(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Magnetic Alloys and Ferrites"

MAGNETIC 
ALLOYS 

AND 
FERRITES 



ELECTRICAL ENGINEERING PROGRESS SERIES 



\v 



M! 



m 



GNETIC 



l^'B^W 



Editor : 
M. G. SAY 






I I 



\K 



U\ 



FERRITES 



H NES 



MAGNETIC ALLOYS 
AND FERRITES 

Consulting Editor: 
M. G. SAY, Ph.D., A4.Sc,, M.I.E.E. 

The past few years have seen 
remarkable developments in magnetic 
materials, which have in turn 
contributed to progress in electrical 
equipment design. Quite recently, for 
example, among the magnetically -soft 
materials, grain-oriented silicon-iron 
alloys, ferrites, and grain-and-domain- 
oriented nickel-iron alloys have passed 
from the research stage to established 
production. Iron and iron-alloy micro- 
powders and a range of ferrites have 
similarly been added to the already 
considerable list of permanent-magnet 
materials. 

The book is therefore intended for 
electrical engineers specialising in 
applications such as: 

Radar 

Radio 

Television 

Telephony 

Telegraphy 

Industrial Process Control 

Instrumentation and Metering 

Protection 

Rotating Electrical Machines 

Nuclear-particle Accelerators, etc. 

A revealing account of modern views 
on the fundamental processes of 
magnetisation — including the domain 
theory — is given by Professor Brails- 
ford in the first section of the book. 
The related subjects of materials for 
magnetic recording and magneto- 
striction, together with magnetic 
compensation and non-magnetic alloy- 
ing are included. 



Because you are interested! 

The extensive NEWNES list of tech- 
nical books includes a wide range of 
titles covering electrical and engin- 
eering subjects. A handy pocket 
catalogue is available giving details 
and prices. To obtain a copy please 
send a postcard stating your interests 
to: — 

GEORGE NEWNES, LTD. 

(BOOK SALES DEPT.) 

TOWER HOUSE 

SOUTHAMPTON STREET 

LONDON, W.C.2 



s 



y 21/- 






NET 





WIGAN 
PUBLIC LIBRARIES 





ISp 



ELECTRICAL ENGINEERING PROGRESS SERIES 

Consulting Editor 

M. G. Say, Ph.D., lf.Sc., M.I.E.E. 

A aeries of books on aspects of oloctrical engineering in 
which there has boon much recent progress and develop- 
ment. Each book is written by specialist contributors for 
the professional engineer or technician who desires the 
latest information. 

ELECTRICAL EARTHING AND ACCIDENT PRE- 
VENTION 

ROTATING AMPLIFIERS. The Amplidyne, Motadyno, 
Magnicon and Magna volt and their 086 in Control 
Sysi cms 

CATHODE-HAY TUBES 

MAGNETIC AMPLIFIERS AND SATURABLE 
REACTORS 

MAGNETIC ALLOYS AND FER KITES 

CRYSTAL RECTI FIF.RS AND TRANSISTORS 



MAGNETIC ALLOYS 

AND 

FERRITES 



Consulting Editor 
M. G. SAY, Ph.D., M.Sc, M.I.E.E. 

Contributors 
F. BRAILSFORD, Ph.D., B.Sc. (Eng.), M.I.E.E. 
F. KNIGHT, A.M.l.E.E. 
W. S. MELVILLE, B.Sc. (Eng.), M.I.E.E. 

B. W. ST. LEGER MONTAGUE, B.Sc. 
M. G. SAY, Ph.D., M.Sc, M.I.E.E. 

C. GORDON SMITH, M.A., M.I.E.E. 



WITH 115 ILLUSTRATIONS 



WiGAN 
CENTRAL] 

LIBRARY. 



LONDON 

GEORGE NEWNES LIMITED 

TOW£R HOUSE, SOUTHAMPTON STREET 

STRAND, W.C.2 



Copyright- 
All Rights Reserved 



WIGAN U^ p *fe s 



W^ 



First published. . 1954 






%w- 



G?6 7t/.S^ 




PREFACE 

The past few years have seen remarkable developments in 
magnetic materials, which- have, in turn, contributed to progress 
in electrical equipment design. Quite recently, for example, 
among the magnetically-soft materials, grain-oriented silicon-iron 
alloys, ferrites, and grain-and-domain-oriented nickel-iron alloys 
have passed from the research stage to established production. 
Iron and iron-alloy micropowders and a range of ferrites have 
similarly been added to the already considerable list of permanent- 
magnet materials. 

The wide variety of materials now at the disposal of the elec- 
trical engineer and component designer demands an extensive 
knowledge. Properties and limitations must be understood if 
skilful use is to be made of a unique and extending range of 
remarkable technological devices. The object of this book is to 
provide that information. 

A revealing account of modern views on the fundamental 
processes of magnetization— including the domain theory— is 
given by Professor Brailsford in the first section of the book. The 
related subjects of materials for magnetic recording and magneto- 
striction, together with magnetic compensating and non -magnetic 

alloying are included. 

J 8 M. G. S. 



PRINTED IX GREAT BRITAIN BT THE WHITEFRIARS PRESS LTD 
LONDON AND TONBRIDGE 



CONTENTS 



seffaom 

1. 



■2. (a) 

m 

3. (a) 

w 

4. 



7. 



Ferromagnetic theory ..... 
,6,7 Professor F. Brailsford, Ph.D ., B.Sc.{Eng.), 
M.I.E.E. 

Soft magnetic materials . 

By W. S. Melville, B.Sc.(Eng.) : M.I.E.E 

Magnettcally-soft ferrites 

PER]VL\NENT magnet steels and alloys 
By F. Knight, A. M.I.E.E. 

mlcropowder magnets 

Permanent magnet ferrites 

By B. W. St. Leger Montague, B.Sc. 

Magnetic powder cores 

By C. Gordon Smith, M.A., M.I.E.E. 

Non-magnetic ferrous and magnetic compen- 
sating ALLOYS ..... 

ByC. Gordon Smith, M.A., M.I.E.E. 

Magnetic recording materials .... 
By C. Gordon Smith, M.A., M.I.E.E. 

Magnetostrictive materials .... 
By C. Gordon Smith, M.A., M.I.E.E. 

Appendix : introduction to m.k.s. magnetic 
units ....••• 
By Professor M. G. Say, Ph.D., M.Sc, M.I.E.E. 

Index. ...•••■• 



37 

85 
95 

144 
148 

159 

171 
179 
184 

190 

197 



v.! 



1. FERROMAGNETIC THEORY 



By 

Professor F. Brailsford, Ph.D., B.Sc.(Eng.), M.I.B.E., 
Mem.A.I.E.E. 

GENERAL 

From the magnetic point of view all natural substances may bo 
classified as diamagnetic, paramagnetic or ferromagnetic. The 
magnetic permeability of the materials in the first two groups, 
however, differs so little from that of free space that, to the 
technologist, they are " non-magnetic " ; but quite different are 
the ferromagnetic elements, nickel, cobalt and iron, which display 
magnetism to an extraordinary degree. With these elements as 
constituents a wide range of ferromagnetic alloys of practical 
importance also may be made, though there are some, the Heusler 
alloys, which include only non-ferromagnetic components. 

The Magnetization Curve 

If we have a long, uniform specimen in a long solenoid having 
a magnetomotive force of i ampere-turns per metre, then, using 
the rationalized m.k.s. system of units, the magnetic field strength 
H = i amperes per metre (A/m). The flux density in the sample 
is B =iiH = HrHoH webers per square metre (Wb/m 2 ), where 
flo =4^/10' henrys per metre (H/m) is the absolute permeability 
oHrce space, and ^ r (numeric) is the relative, and /x the absolute 
permeability of the material. 

We also have B =pjl +J where J is the intensity of magneti- 
zation in the specimen. 

Fig. 1-1 shows, on suitable scales of H, the lower and upper 
parts of the initial magnetization curve of a sample of ordinary 
dynamo iron, which we may take to be representative in form <>1 
that for any ferromagnetic. The slope of the curve at the origin 
gives the initial relative permeability j* rt , and the slope of the 
tangent to the curve the maximum value n Tm . J approaches a 
saturation valve J s at high field strengths, and if the scale of H 
enables this approach to saturation to be plotted, as in the upper 
curves, a fairly well-defined knee point appears at which it is 
evident that a change in the mechanism of the magnetization 



FERROMAGNETIC THEORY 



UPPER CURVES 

H. amp/m 

60.000 



120,000 * 



2-0 



PoH B |_ T======S =«— — 


v 7 i ^^r- 

5 S^"^ 7 




f J 

-•-KNEE 


B^^~ 


firm/ 





4000 



si 



2P00 



/~\l l rm 








Mr, V 





200 

//, amp/m 

LOWER CURVE 



B, Wb/m 2 



2-0 



Ftgr- I— 1. — Initial magnetization curve and perm k ability of dynamo 

IRON SHEET. 



process occurs. The quantities /z rl , ft rm , J s and sometimes the 
position oi'the knee are important quantities in assessing magnetic 
quality. 

Magnetic Hysteresis 

Fig. 1-2 shows the cyclic hysteresis loops obtained for a slowlv 
alternating applied field : (a), (6) and (c) are for a 4 per cent, 
silicon-iron transformer sheet material for different peak values, 
B max , of the induction. The positive tips of the loops will lie on 
the initial magnetization curve, while the area of the loops is a 
measure of the energy dissipated in heat as a hysteresis loss per 
unit volume, iv h , for each complete cycle. For any alternating 
frequency, /, the'specific energy loss, in watts per kilogramme, is : 



FERROMAGNETIC THEORY 






(l) 



where 3 is the density of the material (kg/m s ). For a given B,„,, : , 
in a specimen w h is not, in general, a constant, but almost certainly 
increases with increasing frequency (see also p. 6). 

Curves (a), (6) and (c) are for a soft magnetic material, that is 
one in which the lowest hysteresis effect is desired. Curve (d) has 
been included for comparison, in particular, of the relative values 
of H. This is for a permanent magnet material (Alcomax II) m 
which the fullest possible hysteresis loop is aimed at. 

The form of the relation between iv h and J max when the alter- 
nating hysteresis cycle is made slowly is shown in Fig. 1-3 at (a). 
The particular values given are for a sample of silicon-iron sheet 
with 1-91 per cent, silicon. As may be seen from Fig. 1-1, B ma . x 
and J will be inappreciably different except at high Mux 

"" max • • * • • i i D" 

densities. For B max < I J s the Steinmetz empirical law, wy=??i*,„„, r . 
fits the observed results fairly well, where rj is the Steinmetz 
coefficient for the particular curve and the exponent n varies from 
about 1-6 to 1-8 for different materials. It will be seen, however, 
that there is a discontinuity in the curve occurring near the knee- 
point already mentioned. Above this region a straight line ol the 
form 

where 6 and c are constants, fits the observed results fairly well 
up to the highest density to which measurements have been 
made, 1 i.e. up to about 0-85J s . The form of the relation in the 
neighbourhood of saturation is not known (but see Reference 39). 
In the armatures of rotating machines a complicated mixture of 
alternating hysteresis loss, w h , and rotational hysteresis loss, 
w' h , occurs. The latter arises when, instead of an alternating flux 
in one direction in a specimen, there is a constant flux density 
which rotates in one plane : w\ is then the loss per unit mass 
for one complete, slow revolution, and its relation to the rotating 
intensity of magnetization, J, is as shown in Fig. 1-3 [b). This 
curve 2 has a knee-point discontinuity, rises to a maximum and 
falls towards zero as saturation is approached. 

Eddy Current Losses 

Further losses occur with alternating or rotating magnetization 



FERROMAGNETIC THEORY 



due to the induced eddy currents. They are kept small in practice 
by employing insulated laminations. If the non-linearity between 




-12,000 



2-0 



-1-0 



1 





-500 



Z/.amp/m 



12,000 -120,000 




500 





H.omp/m 



120.0C 



-2-0 -2-0 

(c) (d) 

Fig. 1-2. — Hystekesis loops. 

(a), (b), (c) Soft magnetic material (4 per cent, silicon-iron), (d) Permanent 
magnet alloy (Alcomax II). 



FERROMAGNETIC THEORY 



1000 



500 



£ 




1000 



3 500 



2 



2-0 




(a) 



w> 

.7, Wb/m 3 

(b) 



2-0 



Fig. 1-3. — Form of hysteresis loss curves. 
(a) Alternating ; (6) rotational. Material : silicon-iron with 1-91 per cent, 
silicon. 

B and H, to be seen in the hysteresis loops, is neglected and /x is 
assumed to be a constant, simple formulae may be derived for 
sinusoidal conditions for the eddy losses in thick or thin plates. 3 

For alternating conditions in thick plates where a I ( — 1 

the eddy loss is given by 



>1, 



We ~ 2a. h * ' 
2a . ^ 3/2 8 



W/kg. 
W/kg. 



where H max and B max are amplitudes at the sheet surface, 2a is 
the sheet thickness and p is the electrical resistivity. 

For thin sheets where a J (—) is small compared with unity, 



W. 



(7T.2a.f.B, 



W/kg. 



(2) 



6 . P . 8 

For a constant magnetic field H and flux density B rotating 
at a uniform speed of / revolutions per second, again neglecting 
complications introduced by hysteresis, the eddy-current losses 



6 FERROMAGNETIC THEORY 

will be precisely double the above values if we put B and II for 

H max Mid II milx in the formula!. 

Total Iron Loss 

The total specific iron loss W t in a thin sheet under A.C. condi- 
tions will therefore be : 



W t = W 



W* 



W/kg. 



which, from equations (1 ) and (2), becomes at a given flux density : 

W, w, 

-f = * • / + f J/kg./cycle 

where k is a constant. 

The two parts of this expression are, respectively, the eddy 
current and hysteresis components of the total loss per cycle. It 



X> 0-03 



0-02 



sTK°-< 




.oj 0*02 



^ V.0-01 



apparent hysteresis 
loss/cycle 



50 

fc/s 



75 



100 




100 



/'»/. 1-4. SEPARATION OF IRON LOSSES. 

(Cold -inlm ■(■:! :!] percent, silicon-iron sheet at B max = 1-5 Wb/m 2 observed 
points.) 

has been common to assume that w A is independent of frequency, 
making the relation botween WJf and fa, straight line from which 
an experimental separation of the losses could be made. This 
assumption, however, appears to be untenable. Fig. 1-4 shows, 
for example, some results obtained at three frequencies, at 
B maa = 1-5 Wb/m 2 , on a sample of cold-reduced 3| per cent, 
silicon-iron transformer sheet, 0-35 mm. thick. This material 
had unusually low hysteresis loss. On the left is the conventional 
method of loss separation. On the right is an alternative method 
which assumes the validity of equation (2). In this particular 
case the - apparent eddy loss " is about double the calculated 



FERROMAGNETIC THEORY 7 

value. The difference between these two quantities is sometimes 
called an :1 extra " or " anomalous loss." The evidence 4 indicates 
that it is due to a hysteresis effect and that the right-hand method 
of separation is, in fact, probably fairly accurate. 



DOMAIN THEORY 

Weber and Ewing considered that a ferromagnetic substance 
was made up of atomic or molecular magnets capable of rotation. 
The magnetic properties, according to Ewing, were determined 
by the magnetic forces between neighbouring magnets. The 
ihape of magnetization curves and hysteresis loops could be 
explained in this way but quantitatively the theory fails com- 
pletely to explain the high permeabilities observed with the 
ferromagnetics. 

Modern Theory 

In the modern theory the atoms or molecules of every para- 
magnetic or ferromagnetic material have, as in the older ideas, a 
magnetic moment. Fig. 1-5 represents, diagrammatically, a 
free atom of iron, with 26 orbital electrons situated in the main 
shells K, L : M and N about the central nucleus. Each electron is 
regarded as spinning gyroscopically on its own axis, by virtue of 
which and the negative charge which it carries, it is equivalent 
to a current in a small circular path. It therefore has a magnetic 
moment of unit amount equal to one Bohr magneton. Indepen- 

ELECTRONS WITH 
PARALLEL AND ANTI-PARALLEL 
SPINS 




Ave / -ve 



FREE IRON ATOM. 
MAGNETIC M0MENT = 4 UNITS 




Fig. 1-5. — Spinning electrons in the iron atom. 



8 



FERROMAGNETIC THEORY 



dent evidence indicates that in a free iron atom there would be 
15 electrons with parallel spins about a particular axial direction 
while the remaining 1 1 spin in the opposite direction. The figure 
indicates that all the atomic sub-shells are magnetically neutral 
except the incompleted outer one of shell M. In this there are 
four uncompensated spins which give the atom as a whole a 
magnetic moment of four units. Cobalt with one more orbital 
electron than iron has + 8 an( i — 2 spins in this sub-shell, and 
nickel -f- 5 and — 3. The free atoms of cobalt and nickel therefore 
have magnetic moments of 3 and 2 magnetons respectively. 

However, the foregoing applies to hypothetical free atoms, and 
when these come together in the metallic state the distribution of 
the electrons in the outer shells is somewhat modified, giving 
average magnetic moments for iron, cobalt and nickel of 2-22, 1-71 
and 0-606 units per atom respectively. 

The Langevin- Weiss Theory 

We suppose now that a magnetic field is applied in a particular 
direction to a substance, at room temperature, whose atoms or 
molecules each have a magnetic moment and are free to turn 
in any direction. Mutual magnetic effects between them of the 
kind envisaged in E wing's theory will be small and may be 
neglected. However, there will be a tendency for the atomic 
magnetic axes to align themselves in the direction of the field, thus 
giving a resultant intensity of magnetization in that direction. 
This process is, however, so very strongly disturbed by the 
thermal motions of the atoms that the magnetic effect is, indeed, 
very feeble. Langevin 5 derived an expression for this case, which 
was later modified by the assumptions of quantum mechanics, to 
give the following : 



T = tanh kT 



(3) 



In this J is the intensity of magnetization produced by the 
field H. J is the value of J corresponding to parallel alignment 
of all the magnetic axes, v is the magnetic moment of each atom 
or molecule, T is the absolute temperature and k is Boltzmann's 
universal gas constant. 

Tho Langevin theory is adequate for the paramagnctics, but 
fails when applied to the ferromagnetics, as we may see, for 
example, by substituting in equation (3) the appropriate values 



FERROMAGNETIC THEORY 



2-0 



1-0 





Jo— -^^ 


<?y 






20°C/ 




400°C/ 










/ C ^ 








\ 

WEISS FIELD 

H m =NJ 










' \ 


MODIFIED LANGEVIN CURVES 


a/ 


^LMf 













H, amp/m 



6xl0 8 



Fig. 1-6. — Langevin magnetization curves for iron at different tempera- 
tures, AND WEISS MOLECULAR FIELD LINE. 



for iron. For iron, in the m.k.s. system we have J — 219 Wb/m 2 , 
v = 2-56 X lO" 29 Wb-m, h = 1-37 x 10~ 23 J/°K, and at 20° C, 
T = 293° K., and hence 

6-38# 

W b/m 2 



J = 2-19. tank 



10 9 



Tke magnetization curve given by this expression is plotted in 
Fig. 1-6, from which it may be seen that saturation is approached 
only for fields of the order of 4 x 10 8 A/m, a figure which, com- 
pared with experimental observations on iron, is too high by a 
factor of about 10 4 . 

Weiss, 6 however, extended the Langevin theory to ferro- 
magnetism. He assumed that in any ferromagnetic there existed 
a " molecular field " which was proportional in magnitude to the 
intensity of magnetization in the material, i.e. the molecular field 



10 



FERROMAGNETIC THEORY 



H m = NJ where N is a constant. We may represent this relation 
by the -straight line OP shown in Fig. 1-6. If the slope of this line 
is less than the initial slope of the magnetization curve calculated 
from the Langevin theory, the material would be unstable in the 
unmagnetized condition at 0. For suppose the material were 
accidentally magnetized by a small amount to a point a. This 
magnetization would automatically produce the molecular field 
corresponding to the point b, which by the Langevin theory 
would produce magnetization in the material corresponding to c, 
and so on until the point P was reached. The material would thus 
spontaneously magnetize itself to a saturation value J s , a little 
lower than J , as indicated for a temperature of 20° C. by the 
point P. 

Experimental Results 

The Weiss theory has been amply confirmed by experiment 
and the origin of the powerful molecular field has been found 7 
by the mathematicians. It is due to quantum-mechanical forces 
of interaction between neighbouring spinning electrons in the 
metal, which may be expected to produce nearly parallel align- 
ment of the spins and hence spontaneous saturation in the elements 
iron, cobalt, nickel and gadolinium. 8 (The latter is a rare element 
which is ferromagnetic at temperatures below 1(3° C). 

From experimental results a value can be assigned to the 
constant N, and in Fig. 1-0 the straight line OP has been put in 
at the correct slope for iron. From equation (3) the Langevin 
curves corresponding to temperatures of 400° and 770° C. have 
also been drawn. It is clear that the intersection of the straight 
line and the curves gives the saturation values corresponding to 
the different temperatures. We can in fact derive the following 
relation : 



fc-'—S 



T 



(4) 



where 9 = 



where J. is the saturation value for absolute temperature T and 

vNJ_ 
k ' 

From equation (4) we see that when T = 9. J s = 0. The 
relation between J s and T for iron is plotted in Fig. 1-7. It will 
be seen that the saturation value falls with increasing temperature, 



FERROMAGNETIC THEORY 




11 



Fig. 1-7. — Calculated 

AND OBSERVED SATU- 
RATION OF IRON. 



1000 



7TK 



becoming zero at 6 = 1043° K. (770° C.) for iron. Then 9 is the 
magnetic change, or Curie point, and is the temperature at which 
a material changes from the ferromagnetic to the paramagnetic 
condition. The broken line indicates the experimentally observed 
saturation values for iron in fair agreement with the theory. 
Similar results are obtained for cobalt and nickel. 

Domains and Crystals 

The theory thus far indicates that any ferromagnetic material 
is spontaneously magnetized to saturation, even in the absence of 
an externally applied field, and we have therefore to account for 
the material in the apparently unmagnetized or partially mag- 
netized condition. Weiss made a second postulate that the 
material was divided up into small volumes, called domains, 
each of which was saturated in a particular direction but that, for 
material in the technically demagnetized condition, these direc- 
tions in neighbouring domains varied at random. Thus on 
applying a field a resultant magnetization in the field direction 
occurred due mainly to the domain vectors redirecting them- 
selves in whole domains, this process requiring a much smaller 
applied field than in the case of a paramagnetic. 

L 



12 



FERROMAGNETIC THEORY 



[0001] 






IRON 



COBALT 



NICKEL 



Fig. 1-8. — Crystal structlties of iron, cobalt and nickel, showing also 

THE MILLER INDICES. 




FERROMAGNETIC THEORY 



13 



Fig. 1-9. — Magnetization processes in an iron-type crystal. 



Now metals are crystalline substances. The crystal structures 
of iron, cobalt and nickel are shown in Fig. 1-8. Iron crystallizes 
on a body-centred cubic lattice, nickel is face-centred whilst cobalt 
takes a hexagonal form. In general, with no applied field a single 
domain will be part of a crystal in a polycrystalline material but 
will contain a very large number of atoms. Earlier experiments 
indicated that in ordinary polycrystalline materials the domains 
varied in size, containing up to 10 15 atoms. However, recent 
work shows that domains may also be very much larger than this. 
In iron it is found that the direction of the spontaneous magnetiza- 
tion of the domain will be, without preference, along any one of 
the cube-edge directions of the crystal and a force is required to 
turn the magnetic axis away from this position. In nickel the 
equilibrium position is along a long diagonal of the crystal, and in 
cobalt along the longitudinal axis. 

We may now consider the magnetization process for a ferro- 
magnetic of the iron type by referring to Fig. 1-9. The squares 
represent a part of a single crystal, the edges corresponding to 
the cube edge directions of the crystal. This is shown divided up, 
at first, into four equal domains magnetically saturated in the 
directions of the arrows. There would be six possible directions 
for these arrows, but four only are shown for the purpose of this 
diagram. Initially the resultant magnetization in any direction 
is zero, corresponding to the point of the magnetization curve. 
Suppose now a small field H is applied at an angle to a cube edge 
as shown. The first effect is that the boundaries between the 
domains move so as to increase the volume of the domains having 
a component of magnetization in the field direction and to 
decrease the others. The initial, almost reversible, part of the 
magnetization curve shown at a is thus obtained. When H is 
further increased some of the oppositely directed domains become 
unstable and the magnetization in these suddenly swings through 
90° or 180°. This occurs on the steeply rising part of the curve 
shown at b. An enlargement of this part of the characteristic 
would therefore show discrete steps, or Barkhausen jumps, as 
indicated in the circle. At c the latter process is complete and the 
magnetization in all the domains lies along the cube-edge nearest 
to the direction of H . This is the knee of the curve. Beyond c, a 
further increase in magnetization can occur only by a reversible 
turning of the domain vectors out of the cube edge direction into 



14 



FERROMAGNETIC THEORY 



that of the field. This is found to require a field strength of a 
higher order than is needed below the point c (see Fig. 1-1) and 
gives the part d of the curve. Eventually this process is complete 
and technical saturation of the material results, as shown by the 
part e. 

Large single crystals of iron, cobalt, nickel and a number of 
their alloys have been grown and the magnetization curves in 
different crystallographic directions determined. The curves for 
iron are shown in Fig. 1-10 for three directions, denoted by the 
Miller indices 1 100], etc. (see Fig. 1-8). It will be seen that the 
cube edge is a direction of easy magnetization, whilst, for the 
[110] and [111] directions, there is a knee near JjVZand JJV$, 
respectively, as we might expect from the theory and from the 













~j//^ Oooj 


1 








/ / 

/ 

IRON 




J5 

s 


/\j 9 
'( V2 

V3 








[in] 






0-5 


O^^m 




NICKEL 







s 



0-5 



20,000 

H,omp/m 



40,000 





(0001] 


f 






/[ioTo] 

and 
[jl20] 




COBALT 







400,000 800,000 

H.amp/m 



Fig. 1-10. — Magnetization curves tor single crystals or iron, cobalt 
and nickel. (The broken lines are theoretical.) 



FERROMAGNETIC THEORY 



15 



geometry of the crystal. Results for cobalt and nickel are also 
shown. 

Experimental Confirmation 

The existence of ferromagnetic domains has been well estab- 
lished by a variety of experiments, and there is a technique by 
which the domain boundaries, at the metal surface, can be made 
visible under the microscope. The metal surface is electrolytically 
polished and a soap solution containing a fine precipitate of 
magnetic iron oxide (Fe 3 4 ) is applied. The magnetic particles 
collect at the domain boundaries, which may thus be examined. 
The resulting patterns are complicated and difficult to interpret. 
The simplest are for crystals of the iron type, where the surface 
being examined is slightly inclined to a cube-face plane of the 
crystal. 10 It is concluded from the observations, in confirmation 
of theoretical work by Neel, 11 that the domains in this case are 
narrow plates with short " closure " domains to form, in the 
absence of an applied field, small closed magnetic circuits in the 

material. This is illus- 
. [ooi] trated in Fig. 1-11, from 

P , which it is seen that, for 

iron, the domain bounda- 

OoqI ^~~~->~. r * es are P araue l to CUD0 

faces, the angle between 

the domain vectors being 
180° on either side of the 
boundary. Alternatively, 
at the ends of the plates, 
the domain boundary sur- 
faces are at 45° to cube- 
face planes, the domain 





Fig, 1-11.— Probable arrange - 

MKNT OF DOMAINS IN AN IRON 
CRYSTAL WITH NO APPLIED 
TTELD. 



16 



FERROMAGNETIC THEORY 



vectors being at 90° on either side. In the former case there is 
no normal component of magnetization across the boundary, 
and in the latter the normal component is constant across the 
surface. There is therefore, in either case, no free pole appear- 
ing at the boundary. This is one of the conditions required 
to make the energy of the whole arrangement of domains a 
minimum. Another is that, for iron, the vectors shall lie along 
cube-edge directions as indicated in the Figure. There will, 
however, also be energy associated with the boundaries them- 
selves and with the magnetostriction of the material, matters 
which will be referred to in the following sections. Neel has 
shown that the total energy is a minimum, leading to a stable 
arrangement of the domains, for the configuration shown, where 
the width of the plate-like domains is proportional to the square 
root of their length. In single crystals of 3-8 per cent, silicon-iron, 
Williams, Bozorth and Shockley have found domain widths of 
the order of 01 mm, although wide variations in domain dimen- 
sions are possible. They have also given results to show that the 
boundaries move, when an external magnetic field is applied, in 
the expected manner, to widen the domains whose vectors are in 
the field direction and to contract the adjacent ones. 

The Upper Part of the Magnetization Curve 

Considering the observed magnetization curves for single 
crystals of iron in Fig. 1-10, it will be seen that a comparatively 
small field in a cube edge direction is sufficient to bring the 
material almost to saturation. The process of magnetization is 
mainly one which involves sudden reversals or swings through 90° 
of the domain vectors in whole domains. In a perfect crystal 
saturation might be expected for a vanishingly small applied 
field. 

However, for the field applied in the [110] direction the process 
just mentioned would be complete when the domain vectors were 
all at 45° to the field direction and the magnetization resolved 
into the field direction would then be JJV2. This gives the 
knee point beyond which the bulk magnetization J increases, 
not by sudden changes in direction of the domain but by a smooth 
rotational process. 

Similarly for the [111] direction the knee point will be at JJV3 
followed by a slow increase in J. 



FERROMAGNETIC THEORY 



17 



Theoretical expressions may be derived for these curves. 12 
If, in a saturated cubic crystal, the saturation vector J a is in 
any position relative to the cube edges there will be crystalline 
forces acting to restore J t , in the case of iron, to a cube-edge 
direction. There is therefore energy stored in the crystal on this 
account and, for cubic symmetry, this is given, very nearly, by 

E = K + K x (Si*S 2 * + S 3 2 £ 3 2 + S* 2 Si 2 ) + K 2 (S^rf . . (5) 
where K , K x and K 2 are constants and S lt S 2 and S 3 are the 
direction cosines of J a relative to the cube edges as co-ordinate 
axes. If now in a domain forming part of an iron crystal a field H 
is applied in the diagonal direction shown in Fig. 1-12, resulting 
in a rotation through an angle a of the saturation vector J s , then 
it may be shown that J s moves in the plane of the cube face and 

E =JL,-|---*(1 -cos 4a). 
o 

Apart from this magneto-crystalline energy there is also, due to 
the presence of J s in the field H, potential energy 



- HJ S cos 

The total energy E T = K Q 4- — ^ (1 — cos 4a) 

o 



■HJ. cos 



(;-)■ 



The value of a which makes E T a minimum, determined from 

dy. 

gives the stable position of../,. 

The magnetization in the field direction is 



J = J. cos 



(H 



(«; 



(?) 



Elimination of a between equations (6) and (7) then gives the 
expression 

1K X J_ 

" J. 



H = 



»r 



(8) 



which represents the magnetization curve for the [110] direction. 
Similarly an expression, which is, however, somewhat lengthy, 
may be derived for the [111] direction, this including the constant 
K z also. 



18 



FERROMAGNETIC THEORY 



FERROMAGNETIC THEORY 



19 



Using the vahies K x = 4-5 X 10 4 , K 2 = 2-25 x 10 4 J/m 3 and 
J s = 2- 16 Wb/m 2 , these expressions are plotted by the broken 
lines in Fig. 1-10 and show reasonably good agreement with the 
experimental results. 

Nickel may be dealt with similarly, using appropriate values of 
K x and K 2 . On the other hand, hexagonal cobalt requires the 
expression 

E = K + JTA" 

for the magnetocrystalline energy where S x is the direction cosine 
of J s relative to the principal axis. 



f H 



Wig. 1-12. — Rotation process in 

A DOMAIN OF IRON. 




Domain Boundaries 

Inside a domain the forces of interaction between neighbouring 
electron spins produce, in effect, the powerful Weiss molecular 
field, which acts to produce spontaneous saturation as already 
described. To rotate the axes of the atomic magnetic moments 
out of this position in the presence of the field would require 
energy which, for a reversal, would amount to H W J S per unit 
volume where H w is the Weiss field. On either side of a domain 
boundary, however, there will be an angle of either 90° or 180° 
between the respective spins, and this involves stored energy at 
the boundary. Bloch has shown 13 that as a result of this, in 
combination with the magnetocrystalline energy represented by 
equation (5), there is a gradual, rather than a sudden, transition 
in the angular position of the electron spins across the boundary 
which therefore has a substantial thickness. 






A 180° boundary layer, or Bloch Avail, is shown diagrammati- 
oaUy in Fig. 1-13. Such a layer in iron would, according to Stoner 
and Wohlfarth, 14 have a thickness of about 8-4 x 10- 6 cm, 
corresponding to about 600 parallel layers of atoms, with energy 
stored amounting to 0-84 x 10" 3 J/m 3 . 

Magnetization in Comparatively Low Fields 

The magnetic quality of a ferromagnetic material depends 
upon the case or otherwise with which domain boundaries can 
move when a magnetic field is applied. This in turn depends 
upon internal strains in the material and upon that property of 
ferromagnetics, known as magnetostriction, whereby the material 
changes its dimensions when its state of magnetization is changed. 
The dimensional changes occurring are small but none the less of 
the greatest theoretical importance. Fig. 1-14 shows how the 
length of an iron crystal 
changes in the field direc- 
tion when magnetized 
along different crystallo- 
graphic axes. 15 Nickel and 
cobalt, on the other hand, 
contract when magnetized 
in any direction. It is 
known that internal 
strains, or crystal lattice 
distortions from other 
causes, also have a pro- 
found effect ; for example, 
a material in the hard, 
cold-worked condition has 



ooma.n A -,- vvall thickness 

ABOUT \Q~S cm 




DOMAIN B 



/•'/;/. I IS. SlKiWINC ( : ha di AL 
QHANGB l-N DIRECTION OI MAO- 
NF.l'K MO.UK-VJ'S ACROSS THE 

BOUNDARY BOTWBBK TWO 

DOMAINS. 




20 



FERROMAGNETIC THEORY 



20X10 



-6 



*<■ 10 



-10 





[100]/ 


/ 




' [110] 






\[111] 


\ 



1-0 



2-0 



Js Wb/m' 



Fig. 1-14. — MAGNETO- 
STRICTION X op 
SINGLE CRYSTALS OF 
IRON. 

A is the increase in 
length per unit length. 



a very much lower permeability and higher hysteresis loss than 
the same material which has had the cold-working strains relieved 
by heat-treatment (for example see Fig. 1-16). 

Becker 16 and Kersten 17 have shown how to relate the per- 
meability and coercivity with the magnetostriction and the 
internal stresses. Becker considered an iron-type crystal, with 
90° domain boundaries, having an idealized distribution of 
compressive and tensile stress. The equilibrium positions of 
domain boundaries with no applied field are at positions of zero 
stress. As the field increases the boundaries move reversibly until 
the boundary reaches a position of maximum stress when a sudden 
irreversible jump occurs. On the other hand, with 180° boun- 
daries the boundary, having a finite thickness as already described, 
reaches an unstable position and jumps forward where the stress 
gradient is a maximum. 

On this basis the following expression, in rationalized m.k.s. 
units, for the initial relative permeability may be obtained : 



FERROMAGNETIC THEORY 



21 



%j:~ 



AM = 



where A s is the saturation magnetostriction along [100] and p is 
the amplitude of the internal stress. For a pure, well-annealed 
material in which all internal strains have been removed except 
those unavoidably present due to magnetostriction, this becomes 

3tt/* .Af.E 

where E is Young's modulus, and this formula represents the 
highest value of initial permeability to be expected. 

Somewhat similarly Kersten derives an expression for coercivity 
of the form 

3A^ 

2 J. 



H r = 



(10) 



where p again denotes the internal stress amplitude. 

The formulae indicate that low values of magnetostriction and 
internal stresses will give a high permeability and low coercivity. 
Since the initial magnetization curve lies closely inside the 
hysteresis loop this condition also corresponds to a low hysteresis 
loss. 

A s can be changed by alloying. Fig. 1-15 shows, in curve (a), 
how it varies with the nickel content in a series of niokel-iron 
alloys. Curve (6) shows observed values of //. rl for air-quenched 
alloys 18 whilst (c) is derived from equation (9) above. At about 
81 per cent, nickel A s = 0, and therefore /x rl theoretically rises to 
infinity. However, the similarity between curves (6) and (c) is 
striking and in confirmation of Becker's theory. 

The Problem of High Coercivity 

In the permanent-magnet materials the characteristic of prac- 
tical importance is the demagnetization curve, examples of which 
are shown in Fig. 1-24. This curve is a part of the hysteresis loop 
for saturation. The maximum hysteresis effect is desired in this 
case. In recent years materials of very high coercivity have been 
produced and the theoretical problem is to find a mechanism 
winch will explain the high values observed. 

The earlier permanent -magnet materials were martensitic 
steels depending upon the carbon present to introduce, after 



22 



FERROMAGNETIC THEORY 



appropriate heat-treatment, high internal strains and so high 
mechanical and magnetic hardness. These materials had 
coercivities up to about 20 : 000 A/m. They were followed by the 
dispersion-hardened alloys having coercivities up to 70,000 A/m. 

These high values of coercivity can hardly be accounted for by 
the Becker-Kersten mechanism. For if, in equation (10), we 
take A, = 20 x 10-« and J s = 1-5 Wb/m 2 , and further, let 
p = 1-5 x 10 9 newtons/m 2 (97 tons/in. 2 ), which is as high as or 
higher than the probable ultimate tensile strength of the material, 
we have H = 30,000 A/m. The theory might therefore embrace 
the martcnsitic steels but not the later materials. A later idea of 
Kersten's, the foreign-body theory, appears to havo an even more 
limited application. 

High coercivity is, moreover, not always accompanied by high 
internal strains. There are, for example, certain copper-nickel- 




eooo 



< 



4000 




40x10 



-20 



-40 



Fig. 1-15. — Initial pkkmkabtlity and saturation magnetostriction of 

NICKEL-IRON ALLOYS. 



FERROMAGNETIC THEORY 



23 



iron alloys having coercivities up to about 40,000 A/m which are, 
at the same time, ductile, malleable and machinable. 10 It is 
reported that coercivities of over 30,000 A/m may be obtained, 
without heat-treatment, from compressed, commercially pure 
iron powder provided that the particles are below about 10~ 7 cm. 
in diameter. 20 This iron in the ordinary solid form would have a 
coercivity of only about 100 A/m. Again, as a matter of theo- 
retical interest, very high coercivities have been observed 21 in 
materials, such as brass, containing ferromagnetic impurities, 
although in this case the retentivity would be exceedingly low. 

Stoner and Wohlfarth 14 have shown theoretically how, in a 
very simple and probable way, values of coercivity of the order 
of those experimentally observed, and even much higher, could 
be accounted for. N6eJ has independently and similarly discussed 
the same problem. The material is considered to consist of fine 
ferromagnetic particles embedded in a dissimilar matrix. This 
might therefore well apply to the powdered-iron .magnets, to 
such instances as brass containing iron as an impurity, and also 
to the dispersion-hardened alloys where one highly ferromagnetic 
phase may be in a particular state of precipitation from a magneti- 
cally dissimilar one. 

It may be shown that if a ferromagnetic particle is below a 
certain critical size it will be a single domain. For if a particle 
single domain it will have energy stored on account of its 
spontaneous magnetization and its self-demagnetizing tteld. If 
the particle divides into two domains with a 180° boundary this 
energy vanishes but there is then energy stored due to the domain 
boundary wall as described on p. 18. Calculation shows that the 
latter energy is greater than the former in particles below a 
certain size which will therefore be, on this account, single domains. 
Stoner and Wohlfarth calculate that for spherical iron particles 
the critical diameter will be about 1-5 X 10~ 6 cm. 

For single domain particles in the form of prolate spheroids it 
may then be shown that the field required to produce reversals 
of the magnetization in the particles, which we may take to be 
the coercivity, is given by 

where N a and N b are, respectively, the polar and equatorial 
demagnetizing factors of the prolate spheroid. H c will be a high 



24 



FERROMAGNETIC THEORY 



value even if the particle departs only by a small amount from 
the spherical form : for example, if the ratio of the polar to equa- 
torial axes is 1-1 then, for iron, H c would be about 60,000 A/m. 
Higher values would be obtained with more elongated particles, 
the limiting value for iron being about 800,000 A/m. 

Strong support for the theory is given by some published re- 
sults 20 on powdered iron magnets. A coercivity of 34,000 A/m 
is quoted for a particle size of 10 -7 cm., but only 800 A/m for the 
same material with a particle size of 5 X 10~ 6 cm. Stoner and 
Wohlfarth's theory would, indeed, hold only for particles below 
the calculated, critical diameter of 1-5 x 10 -6 cm. 



TECHNOLOGICAL ASPECTS 

It will be clear from the preceding that a requirement for high 
permeability .. and low hysteresis loss in a particular material is 
that the material shall be as free as possible from internal strains. 
Careful annealing of the metal is therefore necessary. Fig. 1-16 
shows, for example, magnetization and rotational hysteresis loss 
curves for a sample of 4 per cent, silicon -iron in the annealed and 
cold-worked conditions respectively. 22 

Internal Strains and Impurities 

A further important source of internal strains results from 
impurities, held either in solution or as larger inclusions in the 
metal. Yensen has made an extensive study of the effect of small 
amounts of the common impurities in iron and silicon -iron. 23 
Carbon is particularly harmful, as Fig. 1-17 shows. Cioffi reduced 
the impurities carbon, sulphur, oxygen and nitrogen, in small 
specimens of iron, to a few thousandths of 1 per cent, each, by 
heat -treatment at temperatures a little below the melting point 
in a hydrogen atmosphere. 24 This treat mont raised the maximum 
relative permeability to 280,000, which is about forty times that 
of ordinary dynamo iron, and reduced the hysteresis loss to about 
one-twentieth of that of the commercial material. For a single- 
crystal specimen of iron similarly treated 25 a maximum relative 
permeability of 1,430,000 was obtained for a cube edge direction 
of the crystal. There are limits to the degree of chemical purity 
which can be expected in materials produced by commercial 



2-0 



■■o 



-> 












FERROMAGNETIC THEORY 

2.000 



_a> 1,000 



3 



4,000 

H,amp/m 



(b)/ 




/ i£ 





epoo 



vo 
J. Wb/m 



2-0 



40 000 




Wig. 1-16 (above). — Effect of severe 

COLD WORK ON 4 PER CENT. SILICON- 
IRON. 

