AN-4159 - Fairchild Semiconductor

www.fairchildsemi.com
AN-4159
Green Mode Fairchild Buck Switch FSL336LR
Introduction
This application note describes a detailed design method
and procedure for a buck offline converter. Design
consideration and formula are presented. The FSL336LR is
designed for non-isolated topologies; such as buck, buckboost converter, and non-isolated flyback converter. This
device is an integrated current-mode Pulse Width
Modulation (PWM) controller and SenseFET. The
integrated PWM controller includes: 10 V regulator for no
external bias circuit, Under-Voltage Lockout (UVLO),
Leading-Edge Blanking (LEB), optimized gate turn-on
/turn-off driver, EMI attenuator, Thermal Shutdown (TSD),
temperature-compensated precision current sources for
optimized loop compensation, and fault-protection circuitry.
Protections include: Overload Protection (OLP), OverVoltage Protection (OVP), and Feedback Open-Loop
Protection (FB_OLP). FSL336LR offers stable soft-start
performance during startup. The internal high-voltage
startup switch and the Burst-Mode operation for extremely
low operating current reduce the power loss in Standby
Mode. As the result, this device is able to achieve power
loss less than 25 mW with external bias and 120 mW
without external bias at 230 VAC.
When compared to a linear power supply, the FSL336LR
reduces total size and weight; while increasing efficiency,
productivity, and system reliability. Application of the
FSL336LR is suitable for cost-effective platform designs.
DVcc
FSL336LR
RA
Vcomp VFB
CF1
RF
Bridge
Rectifier
ILIMIT
CF2
Drain
VCC
Drain
GND
CVcc
RPEAK
RB
L
DF
CDC
DFB
CFB
COUT
RDUMMY
VO
AC
Line
Figure 1.
© 2013 Fairchild Semiconductor Corporation
Rev. 1.1 • 12/30/14
Typical Application
www.fairchildsemi.com
AN-4159
APPLICATION NOTE
Device Block Description
Startup Circuit and Soft-Start
Feedback Control
During startup, an internal high-voltage current source (ICH)
of the high-voltage regulator supplies the internal bias
current (ISTART) and charges the external capacitor (CA)
connected to the VCC pin, as illustrated in Figure 2. This
internal high-voltage current source is enabled until VCC
reaches 10 V. During steady-state operation, this internal
high-voltage regulator (HVREG) maintains VCC with 10 V
and provides operating current (IOP) for all internal circuits.
Therefore, FSL336LR needs no external bias circuit. The
high-voltage regulator is disabled when the external bias is
higher than 10 V.
The FSL336LR employs current-mode control with a
transconductance amplifier for feedback control, as shown
in Figure 4. Two resistors are typically used on the V FB pin
to sense output voltage. An external compensation circuit is
recommended on the VCOMP pin to control output voltage. A
built-in transconductance amplifier accurately controls
output voltage without external components, such as Zener
diode and transistor.
Drain
VOUT
VDC.link
VBIAS
Transconductance
Amplifier
VFB
Greenmode
Controller
IPK
3R
4
VREF
6,7
OSC
D1
D2
PWM
LEB
R
Gate
driver
Drain
VCOMP
6,7
VCC
ICH
CC1
CC2
10V HVREG
3
RC1
Figure 4.
ISTART (during startup)
Iop (during steady-state operation)
CA
RSENSE
5
Pulse-Width Modulation (PWM) Circuit
Transconductance Amplifier (gm Amplifier)
VBIAS
Figure 2.
UVLO
The output of the transconductance amplifier sources and
sinks the current to and from the compensation circuit
connected on the VCOMP pin (see Figure 5). This
compensated VCOMP pin voltage controls the switching duty
cycle by comparing it with the voltage across the RSENSE.
When the feedback pin voltage exceeds the internal
reference voltage (VREF) of 2.5 V; the transconductance
amplifier sinks the current from the compensation circuit,
VCOMP is pulled down, and the duty cycle is reduced. This
typically occurs when input voltage is increased or output
load is decreased. A two-pole and one-zero compensation
network is recommended for optimal output voltage control
and AC dynamics.
Startup Block
The internal soft-start circuit slowly increases the SenseFET
current after it starts. The typical soft-start time is 10 ms, as
shown in Figure 3, where progressive increments of the
SenseFET current are allowed during startup. The pulse
width to the power switching device is progressively
increased to establish the correct working conditions for
transformers, inductors, and capacitors. The voltage on the
output capacitors is gradually increased to smoothly
establish the required output voltage. Soft-start also helps
prevent transformer saturation and reduces the stress on the
secondary diode.
IEA [A]
1.25ms
Sinking current 12µA at
2.525V
+24µA
-24µA
ILIM
Sourcing current 12µA at
2.475V
Soft-Start Envelope
0.2ILIM
GM [µmho]
960µmho
Drain Current
8-Steps
Figure 3.
480µmho
t
Soft-Start Function
VFB
2.45V
Figure 5.
© 2013 Fairchild Semiconductor Corporation
Rev. 1.1 • 12/30/14
VREF
(2.5V)
2.55V
VFB
Characteristics of gm Amplifier
www.fairchildsemi.com
2
AN-4159
APPLICATION NOTE
Pulse-by-pulse Current Limit
(40 ms) circuit determines whether it is a transient situation
or a true overload situation (see Figure 6). The currentmode feedback path limits the maximum power current and,
when the output consumes more than this maximum power,
the output voltage (VO) decreases below its rated voltage.
This reduces feedback pin voltage, which increases the
output current of the internal transconductance amplifier.
Eventually VCOMP is increased. When VCOMP reaches 3 V,
the fixed OLP delay (40 ms) is activated. After this delay,
switching operation is terminated, as shown in Figure 7.
Because current-mode control is employed, the peak current
flowing through the SenseFET is limited by the inverting
input of PWM comparator, as shown in Figure 4. Assuming
50 µA current source flows only through the internal
resistors (3R + R = 46 kΩ), the cathode voltage of diode D2
is about 2.4 V. Since D1 is blocked when VCOMP exceeds
2.4 V, the maximum voltage of the cathode of D2 is
clamped at this voltage. Therefore, the peak value of the
current of the SenseFET is limited.
OSC
Leading-Edge Blanking (LEB)
At the instant the internal SenseFET is turned on; primaryside capacitance and the secondary-side rectifier diode
reverse recovery of the flyback application, the
freewheeling diode reverse recovery, and other parasitic
capacitance of the buck application typically cause a highcurrent spike through the SenseFET. Excessive voltage
across the sensing resistor (RSENSE) leads to incorrect
