36 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 52, NO.

1, FEBRUARY 2005

Boost-Chopper-Derived PFC Rectifiers:

Interest and Reality
Jean-Christophe Crébier, Bertrand Revol, and Jean Paul Ferrieux

Abstract—This paper presents an overview of single-phase to the input current HF harmonic content. In boost-derived
boost-chopper-derived power-factor-correction rectifiers. Con- topologies, based on either continuous current mode (CCM)
verters are presented for several control strategies in continuous or discontinuous current mode (DCM) operation, it is mainly
and discontinuous current mode operations. The presenta-
tion highlights the main characteristics of the converters and the differential-mode harmonic contents that may be greatly
underlines harmonic content over a wide frequency range. A affected. Therefore, a more or less large differential-mode filter
methodology is derived to compute these harmonics in order to must be added to comply with the standard regulation and these
facilitate the analysis and performance evaluation of the con- filter needs should be considered as a necessary component in
verters. Based on this methodology, a comparison is conducted in the design routine for optimal converter selection and operation.
order to highlight advantages and drawbacks of the most popular
converters and control strategies. A table is given where main The objectives of this paper are to present and to analyze
converter critical parameters are summarized in order to help the what the price to pay is in terms of design, and also character-
designer , in converter and operating mode selections. Practical istics when simple and low-cost low-power-level single-switch
results are provided to validate the modeling technique. boost-chopper-derived rectifiers are used as a possible solution
Index Terms—Filtering, modeling, power-factor correction, rec- for standard compliance. This paper will first summarize the
tifiers. main topologies with a classical state of the art in a power range
of a few hundred watts, and then, a modeling technique will be
presented in order to help analyze and compare their main char-
I. INTRODUCTION
acteristics. Special care will be paid in order to consider in the

T HE introduction of harmonic and total harmonic dis-

tortion (THD) limitations with regulation EC61000-2-3
in Europe [1] has generated large amount of research and
comparison the impact of required differential-mode filtering in
order to comply with regulations. The modeling technique will
be validated with a practical example and then used as a tool to
development activities around power-factor-correction (PFC) estimate input current harmonic contents. Comments and com-
converters to be connected to the mains. These researches have parisons will be provided in order to help the power electronic
brought numerous of topologies for effective PFC [4]–[19], [22] designers in their choices depending on specifications and ap-
in single- and three-phase applications for a wide power range plications.
from a few watts up to several hundred kilowatts. Depending on
specifications, the PFC rectifiers have been either designed for II. LOW-POWER SINGLE-PHASE CHOPPER RECTIFICATION
low-cost applications or high-quality rectification. However,
Low-power PFC converters are not required to operate per-
in all cases, this is achieved throughout high-frequency (HF)
fectly and most of the current solutions are trying to comply
switching patterns and filtering, bringing us to other regulation
with the regulation at minimum cost and complexity. As a re-
constraints in the medium- and HF range with EC55011 [2]
sult many PFC structures and operating principles have been
for conducted and radiated electromagnetic interference (EMI)
derived cascading a bridge diode rectifier and a dc–dc chopper
regulations. The problems are in fact moved in the upper fre-
without storage element between them (see Fig. 1). The full-
quency range with the assumption that they are easier to handle
bridge diode rectifier is then used to rectify the input sine wave
for compliance. It is then questions of conducted differential
whereas the chopper is used as an input current shaper. The
and common-mode current disturbance filtering with low size
ability in input current shaping offers the opportunity to regulate
passive L–C and sometimes L–C–R circuits. In boost-derived
power flow and, therefore, output voltage. Most parts of chopper
rectifiers, common-mode disturbances are mainly generated
rectifiers are derived from the boost chopper and its associated
by applied to parasitic capacitors. As far as differen-
family. Some exceptions are possible if the constraints on power
tial-mode disturbances are concerned, they are mainly related
flow are limited. In this e paper, we will focus on boost-chopper-
derived rectifiers and we will start with a large overview of ex-
Manuscript received December 31, 2003; revised February 1, 2004. Abstract isting single-phase chopper rectifiers. This will include the pos-
published on the Internet November 10, 2004. sible operating principles and main control strategies used to
J.-C. Crébier and J. P. Ferrieux are with the Laboratoire d’Electrotechnique provide PFC. A short discussion will end this section to present
de Grenoble, INPG/UJF CNRS UMR 5529, F-38402 SMH Cedex, Grenoble,
France (e-mail: [email protected]). three-phase low-power chopper rectification.
B. Revol was with the Laboratoire d’Electrotechnique de Grenoble,
INPG/UJF CNRS UMR 5529, F-38402 SMH Cedex, Grenoble, France. He is A. Single-Switch Boost Chopper Rectifiers
now with SATIE-ENSC/CNRS, UMR 8029, F-94235 Cachan Cedex, France
(e-mail: [email protected]). The simplest and probably the most known structure is the
Digital Object Identifier 10.1109/TIE.2004.841143 single-switch boost chopper rectifier (see Fig. 1). It has a limited
0278-0046/$20.00 © 2005 IEEE
CRÉBIER et al.: BOOST-CHOPPER-DERIVED PFC RECTIFIERS: INTEREST AND REALITY 37