(a) Annealed strip, 0-65 mm. thick. 
(b) After cold-rochicing to 0-30 mm. 
thickness. 



Fly. 1-17 (left). — Effect of carbon as 

AN IMPURITY ON TIIK MAXIMUM PER- 
MEABILITY OF IRON. 



processes, rather than as small samples under laboratory condi- 
tions, but it is of interest to note that the material discussed in 
the next section is produced economically in large quantities with 
a carbon content of loss than 0-005 per cent. 

Preferred Orientation of Crystals 

The iron and silicon-iron sheet materials used in electrical 
machinery and in transformers have, in the past, been produced 
by a hot-rolling process. Smith, Garnett and Randall, however, 
in experiments on nickel-iron alloys found that if the material 
was heavily cold-reduced in thickness, and annealed, greatly 
unproved magnetic properties were obtained along the direction 
of rolling of the strip material. 26 Coss later described a cold- 
rolling procedure for silicon-iron, with silicon up to 3-5 per cent., 

M.A.F. c 




FERROMAGNETIC THEORY 
2-0 



FERROMAGNETIC THEORY 



27 



1-5 



1-0 



S 



/^ [001] 




Djo]^ 




X W^ 






[001] 




<C 


r^ 


^fn] 




rf* 


/ 


^ 


JjlO] 






x 



8,000 

ftA/m 



16,000 



8,000 

tf.A/m 



16,000 



/'/;/. 1—18. — DlKECTIONAT. MAGNETIC PROPERTIES OF POIA cl: VSTALL1NE COLD- 
HKIil OKD 3 PEK CENT. SILICON-IRON (left) AND OF SIN'CI.E ll!VST.\I.S OK 

3-8i> pkk i BUT. sii.k ox-ikon (right). 

which also resulted in a magnetically anisotropic sheet, the 
direction of highest permeability and lowest hysteresis loss bcinu 
again in the rolling direction. 27 This material has been highly 
developed and is now being produced on a large scale. It has 
proved so greatly superior to the hot-rolled material that, in the 
United States at least, the latter is being superseded by cold- 
reduced steel for transformers and in turbo -alternators. 

The directional properties of this material are due to the 
constituent crystals in the sheet taking up a " preferred orienta- 
tion." In silicon-iron the crystals are lined up, approximately, 
with a cube edge direction in the direction of rolling and with the 
diagonal [110] plane in the plane of the sheet. In the best material 
almost the whole of the crystals are so arranged with, however, a 
small spread about this position. The polycrystalline material 
therefore has directional magnetic properties similar to those for 









a single crystal. Fig. 1-18 shows, for example, a comparison 
hot ween the magnetization curves of an early sample of (>oss 
sheet and a 3-85 per cent, silicon-iron single crystal, 28 while 
Fig. 1-19 shows curves of hysteresis loss in the sheet 1 and in a 
2-1 per cent, silicon-iron single crystal. 29 

A useful method 2 of estimating the degree of preferred orienta- 
tion in a specimen of the cold-reduced sheet is to cut a disc from 
it, and to measure the torque required to rotate it slowly in a 
strong magnetic field which is in the plane of the disc. Such a 
disc, if cut from the diagonal [110] plane of a cubic crystal, may 
be shown by making use of equation (5), page 17, to require the 
following torque per unit volume : 

This theoretical curve is shown at (a) in Fig. 1-20 for a 3 J per cent, 
silicon-iron crystal. The observed curve at (b) was obtained for a 
disc cut from a polycrystalline. cold-reduced 3J per cent, silicon- 
iron sheet. The close resemblance in shape of the two curves will 
be noted. Comparison of their amplitudes indicates that about 
80 per cent, of the crystals had the preferred orientation which 



L = - - 



sin 4a. 4- - . sin by.. 
f>4 



aoo - 



200 









54-7 // / 

// 90 ° / 
/ / I " 







400 



200 













[100JX 



1-0 

^Wr-Wb/m 2 



2-0 



JwaxMblm 1 



2-0 



fig. 1-19. — Alternating hysteresis loss in different directions in 

POLYCRYSTALLINE COLD-REDUCED 3 PER CENT. SILICON-IRON (left) AND IN 
SINGLE CRYSTALS OF 2-1 PER CENT. SILICON-rRON (riijllt). 

c 8 



FERROMAGNETIC THEORY 




FERROMAGNETIC THEORY 



29 



Fig. 1-20. — Torque curves. 

(a) Theoretical for [110] plane of 3 J per cent. Silicon iron single crystal. 
(b) Observed for polycrystalline cold-reduced 3.J per cent, silicon-iron disc. 

has been already mentioned and is clearly illustrated in the 
diagrams in Fig. 1-18. 

Some Effects of Alloying 

Silicon-iron with varying amounts of silicon is universally 
used for the laminations of transformers and electrical machines. 
Brittleness limits the amount of silicon that can be usefully 
employed to about 5 per cent. Four per cent, silicon raises the 
electrical resistivity of commercial iron by about four times and 



therefore reduces eddy current losses in this inverse ratio. At 
the same time, probably because of the effect of silicon on the 
impurities present, and on the grain size, the hysteresis loss is 
reduced to about one-half. Aluminium behaves similarly and 
there is no other competitor, at least for materials for use in 
power plant. 

However, a remarkable series of magnetic alloys is obtained by 
alloying iron and nickel together in different proportions. When 
the composition is near to Fe Ni 3 very high initial and maximum 
permeabilities are obtainable with very low hysteresis loss. This 









5,Wb/m 2 










~| 


(a^ 
















/ (h) / 


(c)\ 
(d)\ 










-ao 


/-40 


1 o 1 


\J/ 


40/ 


//.amp/m 


80 



Fig. 1-21. — CoitfPARisoN of hysteresis loops. 
(«) Dynamo iron ; (6) 4 per cent, silicon-iron ; (c) Mumetal ; (d) Supermalloy. 

has already been discussed on p. 21. It may be noted that 
the initial relative permeability of dynamo iron is about 250, but 
nickel-iron alloys of about the composition mentioned have 
values of over 10,000 and, when great care is taken with the heat 
treatment, this figure is raised to about 100,000. A comparison 
of the hysteresis is given in Fig. 1-21 for (a) dynamo iron, (6) 4 per 
cent, silicon-iron transformer sheet, (c) Mumetal with a composi- 
tion of 75 per cent. Ni, 16 per cent. Fe, 5 per cent. Cu and 4 per 
cent. Mo, and (d) carefully prepared Supermalloy 30 containing 
79 per cent. Ni, 15 per cent. Fe, 5 per cent. Mo and 0-5 per cent. 

Mn. 

These nickel-iron alloys are valuable in applications where 



30 



FERROMAGNETIC THEORY 



superior properties in weak fields are required. They are, how- 
ever, too low in saturation value and too high in cost to be used 
in heavy electrical equipment. The way in which the saturation 
value 31 varies with the nickel content in iron is shown in Fig. 1-22. 
The corresponding curves for silicon in iron, and for cobalt in 
iron, are also shown. Cobalt it will be seen raises the saturation 
value to a maximum about 11 per cent, greater than that of 
iron. This higher saturation would be useful, for example, in the 
rotor punchings of electric motors, but the use of the alloys is 



2-0 



¥Uj. 1 22. — Saturation values 

OF SILTCON-IRON, NICKEL-IRON 
AND COit ALT -IRON ALLOYS. 



*? 



1-0 



Co 




\si 


Ni / 

• y 











50 



ALLOYING CONSTITUENTS, % 



100 



limited because of the high cost of cobalt and the difficulty in 
rolling the alloy. 

Heat Treatment in a Magnetic Field 

It has been found with Certain alloys that, if they are allowed 
to cool through the Curie point in the presence of an applied field, 
the materials at room temperature are magnetically anisotropic 
with improved properties in the selected direction. For soft 
magnetic materials this effect is greatest in nickel-iron with 65 to 
70 per cent, nickel. Fig. 1-23 shows results obtained by Dillinger 



FERROMAGNETIC THEORY 



31 



and Bozorth 32 for a specimen with 65 per cent, nickel. Curve (a) 
was for the material annealed in hydrogen at 1,400° C. Curve (6) 
shows the remarkable change in permeability occurring after 
heating the specimen to 650° C. for a few minutes in a magnetic 
field of 800 A/m. The domains form during cooling through the 
magnetic change point. The domain vectors and magneto- 
strictive expansion will be aligned by the field in one direction. 
If the temperature is high enough to allow strain relief in the 
metal during the subsequent cooling, the domain vectors will 
finally have one axial direction which is energetically preferred 
and which thus becomes a direction of easy magnetization. 

It has been similarly found 33 that dispersion-hardened alloys 
of suitable composition can have their demagnetization charac- 
teristic greatly improved by applying during cooling a magnetic 
field some thirty times greater than the above. Fig. 1-24 shows 
curves for Alcomax, («) for the direction selected during cooling, 
and (b) for a lateral direction. The possibility that, in this case, 
the material is made up of particles below the critical size, 
(see p. 23), whose shape during domain formation has been 
influenced by this treatment, and to which, therefore, Stoner 
and Wohlfarth's theory would apply, has been discussed by 
Hoselitz and McCaig. 34 

600,000 



400,000 



200,000 



















L 


(a) 





4 8 

Zf,amp/m 



12 



Fig, 1-23. — Effect of ANNEALING in a mac;neth VJXLD on a sample of 65 FEB 

CENT. NICKEL-IRON ALLOY. 



32 



FERROMAGNETIC THEORY 



Industrial Measurements 

The precision measurement of the magnetization curve and 
hysteresis loop of a magnetic material may be done, for specimens 
in the form of straight strips or rods, in a permeameter of which 
several different types are in common use. These include the 
Illovici, the Fahy, the Webb and Ford, and the Burrows 35 and 
that described by Armour, King and Walley. 36 

The latter type is shown diagrammatically in Fig. 1-25. The 




1-0 



0-5 



«f 



-40,000 



-20,000 

7/,amp/m 



Fig. 1-24. — Demagnetization curves of alcomax. 

(a) In tho direction in which the field was applied during cooling, (b) In a 
lateral direction. 



specimen is magnetized by a heavy copper winding and the 
magnetic circuit is completed by a pair of heavy yokes. Extra 
windings are provided at the ends of the specimen to provide 
additional m.m.f. for the reluctance of the junction between the 
specimen and the yokes. By suitable adjustment of the currents 
in the main and compensating windings uniform field conditions 
are established, at least over the central part of the specimen, 
ff-coils are provided near the specimen, and it is usual to measure 



FERROMAGNETIC THEORY 



33 



FOUR- POLE ,,.,,,,,.,, 



CHANGE-OVER 
iWITCH 




YOl\E 



MAIN 

— MAGNETISING 

COIL 



COMPENSATING 

VVINDING 



SPECIMEN 



iron loss primary *%• 1—26 (above). — Permeameter cir- 

TESTER \ / SPECIMEN OtfJT ARRANGEMENT WITH SPECIMEN 



mowv 




REMOVED FROM YOKE. 



Fig. 1-26 (left). — Connections for 

I RON -LOSS TESTINO. 



34 



FERROMAGNETIC THEORY 



(/{ Pt)B), or •/. directly, employing a 5-coil round the specimen 
and an II coil alongside of identical area-turns connected in 
opposition. When uniform conditions are established H may be 
determined from the ampere turns on the main magnetizing 
winding and ./ from the deflection of a ballistic galvanometer. 

Iron-loss measurements at supply frequencies are normally 
made on a specimen built up to form a closed magnetic circuit. 
There are three types of tester recognized by the B.S.I. 37 : (a) the 
Lloyd- Fisher Square, (b) the Churcher tester, and (c) the Epstein 



SPECIMEN 




Fig, 1 27. ('A.MI'JSKU. A.C. I'OTKNTlOMETKIt BOB RON-LOSS ZESTING. 

Square. These differ mainly in the size and arrangement of the 
specimen. In the Lloyd-Fisher method thirty-two pieces, each 
■ < 7 cm., are used. These are arranged on edge in four packs 
of eight, like the sides of a box, with small corner pieces to complete 
the circuit. The Churcher 1 ester uses sixteen pieces each 30 x 4 in. 
in two packets of eight, the circuit being completed by U-shaped 
end pieces. The Epstein square has four equal packets of lamina- 
tions with butt joints. Tn all cases the tester is provided with 
primary and secondary windings as shown diagrammaticallv in 
1-26. With the circuit shown copper losses in the primary 



FEEEOMAGNETIC THEORY 



35 



winding are not included in the wattmeter reading, but corrections 
must be made for instrument losses and for losses in the cornel' 
pieces. A dynamometer wattmeter is usually employed and. to 
determine B mamt the average voltage is observed using a moving- 
coil voltmeter in conjunction with a rectifier. 

An alternative method of iron-loss measurement, which is 
convenient for flux densities up to about half the saturation value 
and for small specimens, is by means of an A.C. potentiometer. 
A form of this due to Campbell M is shown in Fig. 1-27. In this 
circuit a ring specimen is shown, with primary and secondary 
windings, excited by a sinusoidal current rather than with the 
sinusoidal flux which is aimed at in the preceding methods. The 
magnitude and phase angle of the secondary e.m.f. is measured 
in relation to" the magnetizing current and hence the iron loss 
may be calculated. A balance is obtained by varying the mutual 
inductance M and the potentiometer resistance r. Then the iron 
loss in watts is given by 

the power factor by cos <j> = rj^^ -f- cu 2 M 2 ) 
and the flux density by B max = Ej4faT s 

where T x and T z are the primary and secondary turns respec- 
tively, I x is the primary r.m.s. current, a> = 2irf, a is the cross 
section of the specimen and E av is the average value of the 
secondary e.m.f. 

Various forms of A.C. bridge have been used for measurements 
under incremental conditions, that is when the specimen carries 
an alternating and a steady flux at the same time, and for 
measurements at audio and radio frequencies, but a fuller 
discussion of magnetic measurements is not possible here. 



References 

1. Brailseoho. P. Joum. I. ICE.. 19:59, 84, 399. 

2. BitAiLSFOim, F. Journ. I.E.E., 1938, 83, 566. 

3. RrssKi.i,. ,\. •■ Alternating Currents," Vol. J. 1914. 

4. Stewakt. K. H. Proc. I.E.E., 1950, Pt. IT, 97, 121. 

5. Lanokvin. I\ Ann. tie Chim. >i ,/<■ I'hya., 1905, (8), 5, 70. 

6. Weiss, P. Jovm. de Phye., 1907, (4), 6, 661. 

7. Hkiskmikkc. YV. Zeits.f. Phys., 1928, 49, 619. 

8. Bkthe, 11. Hundb. d. Phys., 1933, 24 2, 595. 



36 FERROMAGNETIC THEORY 

9. Honda, K., and Kaya, S. Sci. Rep. Toh. Imp. Univ., 1926, Series 1, 
15, 721. 
Kaya, S. Sci. Rep. Toh. Imp. Univ., 1928, Series 1, 17, 639 and 1157. 

10. Williams, H. J., Bozorth, R. M., and Siiockley, W. Rhys. Rev., 

1949, 75, No. 1, 155. 

11. Neel, L. Joum. de Phys., 1944, (8), 5, 241. 

12. Akulov, N. S. Zeits.f. Rhys., 1929, 57, 249 ; 1931. 67, 794 ; 1931, 

69, 78. 

13. Bloch, F. Zeits.f. Phys., 1932, 74, 295. 

14. Stoner, E. C, and Woblfarth, E. P. Phil. Trans. Roy. Soc„ 1948, 

A, 240, 599. 

15. Webster, W. L. Proc. Roy. Soc, A, 1925, 109, 570. 

16. Becker, R. Zeits.f. P%s.,1930, 62, 253 ; Rhys. Zeits., 1932, 33, 905. 

17. Kersten, M. Zeits.f. Phys., 1931, 71, 562 ; Zeits.f. tech. Rhys.. 1938, 

19, 546 ; Electrotech. Zeits.. 1939. 60, 498. 

18. Elmen, G. W. Journ. Frank. Inst., 1928, 206, 317 ; 1929, 207, 582. 

19. Dahl, O., Pfaffenberger, .1., and Schwartz, N. MetaWwirtachaft, 

1935, 14, 665. 
Neumann, H. MetaUwirtschaft, 1935, 14, 77S. 
Neumann, H., Buchner, A., and Reinboth, M. Zeits.f. Mekdlkunde, 

1937, 29, 173. 

20. British Pat. Spec. 590,392. 

21. Schroder, H. Ann. Phys., 1939, 36, 71. 

Constant, F. W., and Formwalt, J. M. Phys. Rev., 1939, 56, 373. 

22. Brailsford, F. " Magnetic Materials," 1948. 

23. Yensen, T. D. Trans. Amer. I.E.E., 1924. 43, 145. 

Yensen, T. D., and Ziegler, N. A. Trans. Amer. I.E.E.. 1935, 23, 
556 ; 1936, 24, 337. 

24. Cioffi, P. P. Phys. Rev., 1932, 39, 363 ; 1934, 45, 742. 

25. Cioffi, P. P., Williams, H. J., and Bozorth, R. M. I'hys. Rev., 

1937, 51, 1009. 

26. Smith, W. S., Garnett, H. J., and Randall, W. F. Brit. Pat. Spec, 

366, 523. 

27. Goss, N. P. Trans. Amer. Soc. Metals, 1935, 23, 511 ; Brit. Pat. 

Spec. 442,211. 

28. Williams, H. J. Phys. Rev., 1937, 52, 747. 

29. Wilson, A. J. C. Proc. Phy. Soc, 1946, 58, 21. 

30. Boothby, O. L., and Bozorth, R. M. Journ. App. Phys., 1947, 18, 

173. 

31. Yensen, T. D. Trans. Amer. I.E.E., 1920, 39, 791. 

32. Dillinger, J. F., and Bozoktii. EL -M. Physics, 1935, 6, 279. 

33. Oliver, D. A., and Shedden, J. W. Nature, 1938, 142, 209. 
Brit. Pat. Spec, 522,731. 

van Urk, A. T. Phillips Tech. Rev., 1940, 5, 29. 

34. Hoselitz, K., and McCaig, M. Proc. Phy. Doc, B, 1949, 62, 163. 
Hoselitz, K. " Ferromagnetic Properties of Metals and Alloj's," 

1952. 

35. Astbury, N. F. " Industrial Magnetic Measurements," 1952. 

36. Armour, A. M., King, A. J., and Walley, J. W. Proc. I.E.E.. 1952, 

Pt. IV, 99, No. 2, 74. 

37. B.S. 601. 1935. 

38. Campbell, A. Proc Phys. Soc, 1920, 22, 232. 

39. Brailsford, F., and Bradshaw, C. G. Nature, 1953, 178, 35. 



2. (a) SOFT MAGNETIC MATERIALS 

By 

W. S. Melville, B.Sc.(Eng.), M.I.E.E. 

Magnetically-soft materials constitute by far the major 
proportion of alloys and steels manufactured for use in electrical 
equipment. Although, by comparison with the vast quantities of 
structural steel produced, the total output of iron-based electrical 
alloys is small, nevertheless on the special properties of these 
alloys depends the economical functioning of almost all electrical 
installations. 

The two principal fields of electrical engineering in which these 
alloys find application are : 

(i) Rotating machines (both for electrical power generation 
and for motive power plants) and 

(ii) Static power transformers (which are essential to the econo- 
mical distribution and utilization of electrical energy). 

In addition, there has grown up over the past twenty years a 
smaller, but in some respects even more important, application 
in the electronics and telecommunications industries. This field 
embraces industrial -process control, nuclear-particle accelerators, 
radar, radio, television, telephony and telegraphy, metering and 
protection, small rotating machines with special characteristics 
and a large and growing number of other applications. In many 
of these the quantity of alloy per unit is a tiny fraction of that 
required by a single larger power transformer, but the number of 
units constructed may be several tens of thousands. The total 
requirement in this field is thus by no means negligible. Alloys 
used for these applications frequently require to have magnetic 
properties of the highest-available quality, or special charac- 
teristics that demand that production processes be controlled 
with very great care. 

Main Lines of Development 

There are thus two main lines of development to be examined. 
The first has as its aim the production of large quantities of 

37 



38 



SOFT MAGNETIC MATERIALS 



magnetic steel in which the cost of attaining low loss and good 
magnetic properties on a bulk manufacturing scale must be 
balanced against the long-range economic problems of the losses 
in power generation and distribution apparatus. The second is 
that in which the best possible magnetic properties for specific 
applications are required but in which, because of the compara- 
tively small quantities involved, cost of production is relatively 
u n important and can be subordinated to the over-riding considera- 
tions of characteristics and quality. 

There is a further consideration : certain alloys exhibit their 
optimum properties only when used in particular physical rela- 
tionships to the magnetic field in which they are operating. This 
imposes restrictions on their use which cannot always be recon- 
ciled with economic design of equipment. Other alloys, although 
magnetically advantageous, have mechanical characteristics 
(such as brittleness) which make it difficult to manufacture them 
in convenient shapes and therefore preclude their use in some 
kinds of apparatus. 

These considerations are ever-present in the minds of manufac- 
turers of magnetic alloys and designers of electrical equipment. 
Remarkable advances have been made in the development of 
new materials and alloys, and in adaptation and methods of 
construction to derive optimum benefit from the improved 
characteristics. 

Efforts on the scientific side, aimed at a more comprehensive 
understanding of the fundamental physical process of magnetiza- 
tion, are continually making progress. With recent advances, 
particularly by Neel, Bozorth and Williams, this important 
subject, until comparatively recently largely empirical, is 
beginning to assume the status of an exact science. The extent 
to which these basic investigations, with their laboratory back- 
ground of high purity, exact compositions and heat treatment, 
can be translated into technological practice on an industrial 
scale remains to be seen, but already the grain-oriented iron- 
silicon alloys and grain-and-domain-oriented nickel-iron alloys 
have progressed from the research stage to established production. 

Tn order to make optimum use of the materials at his disposal, 
il is necessary for the electrical engineer and magnetic component 
designer to understand and appreciate their properties and limita- 
tions. He must also take into account the economics of their 



Table 1. Properties of Ferro -magnetic Elements at Room 

Temperature 



Element 


Coercive Force 
(oersteds) 


Initial 
Permeability 


Max 
Permeability 


» Saturation 

Induction 

(gauss) 


Resistivity 
(microhm cm) 


Iron 


0-8-10 


150-300 


3C00- 
10,000 


'",,580 


9-7-11 


Nickel 





300 


2,500 


6,084 


6-3 


Cobalt 


8-12 


GO -BO 


250 


17,870 


G-2 



* For technically pure material 

Table 2. Some Properties of Commercially -a v ailaule 
Iron-n io k el Alloys 



Alloy 
Characteristic 


Approx 
7. Nt 


Commercial 
Name 


Permeability 


Saturation 
Induction 
gauss appro> 


Resistivity 
microhm . em 


Hysteresis 
Loss (B= 5000 e) 
ergs percc 
per cycle 


Initial 


Maximum 


High 

Initial Permeability 


75-80 


Mumeia! 


20,000 to 
3O.0C0 


iic.coo 


7,800 


60 


38 5 


Permalloy C 


10,000 to 
30,000 


lOCC-OC 


a, ooo 


60 


45 


High 
Saturation Density 


45-50 


Raciometal 


2000 


20,000 to 
25, COO 


16,000 


40 


218 


Permalloy B 


2,400 


:c,ooo to 


16,000 


55 


300 


High Resistivity 


35-40 


Rhome'.a! 


1000 


10,000 


12,000 


95 


458 


Permalloy 


litt to 

2000 


:ooc te 

80C0 


13,000 


90 


5 50 


Domain-Oriented 
(Rectangular loop.) 


65 


Permalloy F 


- 


200,000 to 

250,000 


14,000 


:o 


210 "» 


Grain & Domain - 
Oriented 
(Rectangular joopj 


5C 


H.C.R metal 


500 to 
1000 


200,000 to 
250,000 


16,000 


40 


570 " 



» at saturation induction 



Fig, 2-1. — Typical mag- 

NHTI/.ATION CURVES OF 
I (MIUKHCIAI.I.Y PUBS IRON, 
COBALT AND NTCKET.. 



IB 
16 

3 -2 
O 

-l 

* 10 

> 8 

E 6 


3 A 

2 
















IRON 
























l» 






































i 

-1-0. 














COB 
























































NICKEL 




7 




















/ 




s.ooo 

II 




10 


000 

I 


AT/ n 


I 
r, 15.000 

1 1 



20 40 60 80 IOO 120 140 160 180 
MAGNETIC FIELD STRENGTH - OERSTEDS 



40 



SOFT MAGNETIC MATERIALS 



manufacture and use so that equipment is designed by logical 
exploitation of the characteristics most suited to the electrical 
specification. The information given later in this chapter 
summarizes the important characteristics of commonly-used soft 
magnetic materials and indicates their fields of application. 

Basic Soft Magnetic Materials 

Before proceeding to the consideration of specific commer- 
cially-available materials, some general remarks on soft magnetic 
materials and factors affecting their properties may help to 
explain the limitations of the commercial products. 

The three basic ferromagnetic elements are iron, cobalt and 
nickel. Of these, iron is by far the cheapest and most plentiful, 
but all three have important applications as alloying elements in 
soft magnetic materials. Typical magnetization curves for 
commercially-pure samples of these elements are shown in Fig. 2-1 
and other relevant data are given in Table 1. 

Iron 

From these data it is seen that iron of high purity is in several 
respects a material with very good magnetic properties in its own 
right, and indeed when purified and refined under laboratory 
conditions its properties are among the best attained for any 
ferro -magnetic substance. For example, by prolonged refinement 
of iron in a hydrogen atmosphere at high temperature, Cioffi J has 
produced polycrystalline samples having initial permeabilities of 
20,000-30,000 and maximum permeabilities of the order of 250,000, 
the corresponding coercive force being less than 4 AT/m (0-05 
oersted). Bozorth 2 has recorded a maximum permeability of 
1,400,000 for a single crystal of similarly purified iron magnetized 
in the easy direction along the cube edge. Unfortunately these 
superior properties are attainable only when extraordinary 
measures have been adopted to minimize the proportions of 
harmful impurities ; the total percentage in the cases quoted 
being about 0-03 per cent. 

The enormous cost of producing material of this order of 
purity in commercially-useful quantities by vacuum-melting, 
high-temperature hydrogen annealing, or electrolysis, removes 
practical significance from such suggestions. It is also unfortu- 
nately true that the purification of iron to better than say 0-3 per 



SOFT MAGNETIC MATERIALS 



41 



cent, total impurities is difficult by commercially-economic 
processes and that at this figure the magnetic properties, both as 
regards loss and permeability, have deteriorated by some twenty 
to fifty t lines compared with those of technically-pure iron. 



The effects of these and 



{: m ISOO 



O (OOO 



2SO 


\ 












200 


5 












a 












1SO- 


o 












e_ 












- 100- 


O 
















CARBON 


























OXYGEN 
1 






V 



























Effect of Impurities 

The impurities principally responsible for this deterioration are 
carbon, sulphur, oxygen and nitrogen, 
other impurities on magnetic pro- 
perties have been investigated in 
great detail by Yensen 3 - 4 both in 
pure iron and iron -silicon alloys ; 
Elmen 5 has done similar work on | 
alloys of iron, cobalt and nickel. 

In general, the best magnetic 
properties, as exemplified by low 
hysteresis loss and high perme- 
ability, are associated with a strain- 
free material having a regular 
crystal matrix and large grains. 
These conditions ensure that domain 
boundary movements can take place 
freely in response to externally 
applied fields. The presence of 
impurities adversely affects this 
freedom by causing strains and 
irregularities which impede movement and inhibit grain growth. 

In solution, impurities can cause distortion of the crystal lattice 
by taking the place of atoms of the basic material ; they may 
also appear in solution at the interstices of the lattices of the basic 
material and cause severe local strains. When the impurities are 
present in quantities exceeding those supportable in solid solution, 
the formation of non-magnetic inclusions, precipitates and aggre- 
gates occurs throughout the material, and increase its magnetic 
hardness by strains due to deformation of the lattice structure in 
their vicinity. In addition such inclusions cause magnetic 
dilution of the material and reduce the effective saturation induc- 
tion ; they also decrease the effective permeability by local 
demagnetizations. 

Some of Yensen's measurements of the effects of carbon and 



-05 

PERCENTAGE 



Impurity 
Fig. 2 2.- — Effects of impurities 

ON HYSTERESIS LOSS IN IRON 
(YKNSEN). 



42 



SOFT MAGNETIC MATERIALS 



oxygen on hysteresis loss in pure iron are shown in Fig. 2-2. 
The presence of sulphur is estimated to increase the hysteresis loss 
approximately linearly with increasing content to the extent of 
about 15 J/m 3 or 150 ergs/cm 3 per cycle per 0-01 per cent, sulphur. 
This element is harmful also in that hot working properties of the 
material are seriously impaired, causing difficulties in fabrication. 
Nitrogen has the effect of inhibiting grain growth. 

MAGNETIC FIELD STRENGTH -AT/™ 
400 800 1200 1600 2000^400 4000 8000 , « 



1-fO 





c 


-<ANG 


i 

: of 


i 

SCAL 


E — » 








1 






1-8 


















































1-4<M 


ll 


?C 






















fc 


MATERIAL 


PERCENTAGE IMPURITIES 


-0 


1 


/\ 




s 0) 


C 


Mr, 


P 


S 


Si 


1 


ARMCO IRON 


■OIS 


•028 


•OOS 


■025 


•003 


H 


1 




'n 


«> 


PURE SWEDISH 
CHARCOAL IRON 


•04 


•0) 


•028 


•005 


•03 


Z 

D 
a 


LOW CARBON 
CAST STEEL 


%* 


' 8 / 


•06 


•06 


•25, 
•<35 








\ 




D 


LOW CARBON 
ROLLED STEEL 


•) 


•4 


•05 


•OS 


■25 


y. 


1 




|*S 


-t 




MILD STEEL 


-25 


•4 


•OS 


•05 


•25 




1 


























1 


















1 









10 «5 20 25 30 50 IOO 

MAGNETIC FIELD STRENGTH - OERSTEDS 



Fiy. '2 I?. — MaOKBTIZATION ii K\ B8 FOB COMMON II \<;VKTI<- STEKi.s an'ii IRONS. 

From these data, it is evident that very small percentages of 
impurities can contribute very large factors to magnetic losses in 
iron. They have similar effects in other iron-based alloys. 

Magnetic Ageing 

Impurities are also the cause of the phenomenon known as 
magnetic ageing. In materials with excessive impurities this 
manifests itself as a gradual — and in some cases catastrophic — 
increase in the losses and magnetizing current under normal 
working conditions. In the early days of electrical engineering 
before the importance of high purity in magnetic materials was 
recognized, ageing was a common cause of plant failure. The 



I 



SOFT MAGNETIC MATERIALS 



43 



phenomenon is now thought to be due to the precipitation of 
chemical compounds of iron, carbon, oxygen and nitrogen which 
are present in supersaturated solid solution in the basic material. 
The solubility of the impurities in the iron is greatly increased at 
the elevated temperatures employed in the processing of the 
material ; and on cooling, amounts of impurities in excess of the 
stable quantity are held in solid solution. This excess is gradually 
precipitated during normal life, causing severe strains in the 
crystal lattice with consequent increase in hysteresis losses and 
decrease in permeability. The effect is virtually eliminated if the 
percentage of impurities in the basic material is less than that 
which can be held in solid solution at normal working temperatures. 

Low-carbon Steels 

Modern low-carbon steels for magnetic purposes, such as those 
for which magnetization characteristics are shown in Fig. 2-3 
(curves 1,2,3 and 4), have sufficient purity to reduce the effect 
to negligible proportions, but the most effective remedy has been 
achieved by alloying the iron with silicon. The silicon acts as a 
de-oxidizing agent and chemically reacts with any oxides present 
in the basic material to form silicates which are precipitated as 
slag ; in suitable proportions it also has the effect of causing 
carbon to be precipitated as graphite aggregates instead of as 
finely-dispersed cementite which is more harmful to magnetic 
properties. The iron-silicon alloys are discussed in detail later. 

Soft Magnetic Alloys 

The foregoing discussion of the magnetic properties of simple 
unalloyed iron have shown that provided such material can be 
obtained in sufficiently pure form it has considerable potential 
usefulness. For steady-flux applications such as machine frames, 
poles, yokes, and rotors, electro -magnet poles and yokes, relay 
cores and the like in which hysteresis and eddy current losses 
are immaterial, its high permeability and high saturation flux 
density are important advantages. Its low resistivity however 
proves to be a substantial drawback for A.C. applications in which 
cyclic magnetization produces losses due to eddy currents : in 
these applications such materials are restricted to cases in which 
high efficiency is less important than low cost, or where high 
saturation is essential. 




44 



SOFT MAGNETIC MATERIALS 



Fig. 2-4. — Effect of annual- 3' 2 

I NO ON STRAFNED ARM CO * 
INOOT IRON". 1 10 



U 8 

Q 

2 6 




MAGNETIC FIELD STRENGTH - OERSTEDS 



In view of these limitations, it is to be expected that a great 
deal of work has been done to determine whether alloys exist that 
have more suitable properties. At present, only four main groups 
of binary alloys have achieved practical significance and interest. 
These are : 

Iron-silicon, 
Iron-nickel, 
Iron-cobalt and 
Iron-aluminium. 

In addition to these, there are many subsidiary groups of 
ternary and higher order alloys such as iron-cobalt-vanadium, 
nickel-iron-copper, iron-silicon-aluminium, nickel-copper-molyb- 
denum-iron, etc., that have useful soft magnetic properties for 
special purposes. Usually the extra constituents are in the 
nature of additious that improve or facilitate the achievement of 
the intrinsic characteristics of the binary alloys, e.g. by increasing 
resistivity or saturation flux density, or by improving machina- 
bility or fabrication properties. 

Commercially-available Magnetic Irons and Steels 

Fig. 2-3 gives typical magnetization curves and compositions 
for five materials which represent practical approximations to 
pure unalloyed iron and are suitable for D.C. applications. In 
all cases the material has been annealed. 



SOFT MAGNETIC MATERIALS 



45 



toe 


-25-1 












5~ 

2 
X 
O 

o 


v> 

* 

15 

> 

Z 
Id 

a 


2-0 

J 

i 


' »-^ S 
















Q- 




p/ 








> 


* IO 












h- 












a 


z 
o 

£ s 


























5 













Fig. 2-o. — Effect of 

SILICON ON MAGNETIC 
AND FHVS1CAL i'KOi'KR- 
TTES OF IRON. 



PERCENTAGE OF SILICON IN ALLOY (by WEIGHT) 

Armco iron is manufactured in U.S.A. and material of this 
quality and purity is not readily available in this country at 
present. Low-carbon steels either cast or in rolled sections or 
sheets are generally used. Annealing subsequent to fabrication — 
particularly after cold working is of great importance if the best 
magnetic properties are to be obtained from these materials. 
The effect of such heat-treatment on Armco iron previously 
strained by cold working is shown in Fig. 2-4. 



Iron-silicon Alloys 

Iron-silicon alloys were originated towards the end of the 
nineteenth century with the important researches carried out by 
Barrett, Brown and Hadfield. 6 Prior to this the bulk of the 
magnetic material used for electrical purposes was a so-called 
low-carbon-content steel. This material was generally used in 
the form of sheets, hot-rolled to thicknesses of 0-013-0-020 in. 
and it had a total loss of 2-4 W/lb. at 50 c/s and a flux density of 
I Wb/m 2 (10 kilogauss). It was found that the addition of 
2-3 per cent, of silicon to iron which was in other respects 
considerably purer than had hitherto been used had the effect of 
reducing the total loss by a factor of two or three times. This 
improvement was due in part to the higher resistivity of the iron- 
silicon alloy as compared with low-carbon steel, so reducing the 



46 



SOFT MAGNETIC MATERIALS 



SOFT MAGNETIC MATERIALS 



47 



eddy-current losses under alternating magnetization. The improve- 
ment was also due to the greater purity of the basic materials 
used, which helped to increase the permeability and reduce the 
hysteresis losses, and to the chemical reactions of silicon with 
harmful impurities which virtually eliminated ageing. 

A disadvantage for some purposes was the reduction in satura- 
tion density due to magnetic dilution as shown in Fig. 2-5 ; but 
with so many other advantages to their credit and despite their 
higher production cost, these improved alloys effectively super- 
seded low-carbon steel for many electrical purposes. Even now 
they constitute by far the biggest proportion of magnetic materials 
in power-frequency plant. 

Since these early investigations, much systematic research has 
been conducted into the factors affecting the magnetic quality of 
iron-silicon alloys. By controlled variation of the proportions of 
carbon, manganese, sulphur, phosphorous, nitrogen, oxygen and 
other impurities, and by purification of the basic iron as already 
described, a number of investigators (notably Yensen and Ciofn) 
have built up a fairly comprehensive picture of impurity effects 
on iron-silicon alloys. These data have been amassed on a 
laboratory or small-production scale. The superior qualities 
which they reveal for specific alloys and treatments cannot in 
general be attained economically in commercial production. For 
example, Goertz 7 has achieved a maximum permeability of 
3,800,000 for a single crystal of 6-6 per cent, vsilicon-iron alloy 
when magnetized along a cube-edge after heat treatment in a 
magnetic field. Alloys with compositions in this region were 
known to have very low magnetostriction so that it was to be 
expected that very good magnetic softness would be measured 
on such a sample. However, materials with such a high. percen- 
tage of silicon are practically impossible to handle on a commercial 
scale owing to their excessive brittleness and unworkability. 
Practical considerations of manufacture and fabrication limit the 
silicon content to 4£ per cent, for hot-rolled materials in most 
practical cases, although in U.S.A. alloys with up to 5i per cent, 
of silicon have been used occasionally. 

Use of Cold Reduction 

The majority of iron-silicon materials are produced by hot 
rolling from ingots to sheet, but two newer processes are now in 



use which represent important advances in the techniques of 
manufacture. Both involve cold reduction of the material to 
give a final product in the form of continuous thin strip. The 
first method does not produce any marked improvement in 
magnetic properties as compared with hot-rolled sheet : but the 
strip has a better surface finish giving improved space factor when 
assembled as laminations ; its freedom from surface scale- causes 
less wear on tools ; it is suitable for feeding continuous automatic 
presses ; and it allows of improved working conditions for 
personnel in the rolling mills. 



Grain-orientation 

The second method is of greater importance as regards magnetic 
properties. Cold -reduction processes on iron-silicon alloys produce 
a material having a grain-orientation that gives magnetic pro- 
perties, in the direction of rolling of the strip, similar to those 
exhibited by single crystals when magnetized along a direction of 
easy magnetization. 

The pronounced magnetic anisotropy of a single crystal of 
silicon-iron has been recorded by Williams 8 and is illustrated in 
Fig. 2-6. The crystal lattice is a body-centred cube and magneti- 
zation along a cube edge (the easy direction) can be carried out 
without domain rotation — the domains being already aligned in 
this direction or in directions at 90° or 180° to it. This gives 
magnetic properties outstandingly better than polycrystalline 
non-oriented hot-rolled silicon-iron. 

In 1 935 9 Goss described a method of producing strip silicon-iron 
by which magnetic characteristics similar to those obtained for 
single crystals magnetized in the cube-edge direction could be 
achieved. It involved the production of hot -rolled strip followed 
by cold reduction of about 60 per cent., an intermediate anneal 
and subsequent cold reduction of a further 60 per cent, to final 
thickness followed by a final high-temperature anneal to induce 
re-crystallization and grain growth with preferred orientation. 
The method has been perfected by various workers as regards 
amounts of cold reduction and details of heat treatments, and the 
importance of controlled purity in obtaining a high degree of 
preferred orientation has been established. 

The process has not so far been applied commercially to 



48 



SOFT MAGNETIC MATERIALS 




Fig, 2-6. — Magnetization directions in ghain-oiuented iron-silicon alloy 

STKIP. 



MAGNETIZING FORCE — OERSTEDS 
•3 -4 -S -6 -7 




O 10 20 30 40 50 60 70 60 90 100 

MAGNETIZING FORCE, AMPERE - TURN5/METRE 

Fig. 2—7. — Magnetization characteristics of various iron-silicon alloy 

STRUCTURES. 



Curve 

No. 


i i ysful 

Structure 


Direction of 
Magnetization 


Tti'ilncl.ion 

Treatment 


Pet cent. 

Sili. 


I 


Single nysl.nl 


Cube edge 


— 


3-8 


o 


BiBgle crystal 


Face diagonal 


- 


3-8 


3 


Single crystal 


('ill e diagonal 


- 


3-8 


4 


Polyi-ry^tiilliin- 


Direct ion of rolling 


Cold rolled 


8-0 


5 


Polycrystalline 


Acrose direction of rolling 


Cold rolled 


3-0 


6 


Polyrryst alUne 


Direction of rolling 


Sol rolled 


•Hi 



SOFT MAGNETIC MATERIALS 



49 



materials containing more than about 3i per cent, of silicon 
because of the difficulty of working them cold. 

The process results in gram orientation of the kind illustrated 
in Fig. 2-6. The plane of the rolled strip contains a direction of 
easy magnetization along a cube edge in the direction of rolling 
of the strip, i.e. along its length. A magnetization characteristic 
for such a material and magnetization is shown in Fig. 2-7, 
together with a characteristic when magnetized at right angles 



Fig. 2-8. — Hysteresis loops of a 

6-5 PER CENT. SILTCON-IKOn 

.single crystal before and 
after mag n et 1 anneal 
(goertz). 



20 


MAGNETIC riELD STRENGTH- 
>0 25 2S 


AT/m 


5 I° - ? 


1 
















(S 

10 

s 



















1 
















1 














If 


— MAGNETIC 
ANNEAL 








fi 










10 

15 








1 


— 


MAGNETIC 
ANNEAL 








l 


t 












] , 


/y 






1 




.1 




', 


" 



o-e 0-4 0-4 0-2 o 0-2 0-4 o* 

MAGNETIC FIELD STRENGTH - OERSTEDS 



to the direction of rolling, i.e. across the strip or in the direction 
of the cube-face diagonals. This is a direction of less-easy mag- 
netization and inferior properties result. 



