Power Supply Design Seminar

Power Factor Correction Using the

Buck TopologyEfficiency Benefits
and Practical Design Considerations
Topic Categories:
Design Reviews Functional Circuit Blocks
Specific Power Topologies
Reproduced from
2010 Texas Instruments Power Supply Design Seminar
SEM1900, Topic 4
TI Literature Number: SLUP264
2010, 2011 Texas Instruments Incorporated

Power Seminar topics and online powertraining modules are available at:
power.ti.com/seminars

Power Factor Correction Using the

Buck TopologyEfficiency Benefits
and Practical Design Considerations
Bernard Keogh
AbstrAct

I. IntroductIon

operates on average at only a small percentage of

the full rated load for a large proportion of its
operating life. Therefore, high-efficiency
performance is being mandated across the full
load range, down to 25% load, through the same
ENERGY STAR and other initiatives. Many
systems are also demanding better power-supply
efficiencies in the 0 to 25% load range to meet
stringent system-level power-consumption
requirements in sleep/standby/idle modes. And of
course, power supplies for computing and
consumer applications have and always will be
severely cost-challenged, requiring cost-effective
solutions to deliver both improved PF and
efficiency.
To summarize the design challenges facing
power supply designers:
Power factor (PF)PFC is required for most
designs with Pinput > 75 W to meet EN/JIS-610003-2 and Pinput > 100 W to meet ENERGY STAR.
Power densityHigher power density to fit
into smaller case sizes.
Universal inputAbility to operate across
universal AC-mains voltage range efficiently
and cost-effectively, without power derating.
Light loadHigh-efficiency operation at light
loading, since this is the typical operating point
for the majority of the product life.
StandbyVery low standby or no-load power
consumption, effectively heading toward zero
power consumed at a zero load.

The benefits of improving the power factor

(PF) of AC-to-DC power supplies have been well
documented [1, 2]. The drive toward improving
PF was initially mandated by European standard
EN61000-3-2 [3] (and its forerunner IEC555) in
an effort to reduce the harmonic content of the
current flowing in the distribution network.
Ironically, however, EN61000-3-2 (and its Japanese
counterpart JIS C 61000-3-2) does not mandate PF
limits directly, but rather the amount of harmonic
current allowed at each individual harmonic of the
fundamental-mains frequency up to the 39th. More
recently, ENERGY STAR [4] and European
Commission [5] initiatives have started to add PF
requirements alongside efficiency specifications.
For example, ENERGY STAR EPS 2.0 for external
AC power supplies (such as adapters) now requires
a minimum PF of 0.9 at full load at 115 VAC for
power supplies with input power in the range of
100 W to 250 W.
Improved PF by means of a PFC front-end
must also be achieved with high efficiency of the
PFC stage to maintain the same or better overall
system efficiency [1]. For universal-mains-input
power supplies, this high-efficiency performance
must be maintained across the full AC line range
to meet thermal constraints and avoid power
derating at the AC line extremities. The industry
has recognized that much electronic equipment
Texas Instruments

4-1
1

SLUP264

Topic 4

Although active power factor correction (PFC) is typically accomplished with a boost power topology,
this topic will show that a buck power stage offers significant efficiency advantagesparticularly when
universal line operation is required. Specific design and performance issues such as bus voltage choice,
achievable total harmonic distortion (THD) and power factor (PF), control algorithms, and design
practicalities will be discussed. Design choices are illustrated by a practical buck PFC design example
based on a 90-W, high-density, slimline notebook adapter design (90WHD), demonstrating a >0.9 PF over
a 20- to 90-W load range and over 100 to 230 VAC, and >96% full-load efficiency over 100 to 230 VAC.

Topic 4

II. the need for Pfc

(CrCM), boost converter in continuous conduction

mode (CCM), boost in DCM, and boost in CrCM.
The boost converter is so widely deployed as a
PFC front-end, that the term PFC Front-end has
become almost synonymous with the term boost
PFC front-end in industry. Although not widely
used, the buck converter can and has been used as
a PFC Front-end [8, 9, 10, 11].

Many techniques and topologies can and have

been deployed for PFC. Numerous articles have
been published that survey the options available
for both passive and active PFC [6]. A detailed
discussion of the full breadth of options,
architecture and topologies would be beyond the
scope of this topic, but some of the most commonly
used approaches are briefly mentioned here.
To meet only the requirements of EN610003-2, as detailed in Appendix B, very low THD and
unity PF are not required. Therefore, at low power
levels (up to maybe 200 W), various passive PFC
techniques have been employed to spread the
conduction angle of the current waveform [7].
Although passive PFC can be low cost and easy to
add to existing noncompliant designs (literally as
a bolt-on fix) to meet EN61000-3-2, it does add
considerable size and weight, degrades overall
efficiency, and is limited in the extent of achievable
PF improvement. A design that meets EN610003-2 with a 10% margin on all harmonics could
have a PF as low as 0.76. Recent ENERGY STAR
specs also now require a minimum PF of 0.9 at
full load and 115 V/60 Hz. So a design that just
about passes the requirements of EN61000-3-2
may not meet ENERGY STAR requirements. On
the other hand, as cited in Reference [1], a PF of
0.9 could be achieved by drawing a
square-input current waveform, but such
a current waveform could not meet
EN61000-3-2 because all harmonics
above the 11th would exceed the limits.
There are many widely used active
VAC
PFC circuits that can easily meet both
EN61000-3-2 and ENERGY STAR 0.9
PF. Depending on the power level, input
AC voltage range, required THD/PF performance, and cost constraints, many
possible solutions may be adopted.
Commonly used active PFC circuits
include the fly back converter in
discontinuous conduction mode (DCM),
flyback in critical conduction mode

III. buck Pfc toPology overvIew

And oPerAtIon
The basic circuit of the buck PFC power stage
is shown in Fig. 1. This is a conventional buck
(step-down) converter connected to an AC source
and bridge rectifier. The output bus voltage is set
at a level less than the peak AC voltage at the
lowest line voltage. When the instantaneous AC
input voltage is greater than the output bus voltage,
the PFC stage is forward biased and current can be
drawn from the AC input. When the AC input
voltage falls below the bus voltage level, the diode
bridge rectifiers become reverse biased, and no
power can be drawn from the AC line. There will
be an inherent cross-over distortion in the AC
line current when the buck PFC stage is reverse
biased. But in many applications, this distortion
can be made acceptable with adequate line-current
THD and PF performance. Some putative AC line
Diode Bridge
PFC FET
PFC Inductor
HS
Drive
PFC
Diode

Bulk
Cap

Load

Current
Sense
Control
Circuit

Fig. 1. Buck PFC basic power circuit.

Texas Instruments

4-2
2

SLUP264

VAC

VAC

IAC

Conduction
Angle

Conduction
Angle

Fig. 3. Buck PFC line current and voltage at

230 VAC, with 80-VDC output, THD 16.7%,
passing EN61000-3-2, potential PF 0.986.

Fig. 2. Buck PFC line current and voltage at

100 VAC, with 80-VDC output, THD 44.3%,
passing Japan JIS C 61000-3-2, potential
PF 0.914.

Iv. buck versus boost Pfc

comPArIson

voltage and current waveforms are shown in Fig. 2

at 90 VAC and in Fig. 3 at 230 VAC for a typical
bus voltage level of 80 VDC. At a higher input
voltage, the cross-over distortion becomes much
less significant, with correspondingly lower THD
and higher PF, as one would expect.
In a similar fashion to more conventional boost
PFC stages, the buck PFC is typically controlled
using an outer voltage-control loop to regulate the
bus voltage, with an inner current-control loop to
control the average current shape. The outer loop
adjusts the demand applied to the inner loop to
maintain the regulated average bus voltage in
response to line and load changes. This loop is
typically designed for slow response and low
bandwidth so that it will not respond to outputvoltage ripple, thus maintaining good PF of the
input current. The inner current loop controls the
PWM duty cycle of the PFC MOSFET over the
AC half-cycle in response to the demand from the
voltage loop. This controls the average AC line
current to follow a pseudo-sinusoidal shape and
essentially widening the conduction angle to
deliver the required PF performance. These control
loops are explained in more detail later.

Texas Instruments

Vbus

A. Boost PFC Performance Advantages

and Challenges

There are many reasons why the boost converter has become the topology of choice for a
PFC front-end in many applications. Its many
advantages include the following:
Achieves very low THD, offering probably the
best possible PF.
High output voltagevolumetrically-efficient
energy-storage capacitors, good hold-up.
Low-side boost switch allows easy gate drive
and switch current sense.
Direct forward path from AC input to bulk
storage capacitor eases lightning surge
management.
Wealth of available control ICs and design/
analysis literature to aid designers.
The boost converter also has some limitations
and drawbacks, some of which are simply the
corollary of its advantages:
Output voltage must always be higher than the
instantaneous AC input voltagefor universal
or high-line AC input (up to 264 VAC), bus
voltage must be set at about 400 VDC.
Requires a subsequent high-voltage primary
regulation/isolation stage to step down to
practical voltage levels required by most
electronic loads.
High bus voltage causes higher level of commonmode (CM) EMC noise.
4-3
3

SLUP264

Topic 4

IAC

Vbus

 No inrush limitation at start-up (this is the flipside of the surge advantage listed earlier), so a
potentially dissipative or costly inrush limiting
mechanism is required.
Considerable drop in efficiency at low line
because of the high voltage differential between
the AC input voltage and the bus voltage and the
consequent effect on PFC choke design.

Low output bus voltage allows downstream

stage to make use of low-cost, low-voltage
switches compared to conventional boost PFC
bus voltages of ~400 V.
Of course, as with every circuit topology, the
buck PFC has downsides and limitationsits not
an appropriate solution for all applications.
However, in many cases, the limitations can be
overcome or sufficiently managed to offer a
solution with distinct performance advantages
over the boost PFC.

Topic 4

B. Buck PFC Performance Advantages

and Challenges

Detailed operation of the buck PFC is outlined

in the next section, but the advantages and disadvantages of the buck PFC are listed in order to
compare with those of the boost PFC. Advantages
include:
The high-voltage AC input is immediately
bucked to a lower voltage level by the PFC
front-end which results in:
Easier functional safety spacing in the
subsequent regulation/isolation stage.
Lower downstream voltages, improving robustness and reliability.
Low bus voltage results in lower CM noise.
Lower bus voltage for downstream stage,
allowing more efficient designs, using lowervoltage MOSFETs with better figures of merit.
Lower input-to-output voltage differential across
PFC choke benefits include:
The ability to use a lower inductance value
when compared to a boost PFC.
Input-output differential voltage across the
choke varies in the opposite direction compared
to a boostthus a lower voltage differential at
low line where efficiency is most challenged.
The buck has to work hardest at high line
voltages where currents are much lower,
whereas the boost works hardest at low line
voltages when it boosts the most and currents
are simultaneously the highest.
Easy to control. Control flexibility helps achieve
good THD and PF without the need for an AC
line-sense reference and multiplier.
Inherent free inrush limitation at start-up; no
dedicated inrush limiting mechanism is
required.