Fig. 1. Single-switch single-phase boost rectifier topology.

amount of components and it can operate under many types of

control strategies. The main electrical characteristics are compa-
rable to a classical boost chopper which implies that the output
voltage value must be greater than the maximum input voltage
value.
PFC can be obtained at different quality levels depending on
design specifications and operating conditions. Very simple con- Fig. 2. Time- and frequency-domain representations of the input current.
(V = 230 V, V = 385V, P =1 kW, I 1 =2 A, L =24: mH,
trol such as constant duty cycle, constant frequency [5], or line
frequency commutated control [6] can be implemented allowing
F = 20 kHz.)

for compliance with standard IEC 1000-3-2 up to some power

levels for low-cost high-efficiency applications. More complex
control strategies allow us to shape very well the input current as
a sinusoid acting on the duty cycle or the switching frequency
[7]. It can either operate in CCM or DCM. For CCM opera-
tion and constant switching frequency, variable duty cycle of-
fers very nice PFC as can be seen in Fig. 2. The control strategy
remains quite simple and dedicated ICs are produced for this
[8]. The remaining HF ripple sets at switching frequencies ( )
around 20 kHz and above and must be filtered to comply with
standards.
Duty cycle evolution (1) and boost inductor computation (2)
are given as

(1)

(2)

Hysteretic PFC can also be used under continuous current op-

erating mode as it has been presented in [9]. It allows one to
spread HF harmonics while keeping the THD very low. How-
ever, for EMI compliance, hysteretic operation does not provide
a real interest as it will be underlined later.
DCM operation and constant switching frequency and duty Fig. 3. Time- and frequency-domain representations of the input current
(F ;  constants). (V= 230 V, V = 650V, P =1 kW, L = 600 H,
cycle provides limited input current shaping as can be seen
in Fig. 3. Although the current-shaping quality increases
 = 05
: ;F = 20 kHz.)

with greater output voltage levels, constraints on switches

and filtering output capacitors do not allow for very good low-level input voltage such as in the U.S. Besides, DCM
performance. It can be demonstrated that there is a compro- operation moves harmonics in the upper frequency range but
mise to find between power level and output voltage level for large harmonic levels must then be filtered in order to comply
compliance with standards. This solution is more suitable for with standards.
38 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 52, NO. 1, FEBRUARY 2005

Fig. 4. Time- and frequency-domain representations of the input current (

variable). (V = 230 V, V = 385 V, P = 1 kW, L = 200 H, F = 20 Fig. 5. Time- and frequency-domain representations of the input current (F
kHz.) variable). (V = 230 V, V = 650 V, P = 1 kW, L = 660 H, T =
25 s.)
Boost inductor value is derived as a function of all parameters
with (3) are given as a function of (the on-state time).