Magnetic Annealing 

Two further recent lines of development that may lead to 
improvements in the quality of iron-silicon alloys can be men- 
tioned. One has already been noted 7 — -annealing in a magnetic 
field. This is most effective with silicon contents of the order of 
0-4 per cent., and maximum permeabilities of about 200,000 are 
reported for pure poly crystalline specimens of such composition. 
The effect of magnetic annealing on the shape of the B-H loop 
for a 6-5 per cent, silicon single crystal is also noteworthy for its 
marked " rectangularity," Fig. 2-8. 



50 



SOFT MAONKT1C MATERIALS 



SOFT MAGNETIC MATERIALS 



51 



Single-crystal Strip 

lb The other development concerns the possibility of producing 
continuous strip which is in .effect a single crystal. A laboratory 
method has been described by Dunn. 9, 10 

A strip of high purity and controlled composition, with a fine- 
grained polycrystalline struct tire to support grain growth, is used. 
A crystal of the preferred orientation is selected and isolated 
and crystal growth is propagated from this seed crystal by 
drawing the strip through a furnace in which there is an abrupt 



MAGNETIC FIELD STRENGTH - AT/m 

2,000 4,000 6,000 





















AGEC 


600 


HOURS 


AT IC 


o°c 




s 


V 















































































PERCENTAGE OF SILICON IN ALLOY 



fit/. 2 '.). effect of silicon ox magnetic ac f.i no ok iron-silicon alloy. 

oivixg the percentage chance in total loss at # = 1-3 wb/m 2 (13,000 

gauss) after ageing. 

high-temperature gradient. The rate at which the strip passes 
through the gradient is made such that continuous crystal growth 
takes place. Crystals of any chosen orientation can be propagated 
by this method and it is therefore possible to have au orientation 
in which two easy directions of magnetization mutually at right 
angles lie in the plane of the strip. Maximum permeabilities of 
the order of 200,000 have been achieved in laminations cut with 
their legs parallel to the easy directions. 

Investigations and results of such a kind may seem remote 
froin the commercially-available materials now to be described. 
Nevertheless, it is on such researches that future progress depends ; 
they are invaluable in indicating the relative importance of the 
many factors affecting magnetic quality and in guiding magnetic 
material manufacturers towards more satisfactory products. 






16 



Fig. 2-10. 

N BXIZ A T I O V 
CHARACTER!- O 
S T 1 C S FOR 5 12 
VARIOUS COM- ' 

POSITIONS OF 

NON-OKI ENTKI) 
IRON-SILICON 
ALLOYS. 



Mao- § 

< 14 



E io 



I RteAord Tfuma$ tb 

Hahh'iiis /.W.I 



rf 6 




1-0 t 



123456789 
MAGNETIC FIELD STRENGTH - OERSTEDS 



COMMERCIALLY-AVAILABLE IRON-SILICON ALLOYS 

Hot-rolled Non-oriented Alloys 

The importance of minimizing impurities has been discussed. 
Scrap used for magnetic alloys is therefore selected to be as free 
as possible from harmful ingredients. The raw materials are 
usually melted and refined in an electric arc furnace and cast into 
large ingots or slabs, which are soaked in pits at high temperature 









DYNAMO 


GRADES 








TRANSFORMER 


GRADES 


3-0 
































z 










15 KILOGAUSS 
















a 






























M 










13 KILOGAUSS 
| | 










\l5 KILOGAUSS 


< 

9 










r — ' — ' — 

IOKILOGAUSS 










L ' i 

■^13 KILOGAUSS 
I i 


























1 1 . 



PERCENTAGE OF SILICON IN IRON SILICON ALLOY 

Fig. 2-1 I. - EfJ'KCT of silicon ON TOTAL losses of non-oriented IRON-SILICON 



52 



SOFT MAGNETIC MATERIALS 



SOFT MAGNETIC MATERIALS 



53 



and subsequently passed 
through blooming mills and 
reduced to thick sheets. The 
sheets are folded and refolded 
with intermediate reheats 
and passes and hot-rolled as 
a pack of eight sheets to the 
final thickness of 0-014-0-030 
in. ; intermediate reheats may 
take place during pack roll- 
ing. Finally the pack is 
sheared, the sheets then sepa- 
rated and annealed in a con- 
trolled atmosphere to release 
strains and promote grain 
growth. As has already been 
mentioned, the best magnetic 
properties can only be 
achieved in a strain-free 
material and any subsequent 
mechanical strain induced in 
fabrication processes such as 
stamping or shearing should 
be relieved by further anneal- 
ing in a protective atmos- 
phere at 700°-800°C. fol- 
lowed by slow cooling. 

Various grades and thick- 
nesses of material are pro- 
duced in bulk covering the 
range of percentages of silicon 
between 0-3 and 4\ per cent. 
The intermediate grades are 
not widely used but are neces- 
sary to optimize specific 
apparatus design ; most of 
the material supplied is either 
in the region of 0-3-0-5 or 3-5-4 per cent, silicon. The reasons 
for this are discussed later in considering applications of the 
materials. Some typical characteristics for the range of com- 



40 
Q 

z 

lao 

a. 

Ml 

a. 
m 

* 
l-O 






































IS KILOGAUSS 
























13 KILOGAUSS 
























lO KILOGAUSS 
















■o 


2 -014 '016 -018 020 022 025 
(a) 0-4°/o SILICON 


30 

z 
z> 
o 

a. 

t£ 

£2-0 

in 

1- 

* 

l-O 
























IS Kl 


LOGA 


JSS 
























13 Kl 


LOGAl 


SS 


























•o 


2 014 016 -OIB -020 -022 025 
(b) 1-7°/o SILICON 


o 

§*o 

a 

a 
m 
a. 

in 

ti-o 

< 

* 
























15 Kl 


.OGAV 


SS 










13 KILOGAL 


SS 










10 KILOGAUSS 


•O 


2 -014 -OI6 -OI8 -020 -022 -025 
THICKNESS - INCHES 
(C) 3-4% SILICON 



Fig. 2 12. -EFFECT OF THICKNESS ON 
TOTAL LOSSES OF NON-ORIENTED IKON- 
SILICON AI,I,OYS AT VARIOUS I'l.l X 
DKNSITIKS. 



Fig. 2-13. — Effect of 

ANNEALING AFTER CUTTING 
ON TRANSFORMER CRADE 
HOT-ROLLRD IKON-SI I. ICON 
AI..I.OY. 

(J. Sankey &■ Sons, Ltd.) 









































/ 


















J 


/ 




STALLOY 


■014" 


THICK 




> 


// 




























AS 


CUT 


.< 


'A 

RE- 












<J 


<^ 
























0-6 


0;8 


1 





Wb 

1-2 


1 


4 



MAXIMUM FLUX DENSITY - KILOGAUSS 



mercially-available iron-silicon alloys are shown in Figs. 2-9 to 
2-16. 

Fig. 2-9 shows the beneficial effect on magnetic ageing of 
increasing silicon content. Agoing is seen to be virtually elimi- 
nated for transformer grades of sheet above 3 per cent, silicon. 

Fig. 2-10 shows magnetization characteristics covering the 
range of commercially-available non-oriented alloys. The 
improved properties of the higher silicon alloys below about 
1-2 Wb/m 2 (12 kilogauss) and the effect of magnetic dilution in 
km hieing the saturation flux density are clearly visible. 

Kg. 2-1 I shows the variation in total loss with silicon content 
for both dynamo and transformer grades for three values of 
maximum flux density. 



Fig, 2 14. Kfff.ct of 

FREQUENCY ON TOTAL 
LOSSES IN TRANS- 

FORMER GRADE JRON- 
SU.KDN ALLOY. 

(J. Sankey <fc Sons, Ltd.) 



Ill 




















I 




















1 




















•OI4" STALLOY 




1 




WATTS PER POUND 










1 


2 | 3 




\\ 


\ J 


y 


/ 


/ 
















/ 


/ 



























































1-0 $ 

I 



>.* 



400 800 1200 1600 

FREQUENCY- CYCLES PER SECONO 



54 



SOFT MAGNETIC MATERIALS 



MAXIMUM FLUX DENSITY - Wh/mZ 

0-1 0-2 0-3 



Fig. 2-12 gives the variation in total loss with thickness for 
three values of maximum flux density and relate respectively to 
0-4, 1«7 and 3*4 per cent, silicon-content alloys. 

Fig. 2-13 shows the improvement in properties resulting from 
annealing transformer-grade alloy after cutting. 

All the properties except those in Fig. 2-10 relate to magnetiza- 
tion by a sine wave variation 
of flux at 50 c/s. For normal 
power frequency working the 
dynamo grades are normal!}' 
used in sheets about 0-016 in. 
thick, while the transformer 
grades axe usually 0-014 in. 
thick. Thinner material, down 
to 0-007 in. . is made for higher 
frequency applications, but 
is difficult to produce and its 
properties arc generally in- 
ferior to those of cold-reduced 
grain -oriented material or 
some of the nickel-iron alloys. 
Fig. 2-14 shows the effect of 
frequency and flux density on 
total losses in 0-014-in. thick 
transformer grade non- 
oriented iron-silicon alloy for 
three, typical working values 
of total loss. 

Fig. 2-15 shows the effect of lamination thickness on losses at 
three high frequencies. The 0-014-in. thick sample is of 3-9 per 
cent, silicon, while the 0-010 and 0-007-in. samples are of 3-6 per 
cent, silicon. In general, it is not possible to roll alloys of the 
highest silicon content to these small thicknesses, and the curves 
show that the properties of the 3-9 per cent., 0-014-in. samples are 
somewhat better than Ihe 0-010-in. sample at 2,000 c/s. The 
effect of smaller thickness in reducing eddy current losses is shown 
to be very considerable at the higher frequencies. 

Transformer-grade materials are frequently used with combined 
A.C. and D.C. excitations and Figs. 2-16 (a) and (b) show the 
effect of a polarizing field on incremental permeability for various 







0,000 


c/s 


5000 c/s 








1 


// 


' 










/ 












V 


1 






1 






A 

\ -00 

Vot 


7" 

3 "\j4 


\ 200 


3 C/S 


■«V4-^ 













IO 
MAXIMUM FLUX 



20 3-0 

DENSITY - KILOGAUSS 



Fiff, 2 15. -El'-[>'K< II III.- LAMINATION" THICK- 
NESS ON LOSSKS IN TRANSFORMER 
GRADE mON-STT.TCON ALLOY AT HIIIH 

PMSQTJTENOIES. 

(Richard Thcmw .(• Baldwin* Ltd.) 






S ( > !■' T M A G X EC TIC M A T EEIALS 

POLARISATION FIELD STRENGTH- AT/m 
80 <60 240 320 



w 2000 



O IOOO 




4000 



O 1 2 3 4 

DC. POLARISATION FIELD STRENGTH - OERSTEDS 

(a) 

POLARISATION FIELD STRENGTH-AT/m 
BO <GO 240 320 







I 1 

4-1 8 /o TRANSFORMER GRADE 
HOT-ROLLED OI4* THK. 






ALTERNATING EXCITATION: 
400 CYCLES PER SECOND 


























400 CYCLES PER SECOND 
















ALTERNATING 
EXCITATION 








~^^ 


(GAUSS) 


IOOO 
^>^.IOO 











55 



a iooo 



O I 2 3 4 

D.C. POLARISATION FIELD STRENGTH - OERSTEDS 

<k>) 

Fig. 2-16. — Ini i;i:mi:\ rvi permeability of transformer crade iron-sii icun 

ALLOY AT 60 CYCLES PER SECOND AND 400 CYCT.ES PER SECOND. 

(lUchard Thomat .1- SeMwhu lid). 



. 




50 



SOFT MAGNETIC MATERIALS 



SOFT MAGNETIC MATERIALS 



57 





PHECUT STRIP SUB-ASS-EMBLY 




Fig. 2- 1 7 (left). -Section through toans- 

RimiKl! ASSEMBLY WITH PREFORMED 
WOUND AND CUT ORIENTED -STU11' CORE 

(TUUNS'I'INCIL). 

Strip size enlarged. 



Fig. 2 18 (beloic). —Frkcvt and PRE- 
FORMED "i;ii:nti:i>-strii' CORE tiians- 
ron.MKU ASSEMBLY. 



INTERLEAVED JOINTS 




(NOT TO 
SCALE) 




PREFORMED AND ANNEALED CORE SUB-UNIT 



4 CORE UNITS 



CORE 
JOINTS 



CROSS SECTION THROUGH CORE UNIT 



CORE UNIT 





CORE AND COILS ASSEMBLY 






Fig. 2 -19. Tvi'K \i. 

MAGNETIZATION CJIAI!- 
ACTEE1STICS OF ORIEN- 
TED IRON-SILICON 
ALLOY. 

(Hi'-hiti-ii Thomtu ( f: 
Baldwin* Ltd. ) 



/ 








m. 






































(b) 




\ 










1-4 1M 


// 


/\ 






















r 
















I 

> 
1-0 E 

10 

z 
9 




t — 




9EST TRANSFORMER GRADE 
4% Si NON-ORIENTEO 
3°/o Si ; ORIENTED ALPHASI 






/ 
/ 






L 




o-e d 


! 






SO CYCLES PER SECOND 


























0-2 


100 


200 


30( 


400 


soo 


800 


700 AT/m 



(a) MAGNETIC FIELD STRENGTH - OERSTEDS 
fb) VOLT -AMP PER POUND 



CROSS SECTION OF ASSEMBLY OF 4 CORE 
THROUGH CENTRE LEG 



amplitudes of alternating excitations at 50 and 400 c/s respec- 
tively. 

Cold-reduced, Non-oriented Alloys 

These materials have magnetic characteristics similar to those 
described for the hot-rolled alloys ; they have the advantages 
already mentioned and do not require separate description. 
Some manufacturers reserve the right to supply hot- or cold -rolled 
material as alternatives. 

Cold-reduced Oriented Alloys 

Great progress has been made in America during the past 
decade in the commercial production of cold-reduced iron-silicon 
alloys, and more recently a material has been produced in this 
country of comparable quality. The materials contain between 
:J and 3^ per cent, silicon and their production and properties have 
already been briefly described To make best use of the superior 
characteristics it is necessary to magnetize them in the direction 
of rolling of the Btrip : magnetic circuits employing such materials 
have to be designed accordingly. The material is not normally 
applicable to rotating machinery, for directional properties are 
here a disadvantage. It has been used, however, in yokes for 
machines such as largo turbo alternators. The laminations are 
arranged as narrow circumferential segments in which the 



58 



SOFT MAGNETIC MATERIALS 



FLUX DENSITY-Wb/m' 

i-a i*4 i-6 







1 
i 






/ 


II 










i 






I 


y 


/ 










i 
i 






1 


// 










/ 


r 






7 


/ 










/ 

/ 








V 


















// 


f 


















/ 
































Y/ 


/— 


— 


BEST 
4°/o S 


TRANSFORMER GRADE : 
i NON -ORIENTED 










3°/o ! 
SO C 


' . oi 

fCLES 


IEN1 
PEF 


ED ALPHASIL 
SECOND | 





Fig. 2-20. — Typical total-loss 

CURVES FOR TITRF.E (II1AUKS OF 
ORIENTED IRON-SILICON ALLOY. 



FLUX DENSITY KILOGAUSS 



direction of rolling of the steel, and thus the direction of minimum 
loss, is arranged to be tangential to the bore of the stator. This 
permits either a reduction in the core loss occurring behind the 
stator teeth, or a reduction in the outside diameter of the machine 
with the same loss as would be obtained from hot-rolled non- 
oriented material. In large transformers and the like, cores are 
normally built up from plates or strips, the directional properties 
fit in conveniently with conventional manufacturing practice, 
and reduction in size and loss of about 30 per cent, as compared 
with non-oriented material is quite usual. 

For smaller power transformers up to about 4,000 kVA new 
manufacturing techniques have been developed — principally in 
U.S.A. — to take advantage of the properties of oriented materials. 
Some of these techniques have been described by Birnstingl. 11 
Two main methods are used. In one the core is preformed on a 
rectangular mandrel and annealed. After annealing the core 
retains its shape and the strip is cut through every other turn to 
give an interleaved joint pattern. The individual laminations are 
then threaded on to the coil assemblies where they readily take 
up their original preformed shape as shown in Fig. 2-17. This 
method is used for transformers up to about 600 kVA above 
which size threading of the laminations becomes difficult and 
may cause straining of the core material. 









SOFT MAGNETIC MATERIALS 



59 



'E'CORE 




(a)'C'CORE TRANSFORMER ASSEMBLY 



Fig. 2 21 (above). 
— O R I B N X l : I > 

STRIP CORE 

CONSTHI JCTIONS 



Fig. 2-22 (right). 
— Magnetiza- 
tion OHARAC- 

TERIST1CS FOR 
'" C " CORES OF 
ORIENTED IRON- 
SILICON ALLOY 
STRIP 0*013 IX. 
THICK. 

(English Elirlrir 

Co., Ltd.) 



(b) 'E'CORE ASSEMBLY FOR THREE 
PHASE TRANSFORMER 




0-2 0-4 0-6 0-8 1-0 

SCALE A: MAGNETIC FIELD STRENGTH - OERSTEDS 
SCALE B; TOTAL LOSSES - WATTS PER POUND 



SCALE C I MAGNETIC FIELD STRENGTH - OERSTEDS 
SCALED MAGNETISING VOLT-AMP PER POUND 































"\ 










C CORES 


004* 


THICK 






< 
O 




\ 






















2 


V\ 


A 






















>- 


ji 


V 






















a 




s \ 


, \ 














































la. 


WATTS 
PER POUND 





























1 
2 
























* 



lOOO 2000 3000 

hREOUENCY - CYCLES 



4000 sooo 

PER SECOND 



Fig. 2-23 (left).— 
Curves of varia- 
tion of total loss 

with frequency 
and flux density 
for " c " cores of 
ouiknted iron- 
silk "\ alloy 

STRIP 0-004 IN. 

THICK. 
(JSnjli$h Electric Co.. Ltd.) 



B 2 



60 



SOFT MAGNETIC MATK1UALS 





































'c* 


CORE -002"THICK 














































v^- 
































































^ 




/PULSE 
DURATION 




■o-s— 
















































[MICROSECONDS) 




- 2 




— -O — 

■0-2 - 


























- ( 




















>- 




- O-B 




- 0-25 














K^ 


X 




























































1 



IOO 200 500 IOOO 2000 SOOO 

TOTAL PULSE LOSSES - ERGS PER CM 3 PER PULSE 



(a) PULSE MAGNETISATION LOSSES 

1,000 2,000 AT/m 



2 



'C* CORE -002" THICK 








A 

1-0 - 
0-8 - 
0-6 
0-4- 
// 






















' PULSE DURATION 


(microseconds ) 












2 












1 


O -5 










O • 25 










ty 

















"■ O 4 8 12 16 20 24 28 

MAXIMUM PULSE MAGNETISING FORCE -(ft M ) 

OERSTEDS 

(b) PULSE MAGNETISATION CHARACTERISTICS 

Fig. 2-24. — : Pui.se magxetizatiun hkoi-kutiks <>k " c •" cores of otuextkh 

1RON-STT.ICOX AT.T.OY STRIP 0*002 IN. Tllll K. 

{English Electric Co., I.I.I.) 



The other method is used for large assemblies : steps in the 
process are illustrated in Fig. 2 18. The largest lamination is 
cut to length and bent into a circle, the position of the joint being 
located by dowel pins and holes. Further laminations are laid 
within the first, the lengths being progressively varied and the 
locating holes staggered to give interleaved joints when assembled 
on the locating pins. The complete build-up is then pressed on 



SOFT MAGNETIC MATERIALS 



61 



Fig. 2-25. — Ma(!NKTi- o )0 

ZATTOX CHAKA< TKJ.US- 5 

TICS FOR E- AND 1 - * - 

I.AM1XATTONS 01 h 

OK1EXTED TROX-SII.I- !£ ft 

COX ALLOY 0-004 IN. g 

THICK AT 500 t 8. x 4 

(Enylisl, Elect, ic Co. I.ltt.) Z 



'E' i V LAMINATIONS: -004"THICK 
FREOUENCY r 500 CPS j 












6 
a 
S 
1-0 

) 












<* 








WAC- Nt "" | 


■<&> 


■£> 


v*» 


























7 
























r 

7 


/ 


y 



































































SCALE A : TOTAL LOSSES - WATTS PER POUND 



40 60 80 IOO 

MAGNETISING VOLT-AMP PER POUND 



to a steel arbor and annealed in situ. Several strip widths may 
be used to approximate to a cylindrical central leg, and in larger 
units four such cores are assembled in a cylindrical winding. 
The making of preformed cores has been mechanized so that the 
complete process of core fabrication including insulating the 
laminations is carried out without handling. 

Table 3, taken from BirnstingFs paper, compares the per- 
formance of three designs for a typical distribution transformer 
rating under conditions appropriate to the optimum working 
conditions for three methods of core construction. The advan- 
tages, both as regards weight and power saving, of the pre- 
formed design arc apparent. 



Table 3. Comparison of Transformer PERFORMANCE for 
Various Forms op Iron-silicon Alloy Cores (Blrnsttngl) 



K;timatcd performance Ik'ures fur l,i)ui)-kVA 33/0-44 kV 50-6/8 distribution transformer 



Type of core 


i-'iii\ density 


Magnetizing force 


Iron 

Ion 
Watte/lb. 


Relative 

core 
Weights 


Relative 
iron 
loss 


K i i ■ . I i . s, 
ln.= 


YVb lie 


AT/in. 

(R.M.S.) 


AT/.u 
ili.M.S i 


Non-oriented sheet . 

Oriented sheet 

Oriented preformed Btrip • 


85 
90 

Kill 


IS 

I -4 
1 -85 


5-5 
2-5 

2-7 


220 
100 

no 


1-0 

((•8 

(IS 


100 

08 

78 


100 
75 
63 



62 



SOFT MAGNETIC MATERIALS 



Fig. 2-19 shows typical A.C. magnetization and volt-ampere 
characteristics for the best quality of oriented strip at present 
commercially produced in Britain. For comparison, an A.C. 
magnetization characteristic for the best grade of 4 per cent, non- 
oriented silicon alloy is also shown. Fig. 2-20 shows typical total 
loss curves for similar oriented materials and also for the non- 
oriented alloy. The superior properties, both as regards losses 
and ability to operate at high flux density without undue mag- 
netizing current, are apparent. 

C-core Construction 

For miniature transformers a range of cores has been developed 
which consist of two solidified C-shaped units. In one method 
these are made by winding continuous oriented strip on to a 
rectangular mandrel ; the core is annealed to relieve strains duo 
to the winding process and solidified with a suitable impregnant. 
The solid core is then cut into two C-shaped pieces, the cut 
surfaces being ground and etched to ensure minimum air gap 
and to remove cutting burrs. The coil assembly is fitted between 
the halves of the core which are held together by a metal strap, as 
shown in Fig. 2-21 (a). A range of core assemblies suitable for 
use in three-phase components has recently been developed. The 
method of construction is indicated in Fig. 2-21 (b). 

The presence of two airgaps in the magnetic circuit increases 
the magnetizing volt-amperes compared with corresponding 
uncut spiral cores, but the losses are unaffected and the C-core 
construction allows the superior properties of grain-oriented 
material to be utilized in a convenient maimer, with conventional 
windings. Standardization and ease of assembly are important 
advantages of the construction. 

Core assemblies of these kinds are produced by several manu- 
facturers in Britain using strip from the U.S.A. (Hipersil) and 
British materials such as Crystalloy and Alphasil. Strip is avail- 
able in three ranges of thickness : 0-012-0-014 in., 0-004-0-005 in. 
and 0-002 in. The material is also manufactured in U.S.A. down 
to 0-001 in. or less. For detailed information on the characteristics 
of C-cores made up from these materials, the reader is referred to 
the manufacturers' handbooks and pamphlets. 12 Some typical 
characteristics useful in application of the three thickness ranges 
are given in Figs. 2-22, 2-23 and 2-24. 












SOFT MAGNETIC MATERIALS 



63 



In general, these oriented steels are unsuitable for the produc- 
tion of conventional stampings for miniature transformers 
because of their inferior properties across the direction of rolling. 
Nevertheless, the smallest sizes of C-cores are difficult to manu- 
facture with consistent characteristics and it has been found 
advantageous to produce a small range of E- and I-laminations 
0-004 in. thick for use with power supplies of higher frequencies. 
Typical characteristics for these laminations are shown in 
Fig. 2-25. 

IRON-NICKEL ALLOYS 

The iron-nickel group of alloys contains several of the most 
important soft-magnetic materials that are used for special 
applications. Compared with the iron-silicon alloys they are 
relatively expensive, but their special characteristics allow 
magnetic-component designers to achieve performances which 
would be impossible with orthodox materials. In general, they 
are of particular value in applications requiring very low losses 
in association with high permeabilities — both initial and maxi- 
mum — but at comparatively low flux densities. Other alloys 
have constant permeability, rectangular hysteresis loops or other 
special properties. 

Binary Alloys of Nickel and Iron 

Although there had been earlier work on alloys of iron-nickel 
composition, the initial extensive researches on binary alloys of 
these elements were carried out by Yensen (1920) 13 and Elmen 
(101.3-23). 14 These investigators oxplored the whole range of 
compositions and were responsible for segregating the magneti- 
cally-valuable regions ; in particular, Yensen did important work 
on compositions around 50 per cent, nickel which, as shown in 
Fig. 2-26, is a region of relatively high saturation flux density, 
while Elmen — latterly in association with Arnold — found and 
developed the high permeability region around 75-80 per cent, 
nickel and established the vital influence of correct heat treatment 
in optimizing magnetic properties. In this work they showed that 
by rapid and accurately-controlled quenching of these alloys after 
annealing, losses were reduced and initial permeability increased 
by factors of 3 or 4 over slowly cooled material. This treatment 



(54 



SOFT MAGNET rr MATERIALS 



SOFT MAGNETIC MATERIALS 



65 



was called the Permalloy treatment and the alloys — Permalloys. 
The effect was most marked in the region 78-5 per cent, nickel, 
although a smaller maximum was also evident at about 45 per 
cent, nickel ; it is in these two regions that the majority of 
commercially available alloys he although these arc not always of 
pure binary composition. 

High-resistivity Alloys 

In addition to the two major groups of binary alloys just 
mentioned, there are other compositions with special properties. 
Fig. 2-26 shows a maximum in the resistivity curve at about 
35 per cent, nickel and alloys have been developed to exploit this 
feature since the increased resistivity is advantageous in high 
frequency applications even although it is associated with lower 
permeability. 

Temperature-sensitive Alloys 

Alloys with about 30 per cent, nickel have a low Curie point of 
the order of 100° C, and consequently show a marked reduction 
in magnetization as temperature rises towards this value. Among 
other applications, this characteristic can be utilized for flux 
compensation in permanent and electro -magnet systems in which 
the field strength in the air gap tends to reduce with increasing 
temperature. A strip of this alloy shunting the gap diverts less 
flux as the temperature increases and can be arranged to maintain 
a constant flux in the air gap. 

Grain-oriented Alloys 

The development of a preferred grain-orientation in the 
direction of rolling of iron-silicon alloys has already been described, 
and it is interesting to note that a similar phenomenon had been 
reported by Smith. Garnett and Randall 15 for cold-rolled nickel- 
iron alloys as early as 1930. The effect is most pronounced in the 
region of 50 per cent, nickel, and processes have been developed 
by which strip of about this composition can be produced which 
has properties similar to those of a single crystal magnetized 
along an easy direction of magnetization. Preferred grain- 
orientation is achieved by ensuring high purity in the basic 
materials and by carefully controlled cold -reduction and heat 
treatment of the rolled strip. This results in grain -orientation of 









Fig. 2-26. — BlTECT OF COM- 
POSITION ON MAGNETIC 

AM' ki.kctrtcai, PROPER- 

TIKS OF TRON-NH KKL 

AI.I-OVS (KLUEN). 




fig. 2-27.— MAG- 
NETIZATION DI- 
RECTIONS I N 

B K C TA N G O L A It 
STAMPINGS (IN- 
TER I. E A V K l) ) 

MADE FROM 
cKAIN - ORIENTED 
60 60 IRON- 
NICKEL A I, I. O Y 
STRIP. 



13 

s • 

< 


























= 










> - 

2 4 

i 8 

12 


2 






2 



























r 

2-1 » 



(~a\ CHAIN -OBlENTEO (b)OB»IN AND DOMAIN 

V ' OKIENTEO 

MAGNETISING FOBCE- OERSTEDS 

Fig. 2-28.- Hysteresis loops fob 

OUAJN-OTUKNTKD AND GRAIN- 
AND-DOMATN-OKI KM'I'KiJ SO/50 

JKON-NUKJM. AI.I.OY. 

{Telegraph Construction .(• if atnttnana Co. Md.) 



























-= i 


1 ' 



, 
^ - 

Z 4 

.- 1 
12 


PEHMALLOY F 




























/I WITHOUT 
/MAGNETIC FIELD 


- 


3 




/ . 2 3 * 








ANNEALED 
« 1- WITH 










1 








, 




















Fig. 2-29. — Effect of heat 

TRKATMKNT TV A MAGNETIC 
FIELD 08 TIIK IIYSTKKKS1S 

LOOP OK PERMALLOY F 

(BOZOBTH and di lunger). 



66 



SOFT MAGNETIC MATERIALS 



the kind shown in Fig. 2-27. It is seen that the grains have two 
directions of easy magnetization in the plane of the strip surface ; 
one in the direction of rolling, the other at right angles to it. i.e. 
across the strip. 

Magnetization takes place by 90° or 180° domain re-orientations 
a I low fields and losses associated with domain rotation at high 
fields are almost completely eliminated, resulting in a material of 
high permeability and low-hysteresis loss with a rectangular 
hysteresis loop as shown in Fig. 2-28 (a). A material of this kind 
can clearly be used in the form of rectangular punchings since 
orthogonal directions of easy magnetization lie in the plane of the 
strip. To minimize the effects of airgaps the punchings have wide 
yokes and are interleaved as shown in Fig. 2-27. 

Domain-oriented Alloys 

It was found by Bozorth and Dil linger 16 that rectangular 
hysteresis loops could be produced in binary alloys with 60 to 
80 per cent, nickel, if these were cooled from above the Curie 
temperature ill the presence of a magnetic field. The rectangular 
loop characteristic was developed in tho directions parallel and 
anti-parallel to the applied field and the effect was most pro- 
nounced in the region of high magnetostriction and high Curie 
temperature, i.e. about 65 per cent, nickel. Fig. 2-29 illustrates 
it for a material of this composition. 

Grain- and domain-oriented Alloys 

The properties of the grain-oriented materials may be improved 
still further by eliminating 90 per cent, domain re-orientations 
which require more energy than 180° reversals. This is done in 
a similar manner to that described for domain oriented materials 
by annealing grain-oriented strip in a magnetic field which is 
coaxial with the direction of rolling, and cooling in the presence 
of the field, through the Curie point temperature. This process 
leaves all the domains aligned in the easy direction nearest to the 
applied field. Magnetization processes then take place solely by 
180° reversals. A very high degree of rectangularity and low 
coercive force and hysteresis loss result from this treatment as 
shown in Fig. 2-28 (b). 

The same practical limitations arise in the use of these and the 
domain-oriented materials as were mentioned for grain-oriented 



SOFT MAGNETIC MATERIALS 



67 



DIRECTION OF EASY 
MAGNETISATION 



/'7r/. 2-30. -MAGNETIZA- 
TION DIRECTIONS IN 
GRAIN- AND DOM A IN - 
ORIENTED 50.' 50 

TRON -NICKEL ALLOY 
STRIP. 




Fig. 2-31 (a). — Variation ok total 

LOSSES WITH FI.UX DENSITY IN 
I'KKMALLOY " B " FOR VARIOUS HiK- 
QTJENCIES AND THICKNESSES. 

(Standard Telephones & Cables Lid.) 



Fig. 2-31 (6). Variation o» TOTAL 

LOSSES WITH FLL'X DENSITY IN 
PEKMA.Ll,OY " (-' " FOR VARIOUS FRE- 
QUENCIES AN' I) THICKNESSES. 

(Standard Telephones & Cables Ltd.) 



88 SOFT MAGNETIC MATERIALS 

iron-silicon alloys. Since there is only one direction of easy 
magnetization as shown in Fig. 2-30, the best properties can be 
obtained only in the direction of rolling. 

Ternary and Higher Order Alloys 

There arc relatively few commercially-available alloys of simple 
binary composition ; the majority of the materials which have 
compositions in the regions of the two important binary groups 
arc manufactured with one or more additives — usually in minor 
proportions. The purpose of these is generally to decrease the 
sensitivity of the alloys to heat treatment so that the requirement 
for closely-controlled processes such as the Permalloy one 
described by Elmen is. to some extent, relaxed ; and also with 
certain compositions, to increase the resistivity while retaining or 
i nil (loving tne desirable properties of the binary alloys. A 
disadvantage of the alloying process is that it is usually accom- 
panied by a reduction in saturation density. 

Various properties and characteristics of some commercially 
available materials in these groups are given later, and those of 
a few materials with special properties are described below. 

Supermalloy 

The most commonly used additives are copper, chromium, 
molybdenum and manganese, and the highest permeability and 
lowest losses have been achieved by Boothby and Bozorth 17 (1947) 
with an alloy of composition 79 per cent. Ni, 15 per cent. Fe, 
."> per cent. Mo and 0-5 per cent. Mn, known commercially as 
Supermalloy. When heat treated in pure hydrogen at 1,300° C. 
and cooled at a critical rate, initial permeabilities of over 100,000, 
maximum permeabilities of 1,500,000 and hysteresis losses of less 
than 0-1 J/m 3 ("> ergs/cm 8 ) at 0-5 VVb/m 2 (5,000 gauss) have been 
reported ; saturation flux density for this material is about 
0-8 Wb/m 2 (8,000 gauss), and resistivity 00 microhm -cm. 

Constant Permeability Alloys 

Two further groups of ternary alloys have useful applications ; 
both are concerned with achieving constant permeability and low 
hysteresis losses ; one at comparatively low values of flux densities 
and magnetic field strengths, the other at high values. The low- 



so ft MAGNETIC MATERIALS 



69 






value alloys have been called Permivars and the high-value ones 
Isoperms. 

Permivar 

A typical Permivar alloy has a composition 45 per cent. Ni ; 
30 per cent. Fe ; 25 per cent. Co, but materials with various 
composition are made giving a fairly wide range of properties. 
The material mentioned has a constant permeability of about 500 
and negligible hysteresis loss, coercive force or remanence up to a 
maximum magnetic field strength of about 2 oersteds. Above 
this value hysteresis loss increases rapidly, the hysteresis loop 
having a constricted portion around the origin. The properties 
of very low coercive force and remanence are retained so that the 
loop is like a dumb-bell in contour. Near saturation, at higher 
fields, the hysteresis loop is of orthodox shape. A disadvantage 
of the material is that the constant permeability characteristic is 
destroyed if it is subjected to a high magnetizing force and can 
only bo restored by being heat treated again. 

Isoperm 

This material is of German origin and again a wide range of 
compositions and properties exist. The alloys usually contain 
copper up to about 15 per cent, and nickel from 35-50 per cent., 
the remainder being iron. They have constant permeabilities of 
from 50-110 up to field strengths as high as 100 oersteds. A 
direction of easy magnetization at right angles to the plane of 
the rolled strip is induced by drastic cold reduction of about 
98 per cent., followed by annealing and further cold reduction. 
This results in a texture which gives high permeability at right 
angles to the plane of the strip but low and constant permeability 
along it. 

High Initial Permeability Alloy 
" 1040 " Alloy 

This alloy is characterized by very high initial permeability — 
about 40.000 and resistivity (56 microhm-cm). Its composition 
is 72 per cent, nickel, 14 per cent, copper, 3 per cent, molybdenum, 
the remainder being iron. Saturation flux density is low 
(0-62 Wb/m 2 or 6.200 gauss) as also is hysteresis loss (1-5 J/m 3 oi- 
ls ergs/cm 3 per cycle). 



70 



SOFT MAGNETIC MATERIALS 







•OIS" 

1 1. 

■DOS 
























































OOJ" 










~ ' • S- ' * " 


















— 


























































i 






















■ 






























































I 














































J\\ 








































































£ 














RADIO 


MET 


»L 


































HI 

\ 1 






















































OOS 


t 




























































































































































\ 














so 

<o 








































































































^v 


























































••^ 






























































^, 




























































■ 015* 
Hi 



mCOUENCr - CrCLES PEB second 



















































1 


















00« ,k4- 

ooj-: — 


I-U 
































B 00l"- L 


















t^OO 




























\ 






s— - 







































\ 








-■^ 






































r.i 


"" 






X 




OOJ 


< lOO 










RHO 


MET* 


L 




























>l 1 

OO*" 


Z 










--| 


































--- + ' 














































































































I 












































































io 















































«EOUENC»- CYCLES PEP. SECOND 









■ 
MM 


-i — 












ife""-^ 














1 




TTT 










lOOOO 




'r^~^ 


" 
























">. 


V 








fv^~ 




\ 






; ; 
































N 








^"~^ 
















J»ooo 






















■> 
























< 

5 

IOOO 






















\ 






































MUME 


M 














































































. . 


- 




































































































50Q 
































"^ 


















«oo 


























































































\ 




, 

















































































































rOEOUENCY - CYCLES PER second 

Fig. 2-32. Variation of initial permeability of radio-, rho- and mcmetals 

WITH lUKQI KNCV KOK VARIOUS THICKNESSES. 

(Tfhymji/, ConftrveUan <(■ MainUnotmCo, Ltd.) 



SOFT MAGNETIC MATERIALS 7] 

COMMERCIALLY-AVAILABLE IRON-NICKEL ALLOYS 

Manufacture 

It has already been mentioned that the majority of commer- 
cially-available alloys have various small metallic additions 
modifying the binary compositions. They are also subject to 
different heat and reduction schedules. For these reasons a 
detailed account of the preparation of the various alloys will 
not be given, but the following observations relate to manu- 
facturing processes which are commonly in use. 

The basic metals employed for the production of the alloys 
are specially selected to be free from harmful impurities such as 
carbon and sulphur, oxygen and nitrogen, either in free or 
combined states. Melting is usually earned out in a high- 
frequency induction furnace and in a controlled atmosphere ; 
for some alloys the |)rocess is carried out in vacuo, although this 
is expensive and only applied in special cases. 

After melting the material is cast in ingots, which are machined 
bo remove surface irregularities and impurities and subsequently 
rolled to sheet or strip. The first part of the rolling process is 
usually carried out under hot conditions, but the later stages of 
sheet or strip production involve extensive cold-rolling. This 
leaves the surface of the material in good condition and free from 
scale. At all stages in the process care is taken to ensure that the 
material is not exposed to contaminating atmospheres. 

Laminations 

Magnetic anisotropy is caused by the final cold-reduction 
process, the best properties being obtained in the direction of 
rolling. For this reason these materials are frequently used as 
M clockspring " toroids or " C " cores. However, they are also 
used in large quantities in the form of orthodox transformer 
punchings and other laminated structures. The useful range of 
thickness for laminations is from 0-01 5-0-004 in., but strip 
can be rolled down 0-001 in. or even 0-0005 in., although the 
magnetic properties are inferior for these small thicknesses. 
Toroids formed from thin strip are sometimes solidified with 
solventless resin to make them easier to handle. 

Most iron-nickel alloys are sufficiently soft mechanically to be 



72 



SOFT MAGNETIC MATERIALS 



3 a 
* 5 

B§ 
> 



H 
S3 

a 

- 

Q 

s 
- 



Q 



-5 

I 





to 


i 


2 


*< -c 
in u 

a J, 


o 


o 


a 


° 5 j; ^ 

. Ml t 

S O P- u 
h O w v. 

•£ u> ?? a. 
^» '' a. 
K,Jg,0 


a 

10 


* 


io 


g g ^ 

.£ U tin 
V) M 


o 


o 


1 


Permeability 
Initial Maximum 


O 
O 
o 

ef 


o 
o 


o °. 


° 2 

O o° 


o 


O 

vr 


1 


i 


n 

u 

3. 
< 


B 

B 
< 



"■ 




P 




- 




<j 




£ 












gj 




B 


CO 




1* 


CO 




H 
r/) 


i 




<i 


- 


N 




- 


K 




H 


— 


w 




£ 


QQ 




3 


"** 


p^ 




H 


fc 




o 


P 










3 
un 
O 
O 

X 

E 

0) 

*J 

-a 

c 

3 

a 

u 
p. 

nj 

3 

K 

a) 

O 

-J 


IX 

u 
o 
o 
o 


09 


9 


r~ 


p. 

o 
o 


r* 


Z 


to 


IX 

o 
o 


r- 
o 


in 

o 


r\l 


D- 

u 

o 

ID 


in 
o 

o 


O 
O 


1 


C 


r- 
O 
o 


o 


H 

D 

o 


2 


c 


'6 

| 

«2 


•a fx 

" s 

'" c 
K in 









/ 


« 


|| 






/ 


.. 


it* 

s 


. 






- 


6 -■ 


w 4 








"TT* i 
? 1 


9 3 
O 














P 










< 










-*^ 


/ / 








_-^ 


/ / 









2 





< 


la 


-. 


•* 


- 


N 










O 


fc. 

o 




| 


fe 


!h 


p 


j 






-i 






-J 




< 

2 


< 


< 


S5 


x 




3 


> 


- 


i 


3 




a, 


- 




ot 


< 




ft. 






4 






a 


r- 




«j 


ri 


H - 






Ainia»3W03d n»lN3n3<13n 



&l 



SOFT MAGNETIC M A X B B I A L S 



73 







■-F 




g 

o 
o 

§ 

O -l 

82 

CE 

a 

8u 
§S 

i", 



3 
o 

S s 

3 
a 
O 

2 J 

o 
o •- 
o 
















— =2 g - 6 6 6 


— 








\\\A^ 


2 8* 

- o 

I* 




\ (, 






X_\ 




»e 




■ A \ 


/ V / \ 


a Z 
a <- 

?. > 

y o 

2 O- 
O 

tf> <- 

CX UJ 

3 > 

O 4 

»- — 
w J 

!: ;_ 

u 2 
< ft 
a u 



z 




-s \ / 




. i 


i " V 

o x. 