feedback operation in the current-mode control. To counter
this effect, a Leading-Edge Blanking (LEB) circuit (see
Figure 4) inhibits the PWM comparator for a short time
(tLEB) after the SenseFET is turned on.
3R
LEB
Q
R
Q
Gate
driver
VCOMP
RSENSE
40ms
delay
OLP
VOLP
Figure 6.
Overload Protection Internal Circuit
Vcc
HVREG
VSTART
VSTOP
Protection Functions
The protective functions include Overload Protection
(OLP), Over-Voltage Protection (OVP), Under-Voltage
Lockout (UVLO), Feedback Open-Loop Protection
(FB_OLP), and Thermal Shutdown (TSD). All of the
protections operate in Auto-Restart Mode. Since these
protection circuits are fully integrated in the IC without
external components, reliability is improved without
increasing cost and PCB space. If a fault condition occurs,
switching is terminated and the SenseFET remains off. At
the same time, internal protection timing control is activated
to decrease power consumption and stress on passive and
active components during auto restart. When internal
protection timing control is activated, VCC is regulated with
10 V through the internal high-voltage regulator until
switching is terminated. This internal protection timing
control continues until restart time (650 ms) is counted.
After counting to 650 ms, the internal high-voltage regulator
is disabled and VCC is decreased. When VCC reaches the
UVLO stop voltage, VSTOP (7 V); the protection is reset and
the internal high-voltage current source charges the VCC
capacitor via the drain pin. When VCC reaches the UVLO
start voltage, VSTART (8 V); normal operation resumes. In
this manner, auto restart can alternately enable and disable
the switching of the power SenseFET until the fault
condition is eliminated.
20ms
Ids
Power on
40ms 650ms
SS 40ms 650ms
Normal
with SS
Overloading
Overloading
Overloading
Disappear
Overloading
Disappear
Figure 7. Overload Protection (OLP) Waveform
Over-Voltage Protection (OVP)
If any feedback loop components fail due to a soldering
defect, VCOMP climbs up in manner similar to the overload
situation, forcing the maximum current to be supplied to the
SMPS until OLP is triggered. In this case, excessive energy
is provided to the output and the output voltage may exceed
the rated voltage before OLP is activated. To prevent this
situation, an Over-Voltage Protection (OVP) circuit is
employed. In general, output voltage can be monitored
through VCC and, when VCC exceeds 24.5 V, OVP is
triggered, resulting in switching termination. To avoid
undesired activation of OVP during normal operation, V CC
should be designed below 24.5 V (see Figure 8).
OSC
3R
Overload Protection (OLP)
Overload is defined as the load current exceeding a set level
due to an unexpected event. In this situation, the protection
circuit should be activated to protect the SMPS. However,
even when the SMPS operates normally, the OLP circuit
can be enabled during the load transition or startup. To
avoid this undesired operation, an internal fixed-delay
© 2013 Fairchild Semiconductor Corporation
Rev. 1.1 • 12/30/14
PWM
R
5
OLP
S
PWM
LEB
R
2
VCC
OVP
S
Q
R
Q
Gate
driver
RSENSE
OVP
VOVP
Figure 8.
Over-Voltage Protection Circuit
www.fairchildsemi.com
3
AN-4159
APPLICATION NOTE
Feedback Open-Loop Protection (FB_OLP)
In the event of a feedback loop failure, especially a shorted
lower-side resistor of the feedback pin; not only does VCOMP
rise in a similar manner to the overload situation, but V FB
starts to drop to IC ground level. Although OLP and OVP
also can protect the SMPS in this situation, FB_OLP can
reduce stress on the SenseFET. If there is no FB_OLP, the
output voltage is much higher than the rated voltage before
OLP or OVP trigger. When VFB drops below 0.5 V,
FB_OLP is activated, switching off. To avoid activation
during startup, FB_OLP is disabled during soft-start.
OSC
3R
VOUT
As output load condition is reduced, the switching loss
becomes the largest power loss factor. FSL306LR uses the
VCOMP pin voltage to monitor output load condition. As
output load decreases, VCOMP decreases and switching
frequency declines, as shown in Figure 11. Once VCOMP
falls to 0.8 V, the switching frequency varies between
21 kHz and 23 kHz before Burst Mode operation. At
Burst Mode operation, random frequency fluctuation
still functions.
FB_OLP
S
PWM
LEB
R
RH
Green Mode Operation
R
Random Frequency
modulation range
Switching frequency
Q
Gate
driver
Q
VFB
53 kHz
RSENSE
4
47 kHz
FB_OLP
RL
VFB_OLP
Figure 9.
Feedback Open-Loop Protection Circuit
Thermal Shutdown (TSD)
The SenseFET and control IC integrated on the same
package makes it easier to detect the temperature of the
SenseFET. When the junction temperature exceeds 135°C,
thermal shutdown is activated. FSL336LR is restarted after
the temperature decreases to 60°C.
23 kHz
21 kHz
1.9V
VBURL VBURH 0.8V
VCOMP
Figure 11. Green Mode Operation
Adjusting Current Limit
As shown in Figure 12, a combined 46 kΩ internal
resistance (3R + R) is connected to the inverting lead on the
PWM comparator. An external resistance of RX on the ILIMIT
pin forms a parallel resistance with the 46 kΩ when the
internal diodes are biased by the main current source of
50 µA. For example, FSL336LR has a typical SenseFET
peak current limit of 1.8 A. Current limit can be adjusted to
1 A by inserting RX between the ILIMIT pin and the ground.
The value of the RX can be estimated by Equation (1):
Burst Operation
To minimize power dissipation in Standby Mode,
FSL336LR enters Burst Mode. As the load decreases, the
compensation voltage (VCOMP) decreases. As shown in
Figure 10, the device automatically enters Burst Mode when
the feedback voltage drops below VBURL. At this point,
switching stops and the output voltages start to drop at a rate
dependent on the standby current load. This causes VCOMP to
rise. Once it passes VBURH, switching resumes. VCOMP then
falls and the process repeats. Burst Mode alternately enables
and disables switching of the SenseFET and reduces
switching loss in Standby Mode.
1.8 A: 1 A = (46 kΩ + RX): RX
VO
VFB
(1)
Transconductance
Amplifier
4
Voset
VBIAS
VREF
VCOMP
VBURH
VCOMP
VBURL
IPK
3R
PWM
5
R
ILIMIT
IDS
3
VSENSE
RX
Figure 12. Current Limit Adjustment
VDS
time
t1
Switching
disabled
t2
t3
Switching
disabled
t4
Figure 10. Burst Mode Operation
© 2013 Fairchild Semiconductor Corporation
Rev. 1.1 • 12/30/14
www.fairchildsemi.com
4
AN-4159
APPLICATION NOTE
Detail Design Procedure
VDC.min  2VAC.min 2
System Specifications




