Application Note 18

Micrel

Application Note 18
MIC38C43 Off-Line Reference Design
by Jeff Dixon
the power supply designer to increase efficiency, reduce
board space, and reduce the start-up resistors from 1W to
14W in most wide input applications. The MIC38C4x family
has many other value-added features such as 500kHz switching, 40C to +85C operation, 40ns rise and 30ns fall times,
and pin-for-pin compatibility with the bipolar UC3842/43/
44/(A) and also the UC2842/3/4. Factory trimmed oscillator
discharge current and bandgap reference (1%) simplifies
design and builds confidence for the designer where high
volume manufacturing demands repeatability and ultralow
standard deviations in these tolerance-critical areas. The
output section of the MIC38HC4x version has been beefedup with a 1A peak current capability, eliminating the need for
a gate-drive transformer in many applications.
These are applications such as VCRs, power tools, laptop
and notebook computers, digital cameras, appliances, battery charges, video monitors, and dc-dc converters.

Introduction
This application note for the MIC38C43 provides a design tool
for the project engineer. This note also provides a template
for future projects: the user need only scale the appropriate
sections. Designing an off-line discontinuous flyback power
supply for the first time can become overwhelming. The task
involves many disciplines of electrical engineeringmagnetics, filter, feedback, layout, and thermal issues to name a few.
The design procedure can be broken down into two parts.
The reader starts with (1) an in-depth, block-by-block theory
of operation and follows with (2) a step-by-step practical
example. This example includes specifying the major components, including the magnetics. The schematic, PCB layout,
electrical waveforms, and a bill of materials are also included.
The MIC38C4x family of BiCMOS current-mode PWM controllers represents another technological advance in MICRELs
switch-mode power supply line of ICs. Among power supply
designers, the bipolar 384x family is probably the best known
of the current-mode PWM controllers currently in use today.
However, instead of using the standard bipolar technology
used by our competitors, the MIC38C4x ICs use a BiCMOS
process. Comparing the MIC38C4x to its bipolar competitors,
the start-up current has been reduced from 500A to 50A,
and operating current is 4mA instead of 11mA. This allows

This MIC38C43 reference design is an example of an ultralow-cost, off-line, isolated, discontinuous-mode, flyback power
supply using the Micrel MIC38C43BN current-mode PWM
controller IC. This design example features a universal input
(85Vac to 265Vac) in a scalable, manufacturable design.
Figure 2 is the circuit diagram.

VDD
7

35V
5V
Reference

VREF
8

RT/CT

UVLO

OUT

Oscillator

FB

2R

2.5V

R
S

MIC38C43
COMP

GND

ISNS

Figure 1. MIC38C43 Functional Diagram

Micrel, Inc. 1849 Fortune Drive San Jose, CA 95131 USA tel + 1 (408) 944-0800 fax + 1 (408) 944-0970 http://www.micrel.com

May 1998

Application Note 18

Application Note 18

Micrel
rectifier D4 produces a dc voltage. Energy is then stored in
output capacitor C11. C11 is sized to withstand the ripple
current and peak voltage. A low-pass filter formed by L2 and
C12 attenuates the high-frequency harmonics to an acceptable level.
Feedback Blocks (FB1 and FB2)

Discontinuous Flyback Supplys

Building Blocks
The off-line flyback power supply can be broken down into
building blocks to make it easier to understand and design.
The following is a block-by-block description of the Micrel
MIC38C43 reference design.
MIC38C43 Controller Block (Figure 1)
The MIC38C43 uses current-mode control to adjust the PWM
waveform duty cycle that determines the output voltage. The
P-channel and N-channel MOSFETs (output inverter) are
alternately switched at a constant frequency; only the duty
cycle (on time) is adjusted.