44 s 




9 :::: «{ 


\l^r~ 





Sk-Sk A 


-a 


io 




•s 


j 

3 


:: -5— 






---^ t- 


K 


Sp 


- «-\ 


5 


u Z 


-S^ 


" I 


a3 


\ 




\ 


O 

ssn»ooTi» 


- AHSN30 xmj 









1?" 

4|JrJp 

^i o b 6 o 

■niH-fl— 




8 

8S 

* 





V-- 




ftP 


iw 


-I _ 

a o_ 
<o~ 

2 J- 

O 

a z 
«t < _ 

< ui — 
f > 
5 o 
z Q 

UJ < — 

o > 

> — 

IU J — 
tu < 

US 

z 




—ft 


#- 




\-\-X 


-,1-M ^ 




r-V 


J-i-t 




\ \ 


si 




\A) 


!;(;.__ 


■ I 


p::.l 


..,j 






L.X,^^- 


c 


i| ( 


V^^V 


; 




°\ \ x 






-5 ^r 


J 


»! 


-\ V 


r- 


u Z 


•X* 


2 


'3 


\ 


o o 

| SSn»30TI» - 











- 






o 

E 

a 


J 
O 

o ,., 

5 


















„SS I' i' i' i 

oP u o o o 








M.' '-4 








■--i-i^ 










:: ±tf 




j — 
3 ° 

wO- 

< Ck ~ 

3 

o o - 
? z 

< > 

y ° 

* o - 
tfl * ~ 


— 1 


v ^ 


--J-TT 




\ v V 


J-i L 






V^4 


~m~ 






^::s:^ 


#^ 






\ V 


rZ 






\ s \ 


ii 






\) 


t 




X 


s -- -N 


i\ 




I 


I:::::!;:: 





Z ui " ' 


K^x^ 




j 




\ \ 




ul * 

U I 
< a 



z 






4 _\ A 




< 




- ^ 




s 


S 3 


\ 




I 


k'1 


\ 




! 2 

~ ssnvooi 


a - AJ.ISN30 xmd 





H z 



" -j 

a _i 

PS 

it 

t. m 
39 

Si 







Mf 




8 
8. 

uj 

is 

3 

J 
4 

O 

o 






»-8 ;i"in 

Sj S o 6 6 














s 


u r / 


3 - 






I Jk -U 




. :s: 


- A 4-4- 




\\ 


-M- 




"" : "5± 


hii 


Is: 

o " 

< < _ 

y > 

O O 

z a 

U 2 

< 3 




" ^ v\ 


i-ti-i 




\ \ v 


lit 




s \ 


IJiL 


-u 


s \ 




-I 


o 


&$kF- 


Q « 




«A - 




ii 




a 


. : 33 E 


- 




Li 


* 




\ \ 






... ^ 


< 


.5 


\ 




J z 
o 




a 


3 




z 


o « -r « « 

7 ssnvooiix - AHSN3Q xmj 


"" 







x. /. 








n 




« * 




< 




> - 




.. - 












r - 




5 | 

T 5 












-. r 








x u 




« H 








-^ 




p n 








^ s 




< V- 




-- p 




— - 








fl b 












a o 




(x K 




c < 




Si 




E- O 




M 












Bi - 








* :..; 




M 




B 








H cfl 




B p 




v, 




a 3 




< - 








S ~ 


-r 


_ h 












e 


— 




- 


M ^, 


«A s 


7. 


•? 




<- 




F 





74 



SOFT MAGNETIC MATERIALS 



SOFT MAGNETIC MATERIALS 



75 



fabricated into stampings, etc., in the cold-rolled state, but to 
develop the best magnetic properties, the final heat treatment at 
high temperature and in a protective or reducing atmosphere is 
carried out after all working ju'ocesses have been completed. The 
magnetic properties of these alloys are relatively sensitive to 
mechanical strains, and any deformation after final annealing is 
to be avoided. 

Insulating the Laminations 

In A.C. applications it is necessary to provide insulation 
between the lamina) in stacked assemblies so as to avoid excessive 
eddy currents. Three methods are in common use. For low- 
frequency applications where very high resistivity in the inter- 
laminar film is not essential, a thin layer of oxide is allowed to 
form on the surface of the material by exposing it to the air after 
the high -temperature annealing process while it is still sufficiently 
hot for oxidation to take place. 

For higher frequencies, one side of the strip or lamination is 
covered with a thin layer of lacquer. This results in a good 
homogenous film of high resistivity but yields a lower stacking 
factor, particularly with very thin materials. 

The method just described is not suitable for applications in 
which the final high-temperature anneal must take place after 
assembly. In these cases a layer of refractory material, -such as 
magnesium oxide in fine suspension, is deposited on the surface 
of the strip while it is being wound to its final configuration and 
the complete assembly is heat-treated, the insulating properties 
of the refractory material being unaffected by high temperature 
or the protective or reducing atmosphere. 

Non-laminated Parts 

In D.C. magnetic circuits it is sometimes convenient to use 
machined parts or small forgings of high-permeability alloy. 
Some of the materials can be produced in these forms or as 
various rolled or drawn sections or pressings. High -temperature 
annealing is necessary after these processes to develop the best 
magnetic properties, and an upper limit to the size of sections 
which may be usefully dealt with is determined by the inferior 
properties which result due to non-uniform heat treatment 
through large sections. 






Magnetic Properties 

Table 2, p. 39, gives some relevant data for several alloys cover- 
ing the principal categories of special magnetic characteristics. 

Figs. 2-31 to 2-34 show various properties and characteristics 
chosen to illustrate the conditions of magnetization for which 
the several alloys are best suited. 

IRON-COBALT ALLOYS 

Alloys in this group are of interest primarily because of their 
high magnetic saturation induction and their relatively high 
permeability at high field strengths. In comparison with other 
soft magnetic materials they are very expensive owing to the 
high cost of cobalt , but for some purposes their special properties 
are of great value. 

The binary alloys were investigated and the useful properties 
of alloys with 35-50 per cent, cobalt were appreciated by Weiss 
as long ago as 191 3. 18 Nevertheless, in addition to their high 
cost, the simple alloys have the major drawback that they are 
almost impossible to fabricate either hot or cold, so that despite 
having the highest saturation induction of any magnetic material 
(about 2-5 Wb/m 2 or 25,000 gauss), little practical use could be 
made of them. In 1932 White and Wahl patented a range of 
ternary alloys in which a few per cent, of vanadium was added to 
equal quantities of cobalt and iron and the alloys rendered 
workable by hot rolling and quenching. Magnetic properties were 
adversely affected by the addition of the third element, bui a 
compromise between magnetic quality and workability has been 
achieved at about 2 per cent, vanadium when the saturation 
induction is about 23.000 gauss with the advantage of increased 
resistivity. 

More recently (1947) a similar alloy was reported by Stanley 
and Yensen 20 in which the ternary element is about \ per cent, of 
chromium. Only 35 per cent, of cobalt is used in this alloy, 
which is therefore somewhat less expensive on basic materials ; 
in other respects its properties are similar to the Vanadium alloy 
but with a slightly higher saturation induction at very high field 
strengths. 

Fig. 2-35 shows a comparison between the magnetization 



76 



SOFT MAGNETIC MATERIALS 

























































i 


erme 


































1 














Jr, 


















1 










/ 
















.Mil 














^ 












tA 


silicon 














- 


























































































/ 1 


































/ / 


































V 






























































- 





SOFT MAGNETIC MATERIALS 



77 



Fig. 2-85. — Com- 

FABIS ON » 

MAGNETIZATION 
CHARACTERISTICS 
FOB V - T E R - 
MEMUK. AKM< ii 
IRON A NO 4-2 PER 
CENT. I R O N • 

SILICON ALLOY. 



MAGNETISING FORCE (OERSTEOSJ 










































^— 
















015" 










Xo 1 














1 








o.o- 




















































so' 




















1 


/ 




,\* 


































1 


' — 


/ 


r— 


/ 






























I 


/ 


/ 


/ 


/ 
































1 


/ 




/ 
































J 






t— 


/ 








































/ 










































/ 






































/ 










































/ 
























1 











MAGNETISING FORCE ■ (OERSTEOS) 



Fig. 2-30. — Variation 

OF TOTAL LOSSES 

WITH THICKNESS IN 
Y-l'l-'. KM i:\imk FOE 
VARIOUS EM \ i>i')N- 

S1TTES AND I'K !•' - 
QUENCIK.S. 

(Standard Tekpfiones & 
Cablet Ltd.) 









































^ 


».?oooc 
























^^ 
































B-ISOOO 




r 














100, 














6.10000 
























\_^~ 






























>" 










B-2QOOO 


















„ 


y^ 


























B-15000 












































Fig. 2- IS" (abom). — 
Effect of 
thickness on 
macnktization 
f: R V ES OK V- 
FERMENDI K. 



Wig. 2 -38 [left). 

Tykic-u, KAMILY 
ok HYSTERESIS 

LOOKS FOR V- 

V E B M E-N I > ' " B 

(0-010 IN. STK.IK). 



THICKNESS - THOUSANDTHS Of AN INCH 



78 SOFT MAGNETIC MATERIALS 

curves for V-Permendur (the commercial name for the vanadium 
alloy), Armco iron and 4-2 per cent, silicon -iron. The superiority 
of the iron-cobalt alloy at high inductions is immediately apparent. 
Figs 2-36 to 2-38 give some relevant data for V-Permendur. 
which is commercially available in this country. 



IRON-ALUMINIUM ALLOYS 

Alloys in this group have achieved little commercial applica- 
tion, but deserve mention because of their desirable magnetic 
properties. Preliminary investigations into their magnetic 
possibilities were carried out by Barrett, Brown and lladfield at 
about the same time as their work, already mentioned, on iron- 
silicon alloys. The aluminium alloys showed two major practical 
disadvantages. Aluminium was at that time comparatively 
expensive and the formation of a very hard oxide skin on the 
surface of slabs during hot-rolling made fabrication an extremely 
difficult and costly process. For these reasons they did not find 
favour. 

On the credit side the aluminium alloys were found to have 
magnetic properties which as regards losses, saturation density, 
resistivity, inhibiting of ageing, etc., were in every way comparable 
with the corresponding silicon alloys ; also they were not subject 
to the embrittling effect experienced with increased 1 silicon 
content. 

Since these early investigations several workers have re- 
examined the alloys from various aspects and a continual improve- 
ment in properties has been noticeable. In 1940 Bozorth, 
Williams and Morris reported 21 that by cold reduction and high 
temperature annealing of alloys with low aluminium content — of 
the order of 3-| per cent. — properties similar lo those of grain- 
oriented silicon-iron could be produced. These had the additional 
important advantage however that two directions of easy 
magnetization at right angles to each other lie in the plane of the 
rolled sheet — one being in the direction of rolling. 

During and since World War II there has been a revival of 
interest in the alloys as substitutes for iron-nickel alloys since 
they do not require strategically valuable or scarce materials. 
The effect of heat treatment on various compositions up to 
17 per cent, aluminium was investigated in Japan by Masumoto 



SOFT MAGNETIC MATERIALS 



79 



and SaitO 88 and a material with about 16 per cent, aluminium 
known as Alperm was developed. This has properties similar to 
Mil metal. A detailed study of fabrication methods for a similar 
material— 10-Alfenol— has been described by Nachman and 
Buchler. 23 The preferred process which they have evolved is 
briefly as follows. 

Manufacture 

Electrolytically-pure iron and aluminium are melted in vacuo 
and the material ehill-cast into slabs about 1 in. thick and 10 lb. 
weight, the moulds being specially lined. The slabs are trimmed 
and hoi rolled at 1 ,000° C, being reduced about 0-005 in. per pass 
and reheated initially after each four passes and latterly after 
each two passes until the thickness has been reduced to about 
I in. This thick strip is then reduced to thin strip 0-014 in. or 
0-007 in. thick by :i cold " reduction at 575° C. If very thin strip 
is required the thin strip is annealed for 24 hours at 575° C. and 
further cold reduction carried out at room temperature after the 
hard oxide skin formed during previous treatment has been 
removed elect rolytically. By these methods strip as thin as 
0-0003 in. has been produced. The finished strip is annealed at 
1 .Don 0. for about two hours in pure dry hydrogen and cooled in 
the furnace to 000° C. from which temperature it is quenched in 
water. 

Properties 

This process results in a material of high resistivity and low 
anisotropy with D.C. properties which, with the exception of 
initial permeability, are comparable with those of Mumetal, 
Table 4. p. 72. Its high resistivity makes it attractive for A.C. 
applications, and from Table 5 it is seen that even at 2,000 c/s 
and 0*6 W'l) in : - (5 kilogauss) the losses in 0-007 in. and 0-014 in. 
Laminations compare favourably with 0-001 in. silicon-iron strip. 
Also the ability to produce very thin strip may be important in 
reducing losses at even higher frequencies. Interlaminar insula- 
tion of good quality is formed automatically during fabrication. 

It is difficult to predict what the future of such a material 
might be. The fabrication process as outlined above is clearly 
an expensive one and economic progress probably depends on 
the extent to which it can be simplified. 



80 



SOFT MAGNETIC MATERIALS 



APPLICATIONS OF SOFT MAGNETIC MATERIALS 

This summary of the uses to which soft magnetic materials 
may be put is not intended to be comprehensive but rather to 
assist magnetic component designers to choose the magnetic 
materials with the properties most appropriate to their applica- 
tions. Only those materials which have been described in the 
foregoing text will be dealt with. 

Soft Irons and Steels 

These materials, which include electrolytic iron, Armco iron, 
Swedish charcoal iron and low-carbon steel, are characterized 
by low resistivity and high permeability at high flux densities. 
They are therefore most suited to D.C. applications in which 
high Mux densities are required with small magnetizing forces. 
Typical magnetization characteristics are given in Fig. 2-3. 

Applications include relay yokes, armatures and cores, electro- 
magnet yokes and pole pieces. A.C.- and D.C. -machine frames 
and pole pieces, A. C. -generator rotors and a wide variety of 
special equipment such as electromagnetic chucks, clutches and 
the like. The materials are produced in various forms to suit the 
applications ; for example, motor frames are cast in low-carbon 
steel or fabricated from rolled plate and the cores of large electro- 
magnetic structures are frequently forged ; small components 
are usually assembled from stamped, pressed or drawn parts of 
thick sheet. In general, the materials require to be annealed after 
working to develop the best magnetic properties. . 

Applications of Iron-Silicon Alloys 

These alloys may be considered in three groups : low- and high- 
silicon content and grain -oriented materials. Low-silicon alloys — 
up to say 2\ per cent, silicon — are usually called dynamo grades, 
and are widely used in ordinary motor and generator construction. 
Material with less than about \ per cent, silicon is used for cheaper 
grades of small motors for intermittent duty, and in parts of 
magnetic circuits in which high permeability at high flux density 
is of greater importance than low losses. From 1-2|- per cent, 
silicon alloys are used for motors and generators of average 
efficiency, cheaper grades of small transformers and reactors and 
other application in which low cost is more important than high 



SOFT MAGNETIC MATERIALS 



81 



efficiency. From 2^-3 per cent, silicon alloys are applied to 
motors and generators in which high efficiency is the over-riding 
consideration. They are also used for small- and medium-sized 
t lansformers which are not continuously rated but which do not 
merit the use of more costly high-quality grades. 

The high-silicon alloys— from 3-4£ per cent. — are the normal 
materials for power and distribution transformers at power 
frequencies and for some parts of the associated alternators and 
other continuous-running niaclunes and equipment in which low 
losses and high efficiency are of paramount importance. For 
economical construction, low losses must be associated with high 
working-flux densities and these materials represent the best 
compromise between these factors and a magnetic material which 
is not too costly to produce. On account of their high resistivity 
the high-silicon alloys are also suitable for use at audio frequencies 
for small-power and matching transformers. 

The applications of grain -oriented iron-silicon alloy have 
already been described in some detail. The thicker strips arc 
suitable for the highest efficiency transformers of all kinds operat- 
ing at high flux densities and power frequencies, and are also 
used in some parts of associated rotating machinery. Made-up 
cores of thinner strip either as continuous " clocksprings " or 
0-eorea arc applied to a wide variety of purposes, particularly in 
the fields of electronics and telecommunications in which they 
are used for audio-frequency power transformers, chokes, pulse 
transformers, etc., in which low losses are important. They are 
also useful in magnetic amplifier and saturable reactor work in 
which their low losses and comparatively rectangular hysteresis 
loops are advantageous. 

Applications of Iron-Nickel Alloys 

The range of applications of these alloys is too great to catalogue 
in detail. There are four broad classifications other than the 
special materials such as those with constant permeability and 
high-temperature coefficient of induction already mentioned. 

High Initial-permeability Alloys 

This group includes Permalloy and Mumetal. Their special 
characteristics are extremely low losses and high permeabilities 
at low flux densities, which features are of particular value in 



82 



SOFT MAGNETIC MATERIALS 



many electrical engineering and telecommunications applications. 
They are used in wide-band, instrument and other special trans- 
formers, in filter coils, as loading in submarine cables, as saturable 
reactor cores, for magnetic shielding, and in a wide variety of 
special components, such as sensitive high-speed relays, gramo- 
phone pick-ups and instrument armatures. 

High Saturation-density Alloys 

These alloys include Permalloy B and Radiometal. They are 
similar in properties to the best quality grain -oriented silicon-iron 
with somewhat lower saturation density. They are used in small 
power transformers at audio frequencies and for cores in trans- 
formers* chokes, relays, etc., with combined A.C. and D.C. 
magnetization, having comparatively high incremental permea- 
bility under these conditions. 

High-resistivity Alloys 

This group includes Permalloy D and llhometal. These alloys 
are of value where eddy-current losses have to be minimized at 
very high rates of change of flux ; for very thin strip the change 
in initial permeability is small over a wide frequency band up to 
several hundreds of kilocycles. These characteristics are used in 
high-frequency generators, pulse transformers and high-frequency 
inductors and transformers. 



Rectangular-loop Alloys 

This group includes Permalloy F and H.C.R. metal. The 
special features of these alloys are high -saturation flux density 
and remanence. and low losses, coercivity and magnetizing 
current up to saturation which occurs at high flux density for a 
small proportional change in flux density. They are particularly 
useful in circuits in which an "inductive switching'' action or 
"two-state " characteristic is required. Applications of this kind 
arise in mechanical -contact rectifiers, magnetic-pulse generators, 
magnetic amplifiers and magnetic memory-storage devices. They 
are also useful for components such as current transformers 
where low waveform distortion is required over a large change 
in flux density. 



SOFT MAGNETIC MATERIALS 



83 



Construction of Laminated Assemblies of Iron-Nickel Alloys 

It has already been mentioned that the best magnetic properties 
are obtained in these alloys hi the direction of rolling of the strip. 
It is therefore advantageous, wherever possible, to use the 
material in the form of continuous-wound strip cores. This has 
economic as well as technical advantages over the use of stampings 
since there is no waste and no tool costs for punching tools ; also, 
assembly is simplified — particularly where very thin material is 
required. A disadvantage is that the windings have to be threaded 
through the core, but with modern winding techniques this is not 
a serious problem. In some cases C-core constructions are possible. 

Applications of Iron-Cobalt Alloys 

The commercially useful alloy is V-Permendur and is charac- 
terized by high permeability at very high flux densities. It is 
expensive and is used in those parts of magnetic circuits in 
which high flux densities for low magnetizing forces are required. 
When used for pole pieces of D.C. electro-magnet structures, parts 
may be rolled, forged or cast and machined. This material can a lso 
be supplied as thin sheets such as are used in telephone dia- 
phragms, and is useful in A.C. circuits in this form. The best 
magnetic properties are attained when the material is thoroughly 
annealed after working. 

Acknowledgments 

The author is grateful to Messrs. Armco Ltd.. English Electric 
Co, Ltd., Richard Thomas & Baldwins Ltd., J. Sankey & Sons 
Ltd., Standard Telephones & Cables Ltd., Telegraph Construction 
& Maintenance Co. Ltd., and the Steel Company of Wales for 
information about their products which they have generously 
given and permitted to be published, and to the authors and 
publishers of the various papers and articles given in the references 
on which he has drawn freely. He also wishes to thank the 
Directors of the British Thomson-Houston Co. Ltd. for permission 
to publish this chapter. 

References 

1. C'iofki, P. P. "New High Permeabilities in Hydrogen-treat eel Iron," 

I'hys. Rev., 1934, 42. 

2. Bozokth, R. M. "Directional Ferromagnetic Properties of Metals," 

J. App. Phys., 1937, 8. 



84 SOFT MAGNETIC MATERIALS 

3. Yionsen, T. D. " Magnetic Properties of Ternary Fe-Si-C Alloys," 

Tram. A.T.E.E.. 1924, 43. 

4. Yenskn, T. D., and Zitcoler, X. A. •• Effect of C, ami Grain Size on 

Magnetic Properties of Si-Fe." Trans. Amir. Sor. Metals. 1930, 24. 

5. ElmBN, <;. W, '• Magnetic Alloys of iron. Nickel and Cobalt," J. 

Franklin Inst., 1929, 207. 

6. Pakkktt, W. P., Brown, W., and Hapfdbip, R. A. "Electrical 

( 'uiiductivily and Magnetic Permeability of Various Alloys of Iron," 
Sei. Trans. Hoi/. Dublin Soc, 1900, 7 ; J.IJ&JE., 1902, 31. 

7. Gokktz, M. "• Iron Silicon Alloys Hear -treated in a Magnetic Field." 

./. 4pp. Phys., l"l.-)l. 22. 

S. Williams. H. J. " Magnetic Properties of Single Crystals of Silicon 
Iron," Phys. Bev„ 1937, 52. 

9. (Joss. X. P. ''New Developments in Electrolytic Strip Steels Char- 
acterized by fine (!iuin Structure Approaching Properties of a 
Single Crystal." Trans. Amer. Soc. Metals, 1935, 23. 

10. Dunn, C. G. "Controlled Grain Growth Applied to the Problem of 

Grain Boundary Energy Measurements," Joum. of Metals Trans.. 
Jan., 1949. 

11. BiRNSTixox. I>. \V. "Some Recent Advances in American Trans- 

former Manufacture, 11 J.I.E.E., March. 1954. 

12. "A Handbook lor Designers of Transformers for the Electronic 

Industry " issued by English Electric Co. Ltd. 
Pamphlets issued by Telcon-Magnet ie Cures Ltd. and J. Sankey & 
Sons Ltd. 

13. Yensen, T. D. " Magnetic and Electrical Properties oT Iron, Xiekel 

and Cobalt/' Journ. A.I.E.E., 1920, 39. 

14. Ki.mi-;n. C. W., and Arnold, H. D. ! ' Permalloy — An Alloy of Remark- 

able Magnetic Properties," J. Franklin Inst., 1923, 195. 

15. S:\nTir. \V. S.. Carvett, H. J., and Kaxdall, W. F. Brit. Pat. 

366,523—1930. 

16. Dim. inckk, .J. V., and Bo/.orth, II. M. " Heat Treatment of Magnetic 

Ma 1 1 'rials in a Magnetic Field. Survej of Iron-Cobalt-Niekel Alloys," 
Physics, 1935, 6. 

17. BOOTHBY, O. L.. and Bozorth, R. M. " A New Magnetic Material of 

High Permeability," J. App. Ph>/s.. 1947. 18. 

IS. Wkiss. P. • Sue les champs magnet iques oblcnus avee un electro- 
aimant muni de pieces polaires in ferro cobalt," Complex Rendus, 
1913, 156. 

1!). White, J. H.. and \\ aiil, C. V. U.S. Patent 1,862.559—1932. 

20. Stanley, J. K., anil Ykxsen, T. D. " Hiperco — A Magnetic Alloy." 

h'.Uv. Engrg., 1947, 66. 

21. Bozorth, Wn.uwis and Mobbis. "Magnetic Properties of Fe-Al 

Alloys," Phys. Rev., 1940, 58. 

22. Ma/.umoto and Saito. i; On the Effect of Heat Treatment on the 

Magnetic Properties of Iron-Aluminium Alloys," Research Report 
Tohokw Univ., Japan, 1952. 

23. Nuiiuw and BUCKLER. " The Fabrication and Properties of 16- 

Alfenol — a Non-strategic Aluminium-Iron Alloy," Navord Report 
2819, 1953. 

24. Melville. \V. S. " Pulse Magnetization of Nickel Irons from 0-1 to 5 

Microseconds," J. I.E.E., 1950, Part 2. 97. 






2. (b) MAGNETICALLY-SOFT FERRITES 

The present section is devoted to a consideration of magneti- 
cally-soft ferrites which have a low coercivity combined with high 
permeability. The development of these materials was occasioned 
by the need for magnetic materials which could be used as the 
cores of high-frequency coils, chiefly in connection with radio and 
allied equipment. These ferro-magnetic oxides are known, in 
Britain, tinder the name of Ferroxcube. 

The ferro-magnetic oxides are of such high specific resistance 
that the eddy-current losses are reduced to negligible proportions. 
The losses which, occur in them are clue to hysteresis and ferro- 
magnetic resonance absorption (i.e. residual losses). 

ELECTRICAL AND MAGNETIC PROPERTIES 

Ferroxcube. as far as its chemical composition is concerned, 
consists of mixed crystals of simple cubic ferrites with typical 
compositions MFe 2 4 , where M represents a divalent metal (a 
divalent metal can, in this instance, be either Cu, Mg. Mn, Ni, Fe 
or Zn) and crystallizing with the typical spinel structure MgAl 2 4 . 

The material is characterized by high initial permeability /x„, 
small coercivity H c , and of such high resistivity that the need 
for laminating or powder- bonding the core, in order to reduce 
eddy current losses to reasonable limits, is obviated. The 
material is produced by processes similar to those used in the 
pioduction of the common insulating ceramics and possesses 
comparable mechanical properties. 

Ferroxcube is a homogeneous material — not a bonded powdered 
core — and therefore contains no internal air gaps. In certain 
circumstances it may be advantageous deliberately to introduce 
an air gap into the magnetic circuit, say in order to reduce the 
influence of temperature changes or harmonic distortion, or when 
the core is subjected to D.C. polarization superimposed upon the 
alternating induction. Air gaps are also used in Ferroxcube cores 
to obtain coils with the maximum " Q " factors over a given 
frequency range. The electrical and magnetic losses can be 
matched by the introduction of the appropriate air gap, a feature 
which is not always possible with powdered dust cores since in 
these there is always a fixed internal air gap. 

85 



86 



MAGNETICALLY -SO FT FEKRITES 



Ferroxcube is now widely used in a variety of applications 
considerably greater than was originally envisaged. Such 
applications are in high quality coils, in carrier telephony, wide- 
band transformer's., high-frequency power transformers, inter- 
mediate frequency transformers and deflection coils in television 
receivers, pulse transformers and the like. 

There are two main types of Ferroxcube, both of which are 
ferrites having a cubic crystal structure. These types are : 

(1) Ferroxcube A — a range of manganese-zinc ferrites. 

(2) Ferroxcube B — a range of nickel -zinc ferrites. 

Four grades of manganese-zinc ferrites have now been developed 
and five grades of nickel-zinc ferrites. In order to distinguish 
between each grade a colour code is used. 

Ferroxcube B, of which there are five grades, differs consider- 
ably from Ferroxcube A, having in general lower values of initial 
permeability and saturation flux density, increased coercivity and 
Curie temperatures, and very high values of resistivity resulting 
in improved high frequency characteristics. A frequency range 
of from approximately 500 kc/s to 200 Mc/s is covered by the 
five grades. These materials show magnetostrictive properties. 
This makes them suitable for electro-mechanical niters and 
receiving transducers. 

The tables show the electrical and magnetic properties of all 
grades of Ferroxcube. It should be emphasized that the figures 
quoted are nominal, the measurements being taken, for the 
greater part, on pressed ring specimens. 

For extrusions in rod or tube form, somewhat inferior properties 
are obtained since it is difficult to obtain a density equal to that 
of a pressed piece part. In practice this does not represent a 
serious disadvantage as rods and tubes are usually in the " open 
circuit " condition where they will be working with severe 
" dilution " which is determined, for a given material, mainly by 
the dimensions of the core. 

Permeability 

It is found that the initial permeability of Ferroxcube materials 
increases as the flux density increases, up to a certain strength 
and thereafter decreases as the saturation point is approached. 
The point at which maximum permeability is reached depends 
upon the temperature. 



MAGNETICALLY-SOFT FERRITES 



87 





T3 




<y 








GS 








go 




© 




CO 






Ph 




>H 


-q 




+= 




O 


H 




W 


CO 


B 




u 


5 


M 


B 


o 




PS 


o 




o 


ft 










go 


CD 


a 


u 


o 


s 


+3 


PS 

H 
P- 

O 


*3 

c 
c 


PS 




PL, 


c 

-*5 


O 


tj 


M 




H 


h- 




0) 


C3 


tJ 


i 


fl 


r-. 






CO 







< 


00 


A 


<a 


< 


<S 






PS 


*p 


H 


c 


O 


o 


w 


c 


Hi 




K 






a 




CD 








CD 




A 




H 









« 


5 o 






11 1> 
















1 { 1 


1 






1" 














© © © 






rt B i 


■p V 






X 


■# 




© 


>— — * ^H 


1 — 




C C C C 


A d 


Q 


•3 


■* 


< 




© 








10 -t 


ci 


l~ 


'"- 






1" 


XXX 


X 




© CI 
















© tQO 

CI 1- 


■* 




eo ci 
















„ oe 


., 




















1 1 1 

c; o o 


1 

o 




•^ n 


x •*• 






X 


< 




© 


— F-« #— H 


— 




~ =. © © 


— 3 


Q 


© 


•* 




© 


XXX 


. X 




© © 
©_ ©_ 


K 


>o 


© 






-i» 


— " 


© I— t~ 


-* 




■* C5 












2 « 




— CI IC 


















— >-i 






















©^ 






















c © 










































/ /. 






tc 




•o 


















1 




IQ -*■ 












MS 






© 




tC "M 


X •+ 






X 


IN 

< 










© © ©© 


— £ 


© 


© 


■« 


~~ ~~ 






X 




lO ■* 


SO 


11 


£» 












•* 




C5 CJ 
















i T i 


7 




rf eo 
















© © © 


© 




SO <M 


x -* 






X 






© 


•— — 1—1 


— 




© © O c 


— © 


© 


© 


<* 


< 




IC5 


XXX 
to ©c 

CI T 


X 

I- 




© © 


so 


© 


00 








«- 






















o 






















2 . 








, . . . . . 






















'. S 5 












s 


If? 








: i . ,f - 

5_ M> Mj 


• • 










,2 


I'* 






















© 




















Si 

J J 


© 
© • 








•J O 












© 








a 










O 


- 3, 

« Q. 


IA . 








»N © 

3 










~ 


"a*" 




















03 


a. ® 
* S 


d 








¥ R 










c 


_ a 










© 










JO 




o 




d 














O 

e 

g 

9 


a r 

- s 

_ a 

x x 


3 

1 . 
43 




© 
— 
s 

a, 


© 


O 

© 


• • 








&H 


o o 

- - 

O. 3, 

55- 




© © © 
© >- ■-: 

"5 ; : 


° © 

C3 Q 


O 




? 








0. a. 
bp Mi 

£ «- 


1 • 

e 

I 3 

P- = 


S i .. .." 

(ri 

O 


1(5 

© 

II £ 




1 

>; 

i 

- 


a 



.1 


t 

•> 

— 
M 




5 3 


* IO 


i 




1 


w 


< o 


O 
53 


i 




IT ~ 
o o 

w — 


I'- 


% S ! 

o " 


i 


£, 


1 •• 
ft) 


&3 


1 


o 

- 



88 






9 



MAGNETIC ALLY -SOFT I'ERRTTES 



-V 



22 






^ 



it _i_ 



*a 



o 

- = 



II 



II 

If 



i i i I I 

OOOOC 

sssss 

»ni-N us 



XX 



ss 



£ = 



IV 



; ■ 



£ ££ 

;i X M 
M 0$ ".-S 



' — -i a * I-. 



•-- a 



-<IYS 



3 k 



5 S 



■-: SO r ..- 

» — — 



8 8 ° 8 * 8 






5 = - e r c 



t-i n "ji 






a | 
51 






I 



MAGN KTICALLY-SOFT FERK IT BS 




89 



H (OERSTEDS) 



B 
GAUSS 

















s*^^^ 


s*r - -Li-- - 








//* 


/ J R ' 


/ 


/ *_ 


7 j 


/ u 






h it 'i 


, b a 1 1 • 



B 




III 






1 1 1 






- 20°C . -- 


4000 




' 




— »•- -* *i I 










^sp**^ 






Sj" 


■ - HH " 


30OO ^ 


--/{- 7 m2ZZ 


rOO°C J 






" " " 




t i ',4? 






■l-y 




' ' 


i\l _:::: 
















y 






_ t 








i 




II 


t - . 




:- 




--- A3. -- 


-J E 




1 ! 1 


JL 3 




J. I 





















































































































20°C 








































70°C 








































4-1- 


- 








' ( ' 






100* c - 






/ 




"■ 














n 


















— 










—- — 


















IT. 


r »- 




















































































































































A 4 












































.: 1 


I 














i 



HiOERSTEDSI 



B 

GAUSS. 

























3000 
























































,^5" 


r^ 


K^- 


20° C. 












< 


/ 


















/ 


'J 


' 
















2000 




/ 


/ 




















i 




































IOO° c 










• 




















I/ 


















IOOO 




f 




















1 




















































































B 1. 






« 





















H (OERSTEDS) S 



H (OERSTEDS). S K) 

Fig. 2 3!). Tvi-K ,\i. h li loops Kim HEB&OXCUBB " A. " and " b! " ckadks. 

{Milliard Ltd.) 






90 



MAGNETTCALLY-SOFT FERRITES 



















1 - 


















20° 


c. 












































_ 




/ 






s 




-s* 




IOO° 


c. 




1 


i 




' 














I J 


\ 


/ 


















// 


/ 










































/ 






















/ i 




















1 


I 












































1 






















// 














B 2. 






i 
















' 














| 





H (OERSTEDS! IO 





































































































































































"^o«c 


























IOO°C. 








/' . 




^'' 


















/ 




/ 


/• 




















/ 


/ 




















f: 








/ 


















































































1 






1 






















1 




• 


' 














B4. 


— 


























i 




1 





















H10ERSTEDSJ 20 



B 

GAUSS 









































































20»C. 
























/ 






: — - 


IOO<>C 




/ 




" / 












/ ' 




/■ 


















/ 












/ 


1 






































// 


















<'l\ 




































m 












a 3. - 


. 


A 


/ 












1 


/ 


'! 















H.OEBSTEDSl 10 

















































20°C 






^^ri^T^ - 




_, — — ■£TZ~±zz "' " 




s ^ 


*" •£ ** 




J 












Jl 


Jl 




1 






I 




B 5. 


LA 1 L«rf_ 


_i 


1 



ec 



H (OERSTEDS) 20 

/'/;/. 2-39 (continued). — Typical b/b loops BOB eekhoxci be b2, b3, n4 and b5 

CKAUES. 

(Ualhr't 1.01.) 

The initial permeability is also subject to variation according 
to working temperature. 



Curie Point 

The initial permeability depends to a large extent upon the 
temperature, increasing at first as the temperature is raised, then 
decreasing very rapidly at a temperature of 150° C. to 190° C. for 



MAGNETICALLY-SOFT FERRITES 



91 




Wig, 2-40. — Sectional 

ARRANGEMENT OF I.F. 
TRANSFOKMEK. 

S = coils, K = Fer- 
roxcube cores, adjust- 
able by means of screws 
T and glass rods O. The 
cores are held in position 
by glass rods G' and 
springs V. Ferroxcube 
rods /'' constitute a so- 
oa I led " palisade " 
screen for reducing losses 
in the aluminium can A. 
The components on the 
extreme right- and left- 
lumd sides are " drawn M 
wire capacitors. 

(Philips Bbdrical Lid.) 



A grades, and between 120° C. to 550° C. for B grades. The 
temperature at which the initial permeability falls to 10 per cent, 
of its maximum value is known as the Curio point. Ferroxcube 
cores should not. in general, be operated too close to the Curie 
temperature, although no permanent harm to the core results 
from uniform heating above this temperature. 

MECHANICAL PROPERTIES 

Ferroxcube is a hard, black, non-porous material (not a bonded 
powder core) of a ceramic nature, chemically inert and unaffected 
by humidity and other atmospheric conditions, with mechanical 
properties similar to those of the more common insulating 
ceramics. Although not easily broken, it should be handled with 
care, particularly to avoid clipping of sharp edges or corners. 



02 



MAGNETICALLY -SOFT FER1UTES 




Fig. 2-11. — Hkmi-ki riciKSCY TELEVISION receiver t.ine oitimt transformer 

USING MILLARD FKHltOXCUHE FERRITE " " CORES. 

[MvOard Ltd.) 




Fig. '2 42. Modern high-efficiency magnetic television receiver 

DEl'LECTOR-COLL ASSEMBLY USING MULLARD FERROX( I BE 'HUES (shown 

ON bight). 

(Muii'ii'i as. i 



MAGNETICALLY-SOFT FERR1TES 



93 



The table below gives the mechanical properties of Fen-ox-cube 
which in general apply to all grades. 



Specific gravity 



4-8 for A grades. 

4-75 to 3-7 for B grades. 

21 x 10 6 lb/sq. in. 

10- 5 per °C. 

2,600 lb/sq. in. 

10,400 lb/sq. in. 

0- 1 7 cal/gm/°C. 

8 X 10- 3 cal/cm sec./°C. 



Young's modulus 

Coefficient of linear expansion 

Tensile strength 

Crushing strength 

Specific heat .... 

Thermal conductivity 

The processed raw materials are milled into a finely divided 
powder, mixed in the correct proportions, and pressed to shape 
in dies, or extruded in the form of rods and tubes, and then 
sintered in a suitable atmosphere in a high-temperature furnace. 

Shape of Pressings 

Pressings should be of a fairly simple shape to facilitate the 
manufacture of dies, and abrupt changes in the section of the 
pressing should be avoided where possible. Slots and bevels can 
be provided in the piece part where they run parallel to the 
direction of pressing, and provided they do not materially reduce 
the section so as to make the pressing mechanically unsound. A 
uniform cross-section ensures a more satisfactory and consistent 
piece part with consequent less distortion after the firing operation. 

Where large quantities of a given pressed part are required, 
automatic dies are used which are capable of providing a large 
output of piece parts of consistent shape : but where initial 
samples are required to enable experiments to be carried out in 
order to prove or finalize a design, hand tools are used which 
enable only a limited number of piece parts to be produced. 
For samples, the making of dies can sometimes be avoided by 
fabricating a piece part from an existing die, shaping approxi- 
mately to the required size before firing, and then grinding to the 
final dimensions. 

Machining Operations 

After the piece-parts have been sintered, they cannot be 
machined except by grinding, although they can be slit with 
diamond impregnated slitting wheels, or by ultrasonic methods. 
Due to the brittle and hard nature of the material, the large 



94 



MAGNETICALLY -SO FT FEBR1TKS 



shrinkage — some 22 per cent., and the high sintering temperature, 
inserts of any description cannot be moulded into the pressed or 
extruded piece parts. It is. however, possible to thread rods and 
tubes externally. 

Where two faces are to be butted together, grinding of the 
faces can be satisfactorily achieved by using abrasives of the 
aluminium oxide type. During the grinding operation light cuts 
onty should be taken to avoid splintering sharp edges, and the 
piece part should be flooded with a coolant to avoid local heating 
which might cause cracking of the piece part. 

Ferroxcube piece parts can be cemented together very success- 
fully by most of the synthetic resin cements already in use with 
ceramics. 



APPLICATIONS 



Carrier-Telephony 



Ferroxcube may be used as a core for filter coils used in carrier 
telephony. In these coils losses must be small in order to ensure 
sharp separation between the pass and attenuation bands. Flux 
density must be low to avoid distortion and a third requirement 
is that the coils must be of the smallest possible dimensions. 

Other applications in telephony are for the cores of loading 
coils and wide band high-frequency transformers used in amplifiers 
in carrier-telephony systems. Here the main requirements are a 
flat response curve throughout a wide band of frequencies, low 
output impedance, and low overall losses. 

Radio and Television 

Recently, Ferroxcube has been used successfully in the manu- 
facture of radio components, an example of this being intermediate 
frequency transformers (Fig. 2-40). 

Amongst the several applications in television equipment may 
be mentioned the use of Ferroxcube in the apparatus used to 
generate the extra high-tension required for operating cathode- 
ray tubes (Fig. 2-41) and also for the cores of television-receiver 
deflector coils (Fig. 2-42). 



3. (a) PERMANENT MAGNET STEELS AND ALLOYS 

By 

F. Knight, A.M.I.E.E. 

The remarkable advances made in permanent magnet technique 
during the last 50 years probably surpass those made by any other 
trade or art in their relative magnitude. At the beginning of the 
century the only material available for the manufacture of 
permanent magnets was carbon tool steel. An indication of the 
advances made up to the present day is that the latest commer- 
cially available alloys are capable of producing an energy equiva- 
lent to about 40 times that of this earliest steel. 

New Alloys 

Whilst the advances made have been generally in the direction of 
steadily increasing energies, parallel investigations have produced 
alloys having special characteristics for particular applications. 

This rapid expansion has resulted in a very wide range of 
materials suitable for permanent magnets being at the disposal 
of the designer and user. Contrary to what might have been 
expected, the newer alloys have not altogether replaced their 
forerunners, of which it is true to say that none has been com- 
pletely displaced, each retaining its own sphere of application. 
This is not due to conservatism on the part of user or manufac- 
turer but to the fact that for certain applications the earlier and 
less efficient magnet materials may give an overall economy when 
used in conjunction with a particular piece of equipment. More- 
over, the intractability of some of the most modern alloys some- 
times imposes physical and mechanical difficulties which make 
their use impracticable. 

Classification of Permanent Magnet Materials 

This wide range of permanent magnet materials can be sub- 
divided into three main groups, winch are discussed at length 
below. They are the carbide-bearing magnet steels, the composi- 
tions and performances of which have been long established and 
may, therefore, be regarded as standardized ; the diffusion- 

95 



06 



PERMANENT MAGNET STEELS AND ALLOYS 



hardening alloys — the Aim, Alnico, Ticonal, Alcomax materials ; 
and the precipitation-hardening materials, which are of impor- 
tance because of t heir case of niachinability. Many other materials 
exist which exhibit permanent magnetic properties, but only 
those arc mentioned which are commercially available. 

