FERROMAGNETIC-CORE DESIGN

AND APPLICATION HANDBOOK

•
FERROMAGNETIC-CORE
DESIGN
AND APPLICATION
HANDBOOK

M.F. " Doug " DeMaw

Member, IEEE, Graduate Engineer

Prentice-Hall, Inc.. Englewood Cliffs, NewJersey07632

Library of Congress Cataloging in Publication Data
DEMAIY, DOUG
Fcrromagnclk·corc design 2nd appliauion
hondbook .
Bibliogruph) : p.
Includes inde~.
1. Magnctk: cores. 2. Magnetic dc..,ices.
I. Ti1le.
TK n72.M25D4S 62 1.34 80· 16136
ISBN Q. 13·314088·1

Editori:il/ production supervision and interior design: Nancy Moskowitz

Manufocturing buyer: Joyce Le vatino

© 1981 by Prentice-Hall, Inc., Englewood Cliffs, N.J . 07632

All rights reserved. No pa rt of this book

may be reproduced in any form or
by any means wit hout permission in writing
from the publisher.

Printed in the United States of America

10 9 ~ 7 6 5 4 J

PR EN1 lCt-HALL INTER:'IA TlONAL. I NC.• London

PRENTIC!i-HALLOFAUSTRALIA PTY. LlMIT1'D. Syd11ey
PRENTtCE-H ;\LLOFCANADA. LTD.. Tom1110
PR l:NTI Ch- H ALLOF I NDIA PRI V,\TE LI MITl:L>. New Delhi
PR ENT lCl:- H ALL OF JAPAN. I NC.. Tokyo
PR l:NTI CE- H ALLO~· soun1 E;\ ST ASIA PTE L TD .. Si11gapor£'
WH ITWALL B OOKS Ll :VllTED. Wel/i11gt011, New Zealand
CONTENTS

PREFACE ix

CHAPTER 1 THE BASICS OF MAGNETIC MATERIALS

1.1 Classifying M agnetic Materials 1
1.2 Physical and Electrical Traits 2
1.2. I Powdered irons 5
1.2.2 Ferrites 6
1.3 Calculating the Coil Turns 9
1.3. I Determining inductance II
1.3.2 Cirwit Q 12
1.4 Power Capability 14
1.4. I Flux densily versus ue 15
1.4.2 Sample calculation 18
1.4.3 Varia1ions in the equations 21
1.5 Volt- Ampere Ratings 22
1.5.1 A sample K1 core selection 23
1.5.2 Power rating 30
1.6 Inductance versus AC/ DC Excitation 30
1.6. I Employment of the H anna curve 31
1.6.2 C 011st ructing a Hanna curve 33
1.6.3 A ctual air gap 36

CHAPTER 2 APPLICATION OF RODS, BARS, AND SLUGS 39

2. 1 Rods and Bars 39
2.1 . I Characteristics of rods and bars 40
2. 1.2 Q and L versus coil placement 42
2.1.3 Frequency versus core material 48

v
vi Contents

2 .2 Rods and Bars as Antennas 49

2.2.l Loop patlern symmetry 52
2.2.2 Noise immunity 53
2.3 Practical Loop Circuits 54
2.3. l Loop-to-mixer circuits 56
2.3.2 Loop to RF amplifier 57
2.3.3 Loop with sense antenna 59
2.3.4 Loop variations 6 1
2.4 Rod- Core Transformers 62
2.5 Slug Applications 63
2.5. I Performance considerations 65
2.5.2 Practical circuits 68
2.6 RF Chokes 73
2.6. I Choke applications 73

CHAPTER 3 APPLYING TOROIDAL CORES 79

3.1 Low- Level Circuits 79
3. I. I RF amplifiers 80
3.1.2 Capacitive-divider marching 83
3.1.3 Interstage coupling 85
3.1.4 Toroids in filters 87
3. 1.5 Toroids in instruments 93
3.1 .6 Other low-level applications 96
3.2 Broadband Transform ers 99
3.2.l Bifilar-wound transformers JOO
3.2.2 Quadrifilar transformers IOI
3.2.3 Broadband conrentional transformers 102
3.2.4 Variable impedance matching 109
3.3 Antenna Loading 11 2
3.4 Toroidal Distributed Capacitance 11 3
3.5 Power Supplies 114
3.5. I Using poll'er FETs as swi1ches 116
3 .6 Core Doping 11 7
3.6. I Simple doping methods 118
3.6.2 Potting technique 120

CHAPTER 4 BEADS. SLEEVES. AND POT CORES 123

4. 1 Properties of Beads 124
4.1. l Shield-bead allenuation 126
4.1.2 Practical applications for beads 130
4.1.3 Beadsas Qkillers 133
4.1.4 Beads fo r decoupling 135
4.1.5 Reducing incidental radit11ion 136
4.2 Ferrite Sleeves 137
4.3 Ferrite Balun Cores 141
 Contents vii

4.4 Pot Cores 143

4.4. l Pol-core hardware 144
4.4.2 Design co11sideratio11s 146
4.4.3 POI-core designs 153
4.4.4 Tone encoder with pot core 154
4.4.5 Pot cores in filters 157
CHAPTER 5 PERMANENT- MAGNET DATA 159
5.1 The Nature of Permanent-Magnet Materials 159
5.2 Elementary Permanent-Magnet Relationships 165
5.2.1 The B-H curve 165
5.2.2 Recoil loops 166
5.2.3 Working point of a magnet (static
applications) 167
5.2.4 Th e magnetic circuit 167
5.2.5 Th e permanent magnet in an external
field 169
5.3 Magnetization 170
5.3. I Conditions for comp/ere sa111ra1ion 171
5.3.2 Magnetizing equipme111 172
5.4 Demagnetization 176
5.4. l De111agnetiza1ion techniques 176
5.5 Stability and Stabilization 177
5.5.1 M eta/lurgica/ changes 178
5.5.2 Time 178
5.5.3 Temperature 178
5.5.4 Reluctance changes 180
5.5.5 A cfrerse fields 180
5 .5.6 Shock. stresss. and vibrarion 180
5.6 Measurements and Calibration 180
5.7 Handling of Permanent Magnet s 18 1
5. 7.1 U11111agne1ized magners 182
5.8 Glossary of Terms 182

APPENDIX A REFERENCES 189

APPENDIX B MISCELLANEOUS DATA 195
APPENDIX C POT-CORE AND BOBBIN DATA 207
APPENDIX D CYLINDRICAL-CORE ANO COIL DATA 215
APPENDIX E TORROID-CORE DATA 229
INDEX 251
PREFACE

The work within these chapters is dedicated expressly to engineers, techni-

cians, and college students who arc presently or soon to be involved profes-
sionally with electronics. Nearly a ll modern circuits contain magnetic-core
devices of one kind or another. It is not sufficient to the cause of good engineer-
ing practices Lo use inductors and transformers with magnetic cores in a casual
manner. The functional parameters o f toroids, rods, and pol cores are as im-
portant in a composite circuit as a re the operating characteristics of the active
d evices bei ng used. Therefore, more than a basic understanding of ferri tes and
powdered-iron components is essential today during routine design work .
Among the primary considerations when dealing with fcrromagnetics are
proper core selection versus operating frequency, circuit Q, power-handling
capability, and physical mass. T he wrong core material, or a misapplied core
material, can render unusable an otherwise perfect ci rcuit.
Emphasis has been placed on the practical aspects of magnetic core
materials from low frequencies through UHF. The R F engineer or technician
will find this volume invaluable in his or her daily efforts. The student will
value this publication as an important textb ook, and later as a standard refer-
ence for which nothing similar exists. Tedious mathematical procedures have
been omitted in an effort to make comprehension more rapid and enjoyable.
Equations have been used on ly where they a re necessary to illustrate a concept
or to provide a design example .
The chapters include basic theory and practical circuit examples in which
toroids, rods, slugs, and pot cores are used. All of the circuits are proven ones,
based on laboratory research and development by the author. The book con-
tains myriad examples of narrow-band and broadband transformers and in-
ducto rs. Ferrite loop antennas, slug-tuned inductors, RF chokes, and fer rite

ix
Preface x

beads are discussed in depth , also. Magnetic-core devices are highl ighted in
filters, switching types o f power supplies, and impedance-matching net works.
There are live appendices, which contain comprehensive design data, lists
of core manufacturers, a rather lengthy bibliography, design nomographs,
conversion tables, and numerous pages of componen t numbers and
characteristics from a variety of leadi ng man ufacturers.
No credible tech nical author can take credi t for all of the design ap-
proaches or circuits he or she commits to a text. Alm ost without exception the
author treads unknowingly on the past work of someone else, however un-
intentional. Uncredited similarity between any portion of this work and that
of others is uninten tional and without prior knowledge on beha lf of the
author. Credits have been included wherever applicable.
The author acknowledges the gracious goodwill and assistance of the
many manu factu rers of magnetic cores discussed in th is volume. The plethora
of application notes and related reference data which they so readily furni shed
upon request contributed vastly to the relative completeness of the book. A
strong vote of appreciation goes to David Boelio of Prentice-Hall, Inc. He
offered more encouragement and assistance than this writer has ever experi-
enced before. Finally, the au thor wishes to express his profound appreciation
to Jean , his wife, for her encouragement and understanding while the book
was being written.
M . F. " Doug" DeMaw
FERROMAGNETIC -CORE DESIGN
AND APPLICATION HANDBOOK
1
THE BASICS
OF MAGNETIC MATERIALS

In this chapter we discuss the fundamental concepts of magnetic

materials and their significant characteristics and present a generalized
treatment of their applications in electronics for communications engineers
and technicians. The student will find this chapter a valuable base for the
tutorial and practical information that is contained in subsequent sections
of this volume.
One might ask: How do magnetic materials fit into the broad picture of
communications electronics? The answer is simple: There are few ac circuits
that do not in some manner require the use of iron, powdered iron, or fer-
rites. The more common applications include power-supply transformers,
audio transformers, ac and RF filter inductors, broadband RF transformers,
narrow-band RF transformers, and damping networks.
This book treats magnetic materials that are used as core material for
the various types of electronic components just mentioned. A magnetic core
is a configuration of magnetic material (rod, toroidal, pot core, and other
physical forms), which, in its intended application, has a specific physical
relationship to current-carrying conductors and whose magnetic characteris-
tics are essential to its use.

1.1 Classifying Magnetic Materials

The most significant effect resulting from the insertion of ferrous-core

material into an air-core inductor, or one that has been wound on a dielec-
tric form, is a marked increase in inductance. T he core substance used for
increasing the inductance of a given coil might be classified as "hard" or

1
2 The Basics of Magnetic Materials Chap. 1

"soft" magnetic material. These terms do not refer to the physical proper-
ties of the core substance. Rather, they relate to the electrical characteristics
of the chosen material. More specifically, hard compounds have the ability
to retain their magnetism after a magnetizing force has been applied to
them. One application for hard-core material is found in the magnets of PM
loudspeakers. The soft materials, conversely, lose their magnetism im-
mediately upon removal of the magnetizing force. Typical uses for soft
types of cores are in radio-frequency and audio-frequency transformers and
reactors. This book is centered on soft-core materials. Under no cir-
cumstances should the reader assume that the soft substances discussed are
lacking in hardness to the physical touch: a soft magnetic material can be
extremely brittle and durable, but soft in terms of its electrical properties.

1.2 Physical and Electrical Traits

The most common physical formats for magnetic-core materials are rods;
flat bars; toroids; I-E and U cores; and cup or pot cores. Figure 1-1 il-
lustrates some of the shapes under discussion . The physical form used by
the designer is dictated largely by the power level of the circuit in which the
core will be used, the physical dimensions of the assembled transformer or
inductor, and the fabrication time required when using one style of core as
opposed to another type: Some cores lend themselves more readily to the
fast production of a fini shed component than is possible with other core
formats. Generally, magnetic core assemblies that use insulating bobbins to
contain the coil or transformer windings are preferred for high-volume pro-
duction. Pot cores, U cores, and l-E cores fit the foregoing description.
Although toroidal and solenoidal magnetic cores can be wound with con-

Figure 1-1 Various types of magnetic-core materials.

Sec. 1-2 Physical and Electrical Traits 3

siderable rapidity by means of machines, the process is somewhat more

complicated than when dealing with plastic bobbins during mass produc-
tion.
The various magnetic cores are available in a variety of sizes and com-
pound mixes. Laminated cores are fashioned from various grades of steel
and can be obtained in assorted widths and thicknesses. Extremely long,
narrow, thin strips of steel are wound circularly and used in tape-wound
toroid cores. In such an example the core is contained in a plastic housing to
provide insulation between the coil winding and the core. The housing also
keeps the tape core from unwinding while protecting it from moisture and
chemical damage. The steel tape is "grain-oriented." This means that when
the steel tape is manufactured, the rolling process is carried out to ensure
that the individual grains of silicon steel (crystals) are aligned in the direc-
tion of rolling. The advantage of producing a grain-oriented steel lamina-
tion is that magnetization is much easier to realize in the direction of the
grain orientation. Therefore, maximum permeability of the core will occur
when the related magnetic field is parallel to the grain orientation.
Permeability is shown symbolically by the Greek letter µ (mu) or, more
specifically, B = µH orµ = µ,µo. It is a measure of how much superior a
specified material is than air as a path for magnetic lines of force. Air, the
reference, has aµ of l. Permeability is equal to the magnetic induction (B)
in gauss, divided by the magnetizing force (H) in oersteds.
Tape-wound cores other than those which take toroidal form are used in
a variety of modern circuits. Figure 1-2 shows the toroidal and rectangular
forms under discussion. In each case a length of steel tape is wound con-
tinuously until the core has sufficient cross-sectional area to handle the
amount of power required. Cores of this variety have extremely high
permeability factors. Most toroidal tape-wound cores have a thin layer of
paste-type insulating material between the laminations. The rectangular
cores are treated with an insulating varnish between layers. This type of in-
sulating process is necessary to prevent the formation of transverse eddy
currents. When varnish is employed it also makes the laminated tape core
very rigid and stable. The value of this can be seen by examining Fig. 1-3.

Steel Steel
tape tape

(a) {b)

Figure 1·2 Circular and rectangular cores wound with steel tape. The type
seen in !al is toroidal, with the core in (bl shown before it is cut in half to form
two C cores.
 Coil on
bobbin

la) (b)
Figure 1-3 Tape-wound steel core after being cut (a) and with the bobbin in place (b).
A steel band holds the core halves tightly together, as seen in (b).

Part (a) shows the evolulion of a tape-wound core of Lhc lype seen in Fig.
1-2g. 1n Fig. 1-3a the varnished and baked core is cut in half to form a pair
of grain-oriented U or C cores, as they are commonly called. Figure l-3b
shows the final assembly form of the inductor and core. One leg of each
core half is inserted in the insulating bobbin on which the wire winding has
been formed. A steel band is drawn very tight around the pair of U cores to
maintain a minimum gap where the ends of the halves are butted together;
then the band is Jocked in position. In some instances a metal mounting
plate is affixed to the inductor or transformer assembly at Lhe time the lock-
ing band is added: The band not only holds the core halves together tightly,
but secures the mounting plate to the core material. The manufacturer may
elect to encapsulate the completed unit in epoxy compound, in which case
provision is made for external connection to the induclor or transformer
leads by means of terminals set into the potting compound. Similarly,
mounting screws or nuts can be embedded in the potting material to permit
attaching the completed assembly to the chassis of the equipment in which it
will be used.
The example in Fig. J-3b can be extended to transformer use as well as
to inductors. In some instances two bobbins are used to contain the
transformer windings-a bobbin on each side of the pair of U cores . This
general concept can be extended farther by utilizing two rectangular cores
side by side, with a single bobbin shared by the core legs, which are adjacent
to one another. This effectively increases the cross-sectional area of the core
material for greater power-handling capability. Tape-wound cores are used
primarily in power-supply transformers and reactors for 60- or 400-Hz ser-
vice. They are used also in de-to-de converters and audio circuits. Another
application is in magnetic amplifiers, wherein the circuit performance is
based on the saturable-reactor concept.

4
 1.2.1 Powdered Irons

We have established that the various physical shapes of magnetic-core

materials can be applied to powdered-iron and ferrite cores as well as to
those which are fashioned from steel. The choice between ferrite and
powdered iron when designing a circuit is founded primarily on two basic
considerations: (I) Powdered-iron cores do not saturate easily. Saturation
can be defined as Bs, which is the maximum value of induction at a specified
high value of field strength. This is the point at which further increase in in-
trinsic magnetization with an increase in field strength is minimal. (2) Fer-
rite cores saturate easily, thereby making them more suitable for use in cir-
cuits such as de-to-de converters, magnetic amplifiers, and the like. Ferrite
cores offer still another advantage: The permeability of a given core can be
made much higher with ferrite than is possible with powdered iron. The up-
per range of the latter is on the order of 90µ. With ferrite it is possible to ob-
tain µ factors as great as 5000.
In view of the foregoing, one might surmise that there is little advantage
in using powdered iron at all. Why not use ferrites for all applications in
which magnetic-core material is required? Basically, the designer must ef-
fect a trade-off between high permeability and core temperature stability.
As a general rule, the greater the permeability of a material, the less stable it
will be at higher frequencies. The stability factor of an inductor is of para-
mount significance in critical circuits such as narrow-band filters, narrow-
band tuned transformers, and oscillators. Ambient and RF-induced
temperature changes cause the permeability to change, which in some cases
can have a marked effect on stability. The higher the permeability of the
core material, the more pronounced the effect. Therefore, the powdered-
iron core is often the designer's preference in RF circuits that need to be
relatively immune to the effects of saturation and poor stability. Iron-core
inductors offer high values of Q (quality factor) and stability over a wide
range of flux levels and temperatures.
Powdered-iron cores are composed of finely defined and separated par-
ticles of iron which are insulated from one another by means of the binder
compound that is used in developing a particular recipe, called a mix. The
mix is pressed into one of the various core forms (rods, slugs, and toroids)
and baked at very high temperature. This technique ensures an even
distribution of the powder, thereby aiding the requirement of having a
relatively constant effective permeability. The useful frequency range of a
mix versus the Q of the inductor wound on a core made from that mix is
determined by the iron-particle makeup, the size, and the density . For ex-
ample, extremely fine particles which are distributed thinly and uniformly
within the binder medium permit the production of cores that are suitable

5
6 The Basics of Magnetic Materials Chap. 1

for use at VHF and higher. Table 1-1 lists a popular group of powdered-
iron mixes which are manufactured by Micro-Metals Corp. and sold in
small quantities by Amidon Associates of North Hollywood, California.
The listing is for toroid cores of various permeability factors.

TABLE 1-1 Selected powdered- iron mixes.

Temperature
Color Per111cahili1y. S1ahili1y
Material Code µ (ppm/°C) Op1i11111111 Q Range 3

HA (41) Green 75 975 I kHz- IOOkHz

HP (3) Gray 35 370 50k Hz-500k Hz
GS6 ( 15) Red and wh ile 25 190 IOO kHz-2 Mhz
C (I) Blue 20 280 500 kHz-5 MHz
E (2) Red 10 95 I MHz-JOMHz
SF (6) Yellow 8 35 10M Hz-90M l-lz
w ( 10) Black 6 150 60 MHz-150 MHz
IRN -8 ( 12) Green and white 3 170 b 100 M Hz-200 MHz
PH (0) Tan I NC C 150MHz-300MHz

:•Typical r:mgc m ther than optimum Q range. Type I IA no! recommended for tu ned circuits: is best suited to
lo"-Q power transformers. noise filters. and pu lse circuits.
bNot linear.
cNA. information not av:1ilablc.
Cmirt~".I' A111ido11 As.wdote.<.

1.2.2 Ferrites
The term ferrite can be assigned to a large number of ceramic materials
that exhibit ferromagnetic properties. A ferromagnetic compound is one
that has the capability of being magnetized to a high degree.
As is the case with powdered-iron core materials, ferrites are compounded
in various mixes to achieve specific electrical characteristics. Iron oxide is
combined in the binder compound with such element oxides as nickel,
manganese, zinc, or magnesium. The end product is a hard, brittle
substance with a smooth sur face. The salient features of a ferrite core are
relative ruggedness, high available µ factors, and low eddy-current losses.
As mentioned during the discussion of powdered-iron core materials (sec.
1.2.1), there is a trade-off in using high-permeability ferrite cores-a
degradation of stability.
An examination of the curves given in Fig. 1-4 indicates that despite the
very high permeabilities attainable with ferrites, the thermal stability is not
as poor as one might think. It can be seen, however, that as the permeability
increases, the stability declines. The shortcoming can be minimized to a
considerable degree by maintaining a fairly constant ambient temperature
around the equipment area where the ferrite-loaded component is used. If
we consider the portion of the curve that represents the frequency region
 300

250

"-
> 200 Type 62
·-
.0
"'
Q)

E 150
~

"'
c. Type 61
ro
·;; Ty pe 65
c 100

50 Type 63

-40 0 40 80 120 160 200 240 280 320

Temperature (° C)
Figure 1-4 Curves that show the relationship between core temperature and initial
permeability. Stability is best for cores with low permeability.

over which the core will be used (narrow), the magnitude o f the temperature
versus µ shift is not as dramatic.
Perhaps the most significant problem attendant to the use of ferrite
cores is saturation. This phenomenon is defined as the state of magnetism
beyond which the B- H curve levels off to a straight line. The effect, shown
by the curve in Fig. 1-5, is one under which the core material cannot be
magnetized further. One of the effects of saturation is called "lockup." In
that state an inductor undergoes a shift in value and becomes "immobilized,"
so to speak. This will lead to circuit detuning and the generation of har-
monics.
A ferrite-loaded inductor or transformer that is fed excessive drive
power will heat considerably. In some instances the excitation can be so ex-
treme that the core material will suffer irreparable damage. The net effect
is a permanent change in core permeability. In a worst-case condition, the
core can fracture and separate in many pieces.
Table 1-2 lists the principal electrical characteristics of a group of
popular-size ferrite toroid cores. The data relate to materials that are
manufactured by Ferroxcube Corporation and sold in small quantities by
several dealers in the United States and Canada. Dimensional data are pro-
vided in inch.es. Although the core sizes in Table 1-2 are anything but all-
inclusive for the industry, they do represent the most common size grouping
used in audio and RF work . Smaller and larger toroid cores can be obtained

7
 Saturation

8 BL
µ =- =-
• H HL

Linear µ, = effective permeability

region
8 = magnetic flux
8 L 1-------1-~t­ density in gauss
H = magnetization in
lOW·fevel, oersteds
low· loss
reg ion

~
H
Figure 1-5 The curve shows how saturation of the core affects the linearity
of the magnetic-core inductor or transformer. The 8-H region is linear and
the BcHL region is the region of lowest loss.

TA B LE l -2

4C4 Mix JDJ Mi.r 387 Mix JCS M L< JelA Mi.<
Ferroxcubc-
c·()r~ no.
A, µ, AL fl, AL JI, AL µ, AL µ,

1041 T060 25 125 144 725 495 2500 NA NA 890 4495

266TJ25 SS 125 330 750 1100 2500 NA NA 2135 4830

76$T l88 70 12S 415 750 NA NA 1475 2700 2750 5000

S46T250 7S 125 NA NA NA NA 1650 2700 3055 5000

S02TJOO NA NA NA NA NA NA 1740 2700 3225 5000

Ferro.~c11Jn·
core no.
. B c
Jl)41T060 0.2)0 0.060 0.120 -~1 8 1-

@l
0.125 0. 187

Dl
266Tl25 0.375

768Tl88 0.500 CJ.188 0.281

846T250 0.870 0.250 0.540 Dimensions A. Band C :lre in rnchcs.

(25.4 )( inches = mm) NA ~ Nol Available m
502T300 1.142 O.JOO 0.748 specified mix.

8
Sec. 1-3 Calculating the.Coil Turns 9

from some manufacturers on special order. Cores large enough to handle 20

kW in low-impedance balun applications are available. They are several in-
ches in diameter.
The approximate number of wire turns for each of the cores listed in
Table 1-2, versus wire gauge, are listed in Table 1-3.

TABLE 1-3 Approximate turns versus wire gauge and core type.a

AWG F ERROXC UBE ( O RE N UMBER

Gauge 1041 T060 266Tl25 768T/88 846 T250 502T300

16 2 5 10 25 25
18 4 IO 15 35 35
20 5 15 20 45 45
22 8 20 30 55 55
24 10 25 40 75 75
26 18 30 50 95 95
28 20 40 65 115 11 5
30 25 50 80 145 145
32 35 65 100 180 180
34 50 80 130 240 240
"These da ta a re based on the use of enameled solid-conductor wire with Fo rmovar insula tion .
T he nu mber of turns specified is a pproximate a nd is intended as an aid when determining the core
size vers us inductance und wire size. based on the lllrns equation in Table 1-2.

1.3 Calculating the Coil Turns

Examination of Table 1-2 reveals that even though a particular mix of fer-
rite, say type 4C4, has a specific effective permeability (µe), the AL factor
changes with the size of the toroid core. Furthermore, the table shows that
with some core materials (3E2A, for example) the J..le value is different as the
core size is increased. Some designers are not aware of this condition and
are misled by the stated J..le for a given core material, which is usually based
on a fixed size of core-frequently with a diameter of l in . Therefore, it is
an error in judgment to believe that all core sizes from a selected mix have
the same J..le• even though the basic mix recipe has a broadly based J..li
number.
The fundamental equation for determining the number of coil turns for
ferrite toroid cores is

where N is the number of turns, L the desired inductance, and AL the

number assigned by the manufacturer for a given core. When AL is not
known, it can be determined by
where AL is the desired factor, L 0 the inductance in mH for a given number
of core turns (usually based on one turn), and N the number of turns used to
obtain a value for L. This equation is rather awkward to use when attempt-
ing to find an Al index for cores with low permeability, because a single
turn of wire would yield an inductance in nanohenries. Accurate measure-
ment in nH is somewhat beyond the instrumentation capability of some
laboratories. When a single turn on a given core provides inductance in µH,
this equation is more workable:

AL = LµH X IO•
Ni

When the core material is of lowµ, it is often more practical Lo use several
turns around the core to secure lhe AL index. As many as 20 turns can be
useful in some instances. Thus, if a given core upon which 20 turns of wire
were wound yielded a net inductance of 5 µH, the AL would be determined
by

5 x 10,000
125
400

Conversely, if the AL was known to be 125 and the desired inductance was
5 µH, the number of toroid turns would be found from
N = 100 v 5 µH + 125 = 20 turns where N is the unknown number of
wire turns. The results are based on the coil turns being spread linearly
around the entire circumference of the core material rather than being bunch-
ed up in a small area of the core. The equations assume also that a single-
layer winding is employed. It is useful in some applications to arrange the
coil winding to occupy approximately 180° of the toroid core. Final trim-
ming of the inductance in critical circuits can then be done by compressing or
spreading the turns on the core. This assumes that a spacing of at least one
wire thickness exists between the coil turns at the offset. The amount of in-
ductance variation will be relatively small at the lower J-le numbers, but will
increase as the µe of the core is made higher. Once the required inductance is
effected, the turns on the core can be cemented in place by means of
Polystyrene Q-Dope or a similar agent that is high in dielectric quality (Fig.
1-6).
Generally, the AL index for ferrites is based on mH per one turn on a
core, although the number of turns can be any convenient value, as stated
earlier. Powdered-iron cores have their AL indices based on µH per 100
turns. This accounts for the 102 (100) and the 1Ql (1000) multipliers in the
two equations.

10
 (spread)
Decrease
L

Increase
L
(comp ress)
Figure 1-6 Compressing or spread-
ing the coil turns on a toroid has a
marked effect on the inductance.

1.3.1 Determining Inductance

One of the more significant benefits from the use of toroidal inductors
and transformers is the inherent self-shielding characteristics of the toroidal
component. This is of benefit to designers who need to engage in high-
density packaging of RF circuitry without the need to include metal shield
enclosures or use space-wasting separation between the inductors and near-
by components. But when a laboratory inductance bridge is not available
for determining AL or making ordinary inductance measurements, a dip
meter can be used if a specific technique is employed. A dip meter is a wide-
range oscillator whose grid, gate, or base current is monitored by means of
a de meter. At resonances of the circuit under test, there is a sharp, well-
defined reduction in indicated current. Figure 1-7 illustrates an easy
method for using a dip meter with a toroidal inductor. Because the self-
shielding nature of a toroid prevents coupling to a dip meter by normal
means (no reading would result), it is necessary to add a temporary two- or
three-turn coupling loop (L2) to the toroidal inductor, as shown. An exter-
nal link, L3, is connected to L2 and the dip-meter probe coil is placed near
L3. C is a close-tolerance capacitor of known value. When a dip in the
meter reading is found- keeping the dip-meter probe as far from L3 as
possible to obtain a shallow dip- the operating frequency of the dipper is
checked against a calibrated receiver. Once f is known, the inductance can
be determined by assuming that L (unknown) is equal to the reactance of C,
since X L and Xe are equal at resonance. Since the value of C in Fig. 1-7 is
known, this method is straightforward. Let us assume that C in an example
investigation is a 100-pF, close-tolerance silver-mica capacitor. The dip-
meter output at the point where LI and C (known) are resonant is 8.7 MHz,
as observed by means of a calibrated receiver. From this we can obtain Xe
for C (known):

11
 L1

L (unknown)

L2. L3: two turn links

C: close-tolerance capacitor
of known value
Dip meter
Figure 1-7 Resonance of a toroidal inductor can be checked by means of a
dip meter and coupling link. W hen C is known, the dip meter can be used to
determine the coil inductance, as shown.

l
Xc = --
2rrJC
where Xe is the capacitive reactance in ohms, f is in MHz, and C is in µF.
Thus,
1
Xe = 6.28 x 8.7 x 0.0001 = 183Q

Since Xe = XL at resonance, the inductance of LI can be found from

LµH = XL = 183 = 3.35

2rrf 6.28 X 8.7
which yields an inductance value of 3.34774 µH if carried further to the right
of the decimal point.

1.3.2 Circuit Q
An indication of the relative Q of a tuned circuit can be had by observ-
ing the depth o f the leftward needle swing on di p meter. T uned-circuit Q is
the figure of merit. It is the ratio of reactance to resistance. The higher the
tuned-circuit Q, the deeper will be the di p in meter reading. Conversely, the
shallower the dip, the lower the circuit Q. When investigating the relative Q
it is essential that the dip-meter probe and the tuned circuit under test be
separated sufficiently to prevent overcoupling effects. A deep dip can be ob-
tained 8 in . or more away from very high Q resonators, such as strip lines,
cavities, and helical resonators. When checking toroid al inductors, as in
Fig. 1-7, the dip-meter probe is more likely to be less than an inch from L3,
and in some instances where the Q is medium or low, the probe may need to
be inserted into L3.
When it is desired to know the actual Qu (unloaded Q) of an air-core ,
ferrite, or powdered-iron-loaded inductor, a laboratory-grade Q meter,

12
Sec. 1-3 Calculating the Coil Turns 13

such as a Boonton 160 or Hewlett-Packard 4342A, can be used.

Measurements of Q should be made at the intended operating frequency of
the toroid, rod, or pot-core inductor. This rules out the use of an ordinary
RCL type of bridge, which uses a 1000-Hz oscillator. The latter is entirely
acceptable for work at audio frequencies, as would be the case with pot-
core inductors or toroids intended as resonators in audio-frequency filters,
and so on.
When laboratory test equipment is not available, the student or designer
can apply the technique shown in Figs. 1-8 and 1-9 to obtain the unloaded Q
value of a resonant circuit. The magnetic-core inductor is wound with the
required number of wire turns to provide the desired inductance (Ll of Fig.
1-9). Cl is adjusted to provide resonance at the chosen / o. Cl must be of
high dielectric and mechanical quality to ensure a high Q at /o. A signal
generator is set at/o. It should be a 50-Q instrument, and if not, a 50-Q pad
should be inserted in the line. With an oscilloscope or RF voltmeter con-
nected across Rl, trimmers Cin and Co are set for approximately equal
capacitances to establish roughly 30 dB of insertion loss. Low-value trim-
mers in the 5- or 10-pF class are usually suitable for measurements in the
MF and HF parts of the spectrum. Larger amounts of capacitance will be
required for LF, VLF, and audio-frequency measurements.
Once the - 30dB Joss has been obtained at/o, the signal source is varied
until the half-power points on the response curve are identified. Whenfo,J,,
and /2 are known, Q11 can be obtained as shown in Fig. 1-8. Hence, if in a
test circuit the value of fo was 2.8 MHz, /2 was 2.79 MHz, and/, was 2.81
MHz, we would find Ou by
2.8 2.8
Q11=---- = -- 140
2.81-2.79 0.02

r, 12
fo
=- -
r, - '2
Q
u

Figure 1-8 Frequency response of a reso-

nant circuit showing the relationship between
Q and the half-power points on the curve.
 --
.., - 30 dB
Co

5 pF 5 pF
son Scope
load or
L1 II Cl
R1 RF VM

Xu = Xc 1 atf0
Figure 1-9 Method for checking the unloaded Q of a resonant circuit.
The generator should be accurate and have suitable readout resolution
for locating the -3-db points on the response curve.

where f is in MHz. The results of measurements made in the manner il-

lustrated in Fig. 1-9 are entirely suitable for most RF design work.
When considering the available Q of an inductor, one must, of course,
include the resistance of the wire, R. Thus, the larger the cross-sectional
area of the conductor, the lower the resistance and the higher the coil Q.
This is related to the standard equation Q = 211/L I R, where f is in hertz, L
is in henries, and R is in ohms. It can be seen from this that magnetic-core
inductors are more favorably for obtaining high-Q performance than is
typically possible with lumped l/C circuits which utilize inductors with
many turns of small-diameter wire (Fig. 1-1 0). To realize Q values
equivalent to those which are common to powdered-iron or ferr ite-loaded
inductors it would be necessary to employ large air-wound coils with
substantially larger conductor diameters. With the latter it is impossible to
achieve workable miniaturization. Additionally, the self-shielding
characteristic of toroids and pot cores is lost if large air-wound solenoidal
inductors a re used. The dramatic contrast is seen in the photograph of Fig.
1-11, where the relative size of a toroidal inductor is compared to that of a
section of B&W Miniductor coil stock. For identical inductance values
(IOµH) the Q11 of the toroid coil at 8 MHz is 350 and for the air-wound unit
it is 450. It is important to point out, however, that in this example the air-
wound coil is capable of handling considerably more RF power than is
possible with the toroidal-wound coil. Each component has its virtues,
depending upon the application.

1.4 Power Capability

When an inductor is called upon to function in a circuit where considerable
power is present, the effective permeability (µe) of the core material will
vary in a significant manner. This can be seen by referring to Fig. 1-5. In ad-

14
 5 MHz

20µH L R

4.2 .a
50 pF

c
a - x . Q z 628.3 = 149.5
., - R ·· 4.2

Where XL =Xe and XL = 2wfl

Figura 1-10 Typical resonator, showing the

components L, C, and R.

Figura 1-11 How an air-wound and toroidal inductor

compare when the inductance is equal.

dition lo this variation in µ, the core losses will increase as a fu nction of self-
heating. This form of heating is apart from that which is brought on by en-
vironmental or ambient temperature. The self-heating effects are caused by
the excitation, which in turn heats the core material and the conductors on
the core. In addition to the core losses resulting from heat, the Q of the in-
ductor can degrade markedly at peak excitation. Earlier we acknowledged
the effect tha t this condition has on the stability of the inductor: Over a
wide frequency range it can be quite poor. In a practical design application
it becomes important, therefore, to acknowledge heating as an important
design limit.

1.4. 1 Flux Density versus µe

An important parameter for any magnetic material is the flux density.
This is the density of the lines of force as measured at a cross section of their
flow. To be more specific, the nux density is specified in lines/cm 1 in a

15
16 The Basics of Magnetic Materials Chap. 1

.agnetic circuit when measured at a given point. This is specified in units of

gauss, and in the magnetic-core industry as B. Bmax is, therefore, the max-
imum flux density in gauss units. This value of flux density corresponds to
the peak of the ac excitation. The value of Bmax is based on two conditions:
(1) where there is only ac excitation and (2) where ac excitation is accom-
panied by de voltage. The basic equation for the first condition is

X 10s
Bmax (ac) = _Erms
___..;._ _ _ gauss
4.44/N1Y4e

where Ae is the equivalent area of the magnetic path in cm 2 , Erms the applied
voltage, NP the number of core turns, f the frequency in hertz, and Bmax the
maximum flux density in gauss. For the second condition the equation is
written as

gauss

where Ide is the de current through the winding and AL the manufacturer's
inductance for the core material in use. In this case Bmax is classified as
"total" rather than just for ac.
The curves of Fig. 1-12 show the relationship between permeability and
flux density at the peak period of ac excitation for a group of core AL
values. These curves show the typical relationship of a specific shape and
size of magnetic-core material which operates at a selected temperature.
When the core is fully saturated, the permeability falls to a low value.
Next, let us consider the case where ac and de components are present in
a magnetic-core inductor. Figure 1-13 contains a family of curves that were

Ascending AL
values

B,,,,.(peak flux density)

ac excitation only
Figure 1-12 Curves demonstra ting the relation-
ship between permeability and flux density at
peak ac excitation for several A , factors.
(Courtesy of Ferroxcube, Division of Amperex
Corp.).
 Ascending
value H0

B"'"" (peak ac flux density)

with de excitation (clue to H 0 )
Figure 1-13 Curves showing the comparison be·
tween µ and Bmax under various conditions of de
magnetization with simultaneous ac excitation. A
single core material is represented here. (Courtesy of
Ferroxcube, Division of Amperex, Corp.).

obtained from a specific type of core material with a defined shape and air
gap. The curves compareµ to B max for different values of de magnetization
of the core, which is subjected simultaneously to ac excitation. The de
magnetizing force is known symbolically as Ho and uses oersteds as the
reference units. Ho is the unidirectional current flowing in an inductor or
transformer winding. The curves illustrate clearly that de magnetization
lowers the µe of the core material, causing it to saturate at a lower value of
peak flux density than would be the case under the conditions of Fig. 1-12.
The flux density discussed here is a result of ac excitation only.
There is a relationship between core loss and flux density with respect to
operating frequency. This consideration is shown by means of curves in Fig.
1-14. The curves were drawn for a specific core material, but the core struc-
ture and air-gap dimensions are not considered. It can be seen that the core
loss is related directly to the nux density at selected frequencies. Curves of
this variety can be established for core loss versus frequency at a number of
flux densit ies. In this illustration the core loss is given in terms ofmW/cm 2
of magnetic-core volume.
Finally, let us consider the matter of core losses versus temperature and
nux density. Once again a particular type of core material is chosen for the
collection of curve data, but the tests do not rely on the core shape or the air
gap. Rather, the curves of Fig. 1-15 are based on core loss versus flux density
at a number of core temperatures.
It was said earlier that part of the core heating can be related to the cur-
rent flowing through the conductor, which is wound on the core. It is
beneficial to select a wire gauge that will pass the ac and de current with a
minimum contribution to the self-heating characteristic. The de resistance
for an average winding can be determined by

17
 500

400

300
M
E 200
u
......
$:
E 100

0
0 1000 2000 3000 G
Bm•• (peak flux densi t y)
Figure 1-14 Curves that show the relationship among fre-
quency, core loss, and flux density for a specified core
material. The core loss is given in terms of mW / cm' of
magnetic-core volume. (Courtesy of Ferroxcube, Division of
Amperex Corp. J.

10 100 1000 10,000 G

Bmax (peak flux density)
Figure 1-15 These curves illustrate the relationship between core
loss and flux density at various temperatures. Again, a panicular type
of core material is the basis of the curves. (Courtesy of Ferroxcube,
Division of Amperex Corp.).

lwNr
ohms
12,000

where lw is the mean length of a single turn in inches, N the number of

turns, and r the resistance of the wire in ohms per 1000 ft. Table 1-4 pro-
vides the appropriate data for finding the wire resistance, plus other infor-
mation of interest concerning conductors which are used with magnetic-
core materials.
1.4.2 Sample Calculation

Let us carry out a sample calculation for core selection at the input of an
amplifier (Fig. 1-16). The operating frequency is 2.3 MHz and Tl will
match the driver stage to the bases of two power transistors operating in
Class C. Tl will function as a broadband conventional transformer.

18
 TABLE 1-4 Data for Solid Copper Wire

AWG Cross Feel Ohms Per Amps For Tums Tum.~

B&S DIAM Section Per Ohm !OOOf1. / OOOA I Perin. 1 l'erin.2
Gauge (in. J { in. 1 J { 200CJ ( 200 J i11. ! N.F. Wire S .F. Wire

10 0.10 19 0.00815 1001. 1.00 8. 15 89 92

II 0.0907 0.00647 794. 1.26 6.47 11 2 11 8
12 0.0808 0.005 13 630. 1.58 5.13 140 146
13 0.07 19 0.00407 499. 2.00 4.07 176 180
14 0.0641 0.00322 396. 2.53 3.22 22 1 231
15 0.057 1 0.00256 31 4. 3. 18 2.56 259 275
16 0.0508 0.00203 249. 4.02 2.03 327 346
17 0.0453 0.00161 198. 5.06 1.61 407 432
18 0.0403 0.00127 157. 6.39 1.27 509 544
19 0.0359 0.00101 124. 8.05 LO I 634 679
20 0.0320 0.000804 98.5 10.7 0.804 794 854
21 0.0285 0.000638 78. I 12.8 0.638 989 1063
_. 22 0.0254 0.000505 62.0 16.2 0.505 1238 1343
ID 23 0.0226 0.000400 49. 1 20.3 0.400 1532 1677
24 0.0201 0.0003 17 39.0 25.7 0.3 17 1893 2094
25 0.0179 0.000252 30.9 32.4 0.252 2351 2632
26 0.0 159 0.000200 24.5 41.0 0.200 2932 3326
27 0.0142 0.000158 19.4 51.4 0. 158 37 11 41 12
28 0.0 126 0.000 126 15.4 65 .3 0. 126 458 1 52 13
29 0.0 11 3 0.000 100 12.2 81.2 0. 100 562 1 6383
30 0.0100 0.0000785 9.69 104 .0785 7060 8145
31 0.0089 0.0000622 7.69 131 .0622 8455 10.097
32 0.0080 0.0000503 6.10 162 .0503 10.526 12.270
33 0.0071 0.0000396 4.83 206 .0396 13 .148 15.615
34 0.0063 0.00003 12 3.83 26 1 .0312 16.889 19.654
35 0.0056 0.0000248 3.04 331 .0248 2 1, 163 25.531
36 0.0050 0.0000196 2.41 415 .0196 26.389 31.405
37 0.0045 0.0000159 1.91 512 .0159 31.405 39.570
38 0.0040 0.0000 126 1.52 648 .0126 39.567 49,070
39 0.0035 0.00000962 1.20 847 .0096 53 ,855 65,790
40 0.0031 0.00000755 0.953 1080 .0075 65.790 82.180

The actual linear turns per square inch will "ary with the manufacturer and the insulation thickness. Courtesy of Ferroxcubc Corp.
 Amp. 1
2.3 MHz

Tl
1o n
II

Amp. 2

1. X L(pn.) "= 4R ,n

:. x L(p11.1 = 200 n
2. Select a core
(F erroxcube 76Tl88. t ype 4C4l

3. l = .!..!:.._ = 200 = 13.8 H

p ro. 2rrf 6.28 x 2.3 µ

4. Turns = lOOO JL mH + A 1• = lOOOJ0.0138 + 70 = 14turns

2
= E.,0.k x 10 22.2 x 100 _
5. 8
= =
116 75 G
""' 4 44 fnA • 4.44 x 2.3 x 14 x 0. 133 .
figure 1-18 Sampl e core calculation for use at Tl while supplying
excitation to a pair of Class C power amplifiers.

The primary reactance should be approximately four times the im-

pedance of the driver- 50 Q. The required primary XL is, therefore, 200 Q,
step 1 of Fig. 1-16. The next step calls for a core selection. A
0.5-in.-diameter Ferroxcube 76Tl 88 is chosen arbitrarily for Tl. Type 4C4
material should suffice, for it has a relatively high Bs (saturation
density)-3000 G. In step 3 we determine the primary inductance required
to satis fy the 200-Q XL at 2.3 MHz from step I . This computes to 13.8 µH.
T he number of core turns is calculated in step 4, using the equation set forth
earlier in the chapter. The core selected has an A L index of 70; hence it will
take 14 turns to provide an inductance of 13.8 µ H .
To permit the Bmax equation to be more workable for inductance in the
µH range, it has been modified from the basic form to that of step 5 in Fig.
1-16. We are concerned with the peak ac voltage at Tl, which at 4 W across
a 50-Q load is 22.2 V. (Although we have specified peak voltage in the ex-
ample, rms voltage could be used. £peak will allow fo r the worst-case condi-
tion .) The lowest operating frequency {f) is 2.3 MHz, the number of core
turns (n) is 14, and the cross-sectional area of the chosen core in cm 2 is
0.133, as specified in the manufacturer's data sheet. Bmax computes to 117
G. This is well below the rated Bs of 3000 G for the core, indicating the

20
Sec. 1-4 Power Capability 21

choice to be a good one. This will allow Tl to operate well within the linear
region.
To illustrate the manner in which Bmax is affected by changes in the
values of the terms in the equation of Fig. 1-16, let us substitute 0.5 MHz
for 2.3 MHz atf. The gauss becomes 537 rather than 117. Therefore, lower-
ing the operating frequency while using a core that will yield proper XL with
the same number of turns (14) at 0.5 MHz will greatly elevate the Bmax·
Next, by using a core that requires 50 turns instead of 14 to obtain 13.8 µH,
and changing nothing but n in the equation from Fig. 1-16, the Bmax drops
from 117 to 33.1.
Keeping all terms the same, but changing £peak in Fig. 1-16 to 100, Bmax
increases to 526 G. Now, to demonstrate the effect of Ae, we will keep all
terms the same but change Ae to 0.025, which represents a smaller toroid
core. The Bmax now rises to 629 G. Finally, to offer a more dramatic rela-
tionship of the terms in the equation to Bmax• let us change Epeak to JOO and
A e to 0.025. This shows the effect of increased power versus decreased core
cross-sectional area in cm 2 • Bmax under this set of conditions becomes 2798
G. The foregoing exercise demonstrates clearly that core choice is not a
casual matter, particularly in circuits where power is present. The pro-
cedures for determining Bmax• as given here, are applicable to all types of
magnetic-core material , regardless of the shape chosen by the designer. It
must be remembered that the foregoing discussion and the equation of Fig.
1-16 relate to Bmax where only ac is present. The equation must be modified
accordingly to accommodate the presence of ac and de in the circuit. The
basic equation for that condition was presented in Sec. 1.4. I .

1.4.3 Variations in the Equations

When the pertinent factors of a magnetic-core inductor are known, ex-
cept for the value of L , this equation can be applied:

L = 0.4nN'µe [~e] X
1
10
henries

where L is the unknown inductance in henries, N the number of core turns,

1.1.e the effective permeability of the core, Ae the cross-sectional area of the
core in cm 2 , and le the effective magnetic length of the core in cm. The value
of le can be found in the manu facturers' data sheets.
One can find Bmax from known facto rs other than those given earlier in
the basic equation. This variation is useful when specific numbers are
known:
22 The Basics of Magnetic Materials Chap. 1

where/is in hertz, Lin henries, Ve the effective core volume in cm 1 (Ve = le

x Ae), Ethe rms excitation voltage, and µe the effective permeability.
Bmax can be defined also by

Bmax = k vµ;;
i

where k1 = _2500E
_ __
fv Lve
It is useful also to know that µe = Bma// k 1 2 •

1.5 Volt- Ampere Ratings

A significant design parameter for a magnetic-core inductor which must
handle power is the volt-ampere rating. This rating is of interest with
respect to the lowest frequency at which the inductor will be utilized. Fun -
damentally, the volt-ampere rating for an inductor is determined by

£2
El= - volt-amperes
2nfL

where E is in volts rms, I in amperes, fin hertz, and L in henries. It can be

seen from this equation that the ratio of volt-amperes to frequency is

volt-amperes/Hz

where the terms of the equation are in the same units as specified for the
previous volt-ampere equation.
For the purpose of convenience, a new factor can be defined as k 2 :

The maximum value for k 2, as applied to any core material, size, or shape
can be expressed by

_!__:_ =
2n
I 5 .6 X
211
10 - 1
[Ve (Bmax)
[ µe
2
J volt-amperes/Hz

where Bmax in this equation is the maximum recommended flux density for
a specified core material, given by the manufacturer.
The reader may be wondering why the k, factor was brought into this
discussion and what practical purpose it can serve. The kl information helps
tell the designer which core size will fill the need for an inductor or
transformer. Step I is to work the k, equation given previously. Once the
Sec. 1-5 Volt-Ampere Ratings 23

value of k, is determined, the designer can select the smallest core that has a
k, value that is equal to or greater than the operational characteristics re-
quired. The Ferroxcube literature lists the k, value and those of other fac-
tors used in the preceding equations of Sec. 1-5. Tables 1-5 through 1-8 con-
tain this pertinent data for E cores, U cores, toroids, and pot cores.
1.5.1 A Sample k, Core Selection

Applying the k, factor is relatively simple if attention has been paid to

the discussion in the foregoing section. A sample problem appears in Fig.
1-17 to demonstrate how a suitable toroid core can be chosen for a specified
application. The diagram shows a 10-H inductor shunted across a 600-Q
audio line. The power level in the circuit is 0.5 W maxjmum. The lowest fre-
quency of interest is 100 Hz.
Step 1 of Fig. 1-17 determines therms voltage across LI. In step 2 the
VA (volt-ampere) characteristic is calculated. From this we can learn the k1
factor. Step 4 calls for inspection of the core characteristics in Table 1-7.
Select a core k, that is equal to or greater than 0.00074. The nearest value
(higher) is that of a K300502 core of the 3E type. The number of turns need-
ed to obtain IO Hof inductance is provided by the AL equation in step 5.

f = 100 Hz
8.66 v tow
P;n = 0.5 W max.

L1 Jg 10 H Eo
. Goon

1. E RMS = ,jWR = J0.5 X 600 = 8.66 V

74 9
2. · = 0.0119VA
6.28 x 100 x 10

3. k = 2r.EI = 6.28 x 0.011 9 = 0.000 74

2
f 100

4. From Table 1. 7 se lect core with

k 2 ;;;, 0.00074. Nearest choice is a
K300502, type 3E. Core k 1 = 0.001 .
AL = 1688, µ , = 2700

5. N= 1000JLmH 7 AL • 1000J1ox 10 3 71688 = 2434

Figure 1-17 Mathematical progression for selecting a suitable
core and wire turns for a typical audio reactor.
 782E272 3E E-E 2000 2180 2.60 0.0888 0.231 0.208 0.985 0.394
6.60 0.574 3.79 1.34 141 x JO - l 0.94
E-1 1900
4.65 0.588 2.74 0.670 0. 152 0.608 215 x 10- s 1.34
3E2A E-E 3000 3280 2.60 0.0888 0.23 1 0.208 0.985 0.394
6.60 0.574 3.79 1.34 166 X I0- 5 0.94
E-1 2800 4450 1.83 0.0912 0.167 0. 104
4.65 0.588 2.74 0.670 1.07
5690785 41\ E-E ?: 450 ;;:: 1050 2.00 0.137 0.275 0.137
5.08 0.884 4.51 0.884 1.4
5690786 4B E- E ?: 195 ?: 452 2.03 0.1 41 0.288 0.1 41
5. 16 0.909 4.72 0.909 0.323 0.129 408 X 10- s 2.62
K540075 3E E- E ?: 19 10 :?:.2575 3.98 0 163 0.65 1 0.429 2.08 0.833
10.11 1.05 10.67 2.77 0.289 0.115 628 X 10-s 3.25
783E608 3E E-E 2550 5900 3.83 0.28 1 1.08 0.429 1.86 0.756
9.73 1.81 17.7 1 2.77 537 x 10- i 2.28
E-1 2200 7500 2.64 0.280 0.739 0.214
6.71 1.81 12. 12 1.38 0.289 0.115 885 X 10- s 3.25
l\J
3E2i\ E-E 3400 8000 3.83 0.281 1.08 0.429 1.86 0.756
en 9.73 1.8 1 17.7 1 2.77 643 X 10- s 2.28
E- 1 3200 10.800 2 .64 0.280 0.739 0.214
6.7 1 1.8 1 12. 12 1.38 R72 x 10- s 2.74
9F520 JCS E-E (see note c) 3.02 0.251 0.757 0.300
7.67 1.62 12.4 1 1.94 852 x 10- s 2.20
JE E- 1 (see no te c l 2.6 1 0.283 0.740 0. 150
6.63 1.83 12.14 0.968 605 X 10 - s 2.74
3E E-E ;:: 1500 :?:.4200 3.02 0.251 0.757 0.300
7.67 1.62 12.41 1.94 592 x 10 - s 2.20
E-1 ?: 1500 ;?;5200 2.61 0.283 0.740 0. 150
6.63 1.83 12. 14 0.968 917 x 10 - l 4.57
56907 18 3E E-E ?: 1970 ?: 3120 5.52 0.273 I .5 1 0.816
1402 1.76 24.76 5.26 593 x 10- s 3.08
E- 1 ?: 1880 ?: 4500 3.54 0.263 0.930 0.408
8.99 1.70 15.25 2.63 162 x 10 - J 6.12
K540060 3E E-E 2360 6000 14.68 2.66 39. 19 5.73

a unless othernisc ~rccified. ""and .-11. value< :trc nominal. with a +25% toleran~-.:: on .-11, .
bFor power-inductor designs. k 2 = F. 1 .'/ .('. For power-lran<formcr dc<igns. Ki~ 2rrEp 1,.f /11(
' For 3C5 ma1cria l. the roll owing applie< · I' 2: 2100 :n 1000 G and 25•c : µ 2: 3000 a1 2000 G and 1oo•c. ;\ no mi nal ,, ,. \•:t lue of 2000 '""employed in calculu1ing the val ue or k2 fo r
J CS core nrnlcrial .
 TABLE 1 -5 E cores.

