Boost Converter Efficiency
Through Accurate Calculations
Identifying all of the individual loss contribu-
tors enables the development of higher-order
models for designing efficient boost-converter
circuits.
P
ortable battery-operated electronic equipment is
becoming packed with more features requiring
larger amounts of power, leading to decreased
operating times between battery charges.
Innovations such as dynamic backlight
control, where the liquid crystal display’s (LCD) backlight
power is reduced based on screen content, and automatic
voltage scaling, where a microprocessor’s frequency and
supply voltage are reduced based on computing time re-
quirements, help to gain back those lost milliwatts. This
milliwatt level of lost power is important considering the
relatively low amounts of power consumed by a typical
electronic device, and the small battery capacities that can
often be rated at 1000 mAh or less. Essentially, everywhere
that power loss can be accounted for helps in optimizing
the design and aids in extending battery life.
One contributor to power loss in portable electronic
devices is the inductive boost-switching regulator. These
types of converters are typically used to power white light-
emitting diodes (LEDs) in back-lit LCDs, organic LED
panels, high-current flash LEDs, or even audio amplifiers.
Although inductive boost converters can be highly
efficient, improper selection of a converter type, operat-
ing frequency and/or external components can lead to an
inefficient design. Therefore, making accurate efficiency
calculations helps to identify the individual loss contribu-
tors and provides the insight necessary in designing an
efficient boost converter.
To begin with, take a standard asynchronous boost
converter. The simple approximation to efficiency can be
made using a first-order model where the ideal duty cycle
(D) = (VOUT – VIN)/VOUT and the average inductor current,
or input current (IIN), IIN = (IOUT/1 – D) is used to estimate
the dc losses in the NMOS switch, the Schottky diode and
the inductor.
Efficiency then becomes:
Power Electronics Technology September 2008 30
By Travis Eichhorn, Senior Applications Engineer,
National Semiconductor, Grass Valley, Calif.
PIN - (PSWITCH + PINDUCTOR + PDIODE )
eff = ,
PIN
where PSWITCH = RNMOS ´ D ´ (IOUT/1 – D)2, PINDUCTOR =
RL ´ (IOUT/1 – D)2 and PDIODE = RD ´ IOUT2 + VF ´ IOUT .
This model makes for a rough estimate; however, the
initial approximation for duty cycle ignores the circuit’s
internal losses and ends up underestimating the duty cycle
as well as the input current. Essentially, the duty cycle
must increase to overcome the circuit’s internal losses.
Consequently, the initial calculation for the duty cycle
should have involved the component of efficiency or, more
specifically:
VOUT - VIN ´ eff
D= .
VOUT
This can be found by using the equation VIN ´ IIN ´ eff =
VOUT ´ IOUT , substituting IIN = (IOUT/1 – D) and solving for
D. Because of this, the first-order model’s efficiency estimate
will increasingly deviate from the actual efficiency as the
power loss between the input and the output increases.
Accounting for this requires a model that computes the
duty cycle based on the power-loss components and then
uses this value to calculate the circuit’s input current, again
via the ratio IIN = (IOUT/1 – D).
In developing this model, consider the power balance
in the boost regulator PIN = PLOSS + POUT . This states that
the total input power is equal to the output power plus the
power lost in each of the circuit’s components. Again, us-
ing a typical asynchronous boost converter, the dc losses
in the NFET switch, the diode and the inductor are used to
generate a power-balance equation given by PIN = PSWITCH
+ PDIODE + PINDUCTOR + POUT .
Substituting IIN = (IOUT/1 – D), as was done before in
the first-order model, leads to the modified power-balance
equation:
www.powerelectronics.com
 2
VIN ´ IOUT æI ö
= VF ´ IOUT + R D ´ IOUT2 + R N ´ççç OUT ÷÷÷ D + 10 mH Schottky
1- D è1 - D ø LPS4012-103 SDM20U30 VOUT = 19 V
2
æ IOUT ö÷
R L ´ççç ÷ + VOUT ´ IOUT .
è1 - D ÷ø
VIN = 3.6 V COUT
Instead of using the ideal duty-cycle ratio D = (VOUT IN SW OVP 1 mF