CARBIDE-BEARING MAGNET STEELS 

This group of materials comprises all those which can be called 
magnet steels, being essentially carbon steels with added elements. 
Carbides formed during heat treatment produce high inter- 
molecular stresses, which are responsible for their magnetic 
hardness. 

The first special magnet steel to be developed was G per cent. 
tungsten steel. It represented quite a considerable advance 
magnetically over carbon steel. Chromium steel, containing up 
to (5 per cent, of chromium, originally a substitute material for 
tungsten, has retained its own sphere of usefulness in spite of 
its slightly reduced performance. 

A more recent development is 2 per cent, cobalt/4 per cent. 
chromium steel. It has an increased performance and in spite of 
its rather different magnetic characteristics can often be substi- 
tuted with advantage for tungsten steel in existing magnet designs. 

Honda's original discovery in 1920 1 of 35 per cent, cobalt 
magnet steel was followed by the development in England of a 
w hole range of cobalt-bearing steels. These have been more or 
less standardized and manufacturers now usually supply steels 
containing 3. <i. '.». II or 35 per cent, of cobalt. The performances 
of these steels increase with the percentage of cobalt and, together 
with those mentioned earlier, they make a range of steels having 
energies in increasing steps from 2,400 J/m 3 (0-3 mega-gauss- 
oeisteds (m.g.o.) ) to 7, GOO J/m 3 (0'95 m.g.o.) and coereivities from 
0*53 hh. \ m (GG oersteds) to 2x 10* A/m (250 oersteds). 
Table I lists the magnetic properties of these steels. 

DIFFUSION-HARDENING ALLOYS 
Misnima discovered in 1931 2 that an alloy of iron, nickel and 
aluminium approximating to the formula Fe 2 NiAl had, after 
a suitable heat treatment, a coercivity about twice as great as 
that of 35 per cent, eohalt steel. The remanence was low. being 
just more than half that of the older steel and, as was to be 






PERMANENT MAGNET STEELS AND ALLOYS 



32s 



5 BJ 

s 



m 

■- ~z © 

sis 



c_f= 



i?££ 






a? 



aa 



aq 



ti 



- 



o ~ — 

X©~ 



t" © 

6 



23- 



© t- 



a ,r - — 

-Ma 

it: © 



S _ 



gs 



c o 

71 ~ 
© •>! 



i- "M 
t- O 



x« 



© * 
© a 

« ■* 



© 

CI © 



IS O 



CI -J 

6 "■*' 



o o 
<n © 



«2 



x© 

6®. 






4$ 

a 3 

•§£ 
^2 

J 



5,0, 



© ■"*■ 



© a 
— © 



o 6* 



— < ic 

x2 






if! © 
fM ITS 

-* CI 






x ,? 

~;© 



•* irs 

© w 



>r: © 
x io 

-* so 

© ■*" 



© >n 
ec ir: 

fM M 



© 
i- © 



8* 



si 



©-^ 

■"f it 



© © 
© © 

>c © 



x» 



5© 
i- © 



c 
© © 
■ft ©^ 
6 i-':' 



© ^ 
^f x 



© © 
os © 



© is 

t- N 

— *© 



X2 



C«5 X 



US © 
fN If} 

.- -I 



r* X 



X _ 

© "*" 



U5 © 
X lO 
O X 



© — • 
© eo 



© 

x£ 



l^© 

fN © 



« © 

© n 
ir: c; 



© © 

r. '" 



97 



3 § 
"5„ 

> 2 
c © 

II 

te 3 

go 

8.S 

j.r 

2 s 

g.5 
«•« 



M 



W 



s* 



2 - 
o .-e 



«a5 



3 -a 

c c 
— © 
d > 

&2 

© ~ 

to a 

- — 
3 cS 

■a > 
> >. . 

s .? 

« © « 

1^ 
H' 

u C 

2 & c 
SO' 8 

III 

I 1 ! 
Ill 

°2 



>§ 



PERMANENT MAGNET STEELS AND ALLOYS 



PERMANENT MAGNET STEELS AND ALLOYS 



99 



— 






i 
















/fat 

max. reco 

point 
























s 


fesi 


Sa 


2« 


ss 


=>5" 


e 5~ 


i 


re!2 




■< 




— -M 


Cl "I 


— CI 


— Cl 


— Cl 




«3, 






















— 



























•* 


, 


^^ 





^^ 


^ 


^^ 




^.-^ 


B at 
x. rec 

point 




= 


o 


CM O 


ci _ 


^ — 


■~ C 






CO o 






-f 71 


=: ci 






CC Cl 


■ 






i 


6 "4* 


n -t 

C re 


C ci 


6 «9 


6--* 


«* re 


• 


Cl -t 

c ci 
























S -3 






















||| 






,_^ 


^^. 


^^ 


_». 


^^ 


- _. 




^^ 




o 


© t~ 


O T* 


C: -* 


_ 


o •* 


- r 








= 




IO CO 




l~ 1- 


o t- 




■ 




'^2 






ei .- 


re is 


— - 


wo 






O X 




>"9 


rC o 


f O 


•* o 


>r: e 


■- c: 


id o 




l-O 
























E 2 






















>> 









ID 


o 




to 


ts 




m 


Average 

recoil 

permeabilit 






1 

o 


O 


O 


© 


<=> 


o 




O 






— ^^ 




""* *— - 


— ,_^ 


^^ --*. 




"■" ^^. 


£ 








*3 


o 


x co 


X iO 


i 


xS 






«H **» 




ee ** - 


00 














o 


ei 


o 


i- 


IO 




o 






CO 


cc 


o 




i' 


CO 




IO 


s 

si 

a to 


to 


s 


— - _ 


53 


ss 


si 


&2 


x c 




g§ 


< 


x^?L 


CI M 


« :*. 


w 2£ 


04 to 


re -+ 




** 12. 
























„ 1 




M 


^^ 






_^ 












C 


69 © 


x - 


•* © 


— o 


Cl p 


i- O 




1- c 










c: -* 


cc »— * 


l~- CN 








^s 


to 


J 


— C 


«■* 


re _ 


>- CC 


Tf* 1^ 


-»• Cl 


I 


re — 


03 




=: 


© ■* 


c re 


C re 


o 10 


6 ■* 


c -+ 




© ec 






" 


~"~* 


■* 


**^ 


._. 


*"" 




' — ' 


>> 






















I 

9 


to 


S 


sag 


?§ 


*% 


©2 


50 S 


o cj 


sf 


2o 


< 


yM* 


t '-- 


i- - 


CC'O 


- '•- 


•* w 


- JB 


t-o: 


o 












































~ 




S 




its "O 


"O in 










©© 


to 
5. 




=: Cl 


G3 <N 


<r. ci 


it; i~ 


ir: i^ 


10 1~ 


S3 >o 


1- CO 




~ 


c: — 


- — 


99 — 


ce — i 


re — 


ce —i 


Cl*— i 


Cl — 






















s 

41 






















to = 


.a 




?! 


?8 


- c 
v o 


r . Cl 




'?ci 


sS 


a 


2 


-d 


«o 


*»* 


°x 


OfcT 


°CO' 


°o 


°.o 


K 






























. 
















o 

ZC 1 


3 


"5" 






o 
o 

1 

1 




P 

— 
O 


















I 


_^ 


o 

S 

1c 


- 


a) 


8 






1 
3 








o 

H 


1 

o 


I 

i- 
2 

H 

o 


9 

O 

8 


< 

Cl 

o 


— 1 

l-l 

c 
o 

1 








c 


1 


s 


| 


< 


i 


S 








<g 


< ! 


< 


^ 


< 


M 


K 



. 



expected, the (BH) m(lx value was only slightly higher. However, 
the economic advantage from using an alloy containing cheaper 
base elements was very great and intensive research work on 
alloys of similar composition was carried out. 

Improved properties were obtained in Britain, and in 1933 the 
alloy known as Alni was marketed. Basically similar to the 
Mishima invention. Alni has an internal energy some 25 per cent, 
greater than that of 35 per cent, cobalt and, by slight variation 
of composition, the coercivity may be varied between the wide 
limits given in Table 2. 

Further research in Britain 3 led to the discovery of Alnico. 
Based on the earlier work on iron-mckel-aluminium alloys, Alnico 
also contains cobalt and copper which together have the effect of 
increasing the remanence without reducing the coercivity. An 
energy increase of approximately 40 per cent, over the performance 
of Alni was achieved with this alloy and, again, variation of 
characteristics can be obtained by careful control of composition. 
The coercivity may be varied between similar limits, i.e. 
4 X 10* A/m to 5 X IO 4 A/m (500-620 oersteds), and corre- 
sponding remanences of from 0-8 Wb/m 2 to 0-65 Wb/m 2 (8,000 gauss 
to <i.500 gauss) are obtained. 

The special alloys Hynico J 1 and Reco 2A are recent develop- 
ments to satisfy the requirement for an Alnico type of material 
having very high coercivity coupled with a {BH) milx product of the 
same order. They find application where the permissible magnetic 
length is restricted by other considerations, as in small generator 
rotors having four or more magnetic poles. 

Anisotropic Alloys 

In 1938, an experiment was carried out by Oliver and Shedden 4 
which has proved to have been a milestone in the history of the 
development of permanent magnet alloys. They found that if 
the cooling of Alnico from the high-temperature solution treat- 
ment (1,250° C) took place in the presence of a magnetic field 
strength of the order of 24 X IO 4 A/m (3,000 oersteds) it became 
magnetically anisotropic. Its properties in the direction of the 
applied field were improved at the expense of those in other 
directions, the internal energy increase being of the order of 
10 to 15 per cent, above that of specimens having similar heat 
treatments but without the application of the magnetic field. 



100 



PERMANENT MAGNET STEELS AND ALLOYS 




PERMANENT MAGNET STEELS AND ALLOYS 



101 



The significance of this performance improvement did not 
escape the notice of other workers, and the logical investigation 
of other alloys of similar composition which followed led to the 
development in 1940 in Holland of the alloy Ticonal 5 and in 
Britain of Alcomax I. 6 It was discovered that alloys having a 
higher cobalt content than ordinary Alnico could be made 
surprisingly anisotropic, the order of improvement due to the 
application of a magnetic field during heat treatment being 
increased from 15 to 200 per cent. Further composition modifica- 
tions and improvements in manufacturing technique produced 
improved versions of these alloys, namely Alcomax II 7 and 
Ticonal G. 8 

Continued research into these complex alloys resulted in the 
development of the alloy Hycomax, 9 which has a high coercivity 
of the order of 70,000 A/m (875 oersteds) ; unfortunately, this 
was obtained at the expense of a reduction in contained energy. 
Tli is alloy has only a limited application to special designs in 
which the magnet is subjected to high demagnetizing forces or in 
which physical or mechanical considerations limit the permissible 
length of magnet. 



H. OERSTEDS 
200 ISO K30 




0-025 0-02 0-015 OOI 0-005 

Fig. 3-1. — Demagnetization curves of magnet steels. 



102 PERMANENT MAGNET STEELS AND ALLOYS 



8000 




0-09 0-08 OO? 0-06 0-05 0-06 0-03 002 OOI O 

>J //.Wb/m z 

Fig. 3-2. — Demagnetization curves of ISOTROPIC diffusion iiardenin<; 

ALLOYS. 



Improvement in magnetic energy of the anisotropic alloys 
could logically be expected only by increase in coercivity, since 
their remanences approach the saturation density. It was found 
in 1950 that additions of niobium (columbium) or tantalum or both 
were beneficial when added to alloys of the Alcomax II or Ticonal 
types and had the effect of increasing the coercivity without 
reducing the contained energy. Two new materials were marketed 
following this development : Alcomax ITT 10 with very high 
energy and Alcomax IV 11 with very high coercivity. 

The magnetic characteristics of the isotropic and anisotropic 
diffusion-hardening alloys are given in Tables 2 and 3 respec- 
tively. As with all magnetic materials, the number of variables 
affecting the magnetic characters! ics is high, resulting in a fairly 
wide spread of performance between individual magnets. In 
listing the magnetic properties an attempt has been made to give 
the average values likely to be obtained from magnets not 
presenting any serious production problems. Figs. 3-1, 3-2 and 
3-3 give the demagnetization curves to which the listed charac- 
teristics refer. 



PERMANENT MAGNET STEELS AND ALLOYS 



103 



Columnar Crystal Magnets 

By careful control of the cooling of diffusion-hardening alloys 
from the molten condition, selective orientation of the crystal 
structure may be achieved. It has been found that with the correct 
crystal arrangement the anisotropic alloys within this range of 
materials can be made to have phenomenal properties far exceed- 
ing those obtained by normal manufacturing methods. 

A sufficient cooling differential between the sides and base of 
a casting causes a preponderance of crystals to grow from the 
base. If a proportion of such ctystals exceeding 75 per cent, of 
the whole of the casting can be made to grow throughout the 
height, improved properties may be obtained with the anisotropic 
alloys by making the subsequent heat -treatment magnetization 
in the direction of the columnar formation. The difficulties of 
production are great and the principle has not yet found extensive 
commercial application. The limitations on size and proportions 



H. OERSTEDS 
900 800 700 600 SOO 400 300 



aoo kx> 




0O9 0-O8 007 0-06 0-05 004 003* 
^W.Wb/m* 

Fig. 3-3. — Demagnetization curves anisotroi 

alloys. 



10 DIFFUSION UARDENINti 



104 



PERMANENT MAGNET STEELS AND ALLOYS 



of magnets made by the method are severe and the costs of 
production at present are high. 

The magnetic characteristics obtained in production are 
dependent largely upon the shape and proportions of the magnet, 
and whilst energies of over 63,500 J/m 3 (8-0 m.g.o.) have been 
obtained on laboratory-prepared samples, an average result of 
49,700 J/m 3 (6-25 m.g.o.) can be obtained from an ideal shape. 
The best results can be expected from magnets of cylindrical 
section of approximately 1 sq. in. and a length not exceeding fin. 

Materials made by this method are marketed under the trade 
names of Alnico D.G. (U.S.A.) and Columax (Great Britain). 

Sintered Magnets 

The well-known sintering method of manufacture of hard 
metals has been applied with considerable success to the manufac- 
ture of permanent magnets in materials of the diffusion-hardening 
type. The various constituent metals, crushed to 200-mesh 
powders of balanced particle distribution, are mixed according to 
the composition required and pressed into moulds of the desired 
shape. 12 Sintering of the pressing so formed is carried out at a 
temperature upwards of 1,350° C in a reducing atmosphere. This 
produces a hardenable magnet shape of a high order of homo- 
gemty and possessing mechanical properties often superior to 
those of cast magnets, undoubtedly due to the much finer grain 
structure obtained. 

The density of a sintered magnet material is approximately 
10 to 15 per cent, lower than the corresponding cast alloy, depend 
mg upon mass and shape. After correct heat treatment this 
results in a reduction in saturation density and residual induction 
of the same order. The (BH) max> or internal energy, is similarly 
reduced, but coercivity is not affected. 

This method of manufacture is successfully applied to most 
of the diffusion-hardening alloys, Alnico, Alcomax, Ticonal 
Hycomax, and it is particularly suitable for magnets of very 
small dimensions. A much higher degree of accuracy is quit. 
naturally obtained than with normal sand-moulded magnets, 
and often machining operations can be omitted. Greater freedom 
of design is permissible and magnets of more complicated shape 
are possible ; large variations in dimensions in the direction of 
pressing are undesirable, however. 



PERMANENT MAGNET STEELS AND ALLOYS 105 

The production costs of sintered magnets are comparatively 
high, and generally the method offers economic advantage m,!\ 
when the weight of the magnet does not exceed 20 gm. The aver- 
age weight of magnets made by the process is considerably less 
than this. 

Sintering involves the use of expensive tools and is economically 
practicable when quantities of 10,000 or more of a particular 
design are required. The difficulty which this imposes upon 
development and prototype production is overcome, either by 
the machining of one of the many designs for which tools exist or 
by cutting the shape required from a standard block before the 
sintering operation. 

PRECIPITATION-HARDENING ALLOYS 

The group of alloys listed, with their approximate compositions 
and magnetic characteristics, in Table 4 have been developed 
during the years 1931 to 1946, mainly in the U.S.A., where they 
have found limited application because of the ease with which 
they can be machined and cold worked, sometimes even in the 
magnetically hard condition. They are expensive to produce, and 
probably this and the fact that in general their magnetic pro 
perties are no higher than those of 35 per cent, cobalt steel are 
responsible for the lack of serious development of them in Britain 
and elsewhere. 

Comol, Comalloy, or Rcmalloy, as it is variously called, is a 
useful precipitation-hardening alloy having magnetic charac- 
teristics very similar to those of 35 per cent, cobalt steel. How- 
ever, it offers little advantage over the older steel for which the 
manufacturing technique is long 'established, and consequently 
has not been seriously developed in Britain. 

The Cunife alloys I and II— copper-nickel -i ion and copper- 
nickel-iron-cobalt— are easily worked even in the magnetically 
hard state. They are made anisotropic by heavy cold working, 
and the {BH) max in the direction of rolling may reach a value 
as high as 14.700 J/ m 3 (1-85 m.g.o.). 

Alloys of copper, nickel and cobalt— the American Cimico I 
and II materials— have contained energies rather lower than those 
of 35 per cent, cobalt steel, and are easily worked. Relatively low 
energy values and high cost have prevented the serious exploita- 
tion of these iron-free alloys. 



106 




PERMANENT MAGNET STEELS AND ALLOYS 





§2 


















81 

It 

Pi 5. 


si 


_« 


1 


_© 


o 


js 


A 




P 


a 


i 


a 


§ 


§ 


- 






« 


O 


n 


a 


q 


A 


fi 




ri 


















si 


- 


© 


5= 


eo 


w 


1 


i 




X 


00 


SO 


oo 


oc 




i 




>» 




















































1 »B 

8 


3«J 


9o 
'Pes 


51 


32 




*?© 


spg 


01 


x-«S 


■* '-2. 


C-l ™. 


■" 'c. 


M S 


,M M 


re ->■ 




o 
















o 


S? S 






























o 


o 

© S 
S T 


o _ 

Is 


-it '. 




If 


C ^~ 


© ^T 




C5 ^9 


i^^!. 




•e 


ȣ 


.-' s 


1~' ~ 


^rn. 


S 


* — ' 
































$ 


P ei 


















i m 


._ s 




sc © 


** © 


N © 




o 






•c o 


i- © 


M © 


>C © 


© © 










6'~ 


6 s . 


©t 


6« 


©« 


— © 














© 


© 




—* 












— 






"5 


IS 


1 


1 


1 


^ 


J 


■*• 


















B 


















u 

£ 


r* 


1 


1 


1 


| 




o 


c~ 


































JS 


















.be 
3 
is 

d 


a 




1 


1 


1 


1 


1 


1 


















<3 


1 


o 
to 


to 


s 


e<5 


1 


1 


.o 


































8 


a 


w 


1 




© 


-< 






c 


*H 










i^: 


IO 


£ 

6 


















& 


1 


© 


© 












1 


<N 


CI 


m 


<M 


1 


1 










• 


• 




• 


• 




1 


















E 


















1 


















a 


o 


M 


a 


H-l 


3 


>. 


i— i 






"5 


<? 


^ 


8 





o 


o 






5 


J 


i 

5 


J 


1 


1 

> 


.3 
> 



PERMANENT MAGNET STEELS AND ALLOYS 



107 



c - 



II 



3 .2 

'i 1 

18. 



•J C C -O 3 P 

SbT =- 3 

So! & = *j 
- ? tt- - 
§* a 3 - 

| = == 5*j ta 



r be = 

5 tj 



I M T 



"Pap" 



fr'ti 



1 

I 



o 

I 



p 



s§ 



fe 






o 



&J2 • 

c ca be 

.pes 



| r. 

— —■ s 






711 
■^ - - 

.-w _aj — 



>. 

13 



- 2 I 

%& "a — 






1^ 1^ X © — © 



~ f N 

•J: i- i- 



ia '- ic re '7 
r^- 1- i'- t^- 1^- 



i 



II © T»l -H <* « ' 

"t '"• ci ri — < — — 



K) «0 «p >0 

A A pi rt 



I I 



iC © © © © 

©•^©"TKreiricb 



i s 



© c 

sb at 



I I 



I I 



-j , "T' '?©©©©©© 



— bO 



5 - 

O © 



a) • 

JS If 



B ~- 
* o 

3 



.6 .J*, 



o c 



P -^ *i ^ 

P c S C 

c-p °» ©•a 8*a 

6 s * fc -2 ^ ® 



6 


o 

Q 


-fai _ 

P 0> 

°1 


••3 _ 

| 1 


114 





•= •= p " - o o £ o 



It 2 



108 



PERMANENT MAGNET STEELS ANT) ALLOYS 



The cobalt-iron-vanadiuni alloys Vicalloy I and Vicalloy II are 
two further examples of this class of material. Heavy cold 
working produces a high degree of anisotropy in the direction of 
rolling, and energies as high as 23.800 J/m 3 (3-0 m.g.o.) have been 
recorded for small sections such as 0-020-in. tape or wire. The 
very high cobalt and vanadium contents make them very expen- 
sive alloys and the application is limited : there has been no 
serious commercial manufacture in Great Britain. 

MECHANICAL AND PHYSICAL PROPERTIES OF 
P.M. MATERIALS 

An understanding of the mechanical and physical limitations 
of the various special magnet steels and alloys which have been 
discussed is most important in view of the limitations which these 
properties impose on the design and application of magnets. 

Whilst magnetic hardness is by no means synonymous with 
physical hardness, it is nevertheless true to say that, with those 
materials finding the greatest commercial application., good 
magnetic properties often go hand in hand with physical weakness. 
This is not altogether surprising when viewed in the light of the 
modern concept of magnetic hardness and high coercivity being 
due to high internal strain. 

The general physical properties of magnet steels and alloys arc 
given briefly in Tables 4 and "> and are detailed more fully below. 
The various methods of manufacture are also given. 

Carbide-bearing Magnet Steels 

The tungsten, chromium and cobalt steels have good physical 
and mechanical properties. They may be manufactured by the 
normal steel-making processes of casting into ingot form followed 
by subsequent cogging and rolling to the required section, or alter- 
nalively, when very complicated magnet shapes are required, 
these may be made by pouring from the shank into individual 
sand moulds. Forging of rolled bars into the !i U " or " C " 
magnet shapes which are common for magnets in these materials 
is usual practice, and presents no serious difficulties to those 
versed in the art. 

Machining by normal methods is possible in a fully annealed 
state (720°-740°C. slowly cooled) although very small drilled 



PERMANENT MAGNET STEELS ANT) ALLOYS 



1 09 



and tapped holes are to be avoided, particularly in the higher 
grades of cobalt. The 35 per cent, cobalt steel is the mo-i brit I le of 
this group of steels and, to minimize the danger of cracking during 
heat treatment, rapid changes of section in castings and sharp 
corners in forgings are undesirable. 

The Diffusion-hardening Alloys 

The highly crystalline nature of the Alni. Alnico. Akomax and 
Ticonal alloys causes them to have comparatively low mechanical 
strength. Manipulation by normal steel-making methods — 
rolling, forging, etc. — is not possible, and magnets have to be 
cast to the approximate final shape in individual sand moulds. 
The hard, brittle nature of these materials prevents normal 
machining methods, and consequently the shape must permit 
grinding to the required form. Considerable ingenuity is used in 
modern magnet foundries to produce the small and difficult 
shapes sometimes required, and individual sand castings of less 
than only 10 gm. weight are economically manufactured in very 
large quantities. A limited amount of drilling is possible after 
prolonged annealing of some of these allo3'S, but it is a difficult 
operation and is to be avoided where possible. Accurately 
positioned holes may be provided by the grinding of holes cored 
in the casting or for very small holes a larger, cored hole can be 
filled with soft metal and drilled to the size required. 

The Precipitation-hardening Alloys 

Magnets in these materials are usually machined from the 
rolled- or drawn-bar stock ; indeed, as has already been stated. 
with certain of them extensive cold working is necessary to 
produce optimum magnetic properties. In quite a number of 
cases manipulation and machining is possible in the magnetically 
hard state. 

Comalloy tends to be brittle although machinable, and care must 
be taken during manipulation to avoid cracking. Magnets may 
be made from rolled bar or as castings. 

THE PHYSICAL AND MECHANICAL ASPECTS OF 
PERMANENT MAGNET DESIGN 

The physical properties of permanent magnet alloys have been 
dealt with elsewhere. The problems met in the use of the avail- 



110 



PERMANENT MAGNET STEELS AND ALLOYS 



able alloys and methods of assembly and construction are 
discussed here. 

Magnetic design considerations establish the dimensions of 
length and sectional area of a magnet. These tend to be inversely 
proportional to the values of H c and B r respectively, and thus 
magnets in the older types of steel are long and of small section 
whereas magnets in Alni or Alnico tend to be of short length but 
of greater cross-section. Alcomax or Ticonal magnets of optimum 
design and for given performances are of similar length to corre- 
sponding ones in Alnico but have a sectional area of less than 
half. 

These very general considerations will give an indication of the 
general trend of magnet shape and proportions in those cases 
where the older steels have been replaced by the more modern 
alloys. It is an indication only because the newer alloys, having 
higher magnet energies, have made possible higher overall 
performances, and magnet dimensions have not been reduced 
in an inverse proportion to the increased internal energies (see 
p. 114) 

It will be obvious after study of magnetic design considerations 
that theoretically any air-gap performance can be achieved using 
any of the wide range of alloys and steels. However, from prac- 
tical considerations, it will be obvious that this is not so. The 
overall space available in many pieces of apparatus requiring high 
magnetic energies prevents the use of correctly designed magnets 
in the older low-performance steels. On the other hand, 
mechanical considerations sometimes fix the length of a magnet 
approximately, and if this should be much greater than the 
optimum for one of the newer alloys, it is possible that a lower 
coercivity material will be more efficient and economical. 

The choice of material can be made only after a study of all 
the factors involved and the calculation of optimum dimensions 
in several materials. 

Methods of Production 

The methods of production of magnet steels allow a considerable 
latitude in the choice of design. Their machinability and com- 
paratively high strength are such that the mounting of pole 
pieces or the fixing of the magnet into its associated equipment 
is comparatively simple by normal engineering techniques, pro- 



PERMANENT MAGNET STEELS AND ALLOYS 



111 



vided that the reservations made regarding tapped holes in the 
higher cobalt steels are borne in mind. 

With any of the nickel-aluminium or nickel-aluminium-cobalt 
alloys the inherent brittleness and low physical strength must be 
allowed for in design. Whereas with the magnet steels it is per- 
missible to mount associated parts on to the magnet this is not so 
with the newer alloys. 

The shape of magnets in these materials must permit of mould- 
ing in sand, and for ease of production complicated designs 
requiring the insertion of cores should be avoided where possible. 
Rapid changes of section are undesirable, since these are liable to 
cause subsequent cracking in heat treatment. 

Small holes in castings are a source of trouble, and a length/ 
diameter ratio of not more than four is desirable ; where possible, 
external slots should be used instead, since these can be positioned 
more accurately. Such slots permit the use of clamping bolts to 
fix the brittle magnet to its pole pieces or housing. 

Various other methods of assembly other than by the clamping 
of the magnet are possible. It is common practice to mould a 
magnet, together with any associated pole pieces, into a pressure 
die-casting using a non-magnetic zinc-base alloy. This method 
is particularly suitable for the rotors of alternators or magnetos, 
where the magnet -pole pieces are often laminated, and sometimes 
the whole assembly can be moulded around a shaft, final machin- 
ing being carried out after die-casting. With the alloys of low- 
temperature melting-point which are used, the heat transference 
to the magnet is nol sufficient to affect the magnetic properties. 
A particular advantage of such methods of assembly, especially 
with rotating parts, is that the low mechanical strength of the 
magnet becomes of quite secondary importance since it is com- 
pletely encased. 

Pressing of magnets into pole pieces is possible, but is advisable 
only if the stresses to which the magnet is subjected are com- 
pressive. The. pressing of shafts into the ground bores of rotor 
magnets is deprecated, since the stresses created may be sufficient 
to cause cracking, which docs not necessarily occur immediately. 

Brazing is a practicable method of assembly provided that a 
low-temperature brazing medium is used so that the magnet is 
not overheated. High-frequency methods are admirable, since 
the heat can be localized. Soft soldering is also possible, the great 



112 PERMANENT MAGNET STEELS AND ALLOYS 

difficulty being the adequate cleaning of the surfaces. This can 
sometimes be overcome by copper plating the parts to be joined. 

Fastening by means of synthetic resin adhesives is a practical tie 
method of assembly, but the method is usually restricted to 
magnets of small dimensions. Where the magnet block is suitably 
encased, positioning by means of a low-melting-point alloy, 
sulphur, or a thermosetting plastic medium, is sometimes adequate. 
Such methods are often used for the fixing of spindles into the 
central cored hole of small alternator rotor magnets. 

Where unit construction of pole pieces is employed, magnetic 
attraction is often sufficient to retain the magnet. This is parti- 
cularly so with the high-performance anisotropic alloys having a 
high working flux density (Pull a £ 2 ), and when the magnet 
dimensions are small. 

PERMANENT MAGNET DESIGN 

A permanent magnet invariably works under conditions of de- 
magnetization caused by the field it establishes in the surrounding 
space or in an air-gap forming part of a more clearly defined 
magnetic circuit. The sense or direction of the gap field is 
opposite to that of the field inside the magnet, which is conse- 
quently partially demagnetized. An examination of the hysteresis 
loop diagram (Fig. 3-4) shows that the conditions under which 
the magnet maintains a positive flux under demagnetization occur 
when the working point lies in the upper left-hand quadrant. The 
magnet flux density falls as the demagnetizing force increases, 
but in the absence of any externally applied magnetization the 
working condition cannot fall below the point (— H = II c . B = 0). 




REMANENCE 



SATURATION 
DENSITY 




Fig. 8-4. — Hysteresis loop 
SHOWInc PRINCIPLE POINTS. 



PERMANENT MAGNET STEELS AND ALLOYS 



113 



When the magnet supplies a static flux with no recoil, the working 
point of the flux density within the magnet falls from the initial 
density B s to a point on the demagnetization curve at which — U, 
the demagnetizing field per unit magnet length, is sufficient to 
meet the demands of flux-maintenance in the external circuil 

Let a uniform gap of length L g be cut in a magnet -ring having 
the hysteresis loop partly shown in Fig. 3-5. Suppose the working 
point is X, corresponding to a magnet flux density B x and a 
demagnetizing field — H x . If the magnet has a length L m and a 
cross-section A m , its flux Mill be B x A m and the m.m.f. released for 
the gap will be H x L m . 

The flux B x A m will, as a simple approximation, become B g A a 



Fig. 3-5 Demagnetization and 

ENERGY CURVES WITH GEO- 
\H • l'lUCAL CONSTRUCTION FOR 

HH ,„.,.. 




H m H x 



BxH 



in the gap, if the gap area A g is assumed to have a uniformly- 
distributed density B a and leakage is neglected. Consequently 

B 1 A m ^B g A a w<\H 1 L m = H u L = B g LJ l x Q . . (1) 

These mean that the gap flux is the same as the magnet flux 
(which is a natural consequence of the circuital nature of magnetic 
flux) and that the m.m.f. released by the fall of magnet flux is 
applied to maintain that flux in the gap. 

The energy W per unit volume of a gap permeated by a magnetic 
flux density B g associated with a field strength H g is 

W = \B g H g = \B*ly. Q . 
For gap dimensions L g and A a the total energy W stored in the 
gap is 

W - IB'A.LJk 

= h B l H l A m L m 

= $(B 1 H 1 )U m (2) 

using the relations in equation (1) and writing the magnet volume 



114 PERMANENT MAGNET STEELS AND ALLOYS 

as A m L„ = U m . Thus the stored energy in the gap corresponds 
■to that released by the magnet, and the product ^(B^^ is the 
gap energy produced per unit volume of magnet. 

The (BH) mtur Point 

A curve connecting the product (BE) with B has the charac- 
teristic shape shown to the right in Fig. 3-5. The value is zero 
for points B r and 1I C , and rises to a maximum intermediately. 
The position of this point, known as the (BH) miIX point, may be 
found from suitable measurements : it is a most important 
criterion of the usefulness of a permanent -magnet material. 

For a rectangular hyperbola, to which most demagnetization 
turves closely approximate, this point of maximum energy is 
given by the intersection of the diagonal of the rectangle aB r OH c 
in Fig. 3-5, with the demagnetization curve. 

For a given air-gap energy, U m will be a minimum when B 1 
and H x are the co-ordinates of this maximum point : the most 
economical design using the least volume of magnet material will 
be obtained therefore if the working point of the magnet occurs 
at the {BH) mnx point, 

The formulae in equation (1) take no account of leakages 
occurring in practical magnet systems. The path of the magnet 
external flux is not confined to the gap, flux emanating from the 
whole surface of the magnet and not just from its pole faces or 
extremities. The amount of flux external to the air-gap, usually 
called leakage flux, is dependent upon the area and length of 
gap and the flux density within it. It increases as gap flux 
density and length increase and is reduced as the gap area is 
increased. 

The integration of all the circuit reluctances is extremely 
difficult, and since with particular designs there exists a more or 
less constant ratio between the gap or useful reluctance and .ill 
the others, it is usual to make use of a leakage coefficient to 
simplify design. This assumption is further justified since in 
many designs by far the greater part of the flux leakage occurs 
in the immediate vicinity of the air-gap. Usually designated K a , 
this leakage factor is equal to the total magnet flux divided by 
the gap or useful flux (K„ = &J& a ). This coefficient is by no 
means constant and it increases with increase of flux density in 
the air-gap. Its value varies very widely, being seldom less than 






PERMANENT MAGNET STEELS AND ALLOYS 



115 



1-5 even for very short gaps where the leakage is lowest, and 
rising to as much as 30 in extreme cases where the flux density 
is high and the gap length great. An example of how the leakage 
coefficient K„ varies as the flux density in the gap of a loudspeaker 
magnet is increased is shown in Fig. 3-6. 

Effects of Small Caps 

It is customary to introduce a second coefficient into design 
formula} to take into account the effects of small gaps at joints 
in the magnetic circuit, the reluctances of pole pieces and the 
curvat ure of lines of force within the actual gap. This coefficient 
(K t ) is usually combined with the gap length in computing the 

o 



Fig. 3 0. — Grass showing 

THE VARIATION OV IHB 
SECTION KEAKAfiE COKI'l'l- 

cient with 1'i.rx iiknsity 
1N A LOUDSPEAKER iUAHNET 

H AVI Mi AN AlKGAr 1 IN. 

DIA. X 0-04 IN. LONG ami 

0-187 IN. DEEP. 



H 
Z 
UJ 

y 
u_ 

LL 
UJ 

o 
u 

LU 



< 
UJ 



• i 

, ?/- 

1 —- 



0*4 



0-8 



1-2 



1-6 



AIR-GAP FIELD STRENGTH w/m 2 

total gap m.m.f. It does not vary so widely as the leakage 
coefficient and is generally between 1-1 and i-5. 

In the practical case therefore equation (1) cannot be used since 

*„ = K a . g 

therefore B m . A m = B, . A u . K u (3) 

similarly //,„ . L, n = B g . L . K t X l0»/4r .... (4) 

and B 1n . H„ . U m - B,* . A g . L, . K a . K, X 10'/*r . (5) 

It is not possible to give methods for computing these widely 
varying coefficients, and it is generally only by pre-knowledge of 
a particular form of magnet that the design engineer is able to 
make an accurate estimate of their values. With unknown de- 
signs, whilst it is possible to calculate the reluctance of the 
various leakage paths and so to determine what flux will pass 
along them, probably the most satisfactory way is by careful 
measurement of fluxes within various parts of the magnetic 
circuit of an approximate prototype. Once these factors have 



116 



PERMANENT MAGNET STEELS AND ALLOYS 



PERMANENT MAGNET STEELS AND ALLOYS 117 



been established, the design of magnets to give a given per- 
formance within an air gap is a comparatively simple matter, 
using equations (3), (4) and (5). 

Permeance Lines 

The solution of design problems is facilitated if permeances are 
used as opposed to reluctances. The permeance P of a circuit is 
the reciprocal of its reluctance S and is the sum of the individual 
permeances of the various parts of the magnetic circuit. 

Related to the magnet 

P = 1/S = <Z>/m.m.f. = B m . AJH m . L m . (6) 
Unit permeance, 13 p, or the permeance as seen from unit 
volume of the magnet, is given by the equation 

p = P.LJA m = B m IH m (7) 

From equations (3) and (4) 

B m = B, . A, . KJA„ and H m == *^> ■ K ' '' W 



so that 



and therefore 



P 



P 



B g . A, . K a . 4 



L m X 4tt 

X 4tx 



B g .L .K l .A m x 10' 



A„ . L . K 



X 



4tt 

To 7 



L u .A m .K l - 10? • • • W 

For a given magnet design therefore the unit permeance is a 
numerical constant which is independent of magnet charac- 
teristics. For a particular magnet design which is ideal, in that 
the working point is the (BH) max point, it is equal to and passing 
through BJH m . A line drawn through the intersection of the B 
and H axes at this point is known as a permeance line. 

Once the permeance for a magnet design has been established 
the effect on the performance due to a direct interchange of 
magnets of different materials is easy to determine. In Fig. 3-7 
the ideal permeance line passing through the (BH) max point for 
Alnico, flux density 0-472 Wb/m 2 , intersects the demagnetization 
curve for Alcomax III at flux density 0-772 Wb/m 2 . The magnet 
working flux density is increased therefore from 0-472 to 
0-772 Wb/m 2 and the gap flux would tend to be increased in the 
same proportion. However, the internal energy of Alcomax III 
at this particular point is only 3-62 x 10 4 J/m 3 (4-54 x 10 6 m.g.o.) 
whereas the maximum energy obtainable for an ideal design 







XIO 

Fici 3-7 —Comparison of alotco and alcomax hi BKOWJNG mow the PER- 
MEANCE OF A MAGNET DESIGN MliMT HE VAHIED TO ACHIEVE THE MAXIMUM 
UTILIZATION OF THE AIXOY. 

is 3-98 X 10 4 J/m 3 (5-00 X 10 6 m.g.o) and therefore the most 
economic utilization is not obtained. To give the highest efficiency 
the magnet proportions would have to be altered until the 
permeance was 2-36 x 10- (BJH m = 18-8 in the c.g.s. system). 

Recoil Lines. Reversible Permeability 

If a magnet is subjected to a greater demagnetizing force than 
that produced by the negative field of the air-gap the working 
point N (Fig. 3-8) moves down the demagnetization curve to 
some point N 2 determined by the new set of conditions. This 
occurs when the reluctance of the flux path is increased, as for 
instance when the armature of a generator is removed from the 
field magnet. When the additional reluctance is removed (i.e. 



1J8 



PERMANENT MAGNET STEELS AND ALLOYS 



when the armature is replaced) the flux density in the magnet 
does not recover along the original demagnetization curve but on 
a minor hysteresis loop (N X QR in Fig. 3-8) to the value 11 where 
ORN is the normal working permeance line. The effect is rever- 
sible along the top of this minor hysteresis loop (RPNj). 

It is usual to assume that the minor loop is in fact a straight 
line along which this reversible action takes place. The slope of 
such a line is known as the incremental, reversible, or recoil 
permeability of the material ; it varies according to the point of 
origin on the demagnetizing curve but between the points B r 
and H c the limits are sufficiently close to warrant the adoption of 
a single value for this constant. 




Fig. 3-8. — Dema < : \ ktiza- 

TIOS CURVE MINOR 
HYSTKHESIS LOOP. 



Stabilizing 

Deliberate depression of the working point of a permanent 
magnet, by subjecting it to a greater demagnetizing force than 
that to which it is normally subjected, is done in instances where 
the magnet is likely to come under the influence of additional 
stray fields in order that the final working flux density shall not 
be permanently lowered by such fields. 

Due to the reversible action of a magnet working on a recoil 
line, the effect of the application of any field less than that repre- 
sented by the H co-ordinate of the point of origin of the recoil line 
(OH in Fig. 3-8) is to depress temporarily the working point of 







> PERMANENT MAGNET STEELS AND ALLOYS 



119 



flux density. On removal of this excess field the working point 
recovers along the recoil line to its original value. 

The stabilizing effect of such intentionally applied additional 
demagnetizing fields is necessary in many pieces of apparatus 
where a high degree of constancy of magnet performance is 
required. A few instances are moving-coil instrument magnets, 
energy-meter braking magnets, and magnets for magnetron 
oscillators. The degree of stabilizing is usually controlled so as 
to give a flux drop of not less than 5 per cent. It is effected by 
open-circuiting the magnet by removal of its pole pieces, for 
generators and magnetos by the removal of the rotor, or by 
deliberate partial demagnetization by the application of A.C. or 
reversed D.C. fields. Generators are sometimes stabilized by short- 
circuiting the windings whilst the machine is running at speed. 

A secondary advantage of stabilizing is that by selective treat- 
ment and individual measurement it is a means of providing 
magnets of much greater performance consistency than is normally 
obtained. 

Efficient Utilization with Magnets on Recoil 

Controlling the working point of a magnet supplying static flux 
with no recoil, so that it coincides with the (BH) max point, is a 
comparatively straightforward matter. With magnets working 
on recoil, control of leakage is important if the most efficient 
utilization is to be obtained. This will not necessarily be obtained 



Fig. 3-9. — Working of 

MAGNEI I MJEU roNDl- 
TIONS OK RECOIL. 




120 



PERMANENT MAGNET STEELS AND ALLOYS 



with a design in which this is a minimum. The useful energy 
developed by a magnet working at point R in Fig. 3-9 is propor- 
tional to ST X UT, and it is easily shown that this will be a 
maximum for the particular recoil line if the point R is midway 
between Nj and V. This maximum value of recoil energy 13 is 
proportional to CX/4. 

Recoil energy contours plotted for a series of recoil lines have 
the general appearance of Fig. 3-10, which is drawn for Alcomax 
III. -Maximum recoil energy is obtained at the point marked and 
is realized when the leakage and working permeance lines are 
respectively ON t and ON. Any other leakage and useful 
permeance lines will give a useful energy of less than this value 
and the necessity for controlling the leakage as well as the 
working permeance will be obvious. Comprehensive data on 
recoil energies has been compiled for most of the materials shown 
in Tables 1, 2 and 3. 14 




Fig. 3-10. — Recoil energy contours fob \i< omax iii. 



