Step-Down Switching Regulators

Application Note 35
August 1989
Step-Down Switching Regulators
Jim Williams
lost in this voltage-to-current-to-magnetic field-to-current-to-charge-to-voltage conversion. In practice, the
circuit elements have losses, but step-down efficiency is
still higher than with inherently dissipative (e.g., voltage
divider) approaches. Figure 2 feedback controls the basic
circuit to regulate output voltage. In this case switch ontime (e.g., inductor charge time) is varied to maintain the
output against changes in input or loading.
REGULATED
OUTPUT
IN
PULSE
WIDTH
MODULATOR
Figure 1 is a conceptual voltage step-down or “buck”
circuit. When the switch closes the input voltage appears
at the inductor. Current flowing through the inductor-capacitor combination builds over time. When the switch
IN
OUT
AN35 F01
Figure 1. Conceptual Voltage Step-Down (“Buck”) Circuit
opens current flow ceases and the magnetic field around
the inductor collapses. Faraday teaches that the voltage
induced by the collapsing magnetic field is opposite to the
originally applied voltage. As such, the inductor’s left side
heads negative and is clamped by the diode. The capacitors accumulated charge has no discharge path, and a DC
potential appears at the output. This DC potential is lower
than the input because the inductor limits current during
the switch’s on-time. Ideally, there are no dissipative elements in this voltage step-down conversion. Although the
output voltage is lower than the input, there is no energy
+
Basic Step Down Circuit
–
A substantial percentage of regulator requirements
involve stepping down the primary voltage. Although
linear regulators can do this, they cannot achieve the
efficiency of switching based approaches1. The theory
supporting step-down (“buck”) switching regulation is
well established, and has been exploited for some time.
Convenient, easily applied ICs allowing implementation
of practical circuits are, however, relatively new. These
devices permit broad application of step-down regulation
with minimal complexity and low cost. Additionally, more
complex functions incorporating step-down regulation
become realizable.
VREF
AN35 F02
Figure 2. Conceptual Feedback Controlled Step-Down Regulator
Practical Step-Down Switching Regulator
Figure 3, a practical circuit using the LT®10742 IC regulator,
shows similarities to the conceptual regulator. Some new
elements have also appeared. Components at the LT1074’s
“VCOMP” pin control the IC’s frequency compensation,
stabilizing the feedback loop. The feedback resistors are
selected to force the “feedback” pin to the device’s internal
2.5V reference value. Figure 4 shows operating waveforms
for the regulator at VIN = 28V with a 5V, 1A load.
L, LT, LTC, LTM, Linear Technology and the Linear logo are registered
trademarks of Linear Technology Corporation. All other trademarks are the
property of their respective owners.
Note 1: While linear regulators cannot compete with switchers, they can
achieve significantly better efficiencies than generally supposed. See LTC
Application Note 32, “High Efficiency Linear Regulators,” for details.
Note 2: See Appendix A for details on this device.
an35f
AN35-1
Application Note 35
L1
150μH
VIN
VSW
VIN
MBR745
LT1074
VC
GND
5V
OUTPUT
+
A = 20V/DIV
1000μF
2.8k
1%
FB
2k
2.2k
1%
0.1
L1: PULSE ENGINEERING, INC. #PE-51516
B = 1A/DIV
C = 0.2A/DIV
ON 1A DC LEVEL
D = 1A/DIV
AN35 F03
AN35 F04
HORIZ = 5μs/DIV
Figure 3. A Practical Step-Down Regulator Using the LT1074
Figure 4. Waveforms for the Step-Down Regulator at
VIN = 28V and VOUT = 5V at 1A
100
VIN = 10V
90
EFFICIENCY (%)
80
A = 20V/DIV
B = 1A/DIV
C = 0.2A/DIV
ON 1A DC LEVEL
LT1074
LT1086
70
VIN = 15.5V
VIN = 6.5V
60
VIN = 8V
50
VIN = 10.1V
LM317
40
VIN = 12.4V
30
20
D = 1A/DIV
VOUT = 5V, 1A
LM317 DROPOUT = 3V
LT1086 DROPOUT = 1.5V
10
0
80
HORIZ = 5μs/DIV
AN35 F05
90
110
120
100
AC LINE VOLTAGE
130
140
AN35 F06
Figure 5. Waveforms for the Step-Down Regulator at VIN = 12V
and VOUT = 5V at 1A
Trace A is the VSW pin voltage and Trace B is its current.
Inductor current3 appears in Trace C and diode current
is Trace D. Examination of the current waveforms allows
determination of the VSW and diode path contributions to
inductor current. Note that the inductor current’s waveform
occurs on top of a 1A DC level. Figure 5 shows significant
duty cycle changes when VIN is reduced to 12V. The lower
input voltage requires longer inductor charge times to
maintain the output. The LT1074 controls inductor charge
characteristics (see Appendix A for operating details),
with resulting waveform shape and time proportioning
changes.
Figure 6. Efficiency vs AC Line Voltage for the LT1074. LT1086
and LM317 Linear Regulators are Shown for Comparison
Figure 6 compares this circuit’s efficiency with linear
regulators in a common and important situation. Efficient
regulation under varying AC line conditions is a frequent
requirement. The figure assumes the AC line has been
transformed down to acceptable input voltages. The input
voltages shown correspond to the AC line voltages given
on the horizontal axis. Efficiency for the LM317 and LT1086
linear regulators suffers over the wide input range.
Note 3: Methods for selecting appropriate inductors are discussed in
Appendix B.
an35f
AN35-2
Application Note 35
100
90
Dual Output Step-Down Regulator
VOUT = 5V
IOUT = 1A
Figure 8, a logical extension of the basic step-down converter, provides positive and negative outputs. The circuit
is essentially identical to Figure 3’s basic converter with the
addition of a coupled winding to L1. This floating winding’s
output is rectified, filtered and regulated to a –5V output.
The floating bias to the LT1086 positive voltage regulator
permits negative outputs by assigning the regulator’s
output terminal to ground. Negative output power is set
by flux pick-up from L1’s driven winding. With a 2A load
at the +15 output the –5V output can supply over 500mA.
Because L1’s secondary winding is floating its output may
be referred to any point within the breakdown capability
of the device. Hence, the secondary output could be 5V
or, if stacked on the +15 output, 20V.
EFFICIENCY (%)
80
70
60
50
40
30
20
10
0
10 12 14 16 18 20 22 24 26 28 30
INPUT VOLTAGE
AN35 F07
Figure 7. Efficiency Plot for Figure 3. Higher Input Voltages
Minimize Effects of Saturation Losses, Resulting in Increased
Efficiency
The LT1086 is notably better because its lower dropout
voltage cuts dissipation over the range. Switching preregulation4 can reduce these losses, but cannot equal
the LT1074’s performance. The plot shows minimum
efficiency of 83%, with some improvement over the full
AC line excursion. Figure 7 details performance. Efficiency
approaches 90% as input voltage rises. This is due to
minimization of the effects of fixed diode and LT1074 junction losses as input increases. At low inputs these losses
are a higher percentage of available supply, degrading
efficiency. Higher inputs make the fixed losses a smaller
percentage, improving efficiency. Appendix D presents
detail on optimizing circuitry for efficiency.
•2
28V
INPUT
VIN
VSW
GND
Negative outputs can also be obtained with a simple 2-terminal inductor. Figure 9 demonstrates this by essentially
grounding the inductor and steering the catch diodes
negative current to the output. A1 facilitates loop closure
by providing a scaled inversion of the negative output to
the LT1074’s feedback pin. The 1% resistors set the scale
factor (e.g., output voltage) and the RC network around A1
gives frequency compensation. Waveforms for this circuit
are reminiscent of Figure 5, with the exception that diode
Note 4: See Reference 1.
L1
4
MBR745
LT1074
VC
Negative Output Regulators
1
3•
15V
OUTPUT
+
1000μF
24.9k
1%
FB
2k
0.1μF
MUR120
+
LT1086
IN
OUT
ADJ
4.99k
1%
120Ω
1%
+
470μF
47μF
390Ω
1%
L1: PULSE ENGINEERING, INC. #PE-65050
AN35 F08
–5V
OUTPUT
Figure 8. Coupled Inductor Provides Positive and Negative Outputs
an35f
AN35-3
Application Note 35
MBR745
1000μF
L1
55μH
+
12V
INPUT
–5V
OUTPUT
VSW
VIN
LT1074
VC
GND
0.33
4.7k
FB
24k
1%
12k
1%
12V
INPUT
NC
–
A1
LT1006
+
L1: PULSE ENGINEERING, INC. #PE-92116
1N4148
AN35 F09
Figure 9. A Negative Output Step-Down Regulator
current (Trace D) is negative. Traces A, B and C are VSW
voltage, inductor current and VSW current respectively.
Figure 11, commonly referred to as “Nelson’s Circuit,”
provides the same function as the previous circuit, but
eliminates the level-shifting op amp. This design accomplishes the level shift by connecting the LT1074’s “ground”
pin to the negative output. Feedback is sensed from circuit
ground, and the regulator forces its feedback pin 2.5V above
its “ground” pin. Circuit ground is common to input and
output, making system use easy. Operating waveforms
are essentially identical to Figure 10. Advantages of the
previous circuit compared to this one are that the LT1074
package can directly contact a grounded heat sink and that
control signals may be directly interfaced to the ground
referred pins.
The inductor values in both negative output designs are
notably lower than in the positive case. This is necessitated
A = 20V/DIV
B = 1A/DIV
C = 4A/DIV
D = 4A/DIV
HORIZ = 5μs/DIV
by the reduced loop phase margin of these circuits. Higher
inductance values, while preferable for limiting peak current, will cause loop instability or outright oscillation.
Current-Boosted Step-Down Regulator
Figure 12 shows a way to obtain significantly higher
output currents by utilizing efficient energy storage in
the LT1074 output inductor. This technique increases the
duty cycle over the standard step-down regulator allowing
more energy to be stored in the inductor. The increased
output current is achieved at the expense of higher output
voltage ripple.