(6)
(3)
(7)
DCM operation with constant switching frequency, variable
duty cycle offers much better performance in the low-frequency with
range as can be seen in Fig. 4. Here, again, DCM operation re-
quires filtering of large harmonics in the upper frequency range. (8)
Two specific parameter calculations are listed here: the boost
inductor computation (4) and the duty cycle evolution as a func- As it has been underlined, all DCM operation control techniques
tion of time (5) generate high-level HF harmonics levels that need to be fil-
tered. The magnitude of these harmonics may require “com-
plex” filter topologies (complex in the sense of number of el-
(4) ements, volume, and weight). As has been presented in the in-
troduction, supposed cost-effective solutions may not be so in-
teresting when the EMI filter is added to the price. As a result,
(5)
several other topologies have been developed by the community
in order to reduce this problem. Several of them are presented
DCM operation with variable switching frequency and hereafter as possible examples for comparison with a classical
constant on-time seems to offer a nice compromise between single-switch boost rectifier.
low-frequency and high-frequency harmonics generation (see
Fig. 5). It operates as a perfect input current shaper and its par- B. Interleaved Boost Rectifier
ticular operating principle spreads out HF harmonics, giving, The simplest solution consists of interleaving technique.
as a first approximation, lower HF harmonic levels [7], [10]. Fig. 6 presents a boost rectifier made out of two boost choppers
However, taking into account IEC55011 standard testing and in parallel [11]. Several chopper units can be paralleled. The
measuring conditions (CISPR16 [3]), the spectral envelope two switching frequencies are out of phase in order to minimize
used for experimental acquisition summarizes and averages HF ripple and, therefore, filter needs. Any kind of the above
harmonics leading to quite large filter requirements in order to control strategies can be implemented here with respect to
comply. application specifications. More components are required in
Boost inductor computation is given in (6). Minimum and this case but these components are designed to handle the input
maximum switching cycle frequencies computations (7) and (8) current only during half the switching period.
CRÉBIER et al.: BOOST-CHOPPER-DERIVED PFC RECTIFIERS: INTEREST AND REALITY 39

allows one to supply each boost rectifier at half the input line
voltage. This means that with simple control strategy (constant
duty cycle, constant frequency), better performance in terms of
PFC can be expected from this structure.
In Fig. 9 is presented a set of plots showing the interest of
such topology. The operating conditions are constant duty cycle,
constant switching frequency. It clearly appears that the input
current has a reduced HF ripple while boost choppers are still
operating in DCM. In addition, this solution exhibits similar
low-order harmonics compared with the classical solution pre-
sented in Fig. 4. However, in this case, the output voltage level
(325 V) is about half reduced.
Fig. 6. Interleaved boost chopper single-phase rectifier topology. Below are given (9) and (10) to calculate what the sizes of
passive components for such structure are, as functions of oper-
ating conditions and maximum input voltage ripple across the
capacitors

(9)

(10)

D. Multilevel Boost Rectifier

This configuration is presented in Fig. 10 and it takes advan-
tage of output stage multilevel structures [13]. It allows well-
known three-level operation that reduces the HF ripple while
dividing by two the stress voltage on the output components.
Of course it requires an increasing number of active and passive
components and control complexity (two drivers with electrical
insulation). The main disadvantages of this technique are that
the output stage bottom voltage level is not clamped to a dc or
at least low-frequency signal reference. Moreover, control of the
neutral voltage is necessary to guarantee half output voltage on
each capacitor. Finally, it may have two different operating re-
Fig. 7. Time- and frequency-domain representations of the input current
gions, when is lower than and when it is greater. This
(F ;  constants). (V = 230 V, V = 650 V, P = 1 kW, requires complex control strategies and DCM operation gives
L = 2 3 650 H, F = 20 kHz.)  = 0:373. medium results.
For low output voltage, the input current waveform presents
Computations for inductor, control duty cycle, and switching a discontinued phase when the input voltage is lower than half
frequency remain comparable to the above. Fig. 7 presents time- output voltage (see Fig. 11).
and frequency-domain representations of the input current. As Such converter gives medium results for low output voltage
can be seen, the input current HF harmonic content seems to be and power. The increase in the cost induced by the second
reduced but it is a question of operating point. At 650-V output switching cell is not justified for low power. This type of PFC
voltage, the ripple cancellation is very good but if the output rectifier seems more interesting for higher output voltages.
voltage is reduced down to 385 V, this is not any more the case,
and this solution is not as attractive as expected. E. Other Cases
Other low-cost low-power single-phase rectifiers have been
C. Boost Rectifier With Artificial Neutral developed to perform PFC. They are more or less complex and
A boost rectifier with artificial neutral (see Fig. 8) is another some of them include an HF transformer that provides specific
candidate for PFC with HF input current ripple reduction. It was electrical insulation. The most popular converter is probably the
first presented in [12]. It considers two boost choppers in an flyback PFC [14] that can generate nice input current shaping
interleaved configuration and supplied throughout a capacitor and output voltage control with a very simple control command
bridge. Here, the main objective is to minimize the input cur- in DCM operation (constant duty cycle, constant switching fre-
rent HF ripple and also to provide better input current shaping. quency). This topology can also operate in CCM in order to re-
The main advantage of this solution is that the capacitor bridge duce current ratings on components, but in all cases the structure
40 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 52, NO. 1, FEBRUARY 2005