Texas Instruments

Disadvantages include:
Inherent AC line current cross-over distortion
limits achievable THD and PF performance.
Requires use of either a high-side drive for the
buck PFC switch or a high-side bus-voltage sense,
depending on configuration used.
No direct path from AC input to bulk-storage
capacitor, complicating surge management.
Lower output voltage results in less efficient
bulk-capacitor energy storage, so bulk capacitors
need to be larger and/or hold-up time is lower.
A higher percentage bus-voltage ripple compared
to boost PFCs requires voltage loops with lower
bandwidths and a slower transient response.
Limitation of available control IC, expertise and
design/analysis literature to aid designers.

v. buck Pfc desIgn APProAch

A. PFC Switching Frequency Selection

As with any AC/DC converter, switching frequency selection is a tricky compromise between
efficiency, size, power density, EMC constraints,
etc. There are always trade-offs between these
conflicting constraints, and the selection of
switching frequency depends on the priority of
these constraints.
Where possible, the lowest practical switching
frequency should be used for best efficiency and
lowest losses. A lower switching frequency will
lead to lower power-device switching loss, core
loss, drive loss, and other AC losses. However,
this goal is often in conflict with power-supply
size and power-density requirements. A lower
switching frequency will also result in much larger

4-4
4

SLUP264

magnetic components to handle the longer

switching periods and maintain the same levels of
peak and AC flux in the core. Typically, the maximum size of the power supply is constrained such
that a practical minimum switching frequency is
dictated.
The maximum allowed switching frequency is
typically limited by EMC requirements. Because
conducted emissions standard EN55022 and similar standards apply from 150 kHz upwards, the
switching frequency is typically limited to perhaps
135 kHz maximum. This ensures that the second
harmonic of the switching frequency is the first
one of interest for conducted emissions. In some
designs, if the switching frequency is kept below
approximately 70 kHz, both the fundamental and
second harmonic will be below the 150-kHz band,
so that the third harmonic of switching frequency
becomes the first one of interest for conducted
emissions.
In practice, typically employed switching frequencies would be near 65 kHz for designs where
highest power density is not required and 100 to
130 kHz for designs requiring higher power
density, at the expense of possibly more EMC
filtering.
In this topic, a nominal switching frequency of
100 kHz is assumed, unless otherwise stated. This
frequency offers good performance in practical
designs over the range of 50 to 400 W and is a
reasonable compromise between efficiency, power
density, and EMC constraints.

VAC_pk

IAC

Conduction
Angle
Half Cycle (180)

greater than the maximum rated AC-line peak

voltage. However, the bus voltage must be
sufficiently lower than the minimum AC-line peak
to allow a reasonable conduction angle.
Fig. 4 clearly illustrates the trade-off between
the level of the bus voltage and the conduction
angle. Setting the bus voltage level too high will
result in a much shorter conduction angle, and
lower bus voltage gives a wider conduction angle.
Conduction angle can be calculated as the number
of degrees out of each 180 half cycle that the
buck PFC is forward biased, during which line
current can be drawn:
Vbus 180
qcond _ deg = 2 cos 1

.
2 V p
rms

(1)

The conduction angle may also be more

conveniently expressed as a percentage of the AC
line half-cycle (or indeed full-cycle) as:

The output bus voltage is probably the most

important design parameter of the buck PFC, and
is the starting point for a design. Fundamentally,
the bus-voltage level must be lower than the
minimum rated AC-line peak voltage, similar to
the requirement that boost bus voltage must be

Vbus 1
qcond(%) = 2 cos 1
.
2 V p
rms

4-5
5

(2)

SLUP264

Topic 4

Fig. 4. Buck PFC line-current conduction angle.

B. Buck PFC Bus-Voltage Selection

Texas Instruments

Vbus

VAC

Topic 4

Conduction Angle (%)

100
Minimum Conduction-Angle Constraint at
Low Line Voltages
As can be seen in Fig. 5, conduction
approaches the full 180 cycle as Vbus is
75
reduced toward zero, but above 100 VDC, the
conduction angle would drop unacceptably
230 VAC
low for low line voltages of 90 to 120 VAC.
50
To remain practical, the bus voltage should
be set low enough to ensure at least a 50%
conduction angle at the lowest line voltage.
25
90 VAC
115 VAC
A conduction angle of less than 50% will
result in very poor PF and very high peak
100 VAC
currents, and will degrade efficiency at lowest
0
0
40
80
120
160
200
line. For a typical minimum-rated line voltage
V
Bus
Voltage,
(VDC)
of 90 VAC, the upper limit on bus voltage
bus
would be 90 VDC to ensure at least a 50%
conduction angle. In practice, it is desirable Fig. 5 Buck PFC conduction angle (in degrees) versus
to maintain a conduction angle reasonably Vbus and AC line.
greater than 50%, so that good PF can be
maintained. However, bus voltage selection
is more typically constrained by the availability
voltage availabilities are summarized in Table 1,
of bus capacitor ratings and by downstream-stage
together with suggested bus voltage settings based
bus voltage requirements.
on appropriate derating.

Available Capacitor Ratings

The choice of bus-voltage level will be strongly
influenced by the choice and availability of suitable voltage ratings for bus capacitors. This is really
the bottom-line constraint. Standard capacitor

Recommended Bus Voltage Range

For universal input AC mains range from 90 to
264 VAC, the bus voltage should be chosen to
allow a minimum 50% conduction angle at 90
VAC; this limits the maximum recommended bus

tAble 1. summAry of stAndArd cAPAcItor voltAge rAtIngs

And PossIble buck Pfc bus voltAge levels
Capacitor Rating

Maximum Vbus
(with 20% Derating)

50 V

40 VDC

Wide conduction angle, ~80% at

90 VAC, very good PF

Low efficiency, very high downstream currents, very poor energy

storage/holdup

63 V

50 VDC

Wide conduction angle, ~74% at

90 VAC, very good PF

Low efficiency, high downstream

currents, poor storage/holdup

100 V

80 VDC

Good conduction angle, ~57% at

90 VAC, good PF, good effi ciency,
moderate downstream currents,
reasonable energy storage/holdup

Somewhat limited PF, < 100 VAC

160 V

120 VDC

Good efficiency, low downstream

currents

Very poor conduction angle, ~22% at

90 VAC, very poor PF, not suitable
for operation below 120 VAC

200 V

160 VDC

Texas Instruments

Advantages

Disadvantages

Not suitable for operation at low line

voltages, only high line voltages of
180 to 264 VAC

4-6
6

SLUP264

Compared to a boost PFC, the buck PFC will

always offer less bulk-energy storage since the bus
voltage is much lower. Thus, the buck PFC will
always offer less hold-up, or require larger bulk
capacitors, or suffer a greater percentage bus
ripple, or all three. Clearly, for a similar hold-up
requirement, the buck PFC will require more
capacitance than the boost PFC, and the size of the
buck PFC capacitance will be greater, taking up
more space inside the power supply. This trade-off
is difficult to quantify because of variations in
capabilities and design rules of different capacitor
manufacturers. The energy storage difference due
to the voltage range of 100 V versus 400 V is
straightforward to calculatethe higher voltage
capacitors can store 16 times more energy for the
same capacitance value. But from reviewing various capacitors in various can sizes for a few
manufacturers, it seems that for a given can size,
the 400-V capacitor will have maybe 8 to 10 times
less capacitance than the 100-V type. So the net
effect on energy storage capability is that the
400-V capacitors can likely store 50 to 100% more
energy in the same physical can size. Or that the
buck PFC bus capacitors would require 50 to 100%
more physical volume for the same energy storage.

C. Buck-PFC Bulk Capacitor

Depending on the application, the bulk capacitance value required could be dictated by voltageripple requirements, required ripple-current rating
of the bulk capacitors, or hold-up requirements.
For notebook adapter and similar applications
where a battery is present in the system, then holdup is typically not a major concern. In this case,
the bulk capacitor choice will be dictated by either
the allowed peak-to-peak bus ripple voltage or by
the required bulk capacitor ripple-current rating.
Unlike the boost PFC, the bus ripple voltage of
the buck PFC will vary with line voltage, and
moreover will typically be a much higher percentage of the DC regulated value. This is because the
power transfer from the AC line only occurs during
the conduction angle, when the instantaneous AC
line voltage is greater than the bus voltage. When
the buck PFC stage is reverse biased during the
AC line cross-over, the bulk capacitor must supply
all of the load current until the AC line voltage
increases above the bus voltage level again. Since
the dead-time interval will be longer at a lower
line voltage, then the bus ripple voltage will be
higher at a lower line voltage. For this reason, the
100-/120-Hz ripple-current rating required for the
buck PFC bulk capacitor could be substantial, and
could be the limiting factor in the choice of bus
capacitance. In addition, the bus capacitors will
need to carry the high-frequency ripple current
from the PFC choke, as well as some additional
ripple current drawn by the second isolation/
regulation stage.

Texas Instruments

Hold-Up Requirements
If hold-up is the limiting design constraint,
then the amount of bulk capacitance required
depends on the hold-up time needed, AC line
voltage, power level, and the amount of bus voltage
headroom provided in the design. The headroom
is the margin allowed between the nominal bus
voltage level and the minimum input voltage rating
of the second stage.
The worst case hold-up of the buck PFC occurs
when the AC line voltage is removed at the trough
of the bus ripple, i.e., at a phase of the AC cycle
where the buck PFC has just become forwardbiased and can start to draw current from the AC
line again. What is interesting is that if the bus
capacitor is increased in order to increase the
energy storage and provide more hold-up, the peakto-peak ripple voltage will also decrease. Thus, the
trough of the bus ripple will increase, further
improving the resultant hold-up performance.

4-7
7

SLUP264

Topic 4

voltage to 90 VDC. From Table 1, using bus capacitors rated at 100-V is a good compromise, so for a
20% typical derating, a bus voltage of approximately 80 VDC would be appropriate. This also
gives good utilization of component ratings,
particularly 100-V rated power MOSFETs for the
downstream stage.
In practice, the actual bus voltage level may be
varied somewhat to suit the specifics of the
downstream isolation/regulation stage and the
most suitable transformer turns ratio. Bus voltages
in the range of 75 to 85 VDC have been used to
good effect in a variety of designs.

However, one of the main efficiency impacts

of designing for large hold-up is the impact on the
second isolation/regulation stage. For greater holdup, this stage must be designed to operate over a
wider input voltage range, with a consequent
impact on efficiency of that stage.
The bus ripple at any operating point can be
estimated as follows:

Topic 4

Vbus(pp) =

Pload

Cbus Vbus

(1 qcond(%) )
2 f AC

for a 90-W load, assuming 96.5% second stage

efficiency:
Cbus

Vbus(pp)
2

(3)

(4)

where Vbus_nom is the nominal PFC bus voltage.