The operating waveforms for this circuit are shown in
Figure 13. The circuit operating characteristics are similar
to that of the step-down regulator (Figure 3). During the
VSW (Trace A) “on” time the input voltage is applied to one
end of the coupled inductor. Current through the VSW pin
(Trace B) ramps up almost instantaneously (since inductor
current (Trace F) is present) and then slows as energy is
stored in the core. The current proceeds into the inductor
(Trace D) and finally is delivered to the load. When the VSW
pin goes off, current is no longer available to charge the
inductor. The magnetic field collapses, causing the VSW
pin voltage to go negative. At this point similarity with
the basic regulator vanishes. In this modified version the
output current (Trace F) receives a boost as the magnetic
field collapses. This results when the energy stored in
AN35 F10
Figure 10. Figure 9’s Waveforms
an35f
AN35-4
Application Note 35
MBR745
+
VIN
+
22μF
–5VOUT
VSW
22μF
LT1074
VC
GND
L1
14μH
1000μF
+
VIN
FB
2.8k
1k
2.2k
1μF
AN35 F11
L1 = PULSE ENGINEERING, INC. #PE-51509
Figure 11. Nelson’s Circuit...A (Better) Negative Output Step-Down Regulator
VIN
20V TO 30V
3
VIN
+
220μF
150Ω
LT1074
D1
1
L1
•
VSW
•
N 2
+
1
D2
1000pF
5VOUT
10A
C1
3000μF
330pF
GND
FB
510Ω
0.33
50Ω
2k*
150Ω
1μF
* = 1% FILM RESISTOR
L1 = PULSE ENGINEERING, INC. #PR-65282
D1 = MUR110
D2 = 1N5831
2k*
AN35 F12
Figure 12. “Current Boosted” Step-Down Regulator. Boost Current is Supplied By Energy Stored in the Tapped Inductor
the core is transferred to the output. This current step
circulates through C1 and D2 (Trace E), somewhat increasing output voltage ripple. Not all the energy is transferred
to the “1” winding. Current (Trace C) will continue to flow
in the “N” winding due to leakage inductance. A snubber
network suppresses the effects of this leakage inductance.
For lowest snubber losses the specified tapped inductor
is bifilar wound for maximum coupling.
Post Regulation-Fixed Case
In most instances the LT1074 output will be applied directly
to the load. Those cases requiring faster transient response
or reduced noise will benefit from linear post regulation. In
Figure 14 a 3-terminal regulator follows the LT1074 output.
The LT1074 output is set to provide just enough voltage
A = 50V/DIV
B = 5A/DIV
C = 10A/DIV
D = 10A/DIV
E = 10A/DIV
F = 10A/DIV
HORIZ = 2μs/DIV
AN35 F10
Figure 13. AC Current Flow for the Boosted Regulator
to the LT1084 to maintain regulation. The LT1084’s low
dropout characteristics combined with a high circuit input
voltage minimizes the overall efficiency penalty.
an35f
AN35-5
Application Note 35
L1
150μH
VIN
VSW
VIN
MBR745
LT1074
VC
GND
≈6.5V
+
LT1084-5
5VOUT
+
47μF
1000μF
4.12k
1%
FB
2k
2.49k
1%
0.1
L1: PULSE ENGINEERING, INC. #PE-51516
AN35 F14
Figure 14. Linear Post-Reglator Improves Noise and Transient Response
Post Regulation-Variable Case
Some situations require variable linear post regulation.
Figure 15 does this with little efficiency sacrifice. The
LT1085 operates in normal fashion, supplying a variable
1.2V to 28V output. The remainder of the circuit forms
a switched mode pre-regulator which maintains a small,
fixed voltage across the LT1085 regardless of its output
voltage. A1 biases the LT1074 to produce whatever voltage
is necessary to maintain the “E diodes” potential across the
LT1085. A1’s inputs are balanced when the LT1085 output
is “E diodes” above its input. A1 maintains this condition
regardless of line, load or output voltage conditions.
Thus, good efficiency is maintained over the full range of
output voltages. The RC network at A1 compensates the
loop. Loop start-up is assured by deliberately introducing a positive offset to A1. This is done by grounding
A1’s appropriate balance pin (5), resulting in a positive
6mV offset. This increases amplifier drift, and is normally
considered poor practice, but causes no measurable error
in this application.
As shown, the circuit cannot produce outputs below the
LT1085’s 1.2V reference. Applications requiring output
adjustability down to 0V will benefit from option “A”
shown on the schematic. This arrangement replaces L1
with L2. L2’s primary performs the same function as L1
and its coupled secondary winding produces a negative
bias output (–V). The full-wave bridge rectification is
necessitated by widely varying duty cycles. A2 and its attendant circuitry replace all components associated with
the LT1085 VADJ pin. The LT1004 reference terminates
the 10k to 250k feedback string at –1.2V with A2 providing buffered drive to the LT1085 VADJ pin. The negative
bias allows regulated LT1085 outputs down to 0V. The –V
potential derived from L2’s secondary varies considerably with operating conditions. The high feedback string
values and A2’s buffering ensure stable circuit operation
for “starved” values of –V.
Low Quiescent Current Regulators
Many applications require very wide ranges of power supply output current. Normal conditions require currents in
the ampere range, while standby or “sleep” modes draw
only microamperes. A typical laptop computer may draw
1 to 2 amperes running while needing only a few hundred
microamps for memory when turned off. In theory, any
regulator designed for loop stability under no-load conditions will work. In practice, a converter’s relatively large
quiescent current may cause unacceptable battery drain
during low output current intervals. Figure 16’s simple
loop effectively reduces circuit quiescent current from
6mA to only 150μA. It does this by utilizing the LT1074’s
shutdown pin. When this pin is pulled within 350mV of
ground the IC shuts down, pulling only 100μA. Comparator
C1 combines with the LT1004 reference and Q1 to form a
“bang-bang” control loop around the LT1074. The LT1074’s
internal feedback amplifier and voltage reference are bypassed by this loop’s operation. When the circuit output
(Trace C, Figure 17) falls slightly below 5V C1’s output
(Trace A) switches low, turning off Q1 and enabling the
an35f
AN35-6
Application Note 35
“B”
35V
INPUT
EDIODES
(≈1.8V)
“C”
L1
150μH
VIN
VSW
MBR745
LT1074
VC
GND
+
1000μF
1N4148s
EDIODES
(≈1.8V)
FB
0.22
NC
1N4148
LT1085
IN OUT
ADJ
1.2V TO 28VOUT
(SEE OPTION “A” FOR
OUTPUT DOWN TO 0V)
+
110Ω*
47μF
“D”
2.5k
OUTPUT
ADJUST
+
10μF
3.9k
35V INPUT
3k
–
A1
LT1006
5
+
4
OPTION “A”
(FOR OUTPUT
DOWN TO 0V)
SEE TEXT FOR
DISCUSSION
TO POINT “D”
TO LT1085 OUTPUT
35V INPUT
TO POINT “B”
TO POINT “C”
10k*
7
+
L2
4
6
3
0.1μF
A2
LT1006
–
2
1
3
2
250k
OUTPUT
ADJUST
4
4.7k
–V
10μF
LT1004
1.2V
+
1N4148
s4
AN35 F15
L1: PULSE ENGINEERING, INC. #PE-51516
L2: PULSE ENGINEERING, INC. #PE-65050
Figure 15. Adjustable Linear Post-Regulator Maintains Efficiency Over Widely Varying Operating Conditions
12V
INPUT
L1
150μH
VIN
VSW
LT1074
NC
5VOUT
+
MBR745
1000μF
1M
1%
FB
VC
NC
GND
SD
Q1
2N3904
12V INPUT
1M
+
C1
1/2 LT1017
–
470k
LT1004
1.2V
12V
INPUT
340k
1%
AN35 F16
L1: PULSE ENGINEERING, INC. #PE-51516
Figure 16. A Simple Loop Reduces Quiescent Current to 150μA
an35f
AN35-7
Application Note 35
LT1074. The VSW pin (Trace B) pulses at full duty cycle,
forcing the output back above 5V. C1 then biases Q1 again,
the LT1074 goes into shutdown, and loop action repeats.
The frequency of this on-off control action is directly load
dependent, with typical repetition rates of 0.2Hz at no load.
Short on-times keep duty cycle low, resulting in the small
effective quiescent current noted. The on-off operation
combines with the LC filtering action in the regulator’s
VSW line to generate an output hysteresis of about 50mV
(again, see Figure 17, Trace C).
A = 10V/DIV
B = 10V/DIV
C = 0.1V/DIV
AC-COUPLED
ON 5V DC LEVEL
Figure 18’s more sophisticated circuit eliminates these
problems with some increase in complexity. Quiescent
current is maintained at 150μA. The technique shown is
particularly significant, with broad implication in battery
powered systems. It is easily applied to a wide variety of
regulator requirements, meeting an acknowledged need
across a wide spectrum of applications.
Figure 18’s signal flow is similar to Figure 16, but additional circuitry appears between the feedback divider
and the LT1074. The LT1074’s internal feedback amplifier
and reference are not used. Figure 19 shows operating
AN35 F17
HORIZ = 100μs/DIV
The loop performs well, but has two potential drawbacks.
At higher output currents the loop oscillates in the 1kHz to
10kHz range, causing audible noise which may be objectionable. This is characteristic of this type of loop, and is
the reason that ICs employing gated oscillators invariably
produce such noise. Additionally, the control loops operation causes about 50mV of ripple on the output. Ripple
frequency ranges from 0.2Hz to 10kHz depending upon
input voltage and output current.