20 mA
CIN

20 mA
– VIN)/VOUT , the power-balance equation is now used to
1 mF
solve directly for D. Multiplying by (1 – D)2 and collecting LM3528
the terms on one side gives a second-order polynomial for VIO
the duty cycle: MAIN
D2 ´ (VF ´ IOUT + VOUT ´ IOUT + RD ´ IOUT2) + D ´ (RN ´ 10 kW 10 kW SUB/FB
SCL HWEN/
IOUT2 + VIN ´ IOUT – 2 ´ VF ´ IOUT – 2 ´ VOUT ´ IOUT) + (RL SDA PGEN/GPIO
´ IOUT2 + VF ´ IOUT + VOUT ´ IOUT – VIN ´ IOUT) = F(D)2. GPIO PGND SET
10 kW 205 W
This becomes the second-order model that generates 11.9 kW

two solutions for D, one of which is correct (the first zero
crossing of F(D) versus increasing D) and the other (the
second crossing) is not. This can be solved using the qua-
dratic formula.
Since only the dc losses are accounted for in this model,
the efficiency estimate is only accurate when the converter
operates in the continuous-conduction mode (CCM) with
relatively small inductor current ripple and when the dc
losses are much bigger than the switching losses. However,
when this is not the case and the peak-peak inductor current
ripple approaches (2 ´ IOUT)/1 – D and/or switching losses
begin to approach or surpass the dc conduction losses,
Indicator LED
Circuit parameters: VOUT = 19 V, VIN = 3.6 V, RL = 0.3 W, RDS = 0.45 W,
ON
RD = 0.75 W, VS = 0.45 V, CDS =40 pF, CD = 20 pF, tRISE
= 8 ns, tFALL = 8 ns
Fig. 1. Calculating boost-converter efficiency using conventional
first- and second-order models can be inaccurate, as is the case
for this LM3528 white LED driver circuit. Higher-order modeling
yields more accurate results.

High Efficiency Solar Charging ... IXYS SOLAR POWER!

IXOLARTM
Power MOSFET
IXTY12N06T
SOLAR ARRAY CPC3801
With our solar components,
designing a high-efficiency solar
X0B17_12X1
X0B17_04X3 power solution is now
Charging Current simpler than ever!

Applications
Batteries
“Solar Powered • NiMH, NiCad, Alkaline, etc.
Gate Driver” • Laptops, Cell phones, MP3 players
CPC1831
• GPS, Portable test equipment, Flashlights
CPC1832
• Security cameras
• Remote sensors