Fig. 8. Interleaved boost chopper single-phase rectifier topology.

Fig. 9. Input current (000 ) inductor current (- - -) time domain waveforms.

Input current frequency-domain representation. (V = 230 V, V = 325 V,
3
P = 1 kW, L = 2 300 H,  = 0:5; F = 20 kHz.)
Fig. 11. Time- and frequency-domain representations of the input current
(F ;  constants). (V = 230 V, V = 385 V, P = 1 kW, L = 40 H,
F = 20 kHz.)

in order to operate under soft-switching conditions [15]. These

structures remain more difficult to implement and to control.
Low-power three-phase PFC can also be considered. Several
structures derived from the boost chopper topology have been
studied [16], [17]. However, PFC capability of these structures
remains quite limited (only one control for three phase currents)
and they are still generating quite large conducted EMI. Studies
have been conducted in order to minimize them [18], [19]. Al-
though it is not usual to consider three-phase rectification for
low-power applications, it can be useful to increase power flow
Fig. 10. Three-level boost chopper single-phase rectifier topology.
reliability with three-phase sources.
Afterthesestate-of-the-artclassicalboostchopperrectification
will generate large HF input and output current ripples to be fil- techniques and topologies, it would be of interest to define which
tered. Other PFC structures take advantage of HF transformers solutionoffersthebestperformance.However,thisisalwaysavery
CRÉBIER et al.: BOOST-CHOPPER-DERIVED PFC RECTIFIERS: INTEREST AND REALITY 41

specific function of the application where the low-power low-cost

PFCconverterwilltakeplace.Inordertohighlightadvantagesand
drawbacks of these solutions, we believe that another parameter
needs to be considered. It is the HF filter that needs to be added. As
hasbeenunderlinedintheprevioussection,allstructuresoperateat
highswitchingfrequencyinordertoshapesinusoidaltheinputcur-
rentwhileprovidingoutputvoltagecontrol.Thiswillmovethehar-
monic filtering issue into the upper frequency range and it seems
realistictoestimateHFfilteringneedsbeforecomparisonandpos-
siblestructureselections.Thefollowingsectionisgoingtopresent Fig. 12. Interleaved boost chopper single-phase rectifier topology.
briefly a simple methodology that has been used in order to de-
rive HF differential-mode harmonics generated by the above con- For each switching period, three causal ramps are required.
verters and operating modes. The study can be limited to differen- Their sum gives a time-domain mathematical representation of
tial-mode harmonic estimation because it is the most susceptible the high-frequency input current
to control strategies and topologies. This modeling will also offer
the computation of THD generation in order to help compare the
PFC capability of each converter.