Core
/'art
Core
Material
Co11-
[igura- ('
A1_
f mH fli'I'
le
( ill.
,, ('

l in. 1
ve
I in. J
A ,.
l in. 1
Acn
/i11. l
0.4!1c8
(in. 2
APPROX. D ,\T,\
Po
N 11111bl'r T vpe 1io11 Re{. 0 !0001/11'11.<) " cm ) c 111 2 1 cm JI cm 2 1 cm 1 / cm 1 / K b (W)
2

8 1'1E250 JE2A E-E 2800 2600 1.02 0.0298 0.0304 0 .039 0.008 J0 0.0032 303 x t0 - 6 0.32
2.59 0. 192 0.499 0 .252 0 .0523 0.0209
E- 1 2200 2900 0. 759 0.0314 0.0238 0.0 195 30 1 x I0- 6 0.23
1.93 0.202 0.390 0 . 126
81J E 187 3E2A E-E 3200 2300 1.54 0.0348 0.0538 0.082 0.0645 0 .0258 468 x 10- 6 0.48
3.91 0.224 0.882 0.529 0.416 0. 166
3.91 0.224 0.882 0.5 29 407 x t0 - 6 0 .39
E- 1 2600 2600 1.1 0 0 .0346 0.038 0 .041
2.79 0 .222 0.623 0.264 0 .0320 0.0128 860 X 10 -s 0 .65
813E34J 3E2A E-E 3200 4250 1.54 0.0639 0.0987 0 .082 0.2060 0 .0825
3.9 1 0.413 1.62 0.529 758 x J0 - 6 0.48
E- 1 2600 4800 I. I I 0 .0639 0.0707 0 .041
2 .82 0.4 13 1.1 6 0 .264 0.0470 o.oi 8& 577 x J0- 6 0.60
N
873E l 89 3E E-E 2080 1730 1.69 0 .044 0.0745 0.0842 0 .303 0 . 12 1
4 .29 0.284 1.22 0.543 468 X I0 - 6
"" E-1 1850 2220 1.20 0.0452 0.0541 00421
0.43

3.05 0 .29 1 0.887 0 .272 0 .0470 0.0188 74 1 x 10- 6 0.60

3E2A E-E 2800 2300 1.69 0 .044 0.0745 0.303 0.12 1
4.29 0 .284 1.22 0.543 569 x J0- 6 0.43
E-1 2650 3200 1.20 0.0452 0.0541 0 .0421
3.05 0.29 1 0 .887 0 .272 0. 1034 0 .04 14 993 x 10 - 6 0.79
812E250 3E2A E-E 3320 3500 1.90 0 .062 1 0 . 118 0 . 125 0.667 0 .267
4 .83 0.400 J.94 0.806 832 x 10 - 6 0.58
E- 1 2900 4 100 1.0 0 .0622 0 .0867 0 .0625
3.56 0 .400 1.42 0.403 0.0472 0.0189 336 x 1 0-~ 1.2 1
206F440 3C5 E- E (see note c) 1.89 0. 154 0.292 0.0992 0.304 0. 122
4.80 0 .993 4.79 0.638 261 x 10- s 0.95
E- 1 (sec note c ) 1.45 0. 157 0.227 0 .0496
3.68 1.01 3.72 0.319 0.0472 0.0189 246 x J0 - 5 1.21
3E E-E 1900 4800 1.89 0. 154 0 .292 0.0992 0.304 0 .1)2
4.80 0 .993 4 .79 0.638 207 x l(l - 5 0 .95
E- 1 1750 6 100 1.45 0.157 0.227 0 .0496
3.68 I.O J J .72 0 .319 0.0472 0.0 189 285 X 10 - s l .2 1
3E2A E-E 2800 0 .304 0.122
4 .80 0 .993 4.79 0.638 253 X I0 - 5 0 .95
E-1 2500 8650 1.45 0 . 157 0.227 0.0496
3.68 I.O J 3.72 0 .3 19 0. 152 0.608 185 x J0- 5 1.34
 TABLE 1-s U Cores. APPROX.DATA

Core Core Con- Al le Ae Ve Ac Aca 0.4ACIJ

Parr 1\lfateria/ (rgura- µe (111Hper (ill. ( in. 1 (in. J (i11. l (in. 1 (i11 . 1 Po
Number Type tio11 Ref 0 /000 llll'llS) a cm) cm 2) cm 3 J cm') cm 1) c111 2) k1b (W)

702U295 3E2A U-U 1950 410 0.797 0.00525 0.00419 0.0265 602 x 10 7 0.09
2.02 0.0338 0.069 0.17 1
3E2A U- 1 1650 370 0.613 0.00432 0.00265 0.0132 0.00740b 0.00296b 448 x 107 0.06
1.56 0.0279 0.0435 0.0851 0.0477 0.0191
376U250 3E2A U-U 3600 2200 3.29 0.0625 0.205 0.375
8.37 0.400 3.36 2.418 159 x 105 1.21
3E2A U- 1 3200 2500 2.53 0.0625 0.158 0.188
6.43 0.400 2.60 0.217 138 X 10 1 0.98
105F250 3E2A U-U 2200 1100 l.16 0.0180 0.0209 o.os2c
2.95 0.116 0.343 0.339 265 x 106 0.23
3E2A U- 1 1950 1150 0.979 0.0180 0.0176 0.0263
2.49 0.1 16 0.289 0.1696 252 x 10• 0.18
lF30 JCS U-U (see noted) 4.31 0.134 0.579 0.562
10.9 0.864 9.50 3.624 667 x 105 2.32
N
en 3C5 U- 1 (see no te d) 3.33 0.137 0.456 0.281
8.46 0.884 7.48 1.812 525 X 105 1.74
IF31 JCS U-U (see no te d) 4.45 0. 161 0.719 0.616
I 1.3 1.04 11.8 3.97 828 x 101 2.82
3C5 U-1 (see no ted ) 3.59 0. 165 0.594 0.308
9. 12 l.06 9.74 1.986 684 x 10~ 2.19
l Fl9 3C5 U-U (sec noted) 7.13 0.227 1.62 1.950
18.1 1.46 26.6 12.577 187 x 104 5.49
3C5 U-1 (sec noted) 5.73 0.23 1 1.32 0.975
14.6 1.49 21.6 6.288 152 x 10' 4.08
IFIO 3C5 U-U (see no ted) 6.77 0.316 2.14 l .500
17.2 2.04 35.1 9.675 246 x IO' 5.42
JCS U-1 (see no ted) 5.76 0.316 1.82 0.750
14.6 2.04 29.8 4.837 209 x 104 4.36
IFS 3C5 U-U (see no ted) 12.4 l.00 12.1 5.000
31.5 6.45 198. 32.25 139 x 103 17.7
JCS U-l (see noted) 9.64 l.00 9.64 2.500
24.5 6.45 158. 16..125 Ill X 101 I J.45

"Unless otherwise specified, Ref. and AL values are no minal with a + 25% tolerance on AL·
bror Power-I nductor designs. k 2 = £ ' / Lf'. For power-trnnsfo rmcr designs. k2 = 2rr£p lp/Mf
coobbin
rl _.
702Fl82 mounts on I core.
- ..
 TABLE1·7 Toroids.
APPROX . DATA
AL
v,.
Corp
Pan
N11111bN
Core
Material
Ty pe
µ
R!!f
e
( mH pl.'r
/000 11/TllS)
( + 20% )
'·
( ill.
Cl/I)
A,
( i11. l
cm 2 )
( in. 1
cml I
Ac
( ill.
c111 : 1
0.4Ac
( in. 1
cm1 ) k2:1
p
(I~)
2 13T050 4C4 125 24 0.422 0.00249 0.00105 0.00635 0.00254 b b
1.07 0.0 161 0.0172 0.0410 0.0 164 b b
303 725 140 1.07 0.0 161 0.0172 0.0410 0.0164
3E2A 4470 850 1.07 0.0161 0.0172 0.0410 0.0164 653 x 10- 1 0.018
1041T060 4C4 125 25 0.532 0.00330 0.00176 0.0116 0.0463 b b
1.35 0.0213 0.288 0.0748 0.0298
303 725 144 1.35 0.02 13 0.288 0.0748 0.0298 b b
3E2A 4495 890 1.35 0.02 13 0.288 0.0748 0.0298 109 x 10- 1 0.025
266Tl 25 4C4 125 55 0.852 0.01 18 0.0 100 0.0275 0.0110 b b
2.16 0.0760 0.164 0. 177 0.0710
303 750 330 2. 16 0.0760 0.164 0.177 0 .0710 b b
,.,, 3E2A 4830 2135 2. 16 0.0760 0.164 0. 177 0.0710 577 x 10- 1
0.05
..... 768Tl88 303 750 415 1.19 0.0206 0.0247 0.06 18 0.0247 b b
3.03 0.133 0.403 0.400 0.159
3E2A 5000 2750 3.03 0.133 0.403 0.400 0. 159 137 x 10-• 0.14
846T250 3E2A 5000 3055 2.18 0.04 18 0.0908 0.228 0.09 12
5.52 0.270 1.49 1.4 7 0.587 506 x 10-6 0.32
K300502 3E 2700 I688 2.92 0.0582 0.169 0.440 0. 176
7.43 0.375 2.78 2.84 I . 14 IOI x 10-) 0.50
K300500 3E 2700 2422 3.58 0.101 0.360 0.643 0.257
9.1 1 0.650 5.92 4.15 1.660 214 x 10- i 0.78
K300501 3E 2700 3639 3.58 0. 151 0.542 0.64.1 0.257
9. 11 0.975 8.88 4.15 1.660 321 x 10- 1 1.2
528T500 JCS 3.40 0.187 0.634 0.465 0.186
8.63 1.21 10.4 3.00 1.20 733 x 10- 1
1.3
400T750 JCS 5.00 0.281 1.40 1.27 0.508
I 2.7 1.8 1 23.0 8.20 3.28 162 x 10- J 2.2
144T500 JCS 6.7 0.342 .30 1.84 0.736
17. I 2.2 1 37.8 I 1.86 4.744 265 x 10- J 3.0

"For power-inductor designs. k 2 • £l/ L{'. For power-transformer designs. k2 = 21r£1,tp/Mf

bsince this malcrial is nm n.'(;ommcnded for high-Oux density applicnlions. no values Mc given for k2 and Po.
cFor JCS malcrial. the following applies: µ > 2100 al 1000 G and 25°C: µ > 3000 at 2000 G and 100°C.
Courtesy of Ferroxcube Corporation.
 TABLE1-8 Po t cores
ArrRox DATA
AL
Cort• Core Core ( 111H per le l ie Ve Ac Aco 0.4ACIJ
/'arr Material Diam. µea I000111msJ 0 (in. (in. I ( ill. J (in. 1 ( i11. 1 ( i11. l k/' Po
N 11111hN Type 111111 R ef a + 25% cm / /'/Il l ) c 111 3) c 111 1 ) c 111 2 ) c 111 2J (W)

743PIJ3-3E 3E 7 1100 576 0.408 0.0067 0.0027 0.0 11 6 0.0056 0.0022 399 x 10- 1 0.57
1.04 0.0433 0.045 0.0747 0.036 1 0.0 144
332 Pl33134-3 E 3E 9 1100 1350 0.559 0.0216 0.0120 0.0124 0.0057 0.0021
1.42 0. 139 0.197 0.080 0.0368 0.0137
174 x 10-' 0.86
l 107P-L00-3B7 387 11 1430 1940 0.608 0.0259 0.0157 0.0164 0.0083 0.0033 378 x 10- 6 0.103
1.54 0.167 0.257 0.105 0.0535 0.0214
l 107P-L00-303 303 11 635 865 1.54 0.167 0.257 0. 105 0.0535 0.02 14 c c
N l 107P-L00-4C4 4C4 II 114 155 1.54 0.167 0.257 0.105 0.0535 0.02 14 c c
0)
1408 P-L00-3 R7 387 14 1400 2240 0.77 1 0.0386 0.0297 0.0267 O.Ql5 0.0060 665 x 10 - 6 0. 161
1.96 0.247 0.487 0.172 0.09b8 0.0387
1408 P-L00-3139 3B9 14 11 90 1910 1.96 0.247 0.487 0. 172 0.0968 0.0387 c c
1408 P- L00-3 0 3 303 14 705 11 30 1.96 0.247 0.487 0. 172 0.0968 0.0387 c c
1408 P- L00-3 E 3E 14 l300 2080 1.96 0.247 0.487 0.1 72 0.0968 0.0387 365 x J0- 6 0.16 1
1408P- L00-4C4 4C4 14 125 200 1.96 0.247 0.487 0.172 0.0968 0 0387 c c
18 11P-L00-3B7 387 18 1740 3680 1.02 0.0672 0.0683 0.0458 0.029 0.01 16 123 X 10- 5 0.265
2.59 0.433 1.1 2 0.296 0. 187 0.0748
1811 P-L00-309 389 18 1250 2630 2.59 0.433 I. I 2 0.296 0.187 0.0748 c c
1811 P- L00-303 303 18 735 1550 2.59 0.43~ 1.12 0.296 0. 187 0.o7-t8 c c
1811 P-L00-3E 3E 18 1660 3500 2.59 0.-03 1.12 0.296 0.187 0.0748 658 x 10- 6 0.265
1811 P-L00-4C4 4C4 18 125 265 2.59 0.433 1.12 0.296 0. 187 0.0748 c c
2213 P- L.00-3 B7 387 22 1825 4650 1.23 0.0985 0.122 0.0652 0.046 0.0184 208 x 10-i 0.392
3. 12 0.6JS 1.99 0.420 0.297 0. 118
22 1J P-LOO-J B9 389 22 1275 3250 3.12 0.6J5 1.99 0.420 0.297 0.11 8 c c
 2213P-L00-303 30 3 22 705 1800 3. 12 0 .635 1.99 0.420 0.297 0 . 11 8 c c
22 I 3P-L00-3E 3E 22 1725 4400 3. 12 0 .635 1.99 0.420 0.297 0.1 18 11 2 X 10- s 0 .392
22 13P-L00-4 C4 4C4 22 120 JOO 3. 12 0 .635 1.99 0.420 0.297 0 . 118 c c
26 16P-L00-3 B7 38 7 26 1880 6000 1.47 0.1 47 0.21 7 0 .0890 0.063 0.025 362 X 10 - s 0.555
3.73 0 .948 3.56 0.574 0.406 0 . 163
2616 P-L00-3B9 389 26 1380 4390 3.73 0.948 3.56 0.574 0.406 0.163 c c
26 16 P- L00-303 303 26 735 2340 3.73 0.948 3.56 0.574 0.406 0.163 c c
2616 P-L00-3E 3E 26 1830 5850 3.73 0.948 3.56 0.574 0.406 0 .1 63 190 x 10 -s 0.555
2616P-L00-4C4 4C4 26 120 390 3.73 0.948 3.56 0 .574 0.406 0. 163 c c
3019P-L00-3 B7 38 7 30 2020 7580 1.77 0. 214 0.380 0 . 124 0.091 0.0364 608 X 10 - 5 0 .767
4.50 1.38 6 .23 0 .800 0.587 0.235
3019 P- L00-3 B9 J B9 30 1480 5750 4.50 1.38 6.23 0 .800 0 .587 0. 235 c c
3019P-L00-303 303 30 745 2820 4.50 1.38 6.23 0.800 0.587 0 .235 c c
3622P-L00-3B7 387 36 2000 9660 2.08 0 .313 .65 1 0.167 0.1 16 0.0464
N 5.28 2.02 18.7 1.08
102 x 10- · 1.06
0.748 0.299
U>
3622 P-L00-3B9 3 B9 36 1440 7050 5.28 2 .02 10.7 1.08 0.748 0.299 c c
3622 P-LOO-JD3 303 36 745 3580 5.28 2.02 10.7 1.08 0.748 0.299 c c
3622 P-L00-3 E 3E 36 2250 10.ROO 5.28 2.02 10.7 J.08 0.748 0.299 464 x 10-1 1.06
4229P- L00-3B7 3 87 42 2100 10,300 2.68 0 .41 3 l. I I 0.301 0.2 17 0.0868 165 x 10- • 1.60
6 .81 2.66 18 . 10 1.94 1.40 0 .560
K5 350 56-3E JE 45 2400 I J.5 00 2 .56 0.45 1 1. 15 0 .25 1 0. 160 0 .0640 764 x 10- 1 1.59
6.50 2.9 1 18.8 l.62 1.03 0.412
K5 350 I l-3E 3E 66 2440 18.200 4.85 I. II 5.38 .857 0.620 0.248 352 x 10 - • 4.42
12.30 7.16 88.2 5.53 4.00 1.60

a Per pair of cores.

bFor power-inductor designs. k 2 = £'IL/'. For power-transformer dc~isns. K 2 ; hEplp/ M.f.
cSince this mntcriul is 110 1 rcco111111cndecl ror high-"ux densit y applicn1io 11s. no values arc given for k 2 und P0 .
Courtesy or Ferroxcuhc Co rrorn ti on.
30 The Basics of Magnetic Materials Chap. 1

The design objective has been satisfied when 2434 turns of wire are layer-
wound on the toroid core. The progression of design is patterned after in-
formati on contained in Ferroxcube Bulletin 330-A. In a practical situation
the core used should be physically larger than that chosen in Fig. 1-17, even
if the k1 rating is substantially higher. Thjs will facilitate easier winding and
perrrut the use of larger wire gauges than would be required for the core
specified.

1.5.2 Power Rating

Tables 1-5 through 1-8 contain a set of Po factors (power dissipation fac-
tor) in watts. This is the power-dissipation level which will cause a
temperature rise of 50 °C above the ambient level within a specified core. P 1
(total power dissipation) is a significant factor also. It is the total power
dissipated in an inductor, constituting winding loss and core loss . The
dielectric losses are not part of the P1 factor.
Once the wire size and number of coil turns have been chosen, the
designer can find the resistance of the winding (Table 1-4) and estimate the
total winding copper loss. The core-loss factor can be obtained from the
manufacturer's published curves for the material used. These data, plus the
winding-loss data, provide the information needed to calculate the P 1•
Once the power is known, the temperature-rise trait can be taken from

where the factor Po is given for each core type in the tables. Once this info r-
mation is acquired, the designer can make final adjustments (empirical) to
compensate for the temperature-rise effects on µe and the inductor Q. It
may at this juncture be necessary to modify the number of coil turns slightly
to meet the initial design objective. Typically, assuming that the design was
done carefully at the beginning, the changes caused by heating should be of
low percentage.

1.6 Inductance versus AC/DC Excitation

It is a common design characteristic to have a large component of de excita-

tion present along with the ac excitation related to a given pbwer inductor.
Figure 1-18 illustrates a typical application, where Vee is supplied to an RF
power transistor through a magnetic-core choke inductor, Ll. The same
condition does not prevail with respect to L2 and L3 of QI because C l
serves as a blocking capacitor. L2 and L3 are subjected to ac excitation only.
It is not uncommon for the de excitation component to exert con-
siderably more magnetomotive force than is applied by the ac-derived ex-
 RF ampli fier L2 L3
Cl

1uT 1_ 7

+Vee
Figure 1-18 Example of a magnetic-core inductor used to
supply de operating voltage to the collector of an RF-
amplifier transistor. Ac and de components are present in
L 1, but only ac is a consideration at L2 and L3.

citation. In a design of this variety the attendant problems can be minimized

to a considerable extent by introducing an air gap in the magnetic circuit.
The advantage is that a smaller core can be used than would be possible with
a closed magnetic circuit, such as a classic toroid core. The principle of air-
gap inclusion is shown in Fig. l-19a.
The designer must be aware that the introduction of an air gap has a
marked effect on the inductance that would be present in a closed-core cir-
cuit. The air gap creates what can be termed a "hybrid flux path." This is
because part of the flux path is via a fairly long ferrite circuit of high µ,
while another part of the path is through a fairly short low-µ air gap. The ef-
fective length of the magnetic path under these conditions can be expressed
as
le = Im + µ 18
where µ is the permeability of the material, 1111 the length of the magnetic
path (normally specified as le for zero gap), and lg the length of the core
gap. All lengths are in centimeters. The air-gapµ is considered as I, or unity.

1.6.1 Employment of the Hanna Curve

Air gaps in cores are used with regularity in the design of power chokes,
filter inductors, and transformers. We have just outlined the fundamentals
of designing with air gaps, but finding the exact air-gap dimension for a
given optimum condition versus core material, inductance, and de current is
a tedious undertaking at best. A simplified method was disclosed by C.R.
Hanna in his 1927 AlEE Winter Convention paper. His method enables the
designer to derive the best air gap for a speci fic group of requirements.
From his technique, Fair-Rite Products Corp. constructed a Hanna curve
for use with its Fair-Rite No. 77 core material, a popular type for use in
power inductor and transformer applications.

31
 µ

Hmax

!al (bl
Figure 1-19 The concept of a magnetic core with air gap is shown in (al. The curve in (bl
illustra tes how µav departs from a straight line. However, the straight line is used as the
ba sis for µav in a practical situation.

During our discussion earlier, we established that de current flowing

through the winding of a ferromagnetic inductor would saturate the core to
some extent, thereby reducing the inductance. This effect is demonstrated
clearly by the curve marked "Air gap = O" in Fig. 1-20. Not only does an
air gap lower the inductance, it decreases the inductance more and more as
the gap is made wider. But the benefit realized from increasing the width of
the gap is greater direct-current capability before core saturation takes
place. The various air-gap curves in Fig. 1-20 illustrate this principle rather
clearly.
A typical method for finding the optimum core gap of a given inductor
is that of trial and error-a truly empirical approach. The usual approach is
to chose a core that will yield the required amount of inductance, based on a
wire gauge that will handle the de current without excessive heating or IR
voltage drop. Next, the assembled inductor is subjected to the normal level
of de current to learn if saturation is occurring. If saturation is noted,
assorted air gaps are tried experimentally until one dimension prevents
saturation. The shortcoming is a need to experiment further-adding turns
on the coil to compensate for the lost inductance. This can require a larger
core in some instances. The foregoing routine is repeated until the proper
core gap and core size is found . This means a core that will not saturate
when de bias is applied. The process is a tedious one, indeed!
T he procedure is simplified by using the curves of Fig. 1-20 for the core
that is chosen fi rst. The designer can pick a curve that is flat almost to the
maximum amount of de bias current in the coil. The curve chosen will show
the available inductance per coil turn versus the air gap selected. From this
information it is a simple task to calculate the total turns required for the
target inductance. If, of course, the speci fied number of turns will not fit on
the core, a larger one must be chosen. This will necessitate a repeat of the
selection procedure and the use of a new set of curves.

32
 10.000 ------------~~-~~.--.-
. ~~~

:i:

"'
S? 1000
~
...
~
E
2
~
Q)
0.
Q)
u
c(\) 100
v
:J
'O
.:

0.040 1n.
10 L-----L.---l'-J...W...W...-'---'-----'-....L.L.1....J....l....L.-_..._....._......._...........1....LI
0.01 0. 1 10

DC bias !Al
Fig ure 1-20 Curves that show the effect of increasing air gap versus induc-
tance. (Courtesy of Fair-Rite Produ cts Corp.).

Air-gap selection by means of the Hanna curve eliminates the need for a
separate set of curves for each core size, with respect to de bias and induc-
tance. But, of greater significance, it negates the need for trial-and-error
determinations.

1.6.2 Constructing a Hanna Curve

Observation of the curves in Fig. l-21 illustrates the construction of the
Hanna curve. Note the similarity between these curves and those of Fig.
1-20. The vertical scale is different, however, as a result of multiplying it by
the de current squared. The dashed line of Fig. 1-21 provides a point of
tangency on each of the curves which indicates the optimum inductance for
the given air gap and de bias. The dashed line is shown with slight altera-
tions in Fig . 1-22. In this presentation it is the classic Hanna curve.
In Fig. 1-22 the scales are normalized as a result of dividing the vertical
scale by the effective core volume. The horizontal has been changed to in-
dicate the magnetizing field in oersteds. T he curve is characteristic of the
Fair- Rite No. 77 core material. It can be used for any core size in that grade
of material. The air-gap tangency points indicated in Fig. 1-23 are identified
in Fig. 1-22 by means of the short marker lines which are perpendicular to

33
 -
O> = 0.040 in. /
0
x

-
"'
5
_,
1000

-0
~
"'cr
:>

;:"'
~ 100
:>
tJ

x
E ; 0 .001 in.
3
....
"'a.
G>
0
c
~
tJ
10
Ai r gap = 0
I
::>
-0
c

1 ..___,_~.__'-'-
l ul'-'-l....._~.....__,__,_~....._,_.._._~_._--'-~..J...J-..1...J..JU

0.01 0. 1 10
DC bias (A)
Figure 1-21 This group of air gaps form a Hanna curve. The curve is strik -
ingly similar to that of Fig. 1-20. (Courtesy of Fair-Rite Products Corp. ).

the curve. These points are based on the value of the air gap, divided by the
effective path length of the core (le) in centimeters.
Normally, the first step in using the Hanna curve dictates selecting a
core that will yield the required amount of inductance versus an acceptable
wire gauge. Following this it is possible to calculate the vertical factor of the
curve by

VI= L x Ii
Ve
where L is the inductance in henries, I the de current in a mperes, Ve the ef-
fective core volume in cm\ and Ve the vertical factor.
Once the vertical factor is derived, a horizonta l line is drawn from the
vertical axis until it meets the Hanna curve. At this point of intersection we

34
 10 2

!!./ : air gap

le : elfective path
,.., length of core
E
u
~

10 3
"'a.
..
"'.,,
a;
a.
E
"'
"'
·~
c
£"'
10 •
":::>"
.........
-....J

10 100 1000
H: 0.4r.Nl/lc ( 0 )
Rgure 1-22 Classic Hanna curve as extracted from the significant points on
the curve in Fig. 1-21 . (Courtesy of Fair-Rite Products Corp.).

have the value of the magnetizing field in oersteds (produced in the core).
T he equation for this is

H = 0.4 x N x I
le
where H is in oersteds, L the inductance in henries, I the de current in
amperes, and le the effective path length of the core in centimeters.
Once this value is obtained, it is possible to calculate the number of
turns required. As mentioned earlier, if the turns will not fit on the chosen
core, a larger one must be selected and the process repeated.
The air-gap factor is found at the intersection point of the horizontal
line and the Hanna curve. The air-gap factor can be determined by

where AG/ is the air-gap factor, ~! the air gap in inches, and le the effective
path length of the core in inches.
The air-gap factor varies in a nonlinear manner. For this reason it is dif-
ficult to use interpolation between the defined points on the curve. The

35
36 The Basics of Magnetic M aterials Chap. 1

foregoing calculations are based on the premise that negligible ac current is

present respective to the amount of de current in the circuit. lf the ac cur-
rent is appreciable, however, its peak value must be added to the / de· This is
because the total current will produce the maximum flux density in the core
material .
Figure 1-23 contains air-gap curves for a group of Fair-Rite E cores
made from No. 77 material. Similar curves for the manufacturer's pot cores
are presented in Fig. 1-24. Both sets of curves were developed from the
H anna curve procedure.

1. 6.3 Actual Air Gap

We are now concerned with the total air gap, as obtained from the
curves. It can be effected by grinding away a part of the E-core center post.
Alternatively, most manufacturers will supply cores with the prescribed air
gap. The same technique is used with pot cores.
When the designer uses two E cores or two pot cores, it is more practical
to grind only one-half of the core pair rather than removing equal amounts
of material from each core half. Alternatively, one can insert nonmagnetic

:x:
"' 9477025002
~ 94770 3600 2
1000
=_, .:94 77032002
:!
E 9477015002
z 9477016002
~ - 94 77020002
a.
"'c:
(.)

100
9477019002
"'::J
u
'O
=

10 .__~...__~~....................__~~~~~~~~~.__..................~

0.00 1 0.010 0. 100 1.00

Air gap (in.)

Flgure 1-23 Curves for a group of Fair-Rite E cores which are made of No.
77 (1800µ;) ferrite material. (Courtesy of Fair· Rite Products Corp.).
 :r:
5677362221
"'0
1000
c:
_,
5
E
2
~
OJ
5677140821
.,0.
(J
c: 100
"'
;:;
:i
i:::>
c:

10L-~-'---'-....J....U....LJ..JLL~-'-_._....!......LI..J...J....l-'-~--'-----'----'-...Ll...LJ..U
0.001 0.010 0. 100 1.00
Air gap (in.)
Figure 1-24 Fair-Rite Products Corp. curves for No. 77 ( 1800µ;) pot cores.
(Courtesy of Fair-Rite Products Corp.).

shim stock between the core halves to provide an air gap. In such cases the
stock is inserted between the outer posts of the E cores and between the
halves of the pot cores along the outer perimeter of the cores. When this
technique is applied, it is necessary to use shims which are half as thick as
the desired air gap. This is because the center post of the core is elevated by
the same amount as the outer legs or walls, creating an effective gap which
is twice the thickness of the shim stock.

37
2
APPLICATION OF RODS,
BARS, AND SLUGS

We discussed in Chapter l the basic principles and design considerations

for the use of magnetic cores. Those details will not be repeated except when
it is useful to demonstrate a design procedure. In this chapter we address the
practical applications in which one might use magnetic-core materials in the
form of rods, bars, and slugs. For the most part, circuits that employ these
physical forms are in the narrow-band category, principally for use in the
radio-frequency spectrum. For the purpose of design simplification we will
work with approximations and rules of thumb where it is practical, side-
stepping the more rigorous mathematical solutions. Fortunately, the perform-
ance results obtained from circuits that are based on well-founded rules of
thumb are entirely satisfactory for most prototype work. Refinement of the
basic design can always be accomplished, if needed, just prior to the
preproduction stage, or the "proof-build" stage, as it is sometimes called.

2.1 Rods and Bars

Ferrite rods and bars are available in standard lengths up to 8 Y2 in. The rods
can be obtained with diameters from ~ in. to V4 in. Bar stock is available in
thicknesses from Ys to ~ in., with widths from % to Ya in. The bar stock is
sold in the flat or rectangular format (Fair-Rite Classes 34 and 35). Rod
materials can be purchased in the solid or hollow formats (Fair-Rite Classes
30 and 31, respectively).
Ferrite rods and bars are used principally as magnetic cores for radio
antennas from the VLF spectrum well into the VHF range. The advantages
of these physically small antennas are high Q, compactness of assembly,

39
40 Application of Rods, Bars, and Slugs Chap.2

and portability of equipment. The rods and bars manufactured by Fair-Rite

Products Corp. are in materials 61 (µi = 125, 2350 G) and 64 (µi = 250,
2200 G). Specific rod and bar sizes are available in materials 33 (µi = 800),
63 (µi = 40), and 68 (µi = 20). Number 63 stock is optimum for use at ap-
proximately 30 MHz and the 68 material is best suited for applications at
roughly 150 MHz. Figure 2-1 shows various types of rod and bar material in
photographic form.

2. 1. 1 Characteristics of Rods and Bars

Rod material is available in small dimensions for use as core material in

slug-tuned inductors. Although the general rules for application of slugs
and rods are the same, this secti.on will treat the larger rods and bars. Slug-
tuned inductors and transformers will be discussed in detail later in this
volume.
It is important that we keep in mind that the µi rating is that of the
material from which a core is made. The size and shape of the finished prod-
uct dictates the more significant parameter, µe, or effective permeability.
The factor of interest in establishing the /le of a specified core µ is the
diameter-to-length ratio of the rod or bar. Figure 2-2 contains a family of
curves that illustrates this concept. The curves have been drawn for a stan-
dard group of µi factors for ferrite materials. The vertical scale displays the
Ile for each kind of material versus the rod length/rod diameter (lid). It
becomes apparent when examining the curves that the rods or bars with
high initial permeabilities (40 and greater) decrease the most in Ile when the
length is short. It can be seen in Fig. 2-2 that a 0.25-in.-diameter rod, 8 Yi. in.
Jong, would be needed to obtain an effective permeability of 30, even
though the rod material being used exhibited an initial permeability of 40.

Figure 2-1 Examples of ferrite rod and bar materials.

 5000
2000
1,000
1000
700
500
400
300
"0 200
:i
.~ 150
'.Ci 100 100
"'"'
E 70
"'0.
"O
0
40
a:

20

10 10

10 100 1000
Rod length divided by rod diameter, /Id
f igure 2-2 Family of curves that illustrate the relationship between rod length and
diameter to permeability. (Courtesy of Fair-Rite Products Corp.).

On the other hand, if a lowerµ; is selected (10) and the form factor is reduc-
ed to 15 (0.25-in. in diameter, 3. 75 in. in length), the µe remains the same as
theµ;. At the upper end of the initial-permeability range it becomes a prac-
tical impossibility to realize a rod whose µe is comparable to the µ; of its
material. For example, a rod that measured 0.25 x 25 in. (form factor =
100), and whose µi is 5000, would exhibit a µe of only 1800 . Because of this
phenomenon, the designer will find the curves in Fig. 2-2 of considerable
value.
If we were to extend the curves of Fig. 2-2 to the right, it could be seen

41
42 Application of Rods, Bars, and Slugs Chap.2

that there is actually an optimum permeability versus lid ratio. Further in-
creases would lead to a decrease in JJe· The effect sets practical size limits for
magnetic rods that are used in loop antennas. Fortunately, the limits are
workable ones with respect to compact portable radios that contain built-in
antennas.
Polydoroff [1] discusses another design characteristic that must be con-
sidered when using magnetic-rod materials: The effective permeabilities in-
dicated by the curves in Fig. 2-2 are somewhat less than the vertical scale in-
dicates, in a practical situation. This is related directly to the differential
between the mean coil area and the core area. The permeabilities are actual-
ly increased as the core length exceeds that of the coil. The rod areas beyond
the ends of the coil are called free ends. The empirical equation Polydoroff
provides is suitable for a wide group of variations in coil length versus rod
length :

whereµ' is the corrected permeability, JJe the effective permeability from

Fig. 2-2, I, the rod length in inches, and le the coil length in inches. This is
based on Ir being equal to or greater than le. From this it can be seen that a
core which is twice as long as the coil (coil centered on the rod) will provide
an increase in JJe of 26%. If the coil dimensions are kept the same, but if the
rod is made eight times the coil length, the effect will be doubled.
Because of the foregoing it is difficult, if not impossible, to construct a
set of AL factors for rods and bars. This is complicated further by dif-
ferences in inductance characteristics for an equal number of turns being
placed at different points along the rod. The spacing between the turns has a
marked effect on the Q and on the inductance as well. Figure 2-3 shows
three conditions of interest with respect to a ferrite rod with diameter 0.5 in.
and length 7 .5 in. The JJ; of the core material is 125, and the JJe is 60, from
Fig. 2-2. Using the AL equation from Chapter 1, which is given in Fig. 2-3,
various approximate AL factors emerge from the manner in which the coil is
placed on the rod. These approximations hold for a rod of the stated dimen-
sions and JJi· Rods of different dimensional format would result in radically
different AL numbers. The measurements were made by the author while
using a Hewlett-Packard 4342A Q meter. The rod material for these tests
was an Amidon Associates R6l-500X7.

2.1.2 Q and L versus Coil Placement

There are some schools of thought which suggest that a coil placed at
one extreme of a selected rod will yield greater Q than when the coil, pruned
for the same inductance, will exhibit at the physical center of the rod (Fig.
2-3a and b). The graphical exhibit of Fig. 2-5 shows that this is a false
assumption. While using the same rod that served in the tests of Fig. 2-3 , an
 - - -- - - 7. 5 on.- - - - --
28 µH
- ·-==----=----
AL "' 750

- 20 turns
(al

0 _,.,_
20 turns
) 50 µH
AL "' 1250

{bl

0~JIID'.)))'.Jmj!)jjj l~I),.·" _ AL ::: 448

~---- 25 turns - -- - --
4
l "H x 10
AL = 2
N
mm ; 25.4 x in.
(cl
Figure 2-3 Coil placement on a ferrite rod of given µihas a direct bear-
ing on the AL factor.

investigation of the unloaded Q was carried out. Figure 2-5 shows a

dramatic difference in Qu between end placement and center positioning of
the coil. With the coil set to provide 27 µH of inductance at the end of the
rod, the Q is 80 for 20 turns of No. 20 Form var wire, close-wound tightly on
the core. The coil was trimmed to yield 27 µH at the center of the rod, and
the Q rose to 190 for 15 turns.
Another experiment was conducted with the Q meter. The results are
seen in the graphic illustration of Fig. 2-4. A 28-µH inductance was obtain-
ed by spreading 25 turns of No. 22 Teflon-insulated hookup wire across the
entire length of the ferrite rod, as in Fig. 2-3c. The effect shown in Fig. 2-5
was reversed. A Q11 of 240 was obtained under the spread condition,
whereas the Qu declined only slightly to a value of 230 when 15 turns of the
same wire were placed at the center of the rod in close-wound fashion. In
this example the copper conductor was spaced by the thickness of the
T eflon insulation. This accounts for the higher Q reading for the center-
positioned coil respective to the one of Fig. 2-5, which used thin insulation
(Formvar). It can be seen from this that the Q dropped only 4% when the
coil was moved to the center. The Q declined 58%, however, in the tests of

43
 0.5 x 7.5 in. ferri te rod ( R6 1- 500X7}
µ, = 125
250 -
!(test} 2.5 MHz
245 ,_
To ta l rod (spread turns)
240 1--------------.--.--.--.--.--.--.--.--.--.--.--.--.--.--.--.---.
28 µH
235 ,_
Rod center (spaced one wire diameter}
~ 230 1--------------------------------------.
0 28µ H
225 -

220 -

215

210 -
I I I I I I I I t

2 4 6 8 10 12 14 16 18 20 22 24 26
Turns no. 22 Teflon
Fig ure 2-4 Comparison between coil turns and Q for a specified inductance on a
given ferrite rod.

f(tesc) 2.5 MH z
0.5 x 7.5 in. ferrite rod
220 -
µ, = 125
200 - 27 µH

180- Rod center

(close·wound)
160 -

d' 140 I-

120 ,_

Rod end (close·woundl 27 µH

80 - - - - - - - - - - - - - ----, I
60'--~'--
' --'''-----''~---''~---'-~--'--~-'---'---'-~-'-~-'-
1
~'--
1
~'--
1
~'--
1
-
2 4 6 8 10 12 14 16 18 20 22 24 26

Turn s no. 20 Forrnvar

Figure 2-5 The Qu of a 27-uH inductor is 80 when the winding is placed at the end
of a ferrite rod. A Ou of 190 was obtained by placing the 27-uH winding at the center
of the rod .

44
Sec. 2-1 Rods and Bars 45

Fig. 2-5, where the actual conductor turns were closer together and in-
sulated from one another by a material with a lower dielectric constant.
Further investigation was conducted to compare the placements of coils
with identical turns and dimensions, versus end and center location. First, a
one-turn coil was moved from the rod center to the extreme end of the core.
Figure 2-6 displays the results. In this case the Q11 was slightly higher at the
end (155) as opposed to the center (145), a variation of only 6% . The induc-
tance at the center was 8.4% greater than at the end-1.19 and l.09 µH,
respectively. This demonstrates that there is a definite relationship between
µi and the number of turns when attempting to find a workable AL, along
with the factors discussed earlier. If one were to pick an AL value referenced
to a single turn at the center of the rod, a value of 11,900 would result, based
upon AL = LµH N2. The difference between this and the value arrived at
when using 20 turns (Fig. 2-3b) is startling proof of the general prob-
lem.
Figure 2-7 shows the inductance and Q11 of a close-wound 20-turn coil of
No. 20 Formvar wire based on center and end placement on the same test
rod . Moving the coil from center to the rod end caused an inductance
decrease of 46% and a Q11 reduction of 47% .
There is still another consideration with respect to the coil geometry and
the ferrite rod. The inductance for a coil wound tightly on the core will be
considerably higher than when the coil is spaced away from the core. These
effects are highlighted in Figs. 2-8 and 2-9. An investigation was conducted
with a 4-in. ferrite rod, 0.5 in. in diameter, the µi of which was 125. By

f(tesl} = 8.0 MH z
1.4 ~
0.5 x 7.5 in. ferrite rod
µ ; = 125
1.3 -

1.2 Rod center

~Zone 1urn
:r 1. 1 ~
~
..... ------------ Rod end
-- ..,
1.0 I
I
I
0.9 I-
I
I
0.8 I
I
0.7 I
' ' I I
' '
120 125 130 135 140 145 150 155 160
au
Figure 2-6 Graphic illustration of Q versus inductance for
one turn of w ire at the end and center of a ferrite rod.
 f(test) = 2.5 MHz
0. 5 x 7. 5 in. ferr ite rod
60 1--
µi = 125
so i--~~~~~~~~~~~~-
Rod center

40

=i 30 -
-.J 20 turns
(close-wound)
20 ~
Rod
end
10>-

90 100 120 1 10 140 160 180 200

av
Figure 2-7 Comparison between the Q and inducta nce of a
20-turn coil when moved from the center of a rod to the end
of the rod .

L = 10 turns no. 22 Teflon L ID=~ in.

wire, close-wound L length = ~ in.
L~H - w/core = 10.4 µ H, mm = 25.4 x in.
au =350
LpH - no core = 1.8 µH
au =180
Agure 2-8 Effects of a ferrite rod on an air-wound coil
that has an inductance of 1.8 µH . Insertion of the ferrite rod
raises the inducta nce and the Q. The coil is tightly wound on
the rod .

means o f Q-meter evaluation, the arrangement of Fig. 2-8 was investigated.

A close-wound, 10-turn coil of No. 22 Teflon-insulated wire was placed at
the center of the rod, as shown. Because of the wire insulation, the conduc-
tor turns were effectively spaced one wire diameter apart, or nearly so. The
resultant inductance was 10.4 µH and the Qu was 350. The coil inductance
without the magnetic core present was 1.8 µH.
Next, a coil was wound with the same type of wire, close-wound, to a

46
 L

L = 6 turns no. 22 Teflon L ID = ~ in.

wire, close·wound
L length = ~ in.
L,,H - w/ core = 4.6 µH
au = 190
Coil·to·rod spacing =~ in.

L,,1-1 - no core = 1.8 µH,

au = 180
Rgure 2-9 Conditions similar to those of Fig. 2-8 except
that the air-wound coil is not wound tightly on the rod.
Rather, an air gap of 1/8 in. exists between the coil and the
core. The inductance and Q increases, but not as markedly
as when the coil is tight against the core material.

%-in. diameter. It was configured to an inductance of 1.8 µH without the

core inside it. With this inductor located centrally on the rod, spaced Ys in.
from the rod (coaxially), the measured inductance was 4.6 µH-somewhat
less than in the example of Fig. 2-8. Furthermore, the Q11 degraded to 190.
Figure 2-9 shows the placement of the coil, L. In both examples, the initial
1.8-µH coils exhibited an unloaded Q of 180. The coil of Fig. 2-9 required
only six turns of wire for an inductance of 1.8 µH, owing to the larger
diameter. Since the %-in. diameter coil was necessarily shorter in length
than the Yz-in.-diameter one of Fig. 2-8, the l//c effects were present and
must be taken into account, as shown by the Polydoroff equation.
However, for the purpose of comparison the results are ample for
demonstrating the reduction of Q and inductance as the coil is spaced away
from the rod.
As a result of numerous laboratory tests the author concluded that the
best values of Q11 are obtained when the coil turns are spaced one wire
diameter apart rather than being close-wound, when the coil is lumped at
the center of the rod. The graph of Fig. 2-4 shows clearly that spacing the
coil turns over the entire length of the rod achieves excellent Q11
characteristics also, offering slightly higher values of Q than when a shorter
coil with spaced turns is lumped centrally on the rod.
Tests conducted with Litz wire versus solid, enameled magnet wire in-
dicated that the Litz wire provided superior Q 11 values in all instances.
However, there seemed to be little difference between the results when No.
22 stranded wire with Teflon insulation was compared to Litz wire of com-
parable gauge.

47
 2.1.3 Frequency versus Core Material

Narrow-band circuits require inductors or transformers which exhibit

high orders of Q. Therefore, it is imperative to select the proper core
material for the application. It is a fortunate fact that as the operating fre-
quency is increased in a narrow-band circuit, the permeability of the core
material must be lowered. Were this not true, a small inductance, such as I
µH, might require less than one wire turn on the core used, assuming that a
relatively high µe predominated. The consequences would be degraded Q
and impractical coil dimensions.
The notable exception to the foregoing discussion of core selection ver-
sus operating frequency is in the case of broadband transformers. In that
type of application it is quite desirable to use high-µ cores, because as the
operating frequency is increased, the core permeability effectively "disap-
pears," whereas at the lower end of the chosen frequency spread the core µe
ensures ample inductance . This subject will be treated in greater depth when
we examine broadband transformer concepts. In this chapter we will ad-
dress the subject of narrow-band circuits for the most part.
Some of the significant properties of various ferrite core materials are
listed in tabular form in Table 2-1. Of particular significance to this discus-
sion is the column that lists the recommended frequency ranges of the
materials. The permeabilities contained in the table are standard ones of-
fered by Fair-Rite Products Corp. and are similar to those of other
manufacturers. It can be seen that as the recommended operating range is
elevated, the µi of the core material becomes lower. This illustrates why inat-
tention to core characteristics can lead to unworkable Q values. These same
general rules apply to powdered-iron cores also.

TABLE 2 -1 Ferrite-core characteristics

/11itial 1\,faxi11111111 Sat11ra1io11 Flux Reco111111e11ded

Permeability. Permeability. Density, Bs· at Frequency Fair-Rite
"i 13 Oe Range 0 ( MHz ) Material
""'
20 2000at400e so.100 68
40 3000at200e 10.so 67
100 370 2150 5-15 65
125 450 2350 0.2-10 61
175 400 2550 0.1-5 62
250 375 2200 0.05-4 64
300 3600 3900 0.001-5 83
800 1380 2500 0.01-1 33
850 3000 2750 0.01-1 43
1200 2500 3400 0.01-1 34
1800 4600 1150 0.001·2 77
2000 3500 3500 0.001-1 72
Courtesy of F:1 ir-Rite Products Corr.
nFrcquency ra tings for optimum Qin narrow·bnnd tuned circuits.

48
2.2 Rods and Bars as Antennas

Ferrite rods and bars are used in numerous loop-antenna applications such
as broadcast-band receivers, low- and medium-frequency direction-finder
receivers, in the receivers used for airborne navigation, and even in pocket-
size VHF receivers. A closed-core magnetic device cannot serve the same
purpose as the rod or bar when used as a loop-antenna core. This is because
the toroidal configuration is self-shielding, thereby preventing the desired
signal pickup. Similarly , there is a degrading effect in rod-antenna efficien-
cy when the length versus diameter (lid) is carried beyond approximately
35. As the length is increased past that point, the "toroidal effect" becomes
more and more pronounced. In this context an infinitely long ferrite or
powdered-iron rod becomes equivalent to a closed circuit at full permeability.
The electrical parameters of a magnetic-core loop antenna are roughly
equivalent to those of the early-day air-core loop or "frame" antenna seen
in Fig. 2-10, which was developed by Hertz in 1888. It was known then as
the "loop collector." Various forms of this type of antenna have been used
since for reception of broadcast-band signals, navigational signals aboard
vessels and aircraft, and amateur-radio signals. The air-core loop can be
used as an open system or as an electrostatically shielded device, as seen in
Fig. 2-11. In the latter form, the loop discriminates various types of noise
and corona static. Ferrite-rod loops can also be equipped for electrostatic
shielding. This subject will be covered later in the chapter.
Loop antennas, whether the air-core or magnetic-core variety, have a
height factor, specified as h. This is called the effective height, he. This is
expressed in meters as a factor, which, multiplied by the field strength in µV
per meter provides the loop-induced voltage in µV:

where d/A<l, N is the number or loop turns, rr = 3 .14 and A is the area of
each turn in square meters. The effective height obtained from the equation
is for signals with horizontal polarity impinging on a loop that is orientated
for maximum response.
The magnetic-core loop exhibits an interesting and superior trait when
compared to an air-core loop: The he increases as a function of 1-'e· Thus,

he = 2rrNAµe
A
where, for practical considerations, A is in centimeters and A is in cm 2 •
During laboratory and outdoor test-site investigations, the author
observed that signal pickup over an efficient air-core loop is somewhat less

49
 Balanced air-core
loop antenna

1on

0.1
100 n µF 100 .Q

Tl====
so n o-----------fr)
to ..
receiver
J_ 1 0.1 µF
(a)

x 0- ·- ->./INv- - --0 y
RL

(b)

To preampli fier

{c)

Figure 2-10 Example of an air-core loop or frame antenna {a). A

balanced preamplifier will ensure minimum pattern distortion. Ru as
seen by the preamplifier, is depicted in lb). It is normally on the
order of 1 n or less. The feed system in (c) is sometimes used, but
imbalance will cause pattern skewing and poor nulling.

50
 Small shield
gap

Balanced
shielded
loop

Ground
plane

To
balanced
preamplifier
Figura 2-11 Format for a bal-
anced, magnetically shielded loop
antenna.

than µe with very short rods (less than lid = 5). But, when the ratio is
greater than 5, the full benefits of µe are realized.
Tests were conducted with lumped turns at the center of rods, at the
ends of the cylindrical cores, and with the coil turns spread uniformly over
the entire rod. Irrespective of the resultant Q values obtained from the three
formats, the greatest he was noted when the turns occupied the entire core
length. This is because the maximum number of turns was necessary to
secure the required inductance, as seen in Fig . 2-4. These results are com-
patible with the equations for he, where a significant term is N. These
evaluations were performed with cylindrical ferrite cores which had an lid
ratio of 30 (0.5 x 15-in. rod), which was obtained by joining two 7.5-in.
rods, end to end, by means of epoxy cement. The JA; of the material was
125.
The test method of Fig. 2-12 was employed during the tests. The signal
source was a commercial AM broadcast station, located I mile from the test
site. No conductive objects were closer than ~ mile from the loop under
test with respect to the signal path. The loop was oriented for maximum
figure-eight response for each test.
An air-core frame loop of comparable a rea to the 0.5 x 15-in . rod loop
was fabricated and tested as in Fig. 2-12. The magnetic-core loop with
spread turns exhibited a gain of approximately 50 over the air-core version.
However, when the rod loop was compared to a free-standing vertical
antenna with buried radials, the loop was some 30 dB less effective when
oriented for the same polarity. The vertical was Ya wavelength high but ad-

51
 Loop
antenna ,.- - -7
- 40 dB max. + 30 dB
Reciever I I
I I
50 n --' I
I
Step
I
atten. -1
•••• _ ·1
_/

Figure 2·12 Test method used by the author during evaluation of a ferrite rod antenna.

justed electrically to \14 wavelength by means of top loading. The feed point
was matched for 50 Q at the operating frequency.
A transmitting loop can be used as a signal source for more precise
testing. In a n exercise of this type, current in the tra nsmitting loop is im-
posed by a single generator. For spacings of A/2 or less, the field strength is
µ V per meter at the test loop is

l 8.85Nr 2J
E= xi
where r is the transmitting loop radius, N the number of turns, I the current
in mA, a nd X the spacing between the two loop antennas in meters, as
discussed by Polydoroff[ l ]. He adds that the results can be expressed alter-
natively by

l ,180Nr2e
E=
X 1JL
where with regard to the transmitting loop, e is the voltage across the loop
in µV ,/the frequency, and L the inductance of the transmitting loop in µH.
In this case E is the voltage across the receiving loop, as in the previous
equation.

2.2.1 Loop Pattern Symmetry

The primary purpose of loop-antenna utilization, a part from effecting a

physically small component, is to take advantage of the nulls in response
which exist off the sides of the classic figu re-eight response pattern. The
characteristic minima and maxima are useful in radio direction-finding ap-
plications. T his need d ictates a requirement that the pattern be as sym-
metrical as possible (Fig. 2-13a). In this context an ideal response produces
a pair of zero points that are precisely 180° apart. If the disturbances in
symmetry are only minor, a pattern d istortion of many degrees can result
(Fig. 2- 13b).

52
 2rrNAE
A. l oop
pattern
2rrNAE .
e• =A.- - cos() sin wt

{a) (b)

Figure 2-13 Classic response pattern of a properly balanced loop antenna (a). In (b) the pat-
tern can be distorted many degrees by only a slight imbalance in the system.

An ideal loop should serve entirely as the antenna, and in accordance

with its intended characteristics. Imbalance in the physical layout (excessive
asymmetry, long connecting leads, etc.) can lead to what is known as anten-
na effect. In such an example the conductors apart from the main body of
the loop will act as unwanted parts of the antenna and will respond to in-
coming energy differently than the loop will. This is represented by the cir-
cle with dots in Fig. 2-13b. Other factors can degrade the ideal response pat-
tern, such as displacement of current. This can result from excessive axial
length in a loop antenna. It is important, therefore, that close attention be
paid to the physical attributes of a loop antenna. Also, balanced loops will
yield superior results to those which do not feature electrical balance (Figs.
2-10 and 2-11 ). Electrostatic shielding of the type illustrated in Fig. 2-11 is
extremely helpful in reducing the antenna effect. The major limitation in us-
ing shielded loops is that the loop must be physically close to its ground
plane for best results. Electrostatic shielding of a rod or bar type of loop can
be carried out by locating the loop at the physical center of a U-shaped
channel of aluminum. This is shown in Fig. 2-14. Most portable direction-
finder receivers utilize this method. Test-range results indicate no degrada-
tion in loop efficiency when the shield is added, provided that the top open-
ing of the U channel is kept entirely open.
2.2.2 Noise Immunity

Loop antennas are considerably better than vertical antennas with

respect to noise immunity. Weak signals that are buried in the noise to
which a vertical antenna is highly responsive are not capable of conveying

53
 Rod
loop

Aluminum
U-channel
Figure 2-14 Method for providing magnetic
shielding of a ferrite-rod loop antenna. An
aluminum U-channel is employed.

intelligence reliably. T he nigh-Q properties of a well-designed loop antenna

(narrow bandwidth) are not as responsive to noise as is true of low-Q loops
or other types of antennas. Furthermore, the loop nulls can be used to ad-
vantage in reject noise from a specific direction. Figure 2-15 gives a com-
parison in signal-to-noise ratio between a low-Q loop and one that has high
Q. The advantage of designing for high Q is readily apparent from examina-
tion of the two response curves. The trade-off is, of course, in bandwidth.
A fixed-tuned loop would not be especially useful if a frequency excursion
beyond/ and.f were required. In situations where the loop is continuously
tunable over the desired frequency range, or where it can be pretuned to a
separate frequency and left in that state, undercoupling to the load is
recommended in the interest of improved noise ratio. A low-noise
preamplifier is a simple and economical way to compensate for the loss
brought about by light load coupling. The higher QL (loaded Q) can be very
advantageous.

2.3 Practical Loop Circuits

Some assembled magnetic-core loops are shown in Fig. 2- 16. Most of the
commercially built loops employ Litz wire for the purpose of elevating the
circuit Q as much as possible. However, the mere use of that preferred type
of wire does not ensure optimum Q. The results of using magnet wire can be
seen by examining Fig. 2-17, where Qu (unloaded Q) measurements were
made on a short bar type of loop which was manufactured in the Orient. A
sample AL calculation indicates the!"; of the material Lo be on the order of
250, which may account in part for the midrange Q value. The winding con-
sists of small-diameter (No. 28 enameled) wire. Although the Qu is not spec-
tacular, it is adequate for most pocket-size AM radios-the type from
which the loop was taken . Most small AM radio loops are overcoupled to
the load (the mixer stage) to minimize loss. Were an RF amplifier employed,
lighter coupling would be practical, and the QL would be much higher. The
benefit of higher Q would be seen in improved receiver front-end selectivity

54
 fo
+ 50 f
OL = -o-
'1 - f2

I
CD 1
"O
- '-
Noise

0 I -t
Figure 2-15 Comparison of signal-to-noise ratio for a low-Q and
a high-Q loop antenna. The trade-off is in bandwidth.

D llWlllll11111llll1l1111JJJ11JlJJJJJJllJJJfil._
\

Figure 2-16 Assembled magnetic-core loop antennas .

. . . an aid in separating strong broadcasting stations in metropolitan loca-

tions.
The turns of section Bin Fig. 2- 17 are spaced across roughly two-thirds
of the bar, with the turns of section A being close-wound. This is a common
scheme in foreign-made radios, presumably to allow section A co be used
for final inductance trimming by spreading the turns slightly. The tap point
between sections A and B is for connection to the input of a bipolar tran-
sistor mixer. The characteristic impedance of a mixer of that type is a pprox-
imately 600 Q at the base element. It can be seen from the tabular data in
Fig. 2-17 chat despite the LC ratio of the loop at any frequency in the broad-
cast band, the Qe remains relatively constant. T his enables the designers lo
use miniature variable capacitors in some radios, which have as little as 150
pF of maxim um capacitance. The increase in the required inductance under
those conditions makes ic necessary to use more turns of wire in the loop.
The advantage of this was discussed in Sec. 2.2. It is likely that the loop of
Fig. 2-17 would have exhibited somewhat higher Q if Litz wire had been used

55
 -----2 ~ in. - - - - -<....

-+
~ in.
__L'--+'"'41<--+-.._._._...........U-L-J.............-W-L.U...,.--YI
f-s-·~-
-1I- A i3G in.

A = 15 turns No. 28 enam., mm = 25.4 x in.

close· wound
L = 230µH
8 = 55 turns No. 28 enam..
spaced one wire diameter J.I, = 250
between turns µ. < 250

Frequency
(kH z) Ou CPF

550 (kHz) 180 360

800 (kHz ) 175 175
1000 (kHz ) 180 110
1600 (kHz) 180 40

Figure 2-17 Physical details and test resul ts for a bar

type of ferrite loop used in an AM broadcast-band
receiver. Antennas of this type are found in pocket-
size radios built in the Orient.

instead of No. 28 magnet wire. But, when pocket-size radios are built for
retail at less than $10, a trade-off between performance and economy is the
rule more often than not.

2.3. 1 Loop-to-Mixer Circuits

An example of how a ferrite-rod loop can be used directly into the mixer
stage of a broadcast-band receiver is given schematically in Fig. 2-18. The
rod characteristics and winding information for this example and the one in
Fig. 2-19 are supplied in Fig. 2-20. The circuit of Fig. 2-18 shows the loop
with a tap near one end. The tap is set fo r approximately 600 Q to match the
characteristic input impedance of the bipolar transistor mixer, QI . With the
specified loop inductance, 240 µH, the entire broadcast band can be covered
by means of a single-sectio n 365-pF variable capacitor. In a typical design
such as one finds in low-cost AM radios, the benefits of the high Q11 are
negated to a greater extent by the tight coupling to the mixer.

56
 550- 1600 kH z

Mixer
; 365
PF 2N2222A
To
Ll
z ~ 600 S2 C1-F ampli fier
Ant.
240 µ H
Ou = 350 5600 !"2
0. 1 µF
1
To Lo

' - - - -- - -- - . - -- --0 +Vee

Agure 2-18 Typical example of a ferrite-rod antenna as used with the mixer of an AM broad·
cast receiver. The mixer base is tapped at approximately 600 n on the rod antenna.

An alternative coupling method between the loop and QI would be to

employ a secondary winding, or link, in place of the tap on LI. If a QL ver-
sus gain trade-off can be used without too much sacrifice in receiver sen-
sitivity, the coupling link can be made small or the tap on LI can be located
closer to the grounded end of the loop.
A method for using a loop antenna with a dual-gate MOSFET mixer is
provided at Fig. 2-19. A 3N21 l, 40673, or any of the similar high-frequency
MOSFETs would be suitable in this circuit. The signal gate is tapped down
partway on the loop, LI, to enhance circuit stability. This technique lowers
the input seen by the gate impedance of the FET progressively as the tap is
moved toward the ground end of LI. The most likely manifestation of mix-
er self-oscillation is experienced when the receiver is tuned to the low-
frequency end of the broadcast band. This places the FET Input at a fre-
quency near the 455-kHz IF, thereby encouraging tuned gate/tuned drain
oscillation. Strapping gate l nearer to the cold end of LI eliminates the
problem. Since gate 1 of the FET is greater than 1 MQ in impedance, the
mixer does not degrade the Q11 of the loop significantly. This means that a
loop used with a FET mixer is capable of providing much greater receiver
front-end selectivity than would be likely with the circuit in Fig. 2-18. Fur-
thermore, the FET mixer is capable of much better front-end dynamic
range-an objective of many receiver designers.

