Using a Modified Charge-Pump Technique to Reduce Line
Harmonics
M.
Selim Göksu
A
solution is introduced to solve the problem of high power
dissipation in the charge-pump method.
Since
January 2001, EN 61000-3-2 has regulated the European
line-current harmonic levels of equipment up to 16 A per
phase from the public 230-V supply.1 Active
and passive power factor correction (PFC) circuits are
used to comply with this standard.
When
the charge-pump PFC circuit is used in flyback convertors,
it provides a very cost-effective solution with enough
performance to fulfill the Class D requirements of EN
61000-3-2, the class applicable to TV sets.2
Because of their low cost and simplicity, conventional
dc-to-dc flyback convertors are commonly found in consumer
electronic products consuming less than 250 W, such as
TV sets, digital versatile disc players, and satellite
receivers.3, 4
The
basic charge-pump circuit has some heat problems related
to power consumption in the switching element. The pump
current that flows through the PFC capacitor (PFC cap
in Figures 4 and 6) influences the voltage-current (V-I)
characteristic of the metal-oxide semiconductor field-effect
transistor (MOSFET) switching element and increases the
temperature (caused by the switching loss) to an unacceptable
level, leading to reliability problems.
The
Existing Charge-Pump PFC Technique
In
classic ac-to-dc convertors without PFC, the current is
drawn through the mains during the charging process of
the bulk capacitor, which is connected after the bridge
rectifier to filter the rectified line signal (see Figure
1). This current has a pulse shape that provides a power
factor of less than 1/2.
 |
|
Figure
1. Basic bridge rectifier and current-voltage waveforms.
|
The
mains voltage is rectified by the bridge rectifier and
smoothed by the bulk capacitor to obtain the dc bus voltage.
The dc bus voltage is switched across the primary winding
of the switch-mode transformer (SMT) via the power switch
Q in the flyback switch-mode (FBSM) power supply topology
generally used in TV sets. A general block scheme of a
one-output FBSM power supply without PFC application is
shown in Figure 2.
 |
|
Figure
2. General block diagram of a one-output FBSM power
supply without PFC application.
|
An
FBSM power supply used in a large-screen TV set includes
an SMT with multiple windings, allowing the generation
of various dc output voltages as shown in Figure 3.3
This type of switch-mode power supply is known as a multiple-output
ac-to-dc flyback convertor.
 |
|
Figure
3. Block diagram of an FBSM power supply for TVs.
|
The
basic charge-pump PFC circuit shown in Figure 4 can be
used to fulfill EN 61000-3-2 for large-screen TV sets
with nominal power consumption (>75 W).2
However, the power loss of the switching element Q1
is quite high. According to measurements on a 33-in. TV
set that consumes 230 W and uses a multiple-output FBSM
power supply, the body temperature of Q1
was 100°C, with the measurement conditions shown
in Table I.
 |
|
Figure
4. Flyback convertor primary-side schematic diagram
with basic charge-pump circuitry.
|
|
Parameter
|
Condition
|
| Vmains |
220
V rms
|
| Power
consumption |
200
W
|
| Operating
time |
4
hours
|
| Ambient
temperature |
40°C
|
| Q1 |
SPP20N60
low-Rds MOSFET
|
| C1 |
10
nF at 1.6 kV
|
|
|
Table
I. Initial measurement conditions of Q1.
|
The
switch current IQ is composed
of Ip and IPFC
cap, as shown in Figure 4. The V-I waveform
of Q1 is shown in Figure 5 and is
directly influenced by the momentary current spikes of
PFC capacitor C1. Especially
at the beginning of the MOSFET power-on cycle, C1
has almost no charge inside. Therefore, the amplitude
of current spikes is more or less equal to the steady-state
peak current amplitude. This behavior leads to an increase
in the turn-on loss of Q1 and its
body temperature.
 |
|
Figure
5. V-I characteristic of the switching element with
basic charge-pump circuitry. Channel 1 shows the
drain voltage of Q1 at 100
V per division, and Channel 2 shows the drain current
of Q1 at 1 A per division.
|
Of
course, the temperature of Q1 can
be limited using a variety of methods, such as increasing
the size of the heat sink surface or using a switching
element with a larger drain-to-source current. Investigation
of the known charge-pump circuitry showed that solutions
often involved changing the charging path of C1,
which would increase the required printed circuit board
(PCB) surface area and the final cost of the product.
To overcome the temperature rise problem of Q1
without increasing the cost and PCB surface area, the
following technique was developed.