III. MODELING AND ANALYSIS OF THD AND EMI GENERATION

A. Modeling Method (11)
The objective of this simple and approximate modeling tech- where is the modulation factor with
nique is to derive, in a limited amount of time, steady-state PFC
rectifier input current harmonic content. It is based on a mathe- (12)
matical representation of input current shape over a line period.
This representation is then used to derive the frequency-domain (13)
representation of the input current, giving us the opportunity to
verify and quantify possible needs to comply with IEC or other (14)
standards.
This kind of method can be applied anytime input current can The following step consists of a summation of functions
easily be characterized (boost- or buck-derived converters). Any over a line period :
kind of control strategy can be used as long as switching events
and the input current waveform of each switching period can be (15)
theoretically forecasted. It can be valid up to several megahertz
depending on the accuracy of the current waveform and propa- Then, using the Laplace/Fourier transform mathematical tools,
gation paths modeling. it is simple to obtain the mathematical expression of the fre-
The description of the methodology is: knowing the switching quency-domain representation of the input current. This is given
events (duty cycle evolution “ ” and switching frequency “ ” by the Laplace transform over a line period (16)
for example) as well as switching period current waveforms (see
Fig. 4 for example), a simple model based on the Laplace formula-
tion can be derived [20]. All these switching period mathematical
formulations are then summed up over a line period. This gives the
complete input current pattern in a mathematical form (11). Using (16)
the Fourier transform, a frequency-domain representation of this
current is derived. It directly gives a result in steady state and it can The Fourier transform is then derived simply replacing:
easily be included in an optimization routine. where represents the harmonic number from the
As an example for validation, the modeling technique is going fundamental up to the switching frequency multiples.
to be applied to the single-phase boost rectifier operating in
DCM with constant switching frequency, constant duty cycle C. Validation
control command. Validation with time-domain simulation software Simplorer
[21] is proposed first. For low-order harmonics, compliance
B. Modeling Description with IEC 1000-3-2 is obtained. For both low- and high-fre-
Time-domain and frequency-domain representations in quency ranges results, the spectrums are in good accordance,
Figs. 2, 3, 4, 5, 7, 9, and 11 have been derived using this mod- as can be seen in Fig. 13.
eling technique. Looking in detail at the input current-switching Only main harmonics are computed with the mathematical
waveforms given in Fig. 12, it clearly appears that it is a sum model in order to save computation time and to clarify the com-
of triangles that can be fully characterized with respect to parison. Validation with experimental results has been carried
operating conditions. out in order to qualify modeling accuracy. Chosen operating
42 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 52, NO. 1, FEBRUARY 2005

These results validate the theoretical forecast and the interest

of the method. This modeling technique is valid at least one
decade above the switching frequency of the studied converters.,
because with higher switching frequencies, greater care is paid
to parasitic impacts on converter operation [22]. As this is going
to be underlined later in the paper, this is usually an important
first estimation of differential-mode filtering requirements at
switching frequency levels and above it. Indeed, low-frequency
harmonic filtering requirements are in general the harder to
reach (in terms of volume and values) because the lower the
filter corner frequency, the greater the passive components will
have to be. It is mentioned that the methodology can of course
be applied to hysteretic control strategies, variable switching
frequency [20]. This is possible any time a mathematical model
can be derived for duty-cycle and/or on-time values and current
waveforms. In addition, it simplifies modeling and time simu-
lation consumption when closed-loop control is considered. In
this case, depending on control loop dynamic parameters, an
equivalent duty-cycle pattern can be derived. Then, the impact
of control on conducted EMI can be observed [19] with great
time savings compared with time-domain simulations. It is now
Fig. 13. Low- and high-frequency harmonics comparison between both going to be used in order to forecast filtering requirements for
method results.
the converters and operating modes that have been presented in
the previous paragraphs.