Feedback for this design is current mode. This means there

are two feedback loops: one that monitors the output voltage,
and a faster loop that monitors the current through the
inductor. Whether it is a forward, buck, boost or a flyback
topology, the current-mode controller will require monitoring
of both loops for standard operation.
Historically, in contrast to current-mode control, the older
voltage-mode control monitors only the output voltage and
current. This secondary-control scheme causes propagation
delays from the supply output to the controller feedback pin.
These delays cause second-order effects that are not acceptable, especially in todays high frequency switching supplies
that operate at ever higher switching frequencies. The advantages of current-mode control over voltage-mode control are
improved line and load regulation, a simpler compensation
network, and better overall transient and load response.
Inner Feedback Loop (Figure 2, Feedback Loop 1)

A switching cycle begins when the oscillator generates a

reset pulse. This pulse resets the RS latch and turns on Q1
(the power supply main switch) via the logic formed by the OR
gate and inverter output. During this time, the inductor current
(T1 primary) is increasing and stores energy in the inductor.
The PWM comparator measures the voltage at ISNS (current
sense pin) and compares it to the output of the error amplifier.
When the current-sense waveform intersects the error amplifier output waveform, the PWM comparator will set the RS
latch, which turns off Q1 via the logic and inverter. Energy is
then discharged from the inductor until the next switching
cycle begins. By varying the P-channel MOSFET on-time
(duty cycle), the inductor current is adjusted to maintain the
output voltage.
Input Block (Figure 2)

The inner loop (FB1) consists of U1, T1 primary, Q1, R1,

R8,and C7. When Q1 turns on, the current in the primary
ramps up to some peak current, based on the line and load
conditions at that time. Next, current flows through the T1
primary, Q1, R8 and finally to ground. The current flowing
through sense resistor R8 develops a voltage drop proportional to the on-time of Q1. This signal is then fed back to the
ISNS pin of U1. The R1 and C7 form a low pass filter used to
attenuate turn-on spikes and any noise that might corrupt the
ISNS signal. Since this is a current-mode controller, current is
sensed on a cycle-by-cycle basis, allowing for tighter line and
load regulation. When the supply goes into current limit due
to a fault condition, output power is limited by folding back the
current to less than its maximum operating point. Adjusting
R8 sets the current limit. If the current limit is too high,
increasing this resistor value will decrease the current limit
and vise versa.
Outer Feedback Loop (Figure 2 Feedback Loop 2)

The input stage of the supply consists of a common-mode

choke, an X-type (line-to-line) safety film capacitor, two Ytype safety ceramic capacitors, one (line-to-ground) and one
(neutral-to-ground). The input stage also contains a slowblow fuse, bulk capacitor, and a diode bridge.
A slow-blow type fuse is used to avoid potential nuisance
blowing of the fuse. This is due to the instantaneous current
drawn by the bulk capacitor at initial turn-on. The slow-blow
fuse also allows for a lower, safer current rating.
Noise/EMI (common-mode currents) normally reflected back
onto the ac line are attenuated by L1 and Y-type capacitors
C2 and C3. The Y-type capacitors are sized as large as
possible to reduce the common-mode noise and still comply
with UL/CSA/VDE leakage current requirements. This is why
off-line supplies that must pass safety agency testing usually
use either 4.7nF or 2.2nF for 120Vac or 230Vac operation,
respectively. X-type capacitor C1 and the leakage inductance of L1 attenuate the differential-mode EMI.

The output voltage is regulated using a TL431 (U3) shunt

regulator that senses any perturbation in output voltage via
the R10-R11 voltage divider. The voltage divider is configured for a divided down voltage of 2.5V for an output voltage
of 5.0V. The divider output voltage is compared to the internal
2.5V reference of the TL431 that sinks the LED cathode of an
optoisolator. R12 sets overall dc gain and keeps the LED
current within its safe operating area. The TL431 biases the
cathode of the optoisolators diode more negative by pulling
it closer to ground, and lets up when the output voltage comes
into regulation. The optoisolator transistor is biased within its
linear region causing a voltage drop across R6, which is seen
by the controllers error amp input. This is compared to the
internally generated 2.5V reference. Components C13, C14,
and R15, via the CTR (current transfer ratio) of the optoisolator set the overall control-loop frequency response.