Knowing the minimum allowable bus voltage
at which the regulation stage can keep the output
voltage in regulation (Vbus_min_reg), the hold-up
capability can be estimated as:
t holdup =

2
(Vbus
_ min

2
Vbus
_ min _ reg )

The peak-to-peak bus ripple is estimated by

Eq. 5-3, so the percentage bus ripple can be
expressed as:
Vbus(%) =

Cbus
, (5)
2 Pload

t holdup 2 Pload
2
2
Vbus _ min Vbus
_ min _ reg

Cbus =

2
Cbus Vbus
2 f AC

(8)

Pload (1 qcond(%) )
2
Vbus
Vbus(% ) 2 f AC

(9)

For a typical 90-W adapter using a buck PFC, the

required bus capacitance for a 6% maximum
ripple at 90 VAC/50 Hz (12% total ripple) and
80-VDC bus, assuming 96.5% efficiency of the
second stage, would be:

(6)

For example, in order to achieve a 3-ms holdup

with a nominal bus voltage of 80 VDC, 5%
maximum bus ripple, and 70-VDC minimum busregulation level for the second stage, required
bus capacitance would be calculated as follows
Texas Instruments

Pload (1 qcond(%) )

For designs where holdup is not the design constraint, the bus capacitance required to achieve a
desired percentage bus ripple can be expressed as:

where Pload is the maximum load power drawn

from the bus (accounting for efficiency of the
second regulation/isolation stage). Thus, in order
to achieve a required hold-up time, the required
bus capacitance may be calculated as:
Cbus =

(7)

Bus Ripple Percentage

The allowed peak-to-peak ripple on the buck
PFC bus voltage is usually the most restrictive
constraint that dictates the required bus capacitance size. The bus ripple influences performance
in several ways:
Introduces small line-current phase shift and
degrades PF.
High ripple forces the second isolation/regulation
stage to operate over a wider range, degrading
efficiency performance of the second stage.
Ripple peak voltage must be limited to the
required component stress-derating levels.
The ripple-voltage minimum limits the amount
of bus voltage headroom to the second stage.
The bus ripple is predominantly at twice the AC
line frequency (100/120 Hz), so high ripple will
affect the voltage loop design and degrade either
PF or transient response.

where Pload is the load power drawn (usually by

the second regulation/isolation stage), Cbus is the
bus capacitance, qcond(%) is the conduction angle
at the AC line of interest (as a decimal percentage
of total cyclee.g., a 50% conduction angle
expressed as 0.5), and fAC is the AC line frequency.
The maximum bus ripple can be estimated
from Equation 3 at a maximum load, minimum
AC line voltage/frequency, and minimum bus
capacitance value. Knowing the bus ripple, the
minimum bus level at the trough of the ripple can
be expressed as:
Vbus _ min = Vbus _ nom

90
0.965 = 639 F.
=
2
(80 0.95) 702
0.003 2

Cbus
4-8
8

90
0.57
0.965
= 2
= 690 F.
80 0.12 2 50

(10)

SLUP264

Bus Ripple Impact on PF

Typically, the percentage bus ripple of the
boost PFC is very lowless than 1% to 2%
peak-to-peak. This allows simplification of the
analysis and design equations for the boost PFC
stage. By contrast, the buck PFC percentage ripple
is often much more significant, typically 5%. So
this amount of bus ripple can have a significant
effect on performance.
As can be seen in Fig. 6, one noticeable effect
of the bus ripple voltage is that the point in the AC
line half-cycle at which the buck PFC becomes
forward biased occurs earlier in time, as ripple
increases. Similarly, the point at the end of the
half-cycle at which the buck PFC becomes reverse
biased also occurs earlier in time. The net effect is
that the AC line current is slightly phase advanced
with respect to the AC line voltage. This phase
shift causes a reduction in displacement factor
from unity, and as a result a reduction in PF.

In some cases, boost PFC transient response

has been improved by the use of line-synchronized
sample and hold, or notch filters tuned to 100 Hz
or 120 Hz, or a parallel high-bandwidth voltage
control loop that acts during transient events to
drive the bus voltage outside a defined regulation
window. However, none of these techniques can
resolve the issue of buck PFC load transients that
can occur during the AC cycle dead time. For the
buck PFC, the easiest solution is to increase the
amount and size of the bulk capacitance. This has
the added benefit that the increased bulk capacitor
will reduce bus voltage ripple, increase hold-up
time, and improve PF. However, increased bulk
capacitance does add cost, size, and weight to the
power supply.
As has been shown in previous publications
[14], if the buck PFC stage is operated in discontinuous conduction mode (DCM) (i.e., the inductor
current is allowed to go discontinuous in every
switching cycle), then the control of the PFC stage
is very straightforward. Operation with a fixed duty
cycle over the AC line half-cycle can readily
achieve good PF performance, with a very simple
controller. Operation in DCM at all conditions
means that reverse-recovery switching loss is not
a big issue; thus, the more expensive high-voltage
Schottky type diodes (e.g., SiC devices) can be
avoided in favor of low-cost, ultrafast PFC diodes.
However, if the PFC choke were designed to
maintain DCM operation over the full load and
universal AC-line range, then the peak-to-peak
currents at full load could become excessive,
especially at high line voltages. At higher power
levels in particular, very high peak currents could
flow in the choke, requiring quite a large size

Transient Response
As discussed previously in section 2, because
of the higher percentage bus ripple of the buck
PFC, the voltage control loop must be designed
for lower bandwidth, which will impact the speed
of response to load transients. However, the buckPFC dead time near the AC line cross-over has
a more profound impact on transient response.
Consider the condition where a load transient
occurs during the dead-time intervalespecially at
lower line voltages where the dead-time is longer.
In this case, regardless of the voltage control
bandwidth, the bus-voltage transient drop will
essentially be a function of the bulk capacitance.
For this reason, the worst-case transient response
can not be improved by employing many controlloop-based transient improvement techniques
[12, 13].
Bus-Ripple Induced LineCurrent Phase Shift

Rectified AC Line Voltage

PFC Bus Voltage Ripple
Rectified AC Line Current

Fig. 6. Phase shift of line current due to the effect of bus

ripple voltage.
Texas Instruments

4-9
9

SLUP264

Topic 4

D. Buck PFC Choke Design Considerations

Topic 4

choke to handle these peak currents and maintain

reasonable flux levels.
For moderate power levels (up to 150 to
200 W), the PFC choke can be designed to operate
in DCM at high line voltages over the full load
range, and in CCM at low line voltages, heavy
load conditions. This allows a more compact PFC
choke design. Operating in CCM only at low line
voltages can be seen as a reasonable compromise.
At high line voltages, DCM operation avoids
PFC-diode reverse-recovery issues, still allowing
usage of a low-cost ultrafast rectifier. At low
line voltages, reverse-recovery losses in CCM are
more contained at the lower voltage levels and
are more easily managed. This approach has been
found to offer good performance for power levels
from 50 to 150 W.
For higher-power designs (>200 W), this
approach becomes more difficult. Designing the
PFC choke to ensure DCM at high line becomes
more challenging. The peak currents become very
large. The required core air gap can become very
significant to support the higher peak current.
Turns count then has to increase and the physical
core size has to get much larger. For >200-W
designs, the choke and overall performance could
be made more efficient by designing the PFC
choke for CCM operation even at high line voltages, when load current increases beyond 60% of
the full-load rating. In this case, the increased
switching losses due to reverse recovery are more
than compensated for by reduced choke air-gap
fringing effects, reduced peak, and rms choke
currents.

The buck PFC has previously been analyzed to

establish the limit or boundary of the choke inductance value to ensure CCM operation at low line
voltages (see Reference [9] for full details). In a
similar fashion, the following analysis determines
the value of inductance required to set the boundary
of CCM/DCM at a desired line voltage level for a
given power level.
As shown in Fig. 7, if the AC line current is
initially assumed to be proportional to the AC line
minus the bulk voltage differential, then the peak
AC line current can be estimated. In Fig. 7, input
current follows a sinusoid of amplitude IM, but
with an offset during the conduction angle, and is
zero outside the conduction angle. The angle at
which conduction starts in the AC cycle is indicated
as qstart, and the waveform is assumed to be
symmetrical within the AC half cyclei.e., it is
assumed that qstart = qstop. The current can be
expressed for two conditions:
1. When qstart < q < (p qstart), then
iin(q) = IM (sinq sinqstart) = Iinpk sinq, (11a)
where Iinpk = Im (1 sinqstart) and
V
Vbus
qstart = sin 1
= sin 1 bus

2V
Vinpk
rms

2. When q < qstart or q > (p qstart), then

iin(q) = 0.

VAC_pk x sin

(11b)

VAC_pk

PFC Choke Inductance Value

Vbus
For modest power levels up to approximately 150 W, the PFC choke is typically
IAC_pk
designed for an inductance value that ensures IAC_pk x sin
DCM operation at high line (at all line voltages
>160 VAC), right up to full peak load. With an
appropriate value of inductance to ensure
DCM at full load at the high end of the lineIM
voltage range, operation at low line should
then result in CCM operation over most of the
conduction angle at full load. At least some
start
CCM operation should occur at low line
voltages for loads greater than approximately Fig. 7. Idealized input current and voltage waveforms
of a buck PFC.
half the full-load rating.
Texas Instruments

4-10
10

SLUP264

By integrating the product of voltage and current,

the average input power may be expressed as:
1
=

p2

Pinavg

vin (q ) iin (q) dq.

CCM/DCM. Knowing Iinpk, the required PFC choke

inductance can be calculated to set this CCM/
DCM boundary operating point at any desired line
peak voltage. If the line peak voltage at this
boundary point is designated Vinpk_BP, then the
switch duty cycle at this point (designated DBP)
will be the ratio of Vbus to Vinpk_BP, or:

(12)

Substituting the expressions for current and voltage

and recognizing that there is zero current outside
the conduction angle, this can be written as:

D BP =

2
p
p

qstart

Vinpk sinq I M (sinq sinqstart ) dq

I L _ avg =

and simplified to

I L _ pp =

2 Vinpk I M

Pinavg =

p
p

qstart

sin 2q sinq sinqstart dq.

(1 sinqstart )

sin 2q sinq sinqstart ) dq

start

q 2

(15)

(1 sinqstart )
.
p qstart cos(qstart ) sin(qstart )

4
2
2

(15a)

Iinpk

Assuming that the PFC choke is chosen so that

operation is in CCM at peak of nominal low line
(115 VAC) and in DCM at peak of nominal high
line (230 VAC), then at a particular intermediate
voltage, operation will be at the boundary of

Texas Instruments

(18)

(19)

For example, to ensure DCM operation at >160

VAC for a 100-W buck PFC with an 80-VDC bus,
the maximum allowed choke inductance could be
calculated as follows. Initially, the peak AC line
current can be estimated from Equation (15),
(knowing from Equation (11) that qstart at 160
VAC will be 20.7 for a Vbus of 80 V):

Pinavg p
2 Vinpk

Vbus
(1 D BP ).
L pfc fSW

Vbus
(Vinpk _ BP Vbus )
Vinpk _ BP

Equation (15) can be evaluated more easily by

solving the integral and rewriting the equation as:
Iinpk =

(17)

D BP

1
2 fSW Iinpk

L pfc =

Pinavg p
2 Vinpk

Iinpk

Recognizing that at this boundary operating

point, the average current will be exactly twice the
peak-to-peak current, then the maximum allowed
Lpfc can be calculated to ensure DCM at peak line
voltages above any desired Vinpk_BP:

(14)

Rearranging for IM and substituting into the

equation for Iinpk gives:
Iinpk =

(16)

Also at this operating point, the average and peakto-peak inductor-ripple currents will be:

Vinpk _ BP

90
p
= 0.965
2 2 160

1
p

start

sin q sinq 0.354 dq

(20)
(1 0.354 )

= 0.953 A (peak).

