J. Lenk Simplified Design of Switching Power Supplies V. Lakshminarayanan Electronic Circuit Design Ideas J. Lenk Simplified Design of Linear Power Supplies M. Brown Power Supply Cookbook
B. Travis and I. Hickman EDN Designer's Companion J. Dostal Operational Amplifiers, Second Edition
R. Marston Electronics Circuits Pocketbook: Passive and Discrete Circuits (Vol. 2)
N. Dye and H. Granberg Radio Frequency Transistors: Principles and Practical Applications
Gates Energy Products Rechargeable Batteries: Applications Handbook T. Williams EMCfor Product Designers
J. Williams Analog Circuit Design: Art, Science, and Personalities R. Pease Troubleshooting Analog Circuits
I. Hickman Electronic Circuits, Systems and Standards
R. Marston Electronic Circuits Pocket Book: Linear ICs (Vol. I) R. Marston Integrated Circuit and Waveform Generator Handbook I. Sinclair Passive Components: A User's Guide
Simplified Design of Switching Power Supplies
John D. Lenk
Butterworth-Heinemann
Boston Oxford Melbourne Singapore Toronto Munich New Delhi Tokyo
Copyright © 1995 by Butterworth-Heinemann -GL A member of the Reed Elsevier group
All rights reserved.
No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of the publisher.
Recognizing the importance of preserving what has been written, it is the policy of Butterworth-
© Heinemann to have the books it publishes printed on acid-free paper, and we exert our best efforts to that end.
Library of Congress Cataloging-in-Publication Data Lenk, John D.
Simplified design of switching power supplies / by John D. Lenk.
p. cm.
Includes bibliographical references and index.
ISBN 0-7506-9507-2
1. Electronic apparatus and appliances—Power supply—Design and construction. 2. Switching power supplies—Design and construction. I. Title.
TK7868.P6L456 1995
621.381Ό44—dc20 94-32727 CIP
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library.
Butterworth-Heinemann 313 Washington Street Newton, MA, 02158
10 9 8 7 6 5 4 3 2 1
Printed in the United States of America
Thank you for being by my side all these years!
To my lovely family, Karen, Tom, Brandon, Justin, and Michael, and to our Lambie and Suzzie: Be happy wherever you are!
And to my special readers: May good fortune find your doorway, bring
ing you good health and happy things. Thank you for buying my books!
To Frank Satlow, Karen Speerstra, Aliza Lamdan, John Fuller, the U.K people, and the EDN people: A special thanks for making me an inter
national best-seller, again (this is book number 81)!
Abundance!
Preface
This book has something for everyone. No matter what your skill level in electronics, this book will show you how you can immediately experiment with, test, interconnect, and design switching power supplies.
For experimenters, students, and serious hobbyists, the book provides suffi
cient information to design and build switching power supplies "from scratch." The design approach here is the same one used in all of the author's best-selling books on simplified and practical design.
The first four chapters provide the basics for all phases of practical design, in
cluding test and troubleshooting for switching supplies. The final chapter includes well over 100 worked-out design examples, using the techniques described in the first four chapters.
Throughout the book, design problems start with approximations or guide
lines for selecting all components on a trial-value basis, assuming a specific design goal and set of conditions. Then, using these approximate values in experimental circuits, the desired results (input/output voltage and current, line and load regula
tion, ripple rejection, noise, etc.) are produced by varying the test component values.
For service technicians and field-service engineers, an entire chapter is de
voted to practical test and troubleshooting. All of the tests can be performed with basic electronic test equipment.
If you are a working engineer responsible for designing and/or selecting switching power supplies, the variety of circuits and configurations described here should generally simplify your task. Not only does the book describe basic switch
ing-supply designs, but also it covers the most popular forms of IC switching regu
lators. A discussion of heat sinks is included, as are practical mounting and interconnection techniques for switching supplies. Throughout the book, you will find a wealth of information on switching-supply components and component man
ufacturers.
Chapter 1 is devoted to basic switching power-supply circuits, including IC switching regulators and DC-DC converters.
xiii
Chapter 2 covers heat sinks and other temperature-related design problems for switching power supplies.
Chapter 3 is devoted to the inductors and transformers used in switching-regula
tor circuits. Here the emphasis is on simplified design, not on mathematical analysis.
Chapter 4 is devoted to testing and troubleshooting for switching power sup
plies. The procedures can be applied to a just-completed supply circuit during de
sign and experimentation or to a suspect supply as design review.
Chapter 5 is devoted to design examples for switching supplies, using off-the- shelf components. All of the design techniques discussed in the first four chapters are used as needed. These circuits can be put to immediate use as is or, by alternat
ing the component values, used as a basis for simplified design of similar switching supplies. Here the emphasis is on how circuit performance can be changed to meet other application requirements by changing components.
Acknowledgments
Many professionals have contributed to this book. I gratefully acknowledge their tremendous effort in making this work so comprehensive: it is an impossible job for one person. I thank all who contributed, directly or indirectly.
I give special thanks to Syd Coppersmith of Dallas Semiconductor, Rosie Hi- nojosa and Kellie Garcia of EXAR Corporation, Jeff Salter of GEC Plessey, Linda daCosta and John Allen of Harris Semiconductor, Ron Denchfield of Linear Tech
nology, David Fullagar of Maxim Integrated Products, Fred Swymer of Microsemi Corporation, Linda Capcara of Motorola, Inc., Andrew Jenkins and Shantha Natara- jan of National Semiconductor, Antonio Ortiz of Optical Electronics Inc., Lawrence Fogel of Philips Semiconductors, Lorraine Jenkins of Raytheon Company Semicon
ductor Division, Anthony Armstrong of Semtech Corporation, Ed Oxner and Robert Decker of Siliconix Inc., Amy Sullivan of Texas Instruments, and Alan Campbell of Unitrode Corporation.
I also thank Joseph A. Labok of Los Angeles Valley College for help and en
couragement throughout the years.
Very special thanks to Frank Satlow, Karen Speerstra, Aliza Lamdan, John Fuller, the U.K. people, and the EDN people of Butterworth-Heinemann for having so much confidence in me. I recognize that all books are a team effort and am thank
ful that I am working with the New First Team on this series.
And to Irene, my wife and super agent, I extend my thanks. Without her help, this book could not have been written.
x v
Switching Power-Supply Basics
This chapter is devoted to basic switching power-supply circuits (also known as switch-mode power-supply circuits). As in the case of series or linear power sup
plies, it is possible to design switching supplies "from scratch," including the oscil
lator required for a switching regulator. However, switching regulators are available in integrated circuits (IC) form, and it is generally simpler to use such ICs.