Figure 17. The Low Quiescent Current Loop’s Waveforms
5VOUT
L1
35μH
VIN
VIN
VSW
MBR745
LT1074
SD
VC
GND
+
C2
2000μF
FB
NC
0.033
C3
0.01
8.2k
12V INPUT
–
R1
200Ω
C1
47μF
A1
1/2 LT1017
1N748
68pF**
+
A2
1/2 LT1017
R3
5.9k*
+
+
R2
18k*
200k**
74C04
10k
470k
–
L1 = PULSE ENGINEERING, INC. #PE-92103
* = 1% METAL FILM RESISTOR
** = TYPICAL VALUES—SEE TEXT
LT1004
1.2V
12V
+
2.2μF
AN35 F18
Figure 18. A More Sophisticated Loop Gives Better Regulation While Maintaining 150μA Quiescent Current
an35f
AN35-8
Application Note 35
waveforms under no-load conditions. The output (Trace A)
ramps down over a period of seconds. During this time
comparator A1’s output (Trace B) is low, as are the 74C04
paralleled inverters. This pulls the VC pin (Trace D) low,
forcing the regulator to zero duty cycle. Simultaneously, A2
(Trace C) is low, putting the LT1074 in its 100μA shutdown
mode. The VSW pin (Trace E) is off, and no inductor current
flows. When the output drops about 60mV, A1 triggers and
the inverters go high, pulling the VC pin up and biasing the
regulator. The Zener diode prevents VC pin overdrive. A2
also rises, taking the IC out of shutdown mode. The VSW
pin pulses the inductor at the 100kHz clock rate, causing
the output to abruptly rise. This action trips A1 low, forcing
the VC pin back low and shutting off VSW pulsing. A2 also
goes low, putting the LT1074 into shutdown.
This “bang-bang” control loop keeps the 5V output within
the 60mV ramp hysteresis window set by the loop. Note
that the loop oscillation period of seconds means the R1-C1
time constant at VC is not a significant term. Because the
LT1074 spends almost all of the time in shutdown, very
little quiescent current (150μA) is drawn.
Figure 20 shows the same waveforms with the load increased to 2mA. Loop oscillation frequency increases to
keep up with the load’s sink current demand. Now, the VC
pin waveform (Trace D) begins to take on a filtered appearance. This is due to R1-C1’s 10ms time constant. If
the load continues to increase, loop oscillation frequency
will also increase. The R1-C1 time constant, however, is
fixed. Beyond some frequency, R1-C1 must average loop
oscillations to DC. At 7mA loading (Figure 21) loop frequency further increases, and the VC waveform (Trace D)
appears heavily filtered.
Figure 22 shows the same circuit points at 2A loading.
Note that the VC pin is at DC, as is the shutdown pin.
Repetition rate has increased to the LT1074’s 100kHz
A = 0.1V/DIV
AC-COUPLED
B = 20V/DIV
A = 0.1V/DIV
AC-COUPLED
C = 20V/DIV
C = 20V/DIV
B = 20V/DIV
D = 2V/DIV
D = 2V/DIV
E = 10V/DIV
E = 10V/DIV
HORIZ = 0.5 SECOND/DIV
AN35 F19
Figure 19. Low Quiescent Current Regulator’s Waveforms
with No Load (Traces B, C and E Retouched for Clarity)
HORIZ = 20ms/DIV
AN35 F20
Figure 20. Low Quiescent Current Regulator’s Waveforms at 2mA
Loading
A = 0.2V/DIV
AC-COUPLED
A = 0.1V/DIV
AC-COUPLED
B = 20V/DIV
B = 20V/DIV
C = 20V/DIV
C = 20V/DIV
D = 2V/DIV
D = 2V/DIV
E = 20V/DIV
E = 10V/DIV
HORIZ = 10ms/DIV
AN35 F21
Figure 21. Low Quiescent Current Regulator’s Waveforms at 7mA
Loading
HORIZ—TRACES A AND E = 10μs/DIV
TRACES B, C, D = 5ms/DIV
AN35 F22
Figure 22. Low Quiescent Current Regulator’s Waveforms at 2A
Loading
an35f
AN35-9
Application Note 35
clock frequency. Figure 23 plots what is occurring, with
a pleasant surprise. As output current rises, loop oscillation frequency also rises until about 23Hz. At this point
the R1-C1 time constant filters the VC pin to DC and the
LT1074 transitions into “normal” PWM operation. With
the VC pin at DC it is convenient to think of A1 and the
inverters as a linear error amplifier with a closed-loop gain
set by the R2-R3 feedback divider. In fact, A1 is still duty
cycle modulating, but at a rate far above R1-C1’s break
frequency. The phase error contributed by C2 (which was
selected for low loop frequency at low output currents)
is dominated by the R1-C1 roll off and the C3 lead into
A1. The loop is stable and responds linearly for all loads
beyond 10mA. In this high current region the LT1074 is
desirably “fooled” into behaving like a conventional stepdown regulator.
A formal stability analysis for this circuit is quite complex,
but some simplifications lend insight into loop operation.
At 250μA loading (20kΩ) C2 and the load form a decay
time constant exceeding 30 seconds. This is orders of
magnitude larger than R2-C3, R1-C1, or the LT1074’s
100kHz commutation rate. As a result, C2 dominates the
loop. Wideband A1 sees phase shifted feedback, and very
low frequency oscillations similar to Figure 19’s occur5.
Although C2’s decay time constant is long, its charge
time constant is short because the circuit has low sourcing impedance. This accounts for the ramp nature of the
oscillations.
Increased loading reduces the C2-load decay time constant. Figure 23’s plot reflects this. As loading increases,
the loop oscillates at a higher frequency due to C2’s decreased decay time. When the load impedance becomes
low enough C2’s decay time constant ceases to dominate
the loop. This point is almost entirely determined by R1
and C1. Once R1 and C1 “take over” as the dominant time
constant the loop begins to behave like a linear system.
In this region (e.g., above about 10mA, per Figure 23)
the LT1074 runs continuously at its 100kHz rate. Now,
C3 becomes significant, performing as a simple feedback
lead6 to smooth output response. There is a fundamental
trade-off in the selection of the C3 lead value. When the
converter is running in its linear region it must dominate
the loops time lag generated hysteretic characteristic. As
such, it has been chosen for the best compromise between
output ripple at high load and loop transient response.
Despite the complex dynamics transient response is quite
good. Figure 24 shows performance for a step from no
load to 1A. When Trace A goes high a 1A load appears
across the output (Trace C). Initially, the output sags almost 200mV due to slow loop response time (the R1-C1
pair delay VC pin (Trace B) response). When the LT1074
comes on response is reasonably quick and surprisingly
well behaved considering circuit dynamics. The multi-time
constant recovery7 (“rattling” is perhaps more appropriate)
is visible in Trace C’s response.
Note 5: Some layouts may require substantial trace area to A1’s inputs. In
such cases the optional RC network around A1 ensures clean transitions
at A1’s output.
Note 6: “Zero Compensation” for all you technosnobs out there.
Note 7: Once again, “multi-pole settling” for those who adore jargon.
LINEAR REGION
EXTENDS
TO 5.0A
LOOP FREQUENCY (Hz)
20
16
A = 10V/DIV
B = 2V/DIV
12
C = 0.2V/DIV
ON 5V DC LEVEL
8
4
0
IQ = 150μA
0.6Hz
0
2
4
6
8
OUTPUT (mA)
10
12
HORIZ = 5ms/DIV
AN35 F24
Figure 24. Load Transient Response for Figure 18
AN35 F23
Figure 23. Figure 18’s Loop Frequency vs Output Current.
Note Linear Loop Operation Above 10mA
an35f
AN35-10
Application Note 35
Figure 25 plots efficiency versus output current. High
power efficiency is similar to standard converters. Low
power efficiency is somewhat better, although poor in
the lowest ranges. This is not particularly bothersome,
as power loss is very small.
The loop provides a controlled, conditional instability
instead of the usually more desirable (and often elusive)
unconditional stability. This deliberately introduced characteristic dramatically lowers converter quiescent current
without sacrificing high power performance.
100
90
EFFICIENCY (%)
80
TYPICAL OPERATING
REGION
70
60
50
40
30
20mA
5mA
1mA
20
10
0
IQ = 150μA
0
0.5
TYPICAL STANDBY
REGION
1.5
1.0
OUTPUT CURRENT
2.0
2.5
AN35 F25
Figure 25. Efficiency vs Output Current for Figure 18.
Standby Efficiency is Poor, But Power Loss Approaches
Battery Self-Discharge
Wide Range, High Power, High Voltage Regulator
BEFORE PROCEEDING ANY FURTHER, THE READER
IS WARNED THAT CAUTION MUST BE USED IN THE
CONSTRUCTION, TESTING AND USE OF THIS CIRCUIT.
HIGH VOLTAGE, LETHAL POTENTIALS ARE PRESENT IN
THIS CIRCUIT. EXTREME CAUTION MUST BE USED IN
WORKING WITH AND MAKING CONNECTIONS TO THIS
CIRCUIT. REPEAT: THIS CIRCUIT CONTAINS DANGEROUS, HIGH VOLTAGE POTENTIALS. USE CAUTION.
Figure 26 is an example of the LT1074 making a complex
function practical. This regulator provides outputs from millivolts to 500V at 100W with 80% efficiency. A1 compares
a variable reference voltage with a resistively scaled version
of the circuit’s output and biases the LT1074 switching
regulator configuration. The switcher’s DC output drives
a toroidal DC/DC converter comprised of L1, Q1 and Q2.
Q1 and Q2 receive out of phase square wave drive from
the 74C74 ÷ 4 flip-flop stage and the LT1010 buffers. The
flip-flop is clocked from the LT1074 VSW output via the Q3
level shifter. The LT1086 provides 12V power for A1 and the
74C74. A1 biases the LT1074 regulator to produce the DC
input at the DC/DC converter required to balance to loop.
The converter has a voltage gain of about 20, resulting
in high voltage output. This output is resistively divided
down, closing the loop at A1’s negative input. Frequency
compensation for this loop must accommodate the significant phase errors generated by the LT1074 configuration,
the DC/DC converter and the output LC filter. The 0.47μF
roll-off term at A1 and the 100Ω-0.15μF RC lead network
provide the compensation, which is stable for all loads.
Figure 27 gives circuit waveforms at 500V output into a
100W load. Trace A is the LT1074 VSW pin while Trace B is
its current. Traces C and D are Q1 and Q2’s drain waveforms.