4-11
11

SLUP264

Topic 4

Pinavg =

Vbus

Knowing Iinpk, the critical value of Lpfc to

ensure that the boundary operating point is at
160 VAC can be found (assuming a nominal
switching frequency of 100 kHz):
L pfc =

1
2 100000 0.953

Topic 4

80
2 160 80)
= 95.9 H.
2 160

(21)

This is the maximum value of inductance that

will ensure that the boundary of operation from
CCM to DCM at line peaks occurs at 160 VAC. In
practice, the actual inductance value used can be
varied to suit the actual design, but the previous
analysis shows that Lpfc values larger than those
predicted here will result in CCM operation at
higher line voltages, while smaller Lpfc values will
result in DCM operation down to lower line
voltages.
Compared to the boost PFC, much smaller
values of PFC inductance can typically be used in
the buck, because the voltage differential that needs
to be supported across the choke is lower. Practical
inductance values that have been used in various
designs have ranged from about 150 H for a 50-W
circuit, 80 to 100 H for a 90- to 150-W circuit, and
down to 50 to 60 H for a 300- to 400-W circuit.
PFC Choke Losses
At low line voltages, choke losses are dominated by rms currents and conduction losses,
assuming deep CCM. It is important to choose a
core set that is suitably large, with a sufficiently
high core cross-section or effective area, Ae, to
keep turns count low. A lower turns count results
in lower DC resistance (DCR), less layers in
the winding construction, and less associated
proximity-effect AC resistance (ACR).
At high line voltages, assuming DCM, both
core and copper losses are important. With DCM
operation, peak-to-peak current will be high
twice the peak of AC line current at the CCM/
DCM boundary operating point, and even greater
as the operation becomes more discontinuous. The
choke should be designed so that it can sustain the
highest peak current at the highest line voltage
and highest peak load, and with a good saturation
margin where possible. Besides the extra safety
Texas Instruments

4-12
12

margin and reliability that will be achieved, this

will reduce the peak-to-peak flux swing at high
line 230 VAC and full load. This reduces the nominal flux swing and can greatly reduce core loss
and improve efficiency performance further. The
downside is that this extra saturation margin will
translate to a slightly physically larger core and
choke size, but it is a very useful way to get even
better buck PFC efficiency where available space
allows for a slightly larger choke.
At high line voltages, conduction losses are
also very critical. Although average currents are
lower at high line voltages, since operation can be
very heavily into DCM, the rms current in the
choke can be as significant as at low line. Moreover,
the current at high line voltages will have a much
more significant switching-frequency harmonic
content, even for the same total rms current. So a
winding design with a low AC-resistance (ACR)
factor is also very important. For this reason, the
choke should be designed to use stranded-wire
bundles rather than single-strand solid-core wire.
Ideally, the wire diameter chosen for the stranded
wire should be no greater than the skin depth for
the chosen switching frequency, especially since
the discontinuous current waveform at high line
voltages will contain significant harmonic components up to many multiples of the fundamental
switching frequency. The stranded-wire bundle
should then have a sufficient number of these
strands to handle the required rms current.
This is a rule of thumb that has been found
usefulat one skin depth, current density will
have dropped to 63% of the value on the surface,
so a wire thickness that is twice the skin depth
can be used to ensure a current density of >63%
everywhere. But for the higher-order switching
harmonics, skin depth will decrease with the
square root of frequency, so by limiting wire
diameter to one skin depth, a >63% current density
is ensured for harmonics up to four times the
fundamental switching frequency. This is beneficial
at high line voltages, where the DCM triangular
current waveforms exhibit high peaks at maybe
only 20% duty cycle.
True Litz-wound construction is overly
complex and costly. Simple stranded wire or
bundles of stranded wire are sufficient for most
applications. However, it is vitally important that
SLUP264

E. PFC MOSFET and Diode Choice

loadso the diode duty cycle will be very

significant. Thus, it is important to minimize diode
Vf to control diode losses at high line voltages.
Use of very fast rectifiers, in particular SiC devices,
usually comes at the expense of very high Vf ,
between 2 to 4 V. In the 80% diode duty-cycle
range, this level of Vf can lead to significant extra
loss. The fast recovery benefit is only beneficial if
the design is operated heavily in CCM across the
full line rangeeven then, the Vf versus reverserecovery loss trade-off is difficult to call. Factoring
in the high cost of SiC diodes, the design for DCM
at high line with an ultrafast rectifier becomes
attractive on cost grounds.

F. Voltage-Control Loop Design

The bus voltage-control loop for the buck PFC

is very similar to that of the boost PFC (or any
PFC voltage-control loop). The purpose of the
loop is to sense the bus voltage, compare to a
reference, and create a demand signal for the inner
current-control loop such that the bus voltage is
regulated as AC line voltage and load current vary.
The voltage loop is shown in Fig. 8. Note that the
buck power stage is inverted compared to the
power stage shown in Fig. 1.
This voltage-control loop operates (and is
designed) in exactly the same way as a boost PFC
loop [2]. The voltage error-amplifier feedback
components are chosen to achieve low bandwidth
in the error amplifier; in other words, attenuate the

Compared to the boost PFC, the MOSFET rms

current in the buck PFC is higher; thus the buck
PFC will benefit from the use of a slightly lower
rds(on) MOSFET. This can also benefit the size and
thermal performance in the final design. For
example, in an open-frame design with
airflow, the efficiency drop between 115 and
90 VAC may not be so significant. By
comparison, for an enclosed design such as a
notebook adapter, 90 VAC is typically the VAC
most thermally challenged operating point. At
Bus
this operating point, the PFC power stage
Sense
would benefit from the use of a slightly better
PFC MOSFET, allowing for a smaller and
slimmer final design.
For the PFC diode, use of a conventional
CCM
ultrafast or hyper-fast type is adequate; at low
PWM Vcomp
line voltages, the reverse-recovery losses
+ Vref
Voltage-Loop
associated with CCM operation are
Error Amp
manageable at the lower voltages. At high line
voltages, the average MOSFET duty cycle
will be quite lowperhaps 15 to 25% at full Fig. 8. Buck PFC voltage-control loop.

Load

Texas Instruments

4-13
13

SLUP264

Topic 4

the stranded wire is tightly twisted. This is to

ensure that all strands spend an equal and uniform
amount of time throughout the cross-section of the
bundle; otherwise, some strands in the bundle may
be forced to conduct more current than others,
which can dramatically increase ACR.
The final constraint to be aware of is the effect
of air-gap fringing flux, which can induce very
high losses in the copper adjacent to the air gap.
This effect can be very severe if the air gap is
large, as is sometimes necessary to increase the
saturation capability of the inductor. Again, it may
prove beneficial to consider a slightly larger core
Ae with a slightly smaller air gap.
In some designs, it can help to construct the
choke with a deliberately fattened bobbin-wall
thickness around the limb with the air gap
(normally the center leg of an E-core construction).
This increase in bobbin-wall thickness serves to
physically remove the copper winding from the
fringing flux field of the air gap, reducing induced
eddy currents in the windings and improving choke
efficiency.
Topic 5 in this seminar series covers all of
these areas in greater detail.

100-/120-Hz bus ripple at the voltage sense input

so that the Vcomp signal fed to the inner-current
loop is as close to DC as possible. As already
outlined in the Bus Ripple Percentage section,
the gain of the error amplifier for the buck PFC
must be made considerably lower than the
equivalent boost circuit, given the higher
percentage bus ripple present.

Topic 4

G. Buck PFC Current-Control Schemes

As outlined in more detail in Appendix B,

given the inherent dead time of the buck PFC and
the resulting crossover distortion of the AC line
current drawn, there is an upper limit on the
achievable PF. Further degradation of PF will
occur because of circuit practicalities such as buscapacitor ripple, EMC film-capacitor current, etc.
But how should the buck PFC power stage be
controlled to achieve the PFC goal?
As previously analyzed [14], when operated in
DCM, the buck PFC can be relatively easily
controlled to give good PF. For example, a fixed
duty cycle may be employed over the entire
AC cycle, giving reasonable PF performance.
Clearly, a control circuit will be required to adjust
this fixed duty cycle in response to load demand to
keep the output voltage in regulation. In its simplest
form, this could be a basic voltage-mode controller
that generates a DC error signal that is compared to
a sawtooth ramp to produce the required duty
cycle for the buck PFC MOSFET. This approach
was simulated, with the resulting waveforms
shown in Fig. 9. At high line, where operation is
always in DCM, the current is well controlled and
pseudo-sinusoidal in shape, with a good PF of
>0.95. However, at low line, operation transitions
from DCM to CCM for a portion of the AC cycle
around the voltage peaks, depending on load level.
Under these conditions, the current can easily run
away because of the much higher loop gain in
CCM; i.e., a small change in duty cycle can
produce a large change in inductor current. This

Texas Instruments

4-14
14

peaking of the current greatly degrades the PF, and

in extreme cases can cause the buck PFC stage to
behave virtually like a diode rectifier with a
capacitive filter. In applications with a narrow
input range and suitable protection mechanisms
against CCM operation and current runaway, this
control approach may be suitable.
Because of the wide operating range required
for universal range of AC mains, it is difficult to
design the buck PFC to always operate in DCM and
maintain good efficiency with a compact magnetic
size. Moreover, if employing only voltage-mode

Line Voltage
(200 V/div)

Line Current
(1 A/div)

Time (5 ms/div)

a. Operation is always in DCM at high-line

voltages (230 VAC).

Line Voltage
(60 V/div)

Line Current
(2 A/div)
Time (5 ms/div)

b. Operation changes to CCM during current

peaks at low-line voltages (100 VAC).
Fig. 9. Simulated waveforms for buck PFC
operated with only voltage-mode control at
constant duty cycle.

SLUP264

Texas Instruments

4-15
15

sharp rise/fall of the current trapezoid can be

problematic in meeting EN61000-3-2.
One good control approach is a mix of peak
current-mode and voltage-mode control. Currentmode control is required to avoid the current runaway at the peak of AC line when in CCM. A
voltage-mode slope-compensation ramp can be
very usefully added to the current-sense signal to

Line Voltage
(200 V/div)

Line Current
(1 A/div)

Time (5 ms/div)

a. Operation is always in DCM at high-line

voltages (230 VAC).

Line Voltage
(50 V/div)

Line Current
(1 A/div)

Time (5 ms/div)

b. Operation is in CCM at low-line voltages

(100 VAC) during peaks.
Fig. 10. Simulated waveforms for buck PFC
operated with only current-mode control.

SLUP264

Topic 4

control, then with transient conditions such as

reduced AC input voltage (e.g., brownout) or output overload, operation may transition into CCM.
The buck PFC has also been analyzed in detail
in CCM [9, 14]. Current-mode control limits the
current in CCM and prevents the runaway issues
cited earlier. Using current-mode control can also
allow the use of clamped-current techniques,
especially at low line voltages where currents
increase [8, 10]. Clamped-current waveforms are
deliberately non-sinusoidal, approaching a trapezoidal shape, to limit the peak current at the peak
of the AC line voltage. Using clamped-current
approaches can be advantageous in reducing the
peak AC line current that is drawn at lower line
voltages and higher power. This is particularly so
at minimum line voltages, where the normally
high peak currents of a sinusoidal or pseudosinusoidal waveform lead to much higher peak
and rms currents in the PFC MOSFET and choke.
The clamped-current buck PFC can still achieve
good PF, while also reducing losses and improving
efficiency performance at low line voltages.
Fig. 10 illustrates simulated waveforms at both
high and low line when the buck PFC is operated
with only current-mode control. It can be seen that
the current peaking effects at low line during CCM
are eliminated, and the current waveform becomes
more trapezoidal in shape.
If the CCM buck PFC is operated heavily in
current-mode control (with little or no added slope
compensation ramp), then it is possible to practically achieve a line current shape that is very close
to the theoretical clamped-current trapezoid [8].
However, very high current-loop gain can lead
to instabilities and oscillations in the line-current
waveform. Additionally, the harmonic content of
the severely clamped-current waveform with a

Load

VAC
Bulk
Sense

Current
Sense

Ramp

Osc

+
+

S Q

Vcomp

R Q

Current-Loop
Comparator

Voltage-Loop +
Error Amp

Vref

Topic 4

Fig. 11. Buck PFC current-control loop.

optimize this mix of voltage- and current-mode
control. The added slope-compensation ramp can
be used to adjust the shape of the AC line current
to trade off PF performance against the harmonics
limit of EN/JIS-61000-3-2. This control approach
is illustrated in Fig. 11. Simulated waveforms
achieved with this approach are shown in Fig. 12.
Fig. 13 illustrates actual practical circuit
waveforms that were observed when only currentmode control (Fig. 13a) and then only voltagemode control (Fig. 13b) were used. The current
peaking effects are very pronounced in Fig. 13b.
The current loop of the buck PFC has been
successfully implemented with a variety of peak
current-mode PWM controllers, including the
industry-standard 3843, and more recently using
a dedicated buck PFC controller UCC29910 that
greatly simplifies the control-loop design [15].