Note: X-type and Y-type agency approved safety capacitors

must be used in these locations if agency approval is a
requirement.
The fixed impedance of L1 (common-mode choke) provides
sufficient inrush protection for this cost-conscious design.
Other devices, such as NTCs and thermistors, are available
if the instantaneous input current becomes an issue.
Output Block (Figure 2)
On the primary side, when the load increases, the control loop
increases the duty cycle. On the secondary side Schottky

Application Note 18

May 1998

Application Note 18

Micrel

85 to 264Vac
50/60Hz
Hot

F1

Ground

1A

BR1
DBR1

L1
C1
0.1F
250V

Neutral
C2
2200pF
400V

C14
3000pF
50V
R15
47.5k
C13
1%
1000pF
50/63V
R11
4.99k
1%

C4
47F
400V

20mH
C3
2200pF
400V

U2b
2501

R10
4.99k
1%

D4
12CTQ045

C10
U2a
0.1F
50/63V
2501
U1
MIC38HC43BN
R5
1
COMP
VREF
1.21k 1%
2
FB
VDD
R6
3
1.21k
ISNS
VOUT
1%
4

RT/CT

D2
18V

10 1%

R1

3T

C8
22F
25V

6T

C11
1200F
10V

C12
220F
10V

R14
200
1%
Return

T1
U6939-D

Q1
IRFIBE30G

C7
1k 1%
470pF
63V

D1
1N4448

C5
0.1F
50/
63V

VOUT
+5V/2A

L2

45T
R4
34
1%

U3
TL431

5H

4
7

R13

R7
14k 1%
C6
470pF
63V

GND

R2
332k
1%
R3
332k
1%

R12
1k
1%

C9
100pF
1kV

R8
0.68
1/4W
1%

R9
470
1/2W

D3
UF4005

Figure 2. Universal-Input Discontinuous Flyback Supply

Universal-Input Discontinuous Flyback

Design Example
VIN(peak) = 85V

Input Specification
(1) VIN(min) = 85Vac, VIN(max) = 265Vac
(2)

(8a) VIN(peak) = 120.21V

For the average voltage, subtract half of the ripple voltage
and the forward drop of the full bridge rectifier:

fLINE(min) = 50Hz, fLINE(max) = 60Hz

fSW = 200kHz

Output Specification
(3) VOUT = 5.0 1%
(4)
(5)
(6)

(9)

IOUT(min) = 0.2A, IOUT(max) = 2A

Load regulation = 0.2%
Line regulation = 0.2%

May 1998

VRIPPLE
2

VF(diode)

VRIPPLE = 15% VIN(peak)

VF(diode) = 1V
then:
VDC(low) = 120V

Given the need for tight load and line regulation, select the
TL431 precision voltage reference for the feedback loop. The
38C4x reference provides constant voltage for the RT/CT
oscillator.
Calculate the peak voltage at VIN(min) and VIN(max).
Determine the peak voltage at VIN(min)
VIN(peak) = VIN

VDC(low) = VIN(min)

where:

(7) VOUT(ripple) = 10mV

Voltage Reference

(8)

120V 0.15
1V
2

round to 100V
(9a) VDC(low) = 100V
Calculate the DC Rail at High Line AC
Determine the peak voltage at VIN(max). Using formula (8):
(10) VIN(peak) = VIN(max)

Application Note 18

Application Note 18

Micrel
Note: By using a maximum on-time of 48% at 200kHz we
have allowed for less rise and fall time, however, this is
outweighed by the fact that the core volume goes down by
half for every factor of four we increase operating frequency.

