Energy and Power Engineering, 2013, 5, 864-868
doi:10.4236/epe.2013.54B165 Published Online July 2013 (http://www.scirp.org/journal/epe)
Comparisons between CRM and CCM PFC*
Weiping Zhang2Wei Zhang2Jianbo Yang1Faris Al-Naemi1
1Materials and Engineering Research Institute (MERI), Sheffield Hallam University (SHU), Sheffield, UK
2Lab of Green Power & Energy System (GPES), North China University of Technology (NCUT), Beijing, China
Email: jumbo-yang@hotmail.com
Received September, 2012
ABSTRACT
The paper presents detailed comparisons between CRM (critical conduction mode) and CCM (continuous conduction
mode) control schemes used for Boost PFC converter. The two schemes are analyzed and compared under the chips of
L6561 and UC 3854 which are commonly used for CRM and CCM respectiv ely. Both schemes are based on multiplier;
however, the CCM is more complex and needs more periphery components which increase the cost. The Boost PFC
under CRM is easier to be implemented. Nevertheless, the variable switch frequency makes the system (including the
power-stage inductor and capacitor) hard to design. It seems that the CRM PFC is more attractive in low power applica-
tions which only need to meet IEC61000-3-2 D standard. Some experiment results are also presented for the compari-
son.
Keywords: CRM; CCM; PFC; Boost
1. Introduction
In conventional AC-DC conversion, a capacitor follow-
ing a bridge rectifier is used to derive DC voltage from
the AC power source. With this capacitor, however, the
input current pulsates. This pulsating current increases
the input current harmonics and results in a low power
factor less than 0.64[1]. To reduce the input current har-
monics and increase the power factor, a high power fac-
tor technique is desired. Power factor correction is to
make the input to a power supply look like a simple re-
sistor. Therefore, an AC-DC converter based on power
stage of boost configuration as shown in Figure 1 has
been studied as a high-power-factor pre-regulation circuit
[2].
As shown in Figure 1, according to the current of in-
ductor L, the operation modes can be specified as: CCM
(continuous conduction mode), DCM (discontinuous
conduction mode and CRM (critical conduction mode).
The critical conduction mode operates at the boundary of
CCM and DCM.
In recent years, PFC with different control schemes
has been pro posed. Nevertheless, a thorough comparison
is seldom reported.
In this paper, a detailed comparison b etween CCM and
CRM with constant on time is presented includ ing appli-
cation area, components selection and small signal anal-
ysis of the entire system. Some experiments are carried
out for the experimental comparison. The experimental
results match the theoretical analysis well.
2. Comparison of Control
2.1. Control Schemes
The control schemes of the two operating modes for the
identical boost PFC converter are depicted in Figures 2
and 3 respectively. Figure 2 illustrates the boost PFC
under CRM. Correspondingly, a boost PFC under CCM
is detailed in Figure 3.
One point should be made clear that CRM and CCM
are only concern with th e minimum of the input inductor
current. CRM means the input inductor current touches 0
without maintenance in every switching circle. While,
the input inductor current is always above zero in CCM
operating mode.
As depicted, PFC under CRM is actually the peak
current control. The current is sensed from switch and
compared with the programming current. When the input
*Project supported by Natural Science foundation of China (N0.
51277004). The Importation and Development of High-Caliber Talents
Project of Beijing Municipal Institutions (No.IDHT20130501) Figure 1. Boost PF C converter.
Copyright © 2013 SciRes. EPE
J. B. YANG ET AL. 865
Figure 2. CRM (Critical Conduction Mode).
Figure 3. CCM (Continuous Conduction M o de ).
current equals to the programming signal, the switch will
be turned off. The switch turned on signal comes from a
ZCD (zero current detection) function block. This ZCD
block will send a turned on signal to switch when the
input current reaches zero. Therefore, the input current
will touch zero in every switch cycle. In contrast, the
PFC under CCM is the average current control and has
different configuration. The input current is sensed from
inductor instead of the switch. Th is sensed current is still
compared with the programming signal; however, a cur-
rent error amplifier rather than a comparator is employed
here. Thus, the error signal is amplified and compared
with a PWM generator.
It is clear that there is one compensation network in
CRM PFC; nevertheless, two compensators exist in the
CCM PFC. More components are needed in the average
current CCM control scheme. This increases the com-
plexity and cost of the CCM average current control.
2.2. Control Blocks
According to Figures 2 and 3, the small signal block
diagrams of the two implementations can be detailed as
Figures 4 and 5 [3,4]. As shown, G1(s) is for the varia-
tions in the output of compensator to variations to the
output of multiplier. G2(s) in Figure 4 is the ratio of
output variations of the multiplier to the inductor current
variations. Gci(s) in Figure 5 means the ratio of varia-
tions of input current to the variations of the output of
current compensator. The means of other blocks are dis-
cussed in details in [4,5].
Obviously, there is one more loop in the CCM PFC
under average current control. This increases the com-
plexity. The inner loop should be decided before design
the outer loop. However, as the input current is con-
trolled directly, the distortions of the input current are
much better than the CRM PFC.