H. EMC Considerations

Because a buck converter places the active

control switch in the input path, with the inductor
in series with the output, it is commonly assumed
that the pulsating input current will be problematic
from an EMC perspective. In practice, however,
the pulsating input current is no worse than it
would be for a similar power level with other
topologies, such as DCM or transition-mode

Texas Instruments

4-16
16

Line Voltage
(200 V/div)

Line Current
(1 A/div)

Time (5 ms/div)

a. Mixed current- and voltage-control modes at

high-line voltages (230 VAC).

Line Voltage
(50 V/div)
Line Current
(1 A/div)

Time (5 ms/div)

b. Mixed current- and voltage-control modes at

low-line voltages (100 VAC).
Fig. 12. Simulated waveforms for buck PFC
operated in mixed control modes.
SLUP264

flyback, or boost. In those cases, the inductor current

is allowed to go fully discontinuous, so there is also
a high pulsating current in the input path. This
pulsating current is typically filtered by a low-cost
film capacitor as part of the EMC filter. So for the

buck PFC, the pulsating input current and filtering

requirements are no worse than the transition-mode
boost. Fig. 14 shows the schematic of the required
EMC filter for a 100-W, buck-PFC front-end to
meet EN55022 Class B with an 8-dB pass margin.

AC Line
Current
(1 A/div)
12

PFC Bus Voltage,

AC-Coupled
(2 V/div)
Time (2 ms/div)

a. 90-VAC, full-load, clamped-current type waveform

high current-sense signal, low slope compensation.

Topic 4

AC Line Current
(1 A/div)

Time (10 ms/div)

b. 90-VAC, half-load, peaky-current waveformlow

current-sense signal, high slope compensation.
Fig. 13. Line-current peaking and runaway effect when
a buck PFC is operated in CCM with only voltagemode control or with very low current-loop gain.
AC Line
Fuse T 3.15 A
250 V

CommonMode
Inductor
5 mH

X2
33 nF

1.5 M

Sidactor
130 V

1.5 M

Neutral

Varistor
175 V

+
~

DifferentialMode Inductor
100 H

Film Cap
220 nF

Film Cap
1 F

DifferentialMode Inductor
100 H

Fig. 14. Buck PFC 100-W-rated EMC filter.

Texas Instruments

4-17
17

SLUP264

Transition-Mode
Boost-PFC
90-W EMC Filter

Buck-PFC 90-W
EMC Filter
Film Cap
220 nF
(underside)

Differential-Mode
Inductor
300 H

Differential-Mode
Inductor
100 H (each)

Film Cap
220 nF

Common-Mode
Inductor
5 mH (each side)

Film Cap
470 nF
Common-Mode
Inductor
30 H
(each side)

X2 Cap
330 nF

X2 Cap
330 nF

Film Cap
1 F

Common-Mode
Inductor
15 mH
(each side)

Topic 4

Fig. 15. Comparison of EMC filters for buck versus transition-mode boost.
Fig. 15 illustrates the size of this filter as
implemented on a low-profile (16-mm total
height), 90-W notebook adapter. For comparison,
the required filter for a 90-W transition-mode
boost circuit in similar profile is shown alongside.
Note that the EMC filters are of similar size and
complexity. The filter size for the buck PFC is
approximately 15 cm2 (2.33 in2 ), compared to
17 cm2 (2.64 in2 ) for the boost circuit (adjusting
for the looser packing density of the boost filter to
give a fairer comparison).
The conducted EMC performance of this 90-W
buck PFC stage is demonstrated in section VII.
This design passes both quasi-peak (QP) and
average (AV) limits of EN55022 Class-B with
good margin in the worst-case configuration with
an earthed load.

drive the power-MOSFET gate. Alternatively, the

power circuit could be inverted, as in Fig. 17, so
that the power MOSFET is on the low side and
thus easier to drive. This then references the output
voltage to the positive rail of the input voltage, so
that the output voltage is effectively floating. In
this case a high-side, level-shifting voltage-sense
circuit is required to generate a ground-referenced
sense signal for the control circuit. With either
power-stage configuration, a high-side element
will always be required. Depending on the
application specifics, one or the other configuration will often be preferable. In practice, the
choice of power-stage polarity will depend on
required cost, performance, and the nature of the
downstream stage deployed.

vI. buck Pfc desIgn PrActIcAlItIes

And chAllenges

The conventional buck configuration with

high-side MOSFET has advantages because the
output bus voltage is referenced to the same ground
as the control circuit. This allows for straightforward
output-voltage sensing for control, using a simple
resistor-divider network. A common ground can
also be maintained between the PFC stage and

A. Buck PFC Power Stage Configuration

As shown in Fig. 16, the conventional buck

power stage places the power switch in the high
side, as is common in DC/DC converters. This
necessitates the use of a high-side drive circuit to

Texas Instruments

4-18
18

B. Conventional Buck Configuration

SLUP264

PFC MOSFET
Vin

+
HighSide
Driver

PFC Inductor
PFC Diode

+
Load

Bulk Cap

Control
Circuit

Fig. 16. Conventional buck power stagehigh-side MOSFET.

Vin +

Bulk
Cap

PFC Inductor

Load

Topic 4

PFC
Diode

High-Side
Bus
Sense
PFC
MOSFET

Control
Circuit

Fig. 17. Inverted buck power stagelow-side MOSFET.

downstream isolation/regulation stage, allowing a
common control circuit to be deployed for both
stages. This common ground also greatly simplifies
voltage and current sensing of both stages by a
shared combo control circuit. It simplifies handshaking and communications between PFC stage
and isolation/regulation stage controllers, if using
separate control circuits for each stage.
However, this power circuit configuration
requires a floating high-side driver for the PFC
MOSFET. This is a significant pointit requires
the use of either a gate-drive transformer or highside bootstrapped driver integrated circuit (IC). In
either case, the design challenges are not trivial. A
high-side gate-drive transformer may initially

Texas Instruments

4-19
19

seem attractive. However, the practical design

constraints can be challenging:
Functional isolation is required to withstand up
to 400-V isolation at high line voltages.
Transformer parasitics (especially capacitance)
can cause common-mode difficulties with the
drive signal.
There is a trade-off between transformer size/
cost versus performance.
The maximum drive duty cycle will need to be
limited (to maybe 80% or less) to allow the
transformer to reset during the off time. This
limits the buck PFC PF performance at the
lowest AC line voltage.

SLUP264

Diode Bridge
PFC
Diode
VAC

Bulk
Cap Load

PFC Inductor

PFC
MOSFET
Current
Sense

Bulk
Sense

Control
Circuit

Drive

Topic 4

Fig. 18. Buck PFC inverted configuration with low-side

PFC MOSFET.
A high-side boot-strapped driver IC is another
possible solution to implement a high-side gate
drive. This solution overcomes many of the drive
transformer limitations, but also comes with
another design challenge. Because this is a singleended drive, there is a challenge in establishing
the high-side drive rail during startup and in
maintaining the rail during extended periods with
no switching, such as heavy load-dump transients
or light-load burst-mode operation.

as detailed below, with varying levels of performance and cost.

The second drawback with this configuration
is that the buck PFC stage control will not share
the same ground as the downstream isolation/
regulation stage control. So common-ground
combo controllers could not be used, and any
signaling (such as fault communications) between
the two stages will also need to be level shifted in
a similar manner.

C. Inverted Buck Configuration

Level-Shifting V-to-I Converter

The high-side bus voltage of the inverted buck
PFC may be sensed in many ways. As there is no
requirement for safety isolation, just functional
isolation, the task is more straightforward than
sensing across an isolation barrier. Three possible
schemes are briefly outlined here, but other
approaches are possible.
A high-voltage PNP transistor may be connected across the high-side bus voltage as in
Fig. 19 to provide a collector current that varies in
proportion to the bus voltage. This current can be
converted to a sense voltage by feeding into a
ground-referenced resistorwith this connection,
the voltage sense is also referenced to ground.
This circuit is low cost, with only a few components
more than a typical resistor divider. The PNP transistor needs to be rated for at least 500 V for high
line-voltage operation, but these are available in
suitable small surface-mount and leaded packages.

The buck PFC power stage may be inverted,

and configured in an upside-down fashion as
shown in Fig. 18, with the PFC MOSFET on
the low side. In this case, the PFC MOSFET
is referenced to the same ground as the control
circuit, so the MOSFET is simply driven directly
from the controller. Current sense for the for
the low-side switch is also available very easily,
without complications of negative polarity and
the need for pull-up. Because of these advantages,
this power-circuit configuration is deployed more
often than the conventional one with a high-side
MOSFET.
The main drawback of the inverted buck is that
the output bus voltage is floating. The difficulty
with this configuration is that the high-side referenced bus voltage must be sensed and levelshifted to the ground-referenced control circuit.
There are several ways to implement this function,
Texas Instruments

4-20
20

SLUP264

The sense voltage across Rsense will be related

to the bus voltage as in Equation 22 below,
assuming the bus voltage is sufficient to forward
bias Q9.

Vsense

Cbus

D1
Vin

R15
1 M

R17
330 k

LPFC

Q1

Vsense

Q9
Vbus

Rsense

Fig. 19. High-side bus voltage-sense circuit using

high-voltage PNP V-to-I converter.
Vbus+
Cbus
Vin

LPFC

Vbus
R1

R2

Vbias

Vsense

R3

Fig. 20. Differential amplifier for high-side bus voltage

sense and level shift.

R17
Vbus Veb 1 +

R16
=
R15
1 R17
R17
1 +

1 +
+
R sense h FE R16 R sense h FE

(22)

where Veb is the PNP emitter-base forward voltage and hFE is the DC
current gain of Q9. This equation can also be rearranged as follows to
calculate the regulated bus voltage as a function of error-amplifier
reference (Vref) and chosen resistor values:

V
1
Vbus = R15 ref 1 +
R sense h FE

Texas Instruments

R17 Vref R17

+ Veb 1 +
+
R16

R sense h FE

4-21
21

(23)

SLUP264

Topic 4

Differential Amplifier
A differential amplifier may also be used to
sense the high-side bus voltage, as shown in
Fig. 20. The circuit uses two resistor chains
connected to either end of the bulk capacitor
(Cbus) to sense and scale the voltage difference
across the bulk capacitor. The op-amp output
can be readily referenced to ground by a suitable
connection of R3 as shown, giving the required
low-side voltage-sense signal, Vsense. However,
a large number of resistors are required. Since
two resistor chains are connected to highvoltage points, larger sized 1206 or 0805
resistors are required for voltage rating. The
circuit requires a reasonably good op amp with
rail-to-rail input (RRI) capability. Given the
high common-mode signal content, good resistor matching is essential. Resistors with a tight
tolerance of 0.1% or better and a low temperature coefficient are required.
This circuit is more expensive than the previous given the constraints on the op amp and
resistors. The power dissipation of the two
resistor chains will also likely be higher than
the previous circuit. However, the performance
of the circuit will be superior, yielding better
accuracy, repeatability, linearity, and temperature stability.

R16
1 M

Vbus+

advantages, but each one also has inherent drawbacks. Neither configuration is a clear winner over
the otherthe choice of configuration to use in a
particular design depends on many system-level
constraints and considerations.