GREAT FOR • Electric fences

• RV & automotive
CHARGING BATTERIES
• UPS

SOLAR
For more information, visit For technical information,

www.ixyscolorado.com call Gib Bates at

Efficiency By The Sun and www.clare.com (970) 493-1902 ext. 22

www.powerelectronics.com 31 Power Electronics Technology September 2008

 Boost converter
this model begins to deviate from the second-order models, consider an LEDs in an LCD backlight (Fig. 1).
actual efficiency, much the same way asynchronous boost converter such With the given circuit parameters, the
as the first-order model did. as National Semi’s LM3528, which is measured efficiency is 83%.
To show the error in the first- and designed to power two strings of white Using the first-order model, the
Continuous conduction
estimated duty cycle is:
Discontinuous conduction
normalized to IOUT2 VOUT - VIN 19 - 3.6
D= = = 0.81.
NFET dc power PNDC = RNFET ´IOUT ´D 2
PNDC = 0 VOUT 19
NFET power The input current is calculated as:
RNFET ´IAC 2 IAC 2 I
loss due to ripple PNAC = ´D3 ´ (1- D)2 PNAC = ´RNFET ´D3 IIN = OUT = 211 mA.
current 12 3 1- D
Inductor dc The dc conduction losses in the
PLDC = RL ´IOUT 2 PLDC = 0
power loss NFET become:
D ´ IIN2 ´ RN = 18 mW.
IAC 2
Inductor power PNAC = ´RL ´ The diode losses become:
RL ´IAC 2 3
loss due to ripple PLAC = ´D2 ´ (1- D)2 VF ´ IOUT + IOUT2 ´ RD = 19.2 mW.
12 æ 3 I ´ 2 ´ f ´L ö
current ççD + OUT
ççè
SW
´D÷÷÷ And the inductor loss becomes:
V IN
÷ø IIN2 ´ RL = 15.6 mW.
This gives an approximated ef-
Diode resistance PRD = RD ´IOUT 2 ´ (1- D) PRD =0
dc power loss DC DC ficiency of:
PIN - PLOSS
IAC 2 =
PRD = ´RD ´ PIN
Diode resistance RD ´IAC 2 AC 3
power loss due PRD = ´D2 ´ (1- D)3 IOUT ´ 2 ´ fSW ´L VIN ´ IIN - PL - PN - PD
to ripple current
AC 12 ´D = 93%.
VIN VIN ´ IIN
The second-order estimate should
Power required VOUT 2 ´ CDS ´ fSW VOUT 2 ´ CDS ´ fSW yield a closer approximation to the
PCDS = ´ (1- D)2 PCDS = actual efficiency since the circuit’s
for CDS capacitor 2 2
internal losses are included in cal-
Power required VOUT 2 ´ CD ´ fSW VOUT 2 ´ CD ´ fSW culating the duty cycle. Taking the
PCD = ´ (1- D)2 PCD =
for CD capacitor 2 2 converter’s parameters and using the
quadratic formula to solve for D in
IDS and VDS VOUT ´IOUT the equation F(D2) gives a duty cycle
overlap losses PCCDC = ´ PCCDC = 0
2 of 0.825. This results in a slightly lower
due to inductor
dc current (tRISE + tFALL )´ fSW efficiency:
VOUT - D´ VOUT
IDS and VDS PCCAC =
VOUT ´IOUT
´ (tRISE - tFALL )
VOUT ´IAC eff = = 92.4%.
PCCAC = ´ VIN
overlap losses 4 2
due to inductor ´ fSW ´D ´ (1- D)2 ( tFALL )´ fSW ´D However, this efficiency is still quite
ripple current a ways off from the actual value.
Diode
reverse recovery PQRR = VOUT ´ QRR ´ fSW ´ (1- D)2 PQRR = VOUT + QRR ´ fSW Switching Losses
charge loss Where does the problem lie? For
2 boost converters such as white LED
Output power POUT = IOUT ´ VOUT ´ (1- D) POUT = IOUT ´ VOUT
drivers that operate with such low
IAC ´ VIN output currents and high duty cycles,
PIN = IOUT ´ VIN PIN = ´ the switching losses are no longer
2
Input power æI ´ f ´L ´ 2 ö trivial compared to the circuit’s con-
çç OUT SW
ççè + D2 ÷÷÷ duction losses. The switching-loss
VIN ÷ø
components must be included in the
second-order power-balance equation
IOUT VIN to give a more accurate estimate of the
Input current IIN = IIN = ´D2 + IOUT
1- D L ´SW ´2 duty cycle and a closer calculation of
circuit efficiency.
Table. Power-loss components for continuous and discontinuous conduction (asynchro- To start with, the primary switch-
nous boost converter). ing-power loss components in an