The loop gain without the compensator of the both
control scheme have just one pole [5,6]. This decreases
the complexity of design of the compensator. The com-
pensator can have the configurations and the frequency
responses simulated by Matlab are presented in Figures
6 and 7.
Figure 4. Control block of CRM.
Figure 5. Control block of CCM.
-40
-20
0
20
40
60
80
Magnitude (dB)
10
0
10
1
10
2
10
3
10
4
-90
-80
-70
-60
-50
Phase (deg)
Bode Diagram
Frequenc y (HZ)
Figure 6. One-zero, two-pole compensator.
Copyright © 2013 SciRes. EPE
J. B. YANG ET AL.
866
-80
-60
-40
-20
0
M agnitude (dB)
10
0
10
1
10
2
10
3
10
4
-90
-45
0
Phase (deg)
Bode Diagram
Frequency (HZ)
Figure 7. One-pole compensator.
Both of the compensators are suitable for the CRM
and CCM PFC except that the CCM PFC with average
current control has two loops and two compensators.
This makes the CCM PFC more complicated to design.
The CCM design process is detailed in [7]. The CRM
PFC is designed in detail in [8].
3. Comparison of Waveforms
As discussed in section 2, the switch of CRM boost PFC
is turned on when the inductor current reaches zero and
turned off when the inductor current equals to the pro-
gramming signal. Therefore, the envelope of the input
current is the rectified AC line signal since the program-
ming signal is derived from the rectified AC line voltage.
Things turn to be different in the CCM PFC. As a large
gain amplifier is implemented for the input current signal,
the input current is forced to follow the programming
signal which is also derived from the rectified AC line
voltage. As a result, the average value of input current
will follow the programming signal [7]. The inductor
waveforms and duty circle are presented in Figures 8
and 9.
As shown, the peak of input current follows a rectified
AC line signal under CRM PFC. Furthermore, the aver-
age input current equals to half of the programming sig-
nal. The average of the input current under CCM PFC
equals to the programming signal. So the CRM PFC is
not suitable f or the hig h po wer application s .
For power applications higher than 300w, CCM PFC
is widely used. For power applications lower than 300w,
CRM PFC is more popular [10].
There are actually two constants in both implementa-
tions respectively. In CRM PFC, when switch is on:
4212
34
ion
s
i
vT R
Rkvk
lRR

Figure 8. Input current of CRM.
Figure 9. Input current of CCM.
where Vi is the input voltage, Rs is the sensing resistor, l
is the input inductor and R3, R4 refers to the Figure 2; k 2
is output of voltage amplifier (E/A in Figure 2) which is
not change under stable state. Ton is the on-time of the
switch. 4
134
R
kRR
, thus,
12
on
s
kkl
TR
(2)
When the power stage circuit is determined, Ton is
determined too. As a result the turned on time of switch
of CRM PFC is constant. This is quite different with the
CCM PFC. For the average current control with UC3854;
the switch frequency is constant [9].
As for B o o s t PFC conver t e r,
11
11
o
on
in
V
T
vd
T

(3)
Therefore, for CRM boost PFC, the switch frequency
will be minimum when input voltage reaches its peak.
The max frequency happens when the input voltage goes
across zero.
The minimum switch frequency of the CRM PFC can
be obtaine d as, [8]
2,
min
(2
1
2
in rmsoin rms
in o
VV V
fLP V
,
)
(4)
where Pin is the input power, Vin,rms is the RMS value of
the input voltage, Vo is the output voltage. The variable
frequency will bring a little complexity in design of the
i
kv
(1)
Copyright © 2013 SciRes. EPE
J. B. YANG ET AL. 867
circuit. This will be discussed next.
4. Input Inductor Selection
The selections of components of CCM PFC circuit have
been provided in [7]. The approach of selections is al-
most the same with CRM PFC circuit except the input
inductor (L in Figure 1). For CRM PFC, the minimum
switch frequency dominates the design of the input in-
ductor. Mathematically manipulating of (4), the input
inductor can be obtained as.
2,
max min
(2
1
2
in rmsoin rms
in o
VV V
LfP V
,
)
(5)
Vin,rms is the rms value of the low input line voltage.
Such as 90 v when the input voltage is required between
90 – 260 v. The input current distortion is the most im-
portant consideration for the design of input inductor of
CCM boost PFC. When the power is constant, the input
current will be higher for the low input voltage (for uni-
versal input range). Thus,
,max (min),
2in
in in rms
P
Iv
(6)
Iin,max means the maximum input current. The peak-
to-peak ripple current is assumed to be 20% of the input
current. Refers to (3), the input inductor can be obtained
as,
(min),
2(min), (min),
2.5 ()
in peak
s
in peakoin peak
sino
vD
LfI
vVv
fPV
(7)
where Vin(min),peak is the peak of the low input line voltage.
Pin is the input power. Vo is the output voltage and fs is
the switching cycle.