D. Inrush Versus Surge Immunity

Unlike the boost or flyback topology, there is

no direct connection from the input to the output
capacitor in a buck topology. For this very reason,
the buck stage exhibits little or no inrush sparking
at the initial application of AC power. The bulk
capacitor can be very easily soft-started by appropriate drive control of the buck-PFC MOSFET.
Inrush control is achieved for free, without the
overhead and complications of a dedicated inrush
limiter. And because there is a well-controlled
inrush current, the current-sense resistor does not
need a high I2t rating, allowing the use of a small
surface-mount thick-film part.
Yet, this inrush advantage is also a disadvantage when considering power line disturbances
(e.g., lightning surges as defined by EN610004-5). The surge energy can only flow into the
limited capacitance of the EMC filter. Consequently, the buck PFC requires dedicated surgemanagement components. The previous EMC
filter circuit in Fig. 14 is an example of a proven

Summary of Power-Stage Configuration

Table 2 compares the advantages and disadvantages of the conventional and inverted buck
power stage. Both configurations have distinct

Voltage-Loop
Error Amp

VAC

Vref

CCM
PWM

Vcomp

Load

Verror

Topic 4

High-Side Error Amp Plus Optocoupler

In all applications, the scaled bus-voltage sense
signal is used for the voltage regulation loopthe
signal is compared to a reference by an error
amplifier to generate an error signal. Fig. 21
illustrates the option of placing the voltage reference and error amplifier directly across the floating
high-side bulk capacitor. The output of the error
amplifier is level-shifted to the ground-referenced
controller by a low-cost optocoupler. Since the
optocoupler is only used to bridge high-side to
low-side functional isolation rather than primaryto-secondary galvanic isolation, a low-cost device
can be used.
The circuit does not suffer the accuracy or
component-tolerance constraints of the previous
level-shifting schemes because the actual bus
voltage is not level-shifted, just the error signal.
However, the bias rail used to drive the feedback
optocoupler must be a regulated DC railany
100-/120-Hz ripple component on this rail will
introduce a 100-/120-Hz ripple into the inner
current loop that will reduce the PF.

Vbias

Optocoupler

Fig. 21. High-side bus sense with error amp plus reference directly across
the bus capacitor and an optocoupler for high- to low-side level shift.
Texas Instruments

4-22
22

SLUP264

tAble 2. comPArIson of Power-stAge confIgurAtIon AdvAntAges And dIsAdvAntAges

Conventional
(High-side MOSFET)

Inverted
(Low-side MOSFET)

Advantages
Allows use of common ground throughout
system for PFC stage and regulation stage

Requires return-path current sense, inverted

signal requiring pull-up offset or inversion

Easy ground-referenced bus voltage sense

High-side driver adds cost/complexity

Compatible with use of combo control IC

for both stages

Transformer drive CM issues, Dmax limit for

magnetic reset

Choice of transformer or silicon driver for

high-side MOSFET drive

IC drive issues with generation and maintenance of floating supply due to single-ended
topology

Simple low-cost low-side MOSFET drive

Requires high- to low-side, level-shifting

bus-voltage sense

Ground-referenced low-side current sense

No common ground with PFC stage and

downstream regulation stage

High Dmax capability

Not compatible with combo controller,

requires separate PFC and downstream
control ICs

surge-management approach that allows the buck

PFC to safely pass a 1-kV differential-mode surge
per EN61000-4-5. This circuit uses low-voltage
varistors in series with a sidactor switcha diactype device that is open until the voltage across it
exceeds a beak-over level, at which point it fires
like a thyristor. Once the sidactor current exceeds
a latching current level, it remains latched on until
the current drops below a holding level, again just
like a thyristor. The sidactor switch connects the
varistors across the line whenever there is a
sufficiently high surge voltage to exceed the
sidactor break-over voltage, and keeps them
switched out otherwise. The important criterion
about varistor choice is that voltage clamping
occurs below 600 V under surge conditions (to
allow usage of standard 600-V rated MOSFETs
and diodes) and to ensure that the varistors survive
potential worst-case conduction for an AC halfcycle, assuming surge initiation at an AC zerocrossing.

E. Bus Overvoltage Faults

For any PFC topology, in the event of an

open-feedback situation, the control loop will
advance the duty cycle toward the maximum
value, potentially generating an overvoltage condition on the bus. Independent bus-overvoltage
protection (OVP) is required to prevent catastrophic damage to bulk capacitors. In particular, the
main PFC MOSFET is a potential issue for a buck
Texas Instruments

Disadvantages

4-23
23

PFC. If the MOSFET failure is a short from drainsource, the bulk capacitor will be charged towards
the peak of the AC line voltage. OVP control
action that attempts to reduce duty cycle or inhibit
switching will be ineffective in limiting or terminating the overvoltage event. Sustained overvoltage of the bulk capacitor can lead to venting of
the capacitor and electrolyte leakage. This would
be a safety agency concern, so steps must be taken
to protect against such an event.
Many possible OVP schemes could be used.
The challenge is to implement a circuit that is
simple and low cost. The circuit shown in Fig. 22
is a very simple two-component solution. A small
resistor fuse is placed in series with the bulk
Vbus+
Cbus

Vin +

LPFC

Rfuse

Z1
Sidactor
P1100SCL

Vbus

Fig. 22. Buck PFC bus overvoltage protection

using crowbar device Z1.
SLUP264

Topic 4

Power-Stage Configuration

Vac

EMC
Filter

Downstream
Isolation
and
Regulation Stage

EMC
Filter

Vout

High-Side
Bulk
Sense

HV
Start-up
MOSFET

Driver
3.0 V
4

11

10

Fault
Feedback

13

Bias
Winding

LINE VDD VDD PFC

SENSE
DRV
3
NC
CS

9 BIAS

SENSE

UCC29910PW

12 BIAS

CS

VBULK

CTRL

5
2

NC REF IN VSS FAULT

14

Topic 4

Ref

Fig. 23. Buck PFC front end as used in the 90WHD reference design to feed downstream
isolation/regulation stage.
capacitors; this branch is shunted by a sidactor
switch element. Under normal operation, the
voltage on the bus will be below the break-over
voltage of the sidactor, so it will be off. If the bus
voltage increases above the break-over voltage,
the sidactor will latch on, drawing a very large
pulse of current from the bulk capacitance and
causing the series resistor to fuse open, thereby
disconnecting the electrolytic capacitors from the
bus voltage to prevent venting.

a simplified block diagram of the buck PFC powerstage topology as deployed in this design based on
a dedicated buck PFC controller, UCC29910. The
90WHD design specification is summarized in
Table 3, with a photo of the final design shown in
Fig. 24.

vII. buck Pfc PerformAnce revIew:

90-w AdAPter reference desIgn
The buck PFC has been deployed as the PFC
front-end in a 90-W, high-density, slimline
notebook-adapter reference design (referred to as
90WHD). This two-stage design uses a halfbridge isolation/regulation stage to down-convert
the buck PFC bus voltage to an isolated and
regulated 19.5-V output voltage [16]. Fig. 23 shows
Texas Instruments

4-24
24

Fig. 24. 90WHD slimline, 90-W, buck-PFC adapter.

SLUP264

tAble 3. summAry of 90whd slImlIne AdAPter

reference desIgn sPecIfIcAtIon
90/110 W

Adapter dimensions (L x W x H)

90 x 67.5 x 16 mm (3.54 x 2.66 x 0.63 in)

EMI-conducted and radiated

EN55022 Class B

Output voltage, Vo

19.5 V 1 V

Output voltage total regulation

18.5 V to 20.5 V (At end of 6-ft cord, 16 AWG)

Output current

0 to 4.62 A (continuous)

Peak output current

5.62 A for 4 s (max)

Vin

90 to 264 VAC (47 to 63 Hz)

Efficiency (at 100% load, minimum line)

92.50% min (not including cable losses)

ENERGY STAR

Efficiency >89% average at 25%, 50%, 75% and 100% load;

no-load power draw < 0.5 W; efficiency at 0.5-W load > 50%

PF at 115 VAC 60 Hz (90 W)

0.94 (typ)

PF at 230 VAC 50 Hz (90 W)

0.96 (typ)

Output ripple/noise

250 mV (typ), 400 mV (max)

Output transient load step

50% of rated load

Output transient voltage undershoot

1.5 V maximum

Overcurrent inception level

105% to 135% of peak load

Short-circuit protection

Unlimited, unconditional, latching. 10-ms activation delay

Overvoltage protection

26 V (max)

Turn on AC input to Vo

1 s (typ)

Vo rise time

40 ms (typ) (10% to 90%)

Vo fall time

100 ms (max)

A. Buck PFC Front-End Performance

show the efficiency of only the PFC stage, as both

line voltage and load vary. Efficiency performance
is pretty consistent over the full universal mains
range and over the majority of the load range.

Figs. 25 and Fig. 26 demonstrate the efficiency

performance of the buck PFC front-end stage as
used on the 90WHD adapter design. These curves

98

97.5

96.5
96.0

90
86
82

95.5
95.0

Vin = 115 VAC

94

95-W Load
Efficiency (%)

Efficiency (%)

97.0

Topic 4

Output power (continuous/peak)

78
90

110

130

150

170

190

210

230

250

270

AC Line Voltage (Vrms)

10

20

30

40

50

60

70

80

90

Output Power (W)

Fig. 26. Buck PFC front-end efficiency versus

load at 115 VAC.

Fig. 25. Buck PFC front-end full-load efficiency

versus AC line.
Texas Instruments

4-25
25

SLUP264

Fig. 27 demonstrates the line-current wave

shape achieved at full load at various AC line
settings, together with the actual THD and PF in
each case. Fig. 28 summarizes the PF performance
versus line and load.

B. Overall Two-Stage Adapter

Performance

The following data illustrates the overall twostage 90WHD adapter performance for the buck
PFC front-end plus the downstream half-bridge

AC Line
(50 V/div)

AC Line
(50 V/div)

Topic 4

Load Current
(1 A/div)

Load Current
(1 A/div)

Time (2 ms/div)

Time (2 ms/div)

a. 90 VAC/60 Hz at full load with

THD = 50.9% and PF = 0.889.

c. 115 VAC/60 Hz at full load with

THD = 34.1% and PF = 0.944.

AC Line
(50 V/div)

AC Line
(100 V/div)

Load Current
(1 A/div)

Load Current
(0.5 A/div)

Time (2 ms/div)

Time (2 ms/div)

b. 100 VAC/60 Hz at full load with

THD = 47.8% and PF = 0.902.

d. 230 VAC/50Hz at full load with

THD = 16.7% and PF = 0.966.

Fig. 27. AC line-current wave-shapes at full load for the 90WHD adapter.

Texas Instruments

4-26
26

SLUP264

tAble 4. summAry of 90whd AverAge effIcIency

At Pcb level, not IncludIng outPut cAble loss

isolation/regulation stage. Fig. 29

and Table 4 illustrate the efficiency
performance up to full load, and
Fig. 30 shows the efficiency over
only the light load range (up to 10
W). Table 5 provides the no-load/
standby power consumption of the
adapter and it is well below the
ENERGY STAR limit of 0.5 W.
Fig. 31 illustrates typical conductedEMC performance and shows a
good pass margin against EN55022
level B.

Efficiency

Input
Voltage

25%

50%

75%

100%

Average

115 VAC

91.00%

93.00%

93.50%

93.10%

92.65%

230 VAC

88.55%

91.85%

92.65%

92.84%

91.47%

tAble 5. summAry of 90whd

no-loAd Power consumPtIon
Input Voltage

No-Load Input Power

Spec Limit

115 V

0.160 W

0.5 W

230 V

0.230 W

0.5 W

90

100

90
100 VAC
85

75

230 VAC

75
70

Future Industry Target

65
60

230 VAC

80

80

Topic 4

Efficiency (%)

Efficiency (%)

95

115 VAC

85

115 VAC

55
50
0

10

20

30

40

50

60

70

80

90

3
4
5
6
Output Power (W)

Output Power (W)

Fig. 28. Buck PFC PF versus line voltage and load

for the 90WHD adapter.

10

Fig. 30. Overall light-load (<10 W) efficiency

performance.
70

Quasi-Peak Level B Limit

94

60

Average Level B Limit

90

50

Amplitude (dBV)

Efficiency (%)

80

90 VAC

86

100 VAC
115 VAC

82

230 VAC

78

264 VAC

30
20
10
0

74
70

40

10
0

10

20

30 40 50 60
Output Power (W)

70

80

Fig. 29. Overall efficiency performance versus

AC line and load.