VIN(peak) = 265V 2
(10a) VIN(peak) = 374.71V
For the average voltage, subtract half of the ripple voltage
and the forward drop of the full bridge rectifier, using Equation (11):

t ON = Dmax

375 0.015
= 375V
1
2

(11) VDC(high)

t ON = 2.4 10-6 s

where:

(11a) VDC(high) = 345V

Input/Bulk Capacitor

Dmax = 0.48

The input capacitor is used to hold up the rectified line voltage

in a off-line application. The size of this capacitor is based on
ripple current, the peak voltage across the capacitor and the
energy storage requirements for a particular application.
It is a good design rule to keep the ripple voltage on the input
capacitor between 10% and 15% when operating under
worst case conditions (low-line, full-load). This means that at
the minimum line input voltage with a full load on the supplys
output, the peak-to-peak voltage ripple will make up only 15%
of the peak voltage seen across the capacitor. The VP-P ripple
is a function of the capacitors ESR and the ripple current. If
the capacitor is too small, the ripple component will become
too large. This causes the main switch to see a lower average
voltage and results in lower overall system reliability and
efficiency.
Calculate the necessary capacitance in microfarads. Assume 80% efficiency.

fOP = 200kHz
t =

1
fOP

LPRI =

0.9

POUT = 10W
Peak primary current is used to determine the center gap for
the core and the number of primary turns.

IPEAK =

(VIN(min) tON )
LPRI

Use 100V for some margin at the ac low line.

where:
VIN = 100V
tON = 2.410-6

12
= 0.12 joules
2 50

2 EIN

LPRI = 414H

VMIN(avg) 2

then:

2 0.12J
C =
= 54.54F
120 2 100 2

IPEAK = 0.579A
Center Gap Formula (Figure 4a)

Select the closest standard value, a 47F, 400V, 105C

aluminum electrolytic for the C4 bulk capacitor.

lGAP =

Designing the Flyback Transformer

A TDK PC40 EE22-Z core and a TDK BE-22-118CP vertical
bobbin were selected for this design.
The maximum recommended on-time is 48% of the full
period. This keeps the regulator operating in discontinuous
mode under worst case conditions. The on-time is used in
Equation 2 to calculate the maximum primary inductance.

Application Note 18

2.5 t POUT

where:
VIN(min) = 100V
tON = 2.410-6

2 fMIN

VPK

LPRI = 414.72H

PIN

(VIN(min) tON )

then:

PO = 10W
PIN = 10W 1.2 = 12W
Then:

C =

= 5 10 -6

The maximum primary inductance for discontinuous operation under worst case conditions (low line voltage, full load) is:

fMIN = 50Hz

EIN =

fOP

then:

Round to 345V.

EIN =

0.4 L PRI IPK 2 10 8

A e (BMAX )

At this time it is important to note that a flyback transformer

operates in the first quadrant and is considered unipolar,
unlike a push-pull topology that operates in two quadrants.
This subject is greater than the scope of this paper and should
be researched before designing your magnetics. For further
information please consult your magnetics suppliers data
book.

May 1998

Application Note 18

Micrel
Secondary Wire Size
The secondary also needs to be at least 300CMA.

where:
BMAX = 1300 (gauss)
Conversion note: 1300 gauss = 130 millitesla
The core area is found in Figure 4a under Ae.
Ae = 0.41cm2

CMA SEC =

lGAP = 0.0252cm = 9.91710-3 inches

IPEAK(sec) = 7.76A
SECRMS = 3.14ARMS

Number of Primary Turns

Bmax lGAP

CMSEC = 1020CM
The secondary is bifilar 23AWG which is 0.0226 inch diameter and is 510CM 2 = 1020 circular mils.
Therefore;

0.4 IPEAK

then:
NPRI = 45 turns
Number of Secondary Turns
NSEC =

SECRMS

where:
secondary diameter = 0.0228 inch

then:

NPRI =

CMSEC

NPRI VOUT + VDIODE

) (1

(VIN (min) ) Dmax

Dmax

CMASEC = 1020CM/3.14A = 325CM/A [OK]

Primary Turns Per Layer

The number of primary turns/layer is based on wire size and

the bobbin width (BW). This is the C dimension in the TDK
catalog. See Figure 4d.

where:
PRITperlayer (max) =

VDIODE = 0.6V
VOUT = 5V
VCC = 11V

where:
BW = 0.332 inch
PRIAWGDIA = 0.0112 inch

then:
NSEC = 3 turns

Single-build 30AWG is 0.0112 inches diameter.