Power Electronics Technology September 2008 32 www.powerelectronics.com

 Boost converter
Standard and
asynchronous boost converter include
the power due to the MOSFET cross
Using CDS equals 40 pF, CD equals
20 pF and tRISE equals tFALL equals 8 ns Custom Power
conduction (PCC), which is the overlap results in an estimated duty cycle of Transformer
between the current and voltage dur-
ing the switch turn-on and turn-off,
83.8% for the LM3528. This results
in a calculated efficiency of 85.5%,
Solutions
the charging of the Schottky diode’s which is much closer to the measured
capacitance with each switching cycle value.
(PCD), and the charging of the NFET
switch’s drain-to-source capacitance Higher-Order Model
each switching cycle (PCDS). Until now, the duty-cycle calcu-
Adding these to the power-balance lations assumed no ripple current,
equation gives a more complicated which isn’t a problem because an av-
(but accurate) equation of the duty erage inductor current is used instead
cycle. The new second-order power- of root-mean-square (rms) inductor
balance equation becomes: current. And assuming the converter
IN ´ IOUT
VIN OUT
operates in the CCM only changes
= the previous answer a slight amount.
1- D
22 However, with boost circuits operat-
æçVFF ´ IOUT OUT + R D D ´ IOUT
OUT + R N N ´ö
÷
çç ÷÷ ing at such low-output power levels, 4 Laminated and Toroidal
ççæ IOUT ö 22
æ I ö
22
÷÷ it is likely that a large portion of the
çççç OUT ÷÷ D + R LL ´çç OUT ÷÷ OU T ÷÷ Power Transformers,
çèè1 - D ÷ø è1 - D ÷ø ÷ø load current will cause the converter
Inductors, and Custom
to operate in a discontinuous conduc-
æç 1 ´C ´ V 22 ´ f + VOUT OUT ´ö ÷÷ tion mode (DCM). High Frequency
çç 2 DS OUT SW
2 ÷÷ Magnetics
DS OUT SW
çç ÷÷ Unfortunately, the previous cal-
çç IOUT ÷÷
´(t RISE ) culations only applied to the CCM.
OUT
+çç RISE + t FALL
FALL ´ ÷÷ 4 Capabilities from 0.1 VA
çç1 - D ÷÷ If the designer wants to model the
çç 1 ÷÷ to 45 KVA
ççfSW 22 ÷÷ DCM efficiency, then the ripple cur-
SW + ´ VOUT OUT ´ C D D ´ fSW ÷ø
èç 2
SW
rent is no longer trivial. Furthermore, 4 PC and Chassis Mount
+VOUT OUT ´ IOUT
OUT .
to calculate efficiency over the entire
load-current range, ignoring the 4 RoHS-Compliant Products
This results in a modified second- ripple current in the CCM and then
order equation for a duty cycle of: including it in the DCM would lead 4 UL, CSA, VDE, IEC,
æ
çç
ö÷ to a discontinuity in efficiency as the and EN Certified
÷
ççVF ´ IOUT + VOUT ´ IOUT ÷÷ load current transitions through the
ç
ç
÷÷
modes. 4 In-House Agency
D2 ´çç+R D ´ IOUT2 + ÷÷ +
çç ÷÷ The solution is to include the Certification Program
÷÷
çç(C DS + C D )´ VOUT2 ÷ ripple current in the CCM duty-
çç ´ fSW ÷÷÷ 4 Large, Finished Goods
è 2 ø cycle equation and also solve for the
æ ö÷ duty cycle using a separate equation Inventory
çç ÷÷
2
ççR N ´ IOUT + VIN ´ IOUT ÷÷ for the DCM. The DCM equation
çç ÷÷ is developed using a power-balance
çç-2´ VF ´ IOUT - 2´ ÷÷ Custom Designs Welcome!
çç ÷ equation in much the same way as for
D´ççVOUT ´ IOUT - (C DS + C D )´÷÷÷ + the CCM case.
çç ÷÷
ççV 2 ´ f - VOUT ´ IOUT ´÷÷÷
çç OUT SW
2
÷÷
÷÷
As a starting point, the rms value of
a triangle waveform, which is the true PRONTO ®

çç ÷÷ inductor current, is given by: 24 HOUR SHIPPING!

çè(t RISE + t FALL )´ fSW ø Phone: 866/239-5777
DIL 2 Fax: 516/239-7208
æ ö IRMS = d ´ ILDC 2 + ILDC ´ .
ççR ´ I 2 + V ´ I + ÷÷ 3
çç L OUT ÷÷
Signal Transformer
F OUT
ççV ´ I - V ´ I +÷÷÷ ILDC is the dc or average value of the
çç OUT OUT IN OUT ÷÷
çç(C + C )´ V 2 ÷÷ inductor current, which is IOUT/1 – D
çç DS D OUT ÷÷ 2 for the CCM and zero for the DCM.
çç ÷÷ = F(D ).
çç 2 ÷
÷÷ ∆IL is the peak-to-peak inductor cur- www.signaltransformer.com
çç VOUT ´ IOUT ÷÷ rent ripple and is given by (∆IL = VIN sales@signaltransformer.com
çç´fSW - ÷÷
÷÷
çç 2 ´ D/(fSW ´ L)) for both the CCM and
ççè(t + t ÷÷
RISE FALL )´ f SW
÷ø the DCM. Finally, d is the portion of