Texas Instruments

20
0.15

90

Peak Scan
Average Scan
1

Frequency (MHz)

10

30

Fig. 31. 90WHD conducted EMC at 230 VAC with

earthed load (worst-case operating point), showing
minimum pass margin of 8 dB.
4-27
27

SLUP264

Topic 4

vIII. conclusIon
The buck PFC topology can demonstrate
excellent full-load efficiency of better than 96%
at the lowest line voltage (90 VAC), typically the
performance bottleneck operating point for boost
PFC designs over the universal mains range. This
level of high-efficiency performance is maintained
over the full universal line-voltage range. This
enables the design of very small format, low-profile
AC/DC converters such as notebook adapters with
power densities in excess of 16 W/in3.
Furthermore, the topology enables high
efficiency over the full load range. Through use of
the additional smart-burst mode controls provided
by the UCC29910 buck PFC controller and
UCC29900 ICC controller, overall two-stage
efficiency performance can be further enhanced
achieving light-load efficiency of >80% at a mere
5% of full loading.
The inherent line-current cross-over distortion
of the buck PFC topology does limit achievable
PFclearly it is not a suitable topology for all
applications. For those applications that do not
require very low THD performance, it offers a
valuable solution to achieve good PF performance
(>0.9 minimum, 0.95 at nominal line) while
simultaneously delivering excellent efficiency.

IX. references
[1] Isaac Cohen and Bing Lu, High Power Factor
and High EfficiencyYou Can Have Both,
TI Power Supply Design Seminar, SEM1800,
2008/9.
[2] Lloyd Dixon, High Power Factor Preregulator
Design Optimization, TI Power Supply Design
Seminar SEM700, 1990.
[3] EN61000-3-2 Edition 2.2: Limits For Harmonic
Current Emissions (equipment input current
16 A per phase), 2004.
[4] http://www.energystar.gov/.
[Online].
Available: http://www.energystar.gov/ia/
partners/product_specs/program_reqs/eps_
prog_req.pdf

Texas Instruments

4-28
28

[5] European Commission. Code of Conduct on

Energy Efficiency of External Power Supplies
Version 2. [Online]. Available: http://sunbird.
jrc.it/energyefficiency/pdf/Workshop_
Nov.2004/PS%20meeting/Code%20of%20
Conduct%20for%20PS%20Version%202%20
24%20November%202004.pdf
[6] A. Fernandez, J. Sebastian, M. Hernando, P.
Villegas, and J. Garcia, Helpful Hints to Select
a Power-Factor-Correction Solution for Lowand Medium-Power Single-Phase Power
Supplies, IEEE Transactions on Industrial
Electronics, Vol. 52, No. 1, pp. 4655, Feb.
2005.
[7] W.M. Lin, J. Sebastian, A. Fernandez, M.
Hernando, and P. Villegas, Design of the
Basic Rectifier with LC Filter to Comply with
the New Edition of the IEC1000-3-2 Current
Harmonic-Limit Specifications (Edition 2.0),
IEEE Applied Power Electronics Conference,
Vol. 2, pp. 12151220, Aug. 2002.
[8] R. Redl, A.S. Kislovski, and B.P. Erisman,
Input-Current-Clamping: An Inexpensive
Novel Control Technique to Achieve Compliance with Harmonic Regulations, IEEE
Applied Power Electronics Conference and
Exhibition, Vol. 1, pp. 145151, March 1996.
[9] L. Huber, L. Gang, and M. Jovanovic, DesignOriented Analysis and Performance Evaluation
of Buck PFC Front-End, IEEE Applied Power
Electronics Conference, pp. 11701176, Feb.
2009.
[10] G. Spiazzi, Analysis of Buck Converters Used
as Power Factor Preregulators, IEEE Power
Electronics Specialists Conference, Vol. 1, pp.
564570, June 1997.
[11] G. Young, G. Tomlins, and A. Keogh, An
ACDC Converter, World Intellectual Property
Organization, International Publication
WO/2006/046220 A1, May 4, 2006.
[12] J. B. Williams, Design of Feedback Loop in
Unity Power Factor AC to DC Converter,
IEEE Power Electronics Specialists Conference, Vol. 2, pp. 959967, June 1989.

SLUP264

[13] A. Pandey, B. Singh, and D. Kothari, A Simple

Fast Voltage Controller for Single-Phase PFC
Converters, IEEE 28th Annual Conference of
the Industrial Electronics Society, Vol. 2, pp.
12351237, Nov. 2002.
[14] H. Endo, T, Yamashita, and T. Sugiura, A
High-Power-Factor Buck Converter, IEEE
Power Electronics Specialists Conference, pp.
10711076, 1992.

[16] G. Young and J. Schneider, Integral-Cycle

Control Techniques for High Efficiency Power
Conversion, PCIM, 2009.
[17] N. Mohan, T. Undeland, and W. Robbins,
Power Electronics: Converters, Applications
and Design, Hoboken, NJ: Wiley, 1989, 1995.

[15] UCC29910: Buck PFC PWM Controller,

UCC29910 Datasheet, Texas Instruments,
2010.

Topic 4

ENERGY STAR is a registered mark owned by the U.S. government.

Texas Instruments

4-29
29

SLUP264

APPendIX A. Power fActor, dIstortIon fActor, thd

And dIsPlAcement fActor eXPlAIned
The terms power factor, displacement
factor, total harmonic distortion (THD), and
distortion factor and the relationship between
them can be confusing. These terms are very
thoroughly explained in Reference [17] and are
summarized here.
Power factor (PF) is the ratio of real power (in
watts) to apparent power (as a product of voltage
and current, or VA):
PF =

Real Power (W)

.
Apparent Power (VA)

S=

1 2
1 2
v (t )dt
i (t )dt

T0
T 0

Apparent Power = S = Vrms I rms

(A-3)
(A-4)

Because voltage, current and apparent power

are all complex numbers (i.e., phasors or vectors
with both an associated amplitude and phase),
apparent power can be recognized as comprising
real power (the real part of the complex number)
and reactive power (the imaginary part of the
complex number).
Periodic complex voltage and current waveforms may be represented by a sum of sinusoidal
components at each and every multiple of the
fundamental frequency of the waveform. Each
individual sinusoidal component is referred to as a
harmonic, and each harmonic will have an
associated amplitude and phase.
The simplest example is that of a purely
sinusoidal voltage source connected to a purely
resistive load. In this case, both voltage and current

Texas Instruments

4-30
30

Topic 4

(A-5)

where
DispF = Displacement Factor = cos;

(A-6)

thus the PF will be

1
= v (t ) i (t )dt. (A-2)
T0

Apparent power, S, sometimes referred to as

complex power, can be shown to be the product of
rms voltage and current, as in Equations A-3 and
A-4:
T

Pavg = Vrms I rms DispF,

(A-1)

Real power (also referred to as average power) is

the average over a time interval T of the product of
instantaneous voltage and current:
Real Power (W ) = Pavg

will be pure sinusoids, at the same frequency and

in phase, so real power will equal the product of
voltage and current. So in this case, PF will be
unity. If the load is somewhat capacitive or
inductive, however, then the current will still be
sinusoidal at the same frequency as the voltage
and there will be a phase angle between the
current and voltage. In this case the real power
will be:

PF =

Pavg
S

Vrms I rms DispF

= cos.
Vrms I rms

(A-7)

In this case, PF will be equal to the phase

displacement factor. Where the current that flows
is non-sinusoidal, Fourier analysis can determine
the fundamental frequency component (I1). Real
power is only transferred by the product of voltage
and current components that are in phase and at
the same frequency. Only similar frequency
voltage and current harmonics can transfer power,
and that transferred power is reduced as the phase
angle between the voltage and current harmonic
increases. The real power in this case will be:
Pavg = V1 I1 cos.

(A-8)

Apparent power will be Vrms Irms, as before.

Thus in this case PF will be:
PF =

V1 I1
cos.
Vrms I rms

(A-9)

For AC power supplies, it can be assumed that

the AC supply voltage is a perfect sinusoid with no
distortion, such that Vrms V1. This simplifies the
PF relationship to an analysis of the current
waveform:
PF =

I1
I rms

cos.

(A-10)

SLUP264

The distortion factor (DF) of the current waveform

is the ratio of the fundamental frequency component to the total rms value:
I1
I rms

THD =

(A-11)

I rms =

1 2
i (t ) dt =
T
0

i2n ,

I rms =

+ i n2 .

THD =

(A-13)

n =2

The THD of the current waveform can be

defined as:
THD = 100%

Texas Instruments

I 2rms I12
I1

(A-15)

(A-16)

1,

(A-17)

1 + THD 2

(A-12)

n =1

I1
1

DF =

or
I12

n =2

To summarize the most important relationships:

Fourier analysis can express the total rms current

in terms of its individual harmonics:
T

i2n

1
DF2

(A-18)

PF = DF DispF,
and
PF =

1
1 + THD 2

cos.

(A-19)

(A-14)

4-31
31

Topic 4

DF =

or alternatively:

SLUP264

Topic 4

APPendIX b. buck Pfc Power fActor lImItAtIons

Extensive analysis of the buck PFC harmonics
performance has been published, utilizing various
methods of control and current wave shapes [8, 9,
10]. Many of these analyses have primarily
explored the possible THD and line-current
harmonics performance achievable and maximum
power levels that can be delivered within the limits
of EN61000-3-2. However, EN61000-3-2 does
not impose any requirement for a minimum fullload PF. A current wave shape containing 90% of
the maximum allowed harmonic current for each
and every harmonic covered by the standard would
have a THD of 85%, resulting in a PF as low as
0.76 (assuming displacement factor of unity). More
recent initiatives such as ENERGY STAR EPS 2.0
have started mandating a minimum PF value of
0.9, which actually imposes much tougher requirements on THD. This puts a different perspective
and constraint on buck PFC performance. Nonetheless, 0.9 PF is readily achievable with the buck
PFC, even at low line voltages.
In the basic topology overview, some typical
current waveforms were drawn. Three such
waveforms have been simulated and analyzed for
harmonic content, distortion factor and maximum
possible power factor. The simulation assumed a
nominal bus voltage of 80 VDC for the buck
output with an assumed infinite buck-output capacitor (i.e. zero ripple on the buck output). Fig. B-1
illustrates the different waveforms, referred to as
A (modified sine, where current is proportional to
the input/output voltage differential during the
conduction angle), B (truncated sine, where the

current is proportional to the input voltage during

the conduction angle), and C (clamped-current
trapezoid, where current is constant during the
conduction angle). The resulting potential
harmonic content and power factor (neglecting the
effects of phase displacement) are summarized in
Table B-1. The analysis was confined to low-line
100 VAC, since this is the operating point where
PF is most challenged. At 230 VAC, cross-over
distortion is much less and good PF is readily
achieved.
Analysis shows that there is a trade-off between
the PF and individual harmonic content, depending
on the current-control technique used and the
shape of the AC line current. Waveform A
(modified sine) has very low harmonic content at
higher orders, but is dominated by a large third
harmonic. This waveform passes the requirements
of EN/JIS-61000-3-2 very easily, since those
standards permit a high degree of third harmonic.
However, the third harmonic results in a quite high
THD of 44%, limiting potential performance at
100 VAC to 0.91 maximum PF (neglecting phase
shift and other sources of distortion). In effect, the
slow rate of current rise and fall within the
conduction angle is good for limiting high-order
harmonics, but results in poor utilization of the
already limited conduction angle of the buck PFC,
with limited PF.
Waveform B (truncated sine) has a faster rate
of current rise and fall within the conduction
angle and better use of the available window.
Consequently, Waveform B has a much lower

tAble b-1. summAry of sImulAted

wAveform thd And Pf
Parameter

Texas Instruments

Waveform A

Waveform B

Waveform C

Irms (total)

1.0886

1.0362

1.0527

I1(rms)

0.9953

0.9866

1.0008

Real Power (W)

99.53

98.66

100.1

THD (%)

44.3

31.9

32.62

Distortion Factor

0.914

0.953

0.9507

Max Possible PF (no

phase displacement)

0.914

0.953

0.9507

4-32
32

SLUP264

150

2.5

AC Input Current

50

1.0

100
150

Waveform A
Current (Arms)

AC Input Voltage (V)

1.0

50

Limit (Arms ) scaled

for 100 V/100 W
AC Input Current (A)

AC Input Voltage

100

1.000

2.0

0.100

0.010

2.0
0

THD = 44.3%

2.5
10 12 14 16 18 20

Time (ms)

0.001

Maximum Possible PF = 0.914

7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39

Harmonic Order

a. Waveform A. Modified sine and harmonic content.