Line voltage range (VAC.min and VAC.max): standard
worldwide input line voltage ranges are 85-264 VAC for
universal input, 195-264 VAC for European input range
Line frequency (fL): 50 or 60 Hz
Output voltage (VO)
Estimated efficiency: η
(3)
VDC.max  2VAC.max
(4)
where DCH is the DC link capacitor charging duty ratio
defined, as shown in Figure 13; which is typically about
0.15 for full-wave rectification and about 0.3 for halfwave rectification. Equations (2) and (3) are minimum
link voltage of full-wave and half-wave rectification,
respectively, and Equation (4) is maximum link voltage.
Determining AC Input Rectification Type
The typical AC-DC SMPS solution rectifies AC input with
full-wave rectification. However, half-wave rectification can
be selected for under 3 W designs with buck and buck-boost
topology to reduce cost. For designs >3 W, full-wave
rectification is typically selected to reduce the size of the
input capacitor with small ripple voltage.
Determining Operation Mode
Before selecting the inductor, freewheeling diode, and
output capacitor; the operating mode should be determined:
Continuous Conduction Mode (CCM) or Discontinuous
Conduction Mode (DCM). DCM has smaller inductor size,
lower freewheeling diode cost, and higher efficiency due to
lower switching loss in low-power buck applications.
However, DCM requires a higher current limit and increases
output voltage ripple. Therefore, compromised selection is
needed according to the system requirements.
Determining DC Link Capacitor (CDC) and
DC Link Voltage Range
The DC link capacitor is selected by rectification type and
input voltage range. For full-wave rectification, it is typical
to select the DC link capacitor as 2-3 µF per watt of input
power for universal input range (85-264 VAC) and 1 µF per
watt of input power for European input range (195264 VAC). DC link capacitance of half-wave rectification is
twice full-wave rectification: 4-6 µF per watt of input power
for universal input range (85-264 VAC) and 2 µF per watt of
input power for European input range (195-264 VAC).
Figure 13 shows the input voltage waveform of full-wave
and half-wave rectification.
Table 1. Brief Comparison of CCM and DCM
CCM
DCM
Larger
Smaller
Efficiency (Switching Loss)
Lower
(Larger)
Higher
(Smaller)
Output Voltage Ripple
Smaller
Larger
Lower
Higher
Output Inductor Size
Current Limit
Higher current limit means that, potentially, a highercurrent-rated device may be needed to deliver maximum
output power.
VIN,min
TCH
Selecting Freewheeling Diode
DCH = TCH / TL
Although a transformer for buck topology doesn’t exist,
other leakage inductance and capacitance creates a voltage
spike on the freewheeling diode when the SenseFET is
turned off. Since this voltage spike must be considered,
typically 30% voltage derating of maximum DC input is
required, as described Equation (5).
TL
VIN,min
TCH
2 PO  ( 1-DCH )
η  CDC  f L
VRRM  1.3 VDC,max
DCH = TCH / TL
The diode is one of the components generating high
temperature in the SMPS. To decide the current rating of
freewheeling diode, consider thermal performance of 150%
design margin of the output full load current, as
recommended in Equation (6):
TL
Figure 13. Bridge Rectifier and Bulk Capacitor
Voltage Waveform
Selecting AC rectification, the link voltages are obtained as:
VDC.min  2VAC.min 2
2 PO  ( 1/ 2-DCH )
η  CDC  f L
© 2013 Fairchild Semiconductor Corporation
Rev. 1.1 • 12/30/14
(5)
I F(AV)  2.5  I O
(2)
(6)
where VRRM is peak repetitive reverse voltage and IF(AV)
denotes average rectified forward current.
www.fairchildsemi.com
5
AN-4159
APPLICATION NOTE
Reverse recovery time is also an important factor to choose
the freewheeling diode. The smaller the reverse-recovery
time, the lower the switching loss.
where:
Table 2. Quick Selection Guide for Freewheeling
Diode for Universal Input Range
Part #
VRRM
IF(AV)
trr
Package Type
ES1J
600 V
1A
35 ns
DO-204AC
UF4005
600 V
1A
75 ns
DO-204AL
EGP10J
600 V
1A
75 ns
DO-204AL
EGP20J
600 V
2A
75 ns
DO-204AC
ES3J
600 V
3A
45 ns
DO-214AB
EGP30J
600 V
3A
75 ns
DO201-AD
β
PO
ηV DC. min
I LIMIT
Normally, the buck converter designed for CCM at the
minimum input voltage and full-load condition can enter
DCM as the input voltage increases. The maximum input
voltage guaranteeing CCM operation in full-load condition
is obtained as:
VDC .CCM 
(7)
I
IL 
(8)
I2
I ds. peak
P
IL  O
VOUT
I
IL
IL 
I1
I
2
DCM Operation : L < LBoundary
Figure 14. MOSFET Drain Current
Maximum drain current peak (Ids.peak) at full-load condition
is determined by the selected output inductor. If the
maximum drain current peak is larger than the pulse-bypulse current limit, larger output inductance or higher
current rating of the device is needed. Equations (13) and
(14) show the maximum drain current peak of CCM and
DCM operation, respectively. If this maximum drain current
peak is smaller than pulse-by-pulse current limit, the output
inductor size is optimized through connecting a resistor
between the ILIMIT pin and the IC ground pin.
(9)
f S   (  I ds. peak  0.8V )  22kHz
© 2013 Fairchild Semiconductor Corporation
Rev. 1.1 • 12/30/14
I
2
CCM Operation : L > LBoundary
f S   (  I ds. peak  0.8V )  22kHz
2VDC . min  VO 
Lf S
I ds. peak
IL
I1
Since FSL336LR has a Green Mode, the practical operating
switching frequency at full load can be smaller than f S.HIGH.
By two simultaneous equations representing the relationship
between switching frequency and peak drain current, the
operating switching frequency is calculated. Each equation,
(9) and (10), includes two simultaneous equations. These
are for CCM operation and DCM operation, respectively:
I ds. peak 
(12)
I2
where VOUT is the sum of target output voltage (VO) and
freewheeling diode forward-voltage drop (VF), as
determined by Equation (7), and fS.HIGH is maximum
switching frequency in Green Mode operation, as
illustrated on Figure 11.
VDC . min VDC . min  VO VOUT / VDC . min