2.3.2 Loop to RF Amplifier

It is a simple matter to resolve the front-end selectivity problem which is

so common in low-cost broadcast-band receivers. For a slight cost increase,
the designer can incorporate a low-noise RF amplifier ahead of the mixer,
as illustrated in Fig. 2-20. This stage can be used to compensate for the loss

57
 550- 1600 kHz

365
PF 455 kHz
Mixer
Tl
L1
ant.
240µH
a. = 350
100kf1
0.01
µF
-:-

I 390 !1 0. 1 µF -=-

+Voo

To
Lo
(3 V peak to peak)
Figure 2-19 Technique for using a M OSFET mixer with a rod antenna. The FET
gate impedance is on the order of 1M nor greater. This permits placement of the
tap near the high-impedance end of the rod antenna.

that results from undercoupling the loop to its load. This allows L2 to be set
for the desired QL or loop BW L (loaded bandwidth).
Transistor Ql in this circuit is configured as a fed-bac k, broadband
amplifier. The characteristic input and output impedance of the stage is ap-
proximately 50 Q. A combination of degenerative feedback (RI) and shunt
negative feedback (R2) ensure unconditional stability of Ql, even when very
light coupling is used between the loop and the amplifier. A 4: I broadband
toroidal transformer, Tl, steps the 200-Q down to 50 Q. Tl contains 20 bifilar
turns of No. 28 enameled wire on a 0.5-in.-diameter ferrite toroid core. The
µi of the core is 950. A 0 ..37-in.-diameter core of the same µ would be
suitable if smaller-gauge wire were employed. The 2N5179 transistor was
chosen for its high f T rating (> 1000 MHz) and low-noise characteristic.
Typical gain is 15 dB. Another advantage of this type of circuit is that the
50-Q output of the RF amplifier is compatible with diode-ring doubly
balanced mixers, the ports of which are also 50-Q. The gain of Q 1 would en-
sure adequate receiver noise figure if a DBM (double balanced mixer) were
used. Such a mixer would greatly elevate the receiver dynamic range. T he
port-to-port isolation of the diode-ring mixer, plus the inclusion of RF stage
Ql, can virtually eliminate unwanted local-oscillator radiation via the
antenna.

58
 + 12 v

RF amp.
0.1 µF
0.55- 1.6 MHz

3.3kQ To
~----,.----' 11nn-_._--1 !---------o mix er
01 0.1µF Z " 50H
2N5179

• - Indicates
same phase
0. 1 µF
Gain~ 15dB

L2 : see text.

LI : 73 turns No. 22 Teflon

wi re, spaced over
7.5 x 0. 5 in. rod .
au ; 350, µ , = 125.
µ , = 60, L = 240 µH

Figure 2·20 Example of a fed -back bipolar-transistor RF amplifier as used with a rod
loop antenna. Shunt and degenera tive feedback are used to ensure unconditional
stability.

2.3.3 Loop with Sense Antenna

A practical circuit for a ferrite-rod loop antenna used in combination
with a sense antenna is shown in Fig. 2-21. An arrangement of this variety is
useful when a cardiod response pattern is desired. A relatively unidirec-
tional pattern, which is heart-shaped, as seen in Fig. 2-21 b, is needed in
direction-finding equipment to avoid ambiguity of compass headings. A
bidirectional pattern can confuse an operator, especially if the operator is
past the radiobeacon transmitter being sought for navigational purposes.
The sense antenna consists of a short vertical element placed close to the
magnetic-core loop-typically from 6 to 36 in. in front of the rod or bar
loop. The sense-antenna voltage is combined and placed in phase with the
loop voltage by means of R 1 of Fig. 2-2l a. The in-phase relationship results
in the cardiod pattern in Fig. 2-2lb. The factor k is equal to the fractional

59
 _,.....- Sense .,, _ _>. _ ft
JI' ant. 0.0032

Phasing
2200 kHz Rl
Loop ant. 10 k.11 2200 kH z
2200 kH z

100 p F
0. 1
100 .n µF
10 kQ
Cl: dual 100 p F ~---o + 12V

C2: 1000 pF max.

C3, C4 : 150 pF max . L 1: 105 µH. 48 turns No. 22
0 1: U310, 2N44 16. MPF 102 Teflon or Lit z wire spread
over 0.5 x 7.5 1n. ferrite
rod (µ , ~ 1251
L4. L5: 55 µH toroi dal inductor
L2 : 5 turn s. L3, L6: 3 turns

(al

Max.

Cardoid response

(bl
Figure 2-21 Circuit of a ferrite-rod loop antenna with sense antenna and preamplifier (al . The
resultant cardioid response pattern is shown in (b) .

60
Sec. 2-3 Practical Loop Circuits 61

relationship of the sense antenna to loop maximum voltage. A cardioid

response results when k = 1. If k were changed to 0.5, a small back lobe
would protrude. By reducing k to 0.2, a figure-eight pattern would develop
as the loop degraded toward the classic bidirectional characteristic found
when no sense antenna is used.
A backside null depth as great as 40 dB has been measured by the author
on a test range while using the circuit of Fig. 2-21a. This was realized only
after repeated adjustment of Rt and Cl for the best null depth: There is
considerable interaction between the two adjustments. Null depths that
were measured off the ends of the rod loop without the sense-antenna cir-
cuitry were on the order of 23 dB. In theory, and with perfect loop balance,
the two nulls could exceed 30 dB. This was explained to the writer by Henry
Jasik during a discussion on loop antennas in 1975 [2].
Loop balance in the circuit given in fig. 2-21a is provided by using a split-
stator tuning capacitor, ClA and CIB. C2 is used to adjust the coupling
between the loop and the load, consistent with the desired QL character-
istic. The circuit from L2 to L6 is isolated from the loop circuit by means of
shielding to prevent stray coupling between the circuits. The length of the
sense antenna may have to be adjusted empirically to obtain the antenna
voltage required. For the frequency specified, an 18-in. length of conductor
proved entirely adequate.
A low-noise post-loop amplifier, Q 1, is used to ensure an acceptable
noise figure. Degenerative feedback is made available by the unbypassed
resistor, R2. This, plus tapping the FET drain down on L5, prevents self-
oscillation of the RF preamplifier. These measures lower the stage gain
from a possible 20 dB to roughly 10 dB. The loop can be shielded elec-
trostatically if desired. The details were given in Fig. 2-14.
2.3.4 Loop Variations

When it becomes necessary to make a loop antenna variable in terms of

inductance, two common approaches can be taken. These are seen in Fig.
2-22. The need for varying the inductance is encountered when the RF
amplifier or mixer stage of a receiver must track with the local oscillator.
The employment of a two-section variable capacitor (ganged) makes this
almost mandatory.
Figure 2-22a shows a compromise type of loop antenna, used primarily
as a replacement unit for AM radio loops that have become damaged.
This style of adjustable loop is short-usually 2 to 3 in . in length. A
magnetic-core slug is adjustable from one end to obtain the required induc-
tance. The Q of this antenna is relatively low compared to long rod or bar
loops ... in the area of 150. To enhance the pickup capability of the loop, a
short length of wire (1 to 3 ft long) is sometimes connected to the unground-
ed end of the coil and allowed to dangle outside the receiver cabinet. The
 Coil form Slug adj.

\
Mounting
Layer· wound clip
co il Solder
lug

(al

~)r,,
(l

Ferrite
rod
Movable coil on
low·loss lorm

(b )

Figure 2-22 Small ferrite loop antenna (adjustable induc-

tance) is seen in (a). A long movable ferrite rod is contained in-
side a layer-wound coil. The coil inductance in (bl can be ad-
justed by sliding it along the ferrite rod.

directional characteristics of this type of loop are practically useless due to

physical and electrical asymmetry.
When space permits, the technique shown in Fig. 2-22b is superior to
that in (a). The loop coil is wound on a movable insulating form. Tracking
is achieved by moving the coil left or right until the desired inductance
results. The coil is then cemented in place. T he lumped inductance is not as
effective as that which is spread over the entire length of the ferrite rod , but
acceptable results can be had with this general scheme.

2.4 Rod -Core Transformers

Although ferrite rods are used primarily as cores for loop antennas, they are
employed sometimes in broadband transformers. More typically, however,
toroids and pot cores are chosen as the core material for narrow- and broad-
band transformers.
The major disadvantage in using a solenoidal winding on a rod core is
that the inductor or transformer Jacks the self-shielding characteristic of the
toroid or pot-core equivalent. Apart from that shortcoming, the rod is an
entirely suitable magnetic-core material. It can be used as the foundation
element in narrow- or broadband circuits. Some manufacturers prefer the

62
Sec. 2-4 Rod-Core Transformers 63

rod format over that of the toroid because of reduced winding complica-
tions, especially if a toroid-winder machine is not available.
Figure 2-23 portrays in pictorial and schematic form a broadband 4: 1
transformer as it would be configured on a ferrite rod. The rules given in the
chapter on transformer design apply in this example, even though a closed
core is not utilized. Although the wire pair is shown in a parallel placement
format, some designers prefer to twist the wires together before laying them
on the core. Normally, 6 to IO twists per inch is the common rule.

2.5 Slug Applications

Slugs are similar in structure to magnetic-core rods. The essential difference

is that they are smaller in size than loop-antenna rods. Also, they are
manufactured with some type of physical adjustment scheme in mind.
Slugs are used as movable elements inside coil forms that are made of

Tl

c A 0
R,"

{a)

R,,, 200 n

4: 1
Tl
• - Phasing

{b )
Figure 2-23 Pictorial (a) and schematic
(bl illustrations of a ferrite rod as used in a
4:1 broadband transformer.
64 Application of Rods, Bars, and Slugs Chap. 2

ceramic, steatite, phenolic, plastic, or paper. The composite assembly is

called a slug-tuned inductor or transformer, once the winding is in place.
The slug material can be ferrite, powdered iron, or even brass. The latter is
often used at VHF and UHF to effect changes in inductance. Brass slugs
have the opposite effect of ferrite or iron cores. Brass decreases the coil in-
ductance as it is moved deeper into the coil. The opposite is true of the other
two materials.
Figure 2-24 shows an assortment of manufactured slug-tuned inductors
in photographic form. It can be seen that they are available in a host of sizes
and shapes. Some are intended for mounting on printed-circuit boards,
while others can be affixed to metal chassis or panels.
Two of the most common formats for slug-tuned coils are illustrated in
Fig. 2-25. The slug in (a) is for use with the coil form in (c). A screw is
embedded into the core material during the manu facturi ng process. The end
of the screw is slotted, as is the bottom end of the slug. This permits adjust-
ment from either end of the coil form. This style of slug-tuned coil is used
primarily for direct mounting to the chassis or panel.
The slug in (b) and the coil form in (d) of Fig. 2-25 are combined to for m
a printed-circuit-mount variable inductor. The slug is molded with threads
on the outside surface, as shown. The inner wall of the coil form has mating
threads, thereby providing for slug adjustment in and out of the coil. The
slug is either slotted or fabricated with a hex hole completely through its
center. The hex or the slot are used for adjusting the slug positioning. If fur-
ther adjustment is not contemplated, wax can be melted into the top or bot-
tom of the coil form to affix the slug so that vibration does not change the
setting. A jam nut or locking spring can be used with the coil form in Fig.
2-25a and c for the same purpose.

I
-
-~

F1gure 2-24 Assortm ent of typical slug-tuned inductors.

 - Threaded
shank

Threaded

Magnetic
core
slug

Sl ot
or
Adjust. slot hex hole

(a) (b)

,--- inserl
Threaded , slu g
mounting •I
post Coil
I
' I

Coil
Metal
clip ~
I!I 11
~
iI
I I
I

Low-loss form \ //
Sol der
lugs

(cl {d)

Figure 2-25 Breakdown views of two popular slug-tuned inductors. The

slug in (a) mates with the coil form in (c). Similarly, the treaded slug in (b)
mates with the printed-circuit style of coil form in (d).

2.5.1 Performance Considerations

Slug-tuned inductors or transformers are almost without exception

employed in narrow-band circuits. This dictates a need for specific criteria,
notably electrical stability, mechanical stability, and circuit Q. A given
resonator, once adjusted, should remain in the intended state within
reasonable constrictions. T his is irrespective of temperature changes and
vibrations.

65
66 Application of Rods, Bars, and Slugs Chap. 2

Mechanical stability requires that the coil winding be held firmly in place
on the coil form with low-loss coil dope, such as polystyrene cement. The
slug mechanism should be tight in its threaded collet or coil form.
The core material must be selected correctly for the chosen operating
frequency. Too low aµ will require excessive coil turns to realize a given in-
ductance. The greater the number of turns, the higher the de resistance and
the lower the Q. Too high a core permeability will result in a small number
of turns but will cause a degraded Q characteristic. Thus, it is prudent to ex-
amine the core characteristics versus recommended operating frequency
before committing a particular core material to the circuit. Table 2-2 lists
the recommended operating frequencies for most of the popular core

TABLE 2-2 Screw-shaft slug characteristics.

Reco111111e11ded
Freqrie11cy Core Basic Color
M/-/z Material Powder Code

0.2· 1.5 Powdered iron Carbonyl C Yellow

1.0· 20.0 Powdered iron Carbon yl E Red
50.0-200.0 Powdered iron !RN8 White
100.0 and up Brass None

2.0· 40.0 Powdered iron Carbonyl Tl·! Purple

40.0-300.0 Powdered iron Carbonyl SF Blue
0.2· 1.5 Powdered iron Carbonyl C Yellow
1.0- 20.0 Powdered iron Carbonyl E Red
20.0- 50.0 Powdered iro n Carbonyl J G reen
50.0-200.0 Powdered iron IRN8 White
100.0 and up Brass None

2.0· 40.0 Powdered iron Carbonyl TH Purple

40.0-300.0 Powdered iron Carbonyl SF Blue
0.2· 1.5 Powdered iron Carbonyl C Yellow
1.0- 20.0 Powdered iron Carbonyl E Red
20.0. 50.0 Powdered iro n Carbonyl J Green
50.0-200.0 Powdered iron IRN 8 White
100.0 and up Brass No ne

2.0- 40.0 Powdered iron Carbon vl TH Purple

40.0-300.0 Powdered iron Carbonvl SF Blue
0.2- 1.5 Powdered iron Carbonyl C Yellow
1.0- 30.0 Powdered iron Carbonyl E Red
20.0- 50.0 Powdered iron Carbonyl J Green
50.0-200.0 Powdered iron !RNB White
100.0 and up Brass None

2.0· 40.0 Powdered iron Carbonyl TH Purple

40.0-300.0 Powdered iron Carbonyl SF Blue

Courtesy of J.W. Miller Co .• Division of Bell Industries.

Sec. 2-5 Slug Applications 67

materials available for slugs of the kind shown in Fig. 2-25a, as provided by
J. W. Miller, a Division of Bell Industries. Table 2-3 gives the same general
data for slugs of the type depicted in Fig. 2-25b. The core materials are
brass or powdered iron. The recommended frequency ranges will ensure op-
timum Q. The Miller color code for slugs is included in the tables. The
permeability factors are not listed in the tabular presentations, as theµ will
depend on the core dimensions. The manufacturers of slugs can provide
speci ficµ information versus the various slugs, with respect to size.
TABLE 2-3 Threaded-slug characteristi cs.

Reco1111m.'1uled
Freque11c.1• Core Basic Color
MH= M01erial Powder Code

0.2- 1.5 Powdered iron Carbonyl C Yellow

1.0- 20.0 Powdered iron Carbo nyl E Red
20.0- 50.0 Powdered iron Carbonyl J G reen
50.0-200.0 Powdered iro n IRN 8 Wh ite
2.0- 40.0 Powdered iron Carbonyl TH Purple
40.0-JOO.O Powdered iron Carbonyl SF Blue
O.J- 1.0 Powdered iron IRN2 Yellow
.05- 5.0 Powdered iron Carbonyl E Red
4.5- 20.0 Powdered iro n Ca rbonyl TH Green
20.0 and up Powdered iro n Ca rbonyl J White

100.0 and up Brass None

0.3- 1.0 Powdered iron IRN 2 Yel low

0.5- 5.0 Powdered iron Carbonyl E Red
4.5- 20.0 Powdered iron Carbonyl T H Green
20.0 and up Powdered iron Carbonyl J White

100.0 Brass None

0.2- 1.5 Powdered iron Carbonyl C Yellow

1.0- 20.0 Powdered iron Carbonyl E Red
20.0- 50.0 Powdered iron Carbonyl J Green
50.0-200.0 Powdered iron IRN8 White
100.0 and up Brass None

2.0- 40.0 Powdered iron Carbo nyl T H Purpl e

40.0-300.0 Powdered iron Carbonyl SF Blue

Courlcsy of J.W. Miller Co.. Division of Bell Industries.

It is important to consider the effect of the slug placement within a coil

on the Q of the inductor. When the slug enters the coil completely, the Q is
the highest within the recommended frequency range. Table 2-4 shows the
results of laboratory tests made on two series 42 J. W. Miller coils. One is
the µH range and the other is in the mH range. In both cases the Q drops as
the slug is withdrawn from the coil. This effect clearly demonstrates the
68 Application of Rods, Bars, and Slugs Chap. 2

need to choose a coil which has the required Qu (determined by the

designer) at a specified value of inductance.
TABLE 2-4 Results of laboratory tests on J.W. Miller coils.

DC Test
S lug QI/a Core R esistance Frequency
Position L Material (fl ) (MHz )

Full 5.8µ H 155} Curbo nal

Half 3.8 µ H 130 E 0.466 8
Out 2.4 µ H 100

Full 5.8 m H Carbonyl

Half 3.8 mH 85}
75 c 21 2.5
Out 3.4 m H 70
3 Compariso n between Q 11 3nd positi on of the coil slug, indicating max imum Q when
the slug is full y into the coil.

2.5.2 Practical Circuits

We have discussed the importance of Q when a slug-tuned inductor is

used in a narrow-band circuit. Another significant point of concern is
stability versus temperature. The effects of long-term drift are particularly
annoying in most free-run ning oscillators of the type shown in Fig. 2-26a.
Typically, drifts in frequency which exceed, say, 100 Hz are unacceptable
by present-day performance standards. In stringent applications the drift
cannot be permitted to go beyond a few hertz.
In an ideal situation the designer would not use an inductor for frequen-
cy control if that inductor contained a magnetic core. This is because t he
core permeability changes with temperature. Alternatively, the designer
would attempt to utilize a mechanically rigid coil which was air wound or
contained on a low-loss form such as ceramic. The shortcoming of this ap-
proach is brought on by today's emphasis on miniaturization. A good air-
wound, stable inductor for the high-frequency spectrum and the lower
would occupy considerably more volume than a magnetic-core component
which provided equivalent inductance.
We may ask: What alternatives are there with regard to stability? The
classic procedure when magnetic cores are contained in the oscillator tank
coil is to use temperature-compensating devices to correct for long-term
drift. Most core materials have a positive drift characteristic-increasing µe
with an increase in temperature. As this happens, the Q degrades somewhat.
Carbonyl E powdered-iron mix, for example, will have a µe change of
roughly +0.0050Jo/ °C. The Q will degrade by approximately -0.02% /°C.
The change in permeability with heat can be corrected by using fixed-value
capacitors with negative-coefficient properties.
 Osc.
1son + 12 v
1005-2055 kHz 2N4416
, --1 820 pF 01 02
v
8.2
400MW
I
C4 I0.1 µF -=-
/
Shield I
' L1
I
C5
0.002 µF 50 pF To

_Lt __ J 100 k
01

C6
{---o
C?

RF Cl
buffer

0.002 µF 1 mH

27 pF L1 : 110µH nom. (core, Carbonyl E).

Cl C2 C3
NPO
.
T uning
l 30
pF
J. W. Miller 43A 104CBI
Omin = 150 at 2.5 MHz
365 pF

(a)

- ---------- . ,
I
I
I
.L c'" 101 i
Cma. = 214pF
L1
T ~ 6 pF
Cm,n = 55 pF I
I
Assume C2 at I
midrange ( 15 pF)

C1 C2 C3

C=
r -1 -- - - -- -- - - -- -
+ - -- -- - + - - - --
C4 Cl + C2 + C3
C;" + 1 1
- +-
cs C6
(b)

Figure 2-28 Design example for a VFO that contains a slug-tuned inductor (Ll ) in the
frequency-determining part of the circuit. Excellent stability can be obtained with this circuit by
using negative-temperature coefficient capacitors at C3 through C7 inclusive. The capa citors
compensate for the positive drih of the core at L1 (a). The effective capaci tance in parallel with
L1 is demonstrated in (bl.

69
70 Application of Rods, Bars, and Slugs Chap.2

Drift is not a primary consideration when the equipment remains in a

relatively stable temperature environment, where room temperature is fairly
constant, and where the ambient temperature within the equipment case
undergoes minor changes. But when heat from tubes, power transistors,
and transformers is present in the region of the oscillator, long-term drift
will become manifest for certain.
A partial solution to the drift problem is to select a slug-tuned inductor
(Ll of Fig. 2-26) which has a low-µ core. Also, the coil should be set for the
desired inductance with the least amount of core entry into the coil. This
condition will not yield the maximum resonator Q, but, as shown in Table
2-4, the Q degradation is not severe. The principal virtue of high Q in an
oscillator tank is reduced noise bandwidth at the oscillator output. This is
especially important to achieve when the oscillator is used as the LO-
injection source for a receiver mixer.
Apart from changes in ambient temperature in the area of the oscillator,
long-term drift is likely to result from RF circulating current in the coil and
the related capacitors. For this reason it is wise to use a physically large coil,
comparatively speaking, which has a fairly large core and a wire gauge (Litz
or enameled copper) that can pass the current without significant heating.
The more capacitors that are used in parallel at critical RF points (Cl
through C6 of Fig. 2-26a) to obtain a specified capacitance, the lower the
drift from RF heating. This is because the RF currents are distributed
through several capacitors rather than a few , thereby reducing the internal
heating of any one capacitor.
In the proven, high-stability oscillator of Fig. 2-26a, the drift from a
cold start to a period 3 hours later was so slight that it could not be
measured accurately. The maximum shift, as observed by a frequency
counter, was less than 1 Hz in a stable temperature environment of 25 °C.
L1 is housed in an aluminum shield, the walls of which should be one
coil diameter or greater away from the outer surface of the coil form. Closer
spacings degrade the resonator Q and complicate the mechanical stability of
the oscillator. An aluminum or brass enclosure will lower the coil induc-
tance somewhat, requiring that the slug be moved farther into the coil wind-
ing. Iron types of shields have the opposite effect, causing an increase in in-
ductance. The coil shield is useful in keeping unwanted RF energy from
other parts of the circuit (composite equipment) away from the oscillator
tank. Also, the shield enclosure helps to maintain a stable temperature when
short-term changes in ambient temperature occur.
Polystyrene capacitors are suitable for use at C4, C5, and C6. They ex-
hibit a negative temperature coefficient which compensates for the positive
coefficient of the coil magnetic core. NPO ceramic capacitors are suitable
also at those circuit points. Silver-mica capacitors are not recommended in
the interest of optimum stability, unless they have known drift characteristics.
Sec. 2-5 Slug Applications 71

The series-tuned Colpitts format was chosen to permit using higher

values of inductance at L1 than would be possible in a parallel-tuned ar-
rangement. The major advantage of the high L is that circuit-connecting
leads and printed-circuit-board conductors become a smaller part of the
overall L than is true when parallel tuning is used. This aids stability and
helps to prevent Q degradation. This consideration is especially meaningful
at frequencies above 5 MHz.
Oscillator short-term stability is aided by Dl , a gate-clamping diode
(l N914 or equivalent). The diode limits the positive peaks of the sine wave,
in turn leveling the peak transconductance of the FET. This serves as a bias
stabilizer and minimizes changes in junction capacitance. The latter, if
allowed to run rampant, would generate considerable harmonic energy. 02
stabilizes the drain supply voltage at 8.2.
The derivation of minimum and maximum capacitance range for the
oscillator is taken from the equation in Fig. 2-26b. The Land C components
have been arranged to illustrate how ceff is determined in this type of
oscillator. The effective tuning-capacitance range is approximately 55 to
214 pF with the arrangement shown. This circuit can be extrapolated to fre-
quencies up to 10 MHz by using the reactance values of the capacitors and
inductors shown.
An example of high-Q slug-tuned inductors used in a passive bandpass
circuit is given in Fig. 2-27. In effect, we are looking at a three-pole Cohn
type of tunable filter which might be used to obtain a high order of receiver
front-end selectivity. Although the frequency specified is for the standard
AM broadcast band, tunable filters of this type can be used at any frequen-
cy where lumped constants are acceptable-150 MHz and lower. The im-
portant thought to keep in mind is that the loaded bandwidth (BW L)
doubles each time the operating frequency is raised one octave. For this
reason a three-pole filter of the type shown is more useful at the lower end
of the frequency spectrum.
For the frequency specified in Fig. 2-27, carbonyl E core material is
chosen. L2 and LS comprise individual resonators for the filter. L3 and L4
are used in combination to provide the third resonator. These two inductors
have twice the nominal inductance of the end coils so that the net value of
the coils in parallel equals that of the individual end sections of the
filter-480 and 240 µH, respectively. This scheme enables the designer to
use a three-section variable capacitor for tuning, rather than a four-gang
capacitor, which would be required if four individual resonators were con-
tained in the filter.
Ll and L6 are adjusted for the desired input and output coupling. L7
and L8 are used as bottom-coupling inductors between the resonators.
Small toroidal inductors are suggested for use at L 7 and L8. The coupling
factors of the filter are chosen to provide the desired BW Lt consistent with
 550- 1600 kHz
-1---- ----r-- -1
3651 I 3551
I c1s 365 pF lc 1c pF I
pF I I
--+---- I

1
-----1--
I

I
--1
I 50 .n
I I I

If
output
I I I LS
I I I
I I I
_J__ _ __.__ _ _j __
'-------' I
I
I
----1---- / ___ J
Shielding

L 1, L6: See text

L2, LS: Carbonyl E core, 240 µH nom.
L3, L4: Carbonyl E core, 480 µH nom.
L7, L8: ,,,5 µH (see tex t)
Figure 2-27 A high degree of receiver front-end selectivity can be achieved by utilizing a
three-section tunable Cohn filter. High-0 slug-tuned inductors are used at L2 through L5.
Since L3 and L4 are effectively in parallel, their inductances are twice that of L2 and L5, to
permit tracking by means of Cl.

the insertion loss that can be tolerated in terms of noise figure and receiver
sensitivity. An insertion loss of 4 to 6 dB is a suitable target amount for the
frequency specified.
In the test model, the author employed J. W. Miller 43A224CBI induc-
tors at L2 and L5. The inductance range is 138 to 275 µH, with a Qu of 128
with the core set for 138 µH. Inductors L3 and L4 are J. W. Miller 43474CBI
parts, with a range of 240 to 580 µH. Q11 at 240 µHis 135. The Q11 of both
coil types was measured at 790 kHz.
The dashed lines in Fig. 2-27 indicate extensive shielding of the filter
from outside influences. Each resonator is shielded from the others. Cl, the
three-section tuning capacitor, should be constructed with shield baffles
between the rotor sections to minimize leakage from one filter pole to the
next.
This circuit was developed primarily to demonstrate just one application
for slug-tuned inductors. There is no reason why toroidal or pot-core units
could not be substituted for the type of inductors shown. Were this done,
however, it would be necessary to include trimmer capacitors at each section
of the filter to permit tracking.

72
2.6 RF Chokes

It is useful to include magnetic-core RF chokes in many designs, particular-

ly when miniaturization is a keynote. Ferrite or powdered-iron choke cores
permit large inductance values to be realized in much smaller volume than is
true when air-core or dielectric-core chokes are built. Another advantage to
the choke with a magnetic core is that considerably fewer turns of wire are
needed for a specified inductance, as compared to a choke of the same value
which uses no core material. This greatly lowers the de resistance of the wind-
ing and elevates the Q1r Figure 2-28 shows a collection of modern RF
chokes which are built in various physical formats. Some RF chokes are
available in encapsulated form. Others are built with the windings exposed.
The designer decides which style is needed, in accordance with the environ-
ment in which the choke will be used (humidity, chemical effects, etc.)
Figure 2-29 contains pictorial representations of a solenoidal choke (a), a pi-
wound unit (b), and a power type of solenoidal RF choke (c). These illustra-
tions portray only three of the myriad formats that RF chokes take. Some
have two or more pi windings, others are encased in phenolic material, and
so on.

2.6. 1 Choke Applications

RF chokes are not selected in random fashion during the design period.
There is a definite and necessary relationship between the choke
characteristics and the circuit point at which it is used. If the choke is used
as a choke, in the true sense of the word, the reactance should be con-
siderably higher at the lowest operating frequency than the impedance of
the circuit to which it is connected. A viable rule of thumb is to ensure that
the RF choke has an X l which is four times or greater the circuit im-
pedance. This principle is demonstrated in Fig. 2-30, where a cathode-driven
RF power amplifier receives excitation in parallel with a cathode choke,
RFCI. The input impedance of the amplifier at full drive is approximately

Figure 2-28 Group of magnetic-core RF chokes.

73
 Solenoidal PI
winding wind ing

~
Slot / Slot

=~·fb)))))ffi))}))LLc::!*~=t=
t Wi re lead ~ Wire lead
Magnetic core Magnetic core

(a) (b)

Ferrite
rod
µ, ;;;. 125
Bi filar winding J
heavy·gauge wire

(c}

Figure 2-29 A small solenoidal RF choke is seen in {a}. A bank·wound RF choke is il-
lustrated in (b). A bifilar-wound power choke is depicted in (c) . All three chokes are
wound on magnetic-core material.

RF
linear amplifier
3-30 MHz

To
matching
network o.o'i' µJI xL ;:;. 4z.,, :. xL ;;;.500 n
h c1ll]
150 n
Thus.
XL
L H ;;> - - ;;> 32µ H
µ 2rrf
• : Phasing
5 Vac 14.5 A
Figure 2-30 Design example for a bifilar-wound filament choke of
the type seen in (c) of Fig. 2-29.

74
Sec. 2-6 RF Chokes 75

150 Q. Therefore, XL for RFC ! should be equal to or greater than 600Q at

3 MHz-the lowest opera ting frequency. This requirement can be met by
using a ferrite-rod core that is 0.5 x 7 .5 in., with a /Ai of 125. The required
inductance for 600 Q at 3 MHz can be found by

x 600
LµH = 2~ = 6.28 x 3 = 32 µH

where! is in MHz. Once this is k nown, the number of turns to obtain 32 µH

on the core specified can be determined from the AL factor of Fig. 2-3c
(448) . Thus,

N = 100 V LµH + AL = 100 V32 + 448 = 26 turns

where N is the number of turns desired.
The filament/cathode choke uses a bifilar winding. T his can be con-
sidered as a single-wire winding in the calculations . The resulta nt induc-
tance is essentially the same for a single winding versus a bifilar winding if
each winding occupies the same amount of space on the core. The Q remains
the same a lso. So, to obtain 32 µH of inductance we will use 26 bifilar turns
of enameled copper wire, close-wound , as shown in Fig. 2-29c.
The bifilar winding can be extended to fill 6 in. of the core, in which case
the inductance will be roughly 100 µH (45 bifilar turns of No. 14 wire). The
core, thus used, will exhibit a Q11 of 150 at 8 MHz. T he No. 14 wire will pass
the 14.5-A filament current without excessive heating or JR losses. The core
materiaJ chosen for RFC I satisfies the Bmax requirements treated in Chap-
ter 1. The XL ~ 4Zin is necessary to prevent power Joss through the RF
choke. Furthermore, if the XL is too low, an intolerable m ismatch can re-
sult between the signal source and the load when RFC! is placed in shunt
with the 50-Q driving source.
The assembled RF choke can be coated with glyptol varnish to protect it
from damage and moisture. AJternatively, the winding can be encased in
heat-shrink tubing. Formvar-insulated copper wire is recommended for the
bifilar winding on RFCI of Fig. 2-30. This insulating material is tough, as
well as being resistant to damage from chemicals, heat, and oil.
Some additional applications fo r RF chokes are given schematically in
Fig . 2-31. In this example the magnetic-core inductors are used in the base
return of a Class C amplifier (RFC l) and in the Vee line to the amplifier, Qt.
Once again we shall observe the 4 x Z rule of thumb fo r XL. The base-
return choke should exhibit a reactance of 12 Qor more. This requires an in-
d uctance of at least 0.83 µH . A standard choke value of 1 µH would be a
suitable choice.
RFC2, the collector choke o f Q I , has an approximate collector im-
pedance of 3. 7 Q, based on the standard relationship Z = Vcc2 12Po, where
 R F power
amplifier 2.3 MHz 25·W
zcoll "' 3.7 n outpul
To to
input
matching
network
--, l ow-pass
fil ter

RFC 2 II ~Rl
_ _ _J

22 µF

RFC 3 II 1° "1····•' T
01 25 v

V~c
z coll '> -
2Po

X LtRFC ll ;;;. 4Z.,,

X L!RFC2);;;. 4Zcoll +13.6 v

R 1: see tex t
Figure 2-31 Additional uses for magnetic-core RF chokes are given here. RFC1, RFC2, and
R FC3 are discussed in the text.

Po is the desired output power in watts. Thus, XL for RFC2 should be 14.8Q
or greater. At four times Zcou this calls for 1.02 iiH. Once again, a standard-
value choke of 1 µH will suffice.
RFC3 is used as a decoupling choke in the Vee supply line. Since / coll will
be approximately 3.67 A, allowing 500Jo efficiency for the amplifier, the
supply-line impedance is roughly equal to 3. 7 Q (Z = El I), RFC3 should
have approximately I µH of inductance also.
Power transistors tend to self-oscillate at the operating frequency and
elsewhere in the spectrum by virtue of tuned base/tuned collector condi-
tions. The greater the transistor gain, the worse the problem. Typical HF,
VHF, and UHF power transistors exhibit gains of 13 dB or more. Stability
is, therefore, a design problem if care is not taken with layout and compo-
nent selection. In theory, the transistor gain increases 6 dB per octave as the
freq uency is lowered. Because of this phonomenon, self-oscillation at low
frequency and VLF is the principal malady to consider. If a power tran-
sistor is operated somewhat below its rated saturated power-output rating
(the point ·at which no further output can be obtained with an increase in
drive), the beta rises and stability is marginal. This is because as the collec-
tor current drops, the beta rises.
When transistor input and output capacitances are similar, and when
base and collector chokes are of the same or approximate value, the tuned
base/tuned collector syndrome arises. If the RF chokes have high Q, such as

76
Sec. 2-6 RF Chokes 77

they might when ferrite cores are used, the condition is quite apt to happen.
Therefore, it is sometimes necessary to swamp the chokes as shown in Fig.
2-31. RFCI is shown with a single miniature high-µ ferrite bead in series
with the grounded end. A bead of 950 µ is slipped over the pigtail of the
choke to swamp the choke Q. Laboratory measurements proved that a
single bead increased the choke inductance by I µH at 2.5 MHz. The choke
Q fell from 60 to IO when the bead was added.
An alternative Q-killer technique is employed at RFC2. RI is placed in
parallel with the choke to lower the Q. The value of resistance should be
selected to stabilize the amplifier with the highest ohmic value possible. The
lower the resistance of RI, the more RF power will be dissipated in the
resistor. Typically, values from 150 to 470 Q are satisfactory. Bypass
capacitors which are effective at various frequencies in the spectrum are
used at the junction of RFC2 and RFC3. They are an important part of the
decoupling network.
Solenoidal RF chokes need not be used for RFCI through RFC3. Small
toroids can serve as the cores for these chokes. This could enhance stability
by means of the self-shielding qualities of toroids.
A wide assortment of blank coil forms with pigtails affixed are
manufactured for use as RF-choke cores. They are available in various J.I;
values.

REFERENCES

[I] W.J. POLYDOROH, High-Frequency Magnetic Materials. New York: John Wiley
& Sons, Inc., 1960.
[2] HENRY JASIK, Amenna Engineering Handbook. New York: McGraw-Hill Book
Company, 1961.
3
APPLYING TOROIDAL CORES

The rules governing core selection with respect to physical size,

permeability, frequency, and Bmax were discussed in Chapter 1. In this sec-
tion we focus our attention on some of the practical applications of toroids
in communications circuits. Complete design examples will be given to il-
lustrate core selection and actual utilization of inductors and transformers
in typical circuits. Wherever it is possible and practical to do so, rules of
thumb will be offered as time-saving shortcuts for the engineer or technician
who is engaged in developing prototypes. From these plateaus the proof-
build or production-stage circuits can evolve with minimum refinement ef-
fort.

3.1 Low-Level Circuits

For the purpose of definition, low level in this book refers to circuits in
which the power level is less than 1 W. In some engineering circles this term
relates to power levels in the mW region. For the most part, the expression
is a mathematical one. However, it has been defined here as it will be used
by the author.
In communications equipment there are a number of low-level circuits in
which inductors or transformers with magnetic cores find application. Prom-
inent among these circuits are RF amplifiers, IF amplifiers, oscillators,
mixers, detectors, tuned audio amplifiers, and filters. Some of these circuits
operate in narrow-band fashion, while others are classified as broadband in
their functions. The narrow-band circuits have reasonably flat responses
over a range of hertz or kilohertz, depending on the design. Broadband cir-

79
80 Applying Toroidal Cores Chap. 3

cuits exhibit responses as great as several GHz, conversely. In general terms,

narrow-band circuits are not as complicated to design as are broadband cir-
cuits. In either case it is necessary to pay special attention to the core
material used, while also considering the operating parameters of the active
devices in the related circuit. Our treatment of magnetic cores will be ex-
panded to include other important design features for the composite circuits
used as examples.
A narrow-band circuit can be structured to encompass a wide range of
frequencies by making it continuously tunable over the required frequency
spread. On the other hand, some narrow-band amplifiers are fixed-tuned to
operate on a single frequency, such as would be the case with an IF
amplifier. Continuously tunable narrow-band circuits are indigenous,
primarily, to the front ends of radio receivers and the stages of some multi-
band radio transmitters, the latter of which employ vacuum tubes.
Magnetic-core material is seldom used in low-level circuits above ap-
proximately 150 MHz. The same is not quite as true of high-level circuits,
where broadband transmitters contain ferrites in devices called isolators and
circulators. Figure 3-1 shows a number of toroidal inductors and
transformers which are typical of the kind discussed in this chapter.

3.1.1 RF Amplifiers

A fixed-tuned RF amplifier for use at 2.5 MHz is used as a design exam-

ple in this section. Figure 3-2 contains the circuit and some pertinent data
that apply to it.
Let us assume that Ql has been chosen for low noise and full gain
capability at 2.5 MHz. A suitable rule of thumb for ensuring adequate gaiq
atfo is to choose a transistor which has an/rrating at least 10 times/0 • This
calls for an fr of 25 MHz minimum when selecting Ql. The symbol fr
specifies the frequency at which the small-signal forward-current ratio is I

Figure 3-1
•
Various toroidal inductors and transformers.
 RF ampli f ier

f 0 = 2.5 MHz
Sig. T1
in
{50nl

C2

470.n C3
R1 39
0.006 µF C5
4. 7 k.11 krl R2 l 0.006µF
100 .11

Xe 1 • XC4 "'200 .n at 10 le= 1.5 mA

Xe2• XC3. Xes ,;;; 10 .11 at f 0 VE= 0.7
Cl. C4: 318 pF nom.

N( Tl) = N (T2) =
IE.
__£,'.,'.
z,.e
RE= VEB ' "' 0.7 silicon)
Rl "' lORE
R2 : Select to obtain
2VEB (1.4 V) with
XL = Xe at f 0 R1 - R2 pair
:. X1.. Tl {sec}.T2(pri) = 200 = 12.7µH tr (01) ;;;. 1or0
Figure 3-2 Design example for a small -signal RF amplifier in which two toroidal transformers
!Tl and T2) are used. Design progression is discussed in the text.

(unity) in the common-emitter state. This is known as the gain-bandwidth

product. The symbol for small-current forward-current ratio is hfe· The hfe
of Ql must be equal to or greater than the voltage gain required of the
amplifier if

NI
- = 1.
N2
Thus, if we desired a voltage gain of 10 for the circuit of Fig. 3-2, the Ql hfe
should be no Jess than 10 if the primary and secondary windings of T2 have
an impedance ratio of 1. If, with the step-down ratio shown in the diagram,
the circuit must provide a voltage gain of 10 at the output port, the hie will
need to be proportionately higher than 10.
An arbitrary Xe is suggested for Cl and C4. It will be 200 Q, which calls

81
82 Applying Toroidal Cores Chap.3

for an XL of the same ohmic value at resonance. Thus, the secondary of Tl

and the primary of T2 have an XL of 200 Q. The required inductance of
those two windings is

XL
Lµ H =_ ____;~- 12.7
2rrfmHz
when the parallel capacitance is 318 pF at fo.
The characteristic impedance o f the tuned windings can be determined
when the series resistance, R, is known:
2
z -- XL
R
ohms

or when the Q is known, by

Z = XL X Q
where Z is in ohms and Q = XIR.
In this design example we will assume that the factor R is 1 Q, made up
by the ac resistance of the transformer winding. This being the case, Q =
XLIR or 200. From the equation for impedance we note that Z = 200 2 /l =
40,000 Q. Similarly, when XL and Qare known, Z = 200 x 200 = 40,000 Q
Starting with Tl of Fig. 3-2, the turns ratio, primary to secondary, is

N = VZsc/Zpri = v 40,000750 = 28 : 1

The tap point on the secondary winding can be obtained from

N = V40,000/600 = 8: l
as taken from the basic equation for N.
In making a core selection, we find one that is rated for optimum Q at
2.5 MHz, thefo in Fig. 3-2. Carbonyl E will be chosen because of its 1- to
30-MHz recommendation. It has a µi of 10, which will permit an ample
number of c0il turns to obtain a practical realization of the turns ratios
determined earlier. A 0.5-in.-diameter core will be large enough to operate
without saturating in this type o f circuit, assuming that abnormally large
signals are not present at the input to Tl. An Amidon Associates T50-2
powdered-iron core is suitable (see Appendix E for available Amidon
cores).
The AL factor for a T50-2 core is 50, so to learn how many turns are
needed to obtain an inductance of 12. 7 µH, we apply

n = 100 VLµH I A L = 100 V 12.7/ 50 = 50 turns

Sec. 3-1 Low-Level Circuits 83

where n is the number of unknown turns and Al is taken from the

manufacturer's literature, or determined as outlined in Chapter 1.
The number of primary turns for Tl is obtained by n/N (50/28), which
calls for a primary consisting of l. 78 turns. Since this is an impractical
number, a two-turn primary is used. The core size versus wire gauge table
indicates that 50 turns of No. 26 enameled wire will fit on a T50 core. To
ease the winding job somewhat, No. 28 wire will be used, thereby allowing
up to 64 turns to be applied to the core.
The tap point on Tl (secondary) is found by n/N (50/8), wbich calls for
the tap being placed 6 turns above the grounded end of the winding. The
calculations for T2 are performed in an identical manner, taking into ac-
count the different impedance levels. It is worth observing that irrespective
of the computed impedance of the secondary winding of Tl (40,000 Q), the
resultant turns ratio between the primary and the tap point to ground on the
secondary, satisfy the 50- to 600-Q condition. This illustrates the relative
unimportance of knowing the Q of the resonator (provided that it is high) or
the ac resistance, R, provided that the large winding of the transformer has
considerably more turns than the low-impedance winding and tap. Details
for biasing Q 1 of Fig. 3-2 are beyond the intent of this volume, but data are
given in the diagram as guides in approximation. A detailed discussion of
amplifiers and biasing methods is contained in a book by Lenk [ l].

3. 1.2 Capacitive-Divider Matching

It is not always necessary to use the transformer arrangement of Fig. 3-2

for matching the source to the load. The approach seen in Fig. 3-3 simplifies
the winding process for toroidal inductors by eliminating the link winding.
In this example we find a capacitive divider at the input and output ports of
the RF amplifier. Employment of an FET at QI eliminates the need for a
tap point on L1 and L2, since the gate 1 impedance is in excess of 1 MQ and
the drain impedance with 1.2 mA of current is 10,000 Q. However, if the in-
put and output impedances of the active device were less than 10,000 Q,
suitable tap points could be placed on L1 and L2 while still using the
capacitive dividers.
In order for the scheme of Fig. 3-3 to be applied, Rin must be less than
RL. Furthermore, QL has to exceed 14.1, as shown by the equation in the
design example. This means that the BWD of the two resonators cannot be
of such magnitude as to require a QL less than 14. l.
For our design exercise we will specify a BWL of 500 kHz. This satisfies
the QL restriction with a resultant value of 20. Cr consists of the nominal
value of C l, as computed, in series with the calculated value of C2. Step 4
of Fig. 3-3 provides the required Cr to meet our bandwidth and impedance-
transformation parameters. Since X e and X l are equal at/a, and because Xe
= 250, the required inductance fo r LI and L2 is 4 µH. From step 5 we learn
that Cl 0001 = 68.4 pF and C2 = 833 pF, as defined by DeMaw [2].
 RF amplifier

27 kU lOOkU
+12 v
R;"= 10 kQ
C4 10 MHz

10 = 10 MHz
l s
I
RL = 10 kQ"" C3

R'L
L2
son
910 51 C5
R'" L1 10 kQ
so n II

ca
100 n

+ 12 v
Xe (C3, C4, C5, CB) ~ lO i l ;;;. 0.0015 µF

QL > 14. 1

1: Select SW L (0.5 MHz)

2: 0
L•
O'
L
=~
t - f
= 10 MHz = 20
O.SMHz
2 1

3: c T = Cl nom + C2 and c~ = C6nom -~ C7

4: Cr=[ QL
2r.f0 (RL /2 )
J x 106 =63.6pF=Xc 250 :. LµH= 4µ H

5: - C2- "" ~L
- - 1 = 13. 1
Cl nom Rin

: . Clnom = 68.4 pF , C2 = 833 pF, C6 nom = 68.4 pF

and C7 = 833 pF
Figure 3-3 Design data for using capacitive dividers to effect an input and output match
when narrow-band toroidal inductors are used without taps or link windings.

84
Sec. 3-1 Low-Level Circuits 85

From a practical viewpoint, selection of the toroid core is less difficult in

this example than was the case related to Fig. 3-2: It is not important how
few turns are used in the winding when the smaller primary is not required.
The core material should, of course, be suitable for the chosen fo. A
powdered-iron core with an SF mix will result in an optimum Qu (Q11 should
always be considerably greater than QL). An Amidon Associates T50-6 core
is used. It is an SF type, has a µ 1 of 8 and is recommended for use between 2
and 50 MHz. The core diameter is 0.5 in. The AL factor for this toroid core
is 40. Therefore, the required number of turns for LI and L2 of Fig. 3-3 is
31.6, using the equation from Sec. 3.1.1. A winding of 31 or 32 turns will be
acceptable.
Mica compression trimmers are suitable for use at Cl and C6 of Fig.
3-3. Silver-mica capacitors can be used at C2 and C7. Both types will ensure
adequate capacitor Q to prevent Q degradation of the two resonators. Cl
and C6 are specified as variable capacitors to allow peaking at precisely IO
MHz. This makes it possible to use the nearest standard capacitor value at
C2 and C7, which in this design would be 820 pF. Fixed-value capacitors
could be placed at all four points if LI and L2 were replaced by slug-tuned
inductors of the type treated in Chapter 2. However, the toroids are a better
choice in the interest of circuit stability, owing to their self-shielding
characteristics.
3.1.3 Interstage Coupling

Tuned toroidal transformers are excellent as coupling elements between

amplifier stages in a receiver or transmitter, as shown in Fig. 3-4. For the
purpose of illustration we will assume that the gfs of QI and Q2 is 3000 and
the drain impedance (Zd) of QI is approximately 4000 Q. Zin of the source-
driven amplifier, Q2, is roughly 333 Q when g15 is 3000. These charac-
teristics call for a Tl turns ratio of 3.4: l to ensure maximum transfer of
energy from Q l to Q2. Cl is used to tune the primary of Tl to resonance at
f o. Using an arbitrary XL and X e of 200 Q, the nominal capacitance of Cl at
resonance is 454 pF and the inductance of the primary of Tl is 18 l'H. In
order for C2 and C3 to function effectively as bypass capacitors, Xe should
be 10 Q or less. A 0.01-µF capacitor will suffice as a close standard value.
RI is chosen for the bias required at Q2.
An SF type of powdered-iron core will be suitable for Tl to satis fy the
high-Q requirement of the tuned transformer. An Amidon T50-2 core will
be satisfactory (AL = 50), requiring 60 turns of No. 30 enameled wire for
the primary winding. The secondary of Tl will contain 18 turns of wire.
Although we have specified a type SF core, a Carbonyl C core would be
suitable also. It would require fewer turns of wire because of the higher µ
(20). The SF core has a I' of IO. If a C-type core was used, we could select an
Amidon T50-l, which has an AL of 100. That being the choice, the primary
 RF amplifier RF amplifier
fo
3mA
-----+- 1.75 MHz
..,__~--~~Tl .-~~~S~ D

0.01 µF
0 .01 µF
C2 C3

+12 v .

01. 02 (g,, ) = 3000 Xe (C2, C3) ..;; 10 n = 0.009 µF min.

Xc 1 ""'200 n at 10 .. 454 pF nom.
z 0 "' Vos =~ =4000n
10 s 0.003A xL 1T1 pri.l=200 n = 1aµH
106
z, 0 "'
g,, "' 333.o n

Figu re 3-4 Toroidal transformer (T1) used for interstage coupling.

of Tl would contain 42 turns of No. 28 wire and the secondary would con-
sist of 12 turns of wire. Either type of core material will yield excellent Q at
l.75 MHz.
A trifilar-wound toroidal transformer is illustrated in Fig. 3-5. In this
circuit a mixer, Q 1, supplies signal voltage to a half-lattice crystal filter,
FLI. Tl is tuned by means of trimmer CL The transformer phasing is such
that YI and Y2 are supplied with energy that has a 180° differential. RI is
used to improve the mixer IMD and R2 is chosen to provide the filter ter-
mination that ensures minimum passband ripple.
The rule of thumb for XL and Xe in Fig. 3-4 can be used in this circuit
example. Therefore, the windings of Tl will have an inductance of 3.5 µH
and Clnom will be 88 pF. An SF type of powdered-iron core will be ap-
propriate for use at 9 MHz. If we employ an Amidon T50-2 toroid core,
which has an AL factor of 50, the transformer will contain 26 turns of wire
for each winding. Since the three windings are applied in parallel, the core
will require an equivalent of 79 turns. This calls for No. 30 enameled wire.
To ease the winding assignment, one could use a larger core of the same
material, such as a T68-2. The required number of turns would be 25

86
 Fll
9 MHz
Tl
Yl

Rl Cl To
..___-11---<> 1- F
amp lifier
lOOkQ
R2
560 i1
I
Lo input
Y2

• : Phasing
+ Vao
Figure 3-5 A toroidal trifilar-wound transformer {Tl) is excellent for coupling a single-ended
mixer to a half-lattice filter.

because of the higher AL (55). Since a 25-turn trifilar winding is approx-

imately equivalent to 75 turns, No. 28 wire would satisfy the requirement.
C2 and C3 follow the rule set for Fig. 3-4. The application of bifilar and
trifilar windings can be made easier by first twisting the wires approximately
eight times per inch. Identification of the windings can be greatly simplified
by using three different colors of enamel insulation. Red, green, and brown
are the three most common colors available. The wire sizes specified in the
foregoing text are those which will fit on the cores indicated for Tl of Figs.
3-4 and 3-5.

3.1.4 Toroids in Filters

One of the most appropriate circuit applications for magnetic-core in-

ductors is in filters. Toroids or pot cores are the choices of most designers
when the operating frequency is at high frequency and lower. The advan-
tage in using either type of core is the high Q attainable and the self-
shielding properties of toroids and pot cores: Isolation between the filter
poles is essential to proper performance.
A design example for a bandpass filter with a 600-Q characteristic im-
pedance is given in Fig. 3-6. It is based on the image-parameter concept. To
ensure a center frequency (/o) of 250 kHz, LI requires 77 .18 µH of induc-
tance and C2 must have 4336 pF of capacitance. Cl, the top-coupling
capacitor, requires 556 pF of capacitance. Cl and C2 should be made par-
tially variable by placing trimmers in parallel with fixed-value capacitors.
This will allow for final optimizing of the filter to compensate for parasitic
inductances and capacitances which may exist. High-Q capacitors of
temperature-stable characteristics should be used. Polystyrene, silver-mica,
or glass capacitors are suggested.
The toroid cores for use at Ll are chosen for optimum Q at/0 • Carbonyl

87
 Cl

R," n Lt
600 II C2 C2

R = 600r2 ' o = 0.250 MHL BWL = 0.05 M Hz f2 = 0.275 MHz

r, = 0 .225 MHz

7.96 {f1 + f2 ) 104 20,660 = F

2 - - Cl = f 1 R = 37, 125 556 P
1 2

3 - - C2 = 3.18f,(10 )
[ f 2 (f2 - 11 )R
5
J= [71 ,550
- -
8.25
J = 4336 pF

2 2

4 - - Lt = 2
[
.0796 (f2 -f1 ) R
t, 12
J =
2
[ 2.388
.0618
J - 77 18 µH
- .

5 · - Choose a toroid core (Amidon T68- 15). A. 0.196 cm2, =

AL= 180. OD = 0.68 in .. µ , = 25 and 8 mo, = 22 kilogauss.

6--N = 100 J LµHI AL = 100J77.18/ 180 =65turns

8
E,ms x 10 2 x 108
7- · B = = 14. 1 gausses
oP 4.44 A 0 N (f mh l X 106 ) 14, 141 ,400
Figure 3-6 Design progression for using toroid cores in a bandpass type
of two-section filter.

GS-6 is suitable. It has a µ of 25 and is recommended for use from 0.1 to 2

MHz. An Amidon T68-15 will be satisfactory for use at LI. The equations
of Fig. 3-6 demonstrate clearly that B 0 P is only a fraction of the Bmax rating
of the core. Therefore, saturation will not occur.
Another design example is offered in Fig. 3-7. The circuit shows how a
toroidal inductor can be used effectively in a tuned audio amplifier to ob-
tain a peak response at 700 Hz. The value of Cl is chosen so that the com-

88
 AF amplifier AF amplifier

2mA f 0 = 700 Hz
0.03 µF
01 02
C1
3k.Q C2
R1 0 .47 µF

100 .n
Max . 0

+24 v

RP "' Rd ll Rl + Xc 1
Xc 1 = 7583 Q

R0 11•111 "'Rdll Rl + Xc 1 = 9833 Q

Xc 2• X Ll = 483

1. Select R, for 1,1 of 2 mA

2. Q = Ro = 9833 = 20.3 Xu
O=-
XL 483 R2

700 Hz .
3. BW, = -fo =- - = 34.5 Hz With R2 at 0 n
· al 20.3

4. BWL = 143 Hz with R2 at 100 .11

5. Choose a toroid core for L 1 ( 110 mH l. Amidon FT 82 -72 ferrite.

µ, = 2000, AL • 1172, 00 = 0.8 in.

6. N = 1ooojl "'H IAL = 1000 J 11011172 =306 wrns

Figure 3-7 A high-Q toroidal inductor is employed in this design example of a tuned
au dio amplifier. R2 serves as a Q (bandwidth) control.

bination of Cl and RI will be high enough in value to prevent excessive

loading of the resonator, Ll/C2. A value of approximately 9833 Q results
from the values indicated. Once this information is established, we can
choose an arbitrary value for Xc 2 and Xu, which for the purpose of this
demonstration is 483 Q. The design procedure can now progress as shown in
steps 1 through 6 of Fig. 3-7. The bandwidth of the resonator can be varied
from roughly 35 Hz to 143 Hz by means of R2. As resistance is added at R2,
the resonator Q drops and the bandwidth increases.
In order to obtain a practical realization of the inductor, LI , a high-µ
ferrite core is selected. T he tuned circuit will be resonant at 700 H z when C2

89
90 Applying Toroidal Cores Chap. 3

is 0.47 µF and Ll has 110 m H of inductance. An Amidon FT-82-72 toroid

core meets our requirement and will contain 306 turns of wire to provide the
inductance specified. The skirt selectivity of the response curve can be
enhanced considerably by cascading another fil ter section after Q2. A dual
100-Q control could then be employed in place of R2. This would permit
bandwidth control over both resonators, sim ultaneously.
Toroidal inductors are well suited to the harmonic filter depicted in Fig.
3-8. In this example cascaded pi networks are used as a half-wave filter to
reduce the harmonic currents o f a 5-MHz local oscillator. Since the
characteristic impeda nce of the bilateral filter is 50Q, a nd because a QL of
1 is specified, X L for the two inductors will be 50. X e for Cl and C3 is also
50, and X e for C2 is 25. Steps I through 6 illustrate the simple progression
for designing the filter to have an fc ( - 3 dB) of 6 MHz. In most practical
situations where this type of filter fs used, the nearest standard capacitor
value can be employed .