IV. ANALYSIS AND COMPARISON

This section of the paper is probably the most critical since
it is always very subjective to compare solutions and to extract
the best one. As a result, the comparison among the different so-
lutions will be limited to qualitative remarks. A large table will
complete the presentation of the different structures for quanti-
tative estimation of critical and characteristic parameters.
The driving parameters have been defined as follows in
order to outline the interest of each solution. First of all, THD
capability and constraints are defined, and then electrical stress
on components is defined and compared to a reference. Finally,
total passive component requirements (including a simple
Fig. 14. Comparison between theoretical and practical spectral representation differential-mode filtering) are estimated and expressed as a
of input current (V = 385 V, V = 127 V ; P = 500 W, L = storage energy level need.
300 H,  = 0:49; F = 23 kHz.)
The differential-mode filtering capacitor computation has
been realized using a computation routine in accordance with
point and spectral representations are given in Fig. 14. Compar- standard testing conditions and regulation limits [3]. Basically,
ison with experiments is required to consider practical nonideal a capacitor is added between a line impedance stabiliza-
behaviors such as the following [20]. tion network (LISN—the structure of this standard measuring
• Input ac power supply is not perfect and low-order har- device is plotted in Fig. 15) and the converter and its value
monics are applied to the converter. is computed such, that under testing conditions, the first HF
• Discontinuous current patterns need to take into account harmonics entering within regulation limits (see Fig. 15), are
the resonance at the 0-A level. correctly filtered (in our case, all converters are operating
These two nonideal circumstances have been taken into ac- at 20-kHz switching frequency, therefore, the first critical
count in the model in order to improve the harmonics compu- harmonic used for the capacitor computation is at 160
tation accuracy. It appears in Fig. 14 that the modeling results kHz). The differential-mode filter capacitor has been calculated
are close to practical experiments. One interesting point here taking into account the influence of the LISN as presented
is that specific operating conditions can be added very simply in Fig. 15. In this figure, the current source represents the
to the model without having to identify the origin and without high-frequency differential-mode current generated by the
having to derive and to characterize the parameters responsible converter. The capacitor value is computed such that, across
for these nonideal behaviors. In the case of time-domain simu- LISN sensing resistor , the harmonic levels remains below
lations, this characterization would have been required. regulation limits (Class A, 67 dB at 150 kHz and 0-dB margin
CRÉBIER et al.: BOOST-CHOPPER-DERIVED PFC RECTIFIERS: INTEREST AND REALITY 43

TABLE I
MAIN DESIGN PARAMETERS AMONG BOOST-DERIVED RECTIFIERS (V = 230 V, V = 385 V, P = 1 kW, F = 20 kHz)

ever, it has been demonstrated in [22] that, for a PFC rectifier, the
switching frequency is likely to remain limited to several tens of
kilohertz due to the line network pulsating power filtering need
and standard limits frequency range.
Table I provides quantitative design estimations for boost
chopper components. The output stage filter capacitor is not
included in the table because it does not vary much with the
topology or the control strategy used because, at first, it is
filtering the 100–120-Hz network pulsating power flow. The
full-bridge rectifier operating under bipolar control strategy is
used as a reference for PFC quality and HF harmonic contents.
In all cases, with today’s integrated control ICs, the control
part complexity is not any more an issue as long as stability
under closed-loop control can be guaranteed [8].
As can be seen in Table I, where correct PFC is obtained
most of the time, there is a real compromise to find a compro-
mise between maximum input current rating and inductor size.
Global active components stress remains reasonable in all cases
although DCM peak current may require larger silicon dies.
However, input discontinuous current remains much harder to
filter than continuous current. Requirements on inductors are
less important in DCM operation, but the HF ripple current
being much larger, the inductor size (more or less a function
Fig. 15. Computation circuit including LISN and EC55011 standard limits for of ) is not so different in all cases. Looking
class A, B.
at computed filtering requirements, it appears that DCM oper-
ation often requires using excessive capacitor values. This
in this computation). In these computation conditions, the means that the resulting filter will have to include at least an-
impact of the LISN is taken into account. It becomes simple other inductor or a proportional-plus-integral (PI)-type differ-
to compute the capacitor needed to comply with the Class ential-mode filter. Calculations have been realized in order to
A standard limit. The capacitor requirement is computed for estimate the filter inductor size required to change the ca-
several harmonics, and then, the most critical capacitor value pacitor value from 12 to 2.5 F and it is around H.
is selected (the greatest value). This means that the real gain in passive components is really not
Of course, the switching frequency selection is of importance so clear when the differential-mode filter is taken into account.
for conducted EMI generation and high or low switching fre- The same comments are made to interleaved techniques where
quency may not lead to the same comments. It is especially the additional components and DCM operation does not generate
case for HF differential-mode filtering where it can be demon- enough improvements reducing differential mode harmonics. It
strated that the impact of the switching frequency is very impor- is especially the case for the interleaved converters operating
tant when compliance with IEC55011 is mandatory [22]. How- under DCM.
44 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 52, NO. 1, FEBRUARY 2005