VOUT
2 Lf S
VO
2  PO  f S  L
1
2
η  VOUT
where, fS is the operating switching frequency considering
Green Mode.
The inductance operating with Boundary Conduction Mode
(BCM) at minimum input DC voltage is represented in
Equation (8). Smaller inductance than LBoundary can be
selected for DCM operation and larger for CCM operation.
I ds. peak  
(11)
2.4V
V
 VO
 SL  t CLD  DC. min
 t CLD
L
where ILIMIT is the peak current limit; SL is the test slope
(di/dt) of ILIMIT; and tCLD is the current limit delay.
Typically α, ILIMIT, SL, and tCLD are 25.5 kHz/V, 1.8 A,
1.2 A/µs, and 200 ns, respectively.
Selecting Output Inductor

V 
2
η  1- O   VOUT
V
LBoundary   DC.min 
2  PO  f S .HIGH
f S.HIGH  f S.LOW
VGREEN.HIGH  VGREEN.LOW
γ
The forward voltage drop (VF) of the selected freewheeling
diode is an important factor for other equations. Especially
when the equations are related to output voltage, the output
voltage must include forward-voltage drop for more exact
calculation, as shown in Equation (7):
VOUT  VO  VF
α
(10)
www.fairchildsemi.com
6
AN-4159
APPLICATION NOTE
I ds.peak 
PO
ηVOUT
I ds.peak 

VO
1 
VDC.min

2 Lf S

VO
21 
 VDC.min
ηLf S

VOUT

The IC ground is pulsed between input DC voltage and the
ground of output voltage when the SenseFET is turned on and
freewheeling diode is conducted. Output voltage is sensed
through a feedback diode (DFB) during the conduction time of
the freewheeling diode. The feedback diode is typically
selected to remove the difference of forward-voltage drop
between the feedback diode and the freewheeling diode. As
this voltage difference is increased, the output voltage
regulation can degrade.
(13)

 PO

(14)
For an application needs multi-output buck converter with
coupled inductor, refer to the step-by-step design guideline
in Appendix A.
Since the output voltage is sensed only during freewheeling
diode conduction time, a feedback capacitor helps maintain
sensed output voltage, especially for Burst Mode operation. A
value larger than 1 µF is typically recommended. Larger
feedback capacitance results in better output voltage
regulation performance.
Adjusting Pulse-by-Pulse Current Limit
The resistor is determined by Equation (15) and this
adjusted pulse-by-pulse current limit must be higher than
the maximum drain current peak defined by Equations (13)
and (14). This function is disabled by letting ILIMIT pin be
open-circuited, such as:
Two feedback resistors determine output voltage, as in
Equation (18), and, by reducing the voltage difference
between sensed output voltage (VO) and feedback capacitor
voltage (VFB*), more accurate output control is possible:
RX
 I ds. peak
(15)
46kΩ  R X
where ILIMIT is pulse-by-pulse current limit of the FPS,
typically it is 1.8 A. For better noise immunity at ILIMIT
pin, a small capacitor (1 nF~100 nF) is recommended.
I LIMIT.adj  I LIMIT 
RA  RB
(18)
RB
where KREG is regulation factor regarding mismatched
voltage between output voltage (VO) and feedback
capacitor voltage (VFB*). It is typically 2 [V/A].
VFB *  VO  K REG  I O  2.5V 
Selecting the Output Capacitor
Determining Compensation Network
The maximum output voltage ripple is determined by the
output capacitance and the Equivalent Series Resistance
(ESR) of the output capacitor. Since the output voltage
ripple by capacitance is negligibly small when over than
100 µF is selected, the output ripple is mostly determined by
the ESR of output capacitor:
CO.recommend 
Because the FSL336LR employs current-mode control and a
transconductance amplifier (gm amp) internally, a
compensation network can be implemented simply. As
illustrated on Figure 16, a two-pole and one-zero circuit can
secure enough phase margin and bandwidth.
FSL336LR
5
8  ESR  f S
(16)
Vcomp VFB
1
 ESR)  ΔI  ESR  ΔI
(17)
8CO f S
where CO.recommend is the recommended output capacitance,
typically is larger than 100 µF.
Ripple  (
The feedback network is comprised of one diode for sensing
output voltage, one capacitor to maintain sensed output
voltage during the SenseFET turn-on period, and two
resistors to determine output voltage, as shown in Figure 15.
ILIMIT
GND
VCC
Drain
GND
The current control factor, K, is defined as:
DFB
RA
Vcomp VFB
Drain
Drain
Figure 16. Compensation Network
FSL3xx
VCC
CF2
RF
Designing the Feedback Network
Drain
ILIMIT
CF1
RB
CFB
Sensed VO
VFB
K
*
VCOMP