PERMANENT MAGNET STEELS AND ALLOYS 



121 



Non-static Working Conditions 

With many pieces of apparatus the working permeance is by 
no means constant .' 5 In such cases the working point oscillates 
along a recoil line between limits set by the conditions of maximum 
and minimum permeance. In generators, apart from the con- 
tinually changing demagnetizing effects of varying current in the 
armature, the reluctance of the magnetic circuit may \ar\ 
according to the position of the rotor. In Fig. 3—1 1, as load is 
increased the working flux density in the magnet falls from 01^ 
to OB 2 . 

A similar set of conditions is found with lifting magnets and 
magnetic chucks. In these cases open and practically closed 
circuit conditions may be obtained so that the working point may 
move along the full length of the recoil line. 



MAX. 



Fig. :* ] l. -Recoil condi- 

TIm\-; WITH VAtlVINC! WORK- 

1NG BBSOKANGB. 




OPEN CIRCUIT 
OR STABILIZED 
PERMEANCE 
LINE 



MAGNETIZING, TESTING AND DEMAGNETIZING OF 
PERMANENT MAGNETS 

The operations of magnetizing, testing and demagnetizing 
permanent magnets are important to manufacturer and user alike. 

The complexities of magnet manufacture are such that many 
makers prefer to test every magnet made in addition to the 
almost universal practice of sampling casts or batches. On the 
other hand, where large batches of a particular type are involved, 
quality control has certain advantages. It is equally important 



122 



PERMANENT MAGNET STEELS AND ALLOYS 



for the user to be able to test, as well as magnetize, his magnets 
in order to detect unsuitable pieces before expensive assembly 

operations have been carried out. 

It is common practice for the user of permanent magnets to 
carry out his own assembly. Apart from the obvious advantages 
of the ease of assembly of unmagnetized pieces, serious reduction 
of performance is likely to result if magnetization is carried out 
before the magnet is assembled into its completed magnetic 
circuit. This last consideration of course does not arise in the 
comparatively few eases in which the magnet is completely 
finished when it leaves the manufacturer, 

Magnetizing 

The problem here is the provision of a field adequate to produce 
a flux density at least as great as the saturation density (B s in 
Fig. 3-4) uniformly throughout the magnet. A very rough guide 
is that the magnetizing field strength should be greater than five 
times the eoercivity of the magnet in question. 

Probably the simplest form of magnetizer is the straightforward 
solenoid type in which bar magnets are passed through the centre. 
Momentary energization of the solenoid is adequate, the duration 
of current required being generally considerably less than one 
second. This short time rating means that the current loading of 
such solenoids can be several times that of normal electrical 
engineering practice, and up to 15,000 amp/sq. in. is common 
practice. 

A more efficient type of magnetizer uses a soft-iron return 
circuit for the flux. As shown in Fig. 3-12 (a), the magnet is 
allowed to complete the flux circuit, and it is often convenient for 
the magnetizing gap, usually adjustable, to be external to the coil. 

A modification of the solenoid type of magnetizer is shown in 
Fig. 3-12 (d), where a thin coil of very short axial length is used 
for magnetizing cobalt or tungsten steel U-shaped magnets. 
In order to produce a sufficient m.m.f., this type of magnetizing 
coil has to be very heavily stressed, and where possible magnetiza- 
tion of U- or arch-shaped magnets is done using the leakage 
field from core or pillar type magnetizers as shown in Fig. 3-12 (b) 
Multipole magnets are magnetized by means of an apparatus 
similar to that shown in Fig. 3-12 (c). 

When the magnet material forms the greater part of a practi- 






PERMANENT MAGNET STEELS ANT) ALLOYS 123 







Pig. 3 12. V.VHIOl S TYl'KS 0* MACJXICTIZKi:. 

(a) Two-pole yoke typo with adjustable gap. (b) Tillar magnet iaser for "" D " 
rntiKiipls. (/•) Multi-polo mugnctizor. (</) King coil and plate for " D " magneto. 

cally closed magnetic circuit and the gap is so small as to prevent 
the introduction of suitable coils, a heavy-current magnetizing 
method has to be used. Unidirectional impulses of high current 
at low voltage are passed through a coil of one or more turns of 
heavy-section copper threaded through the aperture of the 
magnet. 

A D.C magnetizing transformer is used to provide this surge 
of several thousand amperes which flows in a heavy section 
secondary due to the build up of primary current, or, mote 
commonly, to the collapse of flux in the transformer core when 
t he D.( '. primary current is interrupted. The ampere-turn product 
produced by the single-turn magnetizing coil is adequate to give 
magnetization to the saturation point if correctly designed. 

The overheating of copper and the bulk of such magnetizing 
transformers are a great drawback with this type of equipment, 
which is being superseded to some extent by what is known as 
the " ignition magnetizer." Essentially an A.C. equipment, the 
apparatus permits the passage of several hundreds of amperes 
during one half cycle of the supply frequency. Transformation 
produces currents of up to 60,000 amp in suitable magnetizing 
coils. The apparatus uses an ignitron valve connected to a 



124 PERMANENT MAGNET STEELS AND ALLOYS 

triggering circuit, the peak half-cycle heavy current passing auto- 
matically when the correct circuit conditions have been reached. 
Since the heavy current passes for such a short period of time, t he 
actual power used is quite low. 

Small bar magnets may be very conveniently magnetized in 
the gap of much larger magnets maintaining fields of 24 X 10 4 A/m 
(3.000 oersteds) or more. 

Testing 

The B.S.I. Panel. This equipment affords a means of obtaining 
the demagnetization characterise LCfi of permanent magnets or of 
control test pieces quite accurately and quickly. It is suitable 
for testing on one axis only and pieces should have a direction of 
magnetization which is more or less straight ; except where the 
test is being used for comparison only, the section must be uniform 
throughout the length. The ends of magnets to be tested using 
this equipment must be ground Hat and parallel. 

The apparatus consists essentially of a Faraday homopolar 
generator, the magnetic field for which is supplied by the magnet 






PERMANENT MAGNET STEELS AND ALLOYS 



125 



MILLIVOLTMETER FOR 
MEASUREMENT OF H 



MILLIVOLTMETER FOR 
MEASUREMENT OF B 





Fig. ."{ 18. — Theoretical diaqkam of b.s.i. testing equipment 406 : 



1031. 



Fig. :-J-i 4. — B.S.I, testing lppabatus 406. 

under test, and a series of instruments, switches and resistors for 
measuring and controlling the generator output and the current 
used in the magnetizing coil. 

A connection diagram is shown in Fig. 3-13. and the arrange- 
ment of the instrument panel and generator is shown in Fig. 3-14. 

Flux from the magnet passes through the Permalloy pole pieces 
and disc of the generator. Driven at constant speed by the motor, 
the disc will have generated in it between its centre and periphery 
an c.m.f. proportional to the flux crossing it. This e.m.f. is 
divided by means of a resistance network known as the section 
compensator in such a way that only a fraction of it is measured 
by a sensitive millivolt met cr (shown as a Yi-metcr). The resistance 
network is so calibrated that this fraction is inversely propor- 
tional to the section of the sample, and in consequence this 
millivoltmeter can be calibrated to indicate flux density directly. 

The magnetizing or demagnetizing field is controlled by the 



126 



PERMANENT MAGNET STEELS AND ALLOYS 



magnetizing switch and control resistances, the demagnetizing 

field -H being measured by a second millivolt meter. A second 
resistance network— the length compensator, which is similar to 
that used with the £-meter— enables this instrument to be 
calibrated directly in units of H. 

Other devices of the apparatus include a shunt selector switch 
to extend the range of the //-meter and a potentiometer permitting 
compensation for field flux crossing the disc but not passing 
through the magnet. The apparatus at the present t ime is manu- 
factured and calibrated only for C.G.S. units and measures 
directly only the 4-nJ/H curve. Correction to the true B II curve 
is easy, and the reader is referred to the specification for this 
instrument, which gives a full explanation of other corrections to 
be made to give the most accurate results." 1 

The Fluxmetcr. Probably the most widely u^vd instrument for 
magnet testing, particularly where large numbers are involved, is 
the fluxmeter. l T sed in conjunction with suitable easily-made 
search coils, it affords a. most accurate means of measuring fluxes, 
at a point or over an area within an airspace, or within the section 
of part of a magnet. It is essentially for testing finished magnets 
where the conditions of demagnetization are the self-demagnetiz- 
ing ones and is a menus of measuring fluxes equal or proportional 
to some point on or within the demagnetization curve. 

It is Similar to a moving-coil galvanometer except that it has 
negligible restoring torque, the angular movement of its pointer 
or scale being proportional to the (lux linkage change produced in 
a search coil connected to it. Usual calibration constants are 
3,000. 5,000 and L0,000 maxwell-turns per division, corresponding 
to .'{ K. », 6 x 10-s and 10 x 10~« webcr-turns. the normal 
scale length being 120 divisions with the pointer type and 240 
divisions with a projected scale version. 

Calculations of flux or flux change are made, using the simple 
formulas 

Flux = ^ eflection ■ ,S(,;lI t > <'onsta.nl 
No. of search-coil turns 

Flux Density B = - Deflection x Scale constant 

No. of search-coil turns Mean search-coil area. 
The Oauasmet&r. This instrument is convenient for measuring 
field strengths at a point since it gives a steady deflection and does 
not depend upon relative motion. 




PERMANENT MAGNET STEELS AND ALLOYS 



127 



A probe element consisting of a moving-coil carrying constant 
current is inserted into the field at the point at which measure- 
ment is required. The moving-coil rotates against the restoring 
torque of hair springs and a pointer indicates the deflection which 
is i n oportional to the flux density at the point. Alternatively, the 
moving-coil is replaced by a small jDermanent magnet, also 
controlled by hair springs. Rotation to the point of maximum 
deflect ion is necessary. 

The Ballistic Galvanometer. Testing methods using the ballistic 
galvanometer are usually confined to laboratory work in the 
accurate determination of hysteresis loops or, more particularly, 
demagnetization curves. The most accurate results are obtained 
on ring specimens, but these are inconvenient to use and. with 
modern hirfl-coercivity alloys, only small errors are introduced 
by using rectilinear or cylindrical test pieces with a high-per- 
meability yoke or pole-piece arrangement. It is usual to wind 
the search coil for the measurement of flux density onto the test 
piece and to use an accurately fitting magnetizing coil over this. 
Field or H determinations may be made by a direct computation 
from the product of current and magnetizing coil turns or alterna- 
tively by measurement using #-coils clamped onto the sample. 
These //coils may be of the flat type or the semi-circular Chattock 
potentiometer type, and must be individually calibrated. All 
pole-piece joints should be ground flat and smooth to cut doAvn 
the yoke reluctance. 

The circuit used is the basic one for magnetic measurements, 
and is shown in Fig. 3-15. Calibration is usually effected by a 
mutual inductance method, and points on the hysteresis loop 




f§-( ^GALVANOMETER 



■{THvV^v- 1 



Fii/.X 16. Ballistic calvaxomktkr cihccttay. 



128 



PERMANENT MAGNET STEELS AND ALLOYS 



are obtained by the step-by-step method given in most text -books 
on elcctrotechnology. 

Occasionally the ballistic galvanometer is used for production 
testing of permanent magnets where extreme accuracy is required. 
The flux-linkage changes necessary to produce galvanometer 
deflections and thus to permit measurement are usually obtained 
by the use of search coils as with flux meter testing. 

Special-purpose Tests 

The most important criterion of a magnet's usefulness is the 
performance which it maintains in its final design, or in conjunc- 
tion with its associated equipment. Where possible production 
testing is designed to simulate the final working conditions as 
nearly as possible. This is comparatively simple in the case of 
such magnets as those for loudspeakers, instruments and cathode 
ray tube focusing, but is not practicable for magnets for generators, 
motors and magnetos. For these magnets it is usual to arrange 
an arbitrary test to reproduce the final conditions of demagnetiza- 
tion as nearly as possible and to measure the flux produced by 
the magnet under these conditions. A quite common method is 



■■■■■ 




IB TKSTIN.; M\(;yj:ts I'NDKK CO.NTKdl i.kh Condition.- 
OF DE.MACNKTIZATJON. 



S^iSSSSSl^&SSSSS^^B 



PERMANENT MAGNET STEELS AND ALLOYS 



129 



to measure the flux change in a small iron yoke on removal of the 
magnet. The necessary demagnetization is controlled by adjust- 
ment of the air-gaps in the yoke or by applying a definite fixed 
demagnetizing current to a suitable winding. It is convenient to 
house such a device within the actual magnet izer — as shown in 
Fig. 3-16. 

Demagnetizing 

Demagnetizing of permanent magnets is necessary many times 
during production in order to facilitate handling during finishing 
operations. It is most commonly done by means of a solenoid 
cairying alternating current at mains frequency. Provided that 
the maximum field at the solenoid centre is sufficient to effect the 
reversible magnetization of the magnet to saturation point, the 
gradual weakening of the field on withdrawal of the magnet leaves 
it in a sufficiently demagnetized state. Alternatively, with the 
magnet at the solenoid centre the current may be gradually 
reduced to zero. 

The excessive loadings necessitated by modern high-coercivi ty 
alloys can be dissipated by water circulation in a low-voltage high- 
current coil made of copper tubing. This method is advantageous 
where quantities of large magnets are involved. 

When the magnet has a practically closed circuit, or when its 
assembly is such that it is surrounded by a conducting metal, for 
instance a die casting, the demagnetization possible by the above 
methods may be inadequate. This is because of the closed circuit 
in the one instance and because of the shielding effect of eddy 
currents induced in the conducting metal, in the other. For such 
magnets demagnetization must be done on the equipment 
normally used for magnetizing, the direction of current being 
continually reversed and gradually reduced to zero. 



THE HANDLING AND STORAGE OF MAGNETS 

For a variety of reasons probably 95 per cent, of all permanent 
magnets arc supplied in an unmagnetized state. Where subse- 
quent assembly has to be carried out, as for instance 1 lie assembly 
of magnets in a flywheel magneto, unmagnetized pieces are easier 
to handle. Furthermore, open-circuiting may seriously reduce the 
performance, as is explained under " Permanent Magnet Design." 



130 



PERMANENT MAGNET STEELS AND ALLOYS 



A further point is the difficulty of packing magnetized magnets 
so that one will not demagnetize another. 

Handling of Unmagnetized Magnets 

The safe handling of unmagnetized permanent magnets is no 
more difficult than the handling of other steel parts so long as the 
inherent brittleness of so many of these special highly alloyed 
materials is remembered. The damage which can occur to magnet- 
ized magnets is particularly serious, however, in that there is no 
apparent effect and the damage may not be detected until the 
equipment is in the final stages of production and test. Careful 
precautions are necessary therefore to circumvent the possibility 
of magnetic damage in all operations of manufacture subsequent 
to magnelizat ion. 

Demagnetization and Flux Distortion 

Magnetic spoiling may take the form of direct demagnetization 
or of flux distortion at the pole faces. The former may happen 
due to the presence of ojjposing magnetic fields of other magnets 
(i.e. when they are in repulsion) and occurs to a limited extent if 
any part of a magnet is short-circuited by some iron or steel body. 
The effect of open-circuiting or of removing a magnet from its 
pole pieces constitutes a direct demagnetization. Stray fields from 
electrical equipment may be sufficient to cause a permanent flux- 
reduction. 

Where stabilized magnets are concerned, it will be appreciated 
that permanent harm from any of the above effects will only 
result if they are in excess of the stabilizing field. 

Pole-flux Distortion 

Pole-ilux distortion can occur when two magnetized magnets 
which are in attraction are separated by sliding motion. It also 
happens to a limited extent when magnets are slid from a keeper 
or set of pole pieces. The effect in either case is a cross-magnetiz- 
ing of the magnet in the immediate vicinity of the pole tips and a 
permanent distortion of the magnet flux when the motion is 
completed. Magnets should therefore be separated by a direct 
pull or where this is not possible by a lever and hingeing motion. 

The reduction in performance due to any of these effects is 
instantaneous and is recoverable only by remagnetizing. The 






PERMANENT MAGNET STEELS AND ALLOYS 



131 



seriousness will be obvious with magnets which have been 
stabilized or which have had their flux adjusted within close limits, 
as is the practice with some instrument magnets and with those 
for radar. 

Prevention of any of the above demagnetizing effects is quite 
easy and it will be obvious that the most important factor is the 
adequate separation of magnetized magnets. If the design is such 
that the leakage field is great, a separation of 1 in. or more is 
necessary, and because of the attraction between magnets, this 
is best obtained by the use of non-magnetic trays or boxes. 
When the leakage field is low. as is often so when the air-gap 
length is small, a spacing less than this is sufficient and cardboard 
or paper wrappings may be adequate. Experience is the best 
guide, and one should err on the side of safety by having as great 
a spacing as is practicable. 

Non-magnetic Tools 

Non-magnetic tools and gauges should always be used with 
magnetized magnets. 

MAGNET APPLICATIONS 

Electricity Meter Magnets 

Large numbers of magnets for electricity meters are manu- 
factured in tungsten, chromium and the lower grades of cobalt 
steels, proved designs having remained unchanged in some cases 
for many years. Long-term stability is of paramount importance 
with magnets for electricity meters, and since this can only be 
proved after a long time interval, the reluctance of manufac- 
turers to change is understandable. However, the increased 
coeicivities and consequent greater resistance to demagnetization 
of newei- alloys make them particularly suitable for applications 
such as electricity meters ; and, furthermore, they are not sus- 
ceptible to shift of characteristics due to the gradual precipitation 
of carbides as are the older steels. Consequently more and more 
meter manufacturers are re-designing their meters to use the 
newer diffusion-hardening alloys. 

The very greatly reduced length/section ratio of optimum- 
efficiency magnets in these allo}'S has necessitated a complete 
readjustment of ideas regarding braking-magnet design. 



132 PERMANENT MAGNET STEELS AND ALLOYS 



^^ 





MAGNET 




DIE CASTING 

(a) (b) ( c ) 

Wig. 3-17. — Et/ectricity meter braking magnet system. 

Fig. 3-17 (a) shows the general type of magnet made in the 
< >l< ler steels. The difficulties of providing a sufficient clearance for 
the meter disc when an anisotropic magnet is used is overcome either 
by fixing a magnet in a suitable frame and providing a soft-iron 
return path for the flux or by using two smaller magnets having 
a common flux path as shown respectively by Figs. 3-17 (b) and 
3-17 (c). 

Instrument Magnets 

A conventional form of moving-coil instrument magnet suitable 
for manufacture in any grade of magnet steel up to 35 per cent, 
cobalt is shown in Fig. 3-18 (a). The increased efficiency of 
modern alloys is taken advantage of by designs using block or 
arch-shaped magnets in Alnico, Alcomax or Ticonal fitted to 
soft-iron pole pieces as shown in Figs. 3-18 (b) and 3-18 (c). The 
provision of the many small tapped and drilled holes often 
required for mounting the movement is simplified with these 
forms of assembly since the pole pieces are soft iron. 

Methods of fastening such block magnets are by clamping 
through a central hole or, preferably, a slot, by soldering or 
brazing, or when unit construction of the pole piece assembly is 
use. by magnetic attraction only. 

The high coercive force and working flux densities of the aniso- 
tropic alloys make fchc form of instrument magnet shown in 
Fig. 3-18 (d) possible. All the magnet material is contained 
within the cylindrical centre pole, which is magnetized along a 
diameter. With normal air-gap lengths flux densities with this 
type of design seldom exceed 0*2 Wb/m. 2 (2,000 gauss). 

Generator, Motor and Magneto Magnets 

Although many efficient machines are still made using 15 and 
35 per cent, cobalt steels, the trend with this type of machine is 









PERMANENT MAGNET STEELS AND ALLOYS 



133 



towards the use of the higher-performance alloys. The higher 
energies available enable ever-increasing performance demands 
to be met, and because of the reduction in magnet size the whole 
machine can be scaled down. Another great advantage found 
with the newer alloys is their greatly increased stability at high 
temperatures. It is not uncommon for such machines to work at 
temperatures approaching 100° C. At such temperatures the 
gradual dispersion of carbides in the cobalt steels, or ageing, as it 
is called, is accelerated. If this is not allowed for in design the 
performance of the machine may suffer. All the aluminium- 



/•'"/. .'{ IS. Mm l\c i nil. IN- 
sthi Mt;vr ,m.m;\i:t systkms. 

(a) MagQfil si eel up to 35 por 
iTiii . cobalt . 

(b) and (c) Block or mch- 
sluiped magnets fitted to soft- 

iron pole-piecee. 

(d) Cylindrical centre pole 
magnotized along a diameter. 






nickel-cobalt steels are completely stable at temperatures 
considerably in excess of this. 14 

Various designs for magnet assemblies for such machines are 
shown in Figs. 3-20 to 3-24. A conventional design is illustrated 
in Fig. 3-20 (a). In this the magnets are stationary and are often 
simple rectilinear blocks. No difficulties of fixing arise since it is 
usual to fasten the laminated pole pieces into the machine casing 
and fix the magnet by simple clamping. Typical rotor magnet 
assemblies are shown in Figs. 3-20 (6) and 3-21 : in Fig. 3-20 (b) 
a die-casting is formed around the magnets and pole pieces and 
sometimes also the machine spindle. Fig. 3-21 shows a method 
of producing a multiple rotor with a two-pole magnet. 



134 PERMANENT MAC NET BTEEL8 AND ALLOYS 



PERMANENT MAG-NET STEELS AND ALLOYS 



1S4 




Fig. 8-10. — : Varum stvckhok BUBCfiUcrn mi:tkk hiukk macxktk and moving- 

•Oll. INSIKI MKNT MU.NETS. 

The instrumeni magnet in the centra or the front row is an example of brazed 
construction and consists of an arch-shaped AWrrw mugnot blook with uni.i 

stool pole pieces. 

Magnetized axially, the magnet supplies flux to soft iron or 
laminated fingers which project and are interlaced, so producing 
the desired alternate polarity around the rotor periphery. This 
method of assembly is particularly useful when it is required to 
use an anisotropic alloy, since it simplifies the heat treatment 
magnetization. 




Fig. .'{ 20, Vakum's MAGKSTO 

AMI i;|.;NKKAT(ii: U AliNKT ANTJ 

POSH i-ii:i B a i: ha nckiif.nts, 

(a) Convent ional design. 

(6) Rotor magnet assembly. 
(-) and ((/) Flywheel -type 

rotors. 







Fig. 8-21. — A b.t.h. 

MACXICTO. 

The eight poles of the 
rotor of this magneto axe 
obtained by using " Bn« 
gored " ]>ole pieces in 
conjunction with a two- 
pole cylindrical m»i»iu't 
which is magnetized 

axially. 



A rotor magnet can sometimes be cast directly onto its shaft or 
onto a soft Iron sleeve which, during the casting process, is incor- 
porated in the sand mould, the shaft being suitably keyed to 
prevent rotation and axial movement of the magnet. 



Final 



Fig. 3-22. — Examples 
of good modkkn 
ma<;m:tu anu cknk- 

KATOK HOTOll l)K- 

Mi.\S. 

Simple magnet ahapaa la an 
isotropic or anisotropic alloy 
an- nii.ii with laminated pole 
pieces and an Becnrely assem- 
bled together wit* iplndle or 
boshing in ;i pmonre <lif 
casting. Pinal machining Is 
carried ooi after the casting 
operation. 
[induttrial Magnttc Co. Ltd.) 




136 PERMANENT MAGNET STEELS AND ALLOYS 



PERMANENT MAGNET STEELS AND ALLOYS 



137 




F>!/. 3 20. A si:i.i;< ]'iox OF MAiiXRTO. CMUMJUUXOB, AND MOTOR HAONBT8. 

machining of shaft and magnet must be carried out after heat 
treatment. The method is generally only suitable for two-pole 
rotors in anisotropic alloys ; where a greater number of poles is 
required, one of the isotropic alloys is usually used. 

The flywheel-type of rotor is used on many magnetos, and 
magneto alternators have usually six or more poles — typical 
forms of construction arc shown in Figs. 3-20 (c) and 3-20 (d). 
In Fig. 3-20 (c) segmental magnets and lamination blocks are 
pressed or moulded into a suitable wheel. Only half the number 
of magnets is used in the type shown in Fig. 3-20 (d), in which 



RING MAGNET 



CENTRE-POLE MAGNET 






% m'liii fciig.fci.riii,;,^ ;; 



T 



mm- ■=. 



(c) 



Fig. 3-24. — Types of loudspeaker units. 






the manufacture is simplified by using rectilinear blocks. Magnets 
and laminations arc moulded into a die-casting. 

Magnets for Radio 

The improved performances achieved by the use of anisotropic 
alloys and the reduction in size and scaling down of other parts 
which this has made possible have led to their almost universal 
adoption for magnets for loudspeakers, microphones and for the 
focusing of television tubes. 

All modern loudspeakers are of the moving-coil type, which 
have a diaphragm to which is attached a light coil arranged to 
move freely in the annular air-gap of the magnet, usually a 
permanent one. Magnets are generally of one of the types shown 
in section in Fig. 3-24. 

The first of these uses a ring of magnet alloy magnetized axial ly. 
G lamped between mild-steel pole pieces as shown. The efficiency 
reckoned as the ratio of flux in the gap to total magnet flux is 
about 40 per cent., falling as the gap field exceeds 1 YVb/m. 2 
(10,000 gauss). 

The slug or central-pole type is shown in Fig. 3-24 (b) and 
comprises a magnet block surmounted by a cylindrical soft-iron 




Fig, 8-25.— Various maijnkts FOB kadio. C'omim.ktk assk.mhi.iks aki-; siidws 

IN T1IK KOKECKOU.NI). 



138 PERMANENT MAGNET STEELS AND ALLOYS 




Loud STOAJOms rsivo 

itixci-i vi'j: magnets. 



Fig. 3-26. — Thesk illus- 
trations SHOW LOUD- 

SPKAKERS USING KINO- 
TYPE MAGNETS. 

The two loudspeakers 
shown on tho right are of 
the duplex type. They 
have a conventional voice 
coil and diaphragm at the 
front of the magnot urn I a 
secondary coil for high- 
frequency response at the 
rear. This coil loads the 
central cone through a 
hole m the centre pole of 
the magnot assemhly. 
{Whitelry Electrical Badto 

Co. 1.1,1.) 




PERMANENT MAGNET STEELS AND ALLOYS 139 



SOFT-IRON 
POLE PIECES 



Fig. 3-27. — Typical kihbon- 

MICIIOPHONE ASSIOir.l.Y. 




MAGNET 



tip, the whole being fixed into a cup or yoke with a front plate to 
form the outer gap face and return circuit for the flux : the 
maximum efficiency is around 55 per cent. A modification of this 
type, Fig. 3-24 (c), uses a centre pole made completely of magnetic 
alloy. Efficiencies as high as 65 per cent, are obtained within a 
very limited range of gap flux density, using Alcomax or Ticonal, 
which is directionally magnetized so that the flux lines, normal to 




RING 
MAGNET 



MILD STEEL 
POLE PIECES 



Fig. 3-28. — This illustra- 
tion SHOWS THE GKNKKAI. 
FOCrsiNC MAIiNKT ASSKM- 
BLY. 



K J 



140 



PERMANENT MAGNET STEELS AND ALLOYS 




Fly. 3-29. — Magnets used in radah EQUIPMENT. 

the axis \ hroughout the greater part of the centre pole, are turned 
through 1)0° in the vicinity of the air-gap. These last two types 
are widely used in television receivers, for which they are parti- 
cularly suitable, owing to their almost complete lack of external 
leakage field due to the shielding effect of the soft-iron cup or 
yoke. 

Various assembly methods are used ; ring types arc invariably 
clamped with screws as shown ; centre poles may be screwed, 
soldered, brazed or held solely by magnetic attraction. In this 
latter case a positioning device for centralizing the centre pole is 
necessary. Microphone magnets of the moving-coil type utilize 
a magnet essentially similar to that of a loudspeaker. Ribbon 
types use magnets similar to those shown in Fig. 3-27. 

Focusing magnets for cathode-ray tubes use magnets basically 
of the form shown in Fig. 3-28. Flux densities at the centre of 
the annulus of the order of 0-02-0-03 Wb/m. 2 (200-300 gauss) 
arc usual. .Magnets of short length are required because of space 
considerations, and since they usually work on almost open 
magnetic circuits, and consequently have a low permeance;, the 
alloys having very high coercivity such as Alcomax III, IV or 
Hycomax are most suitable. 






PERMANENT MAGNET STEELS AND ALLOYS 



HI 







Fiff. 3 3l». A M.\<:\KT|i . III! K. 

iii,. work plate is cut away to show the magnet assembly. Blocks ol Alcomas an east Into » 
movable oast-steel yoke arranged i" slide under the work plates by means of a lever, n Is shown 
in the " "ii " position. Movement ><> the left causes Bhort-circniting of the magnets and reduction of 

the Held riiiiiii.ii iii- ir. .in I In- -iiri'rt.r of the work plulc 







Fi<j. 3 31. — A MAGNETIC SHEET BXOATKR. 

An Interesting application of permanent magnets. Suitably boosed " " magnets magnetize steel 

.0 (, placed between them. The induoed polarities are such thai like poles are opposed, and the 

Bheeta repel one another, the top shed Boating quite dear of the others leading to easj handling, 

[Jama \eill A Co. (S*a#«W) Ltd.) 



142 



PERMANENT MAGNET STEELS AND ALLOYS 



Radar Magnets 

The very high performances required of magnets for use with 
magnetron oscillators are achieved only by accurate design. This 
is outside the scope of this book, and Fig. 3-29 is included merely 
to show the types of magnet in use. The methods of construction 
are obvious. In the type having two horn-shaped magnet blocks 
with a soft iron base, attachment is by means of screws passing 
through this base into holes tapped into soft-iron cores which are 
moulded into the magnets during casting ; such inserts are 
grooved and knurled to prevent movement, and the method is 
suitable only where the mass of the magnet is great compared 
with the insert. 

Only magnet materials of the highest internal energy are 
capable of producing the performances required within the 
permissible limits of weight. 

Fig. 3-30 shows a magnetic chuck, with the work plate cut 
away to reveal the magnet assembly. An interesting application 
of permanent magnetization is shown in Fig. 3-31. Steel sheets 
are magnetized by being placed between " U " magnets, the 
induced polarities being such that like poles are opposed. The 
magnetized sheets will thus repel one another and facilitate easier 
handling. 






PERMANENT MAGNET STEELS AND ALLOYS 

References 



143 



1. Honda, K., and Saito, S. " KS Magnetic Steels," Phys. Rev., 1920, 

16, 494. 

2. Misiiima. T. "Nickel Aluminium Steel for Permanent Magnets,' 

Ohm, 1932, 19. 

3. HOBSBUBOH, G. D. L., and Tetley, F. W. British Patent Spec. Nos. 

£31660 and 439543. 

4. Oliver, D. A., and Shedden, J. W. " Cooling of Permanent Magnet 

Alloys in a Constant Magnetic Field," Nature, 1938, 142, 209. 

5. Philips, N. V., Gloe^ampexfabrieken. British Patent Spec. 

No. 522731 (1938). 

6. Edwards, A. British Patent Spec. No. 577135 (1940). 

7. Permanent Magnet Association. " An Improved Permanent 

Magnet Material," Jour, of Scientific Instruments, 1945, 22, 56. 
B. Tyrell, A.J. " The Design and Application of Modern Permanent 

Magnets," Jow. B.I.R.E., 1946, Sept. 
9. Teti.ey. P. W. British Patent. Spec. 583411 (1946). 

10. HAPKiKi.n, I). British Patent Spec. 634686 (1950). 

11. Hadfield, D. British Patent Spec. 634700 (1950). 

12. Garvin, S. J. "The Production of Sintered Permanent Magnets,'' 

Iron & Steel Inst. Special Report, No. 38 (1947). 

13. Edwards, A., and Hoselitz, K. "Permanent Magnet Design," 

Elect. Rev., Aug. 4th, 1944. 

14. Permanent Magnet Association. Brochure 1952. 

15. Desmond, D. J. "The Economic Utilization of Modern Permanent 

Magnets," Jour. I.E.E., 1945, 92, Part II, 229. 

16. British Standards Institution. Spec. 406. 1931. " Apparatus for 

Workshop Testing of Permanent Magnets." 



Acknowledgments 

The author wishes to record his appreciation of the facilities 
afforded by the directors of Marrison and Catherall Ltd.. for taking 
the various magnet and equipment photographs and for permission 
to publish much of the technical data given in the various tables. 
The standard demagnetization and recoil characteristics are 
reproduced by courtesy of the Permanent Magnet Association. 
Thanks are also due to the various firms which have given per- 
mission for specialized magnet designs to be included in the 
various figures and illustrations, in particular : Ferranti Ltd., 
and Officine Galileo (Florence) for energy meter magnet designs ; 
Industrial Magneto Co. Ltd. and the British Thomson Houston Co. 
Ltd., for magneto magnet designs ; Whiteley Electrical Radio Co. 
Ltd., for loudspeaker illustrations ; and James Neill & Co. 
(Sheffield) Ltd., for the photographs of a magnetic chuck and sheet 
floater. 






3. (b) MICROPOWDER MAGNETS 

Permanent magnets made from soft ferro- magnetic metal 
powders are now being manufactured in this county after some 
years of development work. They are moulded by high-pressure 
methods from extremely fine powders of iron or iron-alloys, 
known as micropowders, which derive their permanent -magnet 
properties solely from the fineness of their particle or crystal size. 

Coercive Force and Particle Size 

From the domain theory, Neel during the last war deduced 
that very fine particles of iron of the order of magnitude of a 
magnetic domain would exhibit very high coercive force exceeding 
any value hitherto obtained in conventional magnets. The 
required optimum particle size was about one-hundredth that of 
the finest powder hitherto made for radio cores, but the prepara- 
tion, even on a small scale, was difficult and hazardous because 
such powder has an extremely strong affinity for ox}-gen and 
spontaneously oxidizes when exposed to the air. The magnetic 
quality of this highly pyrophoric powder could be maintained 
only by immersing the micropowder in inflammable liquids such 
as benzine which tended to increase the hazard. 

It was found experimentally that coercive forces approaching 
80,000 A/m (1,000 oersteds) coidd be obtained from pure iron 
reduced in dry hydrogen from ferrous formate and the best 
results were obtained with a particle of crystal size of 0-01 to 
(>•! micron (1 micron = 0-00] mm or 10,000 Angstrom units). 
Below this critical size range, the coercive force is rapidly reduced 
to figures inadequate for permanent magnets. Theoretically. I he 
maximum coercive force of pure iron approaches 800.000 A/m 
(10,000 oersteds) but this assumes that the particles are elongated 
and of optimum dimensions oriented in one required direction. 

These figures with the appropriate particle-size range are shown 
diagrammatically in Fig. 3-32, which indicates approximate 
values of coercive force plotted on a logarithmic scale. This 
diagram shows how the same raw material, pure iron in powder 
form, can give two grades of magnetic material with opposite 
performance. The coarse range above 1 micron is of practical 

144 






PERMANENT 
MAGNET 
RANGE 



M I C K OPOWDKR M A C! N E T S 
10,000 



1,000 



145 



100 



LOW LOSS 
RANGE 




OH - 



PARTICLE OR CRYSTAL SIZE IN MICRONS 



Imagnet 1 

l MICRO- 1 
POWDER 



! RADIO LOW 
I CORE | FREQY j 
POWDERI CORE | 
POWDER 



Fig. 3 32.- Relation BBTWKBH OOKRCIVE mi:, i-: AMD PABTIcm size of v iu ■ 

IRON IN I'OWIT.K 1(H;.M. 

Between 0-01 and 0-1 microns, particle raise is suitable for pressing mtoper- 
manent magnets, while for coarser powders the coercive force falls to values 
associated with low-loss .•..res. 

value on account of its low loss and the micropowder range below 
0-1 micron gives useful permanent-magnet properties. 

The best performance is obtained by making magnets from an 
iron-alloy micropowder containing 30 per cent, cobalt but, on 
account of the high price of this metal, cheaper grades are made 
with lower percentages of cobalt or from pure iron prepared from 
less expensive raw materials. 

Commercially-available Micropowder Magnets 

Permanent magnets are made from micropowders by high- 
pressure moulding. When compressed, the physical density and 
the magnetic flux density, which are closely related, increase 
progressively at the expense of the coercive force. In this way a 
large variety of B II curves ran be obtained. 

The considerable difficulties associated with the low density and 
highly-pyrophoric nature of the micropowders have now been 
overcome on a production scale with new techniques. 

Conventional demagnetization curves for two qualities of 
Gecalloy micropowder magnet are given in Fig. 3—33, the code 
letters being H for high quality (alloy) and M for medium quality 
(iron). Each is available in two grades to meet different design 
conditions, the letters R and C representing respectively the high- 






146 



MICROPOWDER MAGNETS 



MAGNETIZING FORCE - A/m 
5 6,000 48,000 40,000 32.000 24.000 16.000 



8.000 



0-8 



0-6 




0-4 



500 400 300 200 100 

MAGNETIZING FORCE- OERSTEDS 

Fig, :i 38.- Bll cuKVKs (HXKnnm) ros oboaxloy hicbopowdbb maonbtb. 
Curve I BR, Sigh-quality (alloy), high remanence. Energy factor. 1-5. 
Curve 2— HC. High-quality (alloy), high ooercive. Energy factor, M. 
Curve 8- MR. Medium quality (iron), high remanence. Energy Factor, l-o. 
Curve 4 — MC. Medium quality (iron), high ooercive. Energy factor, <>•('>. 
Curve A — Beeri non-oriented oast mngnoi, for comparison. 
(Energy factors (BH ,„„_,) expressed in moga-gmiss-oorstod.) 

(Saffbrd SbetrteU instruments Ltd.) 

remanence and high-coercivity grades. A closed or partly-closed 
magnetic circuit requires a comparatively low coercive force to 
maintain the flux and thus a magnet with higher remanence can 
be used (Types HR or MR). An open magnetic circuit requires 
a higher coercive force which results in lower remanence (Types 
HC and MC) ; hence a larger section of magnet is necessary to 
carry the same flux as in the earlier case. 



MICROPOWDER MAGNETS 



147 



Comparison with Cast-alloy and Steel Magnets 

The high-quality micropowder magnets have an energy factor 
equivalent to that of the best non-oriented (isotropic) cast-alloy 
magnets. At the present stage of development, micropowder 
magnets are not available with the greater energy factors of the 
highly-directional (anisotropic) cast magnets. However, since 
micropowder magnets are nearly half the weight of cast magnets 
of the same size, the difference in performance is comparatively 
small and can be often neutralized by the improved designs 
which the new magnets make possible. 

The medium-quality micropowder magnets have rather better 
magnetic properties than the whole range of tungsten, chrome and 
cobalt steels, especially regarding coercive force which is the 
principal magnetic factor affecting stability. These micro- 
powder magnets are made from pure iron produced with materials 
obtained in this country and effect a saving in imported metals. 

Physical Properties 

Possibly the chief advantage of micropowder magnets consists 
of the much lower physical density in comparison with other 
types. For the high-remanence magnets the density is about 
5 g/cm 3 ., but for the high-coercive force types it is in the region 
of 4 g/cm 3 . The latter are usually made with a plastic binder to 
maintain mechanical strength and this increases the specific 
resistance to that of an insulator and opens up possibilities of 
use under A.C. conditions. High-remanence types are made 
without binders and are thus of low electrical resistivity. 

Due to the softness of micropowders, it is possible to mould 
them into magnets to accurate dimensions without final grinding. 
The production methods for micropowder magnets are limited 
throughout to low temperatures and so it is possible to include 
shafts, pole pieces, or inserts of various other metals into the 
mouldings. 

For small quantities and samples, micropowder magnets and 
magnet systems are made by shaping and grinding pressed 
compacts to the required design. 

By moulding complete magnetic systems with pole pieces and 
inserts and by using the special shapes that moulding makes 
possible, a great improvement in the design of electro -magnetic 
components and equipment can be achieved. 



4. PERMANENT MAGNET FERRITES 

By 

B. VV. St. Legeb Montague, B.Sc. 

The earliest -known substance to exhibit permanent magnetic 
properties was the lodestone or natural ore of magnetite (Fe.,0,). 
It can be regarded as the forerunner of the modem ferrite per- 
manent magnet. The extension of this class of magnetic materials 
has been very recent : it is only since 1945 that appreciable 
development has taken place. 

Early Work 

In 1926 (!. Aminoff observed that a mineral containing lead 
and iron oxides was strongly attracted by a magnet and suggested 
for it the name " magnctoplumbitc." V. Adelskold, in 1988, 
determined from X-ray measurements that the crystal structure 
of magnetoplumbite is hexagonal, and proposed the chemical 
formula Pb().()Fe 2 () :J : he also found that the compounds 
B;i().6Fe 2 0;5 a,1, l SrO.CFegOa have a similar crystal structure. 
Other workers have found that Ba() .6Fe 2 3 contains a ferro- 
magnetic phase. 

Early Japanese work on cobalt ferrites as permanent magnets 
was described in 1933 and 1940. A recent British patent describes 
a method of manufacture of cobalt ferrite magnets yielding a 
better material. 

The most recent work on permanent magnet ferrites of the 
magnetoplumbite group has been described by Went, et id. (1952). 
A genera] survey is given by Brockinan (l!)~>2). 

Properties of Ferrite Permanent Magnets 

The magnetically " hard " ferrites are characterized chiefly by 
their high coercive force, somewhat low remanent induction, and 
an appreciably lower value of [BH) max compared with the more 
conventional metal magnets. They are magnetically very stable 
and difficult to demagnetize, either by external magnetic fields 
or by mechanical shock. They usually possess a negative tempera- 
ture-coefficient of remanence which may be of the order of 0-2 per 

148 



PERMANENT MAGNET FERRITES 



L49 



cent, per deg. C, some ten times greater than that of a metal alloy 
magnet, and although this is often a disadvantage, it can, on 
occasion, be put to practical use. 