I
5.0 - 5.5 Local I/ 50 S1 L1
FL1
L2
To
I

f,_ _c1_~_~f>-- 'c~

_2_ __.f c,'
MHz mixer
50 .n
.____,
ir---1':

XLl , XL 2 • 50 X c2 = 25 2/ .. - 3 1 dB
X c 1· Xc3 = 50 fco = 6 MHz 31 = - 53 dB
OL ( FL1 l =1

XL 50
1. LI, L2 (µ H) = - - = - - = 1.32
27ffMMl 37.68

1 1
2. Cl . C3 (µFJ = f X = - - = 0.00053
211' MHI C 1884

3. C2 (µFJ =
1
= - 1- = 0.001
2;rf MHz X C 942

4. Choose core for 5 MHz (Amidon T50 - 2}

5. N = 100 J L,,HIA L = 100 J 1.32150 = 16 turns

6. Choose nearest standard values for Cl , C2, C3.
: . C1, C3 = 510 pF and C2 = 1000 pF.
Figure 3-8 Toroids are well suited for use in harmonic filters because of
their self-shielding properties and compactness. Here L 1 and L2 are used
in a half-wave low-pass filter.
Sec. 3-1 Low-Level Circuits 91

A five-pole Chebyshev filter with a low-pass response is presented in

Fig. 3-9. In this example there are three toroidal inquctors employed, LI,
L2, and L3. A maximum RF power level of 1000 W is specified, which
results in an rms voltage level of 224 across 50 Q. With the inductive reac-
tances specified the end inductors require I0.1 µH of inductance and the
center coil has an inductance of 14.28 µH. Thefco ( - 3 dB) is 1675 kHz. Cl
and C2 compute to the standard value of 0.002 µF. Silver-mica capacitors
with a rating of 600 V will be suitable at Cl and C2, provided that an SWR
of 1 prevails in the transmission line.
Carbonyl C material is selected for the toroids . It will provide adequate
Q at the operating frequency of the filter. The J.li of the material is 20. An
Amidon Tl06-l core is used. It is 1.06 in. wide and 0.437 in. thick. The ID is
0.56 in. The maximum acceptable flux density of the material chosen is
12,000 G. Design steps 3 and 4 of Fig. 3-9 yield a B0 P of 235 G for Ll and
L2, with a saturation density of 194 G for L2. (B0 P is the operational flux

FL 1

E,m, = 224

50!1 son
P,., = 1000 W l.___ _ _ _..__..___ __.~.___ _ _ _.j
I
5·pole Chebyshev

X u. X LJ = 106.8 X L2 = 150.3
Xc 1 • X c 2 = 45.8 2f = -45dB
f0 = 1600 kH7. 3f =- 65 dB
{co= 1675 kHz 4f=-78dB

L 1, L3 = 10. 1 µ H L2= 14.28µH C1 , C2=0.002µF

1. Choose a toroid core (Amidon T 106- 1,

AL = 280, 12,000 G. A ., = 0. 706 cm 2 )

2. L 1. L3 - 19 turns No. 16 wire. L2 = 23 turns No. ·15 wire

224
3. 8 ( L 1 L3) = x l Os = 235 G
op ' 4.44 X 0.706 x 19 x 1.6 X 106

4. 8 ( L2} = 224 x lOs = 194 G

OP 4.44 X 0.706 X 23 X 1.6 X 106
Figure 3-9 Core selection must be based on the Bop of the circuit. L 1,
L2, and L3 are chosen accordingly in this design example.
92 Applying Toroidal Cores Chap. 3

density and is a more definitive term than Bmax• which was introduced
earlier in the book . Both terms find common usage in the industry, and they
refer to the same condition.)
Both values are well below the level at which saturation takes place.
Some heating of the powdered-iron cores will occur. T herefore, it is impor-
tant that ample ventilation be included in the design.
Some RF engineers prefer to use £pea k rather than Erms for the B0 P equa-
tion. This will allow additional leeway in choosing a core chat will operate in
the linear portion of its curve. However, when B0 P is considerably below
Bsat for a given core, as is the case in Fig. 3-9, the E rms factor is satisfactory.
An example of a high-pass filter that contains toroidal inductors is given
in Fig. 3-IOa. A filter of th is type would be used at the 50-Q input of a
marine-band receiver to prevent broadcast-band signals from overloading
the marine receiver front end. T his design is based on the image-parameter
concept.

' 1800pF •91QpF '1800pF

<>---j {--------<>
Cl C3

- son Lill

Xll , Xl 2 =24.55 = 2.3µH !co= 1.7 MHz {- 3 dB)

X c 1 , Xc 3 = 50 = 1873 pF
X c2 = 100 =936 pF
L1, L2 - 21 turns No. 22 w ire on
SF core (Amidon TS0- 21
• - Nearest standard value
(a)

~ L1

(bl
L2

Figure 3-10 The circuit in (al uses two toroidal inductors

in a high-pass filter. The pictorial view in !bl suggests a
method for discouraging unwanted capacitive coupling
between the filter sections.
Sec. 3-1 Low-Level Circuits 93

The core material for LI and L2 is type SF powdered iron. An Amidon

T50-2 core is used. To provide the required 2.3 µHof inductance, the cores
will be wound with 2I turns of No. 22 enameled wire. Cl, C2, and C3 com-
pute to values slightly different than those given in the schematic diagram,
so the nearest standard values have been specified. The change in filter per-
formance will be minimal when using the values listed in Fig. 3-1 Oa.
Figure 3-lOb illustrates the recommended way to assemble the filter. The
same layout technique is suggested for all the filters described in this sec-
tion. Although toroids have a self-shielding characteristic, it is best to
isolate them from one another in shielded compartments. This will prevent
unwanted stray coupling between the leads. Capacitive coupling is also
discouraged by this method . A die-cast enclosure is suitable for a commer-
cial product line. Double-clad printed-circuit-board material is easy and in-
expensive to use for prototypes and la boratory models. The silver-mica
capacitors can be supported by their pigtails, as shown.

3. 1.5 Toroids in Instruments

There are countless applications for toroids in test instruments. We will

examine a few examples here for the purpose of illustration. The first
presentation is that of a simple 20-dB bidirectional coupler (Fig. 3-11). The
circuit is used as an attenuator of the nondissipative variety. Ports X and Y
are for the signal source and a 50-Q load. A straight length of wire connects
X and Y, passing through the core to consitute a single turn of the current
transformer, Tl. The secondary winding of TI contains 10 turns. A
0.5-in. -diameter ferrite core is used. lt has a µ of 950 (Amidon FT50-43
core), ensuring broadband characteristics throughout the high-frequency
spectrum. When port Z is terminated in 50 Q resistive, this load will be
reflected through TI in accordance with the turns ratio squared. As a result,
the core will appear to be a 0.5-Q resistance in series with the through-line.
This condition will cause the source at port X to look into 50.5 Q. Since the

Tl

x y

Rgure 3-11 A toroidal transformers can be

used as shown here to construct a 20·dB
bidirectional coupler (see the text).
94 Applying Toroidal Cores Chap.3

ratio of the resistances is 5010.5, or J00, the power delivered at port Z will
be - 20 dB of that appearing at port Y. The usefulness of this type of at-
tenuator is seen when the signal source is too great in magnitude to use safe-
ly with some instrumenta tion.
An extension of the circuit given in Fig. 3-11 is shown in Fig. 3-12. In the
example given, we see an SWR indicator for use in an RF transmission line
of 50 Q impedance. Tl is a current transformer with the equivalent of a one-
turn primary. The voltage appearing across the transformer secondary is
proportional to the current in the through-line, as is the voltage across RI.
The voltage appearing at the junction of Cl and C3 is reduced from that on
the through-line by the ratio of the capacitive divider. Cl is adjusted so that
the voltage at Cl , C3, and DI is the same as that across Rl. When J2 is ter-
minated in a 50-Q load, the two voltages will be equal and of opposite
phase, thereby canceling . This will provide a reading of zero at Ml to in-
dicate a VSWR of 1.
Terminals J 1 and 12 a re now reversed and C2 is adjusted to provide a
null at MI with JI terminated in 50 Q. The bridge will at this juncture be ad-
justed properly for a 50-Q characteristic impedance. With the load con-

Tl

J1 J2
5 pF 5 pF
son son
Cl C2
02

1on 10 n
RFC C3 Rl R2 RFC
1 mH 1 mH

0.01 µF 0.01 µF

r f-J_
R F

Sensitivity 0- 100

Figure 3-12 The basic configura tion of Fig. 3-11

can be used as shown here to provide
an SWR indicator or RF power bridge. This circuit is based on the classic design by
Warren Bruene of Collins Radio Co.
Sec. 3-1 Low-Level Circuits 95

nected to J2, the source to J 1 and the meter switched to read forward power
(F), a full-scale reading should be noted at Ml when R3 is adjusted for
greater resistance than was present during the adjustment of Cl and C2.
Typically, R3 is set for maximum meter sensitivity during the nulling pro-
cedure. Thereafter, the control is adjusted for full scale at Ml when max-
imum power is flowing through the bridge.
The circuit is suitable for relative SWR measurements, as shown in Fig.
3-12. Additional switches and trim pots can be inserted between SI and the
two RF chokes to provide calibrated RF power readings in both the forward
and reflected modes. This technique permits actual RF power
measurements in both directions. An accurate means of measuring VSWR
will result.
The arrangement shown at Fig. 3-12 is patterned after the basic RF
power bridge developed by Warren Bruene of Collins Radio in the 1950s.
Most commercial RF bridges that are used for power and SWR measure-
ment (using a current transformer) are founded on the Bruene design.
Another useful application of a toroidal transformer is seen at Fig. 3-13.
This circuit can be used as a return-Joss bridge for measuring impedance. It
can be employed also as a 6-dB hybrid combiner when, for example, two
signal generators must be connected to a test circuit. An example of such an
application can be envisioned where the dynamic range of a receiver must be
measured.
When the circuit of Fig. 3-13 is used as a return-loss bridge, the
unknown impedance is connected to port X. Initial setup calls for applying
power to the bridge while port X is open- or short-circuited. A 50-Q detector
is connected to port Y. Signal power is increased until a full-scale reading is
obtained on detector meter. Next, the unknown impedance is attached at
port X. With the same level of power applied to the bridge, the meter is
observed for scale reading. The ratio of the meter readings, expressed in
decibels, is the return loss (dB = 20 log E.JEi). The greater the return loss,

50.U
x
50 .U
z
50.U

Figure 3-13 A toroidal transformer can

be used effectively as the heart of a return-
loss bridge for impedance measurements.
96 Applying Toroidal Cores Chap.3

the closer the unknown impedance is to 50 Q. Tl is wound on a ferrite

toroid core of 950 Q (Amidon FT50-43). It contains IO bifilar turns of No.
24 enameled wire. This basic circuit was described by Hayward in Solid
State Design for the Radio Amateur [3) .
When the circuit of Fig. 3-13 is used as a 6-dB hybrid combiner, a pair of
signal sources are connected to ports Zand Y. The equipment under test is
attached to port X. If each generator is adjusted to deliver 20 mV to the
combiner, an output voltage of IOmV will result, thereby causing a loss of 6
dB. The isolation between the generators is desirable to prevent one signal
source from phase-modulating the other. If phase modulation were allowed
to take place, the resultant sidebands could render the measurements inac-
curate.
3.1.6 Other Low-Level Applications

Broadband passive and active mixers depend upon toroidal transform-

ers in most instances to provide a balanced-to-unbalanced (balun) transfor-
mation. The same is true of balanced-modulator circuits of the broadband
variety. Figure 3-14 contains details for a basic doubly balanced mixer
(DBM) of the four-diode passive type. The same configuration can be used
for diode-ring balanced modulators. In the latter case the audio is applied at
J3 and the suppressed-carrier RF energy is taken out at 12.
The terminal impedance of the circuit shown in Fig. 3-14 is approxi-
mately 50 Q when one of the ports looks into a 50-Q load. Some modern
balanced mixers of the DBM type used in this illustration are followed at J3
by a diplexer. This can consist of an LC network that is terminated in 50 Q.
The fco of the filter is roughly three times the IF. This technique will help
to ensure a 50-Q termination for the mixer at a.II three frequencies of in-
terest, thereby improving the mixer !MD characteristics.
To realize optimum balance, it is necessary that Tl and T2 of Fig. 3-14
be as symmetrical as possible when installed in the circuit. The same holds
true for the four diodes. They should be selected for dynamic balance
(closely matched forward and reverse resistances): hot-carrier diodes offer a
good match for circuits of this kind.
High-µ core material is used for Tl and T2. This permits the designer to
develop relatively small mixer modules without sacrificing low-frequency
response. When core-winding provisions are available for very light gauge
wire, toroid diameters as small as 0.125 in. become practical. Quarter-inch-
diameter cores are specified in Fig. 3-14 to permit hand winding the toroids.
The material is ferrite and the /Ji is 950. If extreme care is given to sym-
metry, this circuit should be suitable from 500 kHz to as high as 500 MHz.
Many commercially made DBMs are rated from the low-frequency spec-
trum to above 1000 MHz.
 IL~ 8 dB
Jl J2
Sig. Tl 01 - 04 LO
in in 150 rn
(50.11) (+ 7d6m)
-::-

1- F
out
Jl, J2: 0.5- 500 MHz
(50.11)
J3: de - 500 MHz
X LIT!. T21 = 200 .\1 Core: FT·23·43, µ; = 950,
L (Tl , T2) ;;l?o 64 µH AL = 189 :. N = 18 Trifilar
turns No. 34 enam wire

Isolation (midrange) 01 - 04: Matched 1N914 or

hot-carrier type
Jl to J2 "" 45 dB
J2 to J3 "" 40 dB
Figure 3-14 Doubly balanced mixers contain broadband baluns (Tl and T2) as
seen here. These trifilar-wound transformers help to ensure symmetry, which is vital
to balanced mixers.

Port-to-port signal isolation at the spectral midrange of this mixer is ap-

proximately 45 dB from J l to J2. Between J2 and J3 it should approach 40 dB.
XL for the trifilar windings of Tl and T2 should be roughly four times
the circuit impedance at the lowest operating frequency. Therefore, an XL
of 200 Q will suffice, thus requiring some 64 µH of winding inductance.
This calls for 18 turns of No. 34 enameled wire on an Amidon FT-23-43 or
equivalent core.
The basic circuit of Fig. 3-14 is useful as an RF switch or RF attenuator
when bias is applied to the diodes in the appropriate fashion. In such cases
the B0 P of the transformers must include the de bias as well as the ac voltage
applied to them .
Figure 3-l 5a demonstrates how a broadband toroidal transformer (TI)
can be used in a diode type of frequency doubler. In the same circuit a
narrow-band toroidal transformer, T2, is employed. A trifilar winding con-
sisting of 10 turns of No. 28 enameled wire on a 0.375-in.-diameter ferrite
toroid (950 µ) will suffice for Tl. A powdered-iron toroid core is recom-
mended for use at T2. This will ensure a high Q for the tuned circuit- a
necessary step toward "laundering" the 2/ waveform. If Cl has a nominal
capacitance of 40 pF, the secondary ofT2 will require 0.7 µHof inductance.

97
 Doubler
Tl 01 T2
f
40 p F (nom.)
30 MHz
Cl
1-=-
2f
60 MHz

-=- 7 -::-
D2 0.7 µH

(a)

Broadband +12 v
ampli fier

(b)
f igure 3-16 In (a), T1 and T2 are toroidal transformers as used in a
balanced diode doubler. The broadband amplifier in (b) contains a 4:1
toroidal transformer (Tl). The amplifier is suitable for use after the
doubler in (a).

An iron core with aµ; of 8 is a good choice at T2. If the core diameter is 0.5
in. (Amidon T 50-6), a winding of 13 turns is required. No. 20 enameled wire
is recommended. The output tap on the transformer secondary is chosen to
provide an impedance match to the doubler load. If it is followed by a
broadband amplifier li ke the one shown in Fig . 3-15b, the tap should be set
for 50 Q .

98
Sec. 3-2 Broadband Transformers 99

By using a broadband transformer in the manner illustrated in Fig.

3-15b, the amplifier will have a flat response over several octaves . One ad-
vantage of this circuit is that it has a 50-Q characteristic at the input and
output ports. It is a stable and predictable circuit by virtue of the shunt
feedback from collector to base, and because of the degenerative feedback
provided by R 1.
Ql is chosen for a high fr if considerable amplifier bandwidth is re-
quired. The device specified has an fr in excess of 1000 MHz. Tl is a 4:1
unbalanced-to-unbalanced transformer which steps the 200-Q collector im-
pedance down to 50 Q. A suitable transformer would consist of 15 bifilar
turns of No. 24 enameled wire on a 0.5-in.-diameter ferrite core which has a
µi of 950. This would enable the amplifier to deliver a reasonably flat output
from 3 to 30 MHz. A 125µ ferrite core would be suitable in this circuit if the
transformer winding had sufficient turns to exhibit an XL of 800 Q at the
lowest desired operating frequency.

3.2 Broadband Transformers

The broadband toroidal transformer is one of the most useful and com-
monplace components found in the RF communications industry. The form
taken by these transformers is pretty much the designer's choice, but he or
she may prefer the transmission-line transformer to the conventional type.
Somewhat better efficiencies are available from the transmission-line ver-
sions, according to some designers. However, the impedance-transforma-
tion ratios obtainable with transmission-line transformers are restricted to
specific values. The designer has relative freedom when working with con-
ventional broadband transformers: He or she can develop any required
turns ratio because the transformer windings are separate from one
another. Transmission-line transformers, conversely, consist of bifilar,
trifilar, or quadrifilar windings of some definite impedance. It is quite com-
mon to find these transformers with 25-Q windings. The wires can be laid
on the core in parallel, twisted a given number of times per inch, or they
may consist of miniature 25-Q coaxial cable.
Most broadband transformers are wound on ferrite cores because for a
specified power-handling application the ferrites provide much higher
permeabilities. This permits the use of considerably fewer turns of wire for
ensuring reduced IR losses and adequate performance at the low end of the
transformer bandwidth curve. In this instance we are considering the basic
rule given earlier, where the transformer windings should have an XL that is
equal to or greater than four times the impedance to which the windings in-
terface.
High-µ cores are ideal for broadband transformer use because the wind-
ings yield considerable inductance at the low end of the operating range,
1 00 Applying Toroidal Cores Chap. 3

while the core material tends to "disappear" as the operating frequency is

increased. Most transformers that are used from 2 to 30 MHz have a µi of
950. Higher values of permeability can be used, and on some occasions it is
practical to employ cores with a lower µi (125). The latter, however, is ap-
plied more commonly at VHF, as are cores of relatively low µ.
Proper identification of the free ends of the windings on multifilar
transformers can become an exercise in frustration. Therefore, it is prudent
to use enameled wires of different color. Some wire manufacturers produce
magnet wire that has red, green, or brown insulation. But, finding a small-
quantity supplier is often impossible. Moreover, the minimum-billing fees
imposed by some suppliers make it impractical to purchase small quantities
of wire. An acceptable alternative to purchasing color-coded enamel wire is
to use aerosol spray-can paint of various colors to set the windings apart
visually. The grease-free lengths of wire are stretched out until taut. The
spray paint is then applied and allowed to dry. Enamel paint is better than
lacquer types of spray paint, as the coating will be less prone to cracking
and faJling off when enamel is used . Blue, green, red, and white are good
colors to use for easy recognition.
3.2.1 Bifilar-Wound Transformers

Figure 3-16 shows four popular forms taken by toroidal transformers

that employ bifilar windings. Since these are broadband transformers, the
core material must be chosen to provide sufficient inductance for the lowest
operating frequency. The core cross-sectional area must be appropriate for
the power level of the circuit. For use from 1.5 to 40 MHz, cores with 125µ
or 950µ are common.
A simple phase-reversal transformer is seen in Fig. 3-16a. This is a 1:1
impedance-ratio component. Both ports are single-ended. The windings can
be formed by twisting the two conductors about 10 to 15 times per inch,
then placing them on the core as one would do with a single conductor.
An unbalanced 4: 1 transformer is illustrated schematically in Fig. 3-16b.
This configuration is especially useful when stepping down from 50 Q to the
base of a solid-state, single-ended RF amplifier. Two of these transformers
can be used in cascade, as shown in Fig. 3-17d, to obtain a 16:1 transforma-
tion. The 4: I transformer is also use ful when stepping up from a low collec-
tor impedance to 50 Q or some other relatively low impedance.
A 4:1 balun (balanced to unbalanced) is presented in Fig. 3- 16c. A
popular application for this device is between a 50-Q transmission line and
the balanced 200-Q terminals of an antenna. Similarly, it can be used bet-
ween a 75-Q coaxial line and the 300-Q terminals of a folded dipole antenna.
A number of TV receivers have used a solenoidal-wound rod type of balun
for use between a 75-Q coaxial feeder and the 300-Q input terminals of the
TV-set tuner. Toroidal-wound equivalents would be ideal for that use.
 1: 1 Phase reversa l 4: 1
(a) (b)
• : Ph asing

R;n
U nbal.
R ,"
R = -
L 2

RL RL : 4R;n
Bal.

4: 1 Balun Combmer
(c) Id)
Figure 3-18 Four classic examples of broadband toroidal transformers. Each contains a
bifilar winding .

A combiner transformer is illustrated in Fig. 3-16d . Two identical im-

pedances can be combined (R 'in and Ri 0 ), as would be the case with two
identical single-ended amplifiers, to provide a single output, RL. For a matched
condition, RL would be equivalent to Ri0 /2. Thus, if two 50-Q driving
sources were used, the load at the remaining port would be 25 Q.

3.2.2 Quadrifilar Transformers

A 9: 1 unbalanced-to-unbalanced transformer is seen at in Fig. 3-l 7a.

This configuration is often employed as a step-down or step-up coupler at
the input or output of a single-ended solid-state amplifier. It is useful also in
matching coaxial transmission lines to low-impedance antennas, such as a
series-fed quarter-wavelength vertical antenna.
In situations where a 9: I transformation is required for balanced-to-
balanced terminations, the transformer of Fig. 3-17b is sui table. It is often
used between push-pull drivers and push-pull amplifiers.
When the I: l transformer of Fig . 3-16a cannot be used , the balanced-to-
unbalanced version of Fig. 3-17c will be suitable. This circuit is idea.I for
matching 75-Q coaxial feed line to a 75-Q dipole antenna.
The cascaded 4: I transformers in Fig. 3-17d provide a 16: I transforma-
tion ratio. Tl and T2 are identical, and are the same as the e~ample given in
Fig. 3-16b.

101
 RL = 9R,n
9: 1 unbal.
to unbal. 9 : 1 bal. to bal.
(a) (b)
• : Phasing

1: 1 bal. 16: 1 unbal.

to unbal. to unbal.

(c) (d)
Figure 3-17 Multifilar windings are used on toroid cores to effect the broadband
transformers in (a) through (c). Two 4:1 transformers can be used in cascade (c) to obtain a
16: 1 impedance step-down.

3.2.3 Broadband Conventional Transformers

Although the consensus among designers seems to be that transmission-

lines transformers are more effi cient than "conventional" transformers
are, the fixed-impedance ratios of the former are somewhat restrictive. We
can avoid those sometimes confining parameters by utilizing the conven-
tional type of transformers seen in Fig. 3-18. The physical format for a
commonly used conventional broadband transformer is shown
photographically in Fig. 3-19. A pictorial representation is provided in Fig .
3-20.
The simple illustration of Fig. 3-18 shows how two broadband con ven-
tional transformers might be used with a single-ended RF power amplifier.
The 4:1 transformer of Fig. 3-16b could be used in place of T l and T2 in
this circuit, but only because the transformation is 4: l at both amplifier
ports. Tl and T2 of Fig. 3- 18 would consist of single toroid cores made of
ferrite. Since the operating frequency is fairly high, and because the large
winding of each transformer should have an X L of 200 Q (1.06 µ H at 30
MHz), a low-permeability material will suffice . A 0.5-in.-diameter core
with a µ of 125 would be a suitable choice for Tl. A pair of 1-in.-diameter

102
 Class C
RF PA
son

son rin ir
30 MHz To
MRF449 harmonic
30 MHz
filter
(30W)

:~·~~' ~ II ':.""

-:- 0.001 µF

L1 11
01;i
Gain "" 13 dB .---4 t+-- ------0 +13.6 v
1. z coll --
v 2
cc - 3n
J- 22 µF
2Po -
2. Turns ratio: Tl =4:1 , T2 = 4: 1
3. XL: Tl (pri), T2 (sec) ""4 x 50 S1 = 200 S1
:. L : Tl (pri). T2 (sec) = 1.06 µH
Figure 3-18 Conventional broadband toroidal transformers IT1 and T2)
are shown in a Class C RF power amplifier.

Figure 3-19 Conventional transformer that might be found in

a variety of RF power amplifiers.

ferrite cores (125µi) can be stacked atop one another at T2 to provide ample
B0 P characteristics. Lt is a decoupling choke. A ferrite core capable of
handling the de current flow and able to provide a few µH of inductance will
satisfy the requirement.

103
 Top view
toroids

End
t ""' End
plate plate
A B

"t/
Toroids
(a) (b)

Etch

End plate A

Tubing

End plate A

Copper
Tubi ng

2
End plate B End plate B

(c) (d)
Figure 3-20 Circuit and structural details for a conventional transformer that uses
rows of toroid cores. The end plates are made of printed-circuit-board material [{a) and
!ell. An assembled transformer is seen in (d).

When push-pull power amplifiers are used, as in the circuits of Figs.

3-21 and 3-22, the construction method seen in Fig. 3-20 is a good choice.
Two rows of high-permeability ferrite cores (950µ; typical for 2 to 30 MHz)
are laid side by side as in Fig. 3-20b. The cross-sectional area of the stacked
cores is selected for the power level of the amplifier. A piece of copper or
brass tubing is passed through the center of each row of cores. The ends of
the tubes are made common at end B of the transformer, but they are
isolated from one another at end A, as seen in Fig. 3-20c. Printed-circuit-
board material is suitable for making the end plates. End A is etched to pre-
vent the tubing ends from being grounded or in contact electrically. This ef-
fective U-shaped tubing conductor constitutes a one-turn winding of the
transformer. End plate B represents the center tap of that winding, or ter-
minal 4 of Fig. 3-20a.

104
 Linear amplifier

R,
50$1 To 50Q
input fil ters
{4W) (100 W)
2-30 MHz

II 01 µF 2-~rr
0. 1 µF

...__-AA.A A------..________,., + 13.6 V

+
Dl
122µF

Agure 3-21 Circuit diagram of a typical push-pull solid-state RF power amplifier in

which broadband toroidal transformers !Tl, T2, and T3) are used.

The larger transformer winding is formed by looping the appropriate

number of wire turns through the two pieces of tubing (terminals 1 and 2 of
Fig. 3-20d). The turns ratio is set by considering the tubing as a one-turn
winding, then using the proper number of wire turns through the core
stacks. This conventional transformer is very compact and symmetrical.
Once the transformer is assembled it is a simple matter to affix it to the
main etched-circuit board: end plates A and B are soldered to the ap-
propriate circuit-board pads.
A circuit example of a broadband linear RF power amplifier is given in
Fig. 3-21. Tl and T3 are conventional broadband transformers of the type
presented in Fig. 3-20. T2 is a phase-reversal transformer that can be wound
on a ferrite toroid. This transformer has a twofold advantage: operating
voltage for the collectors of QI and Q2 is supplied through the windings of
T2, but saturation is not a problem because the current flowing through the
opposing windings produces opposite fields, thereby canceling. T2 also pro-
vides a 180 ° phase reversal, placing the collectors of QI and Q2 in balance.
Forward bias for the transistors is supplied through the secondary center
tap of Tl. DI regulates the bias at approximately 0.7 V. R1 and c1 comprise

105
1 06 Applying Toroidal Cores Chap. 3

the negative feedback networks for the amplifier. Rl swamps the secondary
of Tl to aid stability at low frequencies. Decoupling choke LI and the
associated bypass capacitors also help to ensure stability .
An alternative method for obtaining negative feedback is shown in Fig.
3-22. In this example Tl has a third winding for use in the feedback circuit.
The number of turns for the feedback winding is chosen to deliver the
necessary voltage with respect to the base winding of Tl . Some designers
place the fe.e dback winding in T3 rather than Tl. The effect is the same.
T2 of Fig. 3-22 serves the same purpose as T2 of Fig. 3-21. This
amplifier circuit is shown with a half-wave low-pass harmonic filter con-
nected to the output of T 3. Compensating capacitors (Cc) are connected
across the secondary of Tl and the primary of T3-a necessary step in many
broadband amplifiers that cover several octaves of frequency. RI is includ-
ed to permit precise adjustment of the bias for best linearity. Re at QI and
Q2 establishes degenerative feedback while aiding the dynamic balance of
the two transistors. A typical value for Re would be l Q.
In the example of Fig. 3-23a we find a phase-reversal transformer used
at Tl and a combiner type of transformer at T2. Both transformers are
bifilar wound on ferrite toroid cores . For illustrative purposes we have
specified a collector characteristic of 15 Q for each transistor. T2 combines
the two impedances to provide an output level of 7 .5 Q. In such an instance
the harmonic filter would be tailored to have an input impedance of 7 .5 Q
and an output characteristic of 50 Q for conventional transmission lines .
The hybrid couplers shown in Fig. 3-23a and b (T 1 through T3) can be
wound toroidally or on balun types of magnetic cores. They are used in
many designs to combine the outputs of individual power amplifiers, or to
combine the outputs of groups of power modules. It has become standard
practice to use this style of transformer at power levels as great as 1 kW
throughout the high-frequency bands.
Hybrid transformers can be reversed and used to split a single power
source into two or more branches of equal phase and amplitude. When they
are constructed and used correctly, they provide isolation amounts of be-
tween 30 and 40 dB from 1.8 to 30 MHz between the two like power
sources. This isolation enables the system to continue operating, even
though one of the power sources may fa il. Although failure of one amplifier
will result in reduced power output, a constant load impedance will exist.
This will protect power transistors from damage and will preserve the
linearity of amplifiers which operate in that mode.
The balancing resistors of Fig. 3-23 (R l and R2) keep the VSWR at a
low value even if one of the amplifiers fail. Each balancing resistor must be
twice the ohmic value of the driving source. The resistor should be rated for
at least 0.25 of the output power of the total system, and it must be the
noninductive type. Granberg treats this general subject in depth in
Motorola AN-749 (4).
 RF linear amplifier

2W Harmonic filter
son
output
L1 L2 (100 W)

l·;lff f:r f J
RF T2
in
(4W) •
-=
..L
0.1 µF.._ _ __. -:-
..... -:-
0 Gp "=' 14dB
"'
YOlµf
R,e 0.01 µF Cc: Compensating capacitor
or~ R18 : Feedback resistor

ll~L3
R 1~• ' w.t-. ' VM ' ' 0+13.6V
Bias +
set 01 1 22µ F
1oon
-:- -:-

Figure 3-22 Tl has a third winding which is used in the negative-feedback loop of this amplifier.
 Amplifier

0.1

z=15 .11

T1
T2 To
~lllLr-1,__-0 harrno n ic
RI • fi l ter
30 .11 (7.5 !1)
•

z = 15 .11
y R1(W) = 0.25P0

0.1

(a)

-
25.11
T3
T4
1:4
50 .11

n
R2 ---
12.5

·=

T2
---
25!1
R1 = 2 x 50 .11 = 100 .11
R2 = 2 x 25 n = 50 n
(b)

Agure 3-23 Circuit that shows toroidal transformer Tl as a phase-reversal collector

choke. T2 in this example is employed as a combiner transformer. In lb) is seen a complex
hybrid combiner which consists of four broadband toroidal transformers (see the text).

108
 3.2.4 Variable Impedance Matching

The conventional-transformer concept we have been discussing can be

applied as shown in Fig. 3-24a to effect a matched condition between, say,
50 Q and an unknown load, RL. Su fficient reactance must be present in the
windings to ensure minimum losses at the lowest operating frequency of the
transformer. The secondary winding of the toroidal broadband transformer
can contain as many tap points as the designer deems necessary. T l can be
inserted in the line between a known R in• inclusive of an SWR indicator,
then the taps on the secondary winding are varied until the lowest SWR is
noted- preferably an SWR of l. This type of transformer is especially
useful when matching a 50-Q transmission line to a quarter-wavelength ver-
tical antenna: the feed-po int impedance is usually unknown, owing to the
quality of the ground (conductivity) in the vicinity of the antenna. Values
from a few ohms up to 30 Q are typical. The core material is usually ferrite
(µi of 125 for HF-band applications) and the cross-sectional area is chosen
for the power level in use , in accordance with the B max rules. Teflon-
insulated wire is recommended for the transformer windings.

{a)

N.C.

XL I '=' 315rl.
Xe , -,, 31sn
Xc 2 '=' 530 Q

(b)
Figure 3-24 The transformer in (al has a tapped
secondary to permit matching va rious impedances
t o a specified primary load. In (b) is a T·matching
network that contains a narrow-band format. L1
can be built as a nonresistive rheostat, using a
toroid core as the foundation (see Fig. 3-251.

109
110 Applying Toroidal Cores Chap.3

A narrow-band type of matchi ng circuit is illustrated schematically in

Fig. 3-24b, and photographically in Fig. 3-25. LI is a toroidal inductor
which has been wound on a large powdered-iron core. The completed
assembly resembles a power rheostat in appearance . Bare wire is used for
the winding, and care is taken to prevent the turns from shorting together.
The movable arm of the variable-inductor assembly selects one turn at a
time. It must not short any two turns when it is in position. The lower end
of L1 (marked "N.C.") is left open to prevent a shorted-turn condition .
Under certain conditions the circuit is a high-pass T network. Depending
upon the values of impedance being matched, a bandpass response can
result. Cl, C2, and LI are adjusted in sequence to obtain the lowest
reflected power possible. In terms of network loss, C2 should be set for the
maximum amount of capacitance that will permit a matched condition with
respect to the settings of Cl and LI. Since considerable RF voltage can be
present when Rl is of high impedance, insulating material of high dielectric
quality should be used between the winding and the toroid core. Heavy-
gauge silver-plated wire is ideal for the winding of LI. Rule-of-thumb reac-
tances are listed for Cl, C2, and LI of Fig. 3-24b.
The limitation of the inductor shown in Fig. 3-24b is that the resolution

Figure 3-25 Commercially made toroidal type

of variable inductor. This unit is manufactured by
Ten Tee Corp. for use in their antenna
Transmatch. The wiper is built so that it contacts
one wire turn at a time, thereby preventing the
shorted-turn effect.
Sec. 3-2 Broadband Transformers 111

of inductance must be in one-turn increments. Under some circumstances

this can present a problem, especially if the core has high permeability: one
turn can represent considerable inductance. A continuously variable induc-
tor can be realized by adopting the technique seen in Fig . 3-26. One half of a
powdered-iron toroid core is obtained by sawing the toroid in half. The core
half is affixed to an insulated arm which is connected to a Y4-in.-diameter
brass shaft, as shown . A semicircular air-wound coil is mounted opposite
the core half. It has an ID that permits free passage of the core half when
the tuning shaft is rotated. The overall assembly is similar to that of a small
potentiometer such as one might use as an audio-gain control. For wide-
range inductances the designer can switch additional inductance in series
with the variable one, or fixed-value inductance can be placed in parallel
with Ll to secure lower minimum-inductance values. This type of unit is
ideal for remote tuning by means of a motor. It is useful also as a panel-
mounted control. Circuit connections are made to points X and Y.
A versatile variable-impedance broadband matching transformer is
presented in Fig. 3-27. It was developed by J. Sevick of Bell Labs. It consists
of a fairly large toroid core on which a quadrifilar winding is placed. Im-
pedance transformations of 4: I, 9: I, and 16: 1 are achieved easily with the
basic configuration shown. Additionally, points X, Y, and Z can be attached
to the windings at the appropriate locations to obtain ratios as low as 1.5: I.
When matching between relatively low impedances, say 50 Q to values

Tuning Plexiglass
shaft / header

Air-wound
Toroid inductor
core )~--+--- glued to
half
header

Epoxy -cement
core end to
insulating arm

Figure 3-28 How a variable inductor can be made from one -half
of a toroid core. This scheme was devised by the author in 1970 for
use in a low-power antenna Transmatch for field use. As the core
half enters the 180° air-wound coil, the inductance increases.
Mechanical details were developed by A. Pfieffer of Old Lyme,
Connecticut.
 Tl

Figure 3-27 Circuit diagram of a wide-range

toroidal ma tching transformer which will pro-
vide ratios from 1.5:1 up to 16:1.

below 50 Q, the transformer can handle up to I kW of power. This is based

on a toroid-core diameter of 2.5 in. and a winding that employs No. 14 or
larger enameled wire. For the medium and high frequencies, core
permeabilities of 125 and 40 are common . Cores with 950 µi would be
suitable for the broadcast band and lower. A IO-turn quadrifilar winding on
a 2.5-in.-diameter toroid core (µi = 125) would be satisfactory for use from
2 to 30 MHz. The transformer leads should be kept as short as possible to
minimize unwanted inductances. This is especially significant at very low
impedance levels at the upper part of the high-frequency spectrum.

3.3 Antenna Loading

An interesting application for toroids or ferromagnetic sleeves is illustrated

in Fig. 3-28. A dipole antenna can be made significantly shorter while main-
taining the proper half-wavelength electrical characteristic by loading it
with ferrite. Experiments conducted in the VHF spectrum indicated that
physical-length reductions as great as 2 were entirely possible by using low-µ
ferrites over the two halves of the dipole. The voltage and current distribu-
tion of the system remained linear, and no measurable degradation in effi-
ciency was observed. The larger the cross-sectional area of the toroids, the
greater the size reduction. A similar result would be had by increasing the 1-'e
of the ferrite material, consistent with suitable Q versus frequency .

112
 Low·Z
balanced
line
Figure 3-28 Ferrite cores or sleeves can be placed over the con-
ductors of a dipole antenna to decrease the antenna length. Field-
strength measurements indicated no measureable degradation in
antenna performance as a resu lt of the loaded dimensions.

3.4 Toroidal Distributed Capacitance

When designing the transformers discussed earlier in this chapter, it is im-

portant to consider the distributed capacitance of the inductor or
transformer. Many things relate to the distributed capacitance. Part of the
problem results from the differential voltage between adjacent turns . The
amount of capacitance is dependent upon the type of insulation (dielectric)
on the wire and the core material. If moisture is present-even in small
quantities- it also affects the capacitance. This parasitic capacitance can be
regarded as a fixed-value capacitor in parallel with the toroid winding. The
winding inductance and stray capacitance establish the self-resonant fre-
quency of the inductor.
If the operating frequency of the toroid is too nea r that of self-
resonance, the circuit Q is lowered and the apparent inductance increases.
T his is particularly prominent when high operating frequencies are used and
when the toroid has high inductance. T o illustrate the effects of this un-
wanted condition, assume that the operating frequency is lOOJo of the self-
resonant frequency. The resultant actual Q will be roughly 990Jo of the com-
puted value. The condition worsens as the operating frequency is brought
closer to the self-resonant one: a differential of 50% will degrade the
calculated Q to 750'/o.
Reduction of the capacita nce can be realized by adopting specific wind-
ing techniques. For example, a scramble-wound core that occupies 360° of
the toroid will have relatively high stray capacitance. A core with , say, 900
turns will exhibit a parasitic capacitance of 40 to 70 pF. This can be improved
by using fewer degrees of core circumference, such as 325 °.
Some designers use a progressive or modified-bank style of winding. A
gap of ~ to ~ in. is maintained between the ends of the winding. The stray
capacitance will be reduced some 500'/o over the scramble-wound 360° type

11 3
114 Applying Toroidal Cores Chap.3

of toroid winding. It should be noted that the encapsulation material used

for potting toroidal inductors and transformers will add somewhat to the
distributed capacitance.

3.5 Power Supplies

Toroids, as well as U cores, pot cores, and E cores, are used extensively in
transformers for de-to-de converters and dc-to-ac inverters. The principle
of operation for these power supplies is relatively simple: The transformer is
wound on a core that will provide a hysteresis loop whose shape is nearly
that of a rectangle, as is seen in Fig. 3-29. This type of transformer is found
in Fig. 3-30 (Tl) .
QI and Q2 of the same circuit operate as switches. When one is off the
other is on, and vice versa. So that we may illustrate the operational con-
cept, let us assume that QI is conducting and that Q2 is in cutoff. This con-
dition places the input voltage across the top half of Npri of Fig. 3.30. As
this occurs, a voltage is induced in all the Tl windings. When QI conducts,
instantaneous current and voltage peaks reach maximum in the transformer
windings. This condition will be maintained until the core of Tl saturates.
At this time the flux change rate will drop to zero, and the induced voltages
will decline to zero. As this happens, the base drive disappears from QI. At
this period the current decreases, causing the flux to increase in the opposite
direction . Since a voltage of opposite polarity is induced in the windings of

Ideal saturable Core chara cteristic

Rgure 3-29 An ideal hysteresis loop

for a saturable core would be a perfect
rectangle, as shown here. (Courtesy of
Magnetics Division of Spang Industries,
Inc.)
 f0 "' 800 Hz

30 K
10W +300 v

50 µF C1
R3
450V

P0 = 145 W (483 mA)

Eff. "' 83%@ 145 W

+14 v
20A

• - Heat sink,
40 sq. inches, NPRt - 40TS CT, no. 12enam. (bifilar)
i
inch thick
NFB - 12TS CT, no. 18enam. (bifilar)
or greater
Nsec - 448 TS CT, no. 23 enam.
Wind on magnetics inc. 51001 - 2A
toroid core. Kilogausses "" 15.
01, 02 - Motorola Core type: Orthonol, 2.5" OD, 1.5" ID,
HEP G6006 or equivalent. 0.5" thick, tape wound.
Figure 3-30 Basic circuit for a de-to-de converter in which a saturable transformer (Tl) is
used. This circuit is patterned after one that was published by Delco Division of GM Corp. in
AN-18, an application note.

Tl, Q2 goes into conduction. From this point onward, the cycle keeps
repeating.
A square wave will appear across Nsec of Fig. 3-30. The frequency of
the QJ, Q2 switching action is determined by number of primary turns on
Tl, the de operating voltage, and the flux characteristics of the toroid core.
Hence,

N, . Ebb x 108
pn = 4BsAf

where Ebb is the supply voltage, Npri the number of turns in one-half the
primary winding, B5 the saturation flux density in lines per square inch, A
the core area in square inches, and/ the frequency o f switching in hertz. The

115
116 Applying Toroidal Cores Chap. 3

frequency, from no load to full load, should not change more than 5% if
the design is done correctly.
Oscillation is started by virtue of Rl and R2. These resistances provide
forward bias for Q 1 and Q2. This causes the transistors to operate above the
low-current, nonlinear portion of the curve. A full-wave bridge rectifier is
formed by using Dl through 04, as shown. R3 is a minimum-load resistor.
It is necessary to prevent the 50-µF filter capacitor from charging to the
peak value of the spike voltage which is present. It serves also as a safety
measure against shock hazard by bleeding the de voltage from Cl when the
power supply is inoperative.
The turns on Nsec can be varied in number to obtain the desired output
voltage. The power rating of the composite supply must be kept in mind if
this is done.

3. 5. 1 Using Power FETs as Switches

A circuit similar to that of Fig. 3-30 is given in Fig. 3-31. The major dif-
ference is that Ql and Q2 are VMOS power FETs. The use of power FETs
enables the designer to employ smaller components for a given power level.
This is because the upper frequency limit of FET switches is much higher
(100 MHz or greater), thereby permitting switching frequencies of 50 kHz
or greater. Smaller toroid cores can be used and the oscillator efficiency is
high. The filter capacitor can be much lower in value than at the lower switch-

01

---<>+Vo
+
02 l

+Bias
Figure 3-31 Power FETs (VMOSl are ideal for use with a saturable
transformer. The basic circuit seen here is capable of highly efficient
operation.
Sec. 3-6 Core Doping 117

ing frequencies of a few hundred hertz to 3 kHz. The Siliconix VN66AJs

shown are but one type of VMOS power FET suitable for high-current,
high-speed switching.
Other advantages are offered by enhancement-mode power FETs: they
are not subject to destruction from thermal runaway or secondary
breakdown. They cannot be destroyed when operated into an open or
shorted load. Of more than casual interest to the designer is the low input
and output capacitance characteristics of these FETs: the input amount is
typically about 50 pF and the output is on the order of 35 pF. This explains
the fast switching time that they will provide. Power FETs are somewhat
more efficient at the higher operating voltages, so the circuit shown in Fig.
3-31 would be ideal for use from a 24- to 28-V de supply. Dots indicate the
polarity of the feedback and primary windings of Tl. The same type of core
material specified for TI of Fig. 3-30 (Magnetics, Inc., Orthonol) is recom-
mended for the lower oscillator frequencies . A ferrite toroid core would be
more appropriate for high operating frequencies (10 to 50 kHz).
Figure 3-32 shows how an Orthonol core compares to others which are
used for de-to-de converters and dc-to-ac inverters. The B- H loop closely
approaches the classic rectangle discussed earlier.
The core Joss versus flux density and frequency is shown in the curves of
Fig. 3-33. The curves are for 2-mil Orthonol such as that used in Tl of Fig.
3-30.

3.6 Core Doping

There is seldom an application for toroidal inductors or transformers in

which some form of protection against dirt, abrasion, and moisture is not
used. The final component is usually coated with a low-loss, durable protec-
tive lacquer or compound. This can take the form of simple impregnation
of the core and winding with coil dope, or in a more elaborate situation the
assembly may be encapsulated by means of a potting compound.
There are a number of reasons why we might want to seal the winding of
a toroidal-wound component. First, movement of the coil or transformer
turns has a significant effect on the effective inductance. This is especially
true when a single-layer winding is applied to a core and when there is space
between the turns. Any change in inductance can have a serious effect on
circuits which perform as oscillators or filters. In the case of an oscillator,
the frequency stability can be substandard if the turns are allowed to move.
Furthermore, if the overall piece of equipment in which the toroid is used
happens to be in an environment where vibration takes place, frequency
modulation may result when the turns of the coil vibrate. A narrowband
filter will not hold alignment if the inductors are incapable of being stable.
 8 {kG)

22
8-H loops of Supermendur
saturabl e materials
20

18

16

Orthonol
14

12

10

B
Permalloy 80

6
5000 Perrn.
\er rite
4

0.5 0.4 0.3 0.2 0.1 0 0.1 0.2 0.3 0.4 0.5
H{Oe)
Agura 3-32 Comparison of core types for use in saturable reactors. The Orthonol core has an
almost ideal rectangular hysteresis characteristic. (Courtesy of Magnetics Division of Spang
Industries, Inc.).

Therefore, it is prudent to affix the coil turns to the core material in filters
as well as oscillators.

3.6.1 Simple Doping Methods

Figure 3-34 shows two broadband, trifilar-wound toroidal transformers

being used in a doubly balanced passive mixer. The transformer windings
have been sealed by means of polystyrene-base coil cement (Polystyrene
Q-Dope). This is a high-dielectric substance which has a tough consistency.
Three coats of cement were applied to the toroids shown in the photograph.
The last coating was allowed to flow through the core center and around the

118
 100~~~~~~~~~~~~~~~~~~~~~~~
90
80
70
60
50
40 2 mil orthonol
30

20

1 1--~~~~~+-~~--.f-~~~~+-~~~+-+-~----4
0.9
0.8
0.7
0.6
0.5
0.4

0.3

0.2

CJ "'
0
CJ
"'Gl
Flux density (gauss)
Figure 3-33 Flux density and frequency of operation versus core loss is il-
lustrated here. These curves are based on 2-mil Orthonol core material.
!Courtesy of Magnetics Division of Spang Industries, Inc.).

outer edges. This helps keep the transformers in place on the printed-circuit
board. There is no copper conductor on the component side of the printed-
circuit board. Had there been, each transformer would have been elevated
at least Y4 in. above the copper ground plane, using an insulating pedestal.
Suc.h an elevating insulator serves two purposes: it prevents abrasion of the
windings and consequent short-circuit conditions. Also, placing the trans-
former somewhat above the copper ground plane reduces unwanted
capacitive effects between the windings and ground. In a symmetrical cir-
cuit, such as a balanced mixer, all stray capacitance and inductance must be
kept to a minimum.

119
 r

Figure 3-34 The two toroids shown in this module are trifilar wound. They have been
doped with polystyrene Q-dope 10 seal them against moisture.

An example of toroid doping with polystyrene cement is offered in Fig.

3-35. Each of the three toroids has been given several coats of Q-Dope. The
large toroid is part of the tank circuit of a FET variable-frequency
oscillator. Directly below it are three temperature-stable polystyrene
capacitors. They a re part of the oscillator LC circuit. The capacitors have
also been affixed to the pc board by means of Q-Dope, thereby preventing
unwanted instability from vibration.
Another desirable end result of coil doping is that moisture is kept out of
the windings. As we learned earlier in the chapter, the moisture content in
the air has an effect on the distributed capacitance between the turns of a
toroidal inductor. In the interest of stability, the distributed capacitance
must remain unchanged.
In situations where we might wish to stand the toroid on end for ease of
mounting and to lessen unwanted capacitive effects, the method illustrated
in Fig. 3-36 is suggested. The toroid is first "gunked" with Q-dope and
allowed to dry. Then a generous dab of GE Silastic compound or RTV
sealant is dropped on the printed-circuit board. The toroid, which has
already been soldered to the circuit board, is allowed to "float" in the ce-
ment as shown. Once the compound has set, the toroid is held securely in
place. The compound also functions as an effective shock absorber.

3.6.2 Potting Technique

It is standard practice to imbed low-frequency toroids in potting com-
pound. Many Mil-Spec products require potting procedures to protect the
component from moisture, oil , dirt, and fungus. Although the potting com-
pound always has some effect on the electrical characteristics of a toroidal

120
 Figure 3-35 Three toroids are seen on this pc board. Each has been coated
•' I
generously with Q-Dope to affix them to the circuit board and protect the windings
from abrasion and moisture.

Toroid

PC board
Figure 3-38 A toroid can be edge -mounted by
floating it in a dab of silicone cement such as GE
Silastic or RTV sealant. The cement has a
rubber-like consistency when dry, thereby func-
tioning as a shock mount.

inductor or transformer, the problem is of minor consequence at the lower

frequencies .
Various kinds of chemical compounds are available, and each has a
specific stability factor and dielectric quality. Once the appropriate
substance is selected by the designer, the toroid can be encapsulated in the
manner shown at Fig. 3-37. Although the illustration depicts but one of
many potting formats, it is one of the most common ones used by the industry.
The inductor or transformer is wound, tested, then connected to a piece

121
 Mounting

I/stud~

(b)
Toroid
terminals

Epoxy potting
block

(a)

Figure 3-37 The longevity of toroidal transformers or inductors can be

enhanced greatly by placing them in a potted assembly of the type shown
here. Most Mil-Spec contracts call for this kind of encapsulation.

of circuit board that has been fitted with solder terminals. A mold of
suitable dimensions is chosen next. It is coated internally with silicone
grease to prevent the potting compound from adhering to it. The toroidal
assembly and terminal board is inserted in the mold, then the potting com-
pound is poured. Mounting studs are also contained on the printed-circuit-
board header if mounting is to be done at the terminals. Alternatively, as
seen in Fig. 3-37, the studs can be affixed to one side of the finished pro-
duct. If the latter method is used, the mold must have holes on one wall to
accommodate the studs.
Many potting compounds must be baked for a specified period of time
after they have been poured in a mold. This curing process ensures a stable
and reliable component for field use.
Various dyes are available for making the encapsulating material prac-
tically any color the designer specifies. Some casting resins which are sold in
hobby stores are entirely suitable for one-shot laboratory experiments with
encapsulation. Few of them require oven baking, and dyes can be mixed in-
to the resin as desired.

REFERENCES

[I] JOHN D. LENK, Handbook of Modern Solid-State Amplifiers. Englewood Cliffs,

N.J.: Prentice-Hall, Inc., 1974.
[2] M.F. " Douc" Df.MAw, ARRL Electronics Data Book. Newington, CT: ARRL,
Inc., 1976.
[3] W ES HAYWARD AND M.F. "Douc" DEMAW, Solid State Design for the Radio
Amateur. Newington, CT: ARRL, Inc., 1977.

122
4

BEADS, SLEEVES,
AND POT CORES

Although we covered the subject of toroids in Chapter 3, the ferrite

beads and sleeves treated in this section are similar to their doughnut-
shaped cousins, the toroids. The principal difference between beads and
toroids is that the former are of a different aspect ratio, physically, as can
be seen in the photograph of Figure 4-1 . The diameter /length ratio of beads
is greater than for toroids . Similarly, the ID/ OD ratio is greater, and beads

, I
I I ••
Figure 4-1 Pot core and some ferrite beads.

123
124 Beads, Sleeves, and Pot Cores Chap. 4

are generally much smaller overall than the typical toroid core. Finally, the
applications to which beads are put are somewhat apart from those of
toroids.
We can envision the ferrite sleeve as a long ferrite bead which has a
relatively thin wall, comparatively speaking. It can be used for many of the
tasks assigned to the bead, as we shall learn in this chapter. For most circuit
applications the bead, sleeve, and pot core are used to surround the conduc-
tor (single wire or coil), whereas the conductor is on the outer surface of a
toroid core. Thus, the bead, sleeve, and pot core not only increase the in-
ductance of the conductor, but also serve as shield material for the induc-
tor.
Pot cores and cup cores are two-piece units. One-half of the ferrite
cylinder contains an insulating bobbin on which an inductor or transformer
winding is placed. The remaining core half fits over the bobbin and mates
with the first half of the core. The two sections are tightened together by
means of a nonconductive screw-and-nut assembly. The resultant device is
an enclosed, shielded transformer or inductor.

4.1 Properties of Beads

The physical and electrical characteristics of the ferrite bead are shown in
Fig. 4-2. A blank bead is depicted in part (a). Part (b) shows a single-wire
conductor passed through the bead. Part (c) indicates the existence of Rand
L components that result from the assembly seen at B. When a ferrite bead
is inserted in series with an ac voltage, the condition shown in Fig. 4-3
prevails. Therefore, the impedance exhibited by one or more beads slipped
over a conductor can be used to considerable advantage when it is desired to
suppress a selected frequency or band of frequencies. The amount of sup-
pression is dependent upon the 1-'e of the ferrite, the size of the bead, and the
frequency of the suppressed energy . The impedance presented by the ferrite
material can be increased by winding two or more turns of wire through the
bead. This is seen in Fig. 4-4a. An alternative but more costly method for
increasing the series impedance is shown in part (b), where two or more

I
I
I
I
Bead
Bead Conductor Equivalen t ci rcuit

(a) (b) (c)

Figure 4-2 Physical and electrical comparison of a ferrite bead.
 zbead

R L

Figure 4-3 Schematic representa-

tion of a ferrite bead when inserted in
a signal path.

Zl Z2 Z3

Bead Bead Bead

(b)

(a)
Figure 4-4 The effective inductance of a ferrite-loaded impedance can be increased
(a) by looping one or more wire turns through a single bead, or (b) adding beads in
series on a single conductor.

beads can be used in a string to exceed the inductance of a single compo-

nent.
An increase in bead/conductor inductance can be achieved easily by
utilizing a slightly different form of bead-one which has two or more holes
through it. Figure 4-5 contains examples of three popular multihole beads.
The device seen in part (c) is called a balun core. The two holes are of
somewhat larger diameter than those found in ferrite beads. The bead
shown in Fig. 4-2a, plus those in Fig. 4-5a and b, are not suitable for passing
more than one turn of medium-gauge wire through their bores. The balun
core in Fig. 4-5c is ideal when several turns of wire must be wound through
the core holes.
Some manufacturers will supply ferrite beads that are assembled on No.
22 A WO tinned copper wire. Fair-Rite Products Corp. is one such
fabricator. The beads can be supplied , tapl".d, and reeled for use by
automatic insertion apparatus for printed-circuit boards. They are supplied
in accordance with EIA Standard RS-296-C.
Ferrite beads are available in a variety of permeabilities. Material withµ;

125
 D
t_:Y 0 0
(a) (b) (c)

Figure 4-5 Ferrites are available for use as baluns or single impedances in the
forms seen at la) through (cl. The number of holes will depend on the application .

values of 2500 and 5000 is excellent for frequencies up to 40 MHz. From 40

to 200 MHz a bead with 850 µ;is recommended. Above 200 MHz it is best to
employ bead material that has a 250 I-'; characteristic.
The absorption of ferrite beads can be enhanced measurably by using a
bead and a capacitor in series, but with the load in parallel with the
capacitor (Fig. 4-6). The resultant damping value is equivalent to the bead
impedance multiplied by the reactance of the capacitor. It can be seen from
this that the three beads shown in Fig. 4-6 would by themselves exhibit a
series impedance of 75 Q at 2.2 MHz. Through the addition of Cl, the ef-
fective damping impedance has been elevated to 5422.5 Q. The benefits of
this method are especially significant when it is necessary to filter unwanted
frequencies from power leads. Attenuation amounts of 50 dB or greater are
commonplace at VHF and higher. in RFI and EM I filtering applications it
is standard practice to use feedthrough capacitors in combination with fer-
rite beads at the entrance and exit points of shield enclosures that contain
sensitive circuits. The feedthrough capacitor serves also as a connection ter-
minal for the internal and external circuit wiring.

4.1.1 Shield-Bead Attenuation

In actual design work it is often necessary to specify the attenuation re-
quired of a bead or string of beads. At such times it is a design requirement
to convert this information into an equivalent impedance. We will now ex-
amine the relationship between the two kinds of data. We will learn how to
convert one type of requirement into the other by means of equations.
The insertion loss or attenuation of a bead in decibels is obtained by

attenuation = 20 log,0 EL
ELB
where EL is the voltage across the load without the bead in the signal path,
and ELB is the voltage across the load after the bead has been inserted.
Figure 4-7a and b show the two conditions of the equation. A desired
amount of attenuation can be obtained through empirical means by simply

126
 0.001 µF
Xe = 72.3 D.
C1

Zdamp = Z r x Xe = 5422.5 n
Zl , Z2. Z3: µ1 =5000
Figure 4-8 Signal absorption can be increased
greatly by using a capacitor in combination with fer·
rite beads, as shown here. The load is in parallel with
Cl.