then:

Number of Tertiary Turns

Once the controller has gone through its start-up, the tertiary
winding powers the MIC38HC43 controller.
NBIAS =

NPRI VCC + VDIODE

) (1

(VIN (min) ) Dmax

Dmax

PRIT/Layer(max) = 29.643 T/layer < 45 turns [not OK]

So split the primary in to two layers, (1) 22 turns and (2) 23
turns.
Secondary Turns Per Layer

The secondary uses triple insulation to meet safety agency

approvals. To keep the magnetics compact and efficient we
have gone to a bifilar secondary approach.
SECMAX(dia) = 0.035 inch with triple insulation

then:
NBIAS = 6 turns
Use the primary wire size to minimize the number of different
wire sizes and reduce cost.
Primary Wire Size
All wire needs to be at least 300 circular mils per ampere
(CMA). The area of 30AWG is 100 circular mils.

CMAPRI =

2 SECMAX(dia) = 0.070 inch

SEC T/LAYER (max) =

SECT/LAYER(max) = 4.74 T/Layer [OK < 3T NSEC]

Winding Instructions
At this point we should have two primary winding layers
wound and separated by mylar tape. Tape over the second
primary winding and add the 6-turn tertiary winding. Then,
tape over the tertiary winding and add the 3-turn secondary.
Core Loss
From page 9 of the 96 TDK catalog, the core loss is
0.07mW/cm3 at 200kHz.
Gauss level = 1300 gauss, then: 1300/10 = 130milliteslas

CMPRI = 100CM
PRIRMS = 0.22
The diameter of 30AWG single-build nylese is 0.0112 inches.
then:

May 1998

BW
SECDIA (max)

then:

CMPRI
PRIRMS

where:

CMA PRI =

BW
PRIAWGDIA

100CM
= 454CM/A [OK]
0.22A

Application Note 18

Application Note 18

Micrel
Input Diode Bridge
When selecting a diode bridge for an off-line supply, the peak
current and peak reverse voltage seen by the bridge must be
calculated. The peak current must be calculated at the
minimum input voltage of the off-line supply, with full load
applied to the output. Assume a conservative power factor of
0.6 and efficiency of 80%. We know from calculating the
magnetics that primary IPK is:

TDK uses sine wave data for its core loss charts. The flyback
topology operates in one quadrant and is considered a
unipolar device.
therefore:
130mT
= 65mT
2
CLOSS = 0.07 (watt/cm3)
(see Figure 5)
CV = 1.61cm3
(see Figure 4a)

IPK = 0.579APK
Use a 2 margin.
Therefore:

PLOSS(core) = CLOSS CV

IDIODE(min) = 2 (0.579APK)
IDIODE(min) = 1.12APK

then:
PLOSS(core) = 0.113W
Bobbin Stack-Up
This is the bobbin stack up of windings and tape versus
bobbin window height for a TDK EE22-PC40 core. Designing
magnetics is an iterative process. If the wire does not fit, then
you will have to go to a larger core and recalculate. (See
Figure 4b and 4c for bobbin dimensions.)
BobHMIN =

[(PRIAWG

DIA

(BIASDIA BL)

The reverse voltage rating of the bridge must be calculated at

the high line-input supply voltage. The load is not important
for this calculation.
VPK(rev) = 1.414 VIN(max)
VPK(rev) = 1.414 265
VPK(rev) = 375V
Select a 400V, 1.5A molded diode bridge.