The combination of large coercivity and relatively low rema- 
nence gives rise to demagnetization curves of the form shown in 
Y\g. 4-1 , which also gives comparative curves of a high-grade metal 
magnet. It can be seen that for the ferrite magnet (a) there is a 
large difference between the two points l: ll r and ,//,.. which are, 
respectively, the values of the field H at which the induction B 
and magnetization J are reduced to zero, whereas for the metal 
magnet (b) the two points are almost coincident. The practical 







A^jy^ 










V 










(b) 










y 

s 
* 
























J(*nJ)^ 


' {a L^- 


~~B 






OJ 


3-2 


O-i 








GQ 



Fig. -i I. I)km,\(.mi i- 

z.vnox CORVES im:h- 
nri BD i ROM < ii.MiiiNA- 
TION I IV l. \i;i;i: mi i:> i - 
MTV AND Ri:i.ATlVKl.V 
I.OW B KMAXKXl F. FOB 

.fkkisiti: m\<:m:t («) 
am" comparative 

i i i;vi:s or a men- 
tal aim: MXTAL M.UINCT 

(6). 



jH C II ' '!<■ B^C 

|i H.wi./m'(W"O 4 , oersted) jH C 

implieations of this will be discussed later. It may be noted here 
that the ferrite magnets are resistant to demagnetization by 
heating, provided the Curie temperature is not approached too 
closely. For the barium ferrite, the Curie point is at about 
450° C. 

The electrical resistivity of these materials is very large, 
frequently exceeding 10 6 ohm-m (10 8 ohm-cm). This can be of 
value in applications where a permanent magnet is required to 
polarize the core of an inductor or transformer carrying alter- 
nating fields, for eddy-eurrei it losses in a ferrite magnet are usually 
negligible. 



150 



PERMANENT MAGNET FERRITES 



The most recent developments in ferrite permanent magnets 
have been described by Went, Rathenau, Gorter and van 
Oosterhout (1952). They discuss in considerable detail the 
properties and crystal structure of ferrites in the magneto- 
plumbite group having the formula MO . 6Fe 2 () 3 , where M repre- 
sents one of the metals Ba, Pb or Sr. For commercial reasons the 
barium compound, which may be written BaFc 12 0i 9 , is generally 
preferred, and is marketed in Britain and the U.S.A. under the 
name " Magnadur." Since much more work appears to have been 
published on the structure of the magnetoplumbites than on that 
of the cobalt ferrites, we shall discuss the theory of permanent 
magnet ferrites on the basis of the former group, though there is 
reason to believe that the two classes of material have much in 
common. 

Composition and Manufacture 

The ferrites used for permanent magnets consist of mixed 
oxides of iron and one or more other metals, the heat treatment 
of the mixed oxides producing complex crystals with the required 
magnetic properties. Early work on these materials seems to 
have been devoted almost entirely to mixtures of iron and 
cobalt oxides, and although these cobalt ferrites are apparently 
still being investigated, there is little published information on 
their properties and even less on their structure. 

A cobalt ferrite material is marketed in the U.S.A. under the 
name of " Vectolite." It is prepared from a mixture of iron and 
cobalt oxides. This is pressed and sintered at about 1,000° C. 
and allowed to cool to 300° C, when a magnetic field is applied 
in the required direction and the subsequent cooling to room 
temperature takes place in the field. This magnetic treatment 
makes the material anisotropic, and increases the (BH) mnx 
obtainable to about 4,000 J/m 3 (0-5 m.g.o.). The remanence 
B r is 0-16 Wb/m 2 (1,600 gauss), and the coercive force ^ c is 
about 70,000 A/metre (900 oersteds). 

Cobalt ferrite magnets are described in British Patent Specifica- 
tions Nos. 594,474 and 596,875. The values claimed for coercive 
force and remanence are somewhat higher than the figures above, 
apparently due to the special methods of preparation. 

The method of manufacture of barium ferrites is in principle 
similar to that for cobalt ferrites, but many variations are possible. 






PERMANENT MAGNET FERRITES 



151 



From the point of view of mechanical properties, ferrite magnets 
vary considerably, according to their particular method of 
manufacture. Siiu-e they are made from finely divided oxides 
mixed with a binding agent, compressed and baked, the mecha- 
nical properties depend to a large extent upon the nature of the 
binder, the amount of compression and the subsequent heat 
treatment. In general, however, the ferrites are hard and brittle 
and are frequently classed as ceramics. They are often harder 
than glass and until recently could be machined only by grinding 
or cutting with a diamond wheel* 

Moulding imder pressure a very finely divided mixture leads 
to a product extremely compact and free from cavities and 
inclusions. Thus, ferrite magnets being very homogenous, can 
be used without polepieces and yokes where this would be an 
advantage. 

Theory of Permanent Magnet Ferrites 

In discussing the current theories underlying the behaviour of 
this class of permanent magnet we shall consider separately the 
coercive force and the remanence of the materials. When 
necessary we shall distinguish between the Z?-coercive force ^i c 
and the ./-coercive force jll c . 

There are three main factors which contribute to the coercive 
force of a magnetic material : (a) the magnetic anisotropy of the 
crystal ; (6) the stress anisotropy ; and (c) the shape anisotropy. 
In the case of the oxides M Fe, 2 19 , where M represents one of the 
elements Ba, Pb or Sr, the crystal anisotropy appears to play a 
dominant part in determining the coercive force. The stress 
anisotropy is also considered to be of some importance, but 
owing to the difficulty of estimating the internal stresses the 
contribution of* the stress anisotropy to the coercive force is not 
yet clear. 

Compared with the crystal anisotropy the shape anisotropy is 
small, due to the low saturation magnetization of these oxide 
materials. Tn any case, it is found that the behaviour of BaFe 12 19 
can be explained fairly satisfactorily by considering only the 
crystal anisotropy. It can be shown (Stoner and Wohlfarth, 1 948) 
that for a single crystal of BaFe 12 19 , which has one preferential 

* Noppiras ( I «53) has developed a new method of cutting and drilling such hurd 
materials by means of on ultra-sonically-driven tool. 



152 



PERMANENT MAGNET FKKRITKS 






PERMANENT MAGNET FERRITES 



153 



lb 

14 


S 






















200O0 
































12 

b 
A 

2 


-2 








15 000 Q 

Ul 






























oj 

I00OO ° 














\ 














\ 




bOOO 




































DO -lOO +I0O 


+200 


+300 


+ 400 


+ bOO 



7EMPERATURE,°C 

Wig, 4-2. — Tbm curve en 2X •/.,• aoainsx temi-kkaii re. 

direction of magnetization along the hexagonal axis, the field 
strength necessary to reverse the magnetization direction against 
the crystal anisotropy is approximately 2KfJ 9 , where K is the 
first anisotropy coefficient and ./, is the saturation magnetization. 
a function of temperature. The curve of 2KfJ a against tempera- 
ture is shown in Fig. 4-2. In a sintered aggregate of crystals with 
their preferential directions randomly oriented, the coercive force 
jll,, would be half this value, i.e. K/J K . 

The above result is based on the assumption that only the 
magneto-crystalline anisotropy energy is involved. However, 
for single crystals of approximately spherical shape there exists 
a critical diameter above which the effective coercive force is 
reduced. This was apparently first examined experimentally by 
Kocnigsbergcr (1947) in connection with the permanent magnetism 
of rocks. Xeel (1947) and Stoner and YVohli'arth (1948) show that 
above the critical diameter more than one domain is formed and 
that a reduction of coercive force results from the format ion of 
Bloch walls which modify the energy changes that occur on the 
application of an external field. When the crystal size is reduced 
towards the critical value so that the simpler model should apply 
it is indeed found that the coercive force increases rapidly. 
Kittel (1949) gives a useful survey of domain theory which is of 
interest in this connection. 

Fig. 4-3 shows the curve of coercive force n 0J H c as a function of 



S 



temperature for a very fine-grained sintered specimen of 
BaFe 12 19 . and it will be seen that it does not follow the same 
shape as the curve of Fig. 4-2, as would be expected from the 
simple theory : Went et ctl. attribute this to variation of Bloch 
wall mobility with temperature. 

The critical diameter for single domains in BaFe 12 19 crystals 
is of the order of one micron and is thus much more easily 
approached than the critical diameters for iron and cobalt, which 
are respectively about five and fifty times smaller. 

During the sintering process in the manufacture of the oxide 
materials care is taken to prevent the formation of large crystals 



Fig. 4-3. — Tins snows 
TlfF, COERCIVE IOBOB 
AS A FUNCTION OK 

tumi'i: u.vn BE Fon a 
vkky ii\i:-i:i: \i \ki> 

UUlXBfUfiO SPECIMEN 0> 
BaFo,.0 19 . 



































\ 


Q 
















o 

0-2^ 


y 














S3 














\ 


o 



-200 -I0O 



+IOO +200 +300 
TEMPER AT URE,°C 



+400 +S00 



so that a large coercive force is obtained. Under these conditions 
the high value obtained for the coercive force as compared with 
conventional magnet steels can be attributed partly to the large 
anisotropy of the hexagonal crystal structure and also to the 
closer approach that can be made to the critical particle size. 

The rcmanence of a magnet is of course highly dependent 
upon the saturation magnetization of the crystals. It can readily 
be shown by integration over a hemisphere, that for a homogenous 
isotropic magnetic material composed of domains with random 
orientation the remanent magnetization J r of the material is 
approximately one half the saturation magnetization J s of the 
crystals. The experimental results' given by Went et id. confirm 
this in the case of BaFe 12 19 as shown in Fig. 4-4. 



M.JL.F. 



154 



PERMANENT MAGNET FERRITES 



PERMANENT MAGNET FERRITES 



155 



Saturation Magnetization 

We shall now discuss the factors governing J s for oxides of the 
magnetoplumbite group, using BaFe 12 19 as an example. Although 
we are considering the hexagonal ferrites we can draw -to a 
large extent on the knowledge which has accumulated of the 
cubic ferrites which are magnetically " soft," for there is a close 
similarity between the two groups regarding the origin of the 
saturation magnetization. 

The cubic ferrites crystallize in the structure, known as 



































































\^s 














^5. 


X 


\ 












■^ 


A 



0-2 S 
S 



Fig. 4-4. — Curves SHOW- 
ING HOW THE KEAJA- 
NKNT MAGNKTl/.ATION 
OF MAGNETIC MAT Kit I A I. 

(BaFe ]t O tB ) is mvroxx- 

MATELY ONE HALF TBS 

SAT1 •UATIiiV MACNl.TI- 
ZATTON OP TnE < UYS- 
TAXS. 



-200 -IOO 



o +100 +200 +JOO +400 +SOO 

TEMPERATURE,°t 



" spinel," so called after the mineral MgAl 2 4 . In this arrange- 
ment, the oxygen ions form a closely packed cubic structure, the 
interstices between the oxygen ions providing two distinct kinds 
of site for the metal ions. These positions are known as tetra- 
hedral and octahedral sites, being surrounded by four and six 
oxygen ions respectively. Neel (1948) showed that there is a 
strong tendency for anti-parallel alignment of the free electron 
spins of ions in adjacent dissimilar sites. Since the distances 
between adjacent sites are too large for the normal quantum 
mechanical exchange interactions to be appreciable, Xeel attri- 
buted this phenomenon of " ferri-magnetism " to a super- 
exchange interaction, first proposed by Kramers (1934), in which 






the oxygen anions provide an essential link. Anderson (1 950) has 
extended the theory quantitatively. 

Without going too deeply into details, which may be found in 
the literature referred to in the bibliography, it may be said that 
tho existence of metal ions with different saturation magnetic 
moments in the two kinds of site gives rise to a nett magnetic 
moment for the whole crystal which is less than would be expected 
from simple addition of the individual ionic moments, because of 
the anti-parallel alignment of adjacent dissimilar sites. This 
results in a material with a lower saturation magnetization than 
is customary for metal magnets. 

Application of Ferrite Magnets 

Although ferrite permanent magnets have a value of (BH) max 
considerably less than that of modern anisotropic metal magnets, 
there are many applications where the peculiar properties of the 
former give them distinct advantages over the more conventional 
materials, particularly where the largest magnetic energies are 
not required. 

The chief advantage arises out of the large coercive force 
obtainable with ferrite magnets, and the consequent resistance to 
self-demagnetization. This makes it possible for the magnets to 
be magnetized free of their yokes or polepieces and fitted into the 
latter after the magnetization process, without any appreciable 
loss in flux. The reason for this may be deduced from Fig. 4-1. 
Any short magnet working into a large air gap will produce a 
strong reverse field tending to drive its working point down 
towaft Is tt H c . In the case of a normal metal magnet (6) (Fig. 4-1 ), 
the magnetization J is reduced considerably under these condi- 
tions since jflc and jH c are almost coincident. The reduction of 
J is an irreversible phenomenon and the subsequent closing of the 
air gap by a low reluctance yoke does little to restore J ; the 
operating* point merely moves up a recoil curve on the BjH 
characteristic. In the case of the ferrite magnet with its large 
coercive force however, J is hardly reduced even when the 
operating point moves right back to B H C , which is the limit to 
which the magnet can attempt to demagnetize itself. Conse- 
quently, the introduction of a low-reluctance yoke allows the 
induction B to return almost to the value it would have if the 
magnet had been magnetized in the yoke. In order to demag- 

12 



156 



PERMANENT MAGNET FERRTTES 



ET MAGNET 
,_Jr 



MAGNET 




Pig, 4-5. IjOtmSrKAKKR UNITS INCOItrOKATING TWO TYPES OF MAGNET, i.e. 

METAT. RING MAGNET, \XI> 1ERR1TE MAGNET ASSEMBLY. 

netize the specimen it is necessary to apply a very large reverse 
field. It ifl, in foot, impracticable to demagnetize ferrite perma- 
nent magnets in the conventional manner with an alternating 
Held ; the usual way is to heat the specimen to a temperature; 
above the Curie point. 

Operating Point 

Neglecting the effect of leakage flux, (BH) n ,„ x is obtained when 
the operating point of the magnet is at about half the ^-coercive 
force. This follows from the fact that the BII curve is almost a 
straight line for a ferrite magnet. Under these conditions the 
magnet operates at a low induction B and a large demagnetizing 
field as compared with a normal metal magnet. Thus for efficient 
operation ferrite magnets are much shorter and of larger cross- 
section than equivalent metal magnets. In fact they generally 

take the form of discs rather than of rods, and this necessitates 



an original approach to the design of units to incorporate them. 

An example is given in Fig. 4—5. which shows loudspeaker units 
incorporating the two types of magnet: (a) shows a typical 
arrangement for a normal metal magnet ring, and (6) the form of 
construction used for a ferrite magnet assembly. 

Fig. 4r-5 (/>) also illustrates a property of ferrite magnets related 
to leakage flux. It can be seen that the inside edge of the magnet 
is very close to the centre polepiece. With a metal magnet it is 
necessary to have a fairly large air gap between the side of the 
magnet and any adjacent magnetic material to avoid excessive 
leakage flux. Leakage flux due to adjacent polepieces is much 
lower with a ferrite magnet, for its incremental permeability is 
only fractionally greater than unity and the material thus has a 



PERMANKNT MAGNET FERRITES 



157 




Fig, 4-fi. — Typical pki!mam:mm.\i.\-ct poousihg ourr voa television 

BSH EiVKUs. t sin.: mI'I.i.akh KAOKADtKB Non -METAU.Tc BXNO magnets 
(SHOWN on biobx). 



high reluctance which effectively reduces the leakage under these 
conditions. Normal magnet steels have a permeability of the 
order of four or more when working at their optimum point. 

Ferrite permanent magnets have advantages over cast-metal 
magnets in assemblies where considerable vibration is encountered, 
e.g. in cycle dynamos, for they have greater resistance to demag- 
netization by mechanical shock. This is of particular importance 
in applications where the magnet works into a fairly large air gap. 

A most important advantage of the barium ferritcs is the low- 
cost of the raw materials. A ferrite magnet may replace a metal 
magnet in many applications with an appreciable saving in cost. 
Typical examples of this are magnetic oil filters, and children's 
toys incorporating magnets. 

The high electrical resistivity of the ferrite magnets and their 
consequent low eddy-current losses in alternating fields make 
them very suitable for providing a polarizing field in inductors 
and transformers. Examples are unidirectional pulse trans 
formers, polarized relays, and microphone and telephone units. 
A further advantage is the comparatively short length of magnet 
required, leading to a reduction in the effective air gap introduced 
into the transformer. 



158 



PERMANENT MAGNET FERRITES 



Because of their resistance to demagnetization by external 
fields, ferrite magnets can he used in opposing pairs to provide 
a field of strength variable with their relative position. This 
principle has been used in the focusing of television receivers, 
and could be applied also to the control of inductance in variable 
reactors. 



References 

Adelskold, V. Arkiv fiir Kemi, Minerdhgi och Qeologi (TJpsala) 12a, 

1938, 29, 1-9. 
Aminoff, G. Ueologiska Foreningens Fiirhandliugar, 192;"). 47, III. 283-9. 
Andkhson. I'. YV. Phji.siv.nl Review, 1950, 79, 705. 
Bozorth, R. M. " Ferromagnetism," 1951. 
Brockman, F. G. Electrical Engineering, 1952. 7, 64. 
Chitty. M. W. G. British Patent Spec. No. 594474, 1945. 
Kato, Y., and Takei, T. J. Inst. Elec. Engrs. (Japan), 1933, 53, 408. 
Ktttel, C. Rev. Modern Physics, 1949, 21, 541. 
Koemc;sijkik:ek, J. G. Phil. Mat]., 1947, 38, 640. 
Kramers, H. A. Physica, 1934, 1, 182. 
Neel, L. Ann. Physique, 1948, 3, 137 ; C.R. Anal. Sri. Paris, 1947. 224, 

1488, 1550 ; 1947, 225, 109. 
Neppiras, E. A. ./. Sci. Inst., 1953, 30, 72. 
Soctete d'Electro Chimie d'Uoink. British Patent Spec. No. 598875, 

1943/45. 
Stoner, E. C, and Wohlfarth, E. P. Phil. Trans. Roy. Soc., 1948, 

240A, 599. 
Takei, T., Yasuda, T., and Ishihara, S. Blectrotech. J . (Japan). 1940, 

4, 75. 
Went, J. J., Rathenau, G. W.. Gorter. E. W., and vax Oo.steritot7t, 

(i. W. Philip* Tech. Rev., 1952, 13, 194. 



5. MAGNETIC POWDER CORES 
By 

C. Gordon Smith, M.A., A.M.I.E.E. 

The use of magnetic materials in the form of " dust " or 
compressed powder cores * has been established for some 30 years 
and has been until recently confined to purposes to meet the 
specialized requirements of telecommunication. Such uses, to be 
described below, generally involve the magnetically soft materials! 
developed for use at low inductions, although, as will bo shown, 
the effective properties of these materials become modified as a 
result of the form and condition in which they are used. 

Historical Development 

Reference in 1887 to the use of iron filings embedded in wax 
was made by Heaviside, 1 who found that. the inductance of a coil 
could be increased by such means without causing any appreciable 
dissipation of energy. The magnetic properties of iron powder 
in various forms were investigated in the early twentieth century, 
but little successful application developed until the First World 
War. It was then that the need arose for alternative and improved 
substitutes for the cores of telephone loading coils, which hitherto 
had been made from bundles of steel wire. The development of 
a new production technique for compressed magnetic powder 
cores was successfully completed by the Western Electric Com- 
pany 2 of America, the main features of the process being the 
production of a suitable (electrolytic) iron powder and the use of 
high compacting pressures in conjunction with special insulating 
binders. Improved cores of this type are now used as essential 
components in telephone and radio equipment. Important stages 
of progress in the art were the discovery of the specially suitable 
properties of carbonyl-iron powder in Germany 3 and later, in the 
U.S.A., the application of the nickel-iron alloys. 4 Further 
improvements followed to satisfy the technical requirements of 

* Tho uso of the term " dust " core, although generally accepted, is somewhat 
to be deprecated, particularly owing to the recent application of ultra-fine mag- 
netic particles in the production of permanent magnets. (See Section 3 (6).) 

t The recent applications of ferrites in these fields is described in Section 2 (6). 

150 



160 



MAGNETIC POWDER CORES 



MAGNETIC POWDER CORES 



161 



the widening applications in the development of carrier telephony 
at increasingly higher frequencies. The successful application of 
powder cores to broadcast radio receivers by Vogt 5 in 1932 
was achieved through important new concepts in core design and 
manufacture. 

Elementary Theory of Magnetic Powder Cores 

The principal aim is usually the production of an inductor of 
low power-factor ; the introduction of a ferromagnetic- core assists 
by reducing the resistance of the requisite winding in proportion 
to its magnetic permeability, but this advantage will be partially 
offset by losses occurring in the magnetic material. The two 
factors must thus be carefully balanced to achieve optimum 
results. 

(a) Permeability. Tt will be assumed that the working condi- 
tions involve alternating magnetic fields of low density in the 
region in which the incremental permeability is substantially 
constant. Such conditions usually apply in most practical 
applications, the flux density being in the region of 1 mWb/m 2 
( 1 gauss) or less. 

It is first neeessarj*- to consider the relationship between the 
effective permeability fx c of the core material, and the intrinsic 
permeability ^u,. of the ferromagnetic component. A further 
quantity referred to loosely as effective permeability, /x c , relates 
only to a particular coil assembly, and is defined by the ratio 
L/L , where L is the inductance of the coil plus core and L is 
the inductance of the same coil but with core removed. 

In the more familiar case of a core built up of high permeability 
laminations, fx c ~ /x,- ^ fi e , as the differences are due to flux leakage, 
which is relatively small. In the powder core, however, \i c </u,. 
on account of gaps in the magnetic path. Furthermore, owing to 
the relatively low values of /x r , leakage flux may be appreciable, 
with the result thai /e may be considerably less than fx c * 

The calculation of the relationship between /x,. and /z c has been 
the subject of much study 8 based on various assumptions, e.g. 
that the particles are uniform spheres, cubes, laminae, etc., but 
none of these approach closely the effects obtained in practice 
on account of the wide variation of particle size and shape obtain - 

* For example in (she type of core illustrated in Fig. 5-2 (13), /t, might be 500 
with !•,. 20 and ^,. about 4-0. 






ing in all metal powders. A simple analysis without considering 
such details will, however, be informative. Suppose the magnetic 
circuit to be of unit length and divided longitudinally into a 
portion g of unit permeability (i.e. air-gap), and a portion (1 — g) 
of permeability ti,., then 

fa 



M c = 



</(/*, - i) + i 



Curves of this function are shown in Fig. 5-1, covering a 
practical range of variation in \x c and g. 

Except for low values of the intrinsic permeability u { , which 
are not likely to be of practical interest, the effective permeability 
is mainly inlluenced by the value of g rather than by it t -. This 
explains an important subsidiary property of dust cores, namely, 
constancy of permeability. The usual causes of permeability 
variation, e.g. flux density, temperature, magnetic shock, etc., 
are reduced to small proportions owing to the diluting effect of 
the magnetic gap. The same fact indicates what considerations 
should be made when selecting the most appropriate magnetic 
powders and methods of core manufacture, where effective 
permeability is the prime consideration. The problem is usually 
t he reduction of the gap effect ; this is achieved by using high 
compacting pressures, a minimum of insulating binder, and a 
magnetic powder which is easily compressible, i.e. whose particles 
are soil enough to deform, and so prevent formation of voids in the 
core, and whose degree of subdivision is no greater than necessary 
to restrict eddy-current losses adequately. A high intrinsic 
permeability of the metal is usually of secondary importance. 

(b) Loss Factors. In the majority of applications the power 
factor of the inductor is of major importance and consideration 
must be given to the losses introduced by the ferromagnetic 
powder. To a fair approximation, at the low flux densities for 
which the incremental permeability is constant, for frequencies 
I .clou a lew megacycles per second, and excluding materials with 
unusually large or small losses, the effective series loss-resistance 
R c in ohms due to the core may be expressed as * 

R c = ftrflft (-^ + o + tf) 

* A similar analysis is given by K.rsti'n. 10 Both analyses assume a magnetic 
circuit without leakage (e.g. a toroidal core). 



162 



MAGNETIC POWDER CORES 






MAGNETIC POWDER CORES 



163 







1000 



4000 



2000 3000 

PERMEABILITY OF METAL {1J 

FiiJ. 5-1. RELATIONSHIP BETWEEN INTRINSIC AND KII K( II V K FKKM KA HI 1.ITIES 

FOR DIFFERENT tSAI* EFFECTS. 

where / = frequency, c/s. 

L — effective inductance, H. 
a = hysteresis loss constant* 
c = residual loss constant 
e = eddy-current loss constant 
B = r.m.s. flux density. Wb/m 2 . 

The Q-factor, or " goodness factor," Q of a coil is given by 
Q - 2-nfLj{R e + R ), 
where R is the winding resistance, which for a given inductance 
is inversely proportional to /z c . 

* The exponent x in the Steinmetz formula, loss — -qB 1 has a value of 2-0. 



relating to the 
core material. 






The manner in which both hysteresis and residual losses affect 
the Q-factor is independent of frequency, and the residual loss 
and eddy-current contributions are independent of the si/.e of the 
core. The hysteresis effect is, however, related to the core size 
in that the flux density B is dependent on the core dimensions for 
a given magnetizing current in the inductance. The eddy- 
current loss increases with frequency and tends to preponderate 
as the frequency is raised. The d.c. winding resistance will also 
be inversely proportional to the size of the core. Other factors 
entering into the design of high-Q coils are the further /^losses in 
the winding due to skin and proximity effects, and dielectric losses; 
both may become appreciable even at relatively low frequencies 
(e.g. 10 kc/s). For fuller analyses of loss distribution and general 
principles of design the reader is referred to publications by Legg 
and Given 9 and Kersten, 10 and to the textbook by Welsby. 11 

Applications and Uses of Magnetic Powder Cores 

The original commercial application of magnetic powder cores 
was for telephone loading coils, i.e. inductors inserted at regular 
intervals in telephone cables to decrease signal attenuation. The 
technical requirements for these, which must be met in a limited 
space for economic reasons, are constant inductance and high 
Q-factor. The introduction of high-frequency carrier telephony 
has resulted in a tendency to abandon line loading, but there 
exists a further demand for similar high-grade inductors in 
terminal filtering equipment. A secondary consideration is the 
requirement for low hysteresis loss, which, apart from its eaergj 
dissipation, may produce non-linear distortion and cross-modula- 
tion between carrier circuits. 12 

For all these applications it is usual to use toroidal cores, 
although such a design is not efficient from the point of view of 
d.c. resistance. With any other form of core, even if com- 
pletely shrouded, considerable magnetic leakage will take place 
on account of the relatively low permeability of the compressed 
powder ; this leads to considerable inefficiency in the magnetic 
circuit and to stray fields which may cause coupling with 
neighbouring components. Again, although reasonably strong 
mechanically, powder-core materials do not lend themselves "to 
fabrication or machining ; the toroid is a form easily made by a 
pressing without joints in the magnetic circuit. With available 



164 



MAGNETIC POWDER CORES 



MAGNETIC POWDER CORES 



165 




/'"';/. ."> 2. M.ycnktm: powdkb cokks of vaktoi s tvci;s. 

a, Toroidal core of Mgh-penneabOttjr alloy powdet Ua small loadlng-coIL 

a. Wound oofl on A,. 

\ : , Completed COil in screening can ; Q-faetor of about Kin al i..".iih ,.■ '.-. 

n Luge " pot " core showing winding former for carrier-freqncncy application- \ Q > 300 from 
;>n la - to - Ifc/a. 

i' Small •• (ml " core for carrier mill radio applications : Q > BOO from 80 kc/s to 3 Hb/s, 

I) .Small screw conn for radio applications : effective j»< riin:il>ili1 v about 2 : Q > 150 from 500 EEC - 
to20Mc,'s; > 100 up to 80 Mo/S, 

E Peniieability-tuninir radio con- ; effective permeability about 10. 

F Television line-frcqiieiiey transformer, wound on core of Mgh-permenbility iron-powder pressings. 

(Ill/ courtesy of Siilforil l-llectru-nl buttVOmU Ltd.) 






materials it is easily possible to design toroidal inductors \vi1h 
Q-factors of 200-300 (i.e. power factors of less than £ per cent.) at 
all frequencies above a few hundred c/s up to several hundred kc/s. 

Powder cores in radio receivers can supply the need for high-Q 
inductors for tuning circuits, but their use is based more on 
economic considerations in providing a given selectivity at less 
expense or in a smaller space. 

For broadcast frequencies and above, small cylindrical cores of 
a variety of shapes and sizes are used, giving similar Q-factors, 
reduction in size being possible because of the preponderance of 
eddy-current losses at high frequencies, for which the effective 
resistance is independent of core dimensions. 

For the lower radio (i.e. broadcast receiver intermediate) 
frequencies or for high-frequency carrier telephony, a compromise 
is often made in the form of cores completely or partially shrouding 
a winding (see Fig. 5-2 (G) ). 

Besides these main functions of powder cores, subsidiary but 
important uses have developed mainly in the r.f. fields. These 
depend on the possibilities of inductance variation by the relative 
motion of core and coil. There are inherent limitations on the 
range of possible adjustment ; because the toroidal form is 
impossible, only a fraction of the effective permeability of the core 
materials can be utilized. Where the maximum range is desired 
it is necessary to use a cylindrical core with a large length/diameter 
ratio, which restricts the Q-f actor. However, cores giving an 
inductance ratio of 10/1 have been successfully made for radio 
receiver tuning (see Fig. 5-2 (E) ). 

In other instances it is useful to have a means for the precise 
adjustment of a fixed inductor. This is often required, particularly 
at the higher radio frequencies where any capacitance additional 
to the self-capacitance of the circuit components is undesirable. 
It can effectively be provided by a movable core with moulded 
thread to engage in a suitable coil former. 

Another application in which powder cores may offer important 
advantages is in devices such as receiving loops for direct ion - 
finding equipment, where the introduction of a magnetic core into 
the loop will give increased sensitivity. 6 

High-Permeability Powder Cores 
Attention has recently been given to the development of 



166 



MAGNETIC POWDER CORES 



powder cores of specially high permeability, at the expense of 
increased losses. Such materials may be made from inexpensive 
but readily compressible and relatively coarse iron powder, 
sometimes in the form of flake. 13 They have uses at power 
frequencies as inexpensive replacements for silicon-iron (Stalloy) 
stampings in small transformers, reactors, etc., in which power 
loss is unimportant. They also bridge the gap between dust cores 
and normal laminations. In general their hysteresis losses are 
high compared with those of silicon-iron and similar laminations, 
but eddy-current losses are reduced. An important application 
in television receivers is to line-frequency transformers for 
providing the e.h.t. voltage. 14 A high permeability is required 
to obtain adequate coupling, and under these circumstances lower 
losses can be obtained than with silicon-iron stampings. Another 
use is for the moulded pole pieces for electromagnetic deflectors 
for television cathode-ray tubes. 
A typical selection of coils and cores is shown in Fig. 5-2. 

Magnetic Powder Core Materials 

From the foregoing it will be evident that the properties 
required of a magnetic powder are that the particles shall be 
easily compressible to form a dense core, and that the particle size 
be fine enough to reduce eddy-current losses effectively, but not 
too fine, for otherwise difficulty will be found in producing an 
adequate core permeability. The intrinsic permeability is not of 
first importance ; the metal should, however, have low hysteresis 
loss and a high resistivity in order further to limit eddy losses. 
Constancy of properties with temperature and resistance to 
magnetic shock may be of importance, as, although these factors 
themselves are very largely looked after by the magnetic dilution 
effect, a very high order of excellence may be needed in precision 
applications such as loading coils and inductors for filters for 
carrier telephony. 

In practice the main materials used are special iron powders 
and nickel-iron and related ternary-alloy powders. 

Iron Powders 

(a) Carbonyl Iron Powder. Carbonyl iron powders were among 
the first to be used extensively for high-quality dust cores. Their 
special characteristics 10 include a spherical shape and uniformity 






MAGNETIC POWDER CORES 



167 



of particle size, combined with a peculiar metallurgical structure 
yielding eddy-current losses considerably lower than those of an 
iron-carbon alloy of similar composition. The characteristics are 
ideally suitable for use at radio frequencies. Hysteresis loss is 
also low, but although the spherical shape gives a good packing 
factor, the high mechanical hardness (an estimated diamond 
pyramid number of over 800) makes compression difficult and thus 
it is difficult to obtain high core permeability. This disadvantage 
can be overcome by heat treatment of the powder to remove 
carbon, but even then the small particle size limits the per- 
meability obtainable. Eddy current and hysteresis losses are 
somewhat increased by the heat treatment. Nevertheless, the 
decarburized grades of carbonyl-iron powders are still used for a 
number of low-frequency applications, although in this country 
and the U.S.A. the use of nickel-iron powders 16 has become more 

general. 

(6) Electrolytic Iron Powder. This type of powder is made by 
the mechanical disintegration of a brittle electro-deposit. It has 
a relatively high degree of purity although particle size and shape 
are unfavourable from the point of view of powder-core considera- 
tions. Its use for powder cores has now become restricted to 
purposes for which relatively high losses can be tolerated. 

(c) Hydrogen-reduced Iron Potvder. The reduction of iron 
oxide by hydrogen at suitable temperatures provides an alterna- 
tive means of producing a pure iron powder. Although very fine 
particles may be obtained, their magnetic properties are generally 
inferior to those of carbonyl-iron powder, eddy-current losses 
being higher both inherently and on account of difficulties of 
insulating the characteristically irregular and porous particles. 
These powders are applied for second-grade high-frequency cores, 
particularly when effective permeability is more important than 
high Q-values, the softness of the powder favouring a high 
compressibility at the expense of eddy-current losses. 

{d) Flake-iron Powders. A recent development, 13 intended to 
bridge the gap between powder cores and laminated materials, has 
resulted from the use of iron in the form of thin flakes aligned so 
that their major planes fall parallel to the lines of flux, giving 
permeability dilution controlled by the larger dimensions of the 
flakes, but eddy-current losses controlled by the thickness. Such 
materials have in general rather higher losses than powder cores 



168 



MAGNETIC POWDER CORES 



or thin alloy laminations, but they are finding a number of applies 
tions in new fields as indicated above. 

(e) Miscellaneous. Other methods have been used for the 
production of iron powders but are of little current interest. 
The Hametag process involves the mechanical disintegration of 
chopped wire and produces a rather coarse flattened particle ; 
such material had some application in Germany for the production 
of iron powder cores with permeability rather higher than can 
be obtained from carbonyl powders. 

Other methods include the disintegration of the molten metal. 
Such processes are used for the production of an inexpensive 
powder for metallurgical purposes, but the products are of little 
use for cores. 

Nickel-Iron and Related Alloy Powders 

Although the main feature of these alloj^s. their exceptionally 
high initial permeability, is not of first importance, their excellent 
subsidiary magnetic properties of low hysteresis and residual 
losses have led to successful applications in compressed powder 
form, in the frequency range of line-telecommunication equip- 
ment. 

These alloys are tough and ductile and their preparation in 
powder form have presented some special problems. Two main 
methods have been satisfactorily evolved. The first 4 relies on 
the artificial embrittlement by intergranular penetration of a 
small addition of sulphur followed by mechanical disintegration 
and thermal treatment. The second 17 method involves the 
hydrogen reduction of the mixed oxides followed by a combina- 
tion of thermal and mechanical treatments to ensure thorough 
metallurgical homogeneity. Further problems involve the use of 
refractory insulating binders so that cores may be given a suitable 
annealing treatment after pressing, to restore the optimum 
magnetic properties. Molybdenum-Permalloy and nickcl-coppcr- 
iron alloy powders have both been satisfactorily applied and have 
made possible the production of high -permeability cores of out- 
standing quality for loading coils of small size and high efficiency. 
These form an essential component in telephone distribution net- 
works not only in trunk systems but also in heavily developed 
urban areas. 

A further advantage of the nickel-iron alloys (notably in the 



MAGNETIC POWDER CORES 
Table 1. Magnetic Powder Core Materials 



169 








Relative 










Materia] 


permea- 
bility 


Loss Pactors 


Applications 




ax 10* 


ex 10* 


sxlO" 




Compressed at high pressures, 












e.u. ion timsjin.* 












Electrolytic iron . 


35 


50 


100 


850 


Early loading coils (obsolete). 


Carbonyl-iron, E type . 


15 


8 


200 


2 


Low-loss coils lor filter and load- 
ing application at high fre- 
quencies where minimum 
hysteresis required. 


Carbonyl Iran) C type . 


50 


10 


20(1 


20 


Loading cods general use (mainly 
i ontinental). 


Permalloy, powder . \ 


125 


10 


30 


200 


loading coils witb special bigta 
rjcrmeabflity lor use In re- 
stricted Bpace. 


, , . . \ 


60 


2-5 


5o 


100 


Loading coils, general use. 


(.\ickel-iron-molybdenuni) / 


25 


70 


100 


80"| 


Low-Ion coils for Biter and load- 
ing application at high fre- 


\\iHi dlflereni proportions of 1 








f 


i|iicnck-s where minimum 


binder. / 


14 


10 


ISO 


70 J 


hysteresis required. 


Sendust (a In minimi i--ilin>n- 


00 


5 


200 


30 


Loading coils for general use 


-iron). 










(restricted), 


Fluke-iron and elect rolyt la 


150 250 


300 


— 


30-100 


(Kougn estimation.) Applications 


(Special Insolation techmqoe.) 










at power frequencies and where 
losses unimportant. 


Compressed tit hirer pressure, 












e.g. 20 toitSjin. 1 












Electrolytic iron. --325 me.-h. 


10 


250 


1,200 


200 


(Little used.) 


Carbuiiyl-irou, 10 type . 


12 


3 


200 


Q 


Radio frequencies : first-grade 


Carbonyl-iron, W type (special 


10 


2 


luo 


. J 


cures. 

Kadio frequencies (above 


line particle size). 










lOMc/s), 


ll\ ilriiu'eii-r.-iliiccil iron . 


20 


50 


750 


30 


Radio frequencies, where per- 
meability more important than 












losses, permeability toning, etc. 



range 35 to 50 per cent, nickel) for dust cores is their high resisti- 
vity, which favours low eddy-current losses. These alloys have 
found some application in the high-frequency field, but the 
difficulty of producing sufficiently fine powders has hindered 
competition with carbonyl-iron powders at radio frequencies. 

Iron-Aluminium-Silicon Alloys 

Certain iion-aluniinium-silicon 18 alloys, investigated originally 
in Japan, were found to have the characteristics of M soft " 
magnetic; materials although mechanically hard and brittle, 
facilitating reduction to powder. These alloys were further 
developed in Germany 7 during the Second World War and had 
some measure of success, principally on the grounds «f nickel 
economy. The magnetic properties do not, however, compete 
with the best nickel-iron and related alloys. 

Table 1 gives details of the magnetic properties of the principle 
core materials together with notes on their applications. 



170 



\l \UXETIC POWDBK C'uKICS 

References 



1. Hkavisidk. o. "Notes "ii the Theory of the Telephone and on 

Hysteresis,'' Electrician, 1887, 18, 302 3. 

2. St'KKn. B., and Ei.ukn. «i. W. " Magnetic Properties of Compressed 

Powder Dron, M Trans. A.I.E.E., 1921, 40, 596. 

3. British Patenl 269770 (1926). 

4. Sitacklkton, W. J., and Babbbb, \. (\. "Compressed Powdered 

Permalloy : Manufacture and Blagnetie Properties, Trans.AJJSJS., 
1 Hi's. 47,' 429. 
Lbgg, V. B., andGrvBH, II. J. "Compressed Powdered Molybdenum- 
Permalloy for High Quality [nduotanoe Coils."' lull System Tech. 
./,,!., 1940, 19, 385-406. 

5. Voct. H. "New Development in Tuning Coils; Inductances of 

Remarkably High Efficiency using ' Forrocari ' Cores,"' Wireless 
World, 1932; 31, 272 273. 

6. U.mcki.ky. K. P., el dl. '• Iron Cores, IXF. Loops and ..Manufaclinv of 

Iron Dust," BJ.O.S. Pinal Report No. 1203. H.M.S.O. 

7. HbnSHL, F. R. "Iron Cores," F.l.A.T. Pinal Report No. 792. 

H.M.S.O. 

8. Howe. G. W. 0., tl (A. " Permeability of Iron Dnsi Cores. "' Wireless 

Engineer, Nov. 1946. 23, and Feb. 11)47. 24. 

9. Lego, V. E. " Magnetic Measurements at Low Flux Densities," Bell 

System Tech. Jul., 1936. 15, 39-62. 

10. Kkwstkn. M. "Sptilen mil Massekernen" (Coils with Dust Cores), 

Elektrotechmsche Zeitsckrift, 1937, 58, 1335-38, 1364-67. 

11. WELSBY. V. (',. "Theory of Design of Inductance Coils,'" Macdonald 

& Co., London. 1950. 

12. I'f.tkusox. E. "Harmonic Production in Ferromagnetic Materials 

at Low Frequencies and Lou Flux Densities," Bell System Tech.. I nl., 
l«>2s, 7, 762. 
13 Campbell, GL, and Wood, P. J, "A Laminated Flake Iron! Material 

for use ai Audio and Ultrasonic Frequencies," "Soil Magnetic 
Materials in Telecommunications," ed., Richards, < . E., and Lynch, 

A. 0., pp. 26S-277. Pergamon Press, London. 1943. 

14. Fkikmi. A. W. '" Molded Iron Dust Cores For use in Horizontal 

Deflection Circuits," R.O.A. Review, March 1947, 8, 98-115. 

15. T'kkii., L. IS., POLORBBN, G. R., and Bicki.ky. S. K. Si/ni/xi.siain on 

Powder MetaUurgy, Iron db Steel Tnstitute Special Report No. 38, 

11)47. Seel ion C. 17 59. 

16. Khiiahds, ('. E. A.. BUOBXBY, B. E., 1 L\ iti )i:i.i.. P. H.. and Lynch. 

,\. ('. •Some Properties and Tesis of Magnetic Powders and Powder 

Cores." Symposium of Papers on Ferromagnetic Materials, Session 
IV. I'roc I.E.I-;., 97, Pari 11. No. 56, April 15150, 236-45. 

17. POLGBBBH, Gk E. "The Production and Application of Magnetic 

Powders," QJE.G. Journal, 19. July 1952. No. 3. pp. L52 69. 