. EL
Attenua t ion = 20 log 10 -
ELS

= 20 log 10 -
3 = 16.9 dB
{a)
0.7

{bl
Figure 4-7 How the attenuation of a ferrite bead {Zs) can be determined
(see the text).

127
128 Beads, Sleeves, and Pot Cores Chap. 4

adding beads and making comparative voltage measurements until the

design objective is met. The list of circuit designators in Fig. 4-7 is:

Es source voltage.
EL load voltage before adding a shield bead.
Ew load voltage after adding a bead.
IA circuit current without a bead.
ls circuit current after adding a bead .
Zc source impedance.
Zs impedance of bead.
zl load impedance.

Attenuation in relationship to shield-bead impedance can be extracted from

the following equations, which apply to the foregoing list of symbols and
Fig. 4-7.

EL = I AZL = Es X ZL
Zc + zL
and Eis = laZL = Es x Zi
Zc +Zs + ZL
EL EsZil(Z0 + Zi.)
Also, attenuation = 20 log,o Eia = 20 log ,0 EsZLl(Zc + Zs + Zi.)

Zc + Za + ZL
and attenuation = 20 log, 0 Z
2G + L

It is readily apparent from the equations that attenuation is related to

the shield-bead impedance, the generator impedance, and the load im-
pedance. It is necessary to know or be able to approximate the impedances
of the generator and load in order to specify one in terms of the other.
A typical set of data for a specified ferrite bead is given in Fig. 4-8. T he
curves for impedance, series inductance, and resistance have been drawn
versus frequency for a subminiature bead that has a µi of 5000. Curves for
other types and sizes of beads are available from the manufacturers.
A group of curves for an Amidon Assoc. FB!Ol-43 ferrite bead are
presented in Figs. 4-9, 4-10, and 4-11. The bead has a µi of 950, an OD of
0.138 in., an ID of 0.051 in., and a length of 0.118 in. Measurements were
made with a 0.5-in. length of wire through the bead. Figure 4-9 contains a
curve for operating frequency versus impedance in ohms.
Two curves are seen in the example at Fig. 4-10. One is for resistance
 Bead Dimensions

Magnetic Path Magnet ic Cross

OD (in.) ID (in.) H (in.)
Length'• (cm) Section A, {cm 2 )

0. 138±0.008 0.051±0.004 0.128 ±0.010 0.64 0.0033

Magnetic Properties

Initial permeability µ , ± 30% 5000

Loss factor (tan ofµ) at 100 kHz 10 x 10 - 5

Hysteresis factor (h / µ2) 1 X 10 -S

Saturation flux density (8, ) at 5 o. 4000 gauss

Remanence (8,) 1000 gauss

Coercivity (He) 0.1 oersted

Resistivity 102 ohm-cm

Curie temperature (Tc) > 140 oc

Series impedance vs. frequency Series inductance and resistance vs. frequency

40 40

~ 30 30 !§
.!::. .c
0 0
t\i 20 20 Q:

10 10

0.1 0 .5 0 .10 2 5 10 20 50 100 300 0 .10.S 0 . 10 I 2 10 20 50 100

Frequency (MHz) Frequency (MHz )

FigU'9 4-8 Electrical specifications for a given ferrite bead which has a µi of 5000. It is assumed
that the bead is slipped over a straight piece of wire with pigtails no longer than ~ in.
(Courtesy of Amidon Associates.)

and the other is for inductance. Both curves are referenced to operating fre-
quency.
The curve of Fig. 4-1 l shows the relationship between attenuation and
frequency . It can be seen that as the operating frequency is increased, the at-
tenuation in decibels escalates almost linearly.

129
 Impedance vs. lrequency
so ~~~~~~~~~~~~~~~~~~~~~~~~~

60

tc 40

".,"'a.
E
20

10 50 100 150 200 300

Frequency (MHz}
Figure 4-9 Operating frequency versus impedance for an Amidon Associates
FB101-43 ferrite bead (950µ) .

Inductance and resi stance vs. frequency

0.14 35

0.12 30
:r:
3- 0.10 25 ]
QJ
0
u Effective series L (µH) and
c 0.08 20 tc
"'
u series R (ohms) of one bead
:J on 1.5 in. or No. 22 wire 15 t;
0.06
"c ·~
a:
0.04 10

O.o2 05

10 50 100 500 1000

Frequency (MHz)
Figure 4-10 Series resistance (Rl and inductance (LI versus operating frequency for the
ferrite bead specified in Fig. 4-9.

4.1.2 Practical Applications for Beads

Many RF circuits are subject to self-oscillation in the VHF and U HF

parts of the spectrum. The common term for this condition is "parasitic
oscillation." The malady is especially common when bipolar transistors
with high f T ratings are used. The same is true o f FETs that have high max-

130
 Attenuation vs. frequency
20

18

16
co
:£ 14
c:
.g 12
"'::>c:
e 10
~

2 3 4 5 6 7 8 9
Frequency (MHz) x 102
Figura 4-11 Attenuation with respect to operating frequency while using an
Amidon Associates FB101 -43 ferrite bead.

imum operating frequencies, such as the 2N4416 class of devices. VMOS

power FETs, because of their extremely high gfs ratings (250,000 µmhos be-
ing typical) and high upper-frequency Limits, are prone to VHF parasitics .
In all cases the anomaly is encouraged by long transistor and circuit leads.
When printed circuits are used, parasitics become manifest most easily
when the etched conductors are quite long between component points that
relate to the active devices. Printed-circuit boards of the double-clad variety
are recommended in situations where the parasitic capacitances between the
ground plane and the etched foils will not impair circuit performance. This
type of board material, after etching, has a solid copper surface on the com-
ponent side of the board. It is connected to system ground and to the
ground foils on the etched side of the board. Parasitic oscillations a re
discouraged because the large ground conductor helps to break up other-
wise existing RF ground loops . Furthermore, small amounts of capacitance
exist between the etched conductors and the ground plane. This aids in damp-
ing VHF and UHF self-oscillation.
When all other preventive measures fai l to stop unwanted oscillations,
insertion of one or more ferrite beads directly at the gate or base of a tran-
sistor will help. Figure 4-12 shows this technique. Z l is a 950µ subminiature
bead. It is located as close to the case of Ql as possible. It will present
negligible reactance to the 5-MHz operating frequency, but will appear as a
large series impedance at VHF and higher. Similar results can be had by
putting the bead in the drain or collector lead, provided that those leads are
"hot" at RF. The oscillator of Fig. 4-12a has a bypassed drain, making it
impractical to insert Zl at that port. Were this circuit afflicted with VHF

13 1
 01 2N4416
5MHz

(-o Eo
50 pF

(a) Oscillator

150 MHz

~ zb.ne = 10 .n

(b) VH F amplifier

6.5 MHz

.RF o----t ,____._ _

in

(cl H F power amplifier

Figure 4-12 A ferri te bead is used in (al to suppress VHF/ UHF parasitic oscilla-
tions. Four beads are employed in (b) as a base impedance of low Q. Four beads are
applied in (c) for parasitic suppression.

parasitics, the VHF energy would be readily apparent by observing the out-
put waveform (£0) by means of a wide-band oscilloscope. The VHF energy
would be superimposed o n the 5-MHz sine wave.
Zl through Z4 of Fig. 4- 12c are used for the purpose j ust discussed . But,
because of the very high gfs of the two VMOS power FETs , and because the
devices are connected in a parallel arrangement (increased lead lengths),
parasitic oscillation is a problem. Complete stability resulted in this circuit

132
Sec. 4-1 Propert1esof8eads 133

only after a bead (950 µ) was placed in both the drain and gate leads. It is
practical to employ a 10- to 22-Q Vi-W noninductive resistor in place of Zl
and Z2 for parasitic suppression. ln fact, many designers use that techni-
que. However, if resistors were situated at Z3 and Z4, the de voltage drop
across them would be prohibitive: the beads offer a better solution.
In Fig. 4-12b we find Zl through Z4 serving a different purpose. Here
they provide an impedance across which the driving energy to Ql is sup-
plied. If we use the beads specified for the curve of Fig. 4-9, each bead will
exhibit an impedance of 38 Q at 150 MHz. A total impedance of 152 Q will
be present with the four beads- somewhat more than the minimum require-
ment of four times the base impedance specified (10 Q). The primary advan-
tage in using the beads instead of a conventional wire-wound solenoidal RF
choke is that the Q will be low for Zl through Z4 compared to a wire-wound
choke. This is a desirable condition in the interest of circuit stability. Many
times a solid-state amplifier will break into self-oscillation by virtue of a
tuned base/tuned collector condition brought on by the RF chokes in the
base and collector leads. The input and output capacitance of the transistor
helps to establish the frequency of oscillation. But, if the base-return im-
pedance is of low Q, and if the collector RF choke (RFCl) of Fig. 4-12b is
of low Q, self-oscillation is unlikely. The cost of four beads versus that of a
commercially made RF choke is similar.

4. 1.3 Beads as Q Killers

If for the purpose of manufacturing expediency we elected to use high-Q

solenoidal or layer-wound RF chokes for impedances, we might easily en-
counter RF instability. This possibility was discussed in Sec. 4.1.2, where we
recognized the potential instability of the circuit of Fig. 4-12b. An RF
power amplifier which contains two conventional RF chokes is shown in
Fig. 4-13a. RFCI and RFC2 could lead to unwanted tuned base/ tuned col-
lector instability if the self-resonant characteristics of the tuned circuits
formed by the chokes and the existing circuit capacitances were similar. One
solution to the problem would be to swamp one or both of the RF chokes by
shunting the inductor(s) with a low-ohmage, noninductive resistor. This
would lower the Q and discourage self-oscillation of Ql. The major short-
coming of using parallel resistance is that part of the ampli fier power would
be dissipated in the resistance: the trade-off of gain and stage efficiency for
stability might be prohibitive as far as the design objective is concerned.
If we were to add high-µ ferrite beads as shown at RFC! and RFC2 of
Fig. 4-13a, the Q of the chokes would drop to a very low value. Z I and Z2 in
such an example would be miniature beads with a µ; of 950. A single bead
would be placed over the ground-return pigtail of each choke, and as close
to the body of the choke as possible. Laboratory investigations, while
 (a)

Sig.
in

(b)
Figure 4-13 Zl is used in (a) and lb) to lower the circuit Q (see
the text) and enhance amplifier stability.

using a Boonton 160-A Q meter, revealed a Q reduction of 87. 5 O/o when a

single miniature 950µ bead was added to the pigtail of a I 0-µH solenoidal-
wound RF choke. The measurement was made at 25 MHz. The Q11 dropped
from 80 to 10, and the effective inductance increased approximately 1 µH.
Next, two beads were used and the measurements repeated. The Q11 degrad-
ed to approximately 3 and the inductance rose to 12 µH-a Q reduction of
some 960Jo.
We might ask ourselves at this juncture: Why not simply use a string of
ferrite beads in place of a bead/ RF choke combination? At VHF and higher
this would be the more practical technique. However, in the HF part of the
spectrum it might require a prohibitive number of beads to realize the im-
pedance required for the circuit. Therefore, it is less difficult to utilize a
wir e-wound choke and lower the Q by adding one or more ferrite beads .
There are some circuits in which the base return of a power tra11sistor
can tolerate a small amount of series resistance without imposing a threat to
the Class C transistor amplifier. In such an example we could substitute a
50- to 100-Q composition resistor for Zl. This would lower the choke Q to

1·34
Sec. 4-1 Properties of Beads 135

an acceptable level. The method is not recommended, as the reverse-

breakdown profile that resulted could damage the transistor during periods
when drive was applied. The destruction would be gradual in a normal
situation, and the fault would be observed as beta degradation.
The Q-killing procedure shown in Fig. 4-l3a can be applied in the case
of the broadband transformer, Tl, seen in Fig. 4- l3b. One or more high-
permeability beads would simply be placed over the return lead of the Tl
secondary winding, as shown. The bead permeability discussed in this sec-
tion (950) is not mandatory. The author has learned through years of circuit
design and application that a 950µ ferrite material represents a good com-
promise for circuits from 2 to 450 MHz. However, beads with lower 1-'i
characteristics can be employed provided that the desired results are realiz-
ed. At the high end of the frequency range just mentioned (100 to 450
MHz), permeabilities of 125 and 250 work quite nicely for use as shown in
Fig. 4-13.

4. 1.4 Beads for Decoupling

At VHF and higher the ferrite bead becomes one of the most effective
components available at low cost for RF filtering and decoupling. Zl
through Z6 in Fig. 4-14 are used in the Vee supply lines to prevent unwanted
migration of signal energy from one stage to the other. If feedback voltage
was permitted to travel along the Vee leads, instability could very easily
result at QI, Q2, or both. C l and C2 prevent Q degradation of RFCI and
RFC2 which would otherwise be caused by the use of Zl /Z2 and Z5/Z6.
These capacitors become part of the RF decoupling networks along with
C3, C4, and Z3/ Z4. T hree different values of capacitance are used in the

VHF VHF
driver 150 MHz

ZS

Z6 I

Vee
0.01 0.1 C4
µF
I µFI

Agure 4-14 Example of ferrite beads in the decoupling networks of th e Vccline. 02

uses fou r beads as a base-return Impedance.
136 Beads, Sleeves, and Pot Cores Chap. 4

decoupling circuit. This provides effective bypassing over a relatively wide

range of frequencies, thereby discouraging self-oscillation in and below the
VHF region. Beads that have I-'; values of 125 to 950 are suitable for the cir-
cuit of Fig. 4-14. Four beads are shown between the base of Q2 and ground.
They funct ion as a base-return impedance, as is the case with the circuit of
Fig. 4-12b.

4. 1. 5 Reducing Incidental Radiation

Certain practical and legal restrictions exist with regard to the maximum
tolerable incidental radiation from specified RF generators and transmitters.
A designer must ensure that equipment will not interfere with other pieces
of apparatus in a composite or interfacing system. Furthermore, excessive
incidental radiation from a specified module or its external wiring could
easily cause interference to other services, such as TV ' receivers, FM
receivers, or commercial two-way radio services nearby. Ferrite beads are
excellent in filtering arrangements of the type illustrated in Fig. 4-15.

Fll
C7
G:}ID Oscillator

Ll.

C7Q§D
G
-:- .L2 C1 C2

'
ca
22

C5

RF Vee Vee Vee

out 03 02 01

C1-C6, incl : 0.001 µF feedthrough

Z 1. 22, 23: 2 each 950µ; bead
F l1 : Harmonic filter
Figure 4-15 Suggested method for utilizing ferrite beads to prevent incidental radiation
or pickup of RF energy when a circuit module must be well shielded.
Sec. 4-2 Ferrite Sleeves 137

The example shows a compartmentized three-stage transmitter that is

structured in a shield-compartment format. There are seven separate sec-
tions to the RF module. It is assumed that each is RF-tight with respect to
the others. Additionally, the top and bottom openings of the assembled
module are enclosed by means of RF-tight covers. Double-clad printed-
circuit-board sections are suitable for constructing the compartments. Die-
cast aluminum boxes are frequently used for the purpose as well, but must
be cast with the desired compartments to assure ease of assembly.
The Vee supply Lines to each of the three stages in Fig. 4-15 contain effec-
tive RF filters which are composed of two feedthrough types of capacitors
(Cl through C6) and two 950µ; ferrite beads (ZI, Z2, and Z3). These net-
wor ks also serve to decouple the stages from one another, as was done at
Fig. 4-14. The feedthrough capacitors provide a mechanicaJ advantage: they
serve also as through-terminals for the operating voltages.
If the equipment being treated for incidentaJ radiation operates in the
HF spectrum or below, Zl through Z3 can be changed to large high-µ ferrite
beads which have several turns of wire looped through them. This will pro-
vide a much higher series impedance than can be secured from the miniature
beads indicated. Figure 4-4a shows the method under discussion. Another
feature of the shielding and filtering procedure of Fig . 4-15 is that unwanted
energy exterior to the RF module cannot enter the circuit via the supply
lines. Stray energy pickup is defeated because of the shield box and its com-
partments.
Table 4-1 contains a listing of various sizes and styles of ferrite beads.
Data on the various permeabilities are provided at the end of the table. The
Class 27 beads can be obtained with wire leads for use in high-production
operations. Having the leads attached permits the use of insertion
mechanisms which require components to be supplied in large tape reels.
Such machines are programmed to extract the components from the reels
(resistors, diodes, capacitors, beads, etc.), bend the pigtails, and insert the
parts in the proper pc-board holes.
The multihole beads are useful when the impedance needs to be increas-
ed over what would be presented by one or more single-hole beads. The
turns of wire are passed through the various bead holes to increase the in-
ductance. The beads with very high µ 1 values listed in Table 4-1 are excellent
for use at medium frequency and lower.

4.2 Ferrite Sleeves

There are a variety of applications for ferrite sleeves . Of significance to the

designer is the fact that a sleeve used over an axial-lead inductor will multi-
ply the inductance by a facto r of 2 to JO, depending upon the µ; of the
material and its bulk. The basic sleeve is seen in Fig. 4-16a. The right-hand
 TABLE 4·1

B
Type 4
Class 27 t.. ~cr- · 1 -1A
~~"7 8
D

Pun Number Alaterial T)'P' A /)im . JJ Dim. C Dim, DDim.

- -- - --
26 7300-ISO I 7J in O.OJ710.0J9 0.0 1910.022 0. 19210 . 198
mm 0.9J/ 0 .99 O.• SIO.l6 J.88/l.OJ
267JOOU>OI 7J in . 0.040/ 0.0JJ 0.02510.029 O. ISS/ 0 .16J
mm 1.0211.09 0.6410 .74 J.9JIJ . IJ
l67J00-1701 in. 0.0Slf0.0S6 0.02710.0J I 0.08SI0.09S
73
nun I J2 I • 2 0.69 0. 79 2 16 2.41
l64JOO..t30 1 JJ in . 0.076 0.084 O.OJJ/ O.OJB 0. 10010. 111
mm I 9J l IJ O.S.110 .91 U • 2.8•
!6JJ002l01 .IJ In . 0.06810.076 0.04110.045 0.40010.JJO
mrn l.7J 1.93 1.0.I- LI .I 10 16110.92
!673002201 in. 0.06810.076 0.04110.0.15 0.JOOI 0.4JO
13
mm I 7J, 1.9J J.CJ..l; I. IJ 10 . 16110 92
l67JOOOSO I 7J In . 0.07410.078 O.OJ J/0.045 O.OSJ I0 .067
mm i.88 J.9S l.OJ l l IJ 1.3 9 1.70
l6JJOOOl01 JJ in. 0.07.110 .078 0.0.1 110.045 0. 14010. 160
mm l.88d .9S l.0.111. 1J ' l61J ,06
l67JOOO?OI 73
in . 0.07410.078 O.OJ 110.0JS 0. 14010. ll\0
mm 1.88 I 98 I .().llJ . IJ 1.l6 • 06
l67J00-190J 7J in. 0. 107 0. 117 o.oos. 0.01 1 0•.10010.J lO
mm ~ . "'l 2 9.., I 65 1.80 10 16 10.67
l6JJOOJIOI J3 In. 0. 118 0.1 48 0.02910.0JJ 0. 16010. 190
mm J lS J -~ o.·J o.SJ • 06 J SJ
264J00Jl01 JJ in. 0. 128 0.1 48 0.0?910.0JJ 0.JJ0 10 .J70
mm ':!,!- 3 ""6 0 7J •0.!W 8 .38 9 JO
.I) in. 0. 13010. 146 0.04710.055 0 .04S r0. 0S~
l6JJOOOJO I
nun t JO P l I 19 · I AO I JJ 1 I JO

16.13000 101 JJ in . 0. 13010. lJh 0.04710.0SS 0. 11 010. 126

mm J )01) 7 1 1. 19/J .JO 2.791).20
in . 0. 1.) 010. 146 0.04 7 10.0~5 O. I IOt 0. 12<•
2"6400010 1 6J
mm J ..10'1 . ~I I 191U O 2. 11)1J,.:!O

2673000 10 1 73 in. 0. 130111. 1411 0.047 10.0SS 0. 11010. 126

mm 3.JO •J .71 l. l'l l .• O ! .i91J 20

267SOOO IOI 75 in . O. IJO I0. 146 O.OJ7t 0.0SS 0.22610.2.U,

rnm J .lO' J.7 1 1.191 1.JO ~ 'iol 6 2~
in . O. IJ0t 0. IJ6 0.041 1 0.055 O. ISOI0. 170
264300060 1 4J
mm J.JO J.7 1 1.1911.40 3.81 J .32
2673000601 in . O. IJOI0. 146 O.OJ 7t 0.0SS 0. 150' 0.170
73
mm .U01J 71 1.19 1.40 .I 81 • J2
lt.43000301 43 In. O. IJOi 0. 146 0.04710.0SS 0.226 0.246
mm UO/J.71 l.lllfi AO ~ . 7J '6.2S

266400030 1 6J in . O. IJOt 0. 1.Uo 0.0J710.0SS 0.226/ 0.24(,

min UOI J .71 1, 191 1.40 ~ . 1416 . Z~

138
 In. 0. 130/ 0. 146 0.047/ 0 .0SS 0.22610.246
2673000.101 73
1, 19/ 1.40 5.74/ 6.25
mm J.30/ 3. 71
in. 0.130 0.04710.55 0.485/ (1.515
2643000701 43 1,19/ 1.40 12.32/ IJ.06
mm UO/J.7 1
in. 0. 130/ 0.146 0.04710.055 0.485/ 0.5 15
2664000701 64 1.19/ 1,4() 12.3211 l.08
mm UO/l.7 1
in . O. IJ0/ 0.146 0.047/ 0.055 0.485/ 0.5 15
267300070 I 7,l 1.19/ 1,.10 12.3211 J .OS
mm 3.30/J .7 I
in. O. IJ5/ 0.14S 0.06510.075 0.224/ 0.244
26-13001301 43 1,65/ 1.91 5.6916.20
mm 3.43/J.68
in. 0 .15310.169 0.074/ 0 .084 0.457/ 0,487
26-13001401 43 1.8812. 13 11.611 12.37
mm 3.89/ 4 ,29
in . 0 . 15710.181 0.063/ 0.079 0.185/ 0.?09
2663001201 63 mm 3.9914 ,60 I .6012.0 1 4, 70/ 5.JI
in . 0.190/ 0.210 0.05710.067 0.240/ 0.2W
2633012401 3.1 mm ·• .S3t 5. JJ 1.4511 70 6. 1016.60
in . 0. 19010.2!0 0.05710.67 0.240/ 0.260
?6-13022401 4J 1,4511.70 6. 10 16.60
mm ·• .S315. JJ
in . 0.190/ 0.210 O.OS7/ 0.0ll7 0.4?210.452
26-13021801 43 1.4511.70 10. 7211 1.48
mm 4.8315.33
in . 0. 190/ 0.?10 0.0S710.0ll7 0.42 210.452
2664021801 64 1.4511.70 10.7211 1.48
mrn 4,8315.JJ
in . 0. 190/ 0.210 0.0S710.0ll7 0.422/ 0.452
2673021801 73 1.4511.10 10.72111.48
mm 4.8315.JJ
in. 0.203/ 0,232 (2) 0.029/ 0.041 0.456/ 0 .488
2664001101 6-1 (21 0.14/ 1.04 I U S/ 12.40
mm l .2815.89
in. 0.12610.?46 (6) O.oJ5/ 0.04 I 0.384/ 0.404
26-13225111 43 l.74/ 6.25 (61 O.S9/ l.04 9.7l/ I0.26
mm
in. 0.226/ 0.246 (6) O.OJ5/ 0.04 I 0.384/ 0.404
2664225111 64 5.74/ 6.25((0) 0.891 1.04 9.751 10.26
mrn
in. 0.24010.260 (6) 0.04510.055 0.45610.486
2664000901 64 llH1l 6. 10/ 6.60 (21 1.14/ 1.40 I U8/12.J4
In . 0.291/0.JO I 0.03910.0?9 0.287/ 0.J07
2643000801 43 7.J911.6S 2.2612.51 7.2917.80
rll01

h1. 0.291/0,JOI 0.08910.41?9 0.28710.307

266400080 IO 6-1 7.J9/7 .65 2.2612.51 7.2917.80
nun
in. 0. 29110.30 I 0.039/ 0.0?'1 0.28710.J07
26 7.100080 I 73 2.2612.51 7.2917.80
mm 7. 3917.65
in. O.J6S/ O.J85 0. 187/ (1.200 Cl.400/ 0.420
264300<180 I 43 4.7515.08 10.16/ 10.67
rnm 0.27/ 9.78
IH . O.J65/ 0.J85 0.187/ 0.200 (1.400/ 0.420
2677006801 77 4.7515.08 10. 16/ 10.67
mm 9,2719. 78
In. O.J10/ 0.J\ll) (1. 190/ 0.205 0. 180/ 0.200
2643002401 4.l 9.40/ 9.91 4 .8315.2 1 4,57/5.08
rum
in. 0.542/0.582 0.240/ 0.260 1.095/ 1.ISS
264340001 43 6 .10/ 6.60 27 .8 1/2.9.34
mm 13.77/ 1-1,78
in. O. l?K/ 0 . 148 2. 72012. 78() 0.160/ 0.1\IU : 22 A WG TCW •
l743lHHll I 43 119.0917().61 •l .0()/4.BJ )( I Y: {38.1111111)
mm 3.25/J .76
In . 0.12810 . 148 2. 720/ 2. 780 0 .JJOI0.370 = 22 r\WG TCW'
274300212 1 4J J .25/ 3.7(> 69.(19170 .6 1 8.)8/9.40 )( I Vi (38. 1mm)
mm

Avai lable bead sizes and permeab ilities. Courtesy Fair- Rite
Corp. Initial permeabilities are given here with material
type: 33 = 800 mu; 43 = 850 mu; 63 = 40 mu; 64 = 250 mu ;
73 = 2500 mu; 75 = 5000 mu; 77 = 1800 mu;

139
 (a)
~
j Sleeve
Coil
(b)

Antenna

Sleeve

Sleeve

Coaxial
line

To
equipmen t

(c)
Rgure 4-18 The basic ferrite sleeve is shown pictorially in (a). A
solenoidal inductor is shown partially inserted into the sleeve in (bl.
Ferrite sleeves can be used to decouple coaxial RF transmission lines
as shown in (cl .

illustration in part (b) shows how the sleeve and a solenoidal inductor can be
used in combination to elevate the coil inductance.
Additional to the feature of increased inductance, the sleeve functions as
an effective shield for the coil. Still another advantage in using the sleeve is
that it offers physical protection to the inductor. This eliminates the need
for a special shield can or the added burden of encapsulation. The wall
thickness of ferrite sleeves is 120/o of the sleeve outside diameter, or greater.
This is a manufacturing requirement that makes it practical to fabricate the
sleeves by means of the compaction-pressing or extrusion techniques. The
net result is beneficial to designers who choose to use the sleeve as a protec-
tive covering for the coil with which it is used.

140
Sec. 4-3 Ferrite Balun Cores 141

Sleeves are available from Fair-Rite Corp., for example, in permeabili-

ties from 40 to 1200. The available sizes range from 0.115 in. OD/ 0.155 in.
long to 1.025 in. OD/2.03 in. Jong.
Another practical use for the ferrite sleeve is seen in Fig. 4-l 6c. They can
be used over a coaxial transmission line to serve as decoupling sleeves. This
helps to prevent RF energy from traveling on the outer conductor of the
line. This technique is especially useful when arrays of gain antennas are
combined. The decoupling sleeves discourage unwanted radiation of the
harnessing sections of the array. Such radiation can destroy the array pat-
tern and gain if it is great enough in magnitude.
Similarly, fe rrite sleeves can be used in RFI/EMI suppression by placing
them over signal cables which enter and leave the various parts of an elec-
tronics system . The sleeve can be affixed at the desired location on the cable
by means of Silastic compound or epoxy cement. Heat-shrink tubing is also
excellent for securing the sleeve to the cable.

4.3 Ferrite Balun Cores

The ferrite balun core of Fig . 4-5c is an excellent substitute for a group of
toroid cores that have been combined for broadband transformer applica-
tions (Fig. 3-20). The practical limit is set by the sizes of available balun
cores. Most of them are relatively small in cross-sectional area. This places a
restriction on the amount of power that can be handled safely with ferrite
baluns.
Although the term "balun" refers specifically to a balanced-to-unbalanced
transformation of impedance levels, there is no reason why balun cores can-
not be used for all manner of conventional or transmission-line
transformers. Figure 4-17 shows a 150-W solid-state linear amplifier for use
from 2 to 30 MHz. The input transformer uses a balun core that is 'Is in.
Jong and Y:z in . wide. Driving power for the amplifier is on the order of 6 W.
The larger transformer on the module is used at the amplifier output. It is
fashioned from two I '!ii-in. lengths of ferrite tubing. Each is Y2 in. in
diameter. End plates have been made from printed-circuit board as shown
in Fig. 3-20. However, the principle of operation for the two transformers
seen in Fig. 4-17 is the same. Both are conventional transformers.
A pictorial view of how we might employ a balun core for a broadband
conventional transformer is given in Fig. 4-18. The example in part (a)
shows the core from one end. The shield braid from a small piece of coaxial
cable has been placed in the core holes, then fanned as shown. The braid is
passed through the holes so that fanning can be done at each end of the
core. Solder is flowed over the fanned areas to secure the braid in the
desired position. At the far end of the core in Fig. 4-18, the two sections of
braid are joined and soldered. This results in the winding C-D-E of Fig.
 I
Flgure 4-17 A 150-W solid-state broadband amplifier in which a balun core is used for
the input transformer. The larger transformer (output) is a conventional type which
coniains two long cylinders of ferrite . This amplifier was designed by Helge Granberg of
Motorola Semiconductor Products, Inc.

D
{join braid)

Coax
shield braid
(a}
c E
( b}

{cl
Flgure 4-18 Method of using coaxial-cable shield braid as the CDE winding of (cl
The braid is passed through the holes of a balun core as shown in (a) and (b).

142
Sec. 4-4 Pot Cores 143

4- l 8c. In effect, we have a single-turn conductor through the core, with 0

being the center tap. The same result can be obtained by using thin-wall
lengths of brass tubing in place of the shield braid.
Winding A-Bis looped through the core after the process shown at A is
completed. Winding C-0-E becomes the basis for the turns ratio of the
transformer, since it is fixed at one turn. Thus, if a 10: I impedance ratio
was required, winding A-B would consist of 3.16 turns of insulated wire. In
a practical situation we would use a three-tum winding for A-Band tolerate
the slight mismatch that would result.
In utilizing this type of transformer construction, we need to pay atten-
tion to the rule established in Chapter 3: the transformer windings must ex-
hibit a n inductive reactance that is at least four times the characteristic im-
pedance of the circuit to which the transformer winding connects.
Therefore, if Tl of Fig. 4-18 were used to match a single-ended 50-Q driving
source to a 5-Q base-to-base impedance in a push-pull solid-state amplifier,
winding C-0-E would need to have an XL of 20 Q or greater at the lowest
operating frequency of the amplifier. Assuming in this case that the lowest
frequency was 2 MHz, winding C- 0 -E would require 1.59 µH of induc-
tance, or greater. The balun core would be chosen for a permeability that
ensured the design objective .
Ferrite balun cores are used widely at the input ports of TV and FM
tuners for the purpose of converting balanced 300-Q feed line to the singlc-
ended 75-Q receiver input terminal. The cable-TV industry employs
numerous balun cores for similar applications.

4.4 Pot Cores

A typical pot core is shown before and after winding and assembly in Fig.
4-19. This device is excellent when the designer requires a large amount of
inductance and a high degree of magnetic self-shielding. For example, we
could fabricate a 4-H inductor on a pot core that had a diameter of only Y.6
in. Furthermore, excellent values of Q can be obtained when using pot
cores .
Another advantage in the use of pot cores is the reduction in manufac-
turing cost with respect to winding equipment and winding time. Pot cores
contain plastic bobbins on which the inductor or transformer windings are
placed. The winding procedure is far less complex than when toroid cores
are used for the same purpose.
Owing to the small size of pot cores, miniaturization becomes a routine
procedure when an inductor or transformer is required in a circuit. The
mounting techniques for pot cores make them compatible with etched-
circuit boards and metal chassis. An assembled pot core can be impregnated
to protect the winding from dirt, moisture, and fungus. T he impregnation
 -
Figure 4-19 Typical pot core before and after winding.

process will also help to keep the core halves tensioned properly. Any reduc-
tion of pressure will cause Lhe inductance of the pot-core winding to
decrease. Similarly, dirt or grease on the mating surfaces of the core halves
must be avoided to ensure maximum inductance. It is important, therefore,
to prevent varnish or encapsulating material from appearing on the mating
sur faces of the core. Even the thinnest of foreign-matter layers will reduce
the core AL.

4.4. 1 Pot-Core Hardware

Two common styles of pot core are depicted in Fig. 4-20. The a ssembly
seen in part (a) uses a metal mounting clamp to hold the core halves tightly
together. The base plate and the metal clamp are held together by means of
screws and nuts , thereby confining the pot core between them. Alternative-
ly, the entire assembly can be snugged against a circuit board or metal
chassis . One or more shims (washers) can be used at the bollom of the
"sandwich" to provide the desired tension . Slots in the core halves permit
bringi ng the transformer o r inductor leads out of the assembly.
The inductance adjuster screw at the cop of Fig. 4-20a is an o ptional
feature of some pot cores. It can be used to vary the core-winding induc-
ta nce by approximately ± 13 OJo. This screw contains a ferrite slug tha t
bridges the air gap within the core to increase the inductance. This feature is
especia lly useful when precise final a lignment of high-Q filter sections is re-
quired.
T he core assembly shown in Fig. 4-20b is entirely suitable for most non-
critical circuit applications. The winding inductance is predetermined within
reasonable limits of tolerance by means of the AL factor which relates to the
core being used. A nylon nut and screw is used to hold the core halves
toget her a nd to affix the completed assembly to the chassis o r pc board.

144
 Ferrite slug
~~ Inductance
'j'- adjustor

Pot core
half

Nylon
bobbin

~
-----e;;?Z77???????2?; I
:2222???2???2;1 I I
Pot core
half

Add washers
for required Base
tension plate
(a)

l
Washer

~
~ ·
Nylon
screw

Core
half

Nylon
bobbin

~
Core
halt

ITf Nut
o=i Washer

(b)

Rgure 4-20 Breakaway view of a pot-core assembly which has a

fine-adjustment screw and metal mounting clamp (a ). The
assembling in (b) mounts by means of a long nylon screw and hew
nut.

145
 II 2-section

I:IC (a)
bobbin

3-section
bobbin

(b)

End
view

(c)
Figure 4-21 Pot-core bobbins are
available with 2 (a) or more lb)
winding sections. An end view of a
bobbin is seen in (c).

Although ferrous or brass nuts and screws can be used for the purpose, they
will have some effect on the winding inductance and Q. Therefore, they
should be avoided whenever possible.
In addition to the single-section bobbin of Fig. 4-20, there are two- and
three-section bobbins available. These are seen in Fig. 4-2la and b. The
slots in the end plates of the bobbins are situated to align with the slots in
the core halves. This facilitates bringing the transformer or inductor leads
out of the core for connection to the circuit. Some manufacturers of pot
cores provide a glass-epoxy header which can be affixed to the top or bot-
tom of the core assembly. The header is fitted with the desired number of
terminal posts, consistent with the number of leads coming through the core
slots . The transformer or inductor leads are soldered to the header posts to
reduce the fragility of the leads.
4.4.2 Design Considerations
When pot cores are employed in narrow-band LC types of circuits, there
are some basic considerations which the designer must address prior to
selecting a core:

146
 1. The operating frequency.
2. Inductor Q at the chosen operating frequency.
3. Operational flux density (B0 p).
4. Required inductance.
5. Temperature coefficient of the inductor.
6. Long-term stability.
7. Available space for the pot-core assembly.

These considerations are affected significantly by the relative loss factor,

which is expressed mathematically as llµiQ. The core is chosen for the
lowest relative loss at the operating frequency, or range of operating fre-
quencies. This will help to ensure the highest Q that can be expected. The
foregoing factor reflects the relative losses in the core. It varies with
operating frequency and the actual core material being used.
Curves are available from the various manufacturers which show the
characteristics of the different types of ferrite with regard to operating Q
versus frequency. There are cross-over points where different materials may
actually coincide, but the choice of ferrite will have to be based on all of the
considerations in the seven-point criterion given earlier.
As is the case with other forms of ferrite and powdered-iron materials,
the number of turns for a given value of inductance can be determined by N
= 10 3 V LI AL or N = 100 V LIAL> where N is the required number of
turns, Lis in mH, and AL is the manufacturer's inductance index for the
core material being used. An inductance/AL/ turns nomograph is given in
Fig. 4-22.
The AL values for pot cores are standard, just as they are for toroids.
For this reason they can be specified for each model of core the manufac-
turer produces. A variety of AL values can be obtained from a given ferrite
mix by carefully grinding a defined gap between the inner centerposts of the
pot cores. Generally, the manufacturer grinds only one of the centerposts
when small gaps are desired. Both posts are milled down when larger gaps
are provided.
Quality factor (Q), as was stated earlier in the book, is a measure of the
effects of the collective losses with respect to circuit performance. The three
major loss factors are copper losses (winding), core losses, and capacitive
losses in the winding. The Q is affected, therefore, by the number of coil
turns and the manner in which they are wound on the pot-core bobbin
(capacitive effects). The gauge and type of wire used is significant in design-
ing for high Q. The larger wire diameters contribute to higher Q through
reduced series resistance. Litz wire is considered superior to solid wire at the
higher operating frequencies . This is because litz wire will reduce the eddy-

147
 105
"'
E
3 J:
E
0 QJ
~
u
"'E
.0 1000 c
;!
u
z
---
:::J
-0
c:

103

25 40 63 100 160 250 400 630 1000

Inductance factor, AL t
Figure 4-22 Convenient nomograph for determining inductance or wire turns when the A L of
pot core is known. (Courtesy of Magnetics Division of Spang Industries, Inc.)

current losses in the winding- an aid to higher Q. Litz-wire data are provided
in Table 4-2.
As an aid to predetermining the size of bobbin needed for a given
number of wire turns, it is convenient to refer to the nomograph presented
in Fig. 4-23 . First , the a rea occupied by the turns in terms of square inches
or square centimeters must be determined. This will be based on the wire
gauge, inclusive of the insulating material on the wire. It is also assumed
that the windings will be machine-wound in layers rather than scramble-
wound by hand . Tables 4-2 and 4-3 ca n be consulted to learn what bobbin
area will be occupied by the chosen wire and number of turns . T his infor-
mation can then be used to select a bobbin of appropriate a rea by consulting
Fig. 4-23 .

148
 TABLE4-2 Litz wire data.

Lit: T URNS Litz T U RNS

WirrSi:e peri11 2 percm 2 Wire Size peri11l f'l'f('/11:

5/ 44 28.000 4.341 72/ 44 1.500 232

6/ 44 25,000 3,876 80/ 44 1,400 217
7/ 44 22.000 3,4 10 90/ 44 1,200 186
12/ 44 13.000 2.016 100/ 44 1. 100 170
20/ 44 7.400 1.147 120/ 44 900 140
30/ 44 4.000 620 150/ 44 700 108
40/ 44 3,000 465 180/ 44 500 77
50/ 44 2.300 356 360/ 44 250 38
60/ 44 1,900 294

10,000

~
-~

0
VI
c
:; 1000
.....

100

10.__~~~~~~~..__~~~~~~~~~~~~~~~~
0.001 0.01 0. 1 1.0
ReQuired bobbin area (in. 2 )
Agure 4-23 Nomograph that shows w ire gauge, number of turns and bobbin area in
inches'. (Courtesy of Magnetics Division of Spang Industries, Inc.I

149
 TABLE4-3 Bobbin data.

MAXI MUM TURNS

i
AWG Wire Area (cir mils) Resistcmce C11rre111 Cap11ci1y(111A) '
Wire Size Hea11y Quad perin. l f/!'/"C/11 2 r n/ IOOO(tJ at 750rirmil/A

lO 11,470 12.230 89 13 .8 0.9987 13.840

II 9. 158 9,821 112 17.4 1.26 1 10.968
12 7,3 10 7.885 140 2 1.7 1.58R 8.705
13 5.852 6.336 176 27.J 2.00 1 6.912
14 4,679 5.112 220 34. 1 2.524 5.479
15 3.758 4,147 260 40.3 J. 18 1 4,347
16 J,003 3.329 330 51.2 4.020 3.441
17 2.42 1 2.704 410 63.6 5.054 2.736
18 1.936 2, 190 510 79.1 6386 2. 165
19 1.560 1.781 635 98.4 8.046 1.719
20 1.246 1.436 800 124 10. 13 1,365
21 1.005 1. 170 1.000 155 12.77 1.083
22 807 949 1.200 186 16.20 853
23 650 778 1.500 232 20.30 68 1
24 524 635 1.900 294 25.67 539
(11 25 424 520 2.400 372 32.37 427
0
26 342 424 3.000 465 41.0 338
27 272 342 3.600 558 51.4 259
28 219 276 4.700 728 65.3 212
29 180 231 5.600 868 8 1.2 17 1
30 144 188 7.000 1.085 104 133
JI 117 154 8.500 1.3 17 13 1 106
32 96.0 128 10.500 1.628 162 SS
JJ 77.4 104 13.000 2.0 15 206 67
34 60.8 82.8 16.000 2.480 261 53
35 49.0 67 .2 20.000 3.100 33 1 42
36 39.7 54.8 25,000 3,876 415 33
37 32.5 44.9 32.000 4.96 1 5 12 27
38 26.0 36.0 37.000 5,736 648 21
39 20.2 28.1 50,000 7.752 847 16
40 16.0 22.I 65.000 10.077 1.080 13
41 13.0 80,000 12,403 1.320 II
42 10.2 100.000 15.504 1,660 8.5
43 8.4 125.000 19.380 2. 140 6.5
44 7.3 150.000 23.256 2.590 5.5
45 5.3 185,000 28.682 3.348 4. 1

Courl.<;.,<1'. Q.lMj)JllJ..CJ.iq,._Q[y_i~i on of_S,na 11 1! Jnpu~tric~. Inc.

Sec. 4-4 Pot Cores 151

Long-term stability (DFe) of the core assembly is a matter of more than

casual concern when magnetic core material is used for narrow-band ap-
plications. Stability drift causes a decrease in inductance. In common
engineering terminology, this phenomenon is referred to as "disaccorn-
modation. ''
Disaccommodation can be calculated for each pot-core size and AL fac-
tor. The drift takes place at a logarithmic rate and the long-term shift in in-
ductance can be determined from

~L T2
__ = DFe X log
L Tl
where MIL is the decrease in inductance between times TI and T2, DFe is
the effective disaccomrnodation coefficient of the chosen core, and Tl and
T2 are time differences. For the purpose of definition concerning the
foregoing equation, Tl is the time lapse between manufacture of the core
(these data are usually marked on the shipping carton) and the time it is
soldered into the circuit. T2, on the other hand, is the time lapse from
manufacture to the period when the product in which the core is used has
ended.
Disaccommodation commences the moment the core is fabricated by the
manufacturer, through its period of cooling (Curie temperature). Subse-
quent to this time frame, if the core is thermally or mechanically shocked,
or if it is demagnetized, the inductance can increse to its original value. If
this occurs, disaccommodation starts anew. In view of these traits, the
designer must take all the possibilities into account when designing a circuit.
But if a relatively normal environment is expected for the equipment, the
Jong-term changes in pot-core inductance will be small. Most of the change
will take place during the first few months after the core is manufactured.
It is recommended, therefore, that the unwound cores be allowed to age a
sufficient length of time before they are placed in the manufactured
product.
Effective permeability (µe) is as important when working with pot cores
as it is with any magnetic-core material. Unlike the toroid core, a pot core
has a defined gap. The air gap determines the 1-'e and hence the AL factor for
any given type of pot core. The larger the air gap, the lower the AL. Con-
versely, the larger the air gap, the better the stability of the inductance with
time and temperature. A definite degradation of the Q will accompany in-
creases in air gap. These conditions do, therefore, conflict somewhat with
one another from a design-objective viewpoint. Since stability and Q are
both highly desirable in narrow-band circuit work, the air gap must be a
compromise value in order to strike an acceptable balance between the two
desired conditions-Q and stability.
A representative circuit for a pot core with an air gap is seen in Fig. 4-24.
 Am Cross section

(Am and A 9 , considered constant)

__ 0.4rrN 2 x 10 -a
L henries
Im /µ;Am + lgfA 9
Figure 4-24 Representation of a pot core with its gap. Design
considerations are detailed in the text.

For an inductor wound in the manner shown, the inductance can be found
by

0.4rrN2 X 10 - s
L henries
Im lg
+
µ;Am Ag

where Am and Ag are assumed constant, Am is the effective magnetic path

area in cm 2 , A g the effective gap area in cm 2 , N the number of coil turns, 1111
the length of the magnetic path in centimeters, lg the length of the air gap,
and µi the initial permeability of the core. The denominator is written
customarily as
I
L-
µeA
which is the reluctance of the magnetic circuit.
We can avoid thought complication by regarding the pot core as a device
with a closed, homogeneous magnetic path which has a µe that enables us to
determine that

L = µ
.4rrN
I
2
X 10-s) henries
e I --
( A

152
Sec. 4-4 Pot Cores 153

or L = µeLo, where L is the self-inductance in henries and Lo the air induc-

tance in henries (µ = I for air). For small air gaps we can consider µi to be
approximately equal to µe (the effective permeability of the pot core), and
hence L does, indeed, equal µe x Lo.
From the foregoing it can be seen that a manufacturer can specify a par-
ticular µe for a given type of core material and core dimensions (inclusive of
gap).

4.4.3 Pot-Core Designs

A common application for broadband transformers that utilize toroid

or pot cores is depicted in Fig. 4-25. For a practical discussion we can con-
sider Tl as a balun transformer, since it converts a single-ended source to a
balanced load. T his type of product detector offers high dynamic range

455 kHz 4· 1N914s

Tl

50 i1
3.3 kU
AF
II out
10.05
-:-
• : Phasing Product
detector

To BFO
(+ 7 dBm)

Tl : Trifilar wound (broadband)

1. XL ;;. 4 X 50 i1 .'. XL # 200 .0.

2. Lmon = 69.9 µH (L"H = _L )

x
211{

3. Select a suitable pot core.

Magnetics G-41107-16, 11 x 7 mm
µ , = 120. AL= 160. Bobbin area = 0.00785 in~

4. Determine N
N = 1000 . jLmH + AL = 1000 . j0.0699 + 160 = 20.9turns

5. Win d 21 t ri f i lar t u rns of No. 30 enam.

wire on bobbin and assemble core.
(Combi ne the 3 wires and twist 10
times per inch before winding bobbin)
Agure 4-25 A typical design procedure is given here tor using a pot core
transformer in a small-signal application.
154 Beads, Sleeves, and Pot Cores Chap.4

which is well beyond the Bsat of the pot core. A trifilar winding is placed on
the core bobbin after first twisting the three wires approximately 10 times
per linear inch. Each wire has a different color of insulation to make iden-
tification of the windings easy when connecting them to the circuit.
The four switching diodes in the circuit are small-signal types that
should be chosen for similar forward resistances. Hot-carrier diodes are a
better choice than ordinary silicon types, but good performance can be had
after careful matching of a set of JN914s or similar types. An RF RC filter
is used at the detector output to prevent energy other than audio frequency
from reaching the subsequent stages of the composite equipment.
As has been the rule throughout this book, the XL of the windings
should be four times or greater the characteristic impedance of the circuit to
which they connect. Since we are dealing with a 50-Q source impedance, XL
should be 200 Q or greater. Therefore, the minimum acceptable inductance
of the individual windings will be 69.9 µH, as shown by step 2 of Fig . 4-25.
A suitable pot core is chosen next. Because Q is not a prime considera-
tion in this application, we select a core that has a substantially high AL fac-
tor for RF work. The manufacturer's catalog is consulted and we learn that
a Magnetics G-41107-16 core has sufficient bobbin area and a J.lc of 120 (AL
= 160).
Step 4 of Fig. 4-25 is used to determine the required number of turns for
the trifilar winding. Since the equation calls for 20.9 turns of wire, we use the
nearest whole number, which is 21. We must take into consideration the
fact that a trifilar winding is used. Therefore, allowance is made for the in-
crease in effective wire gauge caused by three conductors being laid on the
bobbin in a single winding operation. It is determined that 21 trifilar turns
of No. 30 enameled wire will be suitable for the bobbin area of 0.00785 in. 2
This is verified by consulting the graph in Fig. 4-23. In fact, some bobbin
space should be left over after the winding is placed on the bobbin. The
polarity dots on the circuit diagram of Fig. 4-25 must be observed when
connecting Tl to the circuit. If not, the phasing will be wrong and the detec-
tor will not function correctly.

4.4.4 Tone Encoder with Pot Core

Pot-core inductors find common application in all manner of audio

generators. Perhaps the most ordinary type of circuit in which we might
employ a high-Q pot-core inductor is the tone generator. A notable example
of this application is seen in the standard telephone Touch-Tone pad, which
has nearly replaced the old-fashioned dial system. Several pot-core
resonators are used to generate the standard Touch-Tone frequencies. A
single active device is used as the oscillator. Various inductances are placed
in the frequency-determining part of the oscillator circuit by means of the
pushbuttons on the Touch-Tone pad.
Sec. 4-4 Pot Cores 155

Other uses for this kind of tone generator are numerous. One example is
the "tone-on" or access tone for VHF and UHF repeaters used by some of
the land mobile radio services and by radio amateurs. The operator in his or
her automobile must key a specified audio-frequency tone when the mobile
transmitter is actuated. The tone is decoded at the repeater, thereby
directing the repeater transmitter to turn on and be available for service un-
til the duty is performed. When the operator is finished with the repeater, it
shuts down and awaits another tone encoding before it is again actuated.
The frequency tolerances established for the tone decoders at repeaters
are fairly rigid. Therefore, it is imperative that the tone-encoder oscillators
be stable and reliable at all times. Pot-core inductors have proven reliable
for many years, and are therefore excellent for the application.
Figure 4-26 contains the circuit of a typical tone encoder device. A
device of the type shown is suitable for use with telephone answering ser-

Tone encoder

f ~ 1500 Hz
Oscillator

~0.09 -CllµF II~

2N2222A 8 .n
SPKR
lOO kQ

Sl

+ 19V

l
1. T1 pri. = 88 mH. Z P'i ~ 3 k.11· :. Turns ratio= 19: 1

2. Select suitable pot core. Magnetics G·42823·X 1,

28 x 23 mm. µe = 307, AL = 1000.
Bobbin area= 0.0910 in~

3. Determine Np,;_

N = 1000 ..jBB .;. 1000 = 296 turns No. 28 enam. wire

tap at 148 turns.

4. Determine N,.c· N =0.05NP" = 15 turns No. 28 enam. wire

Figure 4-28 Pot cores are commonly used in encoding oscillators
of the type seen here. Design details are given.
156 Beads, Sleeves, and Pot Cores Chap. 4

vices. The operator can dial or tone up his office or personal phone from
some remote location. He or she can then place the speaker of the encoder
near the telephone transmitter and command the a nswering service (unat-
tended) to rewind its tape, then play back any messages that may have been
recorded during the operator's absence.
A simple bipolar-transistor tuned base/tuned collector oscillator is used
at QI of Fig . 4-26. Tl is a pot-core inductor that is wound for an inductance
of 88 mH. There is nothing especially sacred about the inductance value: it
happens to be a relatively common one in telephone circuits that use
toroidal inductors. The circuit is shown with a standard capacitance value
of 0.09 µF in parallel with the primary of Tl . lf 1500 Hz was the precise fre-
quency desired, an adjustable pot core could be used for putting the
oscillator exactly on frequency. Alternatively, a high-capacitance trimmer
might be used in parallel with the 0.09-µF capacitor.
The oscillator is keyed on and off by means of SI, a momentary switch.
Encoding is effected by merely placing the speaker near the mouthpiece of
the telephone (or microphone of a transmitter).
We shall assume a collector current of 3 mA for QI. Therefore, the de
collector resistance will be on the order of 3000 Q . The impedance transfor-
mation ratio for Tl becomes 375: I, which gives us a turns ratio of 19.36:1.
A precise match between the transistor and the speaker is not important.
The required output from the speaker is modest, negating the need for max-
imum power transfer to the load. Step I of Fig . 4-26 shows that we have
chosen a turns ratio of 19: I.
Step 2 of the design exercise requires some intuition. We are aware that
to develop 88 mH of inductance we will need a reasonably large pot core. It
should have a substantially high µe, and hence a high A L factor. However,
the core chosen needs to permit a Q that is high enough to ensure good
oscillator action . At this juncture we are also aware that a large area core in
a low-power circuit such as that of Fig. 4-26 relieves our need to ponder the
Bmax profile of the core: we will be well within the area of linear operation.
After studying the manufacturers' literature we have selected a
Magnetics G-42823-Xl pot core. Equivalent types made by other manufac-
turers would be entirely suitable also. The core measures 28 x 23 mm, has a
µe of 307 and an AL of 1000. The published curves indicate that the Qu of
the inductor should be in excess of 200 at 1500 Hz. The bobbin area for this
core is 0.0910 in. 2 •
Using the standard equation for N we learn that 296 turns of wire are re-
quired to develop 88 mH of inductance (step 3 of Fig. 4-26). No. 28 enamel-
ed wire will be small enough in diameter to enable us to place 296 turns on
the bobbin, with ample space remaining for the secondary winding of Tl.
The design is concluded by moving to step 4. A layer of thin insulating
tape is place over the completed primary winding of Tl; then the secondary
Sec. 4-4 Pot Cores 157

is laid on the bobbin. We shall need 15.57 turns of wire for the secondary in
order to realize the 19: 1 turns ratio of step 1. A 20-turn winding is
employed.

4.4.5 Pot Cores in Filters

An example of a five-pole low-pass audio filter is offered in Fig. 4-27.
The same core number as that specified for the circuit of Fig. 42-6 will be
used to obtain the inductances of LI, L2, and L3, which are 17 mH (LI and
L3) and 24 mH (L2).
In step 2 of Fig. 4-27 we learn that LI and L3 will consist of 130 turns
each, wound on bobbins for Magnetics G-42823-X I pot cores. L2, accord-
ing to the equation in step 3, contains 155 turns of wire.
Finally, the wire size is chosen in accordance with the 0.091 -in. 2 bobbin
area. No. 24 enameled wire will suffice. The self-shielding properties of the

f 00 =1000H z
Ll L2 L3

5-Pole low·pass audio filte r

Xu. LJ = 106.8 Q.= 17 mH X c1. C2 = 45.9 = 3.47 µF

XL 2 =150.34Q = 24mH

1. Select high-µ pot core

Magnetics G-42823-Xl, µ. = 307,
AL = 1000, 28 x 23 mm. Bobbin in area = 0.091 in ~

2. Calculate turns for L1, L3

N = 10oovLmH +AL = 10oov17 + 1ooo = 130 turns

3. Calculate turns for L2

N = 1000 .,,124 + 1000 = 155 turns

4. Determine wire size vs. bobbin area, per Fig. 4·23.

Use 130 turns No. 24 enam. wire for L 1, L3 and
155 turns No. 24 enam. for L2
Figure 4-27 Many filters contain pot-core inductors. The five-
pole low-pass filter seen here is one example. A basic design pro·
cedure is included.
158 Beads, Sleeves, and Pot Cores Chap.4

pot cores will prevent the need for using partitions between the filter sec-
tions.
A seven-pole low-pass filter is illustrated schematically in Fig. 4-28. It
has an fco (cutoff frequency) of 2. 7 MHz. Therefore, the core material for
the three inductors must be selected accordingly. The improper core would
degrade the Q and render the filter ineffective.
Step 1 of Fig. 4-28 specifies a Ferroxcube 1107PA25 pot core of 4C4
material. The dimensions are given in inches and millimeters (11 x 7 mm).
The µe is 19 and the AL is 25. The manufacturer's Q curves indicate than the
core material is suitable for use at 2. 7 MHz and lower.
N for LI and L3 is found in step 2, again using the standard equation for
Nversus AL. 12.6 turns are specified for the end coils of the filter.
The calculations for L2 are conducted in step 3 of Fig. 4-28. The result-
a nt number is 13.8. A wire gauge of No. 22 will be used for all three induc-
tors. The larger wire gauge will reduce the R of each coil, thereby enhanc-
ing the Q.