For simplicity, DCM operation seems very attractive. How- [5] O. Garcia, J. A. Cobos, R. Priet, and J. Uceda, “Single switch AC/DC
ever, this paper has presented that this type of operation is fi- power factor correction converter valid for both three phase and single
phase application,” in Proc. EPE’97, 1997, pp. 1.188–1.193.
nally costly in term of differential-mode filtering requirements. [6] S. Buso and G. Spiazzi, “A line frequency commutated rectifier com-
It has the advantage to eliminate diode recovery induced losses. plying with IEC 1000-3-2 standard,” IEEE Trans. Ind. Electron., vol.
In any case, under DCM operation, the simplest structures with 47, no. 3, pp. 501–510, Jun. 2000.
[7] J. S. Lai and D. Chen, “Design consideration for power factor correc-
one boost converter seems to be a good choice because they offer tion boost converter operating at the boundary of continuous conduction
real simplicity and correct operation. mode and discontinuous conduction mode,” in Proc. IEEE APEC’93,
1993, pp. 267–273.
In CCM, similar comments can be made; the simplest boost [8] A. Patterson, “Modeling and simulation of a power factor correction cir-
chopper solution (Fig. 2) finally offers the best compromise in cuit based on the Unitrode UC3854 integrated circuit,” in Proc. EPE’95,
terms of performance, differential-mode EMI filtering needs, Seville, Spain, 1995, pp. 2.749–2.753.
[9] M. Kazerani et al., “A novel active current waveshaping technique for
simplicity, and reliability. Although the boost inductor is much solid state input power factor conditioners,” IEEE Trans. Ind. Electron.,
larger, peak current is reduced and, in most cases, differential- vol. 38, no. 1, pp. 72–78, Feb. 1991.
mode filtering can be limited to a small capacitor. [10] J. Spangler et al., “A comparison between hysteretic and fixed fre-
quency boost converters used for power factor correction,” in Proc.
CCM hysteretic and DCM sliding modes generate in both IEEE APEC’93, 1993, pp. 281–286.
cases correct PFC. However, the high-frequency harmonic con- [11] L. Balogh et al., “Power-factor correction with interleaved boost con-
tent is not really improved due to the standard testing condition verters in continuous inductor current mode,” in Proc. IEEE APEC’93,
1993, pp. 168–174.
for compliance [3]. In addition, sliding-mode control generates [12] A. Nabae, H. Nakano, and S. Arai, “Novel sinusoidal converters with
a wide harmonic spectrum which is dependent on load level. high power factor,” in Conf. Rec. IEEE-IAS Annu. Meeting, vol. 2, 1994,
Also, it may lead to variable spectrum contents. Finally, voltage pp. 775–780.
[13] M. Zhang et al., “Single-phase three-level boost power factor correction
regulation also acts on switching frequency harmonic spectrum converter,” in Proc. IEEE APEC’95, 1995, pp. 434–439.
and it is a little harder to implement optimal and minimal fil- [14] C. Larouci et al., “Volume optimization of a PFC flyback structure
tering needs. under electromagnetic compatibility, loss and temperature constraints,”
in Proc. IEEE PESC’02, vol. 3, 2002, pp. 1120–1125.
[15] H. Benqassmi et al., “Comparison between current driven resonant con-
verters used for single stage isolated power factor correction,” IEEE
Trans. Ind. Electron., vol. 47, no. 3, pp. 518–524, Jun. 2000.
[16] L. Simonetti, J. Sebastian, and J. Uceda, “Single-switch three-phase
V. CONCLUSION power factor preregulator under variable switching frequency and dis-
continuous input current,” in Proc. IEEE PESC’93, 1993, pp. 657–662.
[17] E. H. Ismail and R. Erickson, “A single transistor three-phase resonant
This paper has presented a general overview of boost- switch for high quality rectification,” in Proc. IEEE PESC’92, 1992, pp.
chopper-derived PFC rectifiers. It has been underlined that, 1341–1351.
depending on operating conditions, active current shaping may [18] J.-C. Crebier et al., “The three-phase interleaved boost rectifier: Study
and comparison,” in Proc. EPE’99, Lausanne, Switzerland, 1999.
generate large conducted differential-mode HF harmonics. In [19] P. Barbosa et al., “Interleaved three phase boost rectifiers operated in
order to compare the different solutions available, it is impor- the discontinuous conduction mode: Analysis, design considerations,
tant to take into account the HF filtering requirements. This is and experimentation,” IEEE Trans. Power Electron., vol. 16, no. 5, pp.
724–734, Sep. 2001.
possible based on a simple and effective modeling technique [20] J.-C. Crebier et al., “Differential mode current harmonics forecast for
that allows us to forecast input current harmonic content from DCM boost rectifiers design,” in Proc. EPE’99, Lausanne, Switzerland,
the line frequency up to switching frequency multiples. In 1999.
[21] B. Knorr, U. Knorr, L. Zacharias, and H. Puder, SIMPLORER, Reference
particular, it has been shown that DCM generates large HF Manual—Version 3.3. Chemnitz, Germany: SIMEC, 1998.
harmonics requiring large differential-mode filters. Interleaved [22] S. Busquets-Monge et al., “Optimization techniques applied to the
and multilevel techniques are interesting solutions but they design of a boost power factor correction converter,” in Proc. IEEE
PESC’01, Vancouver, BC, Canada, Jun. 2001, pp. 920–926.
increase the amount of components and structure complexity
without giving a real solution for HF harmonic reduction.
Table I with the main design parameters has been provided for a
specific operating point in order to highlight differences among
converters and control strategies in order to help designers in
their converter selections.