The disturbance at the leading edges is due to cross-current
conduction, which lasts about 300ns—a small percentage of the cycle. Transistor currents during this interval
remain within reasonable values, and no overstress or
dissipation problems occur. This effect could be eliminated
with non-overlapping drive to Q1 and Q28, although there
would be no reliability or significant efficiency gain. The
500kHz ringing on the same waveforms is due to excitation of transformer resonances. These phenomena are not
deleterious, although L1’s primary RC damper is included
to minimize them.
All waveforms are synchronous because the flip-flop
drive stage is clocked from the LT1074 VSW output. The
LT1074’s maximum 95% duty cycle means that the Q1-Q2
switches can never see destructive DC drive. The only
condition allowing DC drive occurs when the LT1074 is
at zero duty cycle. This case is clearly non-destructive,
because L1 receives no power.
Figure 28 shows the same circuit points as Figure 27,
but at only 5mV output. Here, the loop restricts drive to
the DC/DC converter to small levels. Q1 and Q2 chop just
70mV into L1. At this level L1’s output diode drops look
large, but loop action forces the desired 0.005V output.
Note 8: For an example of this technique see LTC Application Note 29,
Figure 1.
an35f
AN35-11
28VIN
100μF
SOLID
TANTALUM
+
+
FB
VSW
Q3
2N2369
L2
100μH
0.47
MUR8100
1k
1000μF
A1
LTC1050
12V
+
100Ω
100k
OUTPUT
ADJUST
0.15
LT1010
28VIN
+
Q1
D
100k
D
LT1021-7V
S
Q2
100μF
SOLID
TANTALUM
S
28V
10Ω
2W
7
6
5
4
3
2
1
MUR1100
s4
0.1μF
1000V
L3
680μH
1μF
1000V
AN35 F26
13.7k
1% FILM
1kΩ
OUTPUT
CALIB.
1M*
0V TO 500VOUT
200mA
DANGER! LETHAL
POTENTIALS PRESENT IN SCREENED
AREA. SEE TEXT
L1 = TRIAD TY-94
L2 = PULSE ENGINEERING, INC. #PE-92112
L3 = PULSE ENGINEERING, INC. #PE-52649
Q1, Q2 = MOTOROLA MTH15N20
* = VICTOREEN SLIM-MOX-108
0.005
L1
Figure 26. LT1074 Permits High Voltage Output Over 100dB Range with Power and Efficiency.
DANGER! Lethal Potentials Present—See Text
NC
1N914
GND
LT1074
GND
OUT
LT1086-12
VC
VIN
IN
12V
22μF
2k
LT1010
+
AN35-12
–
+V
GND
1Q
2Q
1D
2D
74C74
1Q
2Q
1P 1C 1CK 2P 2C 2CK
28VIN
Application Note 35
an35f
Application Note 35
A = 50V/DIV
A = 5V/DIV
B = 5A/DIV
B = 50mA/DIV
C = 50V/DIV
C = 0.1V/DIV
D = 50V/DIV
D = 0.1V/DIV
HORIZ = 10μs/DIV
HORIZ = 10μs/DIV
AN35 F27
Figure 27. Figure 26’s Operating Waveforms at 500V Output into
a 100W Load
A = 0.05V/DIV
AC-COUPLED
ON 500V LEVEL
AN35 F28
Figure 28. Figure 26’s Operating Waveforms at 0.005V Output
A = 100V/DIV
HORIZ = 20μs/DIV
AN35 F29
Figure 29. Figure 26’s Output Noise at 500V into a 100W
Load. Residue is Composed of Q1-Q2 Chopping Artifacts and
Transformer Related Ringing. DANGER! Lethal Potentials
Present—See Text
The LT1074’s switched mode drive to L1 maintains high
efficiency at high power, despite the circuits wide output
range9.
Figure 29 shows output noise at 500V into a 100W load.
Q1-Q2 chopping artifacts and transformer related ringing
are clearly visible, although limited to about 80mV. The
coherent noise characteristic is traceable to the synchronous clocking of Q1 and Q2 by the LT1074.
A 50V to 500V step command into a 100W load produces
the response of Figure 30. Loop response on both edges
HORIZ = 50ms/DIV
AN35 F30
Figure 30. 500V Step Response with 100W Load
(Photo Retouched for Clarity). DANGER! Lethal
Potentials Present—See Text
is clean, with the falling edge slightly underdamped. This
slew asymmetry is typical of switching configurations,
because the load and output capacitor determine negative
slew rate. The wide range of possible loads mandates a
compromise when setting frequency compensation. The
falling edge could be made critically or even over damped,
but response time for other conditions would suffer. The
compensation used seems a reasonable compromise.
Note 9: A circuit related to the one presented here appears in the LTC
Application Note 18 (Figure 13). Its linear drive to the step-up DC/DC
converter forces dissipation, limiting output power to about 15W. Similar
restrictions apply to Figure 7 in Application Note 6.
an35f
AN35-13
Application Note 35
Regulated Sinewave Output DC/AC Converter
BEFORE PROCEEDING ANY FURTHER, THE READER
IS WARNED THAT CAUTION MUST BE USED IN THE
CONSTRUCTION, TESTING AND USE OF THIS CIRCUIT.
HIGH VOLTAGE, LETHAL POTENTIALS ARE PRESENT IN
THIS CIRCUIT, EXTREME CAUTION MUST BE USED IN
WORKING WITH AND MAKING CONNECTIONS TO THIS
CIRCUIT. REPEAT: THIS CIRCUIT CONTAINS DANGEROUS HIGH VOLTAGE POTENTIALS. USE CAUTION.
Figure 31 is another example of the LT1074 permitting the
practical implementation of a complex function. It converts
a 28V DC input to a regulated 115VAC 400Hz sinewave
output with 80% efficiency. Waveform distortion is below
1.6% at 50W output. This design shares similarities with
the previous circuit. The LT1074 supplies efficient drive
to a high voltage converter despite large line and load
variations. An amplifier (A1) controls the input to the
high voltage converter via A2 and the LT1074 switching
regulator. The high voltage output is divided down and fed
back to the amplifier where it is compared to a reference
to close a loop. In the previous circuit the output is DC;
here the output is AC. As such, A1’s reference (Trace A,
Figure 32) is an amplitude and frequency stabilized 800Hz
half-sine10. The high voltage converter is driven from a
flip-flop clocked by a reference synchronized pulse (negative going excursions just visible in Trace B) via level shift
transistor Q3. The reference synchronized pulse occurs at
the zero voltage point of the half-sine. The flip-flop outputs
(Traces C and D, respectively) drive the Q1 and Q2 gates.
RC filters in the gate line retard the drive’s slew rate.
A1 biases the LT1074’s VC pin via A2 to produce an 800Hz
half-sine signal at L2’s center tap (Trace E). Because Q1 and
Q2 are synchronously driven with the reference half-sine
their drain waveforms (Traces F and G) reveal alternate
chopping of complete half cycles. L2 receives balanced
drive and its secondary recombines the chopped half-sines
into a 115VAC 400Hz sinewave output (Trace H). The diode
bridge rectifies L2’s output back to an 800Hz half-sine which
is fed to A1 via the resistor divider. A1 balances this signal
against the reference half-sine to close a loop. Transmitting
the 800Hz waveform around the loop requires attention
to available bandwidth. The LT1074’s 100kHz switching
frequency is theoretically high enough to permit this, but
the bandwidth attenuation of its output LC filter must be
considered. The unusually low output filter capacitor value
allows the necessary frequency response. A1’s 330k-0.01μF
components combine with the RC lead network across the
16k feedback resistor to stabilize the loop.
A2 closes a local loop around the LT1074 configuration.
This is necessary because L2 blocks DC information from
being conducted around A1’s loop. This is a concern because the waveform presented to L2’s primary center tap
must have no DC component. DC content at this point will
cause waveform distortion, transformer power dissipation
or both. The LT1074’s VC pin operates with substantial
and uncertain DC bias, making A1’s inability to control
DC errors unacceptable. A2 corrects this by biasing the
LT1074 VC pin at its DC threshold so that no DC component
is presented to L2. A1’s output represents the difference
between the AC-coupled circuit output and the half-sine
reference. A2’s output contains this information in addition to DC restoration information. L2 and A1 contribute
essentially no DC error, so A2’s loop may be closed at the
LT1074 configuration’s output. A2’s feedback capacitor
stabilizes this local loop.
The drive to L2 cannot sink current. This means that any
residual energy stored in L2 when the drive waveform goes
to zero sees no exit path. This is a relatively small effect, but
can cause output crossover distortion. The synchronous
switch option shown on the schematic provides such a
path, and is recommended for lowest output distortion.
This optional circuitry is detailed in Appendix E.
Figures 33a and 33b show waveforms in the “turnaround”
region of circuit operation. This is the most critical part
of the converter, and its characteristics directly determine
output waveform purity. Figure 33a (Trace A), a highly
amplitude and time expanded version of L2’s center tap
drive, arrives at 0V (upper cross-etched horizontal line)
and turns around cleanly. This action is just slightly time
skewed from the reference synchronized pulse (Trace B).
The aberration on the rising edge is due to the optional
synchronous switch’s operation. This switch is shorted
during the on-time of Trace C’s pulse (see Appendix E for
operating details of this option). Trace D, Q2’s gate drive,
aligns with Trace B’s pulse. The slew reduction caused by
Note 10: Complete operating details of the half-sine reference generator
appear in Appendix E.
an35f
AN35-14
VC
1N4148
GND
LT1074
800Hz
CHOPPER
SYNC
1k
1k
Q3
2N3904
1k
12V
NC
SYNC
FB
VSW
FOR DETAIL SEE
APPENDIX E
TIMING AND
REFERENCE HALF-SINE
GENERATOR
SYNC OPTION—
SEE TEXT
100μF
VIN
+
A2
1/2 LT1013
28V
0.22
SYNC OPTION—
SEE TEXT
MBR745
10k
0.01
Q
74C74
CLK
D
Q
A1
1/2 LT1013
330k
12V
SYNCHRONOUS
SWITCH OPTION
(SEE TEXT
AND APPENDIX E)
Q1, Q2 = MTP3055E
L1 = PULSE ENGINEERING, INC. #PE-92112
L2 = TRIAD TY-75
0V TO 2.5V
800Hz HALF-SINES
2.2
1k
1k
28V
LT1086-12
IN
OUT
GND
100k
0.01
S Q2 D BK-G
BK
S Q1 D BK-Y
0.01
+
12V
L2
22μF
RED
RED-Y
120pF
470k
AN35 F31
100Ω
OUTPUT
ADJUST
200Ω
16k 1W
WIREWOUND
1%
DANGER! LETHAL POTENTIALS
PRESENT IN SCREENED AREA
(FOR FULLY ISOLATED
OUTPUT SEE TEXT)
110VAC
400Hz OUTPUT
1N4004
s4
110VAC
400Hz OUTUT
Figure 31. LT1074 Based +28 to 110VAC 400Hz Converter. Sinewave Output Shows Only 1.6% Distortion. DANGER! Lethal Potentials Present—See Text
+
+
+
28VIN
–
–
L1
100μH
Application Note 35
an35f
AN35-15
Application Note 35
the 1k-0.01μF filter is clearly visible, and contributes to
Trace A’s low noise turnaround. The LT1074’s 100kHz chopping related components are easily observed in Trace A.