The data sheets for IC switching regulators often show the connections and provide all necessary design parameters to convert the IC to a complete supply by adding external components. This chapter describes the functions and operations of the switching-regulator circuits (to help you understand the data sheet information).
The chapter concludes with a summary of the most common types of IC switching regulators.
1.1 Basic Switching-Regulator Functions
Figure 1-1 shows the block diagram of a basic switching regulator. The func
tion of this circuit is to convert an unregulated direct current (DC) input to a regu
lated DC output. For this reason, switching regulators are often referred to as DC-DC converters.
In a switching regulator the power transistor is used in a switching (or on/off) mode rather than in the continuous mode of a linear supply. As a result, switching regulator efficiency is usually in the 70 to 95% range, which is more than double that of linear regulators. In addition to increased efficiency, switching regulators can provide outputs that are greater than the input, if desired. The output of linear regulators is always lower than the input. Switching regulators can also invert the input (produce a positive output for a negative input, and vice versa), unlike the conventional linear regulator. High-frequency switching regulators offer consider-
1
Unreguloted DC input
1 Oscillator
Power switch
Duty cycle control
Diode clamp
Sampling circuit
Pil r II i ci t Reaulated DC output
Figure 1-1 · Basic switching-regulator functions
able weight and size reductions and better efficiency at high power than do linear supplies.
I . J. J Switching-Regulator Problems
Switching regulators are not without special problems. In addition to requiring more complex circuits, switching regulators produce electromagnetic interference (EMI). However, with proper design, EMI can be reduced to acceptable levels. Such design techniques involve the use of low-loss ferrite cores for transformers and chokes, of high-permeability magnetic alloys for shielding, and of miniature semi
conductor and IC devices for switching and regulation circuits.
1.1.2 Switching-Duty Cycle
The circuit of Fig. 1-1 regulates by switching the series transistor (power switch) to either the on or off condition. The duty cycle of the series transistor deter
mines the average DC output. In turn, duty cycle is adjusted in accordance with a feedback that is proportional to the difference between the DC output and a refer
ence voltage.
1.7.3 Switching Frequency
Switching is usually at a constant frequency just above the audible range, al
though some switching regulators use a variable frequency with changing line and load. With some switching-regulator ICs, it is possible to set or change the switch
ing frequency with an external capacitor. One of the first design trade-offs to re
member is that higher frequencies are generally less efficient because transistor switching losses and ferrite-core losses increase. On the other hand, lower switching frequencies in the audible range may cause certain components to "sing" or may produce interference in audio circuits being powered by the regulator.
1.1.4 Transistor and Diode Characteristics
Switching regulators must use transistors with a gain-bandwidth product (fT) of at least 4 MHz to operate efficiently (an fT of 30 MHz is even better). Darlington transistors and MOSFETs are also used in switching regulators.
A fast-recovery rectifier, or a Schottky barrier diode, is used as a free-wheel
ing clamp diode to keep the switching-transistor load line within safe operating lim
its and to increase efficiency. Other solid-state devices used in some switching regulators include gates, flip-flops (FFs), op-amp comparators, timers, and recti
fiers.
1.2 Typical Switching-Regulator Circuits
Figure 1-2 shows four typical PNP/NPN switching-regulator circuits. All of the circuits have the following common elements: switching transistor, clamp diode, LC filter, and a logic or control block. None of the circuits provide full isolation be
tween the line and load, as would be the case if more than one series transistor is used. However, the one-transistor design is the simplest and most economical.
It is usually desirable to have at least one line in common with the input and output to reduce ground loops. The one-line approach also determines whether the output voltage is considered positive or negative. However, most circuits can oper
ate from either supply because the input and output grounds are usually isolated.
The one-transistor, one-line approach is the most popular switching-regulator de
sign.
In the circuits of Figs. l-2(a) and l-2(b), the logic or control operates from the load voltage. Such circuits are not self-starting, and provisions must be made to operate from the line during start-up (and in the event of short circuits).
In the circuits of Figs. l-2(c) and l-2(d), the logic operates continuously from the line and is isolated from the load voltage. The sense and feedback elements must be electrically isolated (sometimes with an optocoupler).
The circuits of Figs. l-2(b) and 1—2(d) are generally used in line-operated supplies because economical high-voltage NPN transistors are available whereas PNP types are not. Of the two, the circuit of Fig. l-2(d) is most popular, because the logic is tied directly to the series switch and switching is more efficient. Driver transformers are used in some designs to interface between the logic and switching transistors. In such a case, the switching transistor may be either PNP or NPN.
Figure 1-3 shows three typical MOSFET switching-regulator (or -converter) circuits, representing the three basic configurations: buck, boost, and buck-boost (all of which are described more fully in Section 1.5). In brief, each of the three config
urations meets a particular need. When output voltage is greater than the input, the converter is usually operated in the positive voltage-boost circuit (also known as a step-up converter). The buck circuit is used when the input voltage is always greater than the desired output voltage (and is also known as a step-down converter). The
-nsrn^—t 1 — +v„
tfio
• Ground(a) Positive o u t p u t , c o m m o n logic and load
Logic
-nm^-
Ground
(b) Negative o u t p u t , c o m m o n logic and load
/7ftftTv—T T— + v
\^n
Logic
/
7
Remote sense
■ Ground (c) Positive o u t p u t , isolated logic and load
I
t AI 2
Logic 1
Remote sense
\
I
1 1 1
G r o u n d
-νΛ (d) Negative o u t p u t , isolated logic and load
Figure 1 - 2 . Four typical switching-regulator circuits
T V«».u
3H
Control section• »O +
3: $
Voo, > Vb)Boost converter ■+0-
_ = _ Vb(
Control section
- K > H
vout <: vb„,
Buck converter -►O
"Π_Π
Control section
- • O
Buck-boost converter
|Vou,| either or
- K >
Figure 1 - 3 . Three typical MOSFET switching-regulator circuits (Maxim Seminar Appli
cations Book, 1989, p. 115)
buck-boost circuit inverts the input voltage and can be used with an input voltage that is either greater or less than the desired output. For this reason, the buck-boost circuit is sometimes called an inverter.
1.3 Switching-Regulator Theory
Figure 1-4 shows a theoretical switching-regulator circuit (buck configura
tion) and the related waveforms. The high efficiency of switching regulators is the result of operating the series transistor in a switching mode. When the transistor is switched on, the full input voltage is applied to the LC filter. When the transistor is switched off, the input voltage is zero. With the transistor turned on and off for equal amounts of time (50% duty cycle) the DC load voltage is half the input volt
age. The output voltage VQ is always equal to the input voltage V^, times the duty cycle D, or V0 = DV^.