Jean-Christophe Crebier received the Bachelors

degree in electrical engineering from the Institut Na-
REFERENCES tional Polytechnique de Grenoble, Grenoble, France,
in 1995, and the Ph.D. degree in power electronics,
[1] Limits for Harmonic Current Emissions (Equipment Input Current 16 < EMI modeling, and analysis within PFCs, from the
A Per Phase), IEC1000-3-2, Aug. 1995. Laboratoire d’Electrotechnique de Grenoble, Institut
[2] Limits and Methods of Measurement of Radio Disturbance Characteris- National Polytechnique de Grenoble, in 1999.
tics of Industrial, Scientific and Medical (ISM) Radio-Frequency Equip- After one year as a Post-Doctoral Researcher at
ment, ISM550011, Jul. 1991. the Center for Power Electronics Systems, Virginia
[3] Specification for Radio Interference Measuring Apparatus and Mea- Polytechnic Institute and State University, Blacks-
surements Methods, CISPR 16, 2nd ed., 1987. burg, doing research in system integration and EMI,
[4] J. Salmon, “Reliable 3-phase PWM boost rectifiers employing series- he joined CNRS France as a full-time Researcher. Currently, he is conducting
connected dual boost converter sub topology,” in Conf. Rec. IEEE-IAS research in the Laboratoire d’Electrotechnique de Grenoble. His research fields
Annu. Meeting, 1994, pp. 781–788. are system integration, power integration, system-on-chip, and packaging.
CRÉBIER et al.: BOOST-CHOPPER-DERIVED PFC RECTIFIERS: INTEREST AND REALITY 45

Bertrand Revol was born in France in 1976. He Jean-Paul Ferrieux received the Ph.D. degree and
received the Engineer degree of ENSIEG from the HDR, both in electrical engineering, from the In-
the Institut National Polytechnique de Grenoble, stitut National Polytechnique de Grenoble, Grenoble,
Grenoble, France, in 2000, and the Ph.D. degree in France, in 1984 and 1989, respectively.
electrical engineering from the University Joseph He is a Professor at the Institut Universitaire
Fourier, Grenoble, France, in 2003. These principal de Technologie 1 de Grenoble (University Joseph
fields of research relate to the EMC of the power Fourier), Grenoble, France. His research activities
electronics converters. take place in the Laboratoire d’Electrotechnique
Since September 2004, he has been an Assistant de Grenoble (LEG). His major field of interest is
Professor at the High Normal School of Cachan power electronics (static converters and resonant
(ENSC), Cachan, France, where he carries out converters, integration). He has published over 80
research in the laboratory of SATIE and works on the improvement of the EMI technical papers in these areas and is the coauthor of a power electronics book,
filtering optimization. Switching-Mode Power Supplies, Resonant Converters (Paris, France: Dunod,
1999).