Waveforms at the next half cycle’s zero point (e.g., Q1’s
gate driven) are identical.
Figure 33b shows additional details at highly expanded
amplitude and time scales. L2’s center tap is Trace A, Q1’s
drain is Trace B and Q2’s drain Trace C. The output sinewave
(Trace D) is shown as it crosses through zero.
A = 5V/DIV
B = 10V/DIV
A = 200V/DIV
C = 50V/DIV
B = 0.01V/DIV
(1.6% DISTORTION)
D = 50V/DIV
E = 20V/DIV
C = 30VRMS/DIV
F = 50V/DIV
G = 50V/DIV
H = 200V/DIV
HORIZ = 500μs/DIV
AN35 F32
Figure 32. +28 to 110VAC, 400Hz Converter’s Waveforms. The
Optional Synchronous Switch is Disabled in this Photo, Resulting
in Relatively High Crossover Distortion (Trace H). DANGER!
Lethal Potentials Present—See Text
A = 1V/DIV
B = 5V/DIV
C = 20V/DIV
D = 20V/DIV
HORIZ = 50μs/DIV
AN35 F33a
Figure 33a
A = 2V/DIV
AN35 F34
A AND B HORIZ = 500μs/DIV
C HORIZ = 20Hz/DIV—400Hz CENTER FREQUENCY
Figure 34. Distortion and Spectral Characteristics for the
Sinewave Output Converter. Distortion Trace (B) Shows Crossover
Aberrations and the LT1074 Wideband Chopping Residue. The
Synchronous Switch Option is EmpIoyed in this Photo for Lowest
Distortion. DANGER! Lethal Potentials Present—See Text
Figure 34 studies waveform purity. Trace A is the sinewave
output at 50W loading. Trace B shows distortion products,
which are dominated by turnaround related crossover aberrations and LT1074 100kHz chopping residue. Although
not strictly necessary, the LT1074’s switching can be
synchronized to the reference half-sine for coherent noise
characteristics. This option is discussed in Appendix E,
along with other reference generator details. Trace C is
a spectrum analysis centered at 400Hz11. In this photo
the optional synchronous switch is used, accounting for
improved crossover characteristics over Figure 32.
If a fully floating output is desired the output diode bridge
can be isolated by a simple 1:1 ratio transformer. To calibrate this circuit trim the “output adjust” potentiometer
for a 115VAC output. Regulation remains within 1% over
wide variations of input and load.
B = 5V/DIV
C = 5V/DIV
D = 50V/DIV
11Test equipment aficionados may wish to consider how this picture was
HORIZ = 50μs/DIV
AN35 F33b
Figure 33b
taken. Hint: Double exposure techniques were not used. This photograph
is a real time, simultaneous display of frequency and time domain
information.
Figure 33a and 33b. Details of “Turnaround” Sequence.
Switching Characteristics Directly Determine Output Crossover
Distortion. DANGER! Lethal Potentials Present—See Text
an35f
AN35-16
Application Note 35
REFERENCES
1. Williams, J., “High Efficiency Linear Regulators,” Linear
Technology Corporation, Application Note 32.
2. Williams, J and Huffman, B., “Some Thoughts on DC/DC
Converters,” Linear Technology Corporation, Application
Note 29.
3. LT1074 Data Sheet, Linear Technology Corporation.
4. Nelson C., “LT1070 Design Manual,” Linear Technology
Corporation, Application Note 19.
5. Williams, J., “Switching Regulators for Poets,” Linear
Technology Corporation, Application Note 25.
7. Williams, J., “Power Conditioning Techniques for Batteries,” Linear Technology Corporation, Application
Note 8.
8. Pressman, A.I., “Switching and Linear Power Supply,
Power Converter Design,” Hayden Book Co., Hasbrouck
Heights, New Jersey, 1977, ISBN 0-8104-5847-0.
9. Chryssis, G., “High Frequency Switching Power Supplies, Theory and Design,” McGraw Hill, New York,
1984, ISBN 0-07-010949-4.
Note: This application note was derived from a manuscript
originally prepared for publication in EDN magazine.
6. Williams, J., “Applications of New Precision Op Amps,”
Linear Technology Corporation, Application Note 6.
APPENDIX A
Physiology of the LT1074
The LT1074 uses standard (as opposed to current mode)
pulse width modulation, with two important differences.
First, it is a clocked system with a maximum duty cycle of
approximately 95%. This allows a controlled start-up when
it is used as a positive-to-negative converter or a negative
boost converter. Second, duty cycle is an inverse function
of input voltage (DC ≈ 1/VIN), without any change in error amplifier output. This greatly improves line transient
response and ripple rejection, especially for designs which
have the control loop over-damped.
Referring to the block diagram, the heart of the LT1074
consists of the oscillator, the error amplifier A1, an analog
multiplier, comparator C6, and an RS flip-flop. A complete
switching cycle begins with the reset (down ramp) period
of the oscillator. During this time (≈0.7μs), the RS flipflop is set and the switch driver Q104 is kept off via the
“and” gate G1. At the end of the reset time, Q104 turns
on and drives the output switch Q111, Q112 and Q113.
The oscillator ramp starts upward, and when it is equal to
the output voltage of the analog multiplier, C6 resets the
RS flip-flop, turning off the output switch. Duty cycle is
therefore controlled by the output of the multiplier which in
turn is controlled by the output of the error amplifier, A1.
A multiplier is used in the LT1074 to provide a perfect
“feedforward” function. Conventional switching regulators
sometimes use a simple form of feedforward to adjust
duty cycle immediately when input voltage changes. This
reduces the requirement for voltage swing at the output of
the error amplifier as it tries to correct for line variations.
Bandwidth of switching regulator error amplifiers must
be fairly low to maintain loop stability, so rather large
output perturbations occur when the output of the error
amplifier must move quickly to correct for line variations.
Conventional feedforward schemes typically operate well
over a restricted input voltage range or are effective only
at certain frequencies. The multiplier technique is very
effective over the full range of input voltage and at all
frequencies. The basic function is to compensate for the
generalized buck regulator transfer function; VOUT = (VIN)
(DC), where DC = switch duty cycle. This transfer function
has two implications. First, it is obvious that to maintain
a constant output, duty cycle must change inversely with
input voltage. Second, input voltage appears in the loop
transfer function, i.e., a fixed variation in duty cycle gives
different variations in output voltage depending on input
voltage. Loop gain is directly proportional to input voltage,
an35f
AN35-17
Application Note 35
and this can cause loop instability or slow loop response
if input voltage varies over a wide range. The multiplier
takes out all input voltage effects by automatically adjusting loop gain inversely with input voltage. The multiplier
output (VO) is equal to error amplifier output (VE) divided
by input voltage (VIN); VO = (VR) • (VE)/(VIN). VR is a
fixed voltage required by all analog multipliers to define
multiplier gain. It has an effective value of approximately
20V in the LT1074.
The error amplifier used in the LT1074 is a transconductance
type. It has high output impedance (≈500kΩ), so that its
AC voltage gain is defined by the impedance of external
shunt frequency compensation components (ZC) and the
transconductance (gm) of the amplifier, AV = (gm) (AC).
gm is ≈3500μmho. The error amplifier has its noninverting
input committed to an internal 2.3V reference. The inverting input (fb) is brought out for connection to an external
voltage divider that establishes regulator output voltage.
Two other connections are made internally to the fb pin. A
window comparator consisting of C4, C5, and some logic
provides an “output status” function. It monitors the voltage on the fb pin and gives a “high” output only when the
fb voltage is within ±5% of the internal reference voltage.
This status output can be used to alert external circuitry
that the regulator output is “in” or “out” of regulation. The
delay and one-shot circuits ensure that switching EMI will
not cause spurious outputs, and that the minimum time
for an “out-of-bounds” (low) status output is ≅20μs. Also
tied to the fb pin is a frequency shift circuit consisting of
R15 and Q36. The base of Q36 is biased at ≅1V so that
Q36 turns on when the fb pin drops below ≈0.6V. Current
through Q36 smoothly decreases oscillator frequency.
This is necessary for maintaining control of current limit
at high input voltages. A “dead short” on the output of a
switching regulator requires that switch “on” time reduce
to (VD)/(VIN)(f), where VD is the forward voltage of the
output catch diode and f is switching frequency. VD is
typically 0.5V for a Schottky catch diode, forcing switch
“on” time to shrink to a theoretical 0.1μs for a 50V input
and 100kHz switching frequency with a shorted output.
In practical circuits, effective “on” time can stretch to
0.3μs under these conditions due to losses in the inductor
wire resistance and switch rise and fall time. The LT1074
cannot reduce switch “on” time to less than ≅0.6μs in
current limit because it has true pulse-by-pulse switch
current limiting. The current limit circuitry must sense
switch current after the switch turns on, and then send a
signal to turn the switch off. Minimum time for this signal
path is 0.6μs. Full control of current limit is maintained
by reducing switching frequency when the output falls to
less than approximately 15% of its regulated value. This
has no affect on normal operation and does not change
the selection of external components such as the inductor
or output capacitor.