PL) + SECDIA (max) SL

+ (TL TTH )

)] +

Output Rectifier
The output diode in a flyback supply must be chosen carefully
to optimize the power systems total efficiency. There are two
ways in which a rectifier consumes power: forward conduction loss and reverse recovery loss.

where:

Voltage Rating
To select the correct rectifier for a specific application, the
four main considerations are forward voltage drop (VF),
reverse recovery time, maximum reverse peak working voltage (VRWM), and maximum forward current (IF). The maximum reverse voltage is across the rectifier when the supply
is at high line voltage, therefore, the VRWM calculation is:

A = 0.492 inch
B = 0.311 inch
BobHACTUAL = A B/2
PRIAWGDIA = 0.0112 inch
BobHACTUAL = 0.0905 inch
PL = primary layer
PL = 2

N
VRWM = VO + VIN(peak) SEC
NPRI

SL = secondary layer
SL = 1
SECMAX(dia) 0.035 inch

VRWM = 5V + 375
= 30V

45

TL = tape layers
TL = 4

The next standard value with some margin is 40V.

TTH = tape thickness = 0.001 inch

BIASDIA = 0.0112 inch
BL = 1

Current Rating
The calculation for the maximum forward peak current is
done at VIN(min) and, with the supply at full load.

then:
BobHACTUAL is the bobbins maximum window winding area.
The BobHMIN is the actual stack-up of windings and insulation, therefore the BobHACTUAL must be greater than BobHMIN.

Dmax

IPK = VDC(low)

L PRI fOP
Low line dc rail = 100Vdc

BobHACTUAL = 0.0905 inch

BobHMIN = 0.0724 inch [OK]

Application Note 18

48

IPK = 100V
= 0.579A
414 10 -6 200 10 3

May 1998

Application Note 18

Micrel
Calculate Voltage Spike on Output Capacitor
IPEAK(pri) = 0.579A
VSPK = N IPEAK(pri) ESR
15 0.579 0.29 = 2.52VPK
So the output capacitor will see 2.5 + 5.0 = 7.5V maximum
Voltage rating = 10V
A CEEFM1A122M6 is a 1200F, 10V aluminum electrolytic
capacitor.
Output Filter
The output filter is a low pass filter used to attenuate the
residual noise left on the dc bus after the main hold-up
capacitor. The output inductor should be rated for 3A continuous.
VRIPPLE = 800mV. We need to reduce to 10mV.
We know our fundamental is 200kHz, so we set the filter to
have a break frequency of 5kHz. We also know a single-stage
low-pass filter will roll off at 20dB/decade.
We will be attenuating out past 500kHz, therefore, from 5kHz
to 50kHz is 1 decade and from 50kHz to 500kHz is two
decades.
We have the break frequency set so it cuts in before 5kHz. By
the switching frequencys second harmonic of 400kHz, the
noise is attenuated by almost 40dB. Assume a capacitor
value of 220F.

The peak current the rectifier will see is approximately

IPK(pri) n = 0.579 15 = 8.69APK
Therefore, for this application, we will use a 10A Schottky
center tap diode with a minimum VR of 40V to allow for some
margin. The diode must withstand the rms current and the
peak current. The use of a TO-220 package rectifier with a
center tap allows the diodes to be paralleled and heat sinked
for improved reliability. Another design rule to approximate
the current rating for the output diode in a discontinuous
design is:
ID = 4 IOUT
For a continuous design, the rule is:
ID = 3 IOUT
As in any design, there are trade-offs to be made, such as
choosing discontinuous or continuous mode. This supply
was designed to run in the discontinuous mode to simplify the
feedback control system and to reduce reverse recovery
losses in the rectifier. If this were a continuous design, the
rectifier would be forced to recover while there was still
substantial current flowing through it.
Output Hold-Up Capacitor
When reviewing Marcon Capacitors 96 catalog, notice that
at the end of each section there is derating information based
on temperature and frequency (see Figure 3b). At 45C and
100kHz the catalog shows that we can multiply the 105C
ripple current rating for this series of Marcon capacitor
(CEEF) by 2.4 for a ambient of 45C, worst case. Note even
though the catalogs derating graph only goes to 100kHz,
clearly the trend shows that we can multiply by at least 2.4 for
a ambient of 45C.
Calculate ripple current rating:
ISEC(rms) = IPRI(rms) n

1
L =

2 f C

n =

NPRI

= 4.6H

Lets round this to the next standard value, 5H.

fBREAK =

where:
IPRI(rms) = 0.19ARMS

NPRI

fBREAK =

45
= 15
3

1
2 LC
1
2

40dB = 20 log

then:
ISEC(rms) = 0.19 15 = 2.85A Irms

(5 10 -6

220 10 -6

= 4.8kHz

800mV
800mV
10 -2 =
X
X

800mV
= X = 8mV of peak-to-peak ripple
100
A minimum output load of 10% is used to prevent, (1)
continuous operation, and (2) switching spikes from ratcheting
up the voltage on the output capacitor. Remember, if all else
fails, use Klinkoffs variable constant!

