71
D.Eqxrirxnta 1 Verification
An experimental circuit has been built accordmg to Fig. 5(a)
and tested. The line voltage is 110 Vrms and the output
voltage is 220 Vdc. The switching frequency fs is 100 kHz.
The inductance of the boost converter is 520 pH. An LC
EM1 filter was inserted between the diode bridge and the
boost converter with an inductance of 44 pH and a
capacitance of 0.68 pF. Experimental waveforms for a full
load of 350 W. 40% of full load, and 10% of the full load are
shown in Fig. 8 (a), (b), and (c) respectively. The waveforms
in each figure, from top to bottom, are the duty ratio
I/
,,JUT
I1
310 un
I
M 1 0 W O N CH3 l N D V
lime 5 M V
F i g . 1 1 E r g m e n t a l waveforms of the
flyhach conxerter at 100 W power output
Top dut\ ratio (0 61'div). middle h e
xoltage ( I 10 \ mis) bottom lms current
(1,Idn)
two or more diode currents, thus they can be sensed with one
CT, as shown in each figure.
3) The switching ripple current in the output filtering
capacitor is reduced. The boost converter is usually used as a
pre-regulator and the post-regulator is normally trailingedge modulated. This results in less switching ripple current
in the capacitor C as indicated in [ 131.
From Table 1 one can also find the control circuits for
flyback PFC converters. Fig. 10 shows the flyback PFC
converter of trailing-edge average-switch-current control
(the error amplifier is not shown). Under small ripple
assumption, the line current chstortion for the PFC circuit is
negligible. Similar analysis as the boost converter can be
carried out for the current distortion if the current ripple is
not negligible. As indlcated in [lo], this control method is
uncondltionally stable.
An experimental circuit was built and tested. The major
circuit parameters are as follows: fs = 63 kHz, the primary
inductance is 340 FH,flyback transformer turns ratio is 1: 1,
Lf = 100 pH, Cf = 0.68 pF. line voltage is 110 Vrms, the
output voltage is regulated at 50 Vdc. The experimental
IV.Flyback Converter
Compared
72
V. Cuk Converter
Cuk, Sepic, and Zeta converters were classified as the same
category as the flyback converter in [9] and [IO]. Neither
reference clearly addressed the control of the these
converters. The derivation results in t h s paper indmtes that
the control circuits for these converters are not always
equivalent.
To verify the control methods for the Cuk converter,
instant-inductor-current
and
average-switch-current
controlled PFC circuits were built and examined. The
circuits are illustrated in Fig. 12 and Fig. 14 respectively
with their error ampldiers not shown. The same power stage
were used for both control methods. The switchng
frequency is 100 kHz and other major parameters are shown
in the figures. The line voltage was 110 Vrms and both
outputs were regulated at 50 Vdc. Fig. 13 and Fig. 15 are
the experimental waveforms at 150-W power output. From
top to bottom the waveforms are the duty ratio, line voltage,
and line current respectively. Notice that both line currents
are closely following the line voltage in shape.
It was observed that, for the average-switch-current
control, the line current had some ringing at around 2 kHz
when the rectified line voltage changed slope rate abruptly,
i.e. when the line voltage crosses zero or is at its peak for the
dstorted line voltage. The ringing is noticeable in Fig. 15. It
is believed that thls ringing is caused by the high system
order of Cuk converter and the average-switch-current
control does not provide enough damping. With instantinductor-current control. the ringing i s significantly damped
for the same power stage. In other words, Merent control
methods may provide different dynamic performance for the
same power stage.
References
I. Sebastian, M. Jaureguizar, J. Uceda, "An overview of power
VI. Conclusions
A famil! of PFC controllers are presented in this paper
based on a general PWM modulator. The derivation
procedures are addressed in detail with derivation results for
five commonly-used converters listed in table. Both trailingedge and leading-edge modulation can be realized at
constant switching frequenq . Leading-edge modulation can
sometimes lead to simpler control circuitry as demonstrated
in the boost converter example.
Both leadng and trailing-edge modulation have three
basic control circuits for one converter accordmg to the way
that the current is measured. Physically the average-smltchcurrent control is equivalent to the instant-switch-current
control in the sense that the former is the integration or the
latter.
73
1101
111
12)
131
[14]
I151
1161