!co = 2.7 MHz

Ll L2 L3

Low·pass filter

X LI, L3 =68 (4.0 µHJ X L2 = 82 (4. 77 µH)

X c 1.c4 = 69(854pF) Xc2. c 3 =30 (1966 pFJ

1. Case: Fer roxcube 1107PA25-4C4

;i in. ( 11 mm) OD x f2 in. (7 mm ) high
µc = 19, A 1. = 25. Q" "' 1 50 at f co

2. NLl, l.J = lOOOJLmH +A1.

= 1000 J0.004 + 25 = 12.6 turns
3. N L2 = 1000J0.00477+25 = 13.8 turns

4. Select wire gauge

(No. 22 enarn. will fit bobbin
for turns speci l ied)
Figure 4-28 Pot-core inductors are suitable for tuned filters of the variety
depicted here. Since the fco of this filter is 2.7 MHz, the proper kind of core
material must be chosen to ensure proper Q (see the text!.
5
PERMANENT-MAGNET DATA

Chapters 1 through 4 dealt specifically with magnetic-core materials that

were not magnetized permanently. T his chapter addresses another part of
the core-material family- those cores which are permanently magnetized
fo r a host of modern applications. Among the common uses of magnets are
loudspeakers, phonograph cartridges, magnetic latches, and meters. Other
uses include traveling-wave tubes, magnetrons, klystrons, generators, and
alterna tors. This list is by no means complete, but it does illustrate the
widespread use of magnets in the electrical and electronics industry.
In this section the author has borrowed heavily from a booklet entitled,
Permanent Magnet Guidelines, produced by MMPA (Magnetic Materials
Producers Association), with their kind permission. The pamphlet was so
well wr itten, and the contents of such importance, that it was selected as a
valuable reference in this book.
Another book by MMPA is available on the subject matter of this
chapter. It is Testing and Measurement of Permanent Magnets. It is recom-
mended highly as a supplemental text for students, technicians, and
engineers.

5.1 The Nature of Permane nt-Magnet Materials

French physicist Pierre Weiss, some 50 years ago, supplied the central
thought that explained the observed properties of ferromagnetic materials
and today still provides the basis of ou r highly sophisticated body of
knowledge that explains quite satisfactorily the observed properties and
provides an intelligent guide for the search for improved materials . Weiss

159
160 Permanent-Magnet Data Chap.5

postulated that a ferrom agnetic body must be composed of small regions or

domains each of which a re magnetized to saturation level, but the direction
o f the magnetization from domain to domain need not be parallel. Thus, a
magnet when demagnetized , was only demagnetized from the viewpoint of
an observer outside the material. Man-made fields only serve as a control in
changing the balance of field energy within a magnet.
The inherent atomic magnetic moment associated with such elements as
iron, cobalt, and nickel is believed to originate or result from a net un-
balance of electron spins in ccnain electron shells. For example, in iron in
the 3d shell there are more electrons spinning in one direction than in the
other. Having an inherent atomic magnetic moment is a necessary but not a
sufficient condition for ferromagnetism to be exhibited. Additionally, there
must be cooperative interatomic exchange forces that m aintain neighboring
atoms parallel. Little is known of the exact nature or magnitude of these
forces , but observations suggest they are electrosta tic, a nd it has been
pointed o ut that in ferromagnetic materials the ratio of intera tomic distance
to the diameter of the shell in wh ich the unbalance exists is unusually large
compared to this ratio in materials which do not exhibit ferromagnetism .
Also note that these exchange forces produce magnetostrictive effects and
are associated with the crystallographic structure of magnetic materials in
such a way as to produce a directional dependence of magnetization with
respect to the crystal axis. Figure 5- 1 illustrates this rela ti onship in iron; the
easy axis of magnetization is the cube edge of (100) direction. In Figure 5-2
an exploded view of a ferromagnetic volume shows the order of size for the
va rious regions.
We can view the magnetic domain as a region in which the atomic
moments cooperate to allow a common magnetic moment which may be
rotated in man-made external fields. Domain size, not a fundamental con-
stant of physics, varies widely depending on: composition, purity, and state
of strain of the material; and on some very important energy relationships.
Figure 5-3 shows a boundary region between two domains. This boundary
region and its significance were first proposed by Bloch. The Bloch wall is a
transition region containing many atomic planes . The 180 ° change in
magnetization will want to occur over a considerable distance to minimize
the potential energy in the wall. However, the width of t.he wall will be
restricted because of the restraining influence of crystal aniso~ro py (direc-
tional dependence of magnetism with respect to crystal axis). Figure 5-4 il-
lustrates an additional energy relationship which influences the size of the
domain and involves the magnetostatic or field energy surrounding a
magnetized volume. A magnetized volume tends to subdivide. It will be
energetically possible for subdivision to occur as shown in Figure 5-4 until
the decrease in magnetostatic energy is less than the potential energy
associated with the Bloch wall foundation. At this point we m ight say that
 25,000 ~------~-------~

100

20,000

15,000

8
110
10,000

5,000

111

0 200 400 600 800

H
Figure 5-1 Exchange forces affect the structure of magnetic
materials in such a way as to produce a directional dependence of
magnetization with respect to the crystal axis. In iron {shown) the
cube edge of (100) direction is the easy axis of magnetization.

Domain :
Atom 105 atoms

•
0 0

0
00
0

--~
_ - .....
~ -· - - _,..,
~----....

'' '-"""-- ....

(j) Q- I
-~ 0
I
/
0 a_
0 '~
Unit crystal: 1 cub ic cent imeter:
106 domains 102 crystals
Rgure 5-2 Exploded view of a ferromagnetic volume.

161
 -J I-- Bloch wall

,..,
/' I

I
I I
I
I ,; I / '1 /1 / I _...'1 )
I ,,,/" \,;,.../ ~ _,..- [.,./,.. t,.. ,..." ~ /

Figure S.3 Boundary region between two domains.

, _ /

' y/

t ti I
i
I
,..l._
,,,,./~..............

(a) (b) (c)

Figure !>4 Subdivision can occur until the decrease in

magnetostatic energy is less than the potential energy
associated with the Bloch wall foundation.

the magnetization vector arrangement associated with domain volumes in a

ferromagnetic material results from a complex energy balance. Their ar-
rangement is such that the total potential energy of the system is a mini-
mum. External, man-created fields to magnetize or to demagnetize only
disturb the balance of the potential energies involved and our familiar
S-shaped magnetization curves are records of the change in balance with
respect to the external influence. Figure 5-5 shows the action pictorially as a
bar of ferromagnetic material is magnetized . The demagnetized condition,
(A), results from an internal arrangement with mutually canceling direc-

1 62
 8

(a)

(d) +H

Agure 5-5 Magnetization of a bar of ferromagnetic material.

tions of magnetization vectors. In region (B), with low values of external

field, the action is primarily one of domain boundary stretching, usually
around imperfections. This is a reversible process [a reversible magnetiza-
tion process is one in which the magnetization vectors reorient to their
original position after the field (H) is removed]. As the field is increased,
region (C) domain boundaries break away and move through the material.
The more favorably oriented regions grow at the expense of their less
favorably oriented neighbors. A large increase in magnetic induction oc-
curs, an irreversible process [in which the magnetization vectors tend to
keep their new position after (H) is removed]. In region (D) at still higher
values of magnetizing force, the magnetization vectors are rotated against
the forces of strain and crystal anisotropy into alignment with the direction
of the applied field, and saturation occurs. Removing the magnetizing force
causes some relaxation; the domains rotate back to the easy direction of
magnetization (a reversible process). This relaxation can be minimized by
making the direction of easy magnetization coincident with the direction of
magnetization desired.
Subjecting the magnet to a demagnetizing force retu rns the domain
boundaries to a condition similar to their original positions in (A) and hence
the magnet is demagnetized. In normal use a permanent magnet operates in
the second quadrant of the hysteresis loop and since the magnetization force
is negative the usable induction will be B = 8, + ( - H) and the magnet will
operate at some point such as Figure 5-5d, where a magnetic potential - Hd
per unit length and induction + Bd per unit section will be established. For

163
164 Permanent-Magnet Data Chap. 5

outstanding permanent magnets we want to make domain wall motion and

rotation of the magnetization vectors difficult. The more external energy re-
quired to orient the system, the more will be required to demagnetize and
hence a better permanent magnet.
Figure 5-5 describes the magnetization process satisfactorily for
magnetic materials having coercive force value of Ho up to approximately
300 Oe. The coercive force of the early carbon, tungsten, and cobalt-steel
permanent magnets is believed to be a result of impeding domain wall mo-
tion. These quench-hardened magnets have nonmagnetic inclusions
building up at the domain boundaries, obstructing wall motion. This
mechanism is believed to be of significance only at relatively low field
strength.
Today's modern permanent magnets exhibit coercive forces well above
the level explained by domain wall motion. One must therefore look for
magneti zation mechanisms requiring greater energy input. A signjficant
milestone in understanding permanent magnet properties occurred with the
suggestion of Frenkel and Dorfman that, if small particles were prepared
with dimensions less than the width of domain boundary, such particles
would contain no boundaries. This explanation forms the central concept in
fine-particle magnet theory and provides us with a satisfactory picture and
explanation of high coercive force permanent magnets such as Alnico, fer-
rite, and fin e-particle iron-cobalt types. The dimensions of a single domain
volume are predictable from a consideration of wall energy and
magnetostatic energy. For a sphere the wall energy is proportional to the
cross section or to the square of the radius. The magnetostatic energy is pro-
portional to the volume or to the radius cubed. A critical radius exists where
these two energy values are equal, for this value and below, it is energetical-
ly impossible for a domain boundary to exist. Without domain boundaries
the magnetization of a permanent magnet can be changed only by rotation
of the magnetic moments associated with each domain volume. This process
requires higher energy input than domain boundary motion. The degree of
difficulty in reversing the magnetic moments or the coercive force of a
single domain system depends on the forces of anisotropy, or the forces
which give a directional dependence to a domain's magnetic moment.
Crystal and strain anisotropy have been mentioned previously; additionally,
shape anisotropy is of great importance. In fine-particle magnets, shape
anisotropy arises from the fact that, in an elongated single domain, the
magnetization along the major axis is the easy axis and involves minimum
energy. To rotate the magnetization vector and magnetization to the minor
axis involves additional energy. Hence, elongated particles exhibit higher
coercive force than spherical particles. In service, the permanent magnet is a
unique component in the energy-conversion cycle. When the permanent
magnet is magnetized, energy is dissipated in changing the force balance of
Sec. 5-2 Elementary Permanent-Magnet Relationships 1 65

magnelizatio n veclors. For th e permanent magnet to establish external field

energy, input is required to change the vector orientation near the poles and
establish the permanent magnet's magnetomotive force. Energy is only in-
volved in changing a magnetic field, not in maintaining one. Hence, the
field energy established via the permanent magnet is independent of time
unless the permanent magnet is subjected to additional energy input (heat
energy or demagnetizing field energy, for example). A permanent magnet in
a stabilized cond ition is a reversible medium o f energy transform ation.
Potential energy or the permanent magnet is composed of both internal and
external field energy. Permanent magnets often operate over a dynamic cy-
cle in wh ich energy is converted from an electrical or mechanical form to the
magnetic field and then returned to the origina l form.

5.2 Elementary Permanent-Magnet Relationships

The basic problem of permanent magnet design is to estimate the distribu-

lion of magnetic flux associaled with a specific magnetic circuit geometry
that may include permanent magnets, air-gaps, and soft-steel flux conduc-
tion elements. This is a very difficult problem and only approximate solu-
tions are feasible based on certain simplifying assumptions . Determining
how suilable or appropriate these assumptions are in a particular case re-
quires appreciable experience. This seclion attempts to present o nly a few
basic concepts and inter-relationships. It should be emphasized that obtain-
ing an optimum magnet design in a particular situation involves experience
and trade-offs between many different information inputs involving both
the permanent magnet and the equipment in which it is used.

5.2. 1 The 8 -H Curve

T he basis of d esign is the 8 -H curve, or hysteresis loop, typical of all
magnetic ma terials, which is obtained from measu rement on a specimen
under closed-circuit cond itions (Figure 5-6). Solid lines represent the normal
form of the curve, and the arrows indicate the direction in which the curve
traverses a symmetrical cycle of magnetization . The dashed curve shows the
intrinsic hysteresis loop. Although in the us ual utilization of permanent
magnets the normal curve is used for analysis, it is desira b le to note the dif-
ference between these two curves, because it is very significant in permanenl
magnets. These two curves are related by the equation B = B; + H, where
B is the normal induction , B , the intrinsic induction, and H the fi eld
strength. In 1hc second quadrant of the hysteresis loop H is negative, and
hence B, is greater than B. In a permanent magnet the field strength, H, is
directly opposed to the intrinsic induction B 1 and B is the net induction
available for external use. It sho u ld be noted that this is in contrast to the
electromagnet, in which B1 and Hare in the same d irection.
 +B (gauss)

I
I
BH point max. I
I
I
I
I
I
I
I
I
-H I ~H (oersteds )
I
I
I
I
/
/
/
/
/
/
/
~/-

-8
Figure 5-8 The B-H curve, or hysteresis loop is obtained from
measurement on a specimen under closed circuit conditions. Solid
lines indicate the normal form of the curve; arrows indicate the direc-
tion in which the curve traverses a symmetrical cycle of magnetiza-
tion. Dotted curve shows intrinsic hysteresis loop.

The most important characteristics of permanent magnetic materials,

residual induction, Br, coercive fo rce, He, and maximum energy product,
(BdHd) 111 , may be obtained from the demagnetization quadrant of the
hysteresis loop. The residual induction, Br, is defined as the magnetic flux
density of the material in a closed circuit when the applied magnetizing
field, H 5 , has been removed, while the coercive force, He, is the demagnetiz-
ing force, which reduces the retained flux to zero. (BdHd)111 , a criterion of
magnetic properties, may be obtained by constructing the energy product
curve, that is, the product of abscissa Hand ordinate B plotted against the
value of (B) or (H) at appropriate points on the demagnetization curve. A
smooth curve drawn through the plotted points will commence at zero, the
origin, rise to a maximum point, (BdHd)m, and fall again to zero.

5.2.2 Recol'l Loops

lf, at any point such as Pin Figure 5-6, the cyclic change in His reversed
in direction, the flux density changes along a curve such as PR. If, from R ,
the field strength H retraces its values, the induction traverses the upper half
of the loop (PR) until it again reaches P , after which the major hysteresis
loop is followed . The loop (PR) may be coincident with the major loop.
Such a loop is known as a recoil loop, the average slope of which is fairly

166
Sec. 5·2 Elementary Permanent-Magnet Relationships 167

constant whatever the point of origin. This slope is known as the recoil (or
reversible) permeability (µre>· It is sufficiently accurate for engineering pur-
poses to regard the loop as a straight line.

5.2.3 Working Point of a Magnet

(Static Applications)

A permanent-magnet system usually has to provide magnetic flux in an

air gap. The presence of this gap in the circuit produces a magnetic field op-
posing the flux in the magnet material. In other words, magnetic energy ex-
ists in the gap, and in providing this energy the magnet is correspondingly
self-demagnetized. Therefore, if the induction B of the magnet is positive,
the field strength H inside the magnet must be negative, which means that
the magnet works in the second or fourth quadrant of the hysteresis loop
(i.e., the demagnetization curve). If the magnet is demagnetized by the gap
alone, the operating point is on the demagnetization curve. However, if
after saturation, the magnet undergoes additional demagnetization (e.g., by
a "stabilizing" field) the operating point is inside the demagnetization
curve, on a recoil loop.
The whole volume of a magnet does not operate at the same point but,
for simplicity, it is usual to assume that it does. The error inherent in this
assumption is greatest when the air gap is long, but much less with a short
air gap or when the magnet has substantial soft-iron pole pieces. For a
magnet of cross section Am (cm 2) and length Im (cm) , the total flux provided
in external space is B~m and the total magnetomotive fo rce across the ends
of the magnet is Him· Most design problems can be reduced to a specifica-
tion of the total flux and the total magnetomotive force to be provided by
the magnet. For a given alloy, the total volume of magnet material required
(A 111 lm) is inversely proportional to (BdHd) at the chosen operating point,
which enables the length and section of the magnet to be fixed. For
economy, it is usual to choose the operating point, where the energy pro-
duct has its maximum value for the material but, if the magnet dimensions
thus obtained are impractical, another operacing point or a different
material must be selected.

5.2.4 The Magnetic Circuit

In designing a magnetic circuit, it is necessary, in addition to the normal

assessment of the total flux and total magnetomotive force required, to
assume that the magnetic circuit can be divided into various elements and
treated like an electric circuit by applying the magnetic analog of Ohm's
law:

flux = magnetomotive force

(5.1)
reluctance
168 Permanent-Magnet Data Chap. 5

The main difficulty in applying this formula is to assess the reluctance.

For short gaps of large area the reluctance is

gap length (cm)

(5 .2)
gap area (cm 2)

In practice, the effective reluctance is less than this value, since part of
the nux is leakage. This is equivalent to a proportionate increase in the gap
area, which may be considered as a leakage factor (F). This factor is usually
between 1.5 and 20, and, although it may be calculated approximately by
taking the reciprocal of the sum of reciprocals of all leakage path reluc-
tances, it is more reliably based on previous experience of comparable
designs.
Si milarly , it is frequently advisable to apply a correction factor to the
gap length to allow for the losses introduced by curvature of llux path, sub-
sidiary joints in the magnetic circuit, and loss in pole-piece materials. This
factor(/) is usually much smaller than the area factor, and is normally be-
tween I. I and 1.5. For a gap of section Ag (cm 2) and length 1 (cm), the cor-
8
rected reluctance is

(5.3)

The full circuit equation is

HdtmAgf
(5 .4)
lg!
where Bd and Hd are the values at the assumed point of the magnet of sec-
tion Am and length 1111 • This can be transposed as

Bd 1 Agf
-- = -111 -- (5.5)
Hd A 111 1gf
The reciprocal of reluctance (A F!lgl) is known as the permeance.
8
The whole right-hand side of equation (5.5) is known as the unit per-
meance a nd may be represented graphically by a straight line passing
through the origin of a B- H curve. Several such lines are shown on Figure
5-7. Every design has a corresponding operating line, and its slope depends
on the dimensions of the magnet and the gap, with correction of leakage
factors. The slope is independent of the alloy of which the magnet is made.
The intersection of this line with the demagnetization curve of the alloy
gives the ·working point of the magnet.
The foregoing gives the ratio /11/ Am for the magnet and it remains to fix
either Im or Am. If the total flux required from the magnet (including
Sec. 5·2 Elementary Permanent-Magnet Rela tionships 169

lea kage) is known, then A 111 can be determined, as this total flux is Bd ·Am.
If the field strength Hg in the speci fied air gap is known , then

Hc/m = H/;,f (5 .6)

Equations (5 .5) and (5 .6) together give the section a nd length of the
magnet required in an a lloy of known demagnetization curve and of chosen
values of Bd and Hd. T hese values may be: at the (BdHd)m point, to give
minimum volume of material; a t some other point on the demagnetization
curve, if practical considerations of size make it necessary; or on a recoil
loop, if the conditions of operation make this desirable or inevitable. If the
magnet is magnetized before being attached to the pole pieces, it may work
on a recoil loop. In a motor or generator the magnet has no fixed working
point, but traverses a recoil loop.
In many uses of the permanent magnet, a well-defined magnetic circuit
is not involved (e.g. , a simple bar magnet). Under such conditions the
operating point on the demagnetization curve is determined primarily by the
dimension ratio /m/ dia of the bar.

5.2.5 The Permanent M agnet in an External Field

Jn using a permanent magnet the influence of external fields in changing

the flux level is an important consideration. The geometry of the magnet
assembly sets the slope of the operating line and has considerable effect on
the interaction of the permanent magnet and the external fi eld. To predict

-H 0
Figure S.7 Typical demagnetization curves showing various
operating slopes.
170 Permanent-Magnet Data Chap. 5

the change in flux density of a permanent magnet in the presence of an ex-

ternal field, we must have both the normal and intrinsic curves of the per-
manent magnet material. If the operating line slope or unit permeance
Bd/Hd is known, a line with a slope of B/H<i + 1 is drawn through the
origin, (Figure 5-8). The magnitude of an external field influence - Ha is
laid off parallel to Bd!Hd + I and its intersection with the intrinsic curve is
projected down to the normal curve to yield the new level of flux density in
the presence of this external influence. In Figure 5-8, aB is the loss of flux
density for - H 0 external applied field influence.

5.3 Magnetization
A permanent magnet, to exhibit its fu ll properties, must be magnetized to
saturation. Partial magnetization results in reduced properties, a nd the effi-
ciency and stability of the permanent magnet are consequently impaired.
The permanent-magnet manufacturer generally ships demagnetized
magnets to the user. The main reason is that the permanent magnet is usual-
ly incorporated into a magnetic circuit and the operat ing flux levels can only
be obtained by magnetization after assembly into the magnetic circuit. Ad-
ditionally, shipping costs of magnetized permanent magnets are greater and
the danger of contamination due to pickup of magnetic particles is lessened
by shipping and handling demagnetized magnets.

+B or B;
8

B,

-H
Figure 5-8 B is the loss of flux density for - H 8 external applied
field influence.
 5.3. 1 Conditions for Complete Saturation
To fully saturate a permanent magnet requires careful consideration of
the magnitude, the shape, and the time duration of the applied field. In a
previous section, the term "self-demagnetization" was introduced. In
magnetization the free pole or self-demagnetization concept is extremely
important because the self-inflicted, self-demagnetization is a maximum for
a permanent magnet during magnetization.
The saturation magnetizing force (Hs) of a permanent magnet material
may be taken as four to five times its intrinsic coercive force. In designing a
magnetizer it must be remembered that the value of (Hs) must be the net ef-
fective value acting on the magnet, and the presence of self-demagnetization
or shuntings effects will necessitate additional magnetizing field strength to
compensate for these factors. The self-demagnetization influence will be a
maximum for the case of a bar magnet having a small ratio of length to
cross-sectional area magnetized in a solenoid. For example, a relationship
between net effective working force as seen by the permanent magnet (Hs)
and the field strength of the solenoid (H5 p) may be expressed as follows:

Hs = Hsol - K (Bs - HsJ

where K is a demagnetization constant depending on magnet geometry
(Table 5-1); Bs is the flux density associated with field strength H 5 . An ex-
ample, perhaps, best illustrates the importance of free poles and the self-
demagnetization in magnetizing problems. Consider two Alnico 5 rods, one
having a /m/D ratio of 0.1 and the second a ln/D ratio of 10.0. To saturate
Alnico 5,H5 = 3000 and Bs = 17,000 G. Using the relationships described,
the rod with a /m/D = 0.1 would require a value of H 501 = 14,900 Oe, the
rod with a lm/D = 10.0 would require a value of H 501 = 3240 Oe. This ex-
ample shows that the free-pole influence of the magnet geometry having
heavy self-demagnetization can be minimized by the use of soft-iron parts
to complete the magnetic circuit.

TAB LE 6-1 K forrods and bars.

Dimensional ratio
(length/ diametera) 0.0 0.1 0.2 0.5 1.0 2.0 S.O 10.0
K 1.000 0.850 0.730 0.472 0.270 0. 140 0.040 0.017

a Diameter of bar may be taken as 2V a rea/ ir.

In addition to having an adequate magnitude of magnetizing force, it is

necessary that the shape of the applied field coincide with that of the
magnet as closely as possible. Modern permanent magnets are low in
permeability and their presence in a field will not, to any great extent, shape

171
1 72 Permanent-Magnet Data Chap. 5

the fiel d. C urrent-carrying conductors combined with soft-steel members

can be arranged to give almost any field contour desired. As a general state-
ment, it is not objectionable Lo use magnetizing force in excess of the
minimum required. However, when the field contour is very dissimilar to
that of the magnet, magnetization at some angle from the desired axis may
result. The net effect would be one of partial or incomplete magnetization.
The duration of the magnetizing field is often a factor. We know that
the rise of magnetic induction is essentially instantaneous in thin magnetic
films, but in a metallic permanent magnet of appreciable cross section, in-
duced eddy currents can influence the rise of magnetization. In equipment
such as electromagnets, the rise of magnetizing current is not instantaneous,
owing to inductance. Many large electromagnets require buildup times in
the order of seconds.

5.3.2 Magnetizing Equipment

Steady-state field s produced by electromagnetic yokes and permanent-
magnet structures are convenient forms of equipment for magnetizing
si mple-shaped rods, bars, and rings. The conventional electromagnet
magnetizer has one pole movable with respect to the other, is usually
operated from a full-wave rectifier, and is designed for intermittent opera-
tion. Figure 5-9 shows a typical arrangement.
Principal design considerations for an electromagnet to magnetize a par-
ticular magnet are:

I. Core section must be adequate to carry the saturation flux.

2. Ampere-turns must be su fficient to produce the total mmf required
by the magnet length , the magnetizer yoke, and any self-demagneti-
zation influences present.

Because of the leakage a nd losses, the core of the electromagnet should be

at least two times the area of the magnet to ensure adequate flux level in the
permanent magnet. The ampere-turn requirements are typically twice the
value computed to saturate the length of permanent-magnet material.
Very small permanent magnets are often magnetized by means of a large
permanent magnet. The amount of physical effort and the uniformity prob-
lem created by moving the magnetized magnet through a nonlinear field ad-
jacent to the main air gap great ly limit this approach . Since the time interval
required is extremely short, magnetizing can be achieved by a current im-
pulse, provided that its magnitude is sufficient to deliver the peak magnetiz-
ing force required.
The development of material with a high coercive force and high
available energy has led to relatively intricate magnet configurations. The
length of magnet limbs has decreased and circular shapes and parallel cir-
 Magnetizer
pole pieces

MT" j
~~C7
U-shaped
C7~CJ
Assembly
Bar magnets magnets C-assemblies
(a) (bl (c)

Cross section

Length

Figura ~9 When selecting an electromagnetic magnetizer for

your application, consider two factors: area of magnetizer core
should be about twice cross-sectional area of magnet; for
Alnico magnets, ampere turns should equal 6,000 per inch of
magnetic length of magnet. A slight surplus of ampere turns is
desirable when possible. For ceramic magnets, approximately
20,000 ampere turns per inch are required. (a) Bar magnets; (b)
U-shaped magnets; (c) assemblies.

cuits are common. Many of the newer-shaped permanent magnet ar-

rangements cannot be magnetized by placing them in contact with conven-
tional electromagnets. Instead, they are magnetized by the field around a
conductor carrying a high current impulse. Some shapes can only accom-
modate a single conductor threading through the window or the magnetic
circuit. Others must be wound with several turns of heavy wire. In fact,
many magnet designs are materially influenced by how the magnetizing con-
ductor or conductors can be arranged.
Impulse magnetization has become more popular not only for these
newer shapes, but also for nearly all types of permanent magnets, because
of the flexibility, low initial equipment cost, and the ability to deliver higher
instantaneous currents from conventional ac power mains. Figure 5-10
shows typical impulse magnetizer circuits.
The current required for impulse magnetization around a single conduc-
tor can be estimated from I = 5 H/r), where I is the peak current in
amperes, H 5 the saturation magnetizing force in oersteds, and r the maxi-
mum radius of the magnet configuration in centimeters.
Figure 5-11 shows typical conductor arrangements for magnetizing
various permanent-magnet configurations. In impulse magnetization, in-
duced eddy currents in metallic magnets can impede flux penetration in

173
 r-- - ---,
Ignition : I
Charging Capacitor or 1 Pulse i
ci rcu it bank 1 transformer
Fixture
thyristor I I
(SCR ) I
I
I
I
'--- - --- -~

Capaci t or discharge
(a)

r -------,
Ignition I I
Phased
timing
or I Pulse :
Fixture
thyristor 1 transformeq
circuit
(SCR) I I
I I
l..------- ..J
~Cycle
(b)

Figure 5-10 Typical impulse-magnetizer circuits. (a) Capacitor discharge; (bl half·
cycle.

Magnet Conductor

Magnet
" C" or ring magnets So lenoi d magnet izing coil bar magnets
(a) (b)

Conduct ors
Magnet

s
/
N
Magnet

" E" shape magnets Multip le pole magnets

(c ) (d)

Figure 5-11 Typical conductor arrangements for magnetizing various permanent·

magnet configurations. (a) C or ring magnets; (bl solenoid magnetizing coil bar
magnet; (cl E-shaped magnets; (di multiple pole magnets; (e) multiple poles on one
face of disk or ring magnet; (fl multiple poles on ID of stator magnet or assembly; (g)
two-pole on ID of stat or magnet or assembly.

174
 Steel

Conductor

Multiple poles on one face or disc or ring magnet

(e)

Conductor
N

Steel Magnet ____,.....__ _...-

Multiple poles ID of stator magnet or assembly

(f)

~M~ov.et -1A
Steel
1
2-pole on ID of stator magnet or assembly
(g)

magnets of appreciable cross section. It is often necessary to use several cur-

rent pulses to achieve saturation.
How does one know when saturation has been reached? This is a dif-
ficult question. The common practice is to use a surplus of magnetizing
force, but often one encounters what is believed to be a marginal saturation.
Under such conditions it is recommended that the user consult his or her
magnet supplier.

175
5.4 Demagnetization

Demagnetization is defined as reducing to an acceptable level the flux d ensi-

ty, in the space surrounding a permanent magnet, at some predetermined
distance from the magnet surface. At temperatures below the Curie temper-
ature, permanent magnets consist of elementary domains-microscopic or
submicroscopic volumes of material, each of which are tiny saturated per-
manent magnets. In a magnetized magnet, the magnetic moments of these
domains are aligned essentially parallel. In a "perfectly" demagnetized
magnet, they are aligned in a statistically random manner. Thus, even in a
"perfectly" demagnetized magnet, some external fi eld will be present, if
only on a submicroscopic scale . A$ a generality, demagnetization is handled
by the manufacturer of permanent magnets . This is because magnets must
be magnetized for testing, then demagnetized for shipment, in most cases.
There will, however, be some cases in which the magnet user will find it
necessary to demagnetize. Partial demagnetization, or stabilization, to a
measured level below the saturated value is quite common in applications
requiring close tolerance magnetic fields. This is particularly true in applica-
tions where the manufactured part-to-part variations in both the magnet
and associated structure cause variations in the magnetic field strength that
are outside the acceptable end-product tolerances.

5.4. 1 Demagnetization Techniques

AC Fields. The most common method of demagnetizing is to subject the
magnet to an ac field of an intensity sufficient to approximately saturate the
magnet, followed by gradual reduction of the field to zero, either by
withdrawing the magnet to a remote location with the ac power on, or by
gradually reducing the ac voltage. In general, the higher the field and the
more gradual its reduction, the more nearly you approach ideal
demagnetization.
Design of conventional solenoids to provide 60-Hz fields of sufficient
intensity to demagnetize modern permanent-magnet materials offers some
difficulty if the coil must be run on a semicontinuous basis. Fortunately,
only intermittent operation is normally required . Alternatively, one can
design the coil, which consists of a fe w turns of water-cooled conductors,
and run it at low voltage and high current, by means of a step-down
transformer. This type can be used for continuous duty.
For applications requiring less critical demagnetization, the coil can be
connected to a charged capacitor bank. By proper choice of capacitance and
coil inductance, the resonant frequency can be made sufficien tly low to ac-
complish demagnetization as a result of the clamped oscillatory nature of
the current in the coil- capacitor combination. It is difficult to completely
demagnetize large metallic magnets by means of ac methods because theed-

176
Sec. 5·5 Stability and Stabilization 177

dy currents generated near the surface tend to shield the interior from the ac
field. Very high values of the ac field will aid in this regard, as will lower fre-
quencies. AC fields are also applicable to calibration.
DC Fields. Using de, two methods can be employed to perform
demagnetization (although less than ideally). In the first, a magnetizer with
a reversing switch and a means to reduce the de current to zero are required.
The magnet is placed in the magnetizer and the field is applied in alternating
directions by means of the reversing switch and, a t the same time, its
magnitude is gradually reduced by, for instance, a rheostat in the de line or
a variable autotransformer in the ac line of the rectifier system. The quality
of demagnetization is dependent. upon the number of reversals and the field
reduction per step.
In the second method, the saturated magnet is exposed to a de field in a
direction opposite to saturation. The magnitude of the field, to be found by
experiment, must be such that after exposure, the magnet is demagnetized.
This is the least desirable procedure of the various methods herein describ-
ed. Since neither the de field nor the permanent magnet are perfectly
homogeneous, the best that can be accomplished will result in portions of
the magnet being slightly magnetized in one direction and other portions in
the opposite direction. DC fields are also applicable to calibration , although
with less convenience than in the case of ac fields.
Thermal Method. If the temperature of a magnetized magnet is raised
above the Curie temperature of the material and subsequently cooled to
room temperature, it will be found that demagnetization has been achieved .
Unfortunately, this is not practical for any of the metallic materials because
heating these above the Curie point causes metallurgical changes
resulting in unusably low magnetic properties thereafter. On the other
hand, this is quite practical for any of the ceramic grades. The Curie
temperature for these is approximately 460°C, and a slight safety factor of
40 to 50°C should be used to assure that all parts of the magnets will reach
at least 460°C. To avoid cracking because of heat shock, the ceramics
should be heated and cooled relatively slowly, depending upon the size and
shape of the part. Thermal demagnetization is not applicable to calibration.

5.5 Stability and Stabilization

Permanent magnets do not "run down." In this respect they differ fun-
damentally from batteries, radioactive materials, or the like. The magnetic
field surrounding a magnet does not require energy to maintain it,
therefore, there is no theoretical reason for a permanent magnet 10 con-
tinually lose strength . In practice, however, flux changes may occur as a
result of several factors. Proper stabilization will eliminate or reduce these.
 5.5.1 Metallurgical Changes

In the older permanent magnet materials, such as cobalt-steel, some

metallurgical changes take place as a function of time. If such a magnet is
magnetized before these changes have stabilized, flux changes superimposed
on those to be described in the next section will occur. (This effect can be
reduced to a negligible factor by artificial aging.) In the newer materials,
such as Alnico or ceramic, metallurgical changes do not take place to any
measurable degree at room temperature.
5.5.2 Time

A freshly magnetized permanent magnet will lose a minor percentage of

its flux, as a function of time. It has been shown that, if one plots flux loss
linearly against time logarithmically, an essentially straight line results.
Laboratory measurements on some materials are shown in Table 5-2. All
losses are based on measurements made starting at 0.1 after magnetizing.
TABLE 5·2 Laboratory measurements on selected magnet materials (percent).

Loss
Loss at /00,000h
Material Per Log Cycle ( 11.4 yrs }

Ceramic Esse ntially zero Essentially zero

Al ni co 3 (nea r max. energy) 0.4 2.4
Alnico 3 (near coercive) 0.6 3.6
Alnico 5 (near residual) 0.01 0.06
Alnico5(nearmax.energy) 0.15 0.9
Alnico 5 (near coercive) 0.4 2.4
Alnico 8-no d ata (expected to be less than Alnico 5)

This loss in flux can be essentially eliminated by a partial demagnetiza-

tion of the freshly charged magnet, in an amount of 7 to 15%. This is most
conveniently accomplished by means of an ac coil. The ac field should be in
the same direction as was the magnetizing field. It should be red uced to zero
gradually, either by withdrawing the magnet with power applied or by
reducing the ac voltage to zero whlle using a variable auto transformer.
5.5.3 Temperature

Temperature effects fall into three categories:

1. Metallurgical changes may be caused by exposure to too high a
temperature. Such flux changes are not recoverable by remagnetization.
The approximate maximum temperatures that can be used without ex-
periencing metallurgical changes range from 550 °C for Alnico 5 to J080°C
for the ceramics. The effect of metallurgical changes, if present, can be

178
Sec. 5·5 Stability and Stabilization 179

avoided only by long-time exposure of the magnet to the temperature in-

volved, prior to magnetizing.
2. Irreversible losses are defined as a partial demagnetization of the
magnet, caused by exposure to high or low temperatures. Such losses are
recoverable by remagnetization. Merely as examples, T able 5-3 shows
values measured on laboratory specimens, wit h percent flux losses
measured at room temperature after exposure to the indicated
temperatures. Percentages shown in Table 5-3 are not additive for con-
secutive cycles above and below room temperature.

TAB LE 6-3 Temperature effects on selected magnet materials {percent).

.Material 350°C 200°C - 20°c - 60°C

Ceramic5 0 0 0 0
Bd/ Hd> 2(above maxi mum energy)
Ceramic 6 0 0 0 0
Bd/ Hd> I. I (near maximum energy)
Alnico 5(near maximum energy) 1.3 0.8 I 2.5
Alnico 6(near maximum energy) 0.6 0.4 0.5 l.3
Alnico 8 (near maximum energy) 0.3 0.2 0. 1 0. 1
Alnico 8 (near coercive) 3.5 2.0 0.5 0.8

The ideal method for stabilizing magnets against temperature-induced

irreversible losses is installing them in the magnetic circuit for which they
are intended, magnetize, then subject the assemblies to several temperature
cycles which they are expected to experience in service. However, this is a
time-consuming procedure that is normally impractical in production.
Alternatively, the magnetized assembly may be partially demagnetized by
means of an ac field, following the procedure described in the last
paragraph of Sec. 5.5.2. A rule of thumb to follow is: determining by ex-
periment that temperature cycling will cause X% flux loss, the ac field
should be such as to cause a 2X% flux loss, to properly stabilize against
temperature.
3. Reversible losses are changes in flux that are reversible with
temperature. For example, if any of the ceramic grades are heated I °C
above room temperature, they will lose 0. l 9% of room temperature flux.
However, this will be spontaneously regained upon the magnet cooling back
at room temperature. The Alnico and E.S.D. materials have reversible
variations on the order of 1110 as great as the ceramics, depending upon the
material and the operating point on the demagnetization curve. One cannot
eliminate t hese reversible variations by stabilization treatments. However,
use of proper temperature-compensation material in parallel with the
magnet will reduce the effect to a negligible factor. Among others,
180 Permanent-Magnet Data Chap. 5

household watt-hour meter magnets and speedometer magnets are

temperature-compensated in this manner.
5. 5. 4 Reluctance Changes
If a magnet is magnetized in a magnetic circuit and subsequently sub-
jected to permeance change (such as changes in air gap dimensions or open-
circuiting of the magnet) it may be found that a partial demagnetization of
the magnet has occurred. Whether or not such a loss is experienced depends
upon material properties and upon the extent of the permeance change.
Stabilization against such change is accomplished either by several times
subjecting the magnet to such reluctance changes after magnetizing, or by
use of the previously described ac field.
In this section it should be mentioned that contacting the magnet with
ferromagnetic material (screwdrivers, pliers, and the like), at points other
than the poles can cause an appreciable drop in flux at the poles. It is dif-
ficult to stabilize against this type of abuse. The remedy is to avoid such
practices.

5.5.5 Adverse Fields

If a magnet or magnet assembly is subjected to an adverse magnetic
field, a partial demagnetization may result, depending upon material pro-
perties and upon the intensity and direction of the adverse field. Proper
stabilization consists of subjecting the magnet or assembly to a de or ac
demagnetizing field of the same magnitude as it is expected to encounter in
service. The direction should be the same as that of the anticipated
demagnetizing field.
5.5.6 Shock, Stress, and Vibration
The effects of shock, stress, and vibration below destructive limits on
most permanent magnet materials are so minor (a few tenths of a percent)
that little consideration need be given to them. Proper stabilization, as
described in any of the preceding sections, will also stabilize against shock
and vibration.

5.6 Measurements and Calibration

The selection of an appropriate test method for a permanent magnet should

be based on full knowledge of its application. Information should establish
the adverse conditions to which the magnet will be subjected in both the
normal functioning and anticipated abuse of the product. Such information
is frequently based on tests made on the product itself and should determine
the flux density in the magnet, demagnetizing influences from counter
Sec. 5-7 Handling of Permanent Magnets 181

magnetomotive force or variable permeance conditions, temperature limits,

shock and vibration requirements, and other factors.
Properly correlated magnet tests require that the magnet be magnetized
to saturation, in the proper direction, before proceeding with tests. It is im-
perative that the magnet be properly saturated, since improper magnetiza-
tion frequently causes difficulty in obtaining correlation between magnetic
tests and product performance (see Sec. 5.5.4).
Permanent-magnet acceptance tests as agreed upon by customer and
supplier are generally based on tests and information obtained on the final
product or inherent magnetic properties of the magnet.
Magnetic tests on permanent magnets can be either quantitative or em-
pirical. Quantitative tests include measuring of absolute values of flux, flux
density, force, and so on, in the end-use product. Empirical tests include ar-
bitrary tests with magnetimeters, fluxmeters, gaussmeters such as Hall-
effect devices, and other comparison devices. In most cases, quantitative
test data are essential in setting up magnetic test specifications. Once the
minimum standard and the proper test method have been established, the
choice of test will depend largely upon the available equipment.
Although vital to the proper performance of the finished product and to
the economic considerations involved, the design of the permanent magnet
and its associated magnetic circuit is a problem that is fully discussed in
available texts and design manuals.
It is assumed, therefore, that the design is adequate as applied to the
final product and that the proper grade of permanent-magnet material has
been selected and all dimensional factors have been optimized.

5.7 Handling of Permanent Magnets

The full utilization of permanent magnets is conditioned by proper handling

prior to and during final usage. The natural phenomena of permanent
magnets is such that, in many instances, the permanent magnet remembers
what was done to it or what environment it was exposed to. Adverse factors
affecting permanent magnets must be acknowledged and techniques applied
to minimize or eliminate these deleterious conditions. The design, manufac-
ture, handling, and processing of a permanent magnet is based on adequate
control of this sequence to ensure optimum performance in the final ap-
plication. It is no understatement that in a number of cases improper handl-
ing has resulted in poor, or completely substandard, performance.
A manufacturer who purchases and uses permanent magnets in a
product has a very simple choice in the procurement of this essential
material. The permanent magnet is purchased magnetized or unmagnetized.
The initial choice may be simple, but either of the two conditions may result
in quite distinct problems of processing the final device.
 5. 7. 1 Unmagnetized Magnets
From the standpoint of ease in shipping, handling, and storage, the un-
magnetized magnet is preferred. This condition alleviates special problems
of knockdown, iron-chip pickup, and special instructions to manufacturing
personnel.
To ensure magnetic quality, the permanent-magnet supplier has to test
the magnet in a fully saturated magnetized condition. After this inspection
the magnet is demagnetized prior to shipment. This implies that although
the permanent-magnet material is purchased unmagnetized, it has been sub-
jected to a complete magnetic cycle from an initial unmagnetized condition,
through complete saturation, and back to an unmagnetized state. It is quite
important to recognize that this sequence of events is a normal part of the
inspection process for the permanent-magnet supplier.
The accrued benefits of purchasing, shipping, and processing of un-
magnetized magnets are many. A few of the problem areas that can be
eliminated or minimized for a permanent magnet in this unmagnetized state
are:
Keepers are not required.
Proximity effects of other permanent magnets may be neglected.
Proximity of strong ac or de fields pose no problem.
Physical shock or vibration for critical applications may be ignored.
Shape problems of self-demagnetization are not applicable.
Physical handling problems are considerably alleviated.
Storage and shipment problems are minimized.

A final step must be performed on this unmagnetized condition of the per-

manent magnet-magnetizing and inspecting the functional magnetic field
of the permanent magnet after incorporating it in its intended device. This
implies that the manufacturer of the device has adequate means for satura-
tion and control of the magnetizing process for the permanent magnet. In
some cases stabilization must be included for temperature, ac or de electric
fields, and/or other effects.

5.8 Glossary of Terms

Although there may be slight redundancy between this glossary of terms

and the one in Appendix B, it was felt that for convenience it would be pro-
per to include this section. Many of the terms that apply to magnetized core
materials are not defined elsewhere in this volume-hence this important in-
clusion.

182
Sec. 5-8 Glossary of Terms 183

Ar Area of the Air Gap: The cross-sectional area of the air gap perpendicu-
lar to the flux path, is the average cross-sectional area of that portion of
the air gap within which the application interaction occurs. Area is mea-
sured in cmz in a plane normal to the central flux line of the air gap.
Am, Area of the Magnet: The cross-sectional area of the magnet perpendic-
ular to the central flux line, measured in cmz at any point along its
length. In design, Am is usually considered the area at the neutral section
of the magnet.
B, Magnetic Induction: The magnetic field induced by a field strength, H,
at a given point. It is the vector sum, at each point within the substance,
of the magnetic field strength and resultant intrinsic induction. Mag-
netic induction is the flux per unit area normal to the direction of the
magnetic path.
Bd, Remanent Induction: Any magnetic induction that remains in a mag-
netic material after removal of an applied saturating magnetic field, Hs.
(Bd is the magnetic induction at any point on the demagnetization curve;
measured in gauss.)
Bd/Hd, Slope of the Operating Line: The ratio of the magnetic induction,
Bd, to its self-demagnetizing force, Hd. It is also referred to as the per-
meance coefficient, shear line, load line, and unit permeance.
BdHd, Energy Product: The energy that a magnet material can supply to an
external magnetic circuit when operating at a point on its demagnetiza-
tion curve; measured in megagauss-oersteds.
(BdHd),,,, Maximum Energy Product: The maximum energy a material can
supply to an external magnetic circuit.
Bis• Saturation Intrinsic Induction: The maximum intrinsic induction
possible in a material.
Bg, Magnetic Induction in the Air Gap: The average value of magnetic
induction over the area of the air gap, A 8 i; or it is the magnetic induction
measured at a specific point within the air gap; measured in gauss.
B,, Intrinsic Induction: The contribution of the magnetic material to the
total magnetic induction, B. It is the vector difference between the mag-
netic induction in the material and the magnetic induction that would
exist in a vacuum under the same field strength, H. This relation is ex-
pressed by the equation.

where B, is the intrinsic induction in gauss, B the magnetic induction in

gauss, and H the field strength in oersteds.
Bw Remanent Induction: The magnetic induction that remains in a mag-
184 Permanent-Magnet Data Chap. 5

netic material after magnetizing and conditioning for final use;

measured in gauss.
BmHm, Energy Product: The energy that a magnet material can supply to
an external magnetic circuit when operating at the point BmH,,i>·
measured in megagauss-oersteds .
Bo, Magnetic Induction: The point of the maximum energy product
(Bd Hd )m; . measured in gauss.
B,, Residual Induction (or Flux Density): The magnetic induction cor-
responding to zero magnetizing force in a magnetic material after satur-
ation in a closed circuit; measured in gauss.
Bs, Saturation Induction: The induction at saturation; measured in gauss.
f, Reluctance Factor: Accounts for the apparent magnetic circuit reluc-
tance. This factor is required due to the treatment of Hm and Hg as con-
stants.
F, Leakage Factor: Accounts for flux leakage from the magnetic curcuit.
It is the ratio between the magnetic flux at the magnet neutral section
and the average flux present in the air gap. F = 8 11,Am!BgAg·
F, Magnetomotive Force (Magnetic Potential Difference): The line integral
of the field strength, H, between any two points, p , and P2·
P2

F =
P,
J Hdl

where F is the magnetomotive force in gilberts, H the field strength in oer-

steds, and di the element of length between the two points in centimeters.
H, Magnetic Field Strength (Magnetizing or Demagnetizing Force): The
measure of the vector magnetic quantity that determines the ability of an
electric current, or a magnetic body, to induce a magnetic field at a given
point; measured in oersteds.
He, Coercive Force, of a Material: The demagnetizing force corresponding
to zero magnetic induction, B, in a magnetic material after saturation;
measured in oersteds.
He;' Intrinsic Coercive Force, of a Material: Its resistance to demagneti4a-
tion. It is the demagnetizing force corresponding to zero intrinsic induc-
tion in a magnetic material after saturation; measured in oersteds.
Hd: That value of H corresponding to the remanent induction, Bd; mea-
sured in oersteds.
Hm: That value of H corresponding to the remanent induction, Bm; mea-
sured in oersteds.
Sec. 5-8 Glossary of Terms 185

Ha: The magnetic field strength at the point of the maximum energy prod-
uct (Bd H d)m: .measured in oersteds .
Hs, Net Effective Magnetizing Force: The magnetizing force required in
the material, for saturation; measured in oersteds.
lg, Length of the Air Gap: The length of the path of the central flux line of
the air gap; measured in centimeters .
Im, Length of the Magnet: The total length of magnet material traversed in
one complete revolution of the center line of the magnetic circuit; mea-
sured in centimeters.
l m!D , Dimension Ratio: The ratio of the length of a magnet to the diam-
eter of a circle of equivalent cross-sectional area. For simple geometries,
such as bars and rods, the dimension ratio is related to the slope of the
operating line of the magnet, BdlHd.
P, Permeance: The reciprocal of the reluctance, p; measured in maxwells
per gilbert.
R, Reluctance: Somewhat analogous to electrical resistance. It is the
quantity that determines the magnetic flux, ~ resulting from a given
magnetomotive force, F.
F
R =-
~
where R is the reluctance in gilberts per maxwell, F the magnetomotive
force in gilberts , and t the flux in maxwells.
Tc, Curie Temperature: The transition temperature above which a material
loses its permanent-magnet properties.
Vg, Air-gap Volume: T he useful volume of air or nonmagnetic material
between magnetic poles; measured in cubic centimeters.
µ, Permeability: A general term used to express various relationships be-
tween magnetic induction, B, and the field strength , H.
µre• Recoil Permeability: The average slope of the recoil hysteresis loop.
(Also known as a minor loop.)
~. Magnetic Flux: A contrived but measurable concept that has evolved in
an attempt to describe the "flow" of a magnetic field . Mathematically,
it is the surface integral of the normal component of the magnetic in-
duction, B, over an area, A.

~=ff B dA

where t is the magnetic flux in maxwells, B the magnetic induction in

gauss, and dA the element of area in cm1 • When the magnetic induction,
186 Permanent-Magnet Data Chap. 5

B, is uniformly distributed and is normal to the area, A, the flux, ~. =

BA.
A Closed-Circuit Condition: Exists when the external flux path of a
permanent magnet is confined within high-permeability material.
A Demagnetization Curve: The second (or fourth) quadrant of a major
hysteresis loop. Points on this curve are designated by the coordinates
Bd and Hd.
A Fluxmeter: A galvanometer that measures the change of flux linkage
with a search coil.
The Gauss: The unit of magnetic induction, B, in the cgs electromagnetic
system. One gauss is equal to 1 maxwell per square centimeter.
A Gaussmeter: An instrument that measures the instantaneous value of
magnetic induction, B. Its principle of operation is usually based on one
of the following: the Hall effect, nuclear magnetic resonance (NMR), or
the rotating coil principle.
The Gilbert: The unit of magnetomotive force, F, in the cgs electromag-
netic system.
A Hysteresis Loop: A closed curve obtained for a material by plotting (us-
ually to rectangular coordinates) corresponding values of magnetic in-
duction, B, for ordinates and magnetizing force, H, for abscissa when
the material is passing through a complete cycle between definite limits
of either magnetizing force, H, or magnetic induction, B.
A Keeper: A piece (or pieces) of soft iron that is placed on or between the
pole faces of a permanent magnet to decrease the reluctance of the air
gap and thereby reduce the flux leakage from the magnet. It also makes
the magnet less susceptible to demagnetizing influences.
Leakage Flux: Flux, ~. whose path is outside the useful or intended mag-
netic circuit; measured in maxwells.
The Major Hysteresis Loop of a Material: The closed loop obtained when
the material is cycled between positive and negative saturation.
The North Pole of a Magnet: The pole that is attracted by the geographical
North Pole. The north pole of a magnet repels the north-seeking pole of
a compass.
The Maxwell: The unit of magnetic flux in the cgs electromagnetic system.
One maxwell is one line of magnetic flux.
The Neutral Section of a Permanent Magnet: Defined by a plane passing
through the magnet perpendicular to its central flux line at the point of
maximum flux.
The Oersted: The unit of magnetic field strength, H, in the cgs electromag-
Sec. 5-8 Glossary of Terms 187

netic system. One oersted equals a magnetomotive force of 1 gilbert per

centimeter of flux path.
An Open-Circuit Condition: Exists when a magnetized magnet is by itself
with no external flux path of high-permeability material.
The Operating Line for a Given Permanent Magnet Circuit: A straight
line passing through the origin of the demagnetization curve with a slope
of negative Bd/Hd. (Also known as permeance coefficient line.)
The Operating Point of a Permanent Magnet: That point on a demag-
netization curve defined by the coordinates BdHd or that point within
the demagnetization curve defined by the coordinates BmHm.
An Oriented (Anisotropic) Material: One that has better magnetic proper-
ties in a given direction.
A Permanent Magnet: A body that is capable of maintaining a magnetic
field at other than cryogenic temperatures with no expenditure of
power.
A Permeameter: A complex piece of equipment that can measure, and
often record, the complete magnetic characteristics of a specimen of a
magnetic material.
Magnetic Saturation of a Material: Exists when an increase in magnetizing
force, H, does not cause an increase in the intrinsic magnetic induction,
B, of the material.
A Search Coil: A coiled conductor, usually of known area and number of
turns, that is used with a fluxmeter to measure the change of flux linkage
with the coil.
The Temperature Coefficient: A number that describes the change in a
magnetic property with a change in temperature. It usually is expressed
as the percentage change per unit of temperature.
An Unoriented (Isotropic) Material: Equal magnetic properties in all direc-
tions.

BIBLIOGRAPHY

BARTA, G.T., "Permanent Magnet Processing," Elec1ro-Tech110/ogy, May 1966.

C LEGG, A.G., AND M. MCCRAIG, "High-Temperature Stability of Permanent Mag-
nets of the Iron-Nickel-Aluminum System," British J. Appl. Phys., Vol. 9, No.
144, 1958.
DIETRICH, H., "Time and Temperature Dependence of the Structural Changes in
Permanent Magnet Materials," Cobalt, No. 30, March 1966, pp. 3- 18.
188 Permanent-Magnet Data Chap. 5

GORDON, 0.1., "Environmental Evaluation of Magnetic Materials," Electro-Tech-

nology, Vol. 67, No. 1, January 1961 , pp. 118- 125.
GoRDON, D.l., "Irradiating Magnetic Materials," Electro-Technology, June 1965,
pp. 42-45.
GORDON, 0.1., AND R.S. SERY, IEEE Trans. Commun. Electron., Vol. CE-83, No.
73, July 1964, pp. 357-361.
GORDON, D.I., R.S. SERY, AND R.H. LUNDSTEN, "Nuclear Radiation Effects in
Magnetic Core Materials and Permanent Magnets," ONR-5, Materials Research
in the Navy, Symposium, Philadelphia, 1959 (Office of Naval Research, Wash-
ington, D.C., 1959), pp . 253-292.
HADFIELD. D., Permanent Magnets and Magnetism. John Wiley & Sons, Inc., New
York, 1962.
KRONENBERG, K.J., AND M.A. BOHLMANN, "Long Term Magnetic Stability of Al-
nico V and Other Permanent Magnet Materials," WADC Rept. 58-535. "Long
Term Magnetic Stability of Alnico and Barium Ferrite Magnets," J. Appl.
Phys., Vol. 31, May 1960, pp. 825- 845.
PAHKER, R.J. AND R.J. STUDDERS, Permanent Magnets and Their Application .
.John Wiley & Sons, Inc., New York, 1962.
ROBERTS, W.H., "Performance of Permanent Magnets at Elevated Temperatures,"
J. Appl. Phys., Vol. 29, No. 405, 1958.
ScmN DLER, A.I., AND E.I. SALFOVITZ, " Effect of Applying a Magnetic Field during
Neutron Irradiation on the Magnetic Properties of Fe-Ni Alloys," J. Appl.
Phys., Vol. 31 supplement to No. 5, 1960, p. 245-S.
SERY, R.S., AND D.I. GORDON, "Irradiation of Magnetic Materials with 1.5- and
4-meV Protons," J. A ppl. Phys., Vol. 67, No. 4 (Pt. 2), April 1963, pp. 1311 -
1312.
S ERV, R.S., D.I. GORDON, AND R.H. LUNDsn: N, NAVORD Report 6276 (U.S.
Naval Ordnance Laboratory, White Oak, Md.), 1959.
SF.RV, R.S., R.H. LUNDSTEN, AND D.I. GORDON, Naval Ordnance Laboratory TR-
61-45, Silver Spring, Md., May 18, 1961.
SPREADBURV, F.G., Permanent Magnets. Pitman and Sons Ltd., London, 1949.
TENZER, R.K., "Influence of Various Heat Exposures of Alnico V Magnets," J.
Appl. Phys., Vol. 30, supplement to No. 4, 1959, p. 115-S.
TENZER, R.K., "Temperature Effects of the Remanence of Permanent Magnecs,"
Technical Documentary Report ASD-TDR-63-500, Office of Technical Services
AD-420235, U.S. Department of Commerce.
APPENDIX A
REFERENCES

No book can be considered complete without the inclusion of a com-

prehensive list of selected papers, manufacturers' application notes, and
books. Although there are a number of references that relate to magnetic-
core materials and their applications, only those which the author feels are
absolutely pertinent to the subject are listed here. This bibliography pro-
vides an exceptionally concise cross section of the available material which
can be used to broaden the reader's knowledge of magnetic materials and
how they can be applied in practical experience.
The listing that follows immediately is provided by the IEC. It will be of
particular interest to those who plan to, or are presently engaged in com-
mercial design and manufacture of magnetic-core components. These docu-
ments can be obtained from the American National Standards Institute,
Inc., 1430 Broadway, New York, NY 10018.
The following compilation of selected papers is offered as reference
material on a wide spectrum of subjects that deal with magnetic cores and
their applications. The material is useful to the engineer, technician, or stu-
dent as supplemental theory to that provided in this book.