I LIMIT
VCOMP.sat
(19)
where Ids.peak is the peak drain current and VCOMP denotes
the compensation voltage, respectively, for a given fullload condition, ILIMIT is the current limit of the
FSL336LR; and VCOMP.sat is the compensation saturation
voltage, which is typically 2.4 V.
Figure 15. Feedback Network
© 2013 Fairchild Semiconductor Corporation
Rev. 1.1 • 12/30/14
I ds. peak
www.fairchildsemi.com
7
AN-4159
APPLICATION NOTE
To express the small-signal AC transfer functions, the
small-signal variations of compensation voltage (νCOMP) and
output voltage (νO) are introduced as and vˆCOMP and vˆO . For
40 dB
fp
20 dB
CCM operation, the control-to-output function of the buck
converter applying current-mode control is given by:
0 dB
1  s / z
Gvc ( s) 
 Gvc 0
(20)
vˆcomp
1 s / p
where K is specified in Equation (19) and RL is the load
resistance of the output port, defined as VO/IO. The pole
and zero of Equation (20) are expressed as:
fz
vˆo
Low input voltage
-20 dB
fz
-40 dB
1Hz
10Hz
100Hz
1kHz
10kHz
100kHz
Figure 18. DCM Control-to-Output Transfer Function
Variation for Different Input Voltages
Gvc 0  K  RL
(21)
1
1
& p 
ESR  CO
( ESR  RL )  CO
where ESR is equivalent series resistance of the output
capacitor and CO is the output capacitance.
z 
Figure 19 shows the variation of the converter control-tooutput transfer function for variation in the output load
current. Both CCM and DCM operation have similar
variation, where gain is increased and pole is decreased as
output load is decreased.
For DCM operation, the control-to-output transfer function
of the buck converter adopting current-mode control is
given by:
1  s / z
Gvc ( s)  Gvc 0 
1 s / p
Gvc 0  K  VO 
z 
High input voltage
fp
40 dB
fp
fp
20 dB
(22)
Heavy load
0 dB
2   L  f s

V 
PO 1  O 
V
DC 

VDC / VO  1

2  VDC / VO  3
Light load
(23)
1
ESR  CO
-20 dB
fz
-40 dB
fz
1Hz
10Hz
100Hz
1kHz
10kHz
100kHz
Figure 19. Control-to-Output Transfer Function
Variation for Different Output Loads
2  3  VO / VDC
C O 2  ESR  RL  (3  ESR  RL )  VO / VDC 
where η is the efficiency of the converter and VDC is the
input DC voltage.
p 
The transfer function of the compensation network is
obtained as:
Gvc ( s) 
Figure 17 shows the variation of a CCM converter control-tooutput transfer function for various input voltages. DC gain,
pole, and zero do not change for different input voltages.
 pc1 
1  s /  zc
( s /  pc1 ) /(1  s /  pc2 )
(24)
g m  RB
,
(CF 1  CF 2 )  ( RA  RB )
(25)
 1
1 
1

 &  zc 
 pc2

RF CF 1
 CF1 CF 2 
where RA and RB are defined in Figure 15 and RF, CF1 and
CF2 are shown in Figure 16.
40 dB
1

RF
20 dB
fp
Fixed by input voltage variation
0 dB
40 dB
-20 dB
fz
20 dB
fzc
fpc2
-40 dB
1Hz
10Hz
100Hz
1kHz
10kHz
100kHz
0 dB
Figure 17. CCM Control-to-Output Transfer Function
Variation for Different Input Voltages
fpc1
-20 dB
Figure 18 shows the variation of a DCM converter controlto-output transfer function for various input voltages. It has
the lowest DC gain at low-line input condition.
-40 dB
1Hz
10Hz
100Hz
1kHz
10kHz
100kHz
Figure 20. Compensation Network Transfer Function
© 2013 Fairchild Semiconductor Corporation
Rev. 1.1 • 12/30/14
www.fairchildsemi.com
8
AN-4159
APPLICATION NOTE
Design Tips for Compensation Network
a) To secure enough phase margin compensation, the
second pole (fpc2) and zero (fzc) should be separated as
much as possible. Large CF1 and small CF2 are
recommended.
Selecting Dummy Load Resistor
Since the feedback capacitor voltage sensed from output
voltage is not accurately matched with output voltage, the
output voltage regulation can be poor at light-load condition.
The dummy load resistor increases output load and this small
load helps output voltage regulation at light-load condition. A
5~20 kΩ resistor is typically selected.
b) For wide bandwidth of transfer function, compensation
zero (fzc) should be as small as possible.
c)
The recommended minimum capacitance of CF2 is
100~470 pF to avoid noise.
Based on design tips; typically 220 pF, 220 nF, and 75 kΩ
are recommended for CF2, CF1, and RF, respectively.
Design Example
Application
Output Power
Input Voltage Range
Output Voltage / Maximum Current
Home Appliance and Industrial
Auxiliary Power
7.08 W
85-265 VAC
15 V / 0.45 A and 3.3 V / 0.1 A
Description of Schematic




Full-wave rectification is selected for AC line rectification.
For better EMI performance, X-cap (CX1), two fixed inductors instead of line filter (LF001), and Pi-type filter (C1, C2,
L1, L2, and R1) are selected.
For small standby power consumption, VCC is externally supplied from output voltage through D5 and R2.
C8 is used on the ILIMIT pin for better noise immunity.
Small SMD type (1 µF) is used for VCC capacitor.
Coupled inductor is used for 3.3 V output without much loss on positive voltage regulator (U2).
D5
1N4148
R6
75kΩ
0805
C9
220nF
0805
5.Vcomp
R10
3.3kΩ
1206
LF001
330µH // 330µH
BR1
MB6S
C1
10µF
400V
R1
4.7kΩ
1206
L2
Short
R11
3.3kΩ
1206
C10
220pF
0805
C2
10µF
400V
D6
ES1J
Sensed
output
VCC
U1
FSL336LR
L1
330µH
R2
10Ω
0805
R3
120kΩ
0805
D4
ES1J
C4
47µF/25V
4.VFB
C5
2.2µF
0805
3.ILIMIT
7.D
2.VCC
8.D
1.GND
R4
23.2kΩ
0805
VCC
C7
1µF
0805
C6
NC
C8
1nF
0805
R0
NC
1206
R5
NC
0805
C11
47µF/25V
U2
KA78RH33
3.3V output
100mA
12
2
6
10
L3
EFD20
192µH
C3
220µF/25V
v