www.powerelectronics.com 33 Signal/PowerPower Electronics Technology

Elec. 2/4/08.indd 1 September2/4/08
2008 9:23:48 AM
 Boost converter
of the period in the CCM and d = D + (2 ´ IOUT)/ ∆IL for
10 mH the components conducting during the first two portions of
LPS4012-103 VIN = 3.6 V the period in the DCM.
Schottky SW IN Adding the ripple component to the F(D2) equation adds
SMD20U30 LM27313
196 kW a third-, fourth- and fifth-order term for the resistive power-
VOUT = 19 V GND loss components. On the other hand, the DCM equation
4.7 mF FB
SHDN becomes a third-order polynomial. To avoid the complexity
120 pF 2.2 NF of drawing out these equations, the table categorizes the
13.3 kW power-loss components in the asynchronous boost converter
with the factored D terms to the right. In the table, the first
column lists the CCM components and the second column
lists the DCM components.
Circuit parameters: VOUT = 19 V, VIN = 3.6 V, RL = 0.3 7, RDS_ON = 0.8 7,
RD = 0.75 7, VS = 0.45 V, CDS =50 pF, CD = 20 pF, tRISE
Each term in the CCM column is normalized to IOUT  (i.e.,
= 12 ns, tFALL = 19 ns it is divided by (1 – D)2 to eliminate the (1 – D)2 term in the
denominator). The IAC term is the inductor current slope
Fig. 2. This LM27313 boost converter operates in either the CCM or over the entire switching period (or the inductor current
DCM mode over the entire load current range. It provides a more ripple divided by D). Collecting the terms of the power-loss
Figure 2
direct comparison with the developed higher-order model used in components in the CCM column results in the fifth-order
the measured versus calculated efficiency plots of Figs. 3 and 4. polynomial, FCCM (D5) = a ´ D5 + b ´ D4 + c ´ D3 + d ´ D2
+ e ´ D1 + f ´ D0, and collecting the terms of the power-
loss components in the DCM column gives a third-order
1.2 polynomial, FDCM (D3) = g ´ D3 + h ´ D2 + j ´ D1 + k ´
1 D0 , where the a through k coeffecients are functions of the
Efficiency (%)

0.8 circuit parameters VIN , VOUT , IOUT , RN , RL , etc.

0.6 High-order model This higher-order model using both the FCCM(D5) and
Second-order model (DC losses only) FDCM(D3) equations has no simple solution like the F(D2)
0.4 Second-order model (with switching losses)
First-order model equation. Fortunately, as with F(D2), the useful solutions to
0.2 Measured efficiency the fifth- and third-order polynomials are those occurring
0 at the first crossing of zero with increasing D.
0 5 10 15 20 25 30 35 40
IOUT (mA)
Finding this solution is possible, using a spreadsheet like
Excel, by setting F(D5) and F(D3) to zero and plotting F(D)
Fig. 3. Calculated versus measured efficiency levels
High-order model (for themodel
First-order circuit in versus D and varying D from zero to one. Alternatively,
Fig. 2)Second-order model (DC losses
using the higher-order only) and a modified
model Measured efficiency
higher-order implementing the functions FCCM(D5) and FDCM(D3) in C and
modelSecond-order
to accountmodel
for CD(with
andswitching losses)
CDS at light currents. using a “while” loop to test for the first F(D) = 0 crossing
versus increasing D works, also.
Comparing the calculated values with measured values
0.9 using the higher-order model requires that a converter be
0.8 used that operates in either the CCM or the DCM. Unfortu-
Efficiency (%)

0.7 nately, the LM3528 operates in a pulse-skip mode to improve

High-order model
0.6 High-order model efficiency at light loads. This would require a third equation
0.5 (CDS and CD modified) to model the converter in this mode. Additionally, using a
0.4 Measured efficiency FigureLED
3
white driver adds an extra layer of complexity in the
0.3 calculations since the LED voltage changes with IOUT .
0.2
0 5 10 15 20 25 30 35 40 To remedy this, the LM27313 standard asynchronous
IOUT (mA) boost converter will be used to compare the calculated ver-
sus measured results of the efficiency estimations (Fig. 2).
Fig. 4. Calculated versus measured efficiency levels (for the circuit This converter operates in either the CCM or the DCM over
in Fig. 2) using second-order and higher-order models. the entire load current range, thus providing a more direct
comparison with the developed higher-order model.
the switching period that the rms current is conducting. Figs. 3 and 4 use the equations developed so far for the
For example, d = D for the components conducting first-, second- and higher-order models to calculate D and
during the switch-on time in both the CCM and the DCM; solve for efficiency versus output current. The LM27313’s
d = (1 – D) for the components conducting during the off output voltage is held at 19 V over the entire load current
time in the CCM and d = (2 ´ IOUT)/ ∆IL for the components range of 0 mA to 40 mA (Fig. 2).
conducting during the off time in the DCM; and lastly, Fig. 3 compares the first-order model, the second-order
d = 1 for the components conducting during both portions model, both with and without switching losses, the higher-
Figure 4