Switch open
-^&ffiP-
— Λ
♦it
Closed
Figure 1 - 4 . Theoretical switching-regulator circuit (buck type)
Varying the duty cycle compensates for changes in the input voltage. This technique is used to produce the regulated output voltage, as follows. Repetitive op
eration of the switching transistor at a fixed duty cycle produces the steady-state waveforms shown in Fig. 1—4. With the switch closed, inductor current IL flows from the input voltage VIN to the load. The difference between the input and output voltage (VIN - V0) is applied across the inductor. This causes IL to increase during the "switch-closed" period.
With the switch open, stored energy in the inductor forces IL to continue to flow to the load and to return through the diode. The inductor voltage is then re
versed, and approximately equal to V0. IL decreases during the "switch-open" pe
riod.
The average current through the inductor equals the load current. Because the capacitor keeps VQ constant, the load current IQ is also constant. When IL increases above I0, the capacitor charges, and when IL drops below I0, the capacitor dis
charges (as shown by the theoretical waveforms).
The end results of steady-state operation are as follows:
1. The average inductor voltage is zero, but a wide variation from (VIN - V0) to V0 is experienced.
2. The direct current flowing through the inductor equals the load current. A small amount of sawtooth ripple is also present.
3. The DC voltage on the capacitor is equal to the load voltage. A small amount of ripple is also present at the capacitor.
To be effective, a regulator must provide compensation for changes in both VIN and I0. In a switching regulator, the VIN changes are automatically compensated for by duty-cycle variation in a closed-loop feedback system (see Fig. 1-1). Input
regulation and ripple rejection (both of which are discussed in Chapter 4) depend on loop gain.
Changes in I0 are more difficult to offset and load transient-response is gener
ally poor (rapid changes in load are difficult to regulate). Changes in I0 are offset with temporary duty-cycle changes. For example, a change in load from zero to full results in the following:
1. Duty cycle increases to maximum (transistor may simply stay on).
2. The inductor current takes many cycles to increase to the new DC level.
3. Duty cycle returns to the original value.
1.4 P W M versus PFM
Switching regulators (or converters) are also classified as to how they control output voltage. The two most common approaches are pulse width modulation (PWM) and pulse frequency modulation (PFM). Both approaches control output by varying the duty cycle.
With PWM regulators, the frequency is held constant and the width of each pulse is varied. PWM regulators are well established in high-power switching sup
plies that work off the alternating current (AC) line. With PFM regulators, the pulse width is held constant and the duty cycle is controlled by changing the pulse repeti
tion rate.
There are many variations of the two basic approaches. For example, the cur
rent-mode control is a refinement of PFM. Likewise, pulse skipping is a refinement of PFM. These and other variations of PWM/PFM are discussed throughout the re
mainder of this chapter.
1.5 Common Switching-Regulator Configurations
There are many possible switching-regulator configurations (or "topologies"
as engineers like to call them). The choice of which configuration to use is generally narrowed down by such factors as voltage polarity, voltage ratio, and fault condi
tions. For example, if the output voltage must be larger than the input, a buck con
verter cannot be used. If the input voltage is negative, and the output must be positive, some form of inverter is required. If the regulator must be current-limited, the basic boost circuit is of no value.
Even with these obvious limitations, there are still many choices of configura
tion for most applications. For example, to convert +28 to +5 V, the buck, flyback, forward, and current-boosted configurations could be used. The following discus
sion covers most of the configurations, describing both the capabilities and limita
tions of each. This is included to give you a head start when faced with the decision,
"which switching regulator should I use?"
1.5.1 Boost or Step-Up
Figure 1-5 shows the theoretical boost or step-up configuration. Figure 1-6 shows a typical IC switching regulator (the Raytheon RC4190) connected in a prac
tical step-up converter configuration. Figure 1-7 shows the corresponding wave
forms.
The following paragraphs describe how both theoretical and practical func
tions are performed and how the practical functions (Fig. 1-6) relate to the theoreti
cal (Fig. 1-5). Notice that switch S in the theoretical circuit is replaced by transistor Qj in the practical IC circuit. Capacitor C, diode D, and inductor L in Fig. 1-5 are replaced by C,, D,, and Lx in Fig. 1-6.
As shown in Fig. 1-5, when switch S is closed, the battery voltage is applied across the inductor L. Charging current flows through L, building up a magnetic field, increasing as the switch is held closed. While S is closed, diode D is reverse- biased (open circuit) and current is supplied to the load by capacitor C. Until S is opened, the current through L increases linearly to a maximum value determined by the battery voltage, inductor value, and amount of time S is held closed, or IPEAK = VBAT/(L x TON).
When S is opened, the magnetic field collapses, and the energy stored in the field is converted into a discharge current that flows through L in the same direction as the charging current. Because there is no path for current to flow through S, the current must flow through D to supply the load and charge output capacitor C. If the switch is opened and closed repeatedly (by some form of oscillator), at a rate much greater than the time constant of the output RC, then a constant DC voltage is pro
duced at the output.
An output voltage higher than the input voltage is possible because of the high voltage produced by a rapid change of current in the inductor. When S is opened, the inductor voltage rises high enough (instantly) to forward-bias D and adds to the battery voltage. In a practical IC regulator, a feedback-control system adjusts the on-time of switch S (controlling the level of the inductor current), so that the average inductor discharge current equals the load current, thus regulating the output voltage.
When power is first applied to the practical IC circuit (Fig. 1-6), the current in Rj supplies bias current to pin 6 of the IC. This current is stabilized by a unity-gain current-source amplifier and then used as a bias current for the 1.31-V reference (a bandgap-type reference in this case). The stable bias current generated by the refer
ence is used to bias the remainder of the IC components.
Figure 1 - 5 . Theoretical boost or step-up configuration (Raytheon Linear Integrated Cir
cuits, 1989, p. 9-7)
©ILX , — / Y Y T L
Lx
f - ^ ' c R1
|0
Θ'ο-H-
01
4190
1 5*
- = - VßAT
^ ^ g
:£ci I'LOAOL VOUT = V R £ F ( * + 1 )Θ
VOUT= V R E F ( - ^ - I )
Figure 1-6. Practical step-up converter (Raytheon Linear Integrated Circuits, 1989, p. 9-8)
At the same time the IC is starting up, current flows through Lx and Dj to charge C, to the battery voltage, less the drop across Dj (or VBAT - VD). At this point, the feedback (pin 7) senses that the output voltage VOUT is too low, by com
paring a division of VOUT to the +1.31-V reference. If VOUT is too low, then the comparator output changes to a logical zero. The NOR gate then combines the oscil
lator square wave with the comparator signal. If the comparator output is zero, and the oscillator output is low, then the NOR-gate output is high and switch transistor Qj is forced on. When the oscillator goes high again, the NOR-gate output goes low, and Qj turns off.