As discussed, the selection of the inductor for CRM is
mainly concerns the minimum switch frequency. Whereas,
for CCM, the input inductor selects from the maximum
peak-to-peak rippl e c urrent.
5. Experimental Results
The performances of the two implementations are veri-
fied with two 100w proto types. The CRM PFC is carried
out by L6561 and the CCM PFC is implemented with
UC3854. The output voltage is set to be 360v. The re-
sults are shown below. As shown in Figure 10, Both of
the operating modes have almost the same power factor.
The power factor of CCM is 0.98 which is a little higher
than the one of CRM, which is 0.96. The efficiency of
CCM PFC is 84%, while the efficiency of CRM PFC is
about 80%. As the input current of the input current op-
erates in critical conduction mode, the current falls to
zero once during every switching cycle as shown in Fig-
ure 11. This cause the current distortion of CRM is more
(a) CRM: input vo ltage: 220 vac;
current: 0.57 m A (rms); power factor: 0.96
(b) CCM: input voltage: 220 vac;
input current: 0.54 mA (rms); power factor: 0.98
Figure 10. Input waveforms.
(a) Switching cycle
(b) Line frequen c y
Figure 11. Input inductor current of CRM (L in Figure 2).
Copyright © 2013 SciRes. EPE
J. B. YANG ET AL.
Copyright © 2013 SciRes. EPE
868
(a) CRM: 20 v (peak-to-peak)
(b) CCM: 15 v (peak-to-peak)
Figure 12. Output voltage ripple.
serious than the CCM PFC, which leads to a lower effi-
ciency. Besides, the CRM PFC requires EMI filters to
eliminate the high frequency harmonics of the input cur-
rent which further reduces the efficiency and complicates
the design.
Figure 12 shows the output voltage ripples. The peak-
to-peak value of the vo ltage ripple of CRM PFC is about
20 v, and the counterpart of CCM is about 15v, which is
5 v lower.
It seems that the CCM PFC is a better choice than
CRM PFC with respect to the power factor and effi-
ciency. However, there are two control loops which need
to be design in the CCM PFC. This increase the com-
plexity.
6. Conclusions
Boost PFC converters under CRM and CCM are com-
pared with each other in this paper. Some conclusions
can be obtained as:
1) The peak inductor current of the CRM PFC is twice
higher than the average input current, which makes the
CRM PFC more suitable for low power applications
(lower than 300 w). The CCM PFC dominates the high
power applications (higher than 300 w).
2) The inductor current of CRM PFC reaches zero
during each switch cycle and only has a rectified sinu-
soidal envelope. Whereas, the inductor current is almost
the same with the input current except some high har-
monics.
3)CCM PFC has the constant switch frequency and
variable turned-on time. While CRM PFC has constant
switch turned-on time, variable switch frequency which
makes the design of power stage components more com-
plicated.
4) CRM PFC has only one control loop and needs
fewer external components which is easy to be imple-
mented. There are two control loops in the CCM PFC.
the CCM PFC needs more external components which
will increase the costs, however, the input current of
CCM PFC has fewer harmonics and higher efficiency.
Both PFC operating modes have their own advantages
and disadvantages. Selection is depending on the appli-
cations.
REFERENCES
[1] X. W. Mao and D. W. Zhu, “The Control IC, Principles
and Applications of PFC,” 2007, 1st edition, China Elec-
tric Power PRESS
[2] R. Keller and G. Baker, “Unity Power Factor off-line
Switching Power Supply,” in IEEE INTELEC record,
1984, pp. 332-339.
[3] W. P. Zhang, F. Chen, X. S. Zhao and Y. C. Liu, “A Dis-
crete Modeling for Power Factor Correction Circuit,”
PEDS 2009, pp. 160-163.
[4] K. I. HWu, C. H. Wu and C. F. Chuang, “Development of
AC-DC Converter for Laboratory Power Amplifier,”
PEDS 2009, pp. 1534-1541.
[5] W. P. Zhang, “The Modeling and Control of Switch Con-
verter,” M, first edition, 1996.
[6] J. Sun, R. M. Bass, “Modeling and Practical Design Is-
sues for Average Current Control,” in Conf. Rec, APEC
Fourteenth annual, 1999, pp. 980-986
[7] P. C. TODD, “UC3854 Controlled Power Factor Correc-
tion Circuit Design”, UNITRODE application note, U134
[8] S. L. Xie, L. F. Liu and J. B. Liu, “100 W Power Factor
Correction of Basing On L6561,” AIMSEC 2011, 2nd
conference, pp. 3599-3602.
[9] S. L. Liu and Y. G. Yan, “UC3854A/B Power Factor
Boost Controllers Provide the Improvements to UC3854
and Power Limiting with Sinusoidal Input Current for
PFC Front Ends,” Electronic Components Application,
Vol. 3, No. 1, Mar 2001, pp. 31-36.
[10] J. W. Kim, S. M Choi and K. T Kim, “Variable On-time
Control of the Critical Conduction Mode Boost Power
Factor Correction Converter to Improve Zero-Crossing
Distortion” PEDS 2005, pp. 1542-1546.