Power Electronics Technology September 2008 34 www.powerelectronics.com

 Boost converter
order model and the measured efficiency. Fig. 4 also com- time increases and a damped ringing develops between the
pares the measured efficiency with the calculated efficiency switching node capacitances and the inductor. Eventually
using the higher-order model as previously described, and this voltage ringing will decay to VIN.
the higher-order model that is slightly modified to account The result is that as the converter transitions more and
for changes in CD and CDS capacitive energy losses at very more into the DCM, the switching node capacitances will
light loads. give a lot of their energy back to the input and essentially
The first-order model again uses the ideal duty-cycle ratio have to charge up to a smaller and smaller voltage. At the
and calculates efficiency at each IOUT . Both second-order extreme minimum current, when the switching node has
models (for dc losses only) and second-order switching discharged down to VIN before the start of a new switching
losses use the F(D2) equation and solve for D at each IOUT period, the drain-to-source and diode capacitance power
using the quadratic formula. The higher-order model uses C loss becomes closer to:
code and iterates through the FCCM(D5) equation by varying VIN 2 ´(C DS + C D )´ fSW
PC = .
D between 0 and 1 while testing for the first zero crossing. 2
At the end of each FCCM(D5) iteration, a comparison is done If the third-order DCM equation is made to reflect this,
between the calculated inductor current ripple and the aver- then the efficiency versus load current becomes closer to the
age inductor current to determine when: actual efficiency at light currents (Fig. 4). The discontinuity
in the modified higher-order plot occurs at the CCM/DCM
æ1 ö I
çç ´DIL ÷÷ > OUT . boundary where the CDS and CD capacitance voltage begins
è2 ø 1- D shifting from VOUT down to VIN.
When this happens, the converter is assumed to be in the Another source of error that was left out of the duty-cycle
DCM, and the FDCM(D3) equation is then iterated to calculate calculations is that due to inductor switching losses (core
D. This is done at each IOUT to generate the efficiency graph. loss and eddy currents). This would tend to decrease the
As would be expected, the first- and second-order (dc efficiency even further when the converter operates in the
only) models show a lot of variation between the calculated DCM due to the large current ripple. Unfortunately, inductor
and actual efficiency. The second-orderHolyStone_PET_May_08
model, which in- ac switching-loss
4/28/08 11:15 values
AM arePage
not commonly
1 given, so they
cludes switching losses, is the closest at low currents while were not used in these examples.  PETech
the higher-order model is closer at
high currents, but tends to deviate at

Holy Stone
low currents when the device oper-
ates further into the DCM.
This can be explained by un-
derstanding how the boost-circuit
capacitances charge and discharge
each switching cycle as the boost A POWERFUL PRESENCE
converter transitions from the CCM
to the DCM. When the converter
is operating fully in the CCM, the
IN POWER APPLICATIONS
drain-to-source and diode capaci- Holy Stone offers a
tances charge and discharge from variety of MLCCs
zero to VOUT , each cycle causing a specifically designed
power loss of: for power applications.
� High Capacitance X7R
VOUT2 ´(C DS + C D )´ fSW for DC/DC Input /Output Filtering
PC = . � High Voltage Capacitors
2
for a variety of power apps
As the converter transitions from � High Capacitance MLCCs
the CCM to the DCM (i.e., the load for DC/DC smoothing
current decreases), the inductor � High Voltage and Safety

current slope will hit zero before Certified Disc Capacitors

the end of the switching period. If
the inductor current hits zero well
before the end of the switching pe-
riod, the inductor current will begin
to go negative, causing the energy
in the switching node capacitances
to return back to VIN. As the load
current is further reduced, the dead
Holy Stone MLCCs are available
� Stacked High C/V
with Arc Prevention Coating
for Switch Mode Power Supplies and SuperTerm™, a polymer layer that
reduces the risk of internal cracks.
HolyStone International
A subsidiary of Holy Stone Enterprise Co.
26395 Jefferson Ave., Suite H, Murrieta, CA 92562 Tel: 951-696-4300 Fax: 951-696-4301
Email: info@holystonecaps.com Web: www.holystonecaps.com
In Europe: www.holystoneeurope.com In Asia: www.holystone.com.tw

www.powerelectronics.com 35 Power Electronics Technology September 2008