The turning on and off of the switch transistor Qj performs the same function the opening and closing of the switch (S in Fig. 1-5). That is, energy is stored in the inductor during the on-time and is released into output capacitor Cj during the off- time. The comparator continues to allow the oscillator to turn the switch transistor Qj on and off until enough charge is delivered to C, to raise the feedback voltage above 1.31 V.
Once the feedback voltage is above the reference, the feedback system varies the duration of the on-time in response to changes in load current or battery voltage, as shown by the waveforms in Fig. 1-7. If the load current increases (waveform C), then Qj remains on (waveform D) for a longer portion of the oscillator cycle, thus allowing the inductor current (waveform E) to build up a higher peak value. The duty cycle of the switch transistor Qj varies in response to changes in both load and line.
In any switching regulator, the inductor value and oscillator frequency must be carefully tailored to the battery voltage, output current, and ripple requirements of the application. (Chapters 3 and 5 describe this "tailoring" in boring detail.) In brief, if the inductor value is too high or the oscillator frequency is too high, then the
® Λ Α Α / ' S / V V
©"LTLTLT 1_Π_Γ 1_Γ
<a . J ~
® _n_n_jL_ n_Ji_ n_
VBAT VQUT - VB AT
© Γ-ιΡ-ιΡ- ΙΓΗΓ- a
Figure 1-7. Step-up converter waveforms (Raytheon Linear Integrated Circuits, 1989, p.
9-8)
inductor current never reaches a value high enough to meet the load-current drain, and the output voltage drops (or collapses). If the inductor value or oscillator fre
quency are too low, then the inductor current builds up too high, causing excessive output-voltage ripple. It is also possible for the switch transistor to be overdriven, or for the inductor to be saturated, with low inductor values and/or oscillator fre
quency.
1.5.2 Continuous versus Discontinuous
Operation of switching regulators can be continuous or discontinuous. With continuous operation, current through the inductor never drops to zero during the switch (transistor) off-time. With the discontinuous mode, if the load current is low enough, the inductor current can drop to zero in some cases.
Normally, it is not important to avoid the discontinuous mode at light load currents. A possible exception to this is when the on-time of the switch transistor cannot be reduced to a low enough value to prevent the lightly loaded output from drifting high. If this occurs, most switching regulators will begin "dropping cycles"
where the switch transistor does not turn on for one or more cycles. This maintains control of the output, but does produce subharmonic frequencies (which may or may not be acceptable in some applications).
The main problem with the discontinuous mode is when high load currents are involved, because a high ratio of switch current to output current is required for the discontinuous mode and switching "spikes" are produced. Such spikes are shown in Fig. 1-8, which illustrates the voltage and current waveforms for the switch (transistor), diode, and output capacitor of a boost regulator. As a general rule, with either continuous or discontinuous, the diode and output capacitor must be
MV 0.5V (Internal) Umax) OmA
0.72V (Internal)
ov
>MAX
0mA
'MAX
0mA
VOUT + Vo VBAT 0.3V (Q1 SAT)
Cx OSC
'L0A
VDD
ILK
ID
VLX
L1 D1
u
CONTINUOUS MODE
•inn
4ft"
■lrTr:
•ΊΓΈ:
DISCONTINUOUS MODE
) V0 UT - V IN
AA., UL
JUL-
Figure 1 - 8 . Waveforms for switch, diode, and output capacitor of a boost regulator (Lin
ear Technology, Linear Applications Handbook, 1990, p. AN 19-14)
specified to handle peak currents as well as average currents. Of course, the current limits for the switch (Qj in Fig. 1-6) are part of the IC specification.
1.5.3 Buck-Boost or Inverting
Figure 1-9 shows the theoretical buck-boost or inverting configuration. Fig
ure 1-10 shows a typical IC switching regulator (the Raytheon RC4391) connected in a practical inverting converter (or inverter) configuration. Figure 1-11 shows the corresponding waveforms.
| E L i c Voui ' * i 0 +
Figure 1 -9· Theoretical buck-boost or inverting configuration (Raytheon Linear Inte
grated Circuits, 1989, p. 9-54)
The following paragraphs describe how both theoretical and practical func
tions are performed and how the practical functions (Fig. 1-10) relate to the theoret
ical (Fig. 1-9). Notice that switch S in the theoretical circuit is replaced by transistor Q, in the practical IC circuit. Capacitor C, diode D, and inductor L in Fig. 1-9 are replaced by Cp Dp and Lx in Fig. 1-10.
As shown in Fig. 1-9, when switch S is closed, charging current from the bat
tery flows through inductor L, which builds up a magnetic field that increases as S is held closed. When S is opened, the magnetic field collapses, and energy stored in the magnetic field is converted into a current that flows through L in the same direc
tion as the charging current. Because there is no path for this current to flow through the switch, the current must flow through diode D to charge capacitor C. The key to inversion is the ability of the inductor to become a source when the charging current is removed.
In the practical circuit of Fig. 1-10, the feedback circuit and the output capac
itor decrease the output voltage across the inductor to a regulated fixed value. When power is first applied, the ground-sensing comparator (pin 8) compares the output voltage to the +1.25-V reference. Because CF is initially discharged, a positive volt
age is applied to the comparator, and the output of this comparator gates the square- wave oscillator. The gated square-wave signal turns the switch transistor Qx on and off.
The turning on and off of the switch transistor Qj performs the same function as opening and closing of the switch (S in Fig. 1-9). That is, energy is stored in the inductor during the on-time and is released into output capacitor CF during the off- time. The comparator continues to gate the oscillator square wave to Qx until enough energy is stored in CF to make the comparator input voltage decrease to less than 0 V. The voltage applied to the comparator is set by the output voltage, the ref
erence voltage, and the ratio of Rj to R2. 1.5.4 Buck or Step-Down
Figure 1-12 shows the theoretical buck or step-down configuration. Figure 1-13 shows a typical IC switching regulator (the Raytheon RC4391) connected in a practical step-down converter configuration. Figure 1-14 shows the corresponding waveforms.
To-Vs
♦
R4 Paru Litt R1 = R2 = CX = Lx =-5.0V Output 300kn 75kn ISOpF
-15V Output 900kn 75kn 150pF 1.0mHDaleTE3Q4TA -VOUT = (125V) (Q) 'Caution: Use current limiting protection circuit for high values of CF Figure 1-10. Practical inverting converter (Raytheon Linear Integrated Circuits, 1989, p. 9-54)
C eg ?