R7
10k
0805
Sensed
output
15V output
450mA
D3
ES1J
CX1
100nF
R8
NC
1206
R9
NC
1206
VZ1
471KD07
F1
1A/250V
AC
Universal range
Figure 21. Design Example Schematic
© 2013 Fairchild Semiconductor Corporation
Rev. 1.1 • 12/30/14
www.fairchildsemi.com
9
AN-4159
APPLICATION NOTE
Table 3. Bill of Materials for Evaluation Board
Part #
Value
Note
Part#
Value
U1
FSL336LRN
Fairchild Buck Switch
C1
10 µF
400 V Electrolytic Capacitor
U2
KA78RH33
Fairchild Voltage Regulator
C2
10 µF
400 V Electrolytic Capacitor
C3
220 µF
25 V Electrolytic Capacitor
R0
NC
5% 1206 SMD
C4
47 µF
25 V Electrolytic Capacitor
R1
4.7 kΩ
1% 1206 SMD
C5
2.2 µF
0805 SMD
R2
10R
5% 0805 SMD
C6
NC
50 V Electrolytic Capacitor
R3
120 kΩ
1% 0805 SMD
C7
1 µF
0805 SMD
R4
23.2 kΩ
1% 0805 SMD
C8
1 nF
0805 SMD
R5
NC
1% 0805 SMD
C9
220 nF
0805 SMD
R6
75 kΩ
5% 0805 SMD
C10
220 pF
0805 SMD
R7
10 kΩ
5% 0805 SMD
C11
47 µF
25 V Electrolytic Capacitor
R8
NC
5% 1206 SMD
CX1
100 nF
X-Cap 250 VAC
R9
NC
5% 1206 SMD
R10
3.3 kΩ
5% 1206 SMD
D3
ES1J
Fairchild Super-Fast Diode
R11
3.3 kΩ
5% 1206 SMD
D4
ES1J
Fairchild Super-Fast Diode
D5
1N4148
Fairchild Signal Diode
IC
Note
Capacitor
Resistor
Diode
Inductor
LF001
330 µH *2
Axial Type
D6
ES1J
Fairchild Super-Fast Diode
L1
330 µH
Axial Type
BR1
MB6S
0.5 A 600 V Bridge Diode
L2
Jumper Wire
Axial Type
L3
749196521
Flexible Transformer EFD20
F1
1A
Varistor
VZ1
471KD07
Varistor 7Φ 470 V
Fuse
© 2013 Fairchild Semiconductor Corporation
Rev. 1.1 • 12/30/14
250 V Radial Type
www.fairchildsemi.com
10
AN-4159
APPLICATION NOTE
Experimental Results
Table 4. No-Load Input Wattage, Full-Load Efficiency, IC Temperature, Experimental Result
Input Voltage
Input Wattage (No Load)
Efficiency (Full Load)
IC Temperature (Full Load)
85 V / 60 Hz
0.083 W
77.38%
58°C
110 V / 60 Hz
0.083 W
78.35%
54°C
230 V / 60 Hz
0.094 W
77.68%
61°C
265 V / 60 Hz
0.099 W
76.79%
65°C
Experimental Waveforms
CH2:VCC [5V/div]
CH2:VCC [5V/div]
CH1:VDS [100V/div]
CH1:VDS [100V/div]
Figure 22. Normal Operation at Input Voltage 85 VAC
(CH1: VDS, CH2: VCC)
Figure 23. Normal Operation at Input Voltage 265 VAC
(CH1: VDS, CH2: VCC)
CH2:VCC [5V/div]
CH2:VCC [5V/div]
CH1:VDS [100V/div]
CH1:VDS [100V/div]
Figure 24. Burst Operation at Input Voltage 85 VAC
and No Load (CH1: VDS, CH2: VCC)
© 2013 Fairchild Semiconductor Corporation
Rev. 1.1 • 12/30/14
Figure 25.
Burst Operation at Input Voltage 265 VAC
and No Load (CH1: VDS, CH2: VCC)
www.fairchildsemi.com
11
AN-4159
APPLICATION NOTE
Output Voltage Regulation, Experimental Results
Figure 26. 15 V Output Voltage Regulation
Conduction Electromagnetic Interference (EMI) Performance
Att
dBµV
1
100
10
dB
RBW
9
MT
10
PREAMP
OFF
kHz
ms
MHz
10
MHz
90
1
PK
MAXH
2
80
AV
MAXH
TDF
70
EN55022Q
60
PRN
EN55022A
50
6DB
40
30
20
10
0
150
Comment:
Date:
kHz
30
MHz
2-230N
21.JUN.2013
14:27:15
Figure 27. 110 VAC with Full Load Condition
Att
dBµV
1
100
10
dB
MHz
RBW
9
MT
10
PREAMP
OFF
kHz
ms
10
MHz
90
1
PK
MAXH
2
80
AV
MAXH
TDF
70
EN55022Q
60
PRN
EN55022A
50
6DB
40
30
20
10
0
150
Comment:
Date:
kHz
30
MHz
2-230N
21.JUN.2013
14:25:33
Figure 28. 230 VAC with Full Load Condition
© 2013 Fairchild Semiconductor Corporation
Rev. 1.1 • 12/30/14
www.fairchildsemi.com
12
AN-4159
APPLICATION NOTE
Appendix A — Design Guideline of Coupled Inductor

An LDO, which is directly connected to for multi-output as
illustrated on the Figure 29, has poor efficiency and
temperature performance on the LDO itself. To avoid these
problems, a coupled inductor is typically selected (see
Figure 30). An advantage of coupled inductor is achieving
isolation between the two outputs through different ground
connections. This appendix describes step-by-step design
guideline as well as basic operation of a coupled inductor.
FSL336LR
Rectified
AC input
The applied voltage on slave side (VNs):
VNs  (VOm  VFm )  N s / N m .
where VDC is DC input voltage. The number of windings
of master and slave output are represented by Nm and Ns.
VOm and VFm denote the output voltage and the forwardvoltage drop of the freewheeling diode, respectively.
OUT
GND
VNs
VOs
7.Drain
8.Drain
-
LDO
IN
The description of coupled inductor operation when the
freewheeling diode is conducted:
- The applied voltage on master side (VNm):
VNm  (VOm  VFm ) .
VOs
1.GND
t
VOm
Freewheeling diode
for master output
Figure 29. Circuit Diagram of Multi-Output Buck
Converter with LDO
Output diode
for slave output
FSL336LR
tON
VOs
Coupled
inductor
7.Drain
8.Drain
VNs
v
Rectified
AC input
1.GND
Freewheeling diode
for master output
where VFs is forward-voltage drop of the slave output diode.
Rdummy
Figure 30. Typical Circuit Diagram of Buck Converter
Adopting Coupled Inductor
Step 1: Calculate Inductance and Maximum
Drain Current Peak
Step 0: Describe Operation of Buck
Converter Adopting Coupled Inductor
The inductance operating with Boundary Conduction Mode
(BCM) at minimum input DC voltage is represented in
Equation (8) and the inductance is selected as below.
When the internal SenseFET is turned on, the freewheeling
diode is blocked and the VDC-VOm is applied to the master
side of the coupled inductor. According to the winding
notation, the applied voltage on the slave side (VNs) is the
applied voltage to the master side, divided by turn ratio.
Since this VNs is negative, the output diode for slave output
is blocked. During this period, the energy is not delivered to
the slave output.