SELECTED TECHNICAL PAPERS

[l] R. AsTLE, "Optimum Shapes for Inductors," IEEE Trans. PM & P, March
1969.
[2] H. BLINCHIKOFF, "Toroidal Inductor Design," Electro-Technology, Nov. 1964.
[3] B URNELL & Co. ENGINEERING STAFF, "The Application of Iron Powder Cores
in Electric Wave Filters," Progr. Powder Metalf., 1962.

189
 TAB LE A· 1 I EC publications on linear ferrites.

125 (1961) General classification of ferromagnetic oxide materials and definitions of

terms .
Amendment No. I (1965)
Amendment No. 2 (1968)

133 (1967) Dimensions for pot cores made of ferromagnetic oxides and associated parts
(2nd edition).

133A (1970) First supplement to Publication 133.

133B (1971) Second supplement to Publication 133.

Amendment No. I (1975)

205 (1966) Calculation of the effective parameters of magnetic piece parts.

205A ( 1968) Supplement to Publication 205.

2050 (1974) Supplement to Publication 205.

220 (1966) Dimensions of tubes, pins, and rods of ferromagnetic oxides.

221 (1966) Dimensions of screw cores made of ferromagnetic oxides.

Amendment No. I (1968)
Amendment No. 2 (1976)

22 1A (1972) First supplement to Publication 221.

223 (1966) Dimensions of aerial rods and s labs or ferromagnetic oxides.

223A (1972) First supplement to Publication 223.

226 (1967) Dimensions of cross cores (X-cores) made o r ferromagnetic oxides and
associated parts.

226A ( 1970) First supplement to Publication 226.

367-1 ( 197 1) Cores for inductors and transformers for telecommunication.

Part I-Measuring methods.
Amendment No. i (1976)

367-IA ( 1973) First supplement.

367-1 B (1973) Second supplement to Publication 367-1.

367-I C (1974) Third supplement.

367-2 (1974) Part 2-Guides for the drafting of performance specifications.

367-2A ( 1976) Fi rst supplement to Publication 367-2.

401 ( 1972) Information on ferrite materials appearing in mnnufacturers' catalogues of

transformer and inductor cores.

424 (1973) Gui de to the specification or limits for physical imperfections of parts made
from magnetic oxides.

431 (1973) Dimensions of square cores (R M cores) made of magnetic oxides and
associated parts.

525 ( 1976) Dimensions of toroids made from magnetic oxides or iron powders.

190
 Selected Technical Papers 191

[4] E.H. CHANT, JR .. "Moly Permalloy Cores," MPIF Sem., June 1969.
(5) R.A. CHEWIODEN. "A Review of Magnetic Materials-Especially for Com-
munications Systems," Metal Progr., 1948 .
[6] E.G. CRISTAL, "Tables of Maximally Flat Impedance-Transforming Networks
of Low-Pass-Filter Form," IEEE Trans. Microwave Theory Techniques, Vol.
MTT 13 , No. 5, Sept. 1965, Corres.
(7) M.F. " DOUG" DEMAw , "The Practical Side of Toroids," QST Maga·dne
(ARRL, Inc.), June 1979.
(8) R.M . FA NO. "Theoretical Limitations on the Broadband Matching of Arbitrary
Impedances," J. Franklin Inst., January-February 1950.
[9] G.B. FI NKE, "Moly Permalloy Powder Cores-Their Characteristics and Appli-
cations," MP!F Sem., December 1969.
(10) E.A. GAUGLlm, "Soft Magnetic Materials," Product Eng. , July 1949.
(11) A.J. HARli.NDZA.ffARINXMA, "Recent Developments in the Manufacture of Per-
malloy Powder Cores," Western Electric Eng., 1964.
[12] J.H . HOROWITZ, "Design Wideband UHF Power Amplifiers," Electron.
Design, May 24, 1969.
[13) G.A. KELSALL, "Permcameter for Alternating Current Measurements at Small
Magnetizing Forces," Opt. Soc. Amer. J., Vol. 8, Feb. 1924.
[14) BoHDAN KOSTYSHYN AND PETER H. HAAS, "Discussion of Current-Sheet Ap-
proximations in Reference to High-Frequency Magnetic Measurements," J.
Res. NBS, Vol. 52, No. 6, June 1954.
[1 5) KRA USS-ALLEN, "Designing Toroidal Transformers rn Optimize Wideband Per-
formance," Electronics, Aug. 1973.
[16] LEfft::RSON, "Twisted-Wire Transmission Line," IEEE Trans. Parts, Hybrids,
Packaging. Vol. PHP-7, No. 4, Dec. 1971.
[17) V.E. LEGG, "Magnetic Measurements at Low Flux Densities Using the A.C.
Bridge," Bell Systems Tech. J., Vol. I?, Jan. 1936.
(18] V.E. LEGG, "Analysis of Quality Factor of Annular Core Inductors," Bell
System Tech. J., Vol. 39, No. I, Jan. 1960.
[1 9] V.E. LEGG AND F.J. GIV EN, "Compressed Powdered Molybdenum Permalloy
for High Quality Inductance Coils," Bell System Tech. J., Vol. 19, July 1940.
[20] G.L. MATTHAEI, "Tables of C hebyshev Impedance-Transforming Networks of
Low-Pass Filter Form ," Proc. IEEE, Aug. 1964.
(21] E.J . O ELllERMAN, "Moisture Aging of Powder Core Toroids," Electronics,
May 1953 (with R.E. Skipper and W .J. Leiss).
(22] C.D. OwENS, "Stability Characteristics of Molybdenum Permalloy Powder
Cores," Elec. Eng., March 1956.
[23) o. PITZALIS AND R.A. GILSON. "Tables of Impedance Matching Networks
Which Approximate Prescribed Attenuation versus Frequency Slopes," IEEE
Trans. Microwave Theory Techniques, Vol. MTT-19, No. 4, April 1971.
192 References App.A

(24) 0 . P ITZALIS AND C OUSE. "Broadband Transformer Design for RF Transistor

Amplifiers," ECOM-2989, U.S. Army Electronics Command, Fl. Monmouth,
N.J., July 1968.
(25) STA NLEY PRO, "Toroid Design Analysis," Electro-Technology, Aug. 1966.
[26) W. Q UEFURTH , "Turns Computation for Small Core Inductances," Elec. De-
sign News, Sept. 1958.
[27) C.E. RJCHARDS. p .R. B ARDELL, S.E. B UCKLEY, AN D A.C. LYNCH, "Some
Properties and Tests or Magnetic Powders and Powder Cores," Elec.
Commun., Vol. 28, March 195 1.
(28) C.L. R UTHROFF, "Some Broadband Transformers," Proc. IRE, Vol. 47,
August 1959.
[29) C.F. SA LT , W.T. SACKETT, JH., ANO R.C. MCMASTER. "Research and Develop-
ment of Various Configurations of Core Materials for Optimum Transformer
Design," Battelle Memorial Institute, Columbus, Ohio, 1952.
(30] J . SEVICK, " Simple Broadband Matching Networks," QST Magazine (ARRL,
Inc.), January 1976.
[3 1) M .D. SUTTON, "Application of Iron Powder Cores in electromagnetic Delay
Lines," Progr. Powder Mewl/., 1962 .
[32) R. T URRIN, "Application or Broadband Balun Transformers," QST Magazine
(ARRL, Inc.), April 1969.
[33) J .E. WOL F A ND B. KRAMER, "A Production Technique for the Determination
of Inductance for Toroidal Powdered Iron Cores," Prag. Powder Metal/.,
1964.
(34) T.G. WILSON, S .Y. F ENG, AND W .A. SANDER, " Optimum Toroidal Inductor
Design Analysis," Proc. 1970 Electron. Components Conf
[35] Symposium of Papers on Ferromagnetic Materials, IEE, Vol. 97, Part II, No.
56, April 1950.

APPLICATION NOTES

Manufactu rers' a pplication notes are generally available at no cost to those

who request them by mail. Complete lists of each m anufacturer's technicaJ
bulletins and app lica tio n notes a re also available upo n request. The names
and addresses o f vario us manu facturers of magnetic-core materials are
given fo llowing the list of notes.

[I] Broadband Linear Power Amplifiers Using Push-Pull Transistors," by H.

Granberg, Motorola Semiconductor Products, Inc., AN-593.
(2) Design of H.F. Wideband Power Transformers," by H ilbers, Phillips Applica-
tion Information No. 530.
[3] "Get 300 Watts PEP Linear Across 2 to 30 MHz from this Push-Pull Ampli-
fier,'' by H. Granberg, Motorola Semiconductor Products, Inc., EB-27.
 Pertinent Books 193

[4] "Matching Network Designs with Computer Solutions," Motorola Semicon-

ductor Products, Inc., AN-267.
[5] "Notes on Low Impedance H.F. Broadband Transformer Techniques," by
Lewis, Collins Radio Co., November 1964.
[6) "Precision Toroidal Inductors," Sprague Tech. Bull. No. 41000.
[7] "Systemizing RF Power Amplifier Design," Motorola Semiconductor P rod-
ucts, lnc., AN-282.

MANUFACTURERS

Amidon Associates, 12033 Otsego St., North Hollywood, CA 9 1607.

Arnold Engineeri ng Co., Marengo, IL 60152.
Fair-Rite Products Corp., Walkill, NY 12589.
Ferroxcube Corp., 5083 Ki ngs Highway, Saugerties, NY 12477.
Indiana General, Crows Mill Road, Keasbey, NJ 08832.
Magnetics (Division of Spang Industries), P.O. Box 391, Butler, PA 16001.
Micrometals, Inc., 228 North Sunset, City of Industry, CA 91744.
Motorola Semiconductor Products, Inc., P.O. Box 20912, Phoenix, AZ 85036.

PER TINENT BOOKS

The subject& of magnetic core materials and related network design are
treated in the following books. T he engineer or student may find these
references of value as supplemental material to the topics covered in this
volume.

(1] P.R. BARDELL, Magnetic Materials in the Electrical Industry. Philosophical

Library, Inc., New York, 1955.
[2] H .W. B ODE. Network Analysis and Feedback Amplifier Design. D. Van Nos-
trand Co., New York.
[3] R.M. B OZO RT H, Ferromagnetism. D. Van Nost rand Co., New York, 1951.
[4] F.C. C ONNEL LY , Transf ormers. Pitman and Sons Ltd., London, 1950.
(5) M . F . " DOUG" DEMAw, Practical RF Communications Dara for Engineers and
Technicians (No. 21557). Howard W. Sams & Co., Indianapolis, Ind., 1978.
(6) M.F. " DOUG" DEMAW , ARRL Electronics Data Book. American Radio Relay
League, Inc., Newington, Conn., 1976.
(7) P.R. G EFFE, Simplified Modern Filter Design. Hayden Book Co., Inc., New
York.
[8] R. LEE, Electronic Transformers and Circuits. John Wiley & Sons, Inc., New
York, 1955.
194 References

[9] DEPARTMENT OF ELECTRICAL ENGINEERING, M.l.T., Magnetic Circuits and

Transformers. John Wiley & Sons, Inc., New Yor k, 1943.
[10] W.J . POLYDOROFF , High-Frequency Magnetic Materials. John Wiley & Sons,
Inc., New York, 1960.
[I I] H. F. STORM, Magnetic Amplifiers. John Wiley & Sons, Inc., New York, 1955.
APPENDIX B
MISCELLANEOUS DATA

This appendix contains an assortment of useful information for engi-

neers and students who are involved with ferromagnetic components. The
author has endeavored to include a wide assortment of data that would
otherwise be found scattered throughout the countless technical literature
which pertains to components that contain magnetic-core material. Credits
are given wherever possible to those manufacturers and distributors of
magnetic cores from which the information has been gleaned.
Table B-1 contains a comprehensive listing of the symbols and their
definitions for magnetic-core materials. These symbols and terms are stan-
dard within the industry, with the notable exception of the expressions of
permeability. Some manufacturers designate initial permeability as µo,
whereas others assign the symbol /Ji· Similarly, some fabricators prefer 1-lav
to f:.le when symbolizing average or effective permeability.

SELECTED CONVERSION FACTORS

Table B-2 contains a group of important conversion factors which find fre-
quent application among engineers and students who are concerned with the
design and use of components that utilize magnetic cores . Metric/English
conversion data are also included for those areas of engineering that encom-
pass transformer and inductor design.

195
 TAB LE B-1 Magnetic-core symbology.

S ymbol Units T erm Description

Ac in1 Availa ble windimz Cross-sectional area (perpendicular to

cm 1 area of core ~ directi on of wire). ava ilable fo r wi nding
turns on a particular core..
Acn in1 Avai lable windin!! Cross-sectional area (perpendicular to
cm 1 area ofbobbin - direction of wire), ava ilable for winding
turns on a particular bobbin.
A,, cm2 Effective area The cross-sectional area that an equivalent
of core gapless co re (of uniform magnetic and geo-
metric properties) would have.
Ag cm 2 Effective gap Theequiv:ilen t area through which the
area (assumed un iform) flu x in a magnetic core gap
passes. Cor rects for fringi ng effects.
AL mH / IOOO tu rns Inductance index Relates induct;mcc to turns for a particular
core and gap.
A,,, cm 2 Effective area of In a gapped structure, thcequivalent cross-
magnetic path scctional area of the magnetic part oft he
path.assumed uniform.
A,, in 2 Avai lable winding That porti on of the tota l ava ilable winding
cm1 space for primary area allo tted to the transfo rmer primary
win ding.
Ax in1 Cross-sectional The cross-sectional area of th e conductor
cm 2 area of wire part o f the wi re.
B ga uss Magnetic flux The flux density (lincs/ cm 2) in a magnetic
density circui t. measured at :i given point.
Brnax gauss Maximum flux The va lue offlux density corresponding to the
density peak of the applied excitation.
Bo ga uss Magnetic flux Maximum valuesoffl ux density forµ = µ ein
density tran sfo rmer core.
B, ga uss Pulse-excited The va lue offl ux density co rresponding to the
flux densi ty instant of termination of an applied rectangular
pulse.
BW ht: rl 7. Bandwidth Frequency range over which respo nse of a
tuned transformer is considered to be un iform
fo
( BW = 21:>/ = - -
Q
c { farads Equivalent shunt Theequi va lent shunt ca pacitance represented
capacitance by the interwinding capacitance of a
transfo rmer.
£,,, vo lts, RM S Magnet izing Effecti ve va lue of alternating voltage applied
vo ltage across primary magnetizi ng ind uctance
(sec L111).
£,, vo lts. RMS Transformer Effective va lue of alternating voltage ap plied
primary terminal across primary term inals of a tra nsfo rmer.
voltage
Epmax volts. Peak Peak primary Peak va lue of £,,(forsincwavcs.
£,,max = £,,VL ).
Es volts. RMS Transformer Effective value of altern ating \•oltage applied
secondary across secondary (ou tput) te rmin als of a
terminal voltage transfo rm er.
f her tz Frequency Alternations per second of voltage or current.
/1 hert z Low-frequency Lowest frequency :11 whi ch the sinusoidal ampli-
CUt· Off tude vs. frequency cha racteristic of a trans-
former is within 3 dB of the midfreq uency level.

196
 TABLE B-1 (continued)

Symbol Units Term Description

!2 hertz l-ligh-frcqucncy Highest frequency at which the sinusoidal ampli-

cut-off tudc vs. frequency characteristic or a trans-
former is withi n J dB of the midfrequency level.
H oersteds Magnetization The magnetizing force that produces the
magnetic nux in a transformer or inductor core.
Ampere-turns per cm.
H max oersteds Peak magnetiza tion The peak value of the magneti zi ng force.
Ho oersteds De magnetization The de magnetizing force applied to a core
(unidirectional current flowing in its windings).
/(" amperes, RMS Core loss current Component of magnetizing current accounting
for power lost in core material.
1dc amperes Direct current The magnitude of the unidirectional compon-
ent of a current.
Im amperes, RMS Magnetizing Component of primary alternating current
cur rent devoted to magnetization of transformer core.
amperes, RMS Transformer Total altern ating cur rent Ao wing in transformer
primary current primary winding. Vector su m of magneti zi ng
(I,,,) and loaCl (I, ~ 11) currents.
I map amperes, Peak Peak input current The peak value oft he transformer input current.
p
Is amperes, RMS Transformcr Total altern ating current Ao wing in transformer
seco ndary cu rrent secondary winding.
K numeric Dielectric The ratio oft he capacitance of a capacitor
constant ofa using that substance as a dielectric to the
substance capacitanceofthesamestructure with a
vacuum as the dielectric.
k, (gauss)? µ vs. B factor Factor relatingµ and B for a given power
inductor design requirement.
k2 volt-amperes Core-sclcclion Factor rel ating the volt-ampere rating of a
per Hz X 2ir facto r core to the frequency.
Lm henries Self-inductance Self-inductance of a transformer primary winding
(L m = N p1 AL X l0- 9 henry)
Lo henries Air inductance L for µ = I (air). Thus. L = µ 0 L0.
L, henries Leakage Self-induct ance that appears in series with the
inductance windings due to leakage Aux that does not
contribute to the transfer of energy in a
transformer.
IC in. Length or EITec1ive length ofa uniform Aux path.
cm. Aux path
lg in. Length of EITective length of a uniform air gap in a
cm. air gap magnetic circuit.
I,,, in. Length of Effective length of a uniform magnetic path.
cm. magnetic path
M numeric Load / magnetizing The rati o of th e load component(/< 7 11) to the
cu rrent ratio magnetizing component (/111) in the transformer
primary current (IP).
Np numeric Primary turns Total number of complete turns in transformer
primary winding.
NS numer ic Secondary turns To tal number of complete turns in transformer
secondary winding.
II numeric Turns ratio Ratio of transformer primary / secondary turns.
Equivalent to primary/ secondary voltage ratio
(neglecting resistance of windi ngs)

197
 TABLE B-1 (cont inued )

Symbol UniH Term Description

percent Droop Percent droop in the nominall y f1at portion of

the waveform of a pulse deli vered by a trans-
fo rmer, compared with the input pulse waveform.
watts Load power The power dissipated in the external load
co nnected to the transformer secondary terminals.
waits Power dissipation The power dissipation level that will cause a
factor temperature rise of 50°C above ambient in a
given core.
walls Total power The total power dissipated in an inductor
dissipation (winding loss + core loss). Neglects dielectri c
losses.
Q numeric Quality factor Ratio of reactance to equivalent series
resistance in an impedance.
ohms Equivalent core The val ue of resistance across which the magnetizing
loss resistance excitation voltage would dissipate a power equal
to the core loss of a transformer or inductor.
ohms Generator Effecti ve internal series resista nce of the
resista nce exci tation sou rce.
ohms Load Equi valent resistance of external load con-
resistance nected bet ween transformer secondary terminals.
oh ms Resistance of De resistance of tra nsformer primary.
primary winding
ohms Resistance of De resistance of transformer secondary.
seco ndary winding
•c Maximum oper- The sum of the maximum expected am bient
ating temperatu re temperature and the winding temperatu re rise
in the core material.
·c Temperature Rise in tem perature of an inductor or trans-
rise former core due to internal power dissipation.
seco nds Pulse dura tion Duration of the longest nomi nall y f1 at portion
ofa pulse waveform.
seco nds Pulse rise time Longest allowable time for a pulse to rise from
10% to 90%ofits final ampliwde.
ve in.1 Effecti ve core The effective volume of a magnetic co re
cm3 volu me (Ve= lex Ae) definedas the volumeof a
un iform magnetic path le long. havi ng a cross
sectional area A e·
ohms Leakage The reactance of transformer leakage induc-
reactance tance L1 (X1 = 211' /L 1).
a numeric Attenuation Midfrequency reduction in signal amplitude
constant from primary terminal voltage to voltage
es tablished across the load.
a I
numeric Total Mid frequency reduction in signal amplitude
attenuati on from genera tor internal EMF to transfo rmer
output terminal voltage.
C..f hertz Half-bandwidth Diffcrcncc between center frequency and lo,11
frequen cy cutoff point (f1) on reso nance
curve of t tined transformer.
µ numeric Permeability Rati o of magne tic flux density to the exci-
tat ion producing it ( µ = Bf H).
av numeric Average The slope of the stra igh t line between 0, 0.
permeability 11nd B = Bmax for a given co re material
8 max
(µav = - - )
Hmax
11 0
numeric Initial The effective permeability of the core at low
permeability excitation (in the linear regi on) for a uniform
ui
magnetic path.

Courtesy of Fcrroxcubc Coq>.

 TAB LE B·2 Selected conversion factors.

Multiply: Hy: To O/J1ai11:

Wei g In

pounds 453.59 grams

pounds 0.45359 kil ogra ms
gra ms 0.0022046 pounds
kilograms 2.2046 pounds

Length

feet 30.480 cen1imc1crs

in ches 2.5400 centimclcrs
centi meters 0.032808 feet
centimeters 0.39370 inches
inches 0 .025400 meters
meters 39.370 in ches

Arca

squa re feet 929.04 square ccnl imctcrs

squa re inches 6.4516 square centimeters
squa re centimeters 1.0764 x 10- J square feet
square cen1imc1crs 0. 15500 square inches
squa re inches 6A516 X 10- · squ;irc meters
square meters 1.5500 x 10' square inches
squa re centimeters 10-' squa re meters
s<1ua re meters 10' squarccc111i111e1crs

Sinuso id al Wa vefo rm

pea k cu rren t o r vo l1age 0.707 11 rmsc urrcnl o r vo l1agc

peak c ur rem or vo l1agc 0.63662 avcragccurrcnl or
\'Oltagc
r m s cu rren1 or voltage 1.4142 peak current or voltage
rmscurrent or voltage 0 .90032 average current or
\•olta~c
average current o r voltage 1.5708 peak Ct~rrem or voltage
averagecu rrcn1 or voltage 1.1107 rms current or volt:1gc

Magnetic Induction. 8

gauss 6.45 16 lines per sq uare in ch

gauss 6.45 16 x 10- • wcbers per square inch
ga uss 10 -·1 wcbcrs per squa re meter
(lcs las)
lines per square inch 0. 15500 ga usses
lines per square inch I .5500 X JO- s we hers per square meter
(lcslas)
lines per square inch 10- 5 wcbcrs per square inch

199
 TABLE B-2 (continued)

;'v/11/riply: By: To Obf(li11:

we be rs per sqllare inch l .5500 X 10 7 g,1usses

wcbcrs per squa re inch 10s lines pe r square inch
we be rs pe r sq uare inch 1550 wcbers per square meter
(teslas)

Magnetizing Force. N
oersteds 2.0213 ampere-turns per inc h
oersteds 0.79577 ampere- turn s per
cen timete r
oersteds 79.577 ampere-t urns per meter
ampere-turns per centimeter 1.2566 oc rsteds
ampere-t urns pe r centimeter 2.5400 ampere-turns per inch
am pere-t urn s pe r cent imctcr 100.00 ampere-tllrns per meter
ampere- turns per inc h 0.49474 ocrstcds
a111perc-turns per inch 0.39370 ampere-turns pe r
centimeter
ampere- turns per inch 39.370 ampcre-1 urns per meter
ampere-turns per 111cter 0.012566 ocrsteds
a111pere- turns per meter 10- ~ amperc-IUrnS per
cen timeter
a111pere-turns per meter 0.025400 am pere-turns per inch

Permeabi lity
ga uss per oe rsted 3. 1918 lines per am pere-turn
inch
gauss pe r oerste<l 3. 19 18 x 10- • we bers pcr am pcrc-t urn
inch
gauss pe r ocrsted 1.2566 x 10 - • we hers per ampere-t um
meter
webers per ampere-turn meter 7.9577 x 10~ gauss per ocrstcd
webcrs per ampere-turn meter 2.5400 x 10'• Jines per ampere-turn
inch
wcbcrs pe r ampere-turn mete r 0.025400 wcbcrs per ampe re-turn
inch
wcbcrs pcrampere-tu rn inch 3.1330 x 107 ga uss per oersted
webers pera111perc-tllrn inch 10s lines per ampere-turn
inch
webers per ampere-turn inch 39.370 we be rs per ampere-turn
meter
lines per ampere-t urn inc h O.J IJJO gauss per oerstcd
lines per am pcre-t urn inch 39.370 x 10- • wcbcrs per am pere-turn
mete r
lines per ampere-turn inch 10 - • wcbers pe r a111pcre-t11rn
inch

200
 CONVERSION INFORMATION FOR WIRE GAUGES
10 TO 44

Columns A and B of Table 8-3 provide data on the common circular mils
notation and the metric equivalent for each wire gauge. Column C is useful
to the engineer or student, as it provides the equivalent resistance in
microhms/centimeters (µQ/cm or io- 6 Q/cm), Columns D through L
contain important information concerning coated wires and the effect of
the insulation on the size and number of turns, plus the total weight in
g/cm.
The designer can determine the total resistance of a winding by multiply-
ing the MLT (mean length/turn) of the winding in centimeters by the
µQ/cm for the chosen wire gauge and the total number of turns. Hence,

µQ
R = (MLT) x -cm x N

where N is the number of turns employed.

The weight of the copper in a specified winding can be found by
multiplying the MLT by the g/cm (see column L) and by the total number
of turns.

g
wt = (MLT) x - - x N
cm

The turns per square inch and per cm 2 are based on a 60% wire-fill factor .

INDUCTANCE, CAPACITANCE, AND REACTANCE

Although a calculator can be used to learn the inductance, capacitance, or

reactance of circuit elements, it is sometimes more convenient to obtain a
quick approximation by means of a nomograph. Figure B-1 contains the
necessary information for those who prefer the nomograph format. Induc-
tance values from 0.01 µH to 1000 Hare provided. Reactance in ohms is in-
cluded from 0.01 Q to l MQ. Capacitance values from 0.1 µF to 10 µFare
listed and frequencies between IO Hz and l MHz are included.

RESISTANCE/ TEMPERATURE CORRECTION FACTOR

ln many design applications it is imperative to know the resistance of a trans-

former or inductor winding with respect to tav (average temperature). Cor-
rection factors provide suitable resistance values versus temperature for

201
 TABLE B-3 Circular mils notation and metric equivalents.

ll .\l< I I\ RI· \ Rrs tST.\Nc~

-
H HVY SYNTll ETI CS

tlW(i
Wirt• ('Ill : x 10- i:.1
mi/h
/0-., I
cm at 20°(' c1111 X 10
t i t<'{/

cm
Diamett r
1

,.,,_h cm
Turns per:

i11.h
Turt1.!f.f1t'r:

""'!
We1~h1

x/1·111
Si:" l ' IY ' (trllltl"
"'·
10 52.61 10.384 32.70 55.9 11.()..16 0.267 0.1051 3.87 9.5 10.73 69.20 0.-168
II 41.68 8.126 41.37 44.5 8.798 0.238 0.0938 4.36 10.7 13A8 89.95 0.3750
12 33.08 6,529 52.09 35.6-1 7.022 0.213 0.0838 4.85 11 .9 16.81 108.4 0.2977
I3 26.26 5.184 65.64 18.36 5.610 0.190 0.0749 5.47 (j,4 21.15 IJ(>.4 0.2367
14 20.82 4.109 82.80 22.95 -1.556 0.171 0.0675 6.()..1 14.8 2C>.14 168.<· 0. 1879
15 16.51 3.260 1()..1.3 18.J7 3.62-1 0.153 0.0602 6.77 16.6 32.6<> 110.6 0. 1492
I<> 13.07 2.581 I 31.8 1-1.73 2.905 0.137 0.0539 7.32 18.6 40.7.l 262 .7 0.118-1
17 10.J'I 2.052 165.8 11.68 2.323 o.m 0.0-182 8.18 20.8 5U6 HU 0.0943
18 8.228 1.62-1 209.5 9.326 1.857 0.109 0.()..131 9.13 13.2 64 .3.1 414 .9 0.07-172
19 6.531 l.2X'I ~f. .1 •) 7.539 1.-190 0.0980 0.0386 10.19 25 .9 79.85 515.0 0.05940
20 5.188 1.024 ,,~ . J 6.065 1.197 0.0879 0.0346 11.37 28 .9 9!1.93 638 I 0.04726
21 4.1 16 812.J 418.9 4.$37 954.8 0.0785 0.0.109 12.75 32.4 124.0 799.R 0.03757
22 J.2-IJ 640.1 531.-1 3.857 761.7 0.0701 0.0276 1-1.25 36.2 155.5 1.00.1 0.02965
2.1 2.588 510.8 666.0 3. 135 620.0 0.0632 0.0249 15.82 40.2 191..l 1.2.14 0.02372
24 2.11.11 404.0 842 .1 2.514 -197.3 0.0566 0.0223 17.63 44 .8 2.1!!.6 1539 0.0 188-1
25 I 1>2.1 320.4 1.062.0 2.002 396.0 0.0505 0.0199 19.80 50.3 199.7 1.93.1 0.0 1498
26 1.280 252 .8 1.345.0 1.603 Jl6.8 0.0452 0.0178 22.12 56.2 374.2 2.-11·1 0.01 IS5
27 1.021 201 .6 1,687.6 1.3 1J 259.2 0.0409 0.0 161 24.44 62 . 1 456.9 2.947 0.00945
2X 0.8046 158.8 2.142.7 1.0515 207.J 0.0366 0.0144 27.32 69.4 57Q.6 3.1\80 0 0074 7
29 0.6470 127.7 2.660 0.854R 169.0 0,0330 0.0130 30.27 76.9 70 1.9 .1 ,527 0.00602
J() 0.5067 IOO.O J.402.2 0.67$5 I J4.5 0.0294 0.0 116 33.?J 86.2 X8·U 5.703 0.00472
JI 0..1013 79.21 4.294.6 0.5596 I 10.2 0.0267 0.0 105 37.48 95.2 1.072 (i.C) 1·1 0.00372
J2 0.3242 64.00 S.314.9 0.-1559 90.25 0.024 1 0.0095 41.45 105.3 U 16 8,488 0.00305
J.1 0.2554 50.41 6.748.6 0..161>2 72.25 0.02 16 0.0085 46.33 I 17.7 I J1Jll 10.565 0.0024 1
.14 0.201I 39.69 R.572.8 0.2S6.1 ~n.25 0.0 191 0.0075 52.48 IJJ ..1 2.09 5 11.5 12 0.00IR'l
.15 0. 1589 J 1.36 10.849 0.2268 44.l\9 0.0 170 0.0067 58.77 149.3 2"'45 17.060 0.00150
J(l 0.1266 25.00 13.608 0.18 1.1 36.00 0.0 152 0.0060 65 .62 166.7 JJ09 21,.143 0.001 19
37 n. 1026 20.25 16.80 1 0 . 1 53~ 30.25 0.0 140 0.0055 71.57 IR I .8 3.901 25, ((ii 0.000977
JX 0.08 107 16.00 21.266 0. 1207 24.01 0.0124 0.0049 80.3 5 204.1 4.971 32.062 0.000773
.l9 O.Cl6207 12.25 27,775 0.0931 18..19 0.0109 0.004.1 9 1.57 nu, 6.4J7 41 5 1X 0.000593
40 0.04K69 9.61 35.400 0.072.1 14.4-1 0.0096 0.0038 103 .6 26.1.2 R.298 5.t.522 0.000464
d I 0.0.1972 7.84 4 \.40'\ 0.0584 11.56 0.00863 0.0034 115.7 29-1 .1 10.27.1 (>(1.260 0.000379
42 0.0.116(> 6.25 :'-1.429 O.CJ.155X 9.00 0.00762 0.0030 1.11.2 333.J 13.16.l H4.901 0.000299
4.1 0.02452 4.84 70.308 0.0.168.l 7.29 0.00685 0.0027 145.8 .170.4 111.291 105.076 O.OOOB3
44 ,\ 0.0202 0 4.00 c 85.072 0 O.OJIC.5 ~- 6.25 F0.00635 C 0.0025 11 157.4 I 400.0 .I 18.957 I< 122.272 I. O.ll00195

~•-1111;;;
nol;lllon mcm1:-. 1luu 1hc entry in the colunut m1t'' be multiplied b)' IO-'.
h~, hC\C d:ua frnm R1:.· ,1 ,\ filJ!lft'lir w;,... nmnloat<H'
( 'ou11c .. ~ uf \l.tt?nclii..''.', l)f..,i.!lion of Span~ Industries
 100 H 100 H 10 H 1H 0.1 H
1,000,000 n

100,000 n

10,000 n

1000 n

100 .11

10 .n

1 ..11

1n

Figure B-1 No·mograph format.

most engineering requirements. Table B-4 lists temperatures from - 55 °C

through + 155 °C. Initial resistance (Ri) is determined at 25 °C.

DEFINITIONS

Some of the more pertinent definitions that relate to terms used in magnetic-
core-component design are listed in this section. Although some of the
terms may seem somewhat academic to the engineer, they have been includ-
ed for use by the student. Other terms are peculiar to the transformer
designer and should be of interest to readers of all levels.

203
 TA BLE B-4 Resistance temperature correction factors.

To Find T o Find
Resis1a11ce A 1: M11/1iply 8y: R esis1a11ce A 1: MulriplyBy:

-55°C 0.749 55°C 1.11 g

-50°C 0.761 60°C 1.138
- 45°C 0.772 65°C 1.157
- 40°C 0.784 10°c 1.177
-35° C 0.797 75°C 1.197
-30°C 0.810 S0°C 1.2 16
-25°C 0.822 85°C 1.236
-20°C 0.836 90°C 1.256
- 15°C 0.850 95°C 1.275
- 10°C 0.865 100°c 1.295
- 5•c 0.879 1os 0 c 1.314
0°C 0.895 110°c 1.334
5°C 0.928 11s•c 1.354
10°c 0.945 120° c 1.373
1s 0 c 0.963 12s•c 1.393
20°c 0.980 130°C 1.41 J
25°C 1.000 IJ5°C 1.432
3o•c 1.0 197 140° C 1.452
35°C 1.0393 145°C 1.471
40°C 1.590 1so•c 1.491
45°C 1.0785 1ss•c 1.511
50°C 1.0983

Ampere's Law: T his defines the relationship that exists between current
and magnetizing force. The classic equation is

0.4rrNI
}-[ = -M -
l
where H is the magnetizing force in oersteds, N the number of turns, I
the current through the turns, and M l the magnetic path length of
the core material.
Coercive Fo rce: A specific value of magnetizing force which is needed to
lower the flux density to zero (He).
Faraday's Law: T his defines the relationship of the voltage and flux. Thus

E = Ndf
--
dt
x 10 - s

In the case of sinusoidal voltages, it is expressed as

E = 2.22f1 FN x 10-s
or E = 4.44Bm-4cFN x 10 -s

where Eis the desired voltage, Bm the flux density in gausses, 1 the total +
flux capacity of the core, Ac the effective cross-sectional area of the
core, F the frequency, N the number of turns, and 1 = 2Bm x Ac. +

204
App.a Miscellaneous Data

Gauss: This is the unit of magnetic induction used in the cgs electromag-
netic system. One gauss equals I maxwell per cm 2 •
Magnetic Flux: This is the product of the magnetic induction (B) and the
cross-sectional area when B is distributed evenly and is normal to the
plane of the cross section.
Maxwell: Unit of magnetic flux used in the cgs electromagnetic system.
One maxwell = IO - 1 weber.
Oersted: Unit of magnetizing force in the cgs electromagnetic system. One
oersted = a magnetomotive force of I gilbert per cm of path length.
Permeability: Represented by the symbolµ. Broadly, this is the ratio of the
changes in magnetizing force to magnetic induction (B to H).
Residual Flux: The value of magnetic induction that remains in a magnetic
circuit when the magnetomotive force is lowered to zero.
Squareness Ratio: The ratio of the residual flux density to the maximum
flux density (saturation).
Winding Area: Circular-mil area of the hole in a core.
Winding Factor: Ratio of the total area of wire in the center hole of a tor-
oid to the window area of the toroid.

205
APPENDIX c
POT-CORE AND
BOBBIN DATA

This section contains essential information for use by those who engage
in the design and application of pot-core inductors and transformers. In this
appendix we can find dimensional and tolerance data for a popular group
of standard pot cores. The specifications are in accordance with the interna-
tional standard, !EC Publication 133, entitled "Dimensions for Pot Cores
Made of Ferromagnetic Oxides." The tables have been provided through
the courtesy of Indiana General Corp.
Table C-1 contains the physical dimensions for standard pot cores.
Figure C-1 provides the physical outlines for the pot cores.