1 I
σα Q ω©
Φ
ΤΊΓΊΙ
Φ
©
i_n_r i_r
(Internal) } Osc 'LOAD . . ..— OmA11 LT~ LJJrr»)
VBAT , VQUL
VBEQI
- — IMAX ,
^ O m A
J
- — ' M A X \OmA )
V-OJ-OJ- s u r u η.Ξ
+Vs-Vsw 1 Ground J VLX -VOUT -V0 )Figure 1-11. Inverter waveforms (Raytheon Linear Integrated Circuits, 1989, p. 9-55)
+ D *
-ΟΛ
>V0UT
Figure 1-12. Theoretical buck or step-down configuration {Raytheon Linear Integrated Circuits, 1989, p. 9-55)
The following paragraphs describe how both theoretical and practical func
tions are performed and how the practical functions (Fig. 1-13) relate to the theoret
ical (Fig. 1-12). Notice that switch S in the theoretical circuit is replaced by a transistor (within the IC) in the practical IC circuit. Capacitor C, diode D, and induc
tor L in Fig. 1-12 are replaced by Cp Dp and Lx in Fig. 1-13. Also notice that the ground lead of the IC (pin 4) is not connected to circuit ground. Instead, pin 4 is tied to the output voltage. Using this rearrangement of the feedback system, it is possible
to regulate a nonnegative output voltage (as the feedback system senses voltages more negative than the ground lead).
As shown in Fig. 1-12, when switch S is closed, current flows from the bat
tery, through the inductor L, and through the load resistance to ground. After S is opened, stored energy in L causes current to keep flowing through the load. The cir
cuit is completed by catch diode D. Because current flows to the load during charge and discharge, the average load current is greater than in an inverting circuit. The significance is that for equal load currents, the step-down circuit requires less peak inductor current than an inverting circuit. As a result, the inductor of a step-down circuit can be smaller than for inverting, and the switch transistor in a step-down IC is not stressed as heavily for equal load currents.
In the practical circuit of Fig. 1-13, output filter capacitor CF is discharged, so that the ground lead (pin 4) potential starts at 0 V. The reference voltage is forced to +1.25 V above the ground lead, pulling the feedback input (pin 8) more positive than the ground lead. This positive voltage forces the control network to start puls
ing the output transistor.
As the switching action pumps up the output voltage, the ground lead rises with the output until the voltage on the ground lead is equal to the feedback voltage.
At that point, the control network reduces the on-time of the switch to maintain a constant output.
Λ D1 1N914
Important Note This circuit must have a minimum load > 1 mA always connected
Figure 1-13. Practical step-down converter configuration (Raytheon Linear Integrated Circuits, 1989, p. 9-56)
U~LT 1_Γ
(Internal) OscO OmA 'LOAD
1ΓΊΓΊΓ" IT" U~ 1_ΓΙ~Ι
VBEQIVBAT VQ U T - V T A —, M A X '
TO\ (
•LXVQ(-07V) ]
L-fLWLr- n ^ U ΠΞ
Figure 1-14. Step-down converter waveforms (Raytheon Linear Integrated Circuits, 1989, p. 9-57)
1.5.5 Flyback
Flyback switching regulators and voltage converters are based on a two-cycle energy transfer. First, energy is stored in an inductor. Second, the energy is trans
ferred to a load capacitor. Although some flyback converters use only a simple in
ductor, a transformer is generally more common. Figures 1-15 and 1-16 show the theoretical flyback configuration, using both the inductor and transformer, respec
tively. Figure 1-17 shows a typical IC switching regulator (the Raytheon RC4292) connected in a practical flyback configuration.
The following paragraphs describe how both theoretical and practical func
tions are performed, and how the practical functions (Fig. 1-17) relate to the theo
retical (Fig. 1-16). Notice that switch Sj in the theoretical circuit is replaced by external transistor Mj in the practical IC circuit. Also, Mj is connected to ground, not to the negative-supply input. As a result, a simple inductor as shown in Fig.
1-15 cannot be used in a practical circuit to supply a positive output. By replacing Lj in Fig. 1-15 with a transformer as shown in Figs. 1-16 and 1-17, a positive out
put (VOUT) can be produced with a negative input. Thus, the flyback configuration can be used as an inverter (Section 1.5.3).
Transformers can be operated in two input-to-output modes. When the input- current flow and the output-current flow alternates, one preceding the other, the func
tion is a true flyback operation. If the input and output current flow occur at the same time, the function is called a feed-forward, or simply a forward converter, operation.
GNO (IN) VOLTAGE FEEDBACK
VFB SWITCH CONTROL -V|N
D- Ί o
S1|*#— I,N
Ï "*t
VL
T I
- α *νουτ
C1 (LOAD CAP) - Q GND
IOIOOE
S 1 CLOSED OPEN
VOUT- VL 0 - - V , N -
•L 0
•IN 0
IOIOOE O
Figure 1-15. Theoretical flyback converter with inductor (Raytheon Linear Integrated Circuits, 1989, p. 9-40)
As shown in the simple flyback converter (inductor) diagram of Fig. 1-15, when switch S} is closed, charging current from the battery flows through inductor Lj, which builds up a magnetic field that increases as Sj is held closed. When Sj is opened, the magnetic field collapses, and energy stored in the magnetic field is converted into a current that flows through L{ in the same direction as the charging current. Because there is no path for this current to flow through S,, the current must flow through diode Όχ to charge capacitor Cv The key to inversion in a fly
back circuit is the ability of the inductor to become a source when the charging current is removed. (This is the same as with an inverting or buck-boost configura
tion.)
During discharge, the current in inductor Lj decreases. When the current reaches zero, diode Dj stops conducting. Notice that the rate of change (with time) of the current in an inductor is proportional to the voltage across the inductor and in
versely proportional to the inductance. Also, the load voltage and/or current can be regulated by controlling the on-time of switch S,. As in the case of other regulators, the load capacitor C, stores the energy until the latter is used by the load.
The circuit for the transformer flyback configuration (Fig. 1-16) is similar to that of the simple flyback (inductor) configuration (although the waveforms are sub-
C1 VOUT GNDIN
Figure 1-16· Theoretical flyback converter with transformer (Raytheon Linear Inte
grated Circuits, 1989, p. 9-41)
stantially different). In effect, flyback transformer T} stores energy with one winding and removes energy with the other winding. The first cycle starts with the closing of S j. This pulls VIN up to ground. Current starts from zero and ramps up in the Nj wind
ing. This stores energy in the magnetic flux of the transformer core. After a controlled time, switch St opens, and energy is transferred from the core to the N2 secondary, and then to the output.