True pulse-by-pulse current limiting is performed by comparator C7. It monitors the voltage across sense resistor
R52 and resets the RS flip-flop. Current limit threshold is
set by the voltage across R47 which in turn is set by the
voltage on the ILIM pin. The ILIM voltage is determined by
an external resistor or by an internal clamp of 5V if no
external resistor is used. To compensate for the temperature coefficient of R47 (≈ +0.25%/°C), the internal current
source IL has a matching positive temperature coefficient.
Its nominal value is 300μA at 25°C. Current limit can be
set from 1A to 6A with one external resistor between ILIM
and ground. If no resistor is used, the ILIM pin will self
clamp at ≅5V and current limit will be ≈6.5A. A small prebias is added to the negative input of C7 to ensure that
current limit will go to zero (no switching) when the ILIM
pin is pulled to 0V, either by an external short or via Q11
during undervoltage lockout. Soft-start can be achieved
by connecting a capacitor to the ILIM pin. Foldback current
limiting can also be implemented by connecting a resistor
from ILIM to the regulated output.
Switching frequency of the LT1074 is internally set at
100kHz, but can be increased by connecting a resistor
from the frequency pin to ground. This resistor biases on
Q79 and feeds extra current into the oscillator. Maximum
suggested frequency is 200kHz. A comparator, C3, is also
connected to the frequency pin and allows this pin to be
used for synchronizing the oscillator to an external clock,
even when the pin is also being used to boost oscillator
frequency. R35 keeps the frequency pin biased correctly
in a no-function state when it is left open and R36 limits
Q79 current if the frequency pin is accidentally shorted
to ground.
an35f
AN35-18
Application Note 35
INPUT
–
SHUT
+
+
SHUTDOWN
+
6V
REGULATOR
C1
R52
0.036Ω
R47
500Ω
10μA
+
IL
320μA
0.3V
Q111
–
Q112
C7
Q113
R57
400Ω
Q4
D3
R46
10k
–
2.5V
+
VSW
ILIM
Q11
C2
R58
15Ω
+
–
Q79
R35
+
C3
+
SYNC
SPEED
R36
1.75V
R
Q
S
SLOW OSC
FREQ
G1
Q104
R
SYNC
1.2V
Q36
COMOUT
Q87
+
D2
VIN
+
–
MULTIPLIER
(VC) (VR)
VIN
A1
ERROR
+ AMP
VC
+
VREF
2.21V
0.12V
–
C6
VO
VR ≈ 20V
+
+
–
C5
R15
OR
DELAY
ONE
SHOT
STATUS
AN35 FA1
+
FB
0.12V
+ –
C4
Figure A1. Simplified LT1074 Internal Details
an35f
AN35-19
Application Note 35
The shutdown pin on the LT1074 can be used as a logic
control of output switching, as an undervoltage lockout, or
to put the regulator into complete shutdown with ISUPPLY
≈ 100μA. Comparator C2 has a threshold of 2.5V. It forces
the output switch to a 100% “off” condition by pulling the
ILIM pin low via Q11. Undervoltage lockout is implemented
through C2 by connecting the tap of an input voltage
divider to the shutdown pin. Full micropower shutdown
of the regulator is achieved by pulling the shutdown pin
below the 0.30V threshold of C1. This turns off the chip
by shutting down the internal 6V bias supply. An internal
10μA current source pulls the shutdown pin high (inactive) if it is left open.
The Comout pin is an open-collector NPN whose collector
voltage is the complement of the switch output (VSW). Q87
is specified to drive up to 10mA and 30V. It is intended to
drive the gate of an external N-channel MOS switch which
is in parallel with the catch diode. The MOS switch then
acts as a synchronous rectifier, which can significantly
improve converter efficiency in low output voltage applications. The Comout pin can also be used to drive the
gate of an external P-channel MOS switch in parallel with
the internal bipolar switch to provide ultrahigh efficiency
switching at lower input voltages. A slight time-shift in the
Comout signal prevents switch overlap problems.
The combination of these features produces a DC/DC
converter with the electrical characteristics shown in
Figure A2.
PARAMETER
CONDITIONS
UNITS
Input Voltage Range
4.5V to 60V
Output Voltage Range
2.5V to 50V
Output Current Range
Standard Buck
Tapped Buck
0A to 5A
0A to 10A
Quiescent Input Current
7mA
Switching Frequency
100kHz to 200kHz
Switch Rise/Fall Times
50ns
Switch Voltage Loss
1A
5A
Reference Voltage
2.35V ±1.5%
Line/Load Regulation
Efficiency
1.6V
2V
0.05%
VOUT = 15V
VOUT = 5V
90%
80%
Figure A2. LT1074 Electrical Characteristics
APPENDIX B
GENERAL CONSIDERATIONS FOR SWITCHING
REGULATOR DESIGN
Inductor Selection
Magnetic considerations are easily the most common
problem area in switching regulator design. Ninety percent
of the difficulties encountered are traceable to the inductive components in the circuit. The overwhelming level of
difficulty associated with magnetics mandates a judicious
selection process. The most common difficulty is saturation. An inductor is saturated when it cannot hold any more
magnetic flux. As an inductor arrives at saturation it begins
to look more resistive and less inductive. Under these
conditions current flow is limited only by the inductor’s
DC copper resistance and the source capacity. This is why
saturation often results in destructive failures.
While saturation is a prime concern, cost, heating, size,
availability and desired performance are also significant.
Electromagnetic theory, although applicable to these issues,
can be confusing, particularly to the non-specialist.
Practically speaking, an empirical approach is often a
good way to address inductor selection. It permits realtime analysis under actual circuit operating conditions
using the ultimate simulator—a breadboard. If desired,
inductor design theory can be used to augment or confirm
experimental results.
Figure B1 shows a typical step-down converter utilizing
the LT1074 switching regulator. A simple approach may
be employed to determine the appropriate inductor. A very
useful tool is the #845 inductor kit12 shown in Figure B2.
This kit provides a broad range of inductors for evaluation
in test circuits such as Figure B1.
Note 12: Available from Pulse Engineering, Inc., P.O. Box 12235,
San Diego, California 92112, 619-268-2400.
an35f
AN35-20
Application Note 35
Figure B3 was taken with a 450μH value, high core capacity inductor installed. Circuit operating conditions such as
input voltage and loading are set at levels appropriate to
the intended application. Trace A is the LT1074’s VSW pin
voltage while Trace B shows its current. When VSW pin
voltage is high, inductor current flows. The high inductance means current rises relatively slowly, resulting in
the shallow slope observed. Behavior is linear, indicating
no saturation problems. In Figure B4, a lower value unit
with equivalent core characteristics is tried. Current rise
is steeper, but saturation is not encountered. Figure B5’s
selected inductance is still lower, although core characteristics are similar. Here, the current ramp is quite
pronounced, but well controlled. Figure B6 brings some
informative surprises. This high value unit, wound on a
low capacity core, starts out well but heads rapidly into
saturation, and is clearly unsuitable.
acceptable electrical results, and the “best” unit can be
further selected on the basis of cost, size, heating and other
parameters. A standard device in the kit may suffice, or a
derived version can be supplied by the manufacturer.
The described procedure narrows the inductor choice
within a range of devices. Several were seen to produce
HORIZ = 2μs/DIV
TEST
INDUCTOR
VIN
VIN
VSWITCH
LT1074
VCOMP
MBR745
Using the standard products in the kit minimizes specification uncertainties, accelerating the dialogue between user
and inductor vendor.
A = 10V/DIV
B = 1A/DIV
AN35 FB3
Figure B3. Waveforms for 450μH, High Capacity Core Unit
5V
OUTPUT
+
A = 10V/DIV
1000μF
GND FEEDBACK
2.49k
1%
B = 1A/DIV
2k
0.1
2.49k
1%
AN35 FB1
Figure B1. Basic LT1074 Test Circuit
HORIZ = 2μs/DIV
AN35 FB4
Figure B4. Waveforms for 170μH, High Capacity Core Unit
A = 10V/DIV
B = 1A/DIV
HORIZ = 2μs/DIV
Figure B2. Model 845 Inductor Selection Kit from Pulse
Engineering, Inc. Includes 18 Fully Specified Devices
AN35 FB5
Figure B5. Waveforms for 55μH, High Capacity Core Unit
an35f
AN35-21
Application Note 35
Inductor Selection—Alternate Method
There are alternate inductor selection methods to the
one described. One of the most popular is utilized by
those devoid of the recommended inductor kit, time or
adequate instrumentation. What is usually desired is to
get a prototype LT1074 circuit running NOW. What is
often available is limited to a drawer of inductors (see
Figure B7) of unknown or questionable lineage. Selection
of an appropriate inductor is (hopefully) made by simply
inserting one of these drawer dwellers into an unsuspecting LT1074 circuit. Although this method’s theoretical
premise is perhaps questionable, its seemingly limitless
popularity compels us to address it. We have developed
a 2-step procedure for screening inductors of unknown
characteristics. Inductors passing both stages of the test
have an excellent chance (75%—based on our sample of
randomly selected inductors) of performing adequately
in a prototype LT1074 circuit. The only instrumentation
required is an ohmmeter and a scale.
Test 1 consists of weighing the candidate inductor. Acceptable limits range between 0.01 and 0.25 pounds. This test is
best performed at an Inductor Test Facility (see Figure B8),
where precision scales are readily available. To save time
the quick checkout line is recommended (but only if you
have nine13 inductors or less—no cheating).
Figure B9 shows an inductor under test. The 0.13 pound
weight indicated by the scale places this unit well within
acceptable limits.
Note 13: The maximum permitted number of items in the quick checkout
line varies from facility to facility. Please be familiar with and respect local
regulations.
A = 10V/DIV
B = 1A/DIV
HORIZ = 2μs/DIV
AN35 FB6
Figure B6. Waveforms for 500μH, Low Capacity Core Inductor
(Note Saturation Effects)
Figure B7. “Yeah, We Got Some Inductors in a Drawer.
I’m Sure They’ll Work...”