So we need a capacitor that is rated for 2.85ARMS current at

45C ambient.
Then:
IRIPPLE rating = 2.85/2.4 = 1.1875ARMS at 105C
(see Figure 3a)
The corresponding ESR is 0.29, therefore the VRIPPLE will be:
ESR ISEC(rms) = VRIPPLE
0.29 2.85 = 826mV
Note: We will use this later for calculating the output filter.

May 1998

Application Note 18

Application Note 18

9.35 0.15
(0.368 0.006)

22.0 0.3
(0.866 0.012)

5.75 0.25
(0.226 0.010)

0.290

Max. ESR
( )
@120Hz, 20C
0.126

@10C
0.063

@20C

Max. Impedance

+65C
2.19

+45C
2.4

1.73

+85C
1

+105C

1 3 .0
(0.512)

E min.
5.35 0.15
(0.211 0.006)

F
4.3
(0.169)

1,190

@100kHz

8
B
7.9
(0.311)

A
12.5
(0.492)

Dimensions mm (inches)
C
8.45
(0.322)

41.0

39.6

le (mm)

Effective Parameters

10 25

D L (mm)

Nominal
Case Size

1610

Ve (mm3)

Figure 4d. Transformer T1 Core Dimensions

Figure 4b. Transformer T1 Bobbin Dimensions

BE-22-118CP

Part No.

0.970

Ae (mm2)

1,083

@120Hz

Max. Ripple Current

C1 (mm-1)

Figure 4a. Transformer T1 Core Dimensions

5 . 7 5 0 .2 5
(0.226 0.010)

Figure 4c. Transformer T1 Bobbin Dimensions

PC40EE22-Z

Part No.

0.21

Tan

Figure 3a. Capacitor C11 Specifications

360

Max. Leakage
Current (A)
@20C

Figure 3b. Capacitor Maximum Ripple Current at 100kHz

1,200

Capacitance
(F)
@120Hz, 20C

Dimensions mm (inches)

10

CEEFM1A122M6

Rated
Voltage
(Vdc)

Part No.

TDK Ferrite Core and Bobbin Information

Marcon Capacitor Information

2400 25%
2180 25%

AL (nH/N2)

Application Note 18
Micrel

May 1998

Application Note 18

Micrel
(Sine wave data)

Material: PC40
100

500kHz
10

300kHz

Core loss Pc (W/cm3)

65mT @ 200kHz = 0.07W/cm3

200kHz
100kHz
1

50kHz
25kHz

0.1

0.01

0.001
50

65

100

200

300

500

700

Flux density Bm (mT)

Figure 5. Transformer T1 Core Loss

Drain Voltage Waveforms

Transient Load Response

OUTPUT VOLTAGE

Load Transient Response

100% to 10% Load

120Vac Input

120Vac Input

(Q1, VDS)

DRAIN VOLTAGE

VOUT = 100mV/div

IOUT = 10mV/div

200mA

50V/div, 1s/div

100mV/div, 1ms/div | 10mV/div, 1ms/div

Figure 6a.

Figure 7a.