L > LBoundary for CCM operation
L < LBoundary for DCM operation.
 V

2
η  1- Om   VOm  VFm 
where LBoundary   VDC.min 
2  PO  f s
When the freewheeling diode is conducted, the applied
voltage on the master side of the coupled inductor (V Nm) is
the sum of the master output voltage (VOm) and the forwardvoltage drop of the freewheeling diode (V Fm). Since VNm is
negative, VNs is positive and the output diode for the slave
output is conducted. The slave output voltage (VOs) is
determined when the freewheeling diode is conducted.
Based on the operation mode, the maximum drain current
peak is decided as below:


V
1  Om VOm  VFm 
V
PO
DC.min 
I ds.peak 

ηVOm  VFm 
2 Lf S
for CCM operation.
The description of coupled inductor operation when the
gate of FSL336LR turns on:
- The applied voltage on master side (VNm):
VNm  VDC  VOm .
I ds.peak 
The applied voltage on slave side (VNs):
VNs  (VDC  VOm )  N s / N m .
© 2013 Fairchild Semiconductor Corporation
Rev. 1.1 • 12/30/14
tS
VOs  (VOm  VFm )  N s / N m  VFs
VOm
-
tON
The slave output voltage is described as:
VNm

tS
Figure 31. Waveforms of the Applied Voltage on Slave
Side of Coupled Inductor and Slave Output Voltage

V
21  Om
V
DC.min

ηLf S

 PO

for DCM operation.
where VDC.min is the minimum DC input voltage.
www.fairchildsemi.com
13
AN-4159
APPLICATION NOTE
is
Step 2: Determine Core Size of Coupled
Inductor
Is.pk
ΔIs
Before selecting the coupled inductor size, the current limit
is able to be adjusted to optimize the core size as below.
IOs
RX
 I ds. peak. max
46kΩ  R X
where ILIMIT.min denotes the minimum pulse-by-pulse
current limit and RX is the external resistor on the ILIMIT pin.
Since the couple inductor of buck converter operates similar
to the transformer of a flyback converter, as illustrated in
Step 0; typical cores, such as EI, EE, and EF type can be
selected from Table 5.
I LIMIT.adj. min  I LIMIT . min 
t
ton
Figure 33. Waveform of Slave Output Diode Current
in CCM Operation
2
V


 1  Om


  VDC . min
PO
1
 
I L.rms  
(VOm  VFm ) 


Lf s
12
  VOm  VFm  




RMS value of master side inductor current in CCM.
V  VFm 3
(VOs  VFs ) 2 (1  Om
)
2
I Os
VDC . min
I s.rms 

2
V  VFm
12Lf s ( N s / N m )
1  Om
VDC . min
RMS value of slave output diode current in CCM.
2
Table 5. Core Selection Table (for Universal Input
Range, fS=50 kHz and PO=5~10 W)
EI Core
EE Core
2
EF Core
2
2
Size
Ae (mm )
Size
Ae (mm )
Size
Ae (mm )
EI12.5
14.4
EE16
19.0
EF12.6
13.0
EI16
19.8
EE19
23.0
EF16.0
20.1
EI19
24.0
EE20
31.0
EF20.0
33.5
toff
iL
Step 3: Calculate Minimum Primary Turns
With the chosen core, the minimum number of turns for the
master side to avoid the core saturation is given by:
N m. min 
Lmax  I LIMIT .adj. max
Bsat  Ae
t
tON
RX
I LIMIT.adj. max  I LIMIT . max 
46
kΩ
 RX ; L
where
max is the
maximum value of the inductance, Bsat is the saturation flux
density; and Ae denotes the cross-sectional area of the core.
toff
Figure 34. Waveform of Inductor Current of Master Side
in DCM Operation
is
Step 4: Determine the Number of Turns for
Master and Slave Outputs
The numbers of turns for master and slave side are
determined as below:
t
N m  N m. min
Ns 
ton
toff
Figure 35. Waveform of Slave Output Diode Current in
DCM Operation
VOs  VFs
Nm
VOm  VFm
 VOm  VFm  

V
1 
 81  Om  PO 3
VDC.min   VDC.min 
 
VDC .min
9 Lf S η3
Step 5: Determine Wire Diameter for Each
Winding Based on RMS Current
I L.rms
The rms currents of each winding are obtained as below.
RMS value of master side inductor current in DCM.
iL
I s.rms  I Os
toff
Figure 32. Waveform of Inductor Current of Master Side
in CCM Operation
© 2013 Fairchild Semiconductor Corporation
Rev. 1.1 • 12/30/14



RMS value of slave output diode current in DCM.
where IOs is the slave output load.
Current density of 6~10 A/mm2 is typically recommended.
To avoid severe eddy current losses, avoid diameter >0.5 mm.
t
ton