t
H1 H i
t t
A A
(2 p laces)
~~~~+_
i t f
- - j f-s -I f- s
(2 places) (4 places)

Core with 2 slots Core with 4 slots

Figure C-1 Dimensional outlines for standard pot cores.

207
 o, o, DJ D, H, 11, B,1 .u .011 IJu .°'1or, ,1 R
SI ZE TOL.
(mm) mm in mm in mm in mm in mm mm mm m ,,,,,, m ,,,,,, mm in
In
"' "'
9x 5
MIN . 9.0 .3543 .750 .2953 3.7 .1457 2.0 .0787 S.J . 2008 3.6 . 1417 1.6 .063 - - 6.0 .236 - -
MAX , 9 .3 .366 1 7 , 75 . JOSI 3.9 . 1535 2.2 .0866 S. 4 . 2126 J.9 . 1535 2 ..a .O'J.a - - 7. 2 . 283 0.25 .0 10

11 x 7
MIN. 10.9 .429 9.0 .354 4.5 . 1772 2.0 .0787 6.3 .248 4.4 .173 1.6 063 - - () . 5 . 256 - -
MAX. 11.J .445 9,4 .370 4.7 .1 850 2.2 .0866 6.6 . 260 4. 7 . 185 2.6 . 102 - - M.O . Jll 0. 25 .OIO

14 x 8
MJN. 13.8 .54] 11.6 . 4567 s .8 3.0 .118 8.2 .3228 5.6 .2205 2.0 .079 1.6 .063 8.7 ,J 4J - -
MAX. 14.3 .563 12.0 A724 6.0 .2362 J.2 .126 8.S . JJ46 6.0 .2362 4. 1 . 161 2.0 ,079 1(),4 ,409 0. 25 .010

18 • II
MIN. 17.6 .693 14.9 .587 7.3 .2874 3.0 . 118 10.4 .4094 7.2 . 2835 2.0 .079 2.0 .079 11.J .445 - -
MAX. 18.4 .724 IS.4 .606 7.6 .2992 J. 2 . 126 10.7 .4213 7.6 . 2992 4. 4 . 173 3.0 . 118 14.Q .55 1 0. 25 .0 10
N
0 MIN. 21.2 .835 17.9 .705 4A
CX> 9.1 3583 . 173 13.2 .5 197 9.2 .362 2.5 098 2.5 098 IJ. J .524
22 x 13
MAX. 22.0 .866 18.S .728 9.4 .3701 4.7 . ISS 13.6 .5354 9.6 .378 4 . .J . 173 3. S .138 16.S .650 0.35 .0 14

.~tl N . 2S.O .984 21.2 .835 I I.I .437 5.4 . 2 126 15.9 .626 11.0 .433 2.5 .098 2.5 .098 17.0 .669 - -
26 x 16
MAX. 26.0 1.024 22.0 .866 11.S .453 5.7 .224 16. J .642 11.4 .449 4.4 . 17) J.s . 138 20.0 .787 O. JS .014

Ja x 19
MIN. 29.5 1. 16 1 25.0 .984 13 .1 .5 157 5.4 .2126 18.6 . 732 13.0 .S J 18 3.0 . 118 20 .0 .787 - -
MAX. 30.S 1.201 25.8 1.016 13.5 .SJl5 S.7 .2244 19 .0 .748 13.4 .5276 5.3 . 209 4.0 . I 57 23 .0 .9()6 O.Js .0 14

36 . 22
MIN.
MAX. 36.3
35.0 1.378 29.9
1.425 30.9
1.177
1.217
I S.6
16.2
.61 4
.638
5.4
5.7
.2126 21.•
.2244 22.0
.843
.866
14.6
15.0
.5748
.5906
3.s
5.6
. 138
.220
3. 5
4.S
.138
. 177
24 .0
27.2
.945
1.071
- -
0.35 .0 14

42 X 20
MIN . 41.1 1.642 35.6 1.402 17. 1 .673 S.4 .2126 29. J 1.154 20.J . 799 4.0
-
.160 - - - - - -
MAX. 43 . 1 1.697 37.0 1.457 17.7 .697 5.7 .2244 29.9 1. 177 20.7 .8 15 - - - - - 0.40 .016

TABLE C-1 Dimensional outlines for standard pot cores.

 TABLE C-2 Outline dimensions for wound bobbins.

D2 D3 1-12 R1
Si=t,
( 111111 ) Tnlera11ce 111111 in. 111111 in. 111111 in. 111111 ill.

9 x 5 Min. 4.0 0. 157 0.25 0.010

Max. 7.4 0.29 1 3.6 0. 142
II x 7 Min. 4.8 0. 189 0 25 0.010
Max. 8.9 0.350 4.4 0. 173
14 x 8 Min. 6. 1 0.240 0.25 0.0 10
Max. 11.5 0.453 5.6 0.220
18 x 11 Min. 7.7 0.303 0.25 0.0 10
Max. 14.8 0.583 7.2 0.283
22 x 13 Min . 9.5 0.374 0.35 0.014
M;1x. 17.8 0.701 9.2 0.362
26 x 16 Min. 11 .6 0.457 0..J5 0.0 14
Max. 21.1 0.831 11.0 0.433
30 x 19 Min . 13.6 0.535 0.35 0.0 14
M ax. 24.9 0.980 13.0 0.5 12
36 x 22 Min. 16.3 0.642 0.35 0.0 14
Ma x. 29.8 1.1 73 14.6 0.575
42 x 29 Min. 17.8 0.701 0.4 0.0 16
Max. 35.5 1.398 20.3 0.799

R 1 ( typ.)

rF~1~I'i 1-0-1'
Ll -::Y-J
I '

Figure C-2 Outline for a wou nd

bobbin.

Table C-3 contains the dimensions of bobbins that will fit within the
standard pot cores listed in Table C-1. Printed-circuit versions of some of
these bobbins are presented in Table C-4. These dimensions are referenced
to the outlines given in Fig. C-3.
See Table C-3 for complete dimensional information concerning this list
of bobbins . Dimensional outlines are provided in Fig. C-4.

209
 Toblo c.3 Bobbin Dlmonaions.

f::.XAM l'Ll· orSrANUARO

t•lo. A, ,•l z ;I J II o, Di DJ H
s;:., of
l mmJ Sect. Toft,,.ance mml itt, l mm : ;,,: mm: flt! mm hr. ;,,, mm ,.,,, mm in. mm ;,,,

9 x 5 l Min. J. 17 0.00l92 1.6 0.060 7.23 0.285 4.67 O.IR4 4.0 1 0.158 3AO 0. 134

II x 7 I
Max.
Min
-
4.78
-
0.00742
- 0.()(.0
-
I.(,
7..14
8.69
0.289
0.342
4,78
5.59
0.188
0.220
4. 11
4.81
0. 162
0. 189
3.50
4.09
0.138
0. 161
M a•.
Min.
- -
2.16 0 .0 03H
- - 8.89 0.350 5.69 0.224 4 .9 1 0. 193 4,19 0. 165

Ma.,.
Min.
- - I 26 000195

x 8
Ma'-. - -
14 I Min. 8.81 0.0136 1.6 0 ,()(.0 11.J 0 J4J 6.98 0.275 5.117 0.215 5.28 0 .208
Ma·c
Min .
- - 392 000608
- - 11.5 0.J5-J 1.i .i 0.285 6.10 0.240 5.49 0 .216

~t ax.
Min
- - 2.35 0 .00365

18 y 11 I
~l ;,t\ .
,\ ·tin. 17.1 0.0265
- - 1.8 0.070 14.6 0 .574 8.59 0.338 7.70 0 .303 6.88 0 .271
~ht\ ,
Min.
- - 7 6 1 0 01 18
- - 14 .8 0.584 8.84 O.J48 7.82 0.308 7.09 0.279

i\1 a,,
M in.
- - ·1.66 0 00722
N
..... ~ t ax. - -
0 22 X IJ I Min. 26.2 0.().106 I 8 0.070 17.6 0.69.1 10..l 0..107 9.50 0.374 8.89 0.350
Max.
Min.
- - 12.5 0.019·•
- - 17,8 0.702 10.6 0..11 7 9.75 0 .384 9.09 0.358
Ma,,
~-1 in.
- - 7.S7 o.om
x 16
Max. - -
26 I ~1in . JU 0.0582 1.8 0,070 20.9 0.82• 12.-1 0 .489 11.6 0.451 10.7 0.4 21
Max.
~li n .
- -
17.J 0 .0269
- - 21 I 0 .832 12 .7 0 ..199 11.7 0..162 10.9 0.429

~'1 3\.
Min .
- - 10.S 0 .016~

JO x 19 I
\1a-.:.
M in . 53.7 0.0834
- - I 8 0 .070 24.7 0.971 14.6 0.575 IJ.(. O.SJS 12.7 0 .500
Ma1i .
Min.
- -
25. 1 OOJR9
- - 24.9 0.980 14.9 0.585 13.7 0 .5~0 12.9 0 .508

M 3'(.
Min.
- - 15 9 00246

36 x 22 I
Ma\.
Min. 71.3 0. 110
- -
2.S 0 .110 19.5 1. 160 17.9 0.105 164 0 645 14.2 0 .560
Ma,.
M in.
- -
JI 9 0().194
- - 29.8 1.172 18.2 0.715 16,6 0 .653 14.4 0.568

Ma,_
M in.
- -
20 0 0.0310

42 x 29 I
Ma~ .
Min. 136 0.211
- - 2.8 0. 110 35.1 1.)86 19.5 0 .768 18.0 0 .709 19.6 0.772
~lax. - - - - 35.4 1.394 19 7 0776 18 2 0 71 7 19 8 0780
Min. 55.6 00862
~l ax.
 Figure C-3 Dimensional outlines for pot-core bobbins which can be used with the pot
cores of Table C-1.

T A BLE C-4 Printed-circuit-style bobbins.

p s, S2
Si:i'
( 111111 I T o/('mllf<! 11111/ in. II/Ill in. 111111 in. Fig11re

14 x ll Min. 4.62 0.182 3.45 0.136 16.13 0.635 C-4C. D

Max. 4.88 0.192 3.66 0.144 16.38 0.645
18 x 11 Min. 4.57 0. 180 3.43 0. 135 21 .2 1 0.835 C-4A. B
Max . 4.95 0.195 3.68 0. 145 21 .72 0.855
22 x 13 Min. 4.44 0.175 3.43 0. 135 24.54 0.966 C-4A. B
Ma x. 4.95 0.195 3.68 0.145 25.42 1.00 1
26 x 16 M i11. 4.62 0.182 3.45 0.136 2lUJ7 1.105 C-4A , R
Max. 4.82 0.1 92 3.71 0. 146 28.88 1.137

- s,--

A
w inding
6~ ref.
0
area
wind ing
area
(a) (b) ( c) (d)

Figura C-4 Dimensional outlines tor printed-circuit types of bobbins. These outlines are
referenced to Table C-4.

2 11
 CONVERSION INFORMATION
FOR WIRE GAUGES 10 TO 44

The method used here is recommended for the calculation of the dimen-
sional parameters of pot cores and is in accordance with !EC Publication
205, "Calculation of Effective Parameters of Magnetic Piece Parts."
For this method of calculating the dimensional parameters of pot cores,
the pot-core set is substituted by an ideal toroidal core such that a coil
wound on that toroid would give exactly the same electrical performance as
a coil with same number of turns placed on the pot core set.
The dimensional parameters of that substitute toroid are called "effec-
tive" parameters. These are indicated by the suffix "e" added to the sym-
bol.
Magnetic path length
Cross-sectional area
Core volume
For the purpose of the calculation of the dimensional parameters, the closed
magnetic circuit of a pot-core set is divided into five sections. For each sec-
tion the area, flux path length and the core constants C, and C2 are deter-
mined (Fig. C-5).

c, _}_ mm- 1 and Ci ==- .!_ mm- 3

A A2

/'2

Section A - A
Figure C·5 Pot-core set divided into five sections.

212
 Pot-Core and Bobbin Data App.C

The core constants for the total magnetic circuit of the pot-core set are

C, = L_{_mm- 1 and C, ="' _!_ mm -

"- A1
3

From these core constants the effective dimensional pot-core parameters

can be calculated.
c,1
Magnetic path length l =-mm
e C2

Cross-sectional area A = C, mm 2
e C2
c1
Core volume V = l
e
A = -'
e''e C/ mm 3

For each of the five sections of the magnetic circuit of a pot-core set, the
magnetic path length and cross-sectional area has to be determined:

Area of centerpost,
The condition to obtain A ' 1 = A "1 is
Area of outer ring,
The condition to obtain A ', = A ", is
Cross-sectional area of centerpost,
Cross-sectional area for outer ring,
I .!_ for two plates,
_for two plates,
A A1

Mean flux path length at corners,

Cross-sectional area associated with I.,
Mean flux path length at corners,
Cross-sectional area associated with Is.

The calculations above ignore the effects of wire slots, which can be
taken into account by the following corrections:

From A, subtract:
Multiply _12_ by
A1
Multiply·_!:_____ by
Ai 2
Multiply A. by

213
214 Cylindrical-Core and Coif Data App.D

S1 = r2 -y r2 2 + r1 1
2
A1 = A '1 + A"1

52 = V r1 1 + r.' - r 1
2

/2 1 ) T3 0. 733 T3 - I
- = - '1- =---log-mm
A2 nh r2 h r2

/, = /', + /", = ~ (h + 2S2) mm

4

A , = ~ (r, 2 - r/ + 2r1h) mm 2
2

1, = !', + C = (h + 251) mm

A i = ~ (r2 2 - r12 + 2r2h) mm 2

2
ng (r, - rl) n = number of wire slots, g = slot width

ng
2rrr3

- ng z
2rrr1

ng
1 - - -- -
APPENDIX D
CYLINDRICAL-CORE
AND COIL DATA

The information in this appendix should be useful to those who design

circuits that contain fixed -value and adjustable magnetic-core inductors.
The tables offer pertinent design data and list many of the standard induc-
tance values that are available on the commercial market. Outlines of
various coil-form formats have been included to illustrate the universality
of application for slug-tuned inductors.
A chart that shows recommended operating frequency versus core mixes
which are available from Micrometals Corporation is shown in Fig. D-1.

1 kHz 10 100 kH z 1 M Hz 10 100 200

Mix l l l I I I
0 -
22
12 ,q
10
6 :0
7 "'
"'
E
2 Q;
a.
1 en
c
15 ·v;
42 "'eu
8 _c:
3
41
I I
' I I
'
1 kHz 10 100 1 MHz 10 100 200
Figure D·1 Chart that shows various powdered·iron core mixes versus optimum
operating frequency and a.

215
The cores are made from powdered iron and are available in a wide range of
permeabilities.
It is important to understand that the µe of a cylindrical core will be dif-
ferent from that of the same material in toroidal form. The actual µe is
dependent upon the LID (length-to-diameter ratio). To illustrate this effect,
the curves of Fig. D-2 are included in this section. The graph is founded on
a single layer winding (close-wound) occupying 950Jo of the cylindrical rod.
Table D-1 contains a listing of standard solid cylindrical cores that are
suitable for use in slug-tuned inductors and fixed-value RF chokes. The
cores are available in a wide assortment of powdered-iron mixes.
Hollow cylindrical slugs are available in a large range of powdered-iron
core materials. A listing of various standard sizes is given in Table D-2.

400
300

200

100
80
60
50 LID
40
30

> 20

10
8
6
5
4
rCJ
3

1 ~~~---'~~~---'~~~---'~~~---''--~~~'--~~---'
0 10 20 30 40 50 60
µ. 11 (effective permeabi lity)
Figure 0·2 Family of curves that demonstrate the relationship between effective
permeability of rod material and the LID ratio. (Courtesy of Micrometals Corp.)

216
 TABLE D-1 Standard solid cylindrical cores.

OD Length Parr
(in./mm ) (in./111111) Number

0.030/ 0.76 0.155/3.94 p 25-106

0.041 / 1.04 0.100/2.54 p 33-106
0.049/1.24 0.092/ 2.34 p 33-900
0.060/ 1.52 0.375/9.53 p 412-106
0.062/ 1.57 0.500/ 12.7 p 416-112

0.093 / 2.36 0.250/ 6.35 p 68-106

0.103 / 2.62 0.31 2/7.92 p 710-101
0.120/ 3.05 0.312/ 7.92 p 810-106
0.142/ 3.61 0.250/ 6.35 p 912-112
0.152/ 3.86 0.375 / 9.53 p 1012-102

0.177 / 4.50 0.625 / 15.8 p 1120-207

0.185 / 4.70 0.456/ 11.6 p 1214-141
0.195/ 4.95 0.187 / 4.75 p 136-2 07
0 245 / 6.22 0.250/ 6.35 p 168-102
0.304/ 7 .72 1.000/ 25.4 p 2032- 141

0.365 / 9.27 0.500/ 12.7 p 2416- 103

0.370/ 9.40 0.750/ 19.0 p 2424-141
0.495/ 12.6 0.500/ 12.7 p 3216-102
0.620/ 15.7 1.478/ 37.5 p 4047-141
0.745 / 18.9 1.270/ 32.3 p 4840-102
0.995/ 25.2 1.000/ 25.4 p 6432-102

Co urtesy of Micrornetals Corp.

TABLE 0-2 Standard hollow cylindrical slugs.

OD ID Length Part
( in. / 111111) (in. / 111111) (in./111111) Number

0.139/ 3.53 0.062/ 1.57 0.125 / 3.12 H 22-1006

0.257 / 6.53 0.075/ 1.91 0.375 / 9.52 H 46-1006
0.333/ 8.46 0.093/ 2.36 0.364 / 9.25 H 66-1107
0.362 / 9.19 0.125/ 3.17 0.625/ 15.9 H 610-1107
0.370/ 9.40 0.125/ 3.17 0.500/ 12.7 H 68-1006

0.370/ 9.40 0.205/ 5.21 0.500/ 12.7 H 68-1242

0.365/ 9.27 0.201 / 5.10 1.000/ 25.4 H 616-1202
0.370/ 9.40 0.137 / 3.48 1.000/ 25.4 H 616-1 301
0.480/ 12.2 0.073 / 1.85 .500/ 12.7 H 88-1007
0.495 / 12.6 0.200/ 5.08 1.125 / 28.6 H 818-1003

Courtesy of Micrometals Corp .

Powdered-iron cylinders are also available with wire end leads em-
bedded into them. A limited number of mixes are provided for this style of
core material. A listing of the sizes is given in Table D-3. Figure D-3 shows
the outline details .

217
 +0.004
- 0.000
Dia.
_L_ r l -=---1-1~ in.- -1
~ 9
~
8 Plain IP) 8 Notched (NJ

Flgure 0·3 Profile and end views of slug forms w ith w ire
leads. Dimensional data are given in T able D-3.

T AB LE D-3 Standard powdered-iron cylinders.

OD Length Wire E11dD Part

(i11./111111 / (i11. / 111111 / Gauge (.ree Fig. D-3) l\111111ber

0.060/ 1.52 0.187 / 4.75 24 p F 46- 20

0.060/ 1.52 0.250/ 6.35 24 r F 48-20
0.060/ 1.52 0.312/ 7.92 24 r F 10- 20
0.069 / 1.75 0203 / 5. 16 22 p F 57- 29
0.076 / 1.93 0.1 87 / 4.75 22 r F 56- 20

0076 / 1.93 0.250/ 6.35 22 p F 58- 63

0.076/ 1.93 0.3 12/ 7.92 22 p F 510-23
0.076/ 1.93 0.375 / 9 53 22 p F 512-61
0.093/ 2.36 0.250 /6. 35 22 N F 68- 21
0.093/ 2.36 0.3 I 2/ 7.92 22 N F 610- 12

0.093 / 2.36 0.375/9.53 22 N F 612-2 1

0.103/2 .62 0.250/6.35 22 N F 78-20
0.103/2.62 0.3 12/7 .92 22 N F 710· 14
0.103 / 2.62 0.375/9.SJ 22 N F 7 11- 22
0 .1 03 / 2.62 0.437 I 11. I 22 N F 712- 24

0 .1 21/3 .07 O.J 12/ 7.92 22 N F 8!0-20

0. 121 / 3.07 0.346/8.79 22 N F 811 - 70
0. 121 /3 .07 0.375 / 9.53 22 N F 812- 13
0 .1 21/ 3.07 0.437/ 11.1 22 N F 814-20
0.121 / 3.07 0.500/ 1.27 22 N F 816- 20

0.152/ 3.86 0.500/ 12.7 21 N F 10 16-62

0.152/ 3.86 0.750/ 19.0 21 N F 1024-40
0. 186/ 4.72 0.500/ 12.7 21 N F 121 6- 10
0.186/ 4.72 0.750/ 19.0 20 N F 1224-20
0. 19 1/ 4.85 0.375 / 9.53 22 N F 1312-61

0. 196/ 4.98 0.625 / 15.8 21 N F 1320- 20

0.216/ 5.49 0.750/ 19.0 20 N F 1424-60
0.246/ 6.25 0.625 / 15.8 20 N F 1620-10
0.246 / 6.25 0.875 / 22.2 20 p F 1628- 70
0.266 / 6 76 0.875 / 22.2 20 N F 1728-20

0.280/ 7. I I 0.875 / 22.2 20 N F 1828-20

0.370/ 9.40 0.625 / 15.8 20 p F 2420-10
0.370 / 9.40 1.250/ J I 7 16 p F 2440-11
0.3 70/ 9.40 1.250/ 31.7 20 p F 2440-1 2
0 A95/ 12.6 0.500 / 12.7 22 N F 3216-01

Co urtesy of M icromc1al~ Corp.

21 8
 COIL FORMS

A wide variety of insulated coil forms are available on the U.S. market. The
exact format chosen will depend for the most part on the amount of induc-
tance needed, plus the physical space available in the manufactured com-
posite equipment. A number of forms are available for snap-in mounting,
while others are configured primarily for printed-circuit-board applica-
tions. Figure D-4 contains outlines for a standard group of coil-form
shapes, most of which are intended for use as adjustable inductors. Other
· types of slug-equipped coil forms mount by means of a threaded collet and
machine nuts. The slug shafts are also threaded, and are contained in the
center holes of the mounting collets.

(/ CA cc CD CE

CF

Figure 0-4 Various standard coil forms for use with magnetic cores. {Courtesy of J .W. Miller
Co., Division of Bell Industries.)

219
 co CR cs CT

cu CV cw ex CY

CZ DA DB DC DD DE OF
Figure D-4 (continued)

220
 DH 01 DJ DK

DL DM DN DO OP

DO DR OS OT DU

DV
Figure 0-4 (continued)

22 1
 ?REWOUND SLUG INDUCTORS

Standard values of variable inductance are available commercially. Printed-

circuit-mount coils are sold in the United States, along with slug-tuned in-
ductors that mount by means of a collet and nut. Most of these
prefabricated units contain powdered-iron movable cores. Ferrite slugs are
used in some of the assemblies to obtain large amounts of inductance (high-
permeability cores) with a relatively small coil form. Table D-4 lists some of
the standard inductance ranges that are offered in printed-circuit styles of
coil forms. lnformation is included in the table to indicate the Qu of each
coil at some specific test frequency. The de resistance of the coil windings is
also provided. This group of coils uses polyester-impregnated alpha-cellu-
lose tubing as the insulating material. Coil-form dimensions are Ys in. (9.52
mm) in diameter by % in. (19 mm) in length. Each coil has three printed-
circuit-insert mounting feet.
Tables D-5, D-6, and D-7 contain listings of three types of slug-tuned in-
ductors that mount by means of collets and nuts. Three physical sizes are
represented in standard inductance ranges: subminiature, miniature, and
standard. Blank forms for these and the printed-circuit coils of Table D-4
are available commercially. The coils of Tables D-5 through D-7 are wound
on slug forms which are made of silicone-impregnated ceramic.

FIXED - VALUE /RON-CORE INDUCTORS

Subminiature inductors are available with powdered-iron cores and axial

leads. A wide range of standard RF choke values is presented in Table D-8
for use by the designer. These coils are ideally suited to filter and network
applications. They are suitable for use in delay lines and for computer ap-
plications. Coils of this type are available in the United States with fungus-
proof varnish or with epoxy-resin encapsulation.

222
 TABLE 0-4 Standard values of inductance.

MINIMUMCORE POSITION MAXIMUM COi\ EPOSITION

Max. :\.fin.
Minimum Q Trst :\.fnximum Q Test Max. fl C11rre111 Fo
lmlr1 (' IOIJC{' ."'fill. Fn•qm•m·1· lndtu:tan((' Min. Frequem'.1' I UJ MH: I ! Mrl!

0.095 µ l·I 77 25.0 MHz 0. 125 µH 94 25 .0 MHz 0.02 4100 350.

0. 130 µH 68 25.0 MHz 0.170 µH 92 25.0 MHz 0,02 1600 JOO.
0.185 µl·I 88 25.0 MHz 0.265 µH 100 25.0 MHz 0.02 1600 230.
0.285 11H 88 25.0 1\.1H z 0.410 µH 93 25 .0 MHz 003 1000 198.
0.420 ,,H 100 25.0 MHz 0.580 ,, H 80 25 .0 MH z O.oJ 2500 ISO.
0.540 ~, M 101 25.0 MHz 0.850 µH 89 25.0 MHz O.oJ 1600 136.
0.640 11H 10 1 25.0 MHz 1.00 i•H 78 25 .0 MHz 0.0.l 1600 11 8.
0.760 i1H 98 25.0 Ml-lz 1.25 11H 70 7.9 Ml-lz 0.04 1600 11 4.
J.20 µ 1-1 65 7.9 MH1. 1.87 µ f-1 70 7.9 MHz 0.06 1000 89.
1.65 µ ,., 61 7.9 MHz 2.75 11 1-f 65 7.9 MHz 0.1 4 400 77.
2.40 µH 64 7.9 MHz 4.10 µH 60 7.9 MHz 0. 17 400 62
3.40 ,,H 68 7.9 MHz 5.80 µH 60 7.9 MHz 0.24 400 53 .
4.60 µH 64 7.9 MHz 8.50 µH 56 7.9 MH z 0 ..19 250 45
5.60 µH 64 7.9 MHz 10.0 µH 57 2.5 MHz 0.64 160 40.
7.10 ,,fl 68 7.9 MHz 12.5 "l·I 55 2.5 MHz 0.77 160 .lS .
10.0 µH 58 2.5 MHz 18.7 µH 95 2.5 MHz 1.68 JOO I J.7
14.8 µH 61 25 Ml·lz 27.5 µH 90 2.5 MHz 1.9: 100 8.4
22.0 ,,H 60 2.5 MHz 41.0 µH 75 2.5 Milz 2..\4 100 6.7
N
N
31.0 µH 58 2.5 MHz ss.o µH 68 2.5 MHz 2.72 100 5.6
43.5 µH 56 2.5 MHz 85.0 µH 55 2.5 MHz 3.30 100 4.6
w 4.3
61.0 i1 H 4R 2.5 MHz 100.0 µH SR 790. k Hz 3.89 100
76.0 µH 52 2.5 MHz 15.0 µH 90 790. k Hz 4.39 100 3.8
105.0 _t( H 57 790. kHz 187.0 µH 92 790. kHz 5.46 100 J.3
160.0 µH 63 790. kH1. 275.0 111-1 90 790. kMz 6.70 100 2.9
240.0 µ l-1 66 790. kHz 4 10.0 11 H 90 790. k Hz 8.JO 100 2 .5
360.0 11 H 6R 790. kHz 580.0 µH 81 790. kHz 10.50 100 2. 1
530.0 11 H 66 790. kHz 850.0 µH 75 790. kHz 12.90 100 1.75
700.0 ,, H 64 790. kHz 1.00 inH 80 250. k Hz 14.90 100 J.70
910.0 ,,H 66 790. kHz 1.25 mH 85 250. kHz 17. 10 100 1.61
990.0 µH 35 250. kHz 1.87 mh 60 250. k Hz 28 .20 65 0.73
1.60 mH 39 250. kH z 2.75 mH 62 250. kHz 34.80 65 0.62
2.40 mH 41 250. kHz 4.IO ml-I 60 250. kH z 42.90 65 0.60
.l.40 mll 42 250. kHz 5.80 ml-I 57 250. k Hz 51.60 65 0.53
5.15 mH 42 250. k Hz 8.50 mH 50 250. kHz 63 .60 65 0.50
7.40 mH 42 250. kHz 10.00 mH 50 79. kHz 75.60 65 0.40
9.80 mH 40 250. k Hz 12.SO mH 52 79. kHz 87..10 65 0.38
12.00 mH 39 79, kH1 18.70 mH 55 79, kH7. 111.0 65 0.32
12. 10 mH 20 7?. kHz 27.50 mH 51 79. kH z 197.0 33 0.26
18.20 mH 24 79. kHz 41.00 mH 54 79. kMz 244.0 33 0.2 1
27.50 mH 28 79. kHz 58.00 mH 56 79. kHz 302.0 33 0.20
40.00 mH 32 79. kHz 85.00 rnH 56 79. kHz 378.0 33 0.16
50.00 111H 34 79. kHz 100.00 mH 58 79. kHz 423.0 33 0.15
62.00 mH 35 79. kHz 125.00 mH 56 79. kHz 468 .0 33 0.14

courtesy of J. W . Mill~ r Cu .. Divi:s.io r1 of Bt.:11 lrldus1rics.

 TABLE 0-5 Subminiature adjustable inductors.

MIN IMUM CORE POSITION MINIMUM CORE POSITION

Max. 1'vfin.
Mini11111111 Q Test Maximum Q Test Max.R C11rre111 F
!nd11crance Min. Frequency lnducrance Min. Frequency ( fl ) ( MA J ?t.1
( 1'vl

0.190 H 64 25. MHz 0.250 H 53 25. MHz 0.014 1600 288.

0.300 H 57 25. MHz 0.390 H 44 25. MHz 0.017 1600 212.
0.440 H 61 25. MHz 0.620 H 44 25. MHz O.G31 1000 184.
N
0.770 H 62 25. MHz 0.900 H 41 25. MHz 0.054 636 148.
N 1.00 H 40 7.9 MHz 1.40 H 37 7.9 MHz 0.15 256 114.
~
1.60 H 40 7.9 MHz 2.40 H 36 7.9 MHz 0.25 202 92.
2.70 H 40 7.9 MHz 4.20 H 36 7.9 MHz 0.62 100 69.
4.70 H 42 7.9 MHz 6.80 H 33 7.9 MHz 0.9 1 100 60.
7.80 H 30 7.9 MHz 11.0 H 42 2.5 MHz 1.4 100 18.
14.0 H 34 2.5 MHz 19.0 H 37 2.5 MHz 1.9 100 13.
22.0 H JJ 2.5 MHz 31.0 H 38 2.5 MHz 2.2 100 12.
36.0 H 32 2.5 MHz 49.0 H 33 2.5 MHz 3. 1 100 9.6
56.0 H 23 2.5 l'vlHz 97.0 H 24 2.5 MHz 5.6 64 7.3
111 H 26 0.79 MHz 171. H 26 0.79 Ml-17.. 7.9 64 6.0
196 H 25 0.79 MHz 285. H 24 0.79 MHz I I. 64 4.4

Courtesy of J. W . Miller Co .. D ivision of Bell Industries.

 TABLE D· 6 Miniature adjustable inductors.

MI NIM UM CORF POSITION MAX IM UM CORI! l'OSITION

Ma.\ . M in.
Alinimum Q fr.fl A1aximum Q T1'.<I M a-'.R Currell/ F
/11tl11rtam·,~ M in. Fr1•q111•nn· lmluc1t111rt 1
M iu. Frequency (Ill / /lf11J /M~:I
0.440 II 80 25. MH1 0.760 II 52 25. M~(z O.oJ 1600 142.
1. 10 H 72 25. MHz 1.50 H 40 7.9 MH z 0.()(1 1000 96.
1.70 H 51 7.9 MH z 2.70 H 36 7.9 MH z 0.11 636 80.
3.1 0 H 56 7,9 MH1 4.80 H 33 7.9 MH z 0.23 400 58.
5.50 H 60 7.9 MM1 8.60 II J3 7.9 MH z 0.49 256 45.
9.90 H 52 7,'l MHr. 15.0 H 41 25 MHz 1.5 JOO 32.
17.0 H 47 2.5 MH 1. 23.0 H 53 2.5 MH z 2J 100 19.
26.0 II 48 2.5 MH 7 JJ.O H 51 2.5 MH z 2.9 JOO 16.
31>.0 H .so 2.5 MHt 57.0 II 48 2.5 MH z 3.4 IOU 12.
66.0 II 44 2.5 /\! Ht 114. II 40 0. 79 MH z 4.1 100 5.2
120. H 46 0.7•) /\! Hz 190. II 40 0.79 MH1. 5.7 100 4.1
209. H 45 0.79 Miit 314. II 32 0.79 MHL 7.7 100 3.2
350. H ~u 0.79 1'11it 475. II 41 0.79 MHz 10. 100 ~.9
528. H 44 0.79 Miit 760. II 40 0.79 MHz 14. 100 2.5
N
N Counc~y of J W Miller <:o .. 01vi5ion of Bell lndu5tric:~.
(11

TABLE D-7 Standard adjustable inductors.

M1 NIM1JM CORI· l'os1T10:-i MA XIMl >M CORI· POSl'l'ION

,\/(}\ , Mi11.
,\fi11im11m Q Tr.<t jHa.rimum Q r ....., Ma.1. R Curren/ F1
111d11c1011rr M i11. Freq11e11n /11d11cta11cr M in Frrq11e11c.1· 1 11 1 I \IA I (,\ /ff: /

0.990 H 68 25. ~IHt 1.50 H 51 7.9 MHz 0.04 1600 75.

1.60 H 67 7.9 ~IH1 3.10 H 52 7.9 MHz 0.08 1000 51.
3.30 H 70 79 MH1 6.50 H 44 7.9 MHz 0.17 636 41 .
7.30 H 50 7.9 /<!Ht 14.0 H 35 2.5 MHz 0.57 256 26.
16.0 H 52 2.5 MH1 29.0 H 56 2.5 MHz 2.2 100 7.6
13.0 H SI 2.5 l\I H1 66.0 H 37 2.5 MHz 3.1 100 4.9
74.0 H 42 2.5 M il l 124 II 44 0.79 MHz 4.5 100 3.7
138 H 47 0.79 ~ I !It 238 H 44 0.79 MHz 6.6 100 3.0
270 H 57 0.79 MHt 451 H 43 0.79 MHz 8.9 100 2.1
495 H 5(> 0.79 MH1 760 II 41 0.79 f\'1H z 12. 100 2.0
825 H 41 0.79 MHz 1.30 mH 27 0.25 MH z 20. 100 1.7
1.40 mH 40 0.25 MH 2.00 mH 33 0.25 MHz 25. 100 1.6

Courtesy of J. \\1 Miiier Co .. Division of Bcll lndu~uics .

 TAB LE D-8 Iron -core inductors.

1\.li11imum A1aximum Form

Resonant :Haximum Af aximum Winding Length
Nominal ,\.linimum Tes/ Frequency DC mA Diameter + 1/ 32
illd11c1<111ce Q Frequency f M /1: 1 R t'.~istanre Ra1i11g fin. I f in./

0. 10 µH 49 25 M Hz 600 0.013 3922 0.1 56 5/ 16

0. 15 µH 52 25 M Hz 490 0.025 2828 0. 141 5/ 16
0.22 µH 48 25 M Hz 400 0.038 2294 0.141 5/ 16
0.33 µH 47 25 MHz 330 0.070 1690 0 .1 25 5/ 16
0.47 µ I-I 46 25 MHz 280 0. 125 1264 0 .1 25 5/ 16

068 µI-I 48 25 MH1. 240 0 .200 1000 0. 125 5/ 16

0.75 µH 48 25 M Hz 224 0.264 870 0.125 5/ 16
082 µH 48 25 MH z 216 0.290 830 0.125 5/ 16
1.0 11 H 41 25 MH z 118 0.048 2041 0. 165 li4
1.2 µII 45 7.9 MH1. 11 8 0.072 1666 0. 160 1/ 4

1.5 µH 42 7.9 MH1. 102 0.096 1443 0.160 1/ 4

1.8 µH 31 7.9 MH z 89 0.096 1443 0. 160 I/ 4
2.2 µH 43 7.9 M Hz 87 0. 156 11 32 0.160 1/ 4
2.7 µH 34 7.9 MH1. 74 0.168 1091 0.160 1/ 4
3.3 µH 40 7.9 MH ?. 66 0.240 912 0.1 50 1/ 4

3.9 1di 35 7.9 M Hz 61 0.264 870 0.150 I/ 4

4.7 µH 43 7.9 MH z 5.J 0.457 661 0.150 I/ 4
5.6 11 H 41 7.9 MHz 49 0.'192 637 0.1 50 I/ 4
6.8 µIi 40 7.9 M Hz 49 0.624 566 0.150 I/ 4
7.5 ,,H 32 7.9 M Hz 44 0.624 566 0.150 I/ 4

8.2 ,,H 37 7.9 Miiz 41 0.744 518 0.150 I/ 4

9.1 ,,H 41 7.9 MH z 21 1.44 288 0.1 60 I/ 4
10 µH 36 7.9 MJ-l z 19 1.56 277 0.160 1/ 4
12 1•H 52 2.5 MH z 19 1.68 267 0.160 I/ 4
15 µH 52 2.5 MHz 16 1.92 250 0.165 I/ 4
18 µ I-I 52 2.5 MHz 15 2.28 229 0.165 1/ 4
22 µH 51 2.5 MH z IJ 2.28 229 0.165 1/ 4
25 µI-I 48 2.5 MH1. 13 2.64 213 O.liO 1/ 4
27 µH 49 2.5 MHz 12 2.64 213 0.170 1/ 4
33 µH 50 2.5 MHz 10 2.76 208 0.1 70 1/ 4
39 ,,H 48 2.5 MH z 9.3 3.36 188 0.175 1/ 4
47 µH 44 2.5 MH1 9. 1 3.36 188 0.1 75 1/ 4
56 uH 45 2.5 MHz 8.6 3.84 176 0.1 80 I/ 4
68 µH 42 2.5 MH z 8.1 4.20 169 0. 180 1/ 4
75 µH 38 2.5 MH z 7.2 4.56 162 0.185 1/ 4
82 µH 41 2.S MHz 6.7 4.80 158 0.185 1/ 4
82 µH 41 2.5 MH1 6.7 4.92 156 0.185 I/4
100 µH 25 2.5 MHz 3.6 7.68 139 0.165 I/ 4
120 µH 40 790 kHz 3.2 8.16 135 0.165 I/4
150 µH 47 790 kHz 3.0 8.16 135 0.165 I/ 4

226
 TABLE D-B (contin ued)
1\1inimum Moxi11111111 Form
Re.,·ona!ll Maximum :Wa.rimum Wi11di11g l1·11111/1
/Vomlnol lvfinimum Te.rr rreqllell('_I' DC mil lJiam<'fer I 1/32
lntf11c1a11ce Q Frequency I M/1: 1 Re.<i.<1a11ce Ro1i1111 f ill. / l ilr.I

180 µH 48 790 kHz 2.8 8.16 135 0.170 1/ 4

200 µH 47 790 kHz 2.7 10.3 120 0.170 1/ 4
220 µH 46 790 kHz 2.5 11.5 114 0. 170 1/ 4
250 µH 49 790 kHz 2.5 12. I Il l 0. 170 1/ 4
270 1• H 46 790 kHz 2.5 13.2 106 0. 175 I/4
300 µH 46 790 kH z 2.2 13.2 106 0.1 75 I/4
330 µH 41 790 kHz 2.0 13.9 103 0. 175 I/4
350 11H 46 790 kHz 2.0 14.4 102 0.180 I/4
390 µH 45 790 kHi 2.0 15.8 97 0. 180 1/ 4
470 µH 35 790 kHz 1.8 16.3 95 0.185 1/4
µ
500 id~ 49 790 kHz 1.8 18.0 91 0.195 1/4
560 µH 41 790 kHz 1.7 19.2 SS 0.195 I/4
680 µH 37 790 kHz 1.6 19.8 87 0.200 I/4
750 µ l·I 40 790 kH z 1.6 22.9 80 0.210 I/4
820 µH 33 790 kH z 1.6 22.? so 0.210 I/4
9 10 1d-I 32 790 kH z 1.4 24.0 79 0.220 I/4

1.00 mH J() 790 kH z 1.4 24.0 79 0.225 1/4

1.20 mH 34 250 H l 1. 1.2 33.6 66 0.220 I /4
1.50 mH 40 250 kHz I.I 37.2 63 0.225 I/ 4
1.80 mH 40 250 kHz 0.96 42.0 )9 0.235 1/ 4
2.20 mH 40 250 kHz 0.96 45.6 ~7 0.240 1/4

2.50 mH 48 250 kHz 0.96 45.6 57 0.260 J/8

2.70 mH 50 250 kH z 0.88 45.6 57 0.260 J/8
JJO mH 52 250 kH z 0.80 51.6 53 0.260 3/8
3.90 ml-I 53 250 kHz 0.76 57.6 51 0.275 2/ l\
4.70 mH 49 250 kHz 0.68 64.8 48 0.285 3/8

5.60 mH 53 250 kHz 0.68 69.6 46 0 ..100 3/8

6.80 mH 51 250 kHz 0.64 78.0 43 0.310 3/8
7.50 mH 49 250 kHz 0.60 85.2 ~I 0.3 10 3/8
8.20 mH 48 250 kHz 0.60 92.4 40 0.330 J/ R
9. 10 rn H 52 250 kHz 0.56 98.4 39 0.330 J /$

10.0 mH 41 250 kHz 0.52 10 1 38 o.m 3/ 8

12 mM 46 79 kHz 03 6 100 50 0.300 I/2
15 mH 50 79 kl-17 0.32 113 47 0.300 I /2
18 mH 49 79 k llz 0.29 128 44 0.325 1/ 2
22 mH 50 79 kHz 0.27 144 41 0.330 1/2

25 mH 59 79 kHz 0.250 I I5 46 0.340 5/ 8

27 mH 61 79 kHz 0.244 120 45 0.353 5/ 8
33 mH 61 79 kHz 0.232 134 43 0.353 51~
39 mH 59 79 kH z 0.220 147 41 0.370 5/8
47 mH 57 7? kHl. 0.206 168 )8 0.384 5/8

50 mH 57 79 kHz 0. 196 175 )7 0.400 5/ 8

56 mH Si 79 kHz 0. 188 189 36 0.400 5/8
68 mH 57 79 kHz 0.180 215 34 0.4 15 5/8
75 mH 53 79 kHz 0.174 222 33 0.4)0 5/8
82 mH 50 79 kHz 0.168 238 32 0.430 5/ 8
91 mH 51 79 kH z 0.166 250 31 0.430 5/ 8
100 mH 48 79 kH z 0. 157 278 29 0.446 5/ 8

Councsy of J. W. Mi ller (o .. Divi<ion of Rell lndu mic~.

227
APPENDIX E
TOROID-CORE DATA

This material is provided as a supplement to the information on toroids

that was presented earlier in the book. These data will be useful to engineers
and students who undertake the design of toroidal inductors and trans-
formers. Various core sizes and types are specified in this appendix, along
with pertinent operating curves.
Table 1-1 provides data on optimum frequencies and color codes for
powdered-iron materials which have I.I; characteristics from 1 to 75. Table
1-7 lists similar information for ferrite toroids.

PERMEABILITY VERSUS DC BIAS

Figure E-1 contains a set of curves for powdered-iron toroids which have
permeabilities ranging from 10 to 100. These curves demonstrate the H
characteristics of the cores versus other pertinent parameters.

PERMEABILITY VERSUS AC FLUX DENSITY

Figure E-2 presents a family of curves that relate to factor Bin toroid cores.
The initial permeability is rendered in percentage versus the ac flux density.
These data are useful for quick reference in a design exercise. It illustrates
clearly the effect flux density has on the core permeability.

229
 100
-2
90

80

.~ 70 - 15
ii -1
"'
II> 60
-8
.,E
0. 50 -3
~ Where N = turns
I = current (A)
:s 40 I =mean magnetic path (cm)
* 30
-40
20 - 41
-44
'--~~-L-~~~~-'--~~---'~~-L---''--~-'---'-----'~"-''-'-'--'----''---'--
2 5 10 15 20 30 40 50 60 80 100 200
DC MMF oersted (H)
Figure E-1 Permeability versus de bias for various powdered-iron core materials. (Courtesy of
Micrometals Corp.)

- 40
160

150
~------- -41
.~ 140
ii
"'., Where E = rms voltage
A = core cross section (cm 2 )
.,E 130
B = Ex
4.44ANI
1D8
N = turns
0.
f = frequency (Hz)
~
·c: 120
*

0.5 1.0 1.5 2.0 2.5

AC llux density (kG)

Figure E-2 Percentage of initial permeability versus ac flux density for powdered-iron toroids.
(Courtesy of Micrometals Corp.)

2 30
 TEMPERATURE CHARAC TERISTICS

It was established in the early chapters of this book that the core tempera-
ture of a transformer or inductor has a significant effect on the permeabili-
ty, which in turn affects the inductance of the core windings. Figures E-3
through E-6 contain curves for the main group of powdered-iron toroid
cores manufactured by Micrometals Corp. The curves show the various
core mixes versus temperature and percentage of inducta nce change.

TOROID-CORE FILL FACTOR

It is necessary for the designer to know the fill factor of a given toroid core
in order to make a proper core selection. This section explains how this is
determined and provides data on the various wire gauges versus fill factor.

+3.5
Average temperature coefficien t
+3.0
,,, ,,, No. 6
Material 2: 95 ppm/ ° C _,,/
+ 2.5 /
Material 6: 35 ppm/° C /''
+2.0
/
,, /

/
~ +1.5 / No.2
c /
"'
u /
::i
"O +1.0 /
c /
/
/
/
-55 - 25 /

cOJ / +25 + 100 +1 25

/
/
~ - 0.5 /
0.. /
/
- 1.0 / Temperature (° C)
/
/
/
-1.5 /
/
/
-2.0

-2. 5

Figure E..J Temperature/inductance curves for Amidon Associates and Micrometals

Corp. core types 2 and 6. (Courtesy of Amidon Associates.}

231
 +3.5
Average lemperature coefficient
+3.0

Material 10: 150 ppm/QC

+ 2.5
Material 12: 170 ppm/Q C

+2.0
Cl)
No. 10
g +1.5
"'
ti
:J
"O
c + 1.0 No. 12
c /
/
/
/
/
...... /

- 0.5

-1.0 Temperature {° C)

-1.5

- 2.0

-2.5

Figure E-4 Temperature/inductance curves for Amidon Associates and Micrometals

Corp. materials 10 and 12. (Courtesy of Amidon Associates.)

This material is presented through the courtesy of Magnetics, a division of

Spang Industries.
The term "wound cm 2 /usable window cm 2 " ( = K2) is the fill factor for
the usable window area. It can be shown theoretically that for circular-
cross-section wire wound on a flat form, the ratio of wire cm 2 to the area re-
quired for the turns can never be greater than 0.91. In practice, the actual
maximum value is dependent upon the tightness of winding, variations in
insulation thickness, and wire lay. Consequently, the fill factor is always
less than the theoretical maximum.
Random-wound cores can be produced with fill factors as high as 0. 7,
but progressive sector wound cores can be produced with fill factors of
only up to 0.55. Figures E-7 through E-9 are based upon fill factor ratios of
0.50, 0.60, and 0.70, respectively. As a typical working value for copper
wire with a heavy synthetic film insulation, a ratio of 0.60 may be used safely.
The term "usable window cm 2 /window cm 2" ( = K3) defines how much

232
 +8
Average temperature coefficient
+7
Material 1: 280 ppm /° C No. 3
+6 .. Material 3: 370 ppm/° C /
Material 15: 190pprn/° C /
<·5 -
/ /
+4 / No. 1
"'c
u
+3 /
~
u
:::i
v
c +2 - /
// ___ _ _ No. 15
c
..,
Ct>
c +1 0 +25 /
L-___ _ .... -- ... ~--
"'
~
-55 - 25
- I
C' 0
+ 50 + 75 +100 + 125
"'Ci
u
Cl. -1
Temperature (° C)
/
-2

-3

-4

-5

Figure E-5 Temperature/ inductance curves for Amidon Associates and Micrometals
Corp. core types 1, 3, and 15. (Courtesy of Amidon Associates.)

of the available window space may actually be used for the winding. The
charts are based on the assumption that the inside diameter of the wound
core is one-half that of the bare core; that is, K J = 0. 75 (to allow free pas-
sage of the shuttle) (Fig. E-10) . A typical value for the copper fraction in the
window area is about 0.40. For example, for A WG 20 wire, K, x K2 x Kl
= 0.855 x 0.60 x 0.75 = 0.385.

CORE SIZE SELECTION

Upon selection of the transformer core material and material thickness, the
next step is to select the proper size core for a transformer with a given
operating frequency and output power. The power-handling capability of a

233
 +9

+8

+7 - Average
tem pera tu re
coef fici ent
+6
Material 41 : 845 ppmf C
+5

~ +4
c:
"'
;:;
.g +3
c:
c:
"'
O'l
+2
c:
£"' ·~1
c
<1>
~ o~~~-'-~~--'~~-+-~~--'~~~_._~~_._~~~........~-
8:. - 55 - 25 0 +50 +75 + 100 + 125

- 1 Temperature (° Cl

-2

-3

-4

-5

Figure E-6 Temperature/inductance curves for Amidon Associates and Fair-Rite

Corp. ferrite-core material 41. !Courtesy of Amidon Associates.)

transformer can be determined by its W0 A c product, where W0 is the

available core window area in cm 2 , and Ac the core effective cross-sectional
area in cm 2 •
The W(IAc relationships are obtained by solving Faraday's Jaw in the
following manner:

Faraday's law = E = 4BmAcNf x 10- 4 square wave

E = 4.4B,,,,,,AcNf x J0 - 4 sine wave

where E applied voltage (rms)

Bm flux density in teslas
A c = core effective cross-sectional area in cm 1

234
 E
2
t
~
E 1.0
"'
15
.,
~
~

0.5 inside diameter remaining

Wire area) = 0 5
( Wound area · R
effec1ive win dow area =
0 ,_...:___.i::___w:'-1...~J....L"""-.L-~--L~-'--'---'--'-'....U..L-~~'---'---"-'-...l..CLl...J
1
3 2 T
10 100 1000
Turns
Figure E-7 Fill factor 0.5.

1O.O Fi II factor
0.5 inside diameter remaining
wire are<1 ) = O. S
( wound area
• 2
Effective window area =
3rrR

E
2
.,
~

~
E 1.0
.,"'
'ti
'ti
·;;;
.=

0 <-~C--'-""'--..C........L.....w::lc..J.....L..IC;__~_J..~-'---'''--''--''-'-.i...:..-'-~~-'-~-'----'---'--'-,___,
10 100 1000
Turns

Agure E-8 Fill factor 0.6.

235
E
.!:!.
~

~
E"' 1.0
"'
'O

"'
'O

c"'

0 L..,oc..._~c._---1c;:_..i.....CJ...L..L.£1...L_ __,__ _L._--l._J.....L.~l~•u.....___ _ .....__..___,__,_....._._,__,

10 100 1000
Turns
Figure E-9 Fill factor 0.7.

-- --- --- .......... ' Inside diameter

/
/"
/
' _.- of toroid core

Effect ive I
I
I
"\\

window area-., \
I
i-oo
I
I
I
I Clearance for
passage of
shu t tle

Window area = rrR 2 R2

Effective window area = _31T
_ _
4

Figure E-10 Window-area calculation.

236
 Core Size Sefection

N = number of turns
f = frequency in hertz
Aw = bare wire area in cm 2
Wa = window area in cm 2
K = window utilization fact o r
I = current (rms)
J = current density
Po = output power (total)
Pi = input power
Pr = total power
YJ = efficiency

Solving yields

NA _Ex 10'
c - - --
4Bmf
Window utilization factor :

K = NAw
wa
N KWa
= - -
Aw
Multiply both sides by Ac:

NAc = KWaAc =EX 10'

Aw 4B,,,J

Combining and solving for W 0 Ac yields

Ex 10'
4B111 JK

WA = EAw x 10'
a c 4BmfK

J = _!_ A
Aw cm 2

YJ
=Po
P,
Pi =El

237
 Toroid-Core Data App. E

EA _ EI _ Pi _ P0
w - J -1 - Jry

WaAc = W0 Ac + WaAc
(total) (primary) (secondary)
WaAc = P0 X 10" + P 0 X 10• = P0 X 10• (l/ry + l)
(total) J 4Bm JK 4B111 JKJ 4Bm fKJ
Po
P r=Y}+ P.
o

4
WaAc = PT X 10
4B11JKJ
Window utilization factor K:

Lamination and bobbin

toroid 1/2 I.D. remaining Wa (eff.) x fill factor x Aw bare = 0.4

C core and bobbin
Wa Aw total
PT x 10 4
square wave
1.6 x Bmfl
PT x 10·
sine wave
1.77 x Bmfl

The curve in Fig. E-11 shows the required core Wc/lc product plotted
against transformer output power for different frequency. The values held
constant were

Bm = 0.3T
J = 200 A/cm 2
K = 0.40
Y) = 95%
These values were held constant so that one nomograph could be used
with all materials and the engineer could adjust B111 , J , K, and ri to fit the
design. From the equation

WA = Pr x 10"
a c 1.6 X B111 fl

two nomographs were generated to compare output power

P. = Pr
0 l/ry +

238
 Transformer

10

1.0

.,
E
~
<:<."
~

s:
0 .1

0.01 8 ,,, =0.3T

J =200 A/cm 2
K = 0.4

1.0 10 100 1 kW

Output power (W)

Figure E-1 1 WaAc versus output power.

with WaAc and Po with weight. These nomographs were generated from the
lamination and C core in the article. The nomograph in Fig. E-11 compares
Po with W0 Ac; thus the size of the trans former can quickly be determined.
The nomograph in Fig. E- 12 compares power with weight; thus the weight
of a fully wound transformer can quickly be determined. These nomographs
have the following constraints:

Bm = 0 .3T
J = 200 A/cm 2
K = 0.4

239
 "' 100
..c
Ol

·~

10
Bm =0.3T
J = 200 A/crn 2
K = 0.40
1) =95%

0 10 100 1.0
Output power {W)

Figure E-12 Transformer weight versus output power.

Very likely, after the WaAc has been selected, a slight adjustment of the
constraints is required to correspond to the actual core WaAc products
available.
After calculating the inductance and de current, select the proper per-
meability and size powder core with a given L/2/2. The energy-handling
capability of an inductor can be determined by its W0 Ac product, where Wa
is the available core window area in cm 2 , and Ac is the core effective cross-
sectional area in cm 2 • The WaAc relationship is obtained by solving E =
L d/ldt as follows:
where E = voltage, volts
L = inductance, henries
I = current, amperes
N = number of turns
~ = flux, webers
Bm = flux density, teslas
A 'c = core cross section , m 2
µr = relative permeability
µ 0 = absolute permeability
(4rr X 10- 1 )

240
 Core Size Selection

H' magnetizing force, amp turns/m

i'm magnetic path length, m
K = window utilization factor
W'a = window area, m 2
J' = current density, A/m 2
Eng energy, watt seconds

E = L di= Nd+
dt dt
L =Nd+
dl
+= BmA'c

B m = 11rl1o H ' = /1rl1oNI

l'm
+ _ /J.rµ. 0 NIA 'c
l 'm
d+ _ 11r/1 0 NA 'c
dl L'm
L = Nd+ = l1rP.oN2A'c
dl f'm

Energy

If Bm is specified,

2
_ (BnY1~,A'r:
- -·· -- -- watt seconds

l KW'0 J' =Bmt;11

N 11r/J.oN
Solving for /J.rl1o yields

Substituting into the energy equation, we obtain

241
 100

8,., = 0.3T
0.1
J = 200 A /cm 2
K = 0.4

0.01.___._.....__._,._,_........._~~~·~·~' ~'--".....................
0.0001 0.001 O.Dl 0. 1

Agure E-13 WaAc versus Ll' / 2.

Let

W0 = window area, cm'

A, = core area, cm'
J = current density, A/ cm'
W0 = W0 X I0 - 4
A'c = Ac X lQ -4
J' = J x 10·

Substituting into the energy equation yields

Eng = WaA,B,,/K x 10-4

2
Solving for WaAc, we obtain

W0 A, = 2 (Eng) x lo•
B111JK

242
 100
°'....
.s::.
0)
'iii
3:
2u
:>
-0
.= 10

Bm = 0.3T
J = 200 A/c m2
K = 0.4

o,___,____.__._.._._..........___,__,__._~~'--~---'--'-..................
0.0001 0.001 0.01 0.1
2
Energy = LI 12

Figure E-14 Inductor weight versus LP/2.

Let

lm = magnetic path length, cm

I'm = lm X 10- 2
Bmlm I0- 2
X B111 1111 X 10-2
Kµ 0 ( J¥a x I o-•)(J x 1o•) J.loWaJK
1
For µ 0 = 4n x 10- ,

B111 l 111 X 10-2 Bmlm X 10'

J.lr = 4n X 10- 1 WuiK 0.4WalK

From the equation

two nomographs were generated to compare energy or Ll2/2 with WaAc and
energy or LJ212 with weight. These nomographs were generated for 13 com-
monly used powder cores in this article. The nomograph in Fig. E-13
compares L/2/ 2 with WaAc; thus the size of an inductor can quickly be de-
termined. The nomograph in Fig. E-14 compares L/l/2 with weight; thus
the weight of a fully wound inductor can quickly be determined. These
nomographs have the following constraints:

243
 Toroid-Core Data App.E

Bm = 0.3T
J = 200 A/cm 2
K = 0.4
After the core size has been determined, the next step is to pick the right
permeability for that core size, using the following equation:

µ = Bmlm X 10'
r 0.4nW0 JK

Very likely, after the permeability has been selected, a slight adjustment of
the constraints to match the available core sizes and permeabilities must be
made.
The outside diameter of the wound toroid, Fig. E-15 (less the outside
wrapper), may be calculated from the following equation (assuming that
one-half of the ID remains after winding):

(D = diameter)

where A , = n(R6-R~)
A,= A,
core window ID = D0 = 2R0
core ID = Db = 2Rb
core OD = De = 2Rc
n (K!J - R~) = n (R~ - R~)

Figure E-15 Calculation of outside diameter of wound toroid.

244
 Powdered-Iron Toroid Characteristics

R~ - R~

Rh - R~ = R6 &_
2
Rh = R6 (Y,) + Re
Rd R6 (%) + R/
OD 2Rd

POWDERED-IRON TOROID CHARACTERISTICS

Table E-1 contains a comprehensive listing of powdered-iron toroid cores.

The cores specified are available from Micrometals Corp. and Amidon
Associates. The characteristics are the same for both brands , inclusive of
the color codes. The dimensional data are given in English units. To convert
English inches to millimeters, multiply inches by 25.4. These tabular data
were provided through the courtesy of Micrometals Corp.

Dia. ! 0 .005 i n4;h~s MicfOmetals Lµh Typical Operating Color

Ht. · 0.005 inche$ Nufnbcr 100 T F requencv (MHz) Code

0.062
-.-1 1~·57 Std. ht.
Tl2- I
-2
-3
48
20
60
2
8
I
Blue
Red
Gray

®
- 6 17 16 Yellow
_t 0.050 -
-
7
10
18
12
14
40
White
Black
_1.27 - 12 7.5 90 Green/white
t - IS 50 Red/ whi1e
-I 0.125
3.18
1- -0 J. ISO Tan

B ht.
Tl2-2B 18.S 12 Red
+ 0.042 - 6B 13.5 20 Yellow
t 1.07 - 108 10 35 Black

0.078
- I l~.98 Std. ht.
T l 6- I
- 2
44
22
Blue
Red
I Gray
(OJ _J_0.060 - 3
-
-
6
10
61
19
13
12
30
Yellow
Black
~) T 1.52 -
-
12
IS
s
SS
80
I
Green/ white
Red/ white

-1 0.160
4.06
1- - 0 ) 120 Tan

TABLE E-1

245
 - - _r~:70
0.088
2.24 T20 - I 52 Blue
I I - 2
- J
27
90
R<d
Cral

®
-6 21 11 Y<l1o"
-7 2• y Whut
-10 16 JO Bl>:~
_ _ 1.78 -12 10 80 Grccn/ \lrihll<

t - IS 65 I Red/ " hue

-l 1- - 22 9 90 Gr<cn 1or3ngc
-0 3.5 120 T•n
0.200
5.08

ns- 1 70 0.'I UhH·

- 2 34 6 Red
- J 100 0 .6 Grny
- 6 27 II Yellow
-1 29 'I Wlutc
- 10 19 2S lll3ck
- 12 12 6() G1ccn / wh11c
- IS 85 I Rcd l""hllc
-0 J .S 120 f3n

0.151_.J
3.84 ---i
I- Std. ht .
T:?1- 2 33 u R<d

® 1- _f - 6 27 10 Yclln1.1o•
0.128 Black
- 10 19 !O
_ 3.25 - 12 12 60 Grctn/ \lo hllt
t -0 J.5 120 Tan

-1 0.280
7.11

0. 151
3.84 -
I I- Std. ht. rJ0-1
-2
85
~3
0.6
5
Blue
Red
Croy
- J l.W o. ~
- 6 10 Y<llo"
{OJ -• 0.128 -7
- 10
36
Ji
25
6
10
Whuc
Bl>ck
~ -3.25 - 12 16 60 Grttfl / \lri•hitc

1~_t - IS 9J 0.K Red/ white

-I 0.307 - 0 6 120 T~rn

7.80

5.2 1 ®
0205 _ .,..
1 l-
\......._..,)
fttd. ht.
0. 128
3.25
TJ7 - 8
- 2
- J
- 6
-7
-R
-10
HO
40
120
JO
32
12S
2S
10
5
0.6
2S
0.7
J
0 .6
Blue
Red
ti ray
Yellow
Whitr
Oningc
Olock
-12 IS 60 GrttO/ "- h1tc

0.375
9.53
1- -IS
-0
90
' .9
0.8
120
Red ' 04 h11c
Tan

TABLE E-1 (continued)

2 46
0.229-l
5·82
r- Std. ht.
T44-I
- 2
105
52
0.6
2
Blue
Red
0.l Gray

~t
-3 180

®
- 6 42 10 Yrllow
- 7 46 l While
\.....__.,/ 0.159 Orange
- 8 204 0.6
4.04 - 10 )J 18 lllock
- 11 18.l 50 C rten / v.•h11t'

~ 0.440
11 .18
1- Aht.
-1 ~
- 0
160
6.5
0.1
100
Rctl/ whi1c
fon

_ t_
0 .128 T44-2A 36 2.l Red
3 .25
-t
o.303 I
- I 7.70 -
T50- I
-2
100
49
O.l
2
Blue
Red
~. ht . - 3 175 O.• Gray
-6 40 10 Yellow
- 7 43 l White
0.190 - 8 190 0.4 Oran gt
4.83 -10 JI 16 Black
-t - 12
-15
18
IJ5
lO
0.6
Grccn/ whne
Red / while

~ 0.500
12.70
1- - 0 6.4 100 Ton

0.200-i 1-
5 . 0 8 @ 3 _r:.312 T51 - 2B IJ8 Red
- 68 102 Yellow.
_ 7.92

- 1 0.500 r- t
12.70

9.40 ®
0.370--1 1-
1t:.·,:~
_ 4.83
T68- 1
- 2
-3
-6
-1
- S
-10
I ll

52
205
57
195
41

32
0. 4
2
0.3
10
5
0.4
20
IJluc
!led
Gray
Ycllo,.·
White:
Orange
Black
Grccn/ ,,-.·hicc

r-t
- 12 21 JO
R<dl ...•h11c
--1 0.690
17.53
- ll
- 0
180
7.l
O.l
100 Tan

TABLE E-1 (continued

247
 A ht.
T68-2A 70 2 Red
_l__0.250 - JA 260 0.3 Gray
_ _ 6 .35 -6A 62 JO Yellow
-7A 73 5 Whitt

0.280 _
1 I- t
7.1 1 @ 3 rd.ht.
T72-2 128 2 Red
-0.260 360 0.2 Gray
-3
6.60 -7 95 5 White
-t
- I o.no ~
18.29
o.495
12.57 -
I I- T80- J 115 0.4 Blue
-2 55 2 Red
Id.ht.

®
-3 J80 O.J Gray
-6 45 10 Yellow
0.250 -7 so 5 White
-8 J90 0.4 Or~nge
6.35
-t - 10
-12
32
22
20
50
Black
Gree1\/white

--1 20.19
0.795 1- -15
-0
170
8.5
0.5
70.
Red/while
Tan

-I 14.22
o.s6o l--
ltd. ht. T94-I 160 0.4 Blue
0.312 -2 84 2 Red
-3 248 O.J Gray
_7.92 - 6 70 JO Yellow
t -JO 58 20 Black

~
-J S 200
-I
0.5 Red/white
-0 10.6 60 Tan
o.942
23.93
0.570......J L__
14.48 -1 I~
ftd.ht. TJ0<>-1 325 0.3 Blue

®
-2 135 2 Red
- 0.437 -J 450 0.2 Gray
-6 J 16 10 Yellow
-t-1 l.10 -7 IB 5 While
- J5 345 0.3 Red /·~" hice

-j 1.060
26.92
1--- - 0 J9 0.260 Tan

1-
0.780-1
19.81 cm itd~.::7 Tl30-J
-2
-3
200
J JO
350
0 .3
2
0 .2
Blue
Red
Gra>'
~ -6 96 10 Yellow
11.10 - 7 JOJ 5 White
-IS 250 0.3 Red/ white

-I 1.300
33.02
l_.J" -0 JS 50 Tan

TABLE E-1 (continued)

248
 @Std.ht
0.950~
24. 13I-
_i_ 0.570
TISJ-1
-
-
2
3
320
140
420
0.3
2
0.2
Blue
Red
Gray
14.48 - 6 I 15 10 Ytllow
- JS 360 0.3 Rtd/whi1t

-i 1.570 1--1

®
34.88
0.950--j
1--
24.13
Stdht
J_ 0.710
Tl84-I
-2
-3
soo
240
no
0 .2
2
0 .2
lllue
Red
Gray
TS.03 - 6 195 10 Yellow

- I 1.840
46.74
I-
1.250--1
r-
®
3 1.75
Std.ht
_l0.550
T 200-I
- 2
-3
250
120
425
0.3
2
0.2
Blue
Red
Gray
_13.97 -6 100 10 Y-:llow
s
1- t
-7 105 Whitr

--1 2.000
50.80
TABLE E- 1 (continued)

249
INDEX

A Balun cores, 125, 126, 141-43

Balun transformer, 100, IOL
AC excitati-Jn, 30-37
AC fields, demagnetization with, Bandpass fi lter, 87-88
176-77 Bars, 2
Adverse magnetic fields, 180 as antennas, 38-39, 49-62
Air-core loop antenna, 49-5 1 loop pattern symmetry, 52-53
Air gap,: 1-37 noise immunity, 53-54
actual, 36-37 practical loop circuits, 54-62
Hanna curve and, 31-36 diameter-to-length ratio of, 40-42
of permanent magnets, 167 effective permeability of, 40-42
of pot cores, 151-53 frequency vs. core materials, 48
Air-wound solenoidal inductors, sizes of, 39-40
14, 15 uses of, 39-40
Antenna effect, 53 Beads, 123-37
Antennas: for decoupling, 135-36
loading, 112-13 practical applications for, 130-33
rods and bars as, 38-39, 49-62 properties of, 124-26
loop pattern symmetry, 52-53 as Q killers, 133-35
noise immunity, 53-54 reduction of incidental radiation
practical loop circuits, 54-62 with, 136-37
shield-bead attenuation, 126-30
B sizes of, 137, 138-39
Balanced-to-balanced transformer, styles of, 137, 138-39
IOI , 102 vs. toroids, 123-24
Balanced-to-unbalanced B-H curve, 165-66
transformer, I 01, 102 Bidirectional coupler, 93-94

251
Index 252

Bifilar-wound transformers, 100- Dc-to-ac in verters, 114-17

101 De-to-de converters, 114
Bloch wall, 160, 162 Decoupling, beads for , 135-36
Brass slugs. 64 Demagnetization, 176-77
Broadband transformers, 48 DeMaw, M.F. "Doug," 83
pot-core, 153-54 Diameter-to-length ratio, 40-42
rod-core. 62-63 Dip meter, 11-12
toroidal , 99-112 Disaccommodation, 151
bifilar-wound, 100-101 Domain wal l motion, 160-64
conventional, 102-8 Doping, core, 117-22
qu adrifilar, 101-2 Dorfman , 164
variable impedance matching, Doubly balanced mixer~, 96-97
109-12 Drift, 68, 70
Bruene, Warren, 95
E
c E cores, 114
Capacitance: Effective height, 49, 51
approxim ation by nornograph, Effective permeability, 8, 9, 14-15
201, 203 of bars. 40-42
toroidal distributed , 113-14 flux density vs., 15-18
Capaciti ve-divider matching, 83-85 of pot cores, 151-53
Capacitors, 70-71 of rods, 40-42
C cores, 3, 4 Electrostatic shielding, 51, 53
Chebyshev filter, 91-92 External fields, permanent magnets
Circuit Q [see Q (quality factor)) in, 169-70
Coil cement, 118-21
Coil forms , 219 F
Coil placement, 42-47 Fair-Rite Products Corp., 31 , 48,
Coil turns. calculation of, 9-14 125, 141
Combiner transformer, IOI Ferrites (see also Bars; Beads, Pot
Conversion factors, 195, 198 cores; Rods; Sleeves; Slugs;
Core, defined, 1 Toroids), 6-9
Core doping, 117-22 characteristics of, 48
Core loss. 17-18 Ferromagnetic compound, 6
Coupling, interstage, 85-87 Ferroxcube Corporation , 7
Crystal anisotropy, 163, 164 FET switches, 116-1 7
Cup cores, 2, 124 Fill factor, 231-33 , 235, 237
Cylindrical cores and coi ls, 215-27 Filters:
D pot cores in. 87 , 157-58
toroids in , 87-93
DC excitation , 30-37 Fixed-value iron-core inductors,
DC fields , dem agnetization with, 220, 225
177 Flux density, 8, 15-18
253 Index

Free ends, 42 Irons, powdered (see Powdered

Fren kel, 164 irons)
Frequency: Irreversible losses, 179
beads and, 129-30
vs. core material, 48 J

G Jasik, Henry, 61

Gl ossary of terms, 182-87, 204-5 K

Grain-oriented steel-lam inated
cores, 3-4 k i factor, 22-23, 30
Granberg, 106
L
H
Laminated cores, 3-4
Hanna, C. R. , 3 I Lenk , John D., 83
Hanna curve, 31-36 Litz wire, 47, 54-56, 147-48
"Hard" magnetic materials, 1-2 Lock-up, 7
Harmonic filter, 90 Long-term stability, 151
Hayward , Wes, 96 Loop antenna, 49-62
Height factor, 49, 51 loop pattern symmetry, 52-53
High-pass fil ter, 92-93 noise immunity, 53-54
Hybrid flux path, 3 I practical circuits, 54-62
Hybrid transformers, 106, 108 Loop collector, 49
Hysteresis loop, 165-66 Low-level circuits, 79-99
capaciti ve-divider matching, 83-
85
filters, 87-93
1- E cores, 2
instruments, 93-96
Impedance:
interstage coupli ng, 85-87
beads and, 124-25, 128-30
miscellaneous applications, 96-
measurement of, 95-96
99
toroids and , I09-1 2
RF amplifiers, 80-83
Impulse magnetization, 173-75
Inductance: M
vs. ac/ dc excitat ion, 3-37
approximation by nomograph. Magnetic circuit, 167-69
201, 203 Magnetic core, defined , I
beads and , 129-30 Magnetic materials:
calculation of coil turns and, 9-11 bars (see Bars)
co il placement on rods and, 42-47 beads (see Beads)
determination of, 11-12 classification of, 1-2
Inductors, variable, 110-1 2 coi l turns, calculation of, 9-14
Instruments, toroids in , 93-96 inductance vs. ac/ dc excitation,
Interstage coupli ng. 85-87 30-37
Index 254

Magnetic materia ls (Cont'd.) Permanent-magnet (Cont'd.)

physica l and elect rica l traits of", han dling of. 18 1-82
2-9 magnetiza tion, 170-75
pot cores (see Pol cores) cond itions for complete
power capability, 14-22 saturation , 171-72
rods (see Rods) eq uipment. 172-75
sleeves (see Sleeves) measurements and cal ibration ,
slugs (see Slugs) 180-8 1
toroids (see Toroids) nature of. 159-65
volt-ampere ratings. 22-30 stability and stabilization . 177-80
Magnetization, 170-75 Permeability. 3
conditions fo r complete of beads. 125-26
saturation. 171-72 effective (see Effective
eq uipment, 172-75 permeability)
Magnetizi ng equipment. 172-75 of powdered-iron co res, 5-6
Magnet wire, 47, 54, 56 of toroids:
Metallurgical changes. 178-79 vs. ac flu x densi ty, 229. 230
Mica compression trimmers, 85 vs. de bias. 229. 230
Micrometals Corporation. 6. 215. Phase-reversal transformers, I 00,
231, 245 I 0 I , I 06, I 08
Mil ler, J. W., 67 Pi-wound RF choke, 73, 74
Mix. 5 PolydorofT, W. J., 42, 52
Mi xers, 96-97 Polystyrene cement, I 0, 118-21
MM PA (Magnetic Materia ls Pot cores, 2, 143-48
Producers Association), 159 design consideration s, 146-53
designs, 153-54
N dimensional and tolerance data
for, 207-14
Noise immunity, 53-54
in fil ters, 87, 157-58
Nomograph, 199
hardwa re, 144-46
p impregnation of, 143-44
physica l characteristics of, 124
Parasitic oscillation . 130-33 sizes of, 143
Per111ane111 M agm?I Guidelines, 159 to ne encoder with, 154-57
Permanen t-magnet material s, Potting co mpou nd, 120-22
159-87 Powdered irons, 5-6, 10
demag netization . 176- 177 slugs, 216. 217
elementary relati onsh ips, toroids, 229-34, 245-49
165-1 70 Power capability, 14-22
B-H curve, 165-166 Power dissipation factor. 30
external fields, 169-70 Power FETs. I 16- 17
magnetic circuit. 167-69 Power rating. 30
recoil loops, 166-67 Power supplies, I 14- 17
wo rking point of a magnet. 167 Prewound slug inductors, 222-~ 6
glossa ry of terms, 182-87 Push-pull power ampli fiers. 104-7
255 Index

Q Rods (Cont'd.)
coil placement, 42-47
Q (quality factor): diameter-to-lengt h ratio of, 40-42
beads and, 133-35 effective perm eability of, 40-42
coil placement on rods and, 42-47 frequency vs. core materials, 48
of loop antennas, 54-57 sizes of, 39-40
measurement of, 12-14 as transform ers, 62-63
pot cores and, 143, 147-50
powdered-iron cores and, 5-6 s
of R F chokes, 73, 75-77 Saturation, 7, 8
sl ugs and , 65-68, 70 of permanent magnets, 171 -72
toroida l-distributed capacitance Sense antenn a used with loop
and, 113 anten na , 59-61
Q-dope, 10, 118-21 Sevick, J., 111
Q meter, 12-13 Shape anisotropy, 164-65
Quadrifilar transformers, I01-2 Shield-bead attenuation, 126-30
Shock, effect o n perm anent
R
magnets, 180
Radiation, incidental, 136-37 Silver-mica capacitors, 85
Random-wou nd cores, 232 Sleeves:
Reactance: antenna loading, 112-13
approximation by nomograph, applications for, 137, 140-41
201 , 203 physical characteristics of. 124
ratio to resistance [see Q (quality Slugs, 216-27
factor) ] formals of, 64-65
Recoil loops, 166-67 performan ce considera tions,
Relative loss factor, 147 65-68
Reluctance changes, 180 practical circuits, 68-72
Resistance: prewound , 222-26
beads and, 129-30 vs. rods, 63
measurement of Q and, 14 sizes of. 64
ratio of reactance to [(see Q uses of, 63-64
(quali ty facto r)] "Soft " magnetic materials, 2
Resistance/ temperature correction Solenoidal RF choke, 73, 74
factor, 20 1, 203 Steel-laminated cores, 3-4
Reversible losses, 179-80 Strain anisotropy, 163, I 64
R F amplifiers, 57-59, 80-83 Stress, effect on permanent
RF chokes, 73-77 magnets, I 80
RF power bridge, 94-95 SWR indicator, 94-95
Rod-core transformers, 62-63 Symbols, 195-97
Rods, 2
as anten nas, 38-39, 49-62 T
loop pattern symmetry, 52-53 Tape-wound cores, 3-4
noise immunity, 53-54 Temperature/ resistance co rrection
practical loop circuits, 54-62 factor, 20 1, 203
Index 256

Temperature stability, 5-7, 69, 70, Toroids (Cont'd.)

178-79, 231-34 powdered-iron, 229-34, 245-49
Testing and Measurement of power supp! ies, I 14- 17
Permanent Magnets. 159 tape-wound, 3
Test instruments. toroids in, 93-96 tcm pera tu re characteristics,
Thermal demagnetization method, 231-34
177 toroidal distributed capacitance.
Tone encoder, 154-57 113-14
Toroidal effect, 49 Transmission-line transformer,
Toroids, 2, 79-122 99-102
antenna loading, l t 2-13 Tuned-circuit [see Q (quality
vs. beads, 123-24 factor)]
broadband transformers, 99-112
bifilar-wound, 100-101 u
conventional , I02-8
quadrifilar, 10 1-2 U cores, 2, 4, 114
variable impedance matching, Unbalanced-to-u nbalanced
109-12 transformer, JOI, 102
coil turns, calculation of, 9-14 Unbalanced transformer, JOO. IOI
core doping, 117-22 Unit permeance, 168
core size selection , 231-34, 237-45 Unmagnetized magnets, 182
fill factor , 231-33, 235, 236
low-level circuits, 79-99
v
capacitive-divider matching, Va riable-impedance matching.
83-85 109- 12
filters , 87-93 Vibration, effect on permanent
instruments, 93-96 magnets, 180
interstage coupling, 85-87 Volt-ampere ratings, 22-30
miscellaneous applications,
96-99 w
RF amplifiers, 80-83
permeability of: Weiss, Pierre, 159-60
vs. ac flux density, 229, 230 Wire gauges. 201, 202
vs. de bias. 229 , 230 Working point of a magnet, 167
II
I