2.5

AC Input Current

50

1.0

100
150

Waveform B
Current (Arms)

AC Input Voltage (V)

1.0

50

Limit (Arms ) scaled

for 100 V/100 W

2.0

AC Input Current (A)

AC Input Voltage

100

1.000

0.100

Topic 4

150

0.010

2.0
0

THD = 31.9%

2.5
10 12 14 16 18 20

Time (ms)

0.001

Maximum Possible PF = 0.953

7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39

Harmonic Order

b. Waveform B. Truncated sine and harmonic content.

AC Input Voltage (V)

100

AC Input Current

50

1.0

100

1.000

Limit (Arms ) scaled

for 100 V/100 W

2.0
1.0

50

150

2.5

Waveform C
Current (Arms)

AC Input Voltage

AC Input Current (A)

150

0.100

0.010

2.0
0

THD = 32.62%

2.5
10 12 14 16 18 20

Time (ms)

0.001

Maximum Possible PF = 0.9507

7 9 11 13 15 17 19 21 23 25 27 29 31 33 35 37 39

Harmonic Order

c. Waveform C. Clamped-current trapezoid and harmonic content.

Fig. B-1. Buck PFC simulated current waveforms and their harmonic analyses.
Texas Instruments

4-33
33

SLUP264

Topic 4

third-harmonic content than waveform A. Within

the conduction angle, the current shape is approximately sinusoidal, in proportion to instantaneous
AC line voltage. This results in 32% THD, corresponding to a potential maximum PF performance
of 0.952. Waveform B has a greater content of
higher order harmonics due to the higher di/dt of
the current at the shoulders of the waveform.
This simulation assumed a di/dt of 4 A/ms, which
is quite steep, and shows some harmonics that are
close to the JIS61000-3-2 limits, particularly the
15th. Thus, careful design is required to ensure
that all harmonics are kept within harmonic limits
while still maximizing PF. In practice, besides the
action of the current-control loop, the di/dt of the
current is also limited by circuit practicalities such
as filter impedance, PFC choke inductance, and
control circuit Dmax, so that slower di/dt values
will usually occur.
Waveform C (clamped current trapezoid) is a
variation of the truncated sine waveform B, where
the sinusoidal variation is clamped, resulting in a
flat-topped trapezoidal waveform. This type of
waveform has been previously analyzed [10]. This
also has a fast but limited rate of current rise and
fall within the conduction angle. Waveform C has
a very similar 32% THD and 0.952 PF, but the
clamped current nature can offer an efficiency
advantage, since the average and rms current
flowing in the bridge rectifiers, PFC MOSFET,
and choke will be lower. However, while the THD
and PF are similar to waveform B, the harmonic
distribution is differentin this case there is a
greater higher-frequency content, while the EN/
JIS61000-3-2 limits are much lower. This
waveform actually fails harmonic limits at some
frequencies (9th, 15th, and 19th), and comes very
close at several others (11th, 25th, and 29th).

Texas Instruments

4-34
34

Some interesting conclusions can be drawn

from these results. THD and PF are heavily
influenced by the third harmonic, as evidenced by
waveform A analysis. Waveforms with a low third
harmonic but quite high levels of higher order
harmonics can result in better PF performance, as
evidenced by waveform B. Somewhat similar
waveforms, with similar THD and PF, can result
in quite different harmonic contentwaveform B
passes EN61000-3-2 while waveform C fails; yet
both have almost identical PF.
Practical Issues that Degrade Buck PFC PF
Besides the current wave shape, many other
practical factors affect PF. These can cause either
current phase displacement of the fundamental
component or increase current distortion and
harmonic content. Many of these factors are
problematic for all PFC topologies. Where noted,
some are specific to the buck PFC.
EMC-filter film-capacitor displacement current
is problematic for all topologies.
Voltage-loop bandwidth and resultant controlsignal ripple at 100/120 Hz are more of a problem
for the buck PFC due to a higher bus-ripple
percentage.
Bulk-capacitor ripple phase-shift effect is
particular to the buck, where the bus ripple
causes the conduction to be phase shifted, adding
to filter-capacitor displacement effects.
Circuit practicalities that limit current di/dt at
the start/end of the conduction angle:
Source impedance.
PFC-choke inductance value.
Control-circuit maximum duty-cycle limit.

SLUP264

TI Worldwide Technical Support

Internet
TI Semiconductor Product Information Center
Home Page
support.ti.com

TI E2E Community Home Page

e2e.ti.com

Product Information Centers

Americas Phone

+1(972) 644-5580

Brazil

Phone

0800-891-2616

Mexico

Phone

0800-670-7544

Fax
Internet/Email

+1(972) 927-6377
support.ti.com/sc/pic/americas.htm

Europe, Middle East, and Africa

Phone
European Free Call

International
Russian Support

00800-ASK-TEXAS
(00800 275 83927)
+49 (0) 8161 80 2121
+7 (4) 95 98 10 701

Note: The European Free Call (Toll Free) number is not active
in all countries. If you have technical difficulty calling the free
call number, please use the international number above.
Fax
Internet
Direct Email

+(49) (0) 8161 80 2045

support.ti.com/sc/pic/euro.htm
[email protected]

Japan
Phone
Fax

Domestic
International
Domestic

0120-92-3326
+81-3-3344-5317
0120-81-0036

Internet/Email International

Domestic

support.ti.com/sc/pic/japan.htm
www.tij.co.jp/pic

Asia
Phone
International
+91-80-41381665
Domestic
Toll-Free Number
Note: Toll-free numbers do not support
mobile and IP phones.
Australia
1-800-999-084
China
800-820-8682
Hong Kong
800-96-5941
India
1-800-425-7888
Indonesia
001-803-8861-1006
Korea
080-551-2804
Malaysia
1-800-80-3973
New Zealand
0800-446-934
Philippines
1-800-765-7404
Singapore
800-886-1028
Taiwan
0800-006800
Thailand
001-800-886-0010
Fax
+8621-23073686
Email
[email protected] or [email protected]
Internet
support.ti.com/sc/pic/asia.htm
Important Notice: The products and services of Texas Instruments
Incorporated and its subsidiaries described herein are sold subject to TIs
standard terms and conditions of sale. Customers are advised to obtain the
most current and complete information about TI products and services before
placing orders. TI assumes no liability for applications assistance, customers
applications or product designs, software performance, or infringement of
patents. The publication of information regarding any other companys products
or services does not constitute TIs approval, warranty or endorsement thereof.
A122010

The platform bar, E2E and SO REAL are trademarks of Texas Instruments. All other
trademarks are the property of their respective owners.

SLUP264

IMPORTANT NOTICE
Texas Instruments Incorporated and its subsidiaries (TI) reserve the right to make corrections, modifications, enhancements, improvements,
and other changes to its products and services at any time and to discontinue any product or service without notice. Customers should
obtain the latest relevant information before placing orders and should verify that such information is current and complete. All products are
sold subject to TIs terms and conditions of sale supplied at the time of order acknowledgment.
TI warrants performance of its hardware products to the specifications applicable at the time of sale in accordance with TIs standard
warranty. Testing and other quality control techniques are used to the extent TI deems necessary to support this warranty. Except where
mandated by government requirements, testing of all parameters of each product is not necessarily performed.
TI assumes no liability for applications assistance or customer product design. Customers are responsible for their products and
applications using TI components. To minimize the risks associated with customer products and applications, customers should provide
adequate design and operating safeguards.
TI does not warrant or represent that any license, either express or implied, is granted under any TI patent right, copyright, mask work right,
or other TI intellectual property right relating to any combination, machine, or process in which TI products or services are used. Information
published by TI regarding third-party products or services does not constitute a license from TI to use such products or services or a
warranty or endorsement thereof. Use of such information may require a license from a third party under the patents or other intellectual
property of the third party, or a license from TI under the patents or other intellectual property of TI.
Reproduction of TI information in TI data books or data sheets is permissible only if reproduction is without alteration and is accompanied
by all associated warranties, conditions, limitations, and notices. Reproduction of this information with alteration is an unfair and deceptive
business practice. TI is not responsible or liable for such altered documentation. Information of third parties may be subject to additional
restrictions.
Resale of TI products or services with statements different from or beyond the parameters stated by TI for that product or service voids all
express and any implied warranties for the associated TI product or service and is an unfair and deceptive business practice. TI is not
responsible or liable for any such statements.
TI products are not authorized for use in safety-critical applications (such as life support) where a failure of the TI product would reasonably
be expected to cause severe personal injury or death, unless officers of the parties have executed an agreement specifically governing
such use. Buyers represent that they have all necessary expertise in the safety and regulatory ramifications of their applications, and
acknowledge and agree that they are solely responsible for all legal, regulatory and safety-related requirements concerning their products
and any use of TI products in such safety-critical applications, notwithstanding any applications-related information or support that may be
provided by TI. Further, Buyers must fully indemnify TI and its representatives against any damages arising out of the use of TI products in
such safety-critical applications.
TI products are neither designed nor intended for use in military/aerospace applications or environments unless the TI products are
specifically designated by TI as military-grade or "enhanced plastic." Only products designated by TI as military-grade meet military
specifications. Buyers acknowledge and agree that any such use of TI products which TI has not designated as military-grade is solely at
the Buyer's risk, and that they are solely responsible for compliance with all legal and regulatory requirements in connection with such use.
TI products are neither designed nor intended for use in automotive applications or environments unless the specific TI products are
designated by TI as compliant with ISO/TS 16949 requirements. Buyers acknowledge and agree that, if they use any non-designated
products in automotive applications, TI will not be responsible for any failure to meet such requirements.
Following are URLs where you can obtain information on other Texas Instruments products and application solutions:
Products

Applications

Audio

www.ti.com/audio

Communications and Telecom www.ti.com/communications

Amplifiers

amplifier.ti.com

Computers and Peripherals

www.ti.com/computers

Data Converters

dataconverter.ti.com

Consumer Electronics

www.ti.com/consumer-apps

DLP Products

www.dlp.com

Energy and Lighting

www.ti.com/energy

DSP

dsp.ti.com

Industrial

www.ti.com/industrial

Clocks and Timers

www.ti.com/clocks

Medical

www.ti.com/medical

Interface

interface.ti.com

Security

www.ti.com/security

Logic

logic.ti.com

Space, Avionics and Defense

www.ti.com/space-avionics-defense

Power Mgmt

power.ti.com

Transportation and Automotive www.ti.com/automotive

Microcontrollers

microcontroller.ti.com

Video and Imaging

RFID

www.ti-rfid.com

OMAP Mobile Processors

www.ti.com/omap

Wireless Connectivity

www.ti.com/wirelessconnectivity
TI E2E Community Home Page

www.ti.com/video

e2e.ti.com

Mailing Address: Texas Instruments, Post Office Box 655303, Dallas, Texas 75265
Copyright 2011, Texas Instruments Incorporated