OUTPUT VOLTAGE

Load Transient Response

10% to 100% Load

Q1 Voltage Drain Wave Form

10% Load

120Vac Input

120Vac Input

(Q1, VDS)

1s/div
50V/div

VOUT = 100mV/div
DRAIN VOLTAGE

LOAD CURRENT

1s/div
50V/div

2A
LOAD CURRENT

Q1 Voltage Drain Wave Form

100% Load

2A
IOUT = 10mV/div
200mA

100mV/div, 1ms/div | 10mV/div, 1ms/div

50V/div, 1s/div

Figure 6b.
May 1998

Figure 7b.
9

Application Note 18

Application Note 18

Micrel

Bill of Materials
Item

Part Number

Manufacturer

Description

Value

C1

CFKC22E104M

Nitsuko

fim capacitor

0.1F, 250Vac, X-type

Quantity
1

C2, C3

DE7100F222MVA1-KC

Murata

ceramic capacitor

2200pF, 400V, Y-type

C4

KMG400VB47018X20L

Nippon Chemicon

aluminum electrolytic

47F, 400V

C5, C10 ECU-S1J104KBB

Panasonic

ceramic capacitor

0.1F, 63V

C6, C7

ECU-S1J471JCA

Pansonic

ceramic NPO/COG capacitor

470pF 5%, 63V

C8

CEBSM1E220M

Marcon

aluminum electrolytic, 511

22F, 25V

C9

ECK-D3A101KBN

ceramic capacitor

100pF, 1kV

C11

CEEFM1A122M6

Marcon

aluminum electrolytic, 1025

1200F, 10V

C12

CEEFM1A2211M4

Marcon

aluminum electrolytic, 6.311

220F, 10V

C13

ECU-S1H104KBB

Panasonic

C14

ECU-S1H332JCB

D1

1N4448

D2

1N4746

D3

UF4005

D4

12CTQ045

DBR1

KBPO4M

F1

263 001

J1

1725669

Phoenix

connector

J2

1725656

Phoenix

connector

L1

R5-203F7D3

FDK

inductor

20mH, 800mA

L2

CTX-5-3-FR

Coiltronics

radial inductor

5H, 3A

Q1

IRFIBE30G

International Rectifier

N-channel MOSFET

800V, 3

NPO/COG ceramic capacitor

1000pF, 63V

COG ceramic capacitor

3300pF 5%, 50V

Motorola

diode

100V

Motorola

Zener diode

18V, 5W

General Instruments

UFR diode

600V, 1A

International Rectifier

diode

45V, 12A

General Instrument

molded bridge

400V, 1.5A

Littlefuse

Picofuse

250V, 1A

1
1
1
1

R1

metal film resistor

1.0k 1%, 1/4W

R2, R3

metal film resistor

332k 1%, 1/4W

R4

metal film resistor

34.0 1%, 1/4W

R5, R6

metal film resistor

1.21k 1%, 1/4W

R7

metal film resistor

14k 1%, 1/4W

R8

metal film resistor

0.68 1%, 1/4W

R9

metal film resistor

470 5%, 1/2W

R10, R11

metal film resistor

4.99k 1%, 1/4W

R12

metal film resistor

1.0k 1%, 1/4W

R13

metal film resistor

10.0k 1%, 1/4W

R14

metal film resistor

200k 1%, 1/4W

R15

metal film resistor

47.5k 1%, 1/4W

T1

U6939-C

Coilcraft

transformer*

U1

MIC38C43BN

Micrel

current-mode PWM controller

U2

2501

NEC

optoisolator

U3

TL431

Motorola

voltage reference

user dependent

heat sink

* tel: (847) 639-6400, contact Frank Concialadi

Application Note 18

10

May 1998

Application Note 18

Micrel

Printed Circuit Layout

Silk Screen

Component Side

Solder Side

May 1998

11

Application Note 18

Application Note 18

Micrel

MICREL INC. 1849 FORTUNE DRIVE SAN JOSE, CA 95131

TEL

+ 1 (408) 944-0800

FAX

+ 1 (408) 944-0970

WEB

USA

http://www.micrel.com

This information is believed to be accurate and reliable, however no responsibility is assumed by Micrel for its use nor for any infringement of patents or
other rights of third parties resulting from its use. No license is granted by implication or otherwise under any patent or patent right of Micrel Inc.
1998 Micrel Incorporated

Application Note 18

12

May 1998