The circuit for the practical IC flyback regulator in Fig. 1-17 is similar to those of other switching regulators in that the IC contains an oscillator, comparator, error amplifier, reference, and control logic. However, the circuits external to the IC are somewhat more complex (in this particular configuration). The following is a brief discussion of the IC and external-component functions.
The oscillator at pin 1 of the IC generates a time base for the drive pulse at pin 6. Oscillator frequency is set by an external capacitor CX connected to pin 1. The error amplifier compares the feedback and reference signals at pins 2 and 3 and pro-
Figure 1-1 7. Practical nyoack converter (Raytheon Linear Integrated Circuits, 1989, p. 9-43)
duces an amplified error signal proportional to the input difference. The current comparator compares the error-amplifier output to a signal that is proportional to the current in the transformer (measured by the voltage across R4).
If the feedback signal at pin 7 is greater than the error signal, the control logic (an FF and output driver) turns the external transistor Mx off. The control logic uses an FF to make sure that Mj receives only one pulse for each oscillator cycle. The output driver amplifies the FF output to provide a fast switching signal to Mp The voltage reference at pin 4 provides - 5 V for power to the IC components, as well as a reference for the error amplifier. The shunt regulator at pin 5 acts like a zener to clamp the IC, thus preregulating the supply within safe limits.
When power is first applied, the error amplifier senses that the output voltage is lower than required and sends an error signal through the current comparator to the control logic. In turn, the control logic pulses Mj to increase the output voltage.
When the output voltage reaches the desired value, the control logic changes the Mj drive so that the transformer current is maintained at a constant level.
The ratio of Rj and R2 determines the value of VOUT. Typically, the equivalent resistance of this combination should be in the 25- to 100-k range. This minimizes input-bias current and input-noise errors. (The manufacturer recommends an RN55 metal film resistor for Rx and R2.)
The value of R3 sets the shunt-regulator (pin 6) current. The value of R4 sets the maximum switch current through Mr Resistor R5 holds Mj off during start-up or any time the IC is inactive. Resistor R6 provides for a signal loss in the gate drive to Mp thus preventing possible oscillation. Resistor R? cancels input-bias current er
rors at the error amplifier inputs.
Resistor Rg, capacitor C3, and diode D{ form a "scrubber" network that damp
ens ringing on the Mj drain and T, primary, thus reducing voltage spikes that might potentially damage Mr The R8/C3 combination can be omitted, depending on the type of Mp the supply voltage, and the Tj characteristics. Capacitor C2 acts as a fil
ter for the feedback signal. Capacitors C4 and C5 filter the shunt regulator (pin 5) voltage. If the shunt current goes too low to supply the IC properly, the output will alternately shut down and turn on at a low frequency ("motor boating"), at a fre
quency determined by C5. Capacitor C4 must have a low impedance at high frequen
cies to filter switching noise.
Capacitors C6 and C7 filter the output voltage. Diode D2 rectifies the output voltage. The basic limitation of the output load power that can be extracted from the circuit is determined by the gate-to-drain capacitance of Mr Although specifically designed to drive capacitive loads, the drive output at pin 6 will not switch large field-effect transistors (FETs), where drain current exceeds about 10 A. The maxi
mum Mj size is also affected by the ratio of input to output voltage. The manufac
turer recommends an International Rectifier IRF9633 or Motorola MTP5P18 for M,.
A disadvantage of any flyback converter or regulator is the high amount of en
ergy that must be stored in the transformer in the form of DC current in the wind
ings. This requires larger cores than would be necessary with pure AC in the windings. This problem can be overcome by means of a forward converter, dis
cussed next.
« - T
- * - - = : vo u l
— 1 — ► -
Figure 1-18. Forward converter (Linear Technology, Linear Applications Handbook, 1990, p. AN 19-16)
1.5.6 Forward
A forward converter such as shown in Fig. 1-18 avoids the problem of large amounts of stored energy in the transformer core. However, the forward converter does so at the expense of an extra winding on the transformer, two more diodes, and an additional output filter inductor.
Power is transferred from input to load through Όχ during the switch-on time.
When the switch turns off, Dj is reverse-biased, and the Lj current flows through D2. The additional winding and D3 are required to define switch voltage during switch-off time. Without this clamp, switch voltage would jump all the way to breakdown at the moment the switch is opened (because of the magnetizing current flowing in the transformer primary).
The additional, or "reset," winding usually has a 1:1 turns ratio to the primary.
This limits switch-duty cycle to 50% maximum. Above 50%, switch current rises uncontrolled (even with no load) because the primary winding cannot maintain zero DC voltage. Reducing the number of turns on the reset winding allows higher switch-duty cycles at the expense of higher switch voltage.
Output voltage ripple of forward converters tends to be low because of Lp but input ripple current is high because of the lower-duty cycles normally used. A smaller core can be used for Tj (compared to flyback) because there is no net DC current to saturate the core.
1.5.7 Current-Boosted Boost
The circuit shown in Fig. 1-19 is an extension of the standard boost converter (Section 1.5.1). A tapped inductor is used to decrease the switch current for a given
VQUT
Figure 1-19. Current-boosted boost converter (Linear Technology, Linear Applications Handbook, 1990, p. AN 19-16)
ot I 1 Ti 1 *S 1 1 t
X 1
1 <
P I —
•
1 1
Figure 1 -20. Current-boosted buck converter (Linear Technology, Linear Applications Handbook, 1990, p. AN 19-17)
load current. This allows higher load currents at the expense of higher switch volt
age. Significant increases in power are possible when the input-output differential is low. However, care must be used to ensure that the maximum switch voltage is not exceeded.
7.5.8 Current-Boosted Buck
The current-boosted buck converter in Fig. 1-20 uses a transformer to in
crease output current above the maximum current rating of the switch (which is a transistor in a practical circuit). The current-boosted circuit does so at the expense of increased switch voltage during switch-off time. The increase in maximum output current over a standard buck converter (Section 1.5.4) is equal to input voltage di
vided by output voltage, plus turns ratio times the input-output differential. For ex
ample, in a 15-V to 5-V current-boosted buck converter, with a 1:4 turns ratio, the increase in output current is double: 15/(5 + 1/4 x 15 - 5), or 2. This is a 100% in
crease in output current. However, the maximum switch voltage for a current-boost buck is increased from input voltage to input voltage plus output voltage divided by the turns ratio. Using the same 15- to 5-V converter, the maximum switch voltage is 15 + 5/turns ratio, or 15 + 5/0.25 = 35.
1.5.9 Cuk
The Cuk converter in Fig. 1-21 (named after Slobodan Cuk, a professor at Cal Tech) is similar to a buck-boost or inverter in that input and output polarities are re
versed. However, the Cuk configuration has the advantage of low ripple current at
Figure 1-21. Cuk converter (Linear Technology, Linear Applications Handbook, 1990, P.AN19-15)
both input and output. The need for two separate inductors can be eliminated by winding both on the same core, with an exact 1:1 turns ratio.