V
8η1  Om
V
DC.min

(VOm  VFm )
9 PO f s L
www.fairchildsemi.com
14
AN-4159
APPLICATION NOTE
Appendix B — Equation Details
Equation A:VDC.min  V AC.min 2 
Equation 2: Minimum DC Input Voltage
The voltage ripple at the DC link capacitor can be
calculated with the power delivered to converter system.
PO
η  C DC  V AC.min 2
t AC_dis
1
AC_F)
2 fL
where AC_F is 0 for half-wave rectification and 1 for
full-wave rectification.
Equation B:VDC.min  V AC.min 2  cos 2πf L(t AC_dis 
( 1-DCH )
1
2
2
CDC( 2VAC.min  VDC.min )  Pin
2
fL
for half-wave rectification.
Equation 7: Inductance at Boundary
Conduction Mode
( 1/ 2-DCH )
1
2
2
CDC( 2VAC.min  VDC.min )  Pin
2
fL
To be operated in BCM, the average value of inductor
current should be identical to the half of the ripple of
inductor current, as shown below.
for full-wave rectification.
Therefore, the calculation methods for minimum DC
input voltage are:
VDC,min  2VAC.min 2
IL
2 PO  ( 1-DCH )
η  CDC  f L
IL 
I
IL
for half-wave rectification.
VDC,min  2VAC.min 2
I
2
2 PO  ( 1/ 2-DCH )
η  CDC  f L
for full-wave rectification.
t
Since DCH in the above equations is difficult to estimate
exactly, calculating VDC.min through the below two
simultaneous equations is an alternative. Equation A is
about the input voltage discharging waveform by input
power. Equation B is about AC input voltage waveform.
Through these equations, a more exact minimum input
voltage can be calculated without the estimation of DCH.
Figure 38. Inductor Current at BCM
IL
IL 
I
I
2
IL
Vin,min
t
Figure 39. Inductor Current at CCM
tCH
1
ΔiL
2
I
(V
 VO )D
 in  DC . min
D 2 Lboundaryf S . HIGH
IL 
DCH = tCH / tL
tL
Figure 36. Full-Wave Rectification

The inductance to be operated in BCM is:
Vin,min
tCH
VOUT
I in
(V
 VO )VOUT / VDC . min
 DC . min
/ VDC . min
2 Lboundaryf S . HIGH
η( 1
DCH = tCH / tL
Lboundary 
tL
Figure 37. Half-Wave Rectification
© 2013 Fairchild Semiconductor Corporation
Rev. 1.1 • 12/30/14
VO
)VOUT
2
VDC . min
2 f S .HIGH PO
www.fairchildsemi.com
15
AN-4159
APPLICATION NOTE
Equation 8, 11: Operating Switching
Frequency and Drain Peak Current at CCM
Equation 9, 12: Operating Switching
Frequency and Drain Peak Current at DCM
By Green Mode levels (refer to Figure 11):
As above CCM calculation:
fS 
f S .HIGH  f S .LOW
(VCOMP  0.8V )  22kHz
VGREEN .HIGH  VGREEN .LOW
fS 
The relationship between VCOMP and Ids.peak is:
VCOMP 
I LIMIT
VCOMP 
2.4V
 I ds. peak
VDC . min  VO
 SL  t CLD 
 t CLD
L
I LIMIT
I in I L
PO
(V
 VO )VOUT / VDC . min


 DC . min
D
2
VOUT
2 Lf S
For simplicity, some of constants are substituted as:

f S .HIGH  f S .LOW
PO
, 
,
VGREEN .HIGH  VGREEN .LOW
VDC . min

I LIMIT
V
 Vo
1
I ds.peakD1 & I ds.peak  DC . min
D1
2
Lfs
1
Lfs
2
 I in 
I ds.peak
2 VDC . min  Vo
I in 

2.4V
V
 VO
 SL  t CLD  DC . min
 tCLD
L
Po
ηVDC . min
 I ds.peak 
There are two simultaneous equations having two
variables, Ids.peak and fs:
Equation A:f S  α(γ  I ds.peak  0.8V)  22kHz
Equation B:I ds.peak  β
2.4V
 I ds. pk
VDC . min  VO
 SL  t CLD 
 t CLD
L
However, the calculation method of Ids.peak for DCM
operation is:
The calculation method of Ids.peak for CCM operation is:
I ds. peak 
f S .HIGH  f S .LOW
(VCOMP  0.8V )  22kHz
VGREEN .HIGH  VGREEN .LOW
VDC.min VDC.min  VO VOUT /V DC.min

VOUT
2 Lf S

1
Lfs
2
I ds.peak
2 VDC . min  Vo
2(VDC . min  Vo )
Po

Lfs
ηVDC . min
For simplicity, some of constants are substituted with the
same as for CCM operation:

f S . HIGH  f S . LOW
PO
, 
,
VGREEN . HIGH  VGREEN .LOW
VDC . min

I LIMIT
2.4V
V
 VO
 SL  tCLD  DC . min
 tCLD
L
There are two simultaneous equations having two variables,
Ids.peak and fs:
Equation A : f S   (  I ds. peak  0.8V )  22kHz
Equation B : I ds. peak 
2VDC . min  VO 
Lf S
Related Datasheets
FSL336LRN − Green Mode Fairchild Buck Switch
DISCLAIMER
FAIRCHILD SEMICONDUCTOR RESERVES THE RIGHT TO MAKE CHANGES WITHOUT FURTHER NOTICE TO ANY PRODUCTS
HEREIN TO IMPROVE RELIABILITY, FUNCTION, OR DESIGN. FAIRCHILD DOES NOT ASSUME ANY LIABILITY ARISING OUT OF THE
APPLICATION OR USE OF ANY PRODUCT OR CIRCUIT DESCRIBED HEREIN; NEITHER DOES IT CONVEY ANY LICENSE UNDER ITS
PATENT RIGHTS, NOR THE RIGHTS OF OTHERS.
LIFE SUPPORT POLICY
FAIRCHILD’S PRODUCTS ARE NOT AUTHORIZED FOR USE AS CRITICAL COMPONENTS IN LIFE SUPPORT DEVICES OR SYSTEMS
WITHOUT THE EXPRESS WRITTEN APPROVAL OF THE PRESIDENT OF FAIRCHILD SEMICONDUCTOR CORPORATION.
As used herein:
1.
Life support devices or systems are devices or systems
which, (a) are intended for surgical implant into the body, or
(b) support or sustain life, or (c) whose failure to perform
when properly used in accordance with instructions for use
provided in the labeling, can be reasonably expected to
result in significant injury to the user.
© 2013 Fairchild Semiconductor Corporation
Rev. 1.1 • 12/30/14
2.
A critical component is any component of a life support
device or system whose failure to perform can be reasonably
expected to cause the failure of the life support device or
system, or to affect its safety or effectiveness
www.fairchildsemi.com
16