Figure B8. Typical Inductor Test Facility
Figure B9. Inductor Under Test (Don’t Forget to Pick Up a Loaf of
Bread and a Dozen Eggs)
an35f
AN35-22
Application Note 35
The second test involves measuring the inductors DC
resistance. Acceptable limits are usually between 0.01Ω
and 0.25Ω. Inductors passing both tests will probably
function in a prototype LT1074 circuit. Figures B10 and
B11 show typical acceptable and unacceptable inductors.
Graduates tend to be relatively dense, with (where visible)
thick wire. Flunkers are usually less dense, with small
(again, where visible) wire sizes.
When using an inductor selected with this method try
low power first, then gradually increase loading. Observe
inductor and LT1074 heating, making sure their dissipation is reasonable (warm to the touch). Disproportionate
increases in heating as load is increased probably indicate
inductor saturation. Either reduce the load, or go back to
the drawer and try again.
While these two tests are somewhat lacking in rigor they do
increase the chances of quickly getting a circuit to run with
available components. In the longer term, the appropriate
inductor can be decided upon and specified.
For the theoretically minded, test 1 grades out inductors
which are unlikely to have enough flux storage capability
(core mass) to avoid saturation. Test 2 eliminates units with
too high a resistance to efficiently support typical LT1074
operating currents. Expanded discussion and design considerations for inductors will be found in Reference 4.
Capacitors
Think about requirements in capacitors. All operating
conditions should be accounted for. Voltage rating is the
most obvious consideration, but remember to plan for the
effects of equivalent series resistance (ESR) and inductance. These specifications can have significant impact
on circuit performance. In particular, an output capacitor
with high ESR can make loop compensation difficult or
decrease efficiency.
Layout
Figure B10. Typical Acceptable Inductors
Layout is vital. Don’t mix signal, frequency compensation,
and feedback returns with high current returns. Arrange
the grounding scheme for the best compromise between
AC and DC performance. In many cases, a ground plane
may help. Account for possible effects of stray inductor-generated flux on other components and plan layout
accordingly.
Diodes
Diode breakdown and switching ratings must be thought
through. Account for all conditions. Transient events usually cause the most trouble, introducing stresses that are
often hard to predict. Study the data sheet breakdown,
current capacity, and switching speed ratings carefully.
Were these specifications written under the same conditions that your circuit is using the device in? If in doubt,
consult the manufacturer.
Figure B11. Typical Unacceptable Inductors
Switching diodes have two important transient characteristics—reverse recovery time and forward turn-on time.
an35f
AN35-23
Application Note 35
Reverse recovery time occurs because the diode stores
charge during its forward conducting cycle. This stored
charge causes the diode to act as a low impedance conductive element for a short period of time after reverse drive
is applied. Reverse recovery time is measured by forward
biasing the diode with a specified current, then forcing a
second specified current backwards through the diode.
The time required for the diode to change from a reverse
conducting state to its normal reverse non-conducting
state is reverse recovery time. Hard turn-off diodes switch
abruptly from one state to the other following reverse
recovery time. They therefore dissipate very little power
even with moderate reverse recovery times. Soft turn-off
diodes have a gradual turn-off characteristic that can
cause considerable diode dissipation during the turn-off
interval.
Fast diodes can be useless if stray inductance is high in the
diode, output capacitor or LT1074 loop. 20-guage hook-up
wire has 30nH/inch inductance. Switching currents on the
order of 108A/sec are typical in regulator circuits. They can
easily generate volts per inch in wiring. Keep the diode,
capacitor and LT1074 input/switch lead lengths SHORT!
Frequency Compensation
The basic LT1074 step-down configuration is relatively
free of frequency compensation difficulties. The simple RC
damper networks shown from the VC pin to ground will
usually suffice. Things become more complex when gain
and phase contributing elements are added to the basic
loop. In these cases it is often useful to look at the LT1074
as a low bandwidth power stage. The delays are due to the
sampled data nature of power delivery (100kHz switching
frequency) and the output LC filter. In general, complex
loops can be stabilized by limiting the gain bandwidth of
the LT1074 below that of the added elements. This is in
accordance with well know feedback theory. A discussion
of practical techniques for stabilizing such loops, “The
Oscillation Problem (Frequency Compensation Without
Tears),” appears at the end of LTC Application Note 18.
Other pertinent comments appear in the “Frequency Compensation” sections of LTC Application Notes 19 and 25.
APPENDIX C
Techniques and Equipment for Current Measurement
Accurate measurement of current flow under rapidly
changing circuit conditions is essential to switching regulator design. In many cases current waveforms contain
more valuable information than voltage measurements.
The most powerful and convenient current measuring
tool is the clip-on current probe. Several types appear in
Figure C1. the Tektronix P-6042, shown bottom left, is a
Hall effect stabilized current transformer which responds
from DC to 50MHz. The more recent Tektronix AM-503
(not shown) has similar specifications. The combination of
convenience, broad bandwidth and DC response make Hall
effect stabilized current probes the instrument of choice
for converter design. The DC response allows determination of DC content in high speed current waveforms.
The clip-on probe contains a current transformer and a
Hall effect device. The Hall device senses at DC and low
frequency while the transformer simultaneously processes
high frequency content. Careful roll-off matching allows
a composite output with no peaking or response aberrations at the two sensors bandwidth crossover. Sensitivity
ranges from fractions of a milliampere to amperes and is
switch selectable.
Transformer based clip-on current probes are also available. These types lack the DC response of their Hall effect
augmented cousins, but are still quite useful. The Tektronix
type 131 (and the more modern 134) responds from hundreds of hertz to about 40MHz. AC current probes (type 131
appears in C1, upper left) are as convenient to use as Hall
types, but cannot respond at low frequency. AC current
probes are also available with a simple termination (left
foreground, Figure C1). These types are more difficult to
use than the actively terminated models (e.g., type 131
shown) because of complex gain switching. Their low
frequency limitations are also poorer, although their high
frequency response exceeds 100MHz.
an35f
AN35-24
Application Note 35
A final form of AC current probe is the simple transformer
shown in Figure C1’s foreground. These are not clip-on
devices, and usually have significant performance limitations compared to the instruments discussed. However,
they are inexpensive and can provide meaningful measurement results when used according to manufacturers’
recommendations. In use, the conductor is threaded
through the opening provided and the signal monitored
at the output pins.
Figure C1 also shows a wide ranging DC clip-on current
probe. The Hewlett-Packard 428B (upper right) responds
from DC to only 400Hz, but features 3% accuracy over
a 100μA to 10A range. This instrument obviously cannot
discern high speed events, but is invaluable for determining overall efficiency and quiescent current.
in multi-trace displays as time skewing between individual
channels. The current probes transit time delay can be
mentally factored in to reduce error when interpreting the
display. Note that active probes have the longest signal
transit times, on the order of 25ns.
The AC probes low frequency bandwidth restriction must
be kept in mind when interpreting CRT displays. Figure C2
clearly demonstrates this by showing the AC probes inability to follow low frequency. Similarly, remember that
the probe’s stated bandpass is a –3dB figure, meaning
signal information is not entirely present in the display
at this frequency. When working in regions approaching
either end of the probe bandpass consider that displayed
information may be distorted or incomplete.
There are other ways, albeit less convenient and desirable that clip-on current probes, to measure wideband
current signals. Ohm explains that measuring voltage
across a resistor gives current. Current shunts (Figure C3
foreground) are low value (for LT1074 circuits 0.1Ω to
A = 100mA/DIV
B = 100mA/DIV
Figure C1. Various Current Probe Types Provide Different
Capabilities. Selection Criteria is Application Dependent
A great strength of the probes described is that they take
a fully floating measurement. The extraction of current
information by magnetic connection eliminates common
mode voltage considerations. Additionally, the clip-on
convenience makes the probes as easy to use as a standard voltage mode probe. As good as they are, current
probes have limitations and characteristics which must
be remembered to avoid unpleasant surprises. At high
currents, probe saturation limits may be encountered.
Resultant CRT waveforms will be corrupted, rendering the
measurement useless and confusing the unwary. For Hall
types, measurement below a few hundred microamperes
is limited by noise, which is much more obvious on the
display. Keep in mind that current probes have different
signal transit delay times than voltage probes or dissimilar
current probes. At high sweep speeds this effect shows up
HORIZ = 500μs/DIV
AN35 FC2
Figure C2. Hall Stabilized (Trace A) and Transformer (Trace B)
Based Current Probes Responding to Low Frequency
Figure C3. Typical Current Shunts and an “Isolated” Probe
an35f
AN35-25
Application Note 35
0.01Ω is typical) resistors with four terminal connections
for accurate measurement. In theory, measuring voltage
across a current shunt should yield accurate information.
In practice, common mode voltages introduce measurement difficulties, particularly at speed. Making this
measurement requires an isolated probe or a high speed
differential plug-in. The Signal Acquisition Technologies
SL-10 (Figure C3) has 10MHz bandwidth, a galvanically
floating input and 600V common mode capability. This
probe allows any oscilloscope to take a floating measurement across a shunt.
Differential oscilloscope plug-ins, while not galvanically
floating, can measure across a shunt. Tektronix types W,
1A5 and 7A13 have 1mV sensitivity with up to 100MHz
bandwidth and excellent common mode rejection. Types
1A7 and 7A22 have 10μV sensitivity, although bandwidth
is limited to 1MHz. All differential plug-ins have bandwidth
and/or common mode voltage restrictions that vary with
sensitivity. These trade-offs must be reviewed when selecting the optimal shunt value for a particular measurement.
In general the smallest practical shunt value is desirable.
This minimizes the inserted resistances parasitic effects
on circuit operation.
APPENDIX D
Optimizing Switching Regulators for Efficiency
Squeezing the utmost efficiency out of a switching regulator
is a complex, demanding design task. Efficiency exceeding 80% to 85% requires some combination of finesse,
witchcraft and just plain luck. Interaction of electrical and
magnetic terms produces subtle effects which influence
efficiency. A detailed, generalized method for obtaining
maximum converter efficiency is not readily described
but some guidelines are possible.
Losses fall into several loose categories including junction,
ohmic, drive, switching, and magnetic losses.