With slight adjustments to Lj or L2, either input ripple current, or output ripple current (but not both) can be forced to zero. This eases the requirements on size and quality of input and output capacitors without requiring further filtering. Notice that the ripple current in C2 is equal to the output current, so C2 must be large. However, C2 can be electrolytic, so physical size is not normally a problem.
Heat Sinks for Switching Power Supplies
This chapter is devoted to temperature-related design problems. It is often as
sumed that switching regulators do not require heat sinks. Although this is true in many cases, it may be necessary to use heat sinks for high-current switching regula
tors. As discussed, switching supplies contain at least one element (the shunt or se
ries transistor) that must pass load current. Such transistors can be part of the IC (such as Qj in Fig. 1-6) or external to the IC (where an IC operates as a control ele
ment, such as Mj in Fig. 1-17). Either way, if the power dissipated by the element exceeds about 1 W, the elements must be provided with a heat sink. Before we get into the details, let us review some basic temperature-related design problems for switching supplies (and most other electronic equipment).
2.1 Temperature-Related Design Problems
There are two basic temperature-related problems in simplified switching- supply design. First, although data sheets specify component parameters (for tran
sistors, diodes, rectifiers, ICs, etc.) at a given temperature, many of these parameters change with temperature. Because components rarely operate at the exact tempera
ture shown on the data sheet, it is important to know the parameters at the actual op
erating temperature.
For example, in the case of transistors, three critical parameters (from a switching-supply design standpoint) are current gain, collector leakage, and power dissipation. In addition, changes in parameters can affect transistor temperature (an increase in current gain or power dissipation results in a temperature increase).
Second, on top of knowing the effect of temperature on parameters, it is im
portant to know how heat sinks or component mounting can be used to offset the ef
fects of temperature. For example, if a transistor is used with a heat sink or is mounted on a metal surface that acts as a heat sink, an increase in temperature (from any cause) can be dissipated into the surrounding air.
25
The following paragraphs describe methods for approximating important parameters at temperatures other than those shown on data sheets. Methods for de
termining the proper dissipation characteristics are also discussed. (Temperature- related design problems for ICs, such as switching-regulator ICs, are discussed in Section 2.3.1.)
2.1.1 Effects of Temperature on Transistor Collector Leakage
Collector leakage (Icbo) increases with temperature. As a guideline, collector leakage doubles with every 10°C increase in temperature for germanium transistors and doubles with every 15°C increase for silicon transistors.
Collector leakage also increases with voltage applied at the collector. For ex
ample, a typical data-sheet leakage figure could be 2 μΑ at 25°C and 50 μΑ at 150°C. However, the 25°C figure is with a collector-base voltage of 30 V, whereas the 150°C figure is with 5 V. If the temperature is raised from 25 to 150°C with 30 V at the collector, the collector leakage is about 500 μΑ. As a result, always con
sider the possible effects of a different collector voltage when approximating collec
tor leakage at temperatures other than those on the data sheet.
2.1.2 Effects of Temperature on Transistor Current Gain
Current gain (hfe) increases with temperature. As a guideline, current gain doubles when the temperature is raised from 25 to 100°C for germanium transistors and doubles when the temperature is raised from 25 to 175°C for silicon transistors.
It is obvious that silicon transistors are less temperature sensitive than germa
nium transistors. Data sheets usually specify a maximum operating temperature, or ambient temperature. If this temperature is not given or is unknown, the guideline is not to exceed 100°C for germanium transistors or 200°C for silicon transistors.
2.1.3 Effects of Temperature on Power Dissipation
The power-dissipation capabilities of a transistor (and diode, rectifier, or IC) must be carefully considered when designing any supply circuit. Of course, in small-signal circuits (not the series or shunt transistor!), the power dissipation is usually less than 1 W, and heat sinks are not needed. In such circuit, the only con
cern is that the rated power dissipation (as shown on the data sheet) not be exceeded.
Here is a guideline: Do not exceed 90% of the maximum data-sheet power dissipation for small-signal circuits. This includes most of the circuits in switching supplies, except the series/shunt-pass element (transistor or IC), any rectifier diodes, and possibly some zeners.
As with other characteristics, manufacturers specify maximum power dissipa
tion in a variety of ways on data sheets. Some manufacturers provide safe-operating- area curves or graphs for temperature and/or power dissipation. Other manufacturers
specify maximum power dissipation, in relation to a given ambient temperature or a given case temperature. Still others specify a maximum junction temperature, or max
imum case temperature.
2.1.4 Thermal Resistance
Transistors, rectifiers, and ICs designed for power-supply applications (or any power application) usually have some form of thermal resistance specified to show the power-dissipation capabilities. Thermal resistance can be defined as the increase in temperature of the component junction (with respect to some reference) divided by the power dissipated (or °C/W).
Power-transistor data sheets usually specify thermal resistance at a given tem
perature. This is also the case for many diodes and rectifiers. For each increase in temperature from the specified value, there is a change in temperature-dependent characteristics of the component.
Because there is a change in temperature with changes in power dissipation of the component, the junction-to-ambient-air temperature also changes, resulting in a characteristic change. Consequently, the component characteristics can change with ambient-temperature changes and with changes produced by variation in power dis
sipation.
In power transistors, rectifiers, and ICs, thermal resistance is normally mea
sured from the component junction to the case. This results in the term 0JC, where the lowercase Greek letter theta (Θ) indicates thermal resistance (because engineers like Greek letters!).
On those components where the case is bolted directly to the mounting surface with a built-in threaded bolt or stud, the term ΘΜΒ (thermal resistance to mounting base) or 9MF (thermal resistance to mounting flange) is used. These terms take into account only the thermal paths from junction to case (or mount). For power compo
nents in which the junction is mounted directly on a header or pedestal, the total in
ternal thermal resistance from junction to case (or mount) varies from about 50 to less than 1°C/W.
2.1.5 Thermal Runaway
The main problem in operating a transistor or IC near the maximum power limits is a condition known as thermal runaway. Although thermal runaway can apply to diodes and rectifiers, they are generally not as affected by runaway because they do not have current gain.
When current passes through a transistor junction, heat is generated. If all the heat is not dissipated by the case (an impossibility), the junction temperature rises.
In turn, this causes more current to flow through the junction, even though the volt
age, circuit values, and so on remain the same. The increased current flow causes the junction temperature to increase even further, with a corresponding increase in current flow. If the heat is not dissipated by some means, the transistor burns out and is destroyed.