Improving
Switching Performance
To
provide the same current flow to the PFC capacitor, an
additional winding is added to the SMT, and the PFC capacitor
is coupled to this so-called PFC winding as shown in Figure
6. The resulting switching-element current (see Figure
7) is completely independent of the PFC capacitor current.
 |
|
Figure
6. Flyback convertor primary-side schematic diagram
with modified charge-pump circuitry.
|
 |
|
Figure
7. V-I characteristic of the switching element with
modified charge-pump circuitry. Channel 1 shows
the drain current at 2 A per division, and Channel
2 shows the drain voltage at 100 V per division.
|
The
required current was obtained easily because of the voltage
difference at the terminals of capacitor. When the switch
is on, current flows from the PFC winding to the PFC coil.
When the switch is off, the current flows in the reverse
direction (see Figure 8).
 |
|
Figure
8. PFC capacitor current (Channel 1 at 2 A per division)
and VPFC voltage (Channel
2 at 100 V per division) waveforms.
|
According
to measurements on the same 33-in. TV set, the body temperature
of Q1 was 88°C, with the measurement
conditions shown in Table II. This measurement demonstrates
a reduced body temperature despite a 20°C rise in
the ambient temperature.
|
Parameter
|
Condition
|
| Vmains |
220
V rms
|
| Power
consumption |
200
W
|
| Operating
time |
4
hours
|
| Ambient
temperature |
40°C
|
| Q1 |
SPP20N60
low-Rds MOSFET
|
| C1 |
15
nF at 1.6 kV
|
| D3 |
Short
circuited
|
|
|
Table
II. Measurement condition of Q1
with improved circuitry.
|
In
terms of the turn-on loss of the switching element, the
current spikes from the modified circuitry (see Figure
7) are negligible in comparison with the original circuitry
(see Figure 5).
The
redirection of IPFC cap removes
the need to use diode D3,
which is required in basic charge-pump circuits to prevent
IPFC cap from flowing from
the Q1 drain to the SMT during the
charging of the PFC capacitor. At the same time, to keep
standby power consumption to less than 3 W, the connection
pin of the PFC capacitor to the PFC coil is changed to
the cathode pin of D1. Due
to this modification in the connection of the PFC capacitor,
it is necessary to use a PFC capacitor of a larger value
with the same physical dimensions to retain the same current
harmonic performance within the same PCB surface area.
The mains current harmonic values are provided in Table
III for a power consumption of 180 W, which is a standard
condition.
Harmonic
Number |
Measured
Value
(Arms) |
EN
61000-3-2 Limit
(Arms) |
Status |
|
1
|
0.852
|
|
OK
|
|
3
|
0.365
|
0.728
|
OK
|
|
5
|
0.224
|
0.404
|
OK
|
|
7
|
0.119
|
0.212
|
OK
|
|
9
|
0.104
|
0.106
|
OK
|
|
11
|
0.042
|
0.074
|
OK
|
|
13
|
0.018
|
0.063
|
OK
|
|
15
|
0.008
|
0.054
|
OK
|
|
17
|
0.007
|
0.049
|
OK
|
|
|
Table
III. Measured mains current harmonic values fulfilling
EN 61000-3-2.
|
Conclusion
The
problem with the basic charge-pump circuit used in multiple-output
flyback power supplies for PFC is the excessive turn-on
loss of the switching element. The proposed technique
contains an additional winding on the SMT and a new coupling
to the so-called PFC winding (see Figure 6). By employing
this topology, the V-I waveform of the switching element
is reshaped, drastically reducing the turn-on loss. In
addition, the proposed technique reduces the cost by removing
one diode, which is not necessary for the new circuitry,
and keeps standby power consumption to less than 3 W for
a 33-in. TV set. Therefore, the improved version of the
basic charge-pump circuit appears to be a good option
for using PFC circuits to limit line-current harmonics.
This technique minimizes high power dissipation and reduces
line harmonics to meet EN 61000-3-2 requirements.
References
1.
EN 61000-3-2, "Electromagnetic CompatibilityPart 3-2:
LimitsLimits for Harmonic Current Emissions," European
Committee for Electrotechnical Standardization (CENELEC),
Brussels, 2000.
2.
Peter Preller, "A Controller Family for Switch Mode Power
Supplies Supporting Low Power Stand By and Power Factor
Correction (PFC)," Infineon Technologies Application
Note on AN-TDA1684X Version 1.2 (Munich: Infineon
Technologies AG, 2000).
3.
Keith H Billings, Switchmode Power Supplies Handbook (Maidenhead,
Berkshire, UK: McGraw-Hill, 1989).
4.
M Selim Göksu, "Passive Power Factor Correction (PFC)
Considerations," Vest Report 1 (Manisa, Turkey:
Vestel Electronics A.S, 2000).
M.
Selim Göksu joined the Vestel Electronics Co. R&D
hardware engineering team as a project manager in 1998.
His work on the power supply design for 100-Hz TV sets
is the subject of a patent application.
|