Semiconductor junctions produce losses. Diode drops
increase with operating current and can be quite costly
in low voltage output converters. A 700mV drop in a 5V
output converter introduces more than 10% loss. Schottky
devices will cut this nearly in half, but loss is still appreciable. Germanium (rarely used) is lower still, but switching
losses negate the low DC drop at high speeds. In very low
power converters Germanium’s reverse leakage may be
equally oppressive. Synchronously switched rectification
is more complex, but can sometimes simulate a more
efficient diode (see LTC Application Note 29, Figure 32).
The LT1074’s “Com Out” pin is intended to drive external
synchronous switches. See Appendix A for details.
When evaluating synchronous rectification schemes
remember to include both AC and DC drive losses in efficiency estimates. DC losses include base or gate current in
addition to DC consumption in any driver stage. AC losses
might include the effects of gate (or base) capacitance,
transition region dissipation (the switch spends some time
in its linear region) and power lost due to timing skew
between drive and actual switch action.
The LT1074’s output switch is composed of a PNP driving
a power NPN (Figure D1). The switch drop can reach 2V
at high currents. This will usually be the major loss in the
circuit. Its effect on efficiency can be mitigated by using
the highest possible input voltage. Text Figure 7 shows
5V regulator efficiency improving almost 10% for higher
input voltages. Higher output voltages will further minimize
the switch losses.
SWITCH DROP ≈ VCE Q1 + VBE Q2
Q2
INPUT
PIN
Q1
VSW PIN
AN35 FD1
TO INTERNAL
DRIVE CIRCUITRY
Figure D1. Simplified LT1074 Output Switch
Actual losses caused by switch saturation effects and
diode drops are sometimes difficult to ascertain. Changing
duty cycles and time variant currents make determination
tricky. One simple way to make relative loss judgments is
to measure device temperature rise. Appropriate tools here
an35f
AN35-26
Application Note 35
include thermal probes and (at low voltages) the perhaps
more readily available human finger. At lower power (e.g.,
less dissipation, even though loss percentage may be as
great) this technique is less effective. Sometimes deliberately adding a known loss to the component in question
and noting efficiency change allows loss determination.
Ohmic losses in conductors are usually only significant
at higher currents. “Hidden” ohmic losses include socket
and connector contact resistance and equivalent series
resistance (ESR) in capacitors. ESR generally drops with
capacitor value and rises with operating frequency, and
should be specified on the capacitor data sheet. Consider
the copper resistance of inductive components. It is often
necessary to evaluate trade-offs of an inductor’s copper
resistance versus magnetic characteristics.
Switching losses occur when the LT1074 spends significant
amounts of time in its linear region relative to operating
frequency. At higher switching frequencies transition times
can become a substantial loss source. The LT1074’s 100kHz
preset switching frequency is a good compromise (for this
device) and changes should be carefully considered. Raising the switching frequency to gain some desired benefit
necessitates consideration of increased LT1074 losses.
200kHz is the maximum practical operating frequency.
Magnetics design also influences efficiency. Design of
inductive components is well beyond the scope of this appended section, but issues include core material selection,
wire type, winding techniques, size, operating frequency
current levels, temperature and other issues. Some of
these topics are discussed in LTC Application Note 19,
but there is no substitute for access to a skilled magnetics specialist. Fortunately, the other categories mentioned
usually dominate losses, allowing good efficiencies to be
obtained with standard magnetics. Custom magnetics
are usually only employed after circuit losses have been
reduced to lowest practical levels.
A Special Circuit
In cases where input voltage must be low, but may float,
Figure D2’s circuit may be preferable to an LT1074 based
approach. This circuit uses the LT1070, a common emitter
output device. With the emitter connected to the ground
pin this device’s (LT1070 operating details are available
in its data sheet, and in LTC Application Notes 19, 25 and
29) switch loss is significantly lower than the LT1074’s.
Although intended for voltage step-up in flyback configurations the LT1070 can be arranged to perform the step-down
function. The advantage is the efficiency gain due to the
reduced switch loss. The circuit’s primary restriction is
that the input must float with respect to the output. Q1
performs a level shift to get the feedback information
referenced to the LT1070 “ground” pin, which floats with
the input. The LT1070 is effectively “fooled” and behaves
like a flyback regulator. It is oblivious to the fact that the
overall function is step down, because the floating input
is driven to the output potential. The negative side of the
output filter capacitor is connected to the ground ( ) of
the powered system, and the LT1070 input rail becomes
the 5V output. Other voltages are obtainable by altering the
3.9k – 1.1k feedback ratio. Efficiency approaches 85%.
MBR745
RL
+
4700μF
Q1
2N5401
LT1070
VIN = 12V
(10V TO 40V)
3.9k*
VSW
VIN
+
L1
170μH
VOUT
5V
AT 4A
VFB
VC
GND
1k
1.1k*
+
100μF
1μF
AN35 FD2
* = 1% FILM RESISTORS
MBR745 = MOTOROLA
L1: PULSE ENGINEERING, INC. #PE-92113
Figure D2. Floating Input Buck Regulator
an35f
AN35-27
Application Note 35
APPENDIX E
A Half-Sine Reference Generator
Text Figure 31’s half-sine reference must be amplitude
and frequency stabilized to a fairly high degree. It is not
unreasonable to expect a 115VAC 400Hz source to be
within 1V and 0.1Hz. Additionally, Figure 31’s reference
requires a half-sine, as opposed to the more normal fullsine. These requirements are achievable by classical analog
techniques, but a digital approach eases complexity with
no performance trade-off14. Figure E3 shows such an
approach. C1 forms a 1.024MHz crystal oscillator which
is divided down by the 7490. The 7490’s differentiated
÷10 output becomes the LT1074’s 102.4kHz sync. option output. The 7490’s ÷5 output (204.8kHz) is fed to
the 74191 counters. These counters parallel load a 2716
EPROM which is programmed to produce an 8-bit (256
states) digitally coded half-sine. The program, developed
by Sean Gold and Guy M. Hoover, appears in Figure E1.
The 2716’s parallel output is fed to an 8-bit DAC, which
produces 800Hz 2.5V (peak) half-sines.
Figure E1. Half-Sine Software Code for the 2716 EPROM
Figure E2. Full-Sine Software Code for the 2716 EPROM (Bonus)
Those wishing to utilize this reference for full-sines will
find the appropriate software in Figure E2.
Figure E3 also shows the synchronous switch option discussed in the text. The 74C122 monostable forms a simple
delayed pulse generator which drives the Q4 switch. The
20μs delay and 6μs pulse width set at the 74C122 were
empirically determined to produce lowest overall crossover
distortion in Figure 31’s output.
Note 14: The sinewave is probably the paramount expression of the analog
world. The Old Man Himself, George A. Philbrick, once elegantly discussed
analog functions as “those which are continuous in excursion and time.”
He might have viewed digital production of a sinewave with considerable
suspicion, or simply labeled it blasphemous. Such are the wages of
progress.
an35f
AN35-28
Application Note 35
2k
5V
1.024MHz
5V
5
FROM
12V ON
FIGURE 31
+
Q
C1
LT1016
2k
+
22
Q
–
0.0068
LT1086-5
2k
5V
7490
VCC
5V
BIN
CLK
D
A
204.8kHz
P/U
74191
ENABLE
QA
A0
QB
A1
QC
A2
QD
A3
Q0
Q1
Q2
Q3
Q4
Q5
Q6
Q7
RIPPLE CLOCK
22pF
20k
2716
2.5V
800Hz
HALF-SINE
OUTPUT
1N4148
CLK
5V
74191
ENABLE
CLR
AD558
QA
A4
QB
A5
QC
A6
CE
QD
A7
OE
A9 A10
2.56V
SENSE
RIPPLE CLOCK
A8
5V
800Hz CHOPPER
SYNC OUT
102.4kHz
1074 SYNC
OPTION OUTPUT
1000pF
20k
1000pF
12k
12V
12V
RC1
12V
1A
2A
2CE
1CE
RC2
100Ω
74C221
1B
1Q
2B
1C
2Q
2C
Q
Q4
MTP-3055
SYNCHRONOUS
SWITCH OPTION—
TO L2 CENTER TAP,
FIGURE 31
AN35 FE3
SYNCHRONOUS
SWITCH OPTION—
FROM Q3 COLLECTOR, FIGURE 31
Figure E3. Timing and Reference Half-Sine Generator for Figure 31
an35f
AN35-29
Application Note 35
APPENDIX F
The Magnetics Issue
Magnetics is probably the most formidable issue in
converter design. Design and construction of suitable
magnetics is a difficult task, particularly for the non-specialist. It is our experience that the majority of converter
design problems are associated with magnetics requirements. This consideration is accented by the fact that
most converters are employed by non-specialists. As a
purveyor of switching power ICs we incur responsibility
towards addressing the magnetics issue (our publicly
spirited attitude is, admittedly, capitalistically polluted).
As such, it is LTC’s policy to use off-the-shelf magnetics
in our circuits. In some cases, available magnetics serve
a particular design. In other situations the magnetics have
been specially designed, assigned a part number and made
available as standard product.
In many circumstances a standard product is suitable
for production. Other cases may require modifications
or changes which the manufacturer can provide or advise on. Hopefully, this approach serves the needs of all
concerned.
Recommended magnetics manufacturers include the
following;
Pulse Engineering, Inc
P.O. Box 12235
7250 Convoy Court
San Diego, California 92112
619-268-2400
Coiltronics
984 Southwest 13th Court
Pompano Beach, FL 33069
305-781-8900
an35f
AN35-30
Application Note 35
an35f
Information furnished by Linear Technology Corporation is believed to be accurate and reliable.
However, no responsibility is assumed for its use. Linear Technology Corporation makes no representation that the interconnection of its circuits as described herein will not infringe on existing patent rights.
AN35-31
Application Note 35
an35f
AN35-32
Linear Technology Corporation
IM/GP 989 20K • PRINTED IN USA
1630 McCarthy Blvd., Milpitas, CA 95035-7417
(408) 432-1900 ● FAX: (408) 434-0507
●
www.linear.com
© LINEAR TECHNOLOGY CORPORATION 1989