2.7.6 Operating without Heat Sinks
If a transistor or IC is not mounted on a heat sink, the thermal resistance from case to ambient air, 0CA, is so large in relation to that from junction to case (or mount) that the total thermal resistance from junction to ambient air, 0JA, is primar
ily the result of the GCA term.
Table 2-1 lists case-to-ambient thermal resistances for a few common transis- tor/IC cases (both old and new). As shown, heavy-duty cases such as TO-3 have a small temperature increase (for a given wattage) in comparison to such cases as the TO-5 (because heavy-duty cases dissipate heat into the ambient air). Notice that the values shown in Table 2-1 are for 25 °C and must be derated for ambient tempera
tures above 25°C.
The information in Table 2-1 can be used to approximate the maximum power dissipation of transistors (without heat sinks) when such information is not shown on the data sheet. For example, assume that a silicon transistor with a TO-3 case is used and that the absolute maximum power dissipation (without a heat sink) must be found.
The case-to-ambient thermal resistance for a TO-3 case is 30. As discussed in Section 2.1.2, silicon transistors should not be operated above 200°C (under any cir
cumstances). Assuming a 25°C ambient temperature, the transistor temperature should not be allowed to increase more than 175°C maximum (200 - 25 = 175).
With a factor of 30 for the TO-3 case and a 175°C increase, the case must dissipate 5.83 W (175/30 = 5.83).
However, as discussed in Section 2.1.2, silicon-transistor current gain doubles when the temperature is raised from 25 to 200°C. Assuming that the voltage remains constant, the maximum dissipation allowable is then cut in half to 2.69 W. This is an absolute maximum figure, assuming a silicon transistor, TO-3 case, and an ambient temperature of 25°C. (For simplified design purposes, the 2.69-W figure is usually safe if the case is mounted on a metal surface that can act as a heat sink.)
Table 2-1 · Thermal resistance for common transistor/IC cases Case
TO-3 TO-5 TO-8 TO-18 TO-36 TO-39 TO-46 TO-60 TO-66
eCA(°C/W) 30 150 75 300 25 150 300 70 60
2.1.7 Operating with Heat Sinks
After about 1 or 2 W, it becomes impractical to increase the size of the case to make the case-to-ambient thermal-resistance factor comparable to the junction-to- case factor. However, there are IC regulators with special TO-66 cases that can dis
sipate up to about 3 W. Except for these special circumstances, most power transistors, rectifiers, and ICs are designed for use with an external heat sink. Some
times the mounting area serves as the heat sink. In other cases, a heat sink is at
tached to the case. Either way, the primary purpose of the heat sink is to increase the effective heat-dissipation area of the case and to provide a low heat-resistance path from case to ambient.
Section 2.2 discusses the practical aspects of heat-sink design and selection.
The following paragraphs describe the basic calculations involved.
To properly design (or select) a heat sink for a given application, the thermal resistance of both the component and heat sink must be known. For this reason, some power-transistor/IC data sheets specify the 0JA that must be combined with the heat-sink thermal resistance to find the total power-dissipation capability.
Notice that some power-transistor/IC data sheets specify a maximum case temperature rather than ΘΜ. As discussed in Section 2.1.11, maximum case temper
ature can be combined with heat-sink thermal resistance to find maximum power dissipation.
2.1.8 Heat-Sink Ratings
Commercial heat sinks are available for various case sizes and shapes (see Section 2.2). Such heat sinks are especially useful when the components are mounted such that there is little or no thermal conduction (heat transfer) to the PC board.
Commercial heat sinks are rated by the manufacturers in terms of thermal re
sistance, usually in °C/W. When heat sinks involve the use of washers, the °C/W factor usually includes the thermal resistance between the case and sink, 9CA. With a washer, only the sink-to-ambient-air, 0SA, thermal-resistance factor is given. Either way, the thermal-resistance factor represents temperature increase (in °C) divided by wattage dissipated.
For example, if the heat-sink temperature rises from 25 to 100°C when 25 W is dissipated, the thermal resistance is 75/25 = 3. This is listed on the data sheet as 0SA, or simply 3°C/W.
All other factors being equal, the heat sink with the lowest thermal resistance (°C/W) is best. For example, a heat sink with 1°C/W is better than one with 3°C/W.
Of course, the heat sink must fit the case (and space around the case). Except for these factors, selecting a suitable heat sink should be no particular problem.
Table 2-2 is a brief selection guide to heat-sink manufacturers. No attempt has been made to provide a complete list (which is constantly subject to change).
Likewise, the list covers only the basic types of cases that are likely to require heat sinks (TO-66, TO-99, DIP).
Table 2 - 2 . Commercial heat-sink selection guide (Raytheon Linear Integrated Circuits, 1989, p. 9-74-75)
et./(0c/w)
0.31-1.0 I 1.0-3.0
3.0 - 5.0
I 5.0 - 7.0
I 7.0-10.0
I 10.0-25.0
I 12.0-20.0
I 20.0-30.0
30.0-50.0
20 30 32 34 45 60
Manufacturer/Series or Part Number j
TO-66 Package I Thermalloy — 6441. 6443. 6450. 6470. 6560. 6590. 6660, 6690
Wakefield — 641 I Thermalloy — 6123. 6135. 6169. 6306. 6401, 6403, 6421. 6423.6427. 6442. 6463.
6500
Wakefield —621,623
Thermalloy — 6606, 6129. 6141, 6303 IERC — HP
Staver — V3-3-2
Wakefield — 690 I Thermalloy — 6002, 6003, 6004, 6005. 6052. 6053. 6054, 6176. 6301
IERC — LB Slaver— V3-5-2
Wakefield —672 I Thermalloy — 6001. 6016. 6051. 6105. 6601
IERC — LA, uP
Staver — V1-3. V1-5, V3-3. V3-5, V3-7
Thermalloy — 6-13. 6014. 6015. 6103. 6104, 6105, 6117 I TO-99 Package
Wakefield — 260 I Thermalloy—1101,1103
Staver — V3A-5
Wakefield — 209 I Thermalloy—1116. 1121. 1123.1130,1131. 1132.2227.3005
IERC — LP
| Staver — F5-5 Wakefield —207
Thermalloy — 2212, 2215. 225. 2228. 2259. 2263, 2264 Staver — F5-5. F6-5
Dual-Inline Package Thermalloy — 6007
Thermalloy —6010 Thermalloy — 6011 Thermalloy —6012 IERC — LIC I Wakefield —650,651
* All values are typical as given by manufacturer or as determined from characteristic curves supplied by manufacturer.
Staver Co.. Inc.: 41-51 N Saxon Ave.. Bay Shore. NY 11706
