Alland Chee, Elvis Lin, Joe Wong

1
Design a 100W PFC Boost Inductor

• Alland Chee, Elvis Lin, Joe Wong
• and
• Jeffrey H. Hwang

2
Criteria of Selection

• Size
• Efficiency
• Cost

3
Review Important Magnetic VariablesandImportant
Magnetic Constants

• RS Magnetic ReluctanceThe
  resistance of a material to a magnetic field
• F Weber Magnetic FluxA measure of quantity of
  magnetism.
• B Tesla Gauss Magnetic FieldFlux
  DensityMagnetic Induction
• H Magnetic Field
  StrengthMagnetic ForceMMF per length

• 1Tesla1Weber/Meter2104 Gauss1Newton/(Amp
  ere x Meter)
• 1Weber1Wb108Maxwell
• L1henry1Volt x Sec/Ampere1Wb/Amp
• µ0Air Permeability Constant4p x 10-7 Tesla
  Meter/Ampere

4
Important Magnetic Formula

• Rmmf /FBNI/FB
• ?m/(Ac µ)
• ReRl Rair
• ?m/(Ac µ) ?air/(Ac µair)
• FB
• LN FB/IN2/Re

• Bµ H µ x mmf/leff
• FB/Ac µ x N x I/leff
• L x I/(N x Aac)

5
Bmax Constant _at_ given material
PC95Bmax350mT PC44Bmax300mT PC40Bmax270mT
6
NImax Constant _at_ lair, air gap is fixed
7
Sundest, MPP, Ferrite?Experiment
1stSelecting the best Magnetic Material
8
Selecting Magnetic Materials
Ferrite and MPP have the higher efficiency
performance. Ferrite is our selection.
9
Ferrite is our selection for Po100W Ferrite is
Cost Effective and Ferrite Core Loss (AC Loss) is
much less DCR seems dominates the efficiency
with Ferrite Core! Let us prove it here!
10
290uH or 1mH?Experiment 2ndDCR and switching
frequency is fixedWinding Factor is not optimal.
11
With fixed DCR0.46 ohm and fixed fsw67.5Khz,
Po100 W and Vo19V AC Adapter
Higher Efficiency with Higher Inductor? but it is
miss-leading!
12
We have learnedWithout the Space Limit,Higher
inductance will have the higher efficiency.
13
Should we go higher frequency to solve the space
issue?Experiment 3thWith RM8 and RM6, we
fixed crest factor, r0.95andWinding Factor is
Optimal with the giving bubbin.
14
Po100W, PFC boost OnlyConstant rcrest
factorIp-p/Irms0.95
with Constant r0.95 Higher Efficiency with
Higher Inductance? Higher Efficiency with Lower
Frequency? Again, it is miss-leading! From above
data, to improve efficiency, we only know that
we should reduce frequency to trade efficiency.
15
We have learnedWith the given space, frequency
should be as low as possible before the core is
saturated.The Lower Frequency provides the
Higher Efficiency
16
What will happen with fixed switching
frequency?Experiment 4thOptimal Winding
Factor for RM8 and fsw 67.5Khz
17
fsw67.5Khz with CM6800 100W demo board
L442uH is the Highest Efficiency one Crest
Factor, r1
18
fsw100Khz with CM6800 100W demo board
L295uH is the Highest Efficiency one Crest
Factor, r1
19
fsw100Khz with CM6800 100W demo board
L295uH is the Highest Efficiency one Crest
Factor, r1
20
fsw67.5Khz with CM6805 100W AC Adapter
L295uH is the Highest Efficiency one Crest
Factor, r1.55
21
We have learnedCrest Factor, r1 to
1.55L442uH to 295uH for 67.5Khz and L295uH for
100Khz gives the best efficiencyIt means
without saturating the coreReducing DCR,
Reducing lg , Reducing N, all will improve the
efficiency.
22
ConclusionFerrite PC95 RM8 (67.5Khz) --- gt
RM6 (100Khz)295uHwith CM6805 (PFCPWM combo IC)
23
442uH with fsw 67.5Khz Vout 220V, Po 100W,
RM8 and PC 95

• lg 24 mil
• N 52.5 turns
• Pin max 146.56W
• L 400.58uH x 1.09
• Al 0.1584uH/turn2
• Ipeak at Sat 2.3A
• DCR 0.09 ohm

• P core loss 0.39W at Pin max
• P copper loss 0.228W at Pin max
• Total P loss 0.62W at Pin max
• Wire Area 0.44 mm2 gt AWG21

24
295uH with fsw 67.5Khz, Vout 220V, Po
100W, RM8 and PC 95

• lg 20 mil
• N 40.5 turns
• Pin max 138.6W
• L 285uH
• Al 0.1738uH/turn2
• Ipeak at Sat 2.178A
• DCR 0.05 ohm

• P core loss 0.39W at Pin max
• P copper loss 0.12W at Pin max
• Total P loss 0.51W at Pin max
• Wire Area 0.57 mm2 gtAWG20

25
295uH with fsw gt 100Khz, Vout 220V, Po 100W,
RM6 and PC 95

• lg 32 mil
• N 64.5 turns
• Pin max 118.93W
• L 229uH x 1.23
• Al 0.068uH/turn2
• Ipeak at Sat 1.87A
• DCR 0.19 ohm

• P core loss 0.15W at Pin max
• P copper loss 0.33W at Pin max
• Total P loss 0.48W at Pin max
• Wire Area 0.18 mm2 gtAWG25

26
100W AC Adapter without SREfficiency vs.
LoadWith 295uH(RM8), 442uH(RM8) and
303uH(RM6)at fsw 67.5Khz
27
100W AC Adapter without SREfficiency vs.
VinWith 295uH(RM8), 442uH(RM8) and 303uH(RM6)at
fsw 67.5Khz
Po100W
Po66W
28
100W AC Adapter with SR and without SREfficiency
vs. VinWith 442uH(RM8)at fsw 67.5Khz
Measure the Efficiency Data at the end of cables.