Is. Masimoto. EL "Magnetic and Electrical Properties of a New Alloy, 
* Sendust,' " Sci. Repts., Tohoku Imp. I'niw, 1936, 388-402. 



6. NON-MAGNETIC FERROUS AND 
MAGNETIC-COMPENSATING ALLOYS 

By 

C. Cordon Smith, M.A... A.M.I.E.E. 

NON-MAGNETIC FERROUS ALLOYS 

In the construction of electrical equipment where materials 
with non-magnetic properties arc desirable or essential, and 
where the mechanical properties of non-ferrous alloys are inade- 
quate, use is made of Special non-magnetic steels or cast irons. 

Non-Magnetic Steels 

Although the effect of the addition of non-magnetic alloying 
elements to iron is to reduce the magnetic permeability, in general 
appreciable magnetic properties will be retained after the propor- 
tion of alloying element has been increased to such an extent 
that the characteristic mechanical properties are lost. However, 
the desired results may be obtained by making use of the special 
effect of the lowering of the Curie point, particularly remarkable 
in certain nickel-iron and related alloys. The singular charac- 
teristics of these alloys was first observed in 1880 by Hopkinson, 19 
who found a 26 per cent, nickel-iron alloy to be non-magnetic. 
Fig. (> 1. alter Mcrica, 20 gives a simplified version of the diagram 
of the nickel-iron system, and shows the rapid lowering of the 
Curie point in the region of 30 per cent, nickel. It should be 
noted, however, thai complications exist in the region below about 
35 per cent, nickel in that the alloys may exist in two states with 
irreversible or very slow transformation rates. This peculiarity 
was also observed by Hopkinson, who found that his 25 per cent, 
nickel-steel became strongly magnetic on cooling below room 
temperature and remained so until re-heated to a high tempera- 
t ure. The addition of further elements such as carbon, chromium, 
manganese, etc., may however be used to assist in the stabilizing 
of the alloy in the non-magnetic (austenitic) form. Manganese 
is a particularly useful addition as its presence tends to lower the 
(.'uric point still further. Hopkinson's 25 per cent. nieicel-iron can 
be made reasonably stable by the addition of 0-3 per cent, of 
carbon, and such an alloy has been considerably used. 



171 



II 8 



172 



NON-MAGNETIC FERROUS ALLOYS 



1500 




-100 



Fig. 6 1. Simplified diaguam ox the nickel-ikon system (mehk a). 



NON -MAGNETIC FERROUS ALLOYS 



173 



More recently increasing use has been made of a number of 
highly alloyed nickel-steels, including chromium or manganese, 
or both together, which give improved tensile properties in addi- 
tion to the non-magnetic characteristic. Included in this group 
are the well-known austenitic stainless steels, but unless the 
proportion of alloying elements is adequate, the effect of cold 
work, useful in improving the tensile properties, may reintroduce 
magnetic properties 2l by the formation of a certain proportion 
of ferrite particles distributed throughout the austenitic matrix. 
(Compare the use of this phenomenon in wires for magnetic 
recording : see p. 181.) 

The principal uses of non-magnetic steels are for parts of 
electrical machinery, such as retaining rings for alternator rotor 
caps and wedges, armature binding w ire and strip, and armouring 
for a.c. cables. In the majority of such applications alternating 
magnetic fields are involved, and further advantage is often found 
in the resistivity which limits eddy-current losses. 

Non-Magnetic Cast Irons 

In a similar manner, non-magnetic properties may be conferred 

on cast iron by the inclusion of a proportion of alloying elements ; 
the two main commercially established alloys are Nomag and 
Ni-Resist (see Table 1). These find application in the electrical 
industry where a less expensive material than a non-magnetic 
steel is needed and where conditions of use and production call 
for a casting. They have the additional advantage of higher 
resistivities than the steels and are frequently used for heavy 
starter resistance grids, covers for switchgear and other parts in 
which the effect of induced currents is to be limited. The recent 
development of the production of cast iron containing graphite in 
spheroidal form. 22 which gives most pronounced improvement in 
mechanical properties, has been satisfactorily applied to the 
production of non-magnetic; materials. (In this case, however, 
some reduction in electrical resistivity results from the compact 
graphite structure.) 

In all the above alloys the non -magnetic properties are due to 
the particular metallurgical structure occurring above the Curie 
point, so that on cooling to low temperatures the D&gnetic 
properties will reappear. There appears to be little information 
available concerning precise values of the Curie temperatures ; 



174 



NON-MAGNETIC FERROUS ALLOYS 



NON-MAGNETIC FERROUS ALLOYS 



IT.". 



iffl 

-z - E 
<"3 « 



.2 r- 

•o -- 
5 > 
K"5 



_ s 

1*1 

3 5 



a = 

B -3 Q 

.2 g - | 

1 .is 

i 



■a 



b ~ 



WW 



z — i 
C5 to 
•c © 

o X 



32 



8| 

I* 

Li 

3 — 
a 

a 5 



* £ 



J5 * = 
-- 05 C 






— . « — B-' 



ig-a 



a s 

o o 



S 3 

-3 - 



T3-3 o 

§§:^ 

o « — 

G O O 

o 3 c 
* * 2 



IS 2 3 « 



o o 



a £ m--h; 



§§:■£§ 

v S — 5 

-;.-■"_ 

H°l 

O O u_ O 

* * S * 

o U <- O 



§22 



o 



r _i 



o 

— re w "? 

_ a — co 

'g © — I- © © 

c — — — — ~ 

I 



- : ci 



o 



r-, 

c 






6 s - 



x x 



: 
•M 



o 



- 



1 - | 

8-11 



Ife 






7' 
g 

B S 



O 

spa 



II 



fe 



S o 









a 

i ~ 

n.| 

s « So 

J s - 
gd|| 

-- d ■ 



II 

e,2 



u 

* £ 

•Sit 

£>. 



P>« 



some evidence has been found with regard to phase changes talcing 
place at the Temperature of boiling oxygen (— 180° C.) in some 
of the alloys referred to below ; in one of the less stable alloys 
examined, non-magnetic properties were still found at the tem- 
perature of solid carbon dioxide (— 80° ('.). 

Table 1 gives details of materials at present commercially 
available. 

MAGNETIC COMPENSATING ALLOYS 

Magnetic compensating alloys, like the non-magnetic irons and 
steels, depend for their special properties on the effects of their 
relatively low magnetic transformation temperatures. For these 
alloys however, materials are selected with Curie points slightly 
above working temperature, so that a rapid change of magnetic 
properties (e.g. permeability and saturation values) with tempera 
tore results. 

Applications 

Such alloys find application in the construction of electricity 
meters and speedometers whose indications are required to be 
independent of temperature. In the electricity meter, 23 consider- 
able temperature-errors would be introduced by the relatively 
high resistance/temperature coefficient of the eddy-current 
braking disc ; with the usual aluminium disc an increase in speed 
of 0-4 per cent, per °C. would result from the reduced braking 
effect. A further, smaller, error in the same direction is likely to 
be caused by the reduction in flux in the permanent-magnet 
system. The method of compensation adopted is to arrange, 
across the poles of the permanent magnet, a diverter magnetic 
shunt having a reluctance which increases appropriately with 
temperature. The desired characteristics can be obtained from 
a variety of alloys with Curie points not much above normal 
ambient temperatures. The useful characteristics of the nickel- 
iron series can again be employed ; other alloys also used are 
those from the nickel-copper series in the region of 70 per cent, 
nickel.* 

These alloys may be used in a similar manner to compensate 

• 

* The materials should ho specially made to moot appropriate magnetic 
speefcBcationSi as without special precautions the magnetic properties of alloys 
made commercially for other applications may show wide variations. 



176 



MAGNETIC-COMPENSATING ALLOYS 



speedometers of the " drag disc " type. 24 In such cases the main 
deflectional force is that caused by the eddy currents induced in 
an aluminium disc or cup by the rotation of a small permanent 
magnet. Without suitable magnetic compensation, temperature 
errors up to 10 per cent, might be found in automobile speedo- 
meters, and considerably more in aircraft equipment. 

The same method of compensation can sometimes usefully be 
employed in conventional moving-coil electrical indicating instru- 
ments. Normally voltmeters are compensated by a series swamp- 
resistance of manganese or other alloy of low temperature coeffi- 
cient, and ammeters usually include a copper shunt making 
temperature compensation unnecessary. Where high sensitivity 
is called for, as in pyrometer instruments for temperature indica- 
tion from thermocouple output, magnetic compensation has 
advantages : a combined compensation for cold-junction tempera- 
ture and instrument resistance variation can be arranged 25 with 
the use of temperature-sensitive magnetic alloys. 

Low Curie-point alloys have been applied to temperature- 
operated relays. These may operate by the movement of an 
armature of which the position is determined by the counter forces 
of a spring (constant) and a magnet (temperature-dependent). 
Although bimetallic strip and other differential expansion devices 
are more usual, sometimes greater simplicity and convenience of 
design may be gained by the use of the magnetic method. Further 
applications are to the construction of small transformers and 
choke coils with temperature-independent output, or with special 
temperature characteristics, for compensation purposes in 
instrumentation. 



Practical Alloys 

The alloys used for these applications are usually those deve- 
loped from the nickel-iron and nickel-copper systems. For com- 
pensation purposes a Curie point in the region of 80° C. to 100° C. 
will usually give the requisite variation over the normal range of 
ambient temperatures. From Fig. 6-1 it will be seen that the 
composition required for the appropriate Curie point approaches 
the region in which instability may occur. This difficulty has been 
overcome by the addition of chromium and/or other elements, 26 
or by special processing. 

A difficult problem in the production of such alloys is the 



MAGNETIC -COMPENSATING ALLOYS 



17' 




-40 



-20 



20 40 

tEMPERATURE,°C 



60 



80 



Wig, 6-2. — Characteristics of magnetic compensating alloys. 



Tm:I.I: 2. MAli-NKTTC < '<)M I'KNSATING Ail.OYS 



Trade nam 


Manufacturer 


Notes 


Mutemp 


Richard Thomas & Bald- 
wins Ltd. 


Nickel-iron alloy. 


Teloon R2709 alloy 


Telegraph Construction & 
Maintenance Co. Ltd. 


Nickel-iron alloy. 


Temperature Com- 


Carpenter Steel Co.. 


Nickel- iron alloy, available in 


pensator 80. 


I'.S.A. 


th roo grades. 


Hoskins alloy 567 . 


Hoskins Mfg. Co., U.S.A. 


Nickel -iron alloy. 


.JAE metal ■ 


Henry Wiggin & Co. Ltd. 


Nickel-copper alloy. 




C.E.C, of America 


Niokol-coppor alio}'. Originally 






called Thcrmallov. 


N.M.H.G. alloy . 


Acierics d'Imphy. Krance 


Xickel-iron alloy ; choice of 
compositions available for 
different temperature ranges. 


Thermopefm 


Krupps, Germany 


Nickel-iron alloy. 



ITS 



M A ( 3 -N E TIC- M I* E N S A TING ALLOYS 



extension of the operating range to cover the requirements of 
aircraft, etc. in which operating temperatures may vary over 
100° G. or more. Another use with stringent demands is for 
domestic electricity meters lor outside installation, a practice not 
uncommon outside Britain. A solution can sometimes be found 
by the combination of two or more alloys. 

Xickel-copper alloys have also been used extensively for 
magnetic compensation. It has been stated that their properties 
are easier to control during manufacture ; their permeabilities are 
in general lower, but this can usually be circumvented by suitable 
shunt design. The temperature range over which a linear 
temperature permeability relationship is obtainable is consider- 
ably lower ; for this reason future development would appear to 
be with the nickel-iron alloys. 

Characteristics of typical alloys are shown in Fig. 6-2, and 
Table 2 gives a list of materials commercially available. 

As regards special alloys for thermally operated relays, etc., 
with various Curie points, there appear to be few standardized 
materials commercially available, although there has been 
practical application of such alloys. 27 



References 

10. l Impkinson. .1. " Magnetic Properties of Alloys of Nickel and Iron," 
Froe. Royal Soc. fLond.), 1889-90, 47, 23-4. 

20. Mkkkw. P. I). "Constitution of Iron-Nickel Alloys," Amer. Soc. 

Metals Hninll k. 1936, pp. 271-3. 

21. AUSTIN, J. B., and Mii.i.ku, S. D. " Magnetic Perrncabi lit y of some 

Austenitic [ron-Chrominm-Nickel Alloys as Influenced by Heat 
Treatment and Cold Work," Trans. Amer. Soc. for Metals, Sept. 1940. 

22. Kvkkkst, A. H. " Engineering Applications of Spheroidal Graphite 

Cast Iron," Paper to 4th Internationa] Mechanical Engineers 1 
Congress, Engineer, i<)">2. 193, June 13th and 20th, 794-5, 838-40. 

23. KrN\AKi>. J. !•'., and F.us, H. T. " Temperature Errors in Induct inn 

Wait Hour Meters/' Trans. Amer. hist. Elect. Kngrs., 44, 1025, 275. 

24. Hinplky. \V. N. " ' Thermoperm ' and Magnetic Tachometers," 

li.l.O.S. Final Report 1259. II.M.S.O. 

25. Kixnakd. 1. P., and Faus, H. T. " A Self-compensated Temperature 

Indicator," Jul. Amer. hist. Kiev. Kngrs., 1030, 49, 343-45. 

26. JACKSON, L. I. R., and Russell. H. W. "Temperature Sensitive 

Magnetic Alloys and Their Usee," Instruments, 193s, 2, 279-82. 

27. " Magnetic Fire Detector," Engineering, 169, Feb. 10th, 1950, p. 161. 



7. MAGNETIC RECORDING MATERIALS 
By 

C. Gordon Smith, M.A., A.M.I.E.E. 

The original idea of the magnetic recording of sound is attri- 
buted to Poulscn. His invention, the '* telegra phone." was 
covered by a Danish patent in 1899. 28 The essential features of 
this machine consisted of a helical coil of steel wire mounted on 
a rotating cylinder against which rested two small electromagnets 
respectively for recording and reproduction purposes. 

Development of Magnetic Recording 

Magnetic recording apparatus has altered little in principle 
since that time, but many advances have been achieved by careful 
design and the use of improved materials. Attention has been 
given to the theory of magnetic recording, and although the 
quantitative application of magnetic data in design has not 
proved very tractable, simple considerations give an indication 
of the properties required of the magnetic recording medium. 

Essentially both the recording and reproducing mechanisms 
consist of a high-permeability magnetic core with a magnetizing 
winding and small air-gap. together with the recording medium 
in the form of a wire or tape which can be drawn past and in close 
proximity to the air-gap. The magnetic core is usually toroidal. 
as indicated in Fig. 7 1. and composed of nickel-iron high- 
permeability laminations, with a radial gap so that the main 
reluctance controlling the flux in the magnetic circuit is thai of 
the gap. The recording medium is arranged to move at steady 
speed, generally in a direction parallel to the leakage ilttx from 
the gap. Thus the wire or tape becomes magnetized Longi- 
tudinally, and as it moves elemental magnets with like poles in 
juxtaposition are formed, their axial length being dependent mi 
the frequency of the recorded signal and the speed of motion. 
The effective magnetization retained by the elemental magnets 
will be reduced by their self-demagnetization, which becomes 

170 



ISO 



MAGNETIC RECORDING MATERIALS 



HIGH-PERMEABILITY CORE 




PATHS OF 
LEAKAGE FLUX 



CONFIGURATION OF 
ELEMENTAL MAGNETS 



Fig. 7-1. — Diagram of magnetic recording and ok kkfkoducing mechanism. 



greater as the effective axial length is reduced. Thus a reduction 
of response will be produced by a rise in the frequency, and the 
main problem is to obtain a good high-frequency response without 
excessive tape speeds. The demagnetization factor is also affected 
by fundamental magnetic properties and it can be shown to be 
roughly proportional to the ratio of remanence to coercive force. 
The most satisfactory results with materials for which this ratio is 
low have been achieved with those of lesser remanence. Although 
sensitivity is proportional to remanence, low sensitivity can be 
compensated by increased amplification. 

Other causes of distortion lie in the non-linearity of the magneti- 
zation characteristics inherent in most materials. No possibility 
of controlling the shape of the magnetization curve is likely, but 
the introduction of an alternating bias at supersonic frequency 
helps materially in reducing non-linear effects ; this is now 
general practice. 

A number of subsidiary considerations also enter into the 
choice of materials. The effect of an additional air-gap, actual or 
effective, between the recording medium and magnetic head has 
further important influence on the response to the higher fre- 
quencies, and for this reason it is essential to reduce Ihe L r ap by 
running the recording wire or tape almost, if not quite, in contact 
with the pole pieces. The material must thus have a smooth 



MAGNETIC RECORDING MATERIALS 



181 



surface, reasonable resistance to wear, and adequate strength to 
withstand tension in the winding mechanism. 

The question of optimum wire — or tape — thickness 28 has also 
been considered. Thick tapes might be thought to give an 
inferior frequency response, as demagnetization effects increase 
with the cross-sectional area of the elemental magnets. Experi- 
ment, however, has shown that effects due to increase in thickness 
are masked by lack of penetration and that improvement in 
response by decreasing the thickness occurs only with dimensions 
which are mechanically impracticable. 

The Magnetic Materials 

Developments in materials have proceeded along two general 
lines. The original . carbon-steel wire of Poulsen gave way to 
tungsten and other alloy steels following the general trend of 
permanent-magnet material developments. However, the later 
alloys such as Alnico, Alcomax, etc., cannot be produced in 
wire or tape form. Some attention has been given to ductile 
magnetic alloys, and the limited production of the copper- 
nickel-iron ("Cunife") and copper-nickel-cobalt (" Cunico ") in 
Germany and the U.S.A. has resulted. The former has been used 
successfully in tape form for magnetic recording although expense 
and production difficulties have limited its application. Another 
successful development has been a non-magnetic base wire or tape 
which is covered with a magnetic coating by electrodeposition. 
In general, the properties of electrodepositcd magnetic materials 
are favourable, as the characteristic high stress usually leads to a 
large coercive force. An example of such a material is a recording 
medium made by the Brush Development Company consisting of 
a brass wire coated with a nickel-cobalt alloy by electrodeposition. 
Of particular technical interest and of considerable promise are 
the austcnitic nickel-chromium stainless steels. 30 Normally such 
alloys are non-magnetic, but on cold-working, magnetic properties 
appear on account of the formation of small particles of ferrite 
(magnetic) dispersed throughout the austenitic matrix. The 
properties developed are particularly suitable for magnetic 
recording, and can be adjusted and controlled by heat treatment 

of the wire. 

The second class of recording media consists of non-metallic 
(e.g. paper or plastic) tapes, either coated or impregnated with 



182 



MAGNETIC RECORDING MATERIALS 



powdered magnetic materials. In such tapes the subdivision of 
the magnetic constituent by inter-particle air-gaps effectively 
reduces the remanence and thus improves the ratio coercive force/ 
remanence. Limited use has been made of powdered Alnico 
dispersed in a plastic medium, but the materials so far most 
successfully developed and applied are the magnetic oxides of 
iron. These can be produced in the form of very tine particles ; 
these arc desirable for the reduction of background noise, which 
might be caused by magnetic discontinuities in the recording 
medium comparable in size with the elemental magnets produced 
by the recorded impulses. 

The above considerations have weight according to the severity 
of the requirements : e.g. the object may be the faithful reproduc- 
tion of music, or only the intelligible reproduction of speech. 
Further applications of magnetic recording are found, in which the 
required magnetic properties are much less severe. An example 
is in the storage of signals in electronic computers, which has been 
successfully achieved by the use of an electroplated nickel film 
on a metal drum. As a further instance, the principle has been 

Table 1. Magnetic Recokdixc Matetuals 



Muti-riiil 


Remanence 

'v 


Coercive 

fore- //, 


K* 


Form 


Botes 




Wb/m« 


A/im 








Carbon rted 


0-fl 


4,000 


Mi 


W in- t.r taps. 


Barb nsea now obsolete. 

i -i.i in .Marconi-siiiii: re- 


.v\, tungsten stool 


1-0 


5,200 


bfi 


Wire or tape. 












cording equipment by 












B.B.C. until in ml ly. 


Contfe (ooppei 


04 


40.000 


ir.(» 


T;i)ir. 


Limited dm tor recording 


nickcl-lrou ulloy) 










in QermatM and 1 S.A. 

i c-tly and dillli'iilt to 
prodllri'. 


Nlckcl-colmlt 


li.i 


lfl.000 


3-5 


Plated on to 
braaa wire 

or tii]".'. 


Limited us,- in u.s.a. 


Btatnleai rtoel 


uiiii-o-a 


10,000-82,000 


30-00 


Win-, aoM 


DanaOy the normal stab* 


(18% Or, 








ilniwn mid 


less steel as supplied by 


XI). 


» 






beat 

t rent I'll. 


\ uriotis mamiiiK-turiTs. 
Bpeelal grade designated 
Tophci -M produced by 
Wiiimr B. Driver Co., 






















1 .S.A. 


Magnetic oxide 


o-oa o-os 


(,000 82,000 


60-00 


Fftpfc 


Magnet to oxide ii'c,o, or 


Impregnated 










i Tjii,! dbflpersM In cellu- 


pwnt 










lose acetate or p.v.o. tape. 
Produced in this cotmtry 
by K.M.I., O.B.O, and 
other, manufaotmers tad 
In U.S.A. and Qermany. 



* Relative higli-froquency response determined by ratio of coercive force to 

ri'iimiH-nc •. 



MAGNETIC RECORD INC MATERIALS 



183 



used for the indication of the contents of tins in food-canning 
factories, where other means of labelling during processing are 
difficult, the magnetic properties of ordinary tinplate being 
adequate for a number of code patterns to be impressed and 
retained on the cans and subsequently revealed by an electronic 
scanner. 32 

Table 1 gives a list of materials which have found commercial 
application, with notes on properties, etc. 



References 

28. Pouxsen, V. British Patent No. 8961 (1899). 

29. Kornei, O. "Frequency Response <>( Magnetic Recording," Bleo- 

ironies, 1947, 20, August, pp. 124 -28. 

30. Hobson, P. T. "Developments in Magnetic Recording," Electronic 

h'txj., Dec. 1947, 19, 377-S2. 
Hobson, P. T.. Chatt, K. S.. and Osmond. W. I\ " Magnetic Study 
of Stainless Steel Wirefe," Jul. Iron ds Steel Inst., 1048, 159, 145-57. 

31. Putxino, M. J. L. " The Magnet 0| >l ton Bound Recording and Repro- 

dueing System," B.I.O.S. Final Mi -port No. 951, B.M.S.O. 
Tiiiessex, G. J. "The Magnetophon of A.E.G.," B.I.O.S. Final 
Report No. 207. II.M.S.O. 

32. Glmpebtz, D. G. "Magnetic Sorting on UnlabeUed Pood cans," 

Electronics, Sept. 1952, 100-1 05. 



8. MAGNETOSTRICTIVE MATERIALS 

By 

C. Gordon Smith, M.A., A.M.I.K.E. 

The significance of magnetostriction in modern magnetic theory 
and its connection with fundamental properties has already been 
considered in Section 1. 



Principles of Application 

There are certain applications in which the magnetostrictive 
effect is of direct importance. These normally concern the con- 
struction of electro-mechanical transducers for which the utiliza- 
tion respectively of the Joule * and Villari f effects proves to be 
convenient, particularly when oscillations in the higher audio 
and ultrasonic ranges are required 1 . Although magnetostrictive 
effects are small (the change in length of a specimen of nickel, in 
which the Joule effect is relatively large, is only of the order of 
30 parts in L0 (1 when magnetized to saturation), large amplitudes 
may be built up by resonance in a suitably dimensioned magneto- 
strictive core excited by an alternating field of appropriate 
frequency, thus providing a powerful source of ultrasonic; energy. 

In most magnetostrictive applications other magnetic properties 
play an important part. For example, the efficiency of energy 
conversion may be reduced by excessive eddy-current and 
hysteresis losses. The value of incremental permeability may be 
important in the case of a transducer for converting mechanical 
vibrations. Another important requirement is some form of 
constant magnetic polarization of the vibrating specimen. The 
reasons for this will be apparent when studying Fig. 8-1. which 
shows a typical J curve relating change in length to applied 
magnetic field. First, it should be noted that at low field strengths 
the slope of this curve decreases considerably. Thus, to maintain 

* Change in dimensions clue to magnetization. , 

+ Change in magnetization due to stress. 

% Data from different sources show considerable variation ; the curve givon 
represents a lair average. 

184 



MAGNETOSTRICTIVE MATERIALS 



185 



efficiency, bias is important in a transducer dealing with small 
energy levels such as in the reception of ultrasonic signals ; the 
optimum biasing point will be in the region of field strength equal 
to half or rather more of the saturation value. Further, unless a 
biasing field is applied the frequency of mechanical oscillation will 
be double that of the applied magnetic field, which is generally 
unfavourable and may produce an undesirable waveform. 

In many types of apparatus a biasing field may be introduced 
wit hout difficulty, but sometimes there will be technical difficulties 
in the design of the core or practical difficulties in the provision of 
the necessary permanent- or electro-magnet, in such a case the 
permanent magnetization of the magnetostrictive element itself 
may be employed ; success, however, depends on the use of a 
material with adequate magnetic remanence and coercive force, 
properties not usually found in conjunction with an optimum 
magnetostrictive effect. 

Practical Materials 

The commonest material is nickel, which has a large magneto- 
strictive effect. In general, for optimum properties, the metal 
should be of good commercial purity and folly annealed, but if 




-so 



40 



30 



2C 



10 10 20 

APPLIED MAGNETIC FIELD, H 



50 + 
OERSTEDS 



Fiij. 8-1. — Magnetostbictive effect in nickel. 



186 



MAGNETOSTRICTIVE MATERIALS 



adequate remanence for self -polarization is required, the material 
may be used in a half-hard condition. As alternatives, alloys in 
the nickel-iron series are promising ; as will be seen from Fig. 8-2, 
the magnetostrictive effects in the region of 20 and 45 per cent, 
nickel, although of opposite sign to that of nickel, are nearly as 
large. The 45 per cent, nickel-iron alloy would have the further 
advantage of lower eddy-current losses on account of its higher 
resistivity (about eight times that of nickel), lower hysteresis 
losses and higher permeability. Although eddy-current losses 
can be reduced by lamination, the optimum thickness would be 
in the region of 0-005 in. for nickel at 10 kc/s, a value which 
involves considerable expense in fabrication. Thus an alloy with 
lower inherent losses presents practical advantages. Where large 
power outputs are required, the dissipation of the energy consumed 
in losses may present problems, in the solution of which the nickel- 
iron alloys would present distinct advantages. 

The cobalt-iron alloys are also of interest in view of the high 
magnetostrictive effects available with certain compositions (see 
Fig. 8-3). Their use, however, has been commercially restricted 



X10" 



30 



20 



*> 



10 



-10 



-20 



-30 





















/ 


,''"' 


\ 


1 
/ 
/ 


■"'X^ 


N 








/ 

/ 
/ 
1^^ 




\ 
\ 

1 

\ 

\ 


/ 
/ / 


^1 


= 25 




\ 








1 \ 
1 > 




1 
1 
\ 


i // 




H=10 




\ 






W 1 n — 






"""si 




















__ 10 _ 




















L_S5_J 




















\ > 
250 N V 























10 



20 30 



40 50 60 70 80 

% NICKEL 



90 100 



Fig. 8-2. — Magnetostrictive effect of nickel-ikon alloys. The values 

OF H ARK GIVEN IN" OERSTEDS. 



MAGNETOSTRICTIVE MATERIALS 



187 




Wig. S-3. — Magnetostrictive effect of cobalt -iron alloys. The values 
of n AUK i;l\ K\ IN oersteds. 



on the grounds of cost and difficulty of fabrication, although 
considerable progress has recently been made in overcoming the 
latter. Limited use has also been made, more particularly in 
Japan and Germany, of aluminium-iron alloys, investigated 
originally by Masumuto 2 and developed under the name 
" Alfer." These alloys are of lower technical efficiency, their 
development having been due to nickel shortage. 

Applications 

Probably the most widespread application of ultrasonic vibra- 
tions is in their use for the detection and location of underwater 
objects. 3 The original crude methods of echo ranging have been 
extensively developed and improved, advantage being taken of 
the possibilities in reflection and concentration of the shorter 
ultrasonic waves. Magnetostrictive transducers using nickel rings 
are employed both for the generation of pulses of energy (usually 
with a frequency of 10 to 15 kc/s) and for the detection of 
the reflections from a submerged object. The depth, deter- 
mined by the time interval between transmitting and receiving 

S 2 



188 



MAGNETOSTRICTIVE MATERIALS 



pulses, is usually recorded on a moving chart, and the direction 
of the object may be indicated with a focused beam by the usual 
scanning devices. The original application of depth sounding has 
been extended to the detection and location of various underwater 
objects such as submarines, fish shoals and wreckage., and magneto- 
strictive devices are now being installed in practically all sea-going 
vessels. Similar methods, but on a different scale and using in 
general much higher frequencies, may be used for the detection of 
internal flaws in the non-destructive testing of metal objects. 

Other uses of ultrasonics which have aroused considerable 
interest of recent years depend on the peculiar physical effects of 
high-frequency vibrations on various materials. As examples, 
the emulsification of oils, dispersion of solids in liquids, and the 
killing of bacteria may be mentioned. An interesting use is the 
ultrasonic soldering iron developed by Mullard Ltd.. for the 
tinning of aluminium, in which the effects of ultrasonic vibration 
are used to cause the continuous dispersal of the oxide film which 
tends to form on the aluminium surface. In such processes 
considerable energy is usually needed and magnetostrictive 
transducers are particularly suitable on account of their robustness 
and power-handling capabilities. In other applications for which 
high intensities are not called for, piezo-electric and ferro-electric 
transducers are frequently used. These have relatively poor 
mechanical strength. In general, apart from the question of power 
output, the inherent physical characteristics of magnetostrictive 
transducers tend to be more suitable for the lower frequencies, 
whilst those of piezo- and ferro-electric devices favour use at the 
higher frequencies. 

.Magnetostrictive materials have been used in vibrating fre- 
quency standards, in which the frequency of a valve-maintained 
oscillator is controlled by the mechanical vibration of a suitable 
magnetostrictive element. Alloys with special properties are called 
for when a high precision of frequency is needed. It is fortunate 
that alloys already developed on account of their constancy of 
clastic modulus (and thus vibrational frequency) at different 
temperatures show in most cases appreciable magnetostrictive 
effects. These are modifications of the well-known Invar type, 
e.g. Elinvar, Chronovar, Ni-Span " C," etc. The behaviour of 
alloys of this type with and without minor modifications has been 
studied in detail by Ide 4 for frequency-standard applications. 



MAGNETOSTRICTIVE MATERIALS 



189 



References 

1. Bkhcm ann, L. " Dor Ultraschall.'* Pub. S. Hirzel Wring. Zurich. 1949. 

6th edition. (Translation of early edition, Bell & Sons, London, 1938.) 
-2. Maskmoto, H. " Dynamical Characteristics of Magnetostriction Alloy 

4 Alter.' " Sti. Rep. Res. Inst., Tohoku Univ. Series A, June 1950, 2, 

413-19. 

3. Calway. IT. "Echo Sounding at Sea (British Practice)," Pitman, 

London. 1951. 

4. Ide. .1. M. "Magnetostrictive Alloys with Low Temperature Coeffi- 

cients of Frequency," l J roc. I.R.E., 1934, 22, 177-90. 



INTRODUCTION TO M.K.S. UNITS 



191 



APPENDIX 
INTRODUCTION TO M.K.S. MAGNETIC UNITS 

By 

Professor M. G. Say, Ph.D., M.Sc, M.I.E.E. 

The unit system employed in this book is the metre-kilogramme- 
second system in its rationalized form. The M.K.S. system was 
developed in 1901 by the Italian engineer Giorgi, and indepen- 
dently three years later by Professor Robertson of Bristol. It 
was later brought to the attention of the International Electro- 
technical Commission, which in 1935 signified approval of the 
system for international use. Certain details were not finally 
settled until the 1938 and 1939 meetings, and the I.E.C. formally 
adopted the rationalized M.K.S. system in 1950. However, well 
before that date the system had become common in Europe and 
almost universal in the United States. 

Basic M.K.S. Units 

The M.K.S. units form an absolute system in which the electrical 
units are the commonly used " practical " volt, ampere, ohm, 
coulomb, farad, henry, joule and watt. The unit of power on a 
basis of mechanics must also be the watt. This follows auto- 
matically with the kilogramme, metre and second as units of 
mass, length and time respectively, because the unit of force is 
that necessary to endow a mass of one kilogramme with an 
acceleration of one metro per second per second. This unit of 
force is called the newton. A power of one newton -metre pel 
second is one watt, and an energy of one newton-metre is one 
joule. 

The magnetic units are selected to correspond with these 
electrical and mechanical units to give a one-to-one ratio between 
related quantities. For this purpose the magnetic and electric 
properties of free space are essentially recognized, and are not 
suppressed or ignored as has not been uncommon in the use of 
the classic C.G.S. electromagnetic and electrostatic unit systems. 

180 






The C.G.S. Unit Systems 

In the C.G.S. electromagnetic (e.m.) system, the unit pole (and 
thence unit current, charge, etc.) is defined from 

taking/ = 1 dyne, the distance 5 = 1 cm, and fx as unity. In 
the C.G.S. electrostatic (e.s.) system the unit charge (and thence 
unit current, etc.) is defined from 

/ - <M 2 :'*o* 2 
with/ = 1 dyne and s = 1 cm., and taking x as unity. But, by 
Maxwell's electromagnetic theory, in any consistent system the 
product [M Q y. Q = 1/c 2 , where c is the velocity of free-space electro- 
magnetic propagation in the chosen units of length and time. It 
follows therefore that the C.G.S. e.m. and e.s. systems are not 
mutually consistent : a unit quantity in one system differs in 
magnitude from the unit of the same physical quantity in the 
other. Thus 1 e.m.u. of current has the magnitude of 3 X 10 10 
e.s.u. of current, and neither has the magnitude of 1 A. 

For many years engineers and (to a limited extent) physicists 
have been using a mixture of C.G.S. e.m. and e.s. units, together 
with the " practical " system (volt, ampere, ohm, henry, farad) 
based on decimal multiples of the e.m. units. In general, the 
" practical " units have been employed only in circuitry, the basic 
e.m. units only for magnetic -field problems, and the e.s. units 
only for electrostatics. Inconvenient conversions involving 
powers of 10 and of c have had to be emploj^ed to relate field 
quantities to circuit theory. 

The M.K.S. system replaces these three mutually conflicting 
systems by one comprehensive, unified system. It involves 
certain small sacrifices that are greatly outweighed by important 
and valuable advantages. The question of rationalization is a 
separate one. but the reform is most conveniently made simul- 
taneously with the adoption of M.K.S. units. The M.K.S. 
system in its rationalized form has been commended by the 
Institution of Electrical Engineers ; while B.S. 1637 : 1950 defines 
the system and gives a historical note on its development. 

The M.K.S. Unit System 

This is an absolute system with the metre [m], kilogramme [kg] 
and second [s] respectively as units of length, mass and time. 



192 



INTRODUCTION TO M.K.S. I' NITS 



INTRODUCTION TO M.K.S. UNITS 



193 



The two former are defined arbitrarily by international platinum 
standards maintained at Sevres, and the latter is 1,80400 of a 
mean solar day. The electrical units are the familiar practical 
ones — volt [V], ampere [A], ohm [Q], coulomb [CJ, farad [F], 
henry [H], joule [J] and watt [W]. The mechanical units are. for 
force the newton [N], for energy the joule and for power the watt : 
these link conveniently with the established electrical units in 
the manner already described. 

The absolute volt and ampere are fixed (a) by the convention 
that 1 V X 1 A = 1 W. and (6) by the definition of the ampere 
as that steady current which, maintained in two parallel conduc- 
tors of infinite length and negligible cross-section, and separated 
by a distance of 1 m. in vacuo, produces between the conductors 
a mechanical force of 2 x 10 -7 N per m length. The force is 
conceived to arise magnetically in accordance with the " inter- 
action law " based on experiment : the current in one wire is 
taken as lying in a magnetic field produced by the other. For a 
length I metres of a parallel-wire system with spacing s metres, 
a current I x in one wire produces at the other wire a magnetic 
flux density proportional to I v inversely proportional to the 
distance s, and dependent on the magnetic property /x of the 
siuTounding free space. The second current lies in this magnetic 
flux density and the resulting mutual force is 

/ =-- KI^^jiqIs newtons 

where K is a constant. In the terms of the definition of the 
ampere, I x = I 2 = 1 A, l = s = 1 m ; and in consequence 

Kii = 2 x 10- 7 . 

The value of K, and consequently that of fj, , depends on the 
manner in which current is related to magnetomotive force, F. 
The relation * may be taken as F = 4ttI or as F = I : the 
former gives the unralionolized and the latter the rationalized 
system. 

The concept of the isolated magnetic pole is now commonly 
discarded, and magnetomotive force is treated as a line-integral 
of magnetic field strength. Therefore, taking F — I, the m.m.f. 
per metre length of a circular path of radius s is H = I/2tts, the 

* / represents the current-turn product, which is dimensionally the same as 
current, [A]. 






flux density produced by it is B = [XqH = hqI 2tt5, and the force 
equation above becomes 

/ = (I/^tO/jV,,/ -'. 

which gives A' = 1 ;2-n. Tn consequence, /x = 47r/10 7 . (As a 
further consequence of the /x x product, x = 1/367T X 10 9 ). 

If the m.m.f. is taken as 4ttL then /x = 10" 7 (and x is then 
1/9 X 10 9 ). This alternative will not be considered further, here, 
since this book uses rationalized M.K.S. units. 

Rationalized M.K.S. Magnetic Units 

Consider a single-turn resistanceless coil in vacuo. To establish 
a current of 1 A in it requires the application of a voltage for a 
time such that the time-integral of the voltage is proportional to 
the final current and independent of the manner in which the . 
current changes from zero to 1 A. According to the usual concept, 
the space in and around the coil becomes the seat of a magnetic 
flux which is measured by the voltage-time integral : 

=fvdt [volt X second (V.s)]. 

The more usual name for the flux unit is the weber [Wb], which 
has the physical dimensions of [V.s]. The flux is also proportional 
to the current, i.e., 

=LI, 

where L is the self-inductance in henrys [H] corresponding to 
[Wb A] or [V.s/ A]. The recoverable energy stored in the magnetic 
field is 

W - WI [Wb X A == V.s x A = J]. 

Experiment shows that if the coil is arranged so that the 
magnetic field is uniform within it and negligible outside (e.g. by 
completing the magnetic circuit through other coils), the magneto- 
motive force of the coil current accounts completely for the flux 
within its length. Then the total flux for a given coil current is 
proportional to the cross-sectional area A and inversely propor- 
tional to the length I. The flux density is B = 0/A [Wb/m 2 ] and 
the magnetizing force (or m.m.f. per metre) is H = Ijl [A/m]. 
Then the density is proportional to the magnetizing force and can 
be written 

B = ft H. [Wb/m 2 ]. 



194 



INTRODUCTION TO M.K.S. UNITS 



INTRODUCTION TO M.K.S. UNITS 



195 



If A is 1 m 2 and I is 1 m. the inductance of the unit coil is 

"V.s 



_ B 
L= 1 = H = 



Mo 



nr 



m 
A 



_ V.s _ HI 
A.m mj' 



The cubic unit coil has thus an inductance /x = 47r/10 7 [H/m]. 
The term " absolute permeability of free space," or better the 
" magnetic space constant." is used to describe //.„. 

Corresponding to the total energy \<P1 [J] there is the energy 
density 

\BH = \B*lp Q [J/ni 3 ]. 

If the vacuum in which the coil is immersed is filled with 
magnetic matter of relative permeability /t r . the self-inductance is 
found to increase \i r times for the same magnetizing force, so that 
now 

B = w^ = pH [Wb/m 2 ]. 

The additional density may be considered to be due to the 
intensity of magnetization J of the material in [Wb/m 2 ] : 

B = B + ./ = /x # + J [Wb/m 2 ]. 
Analogously, the relative permeability may be expressed as the 
sum of the free-space relative permeability (unity) and the 
susceptibility of the material : 

th = 1 + X- 
Table 1. Basic Physical QrAXTiTnss 



Physical 
quantity 


Equation 


M.K.S. 
unit * 


C.G.S. unit 


Conversion 


Length 


I 


m 


cm 


1 m 


= 10 2 cm 


Area . 


A =/ 2 


m- 


cm 2 


1 m 2 


= 10 4 cm 2 


Volume 


U =P 


m 3 


cm 3 


lm 3 


= 10 6 cm 3 


Mass . 


m 


kg 


g 


1kg 


= 10 3 g 


Time . 


t 


s 


s 






Velocity 


u - J/J 


m/s . cm/8 


1 m/s 


= 10 2 cm/s 


Acceleration 


a = lit* 


m/'s 2 cm/8 2 


1 m/s 2 


= 10 2 cm/s 2 


Force 


f = mu 


N dvne 


IN 


= 10 5 dynes 


Torque 


T =fl 


N-m dyne-cm 


lN-m 


= 10 7 dyne-cm 


Energy 


W 


J 


erg 


1J 


= 10' ergs 


Power 


P = Wjt 


W 


erg/s 


1 W 


= 10 7 ergs/s 



l 



* The abbreviations used for unit-names are explained on pages 191-192. 



i 



I 



2 E 



= 6p e 

ix 7 - 

<s © 



!<S 



1 1 hk% 

h pr jc p p a a 



-_ 3 
- -^ 

■2 £ 

~ 5 

oc 



a — i 

y. B 
=•3 Bjri 

s * a s 3.S 

- * J «■ D 

am aa ffs 



II II II II II II II II II 
=*, a: j=5 «• *■* < i a 



.< 4 U M £ <£- 



?^£ 






fciHj £Gp S "s y - 3 



Ifif 

* 3 *<i — > - 
2 C * ■= 
H •- <B 

C.5.2B 
_ _ _ _ 
© C g"0 



$ .t: 

- n 

a * 

o g go 

>. 5 s 

- I r 
■1 — -' 



D © TJ 1 g - 



■Sg 

-*s o 

g-o 

— - 

» o 
"3 <o 



- s 
00 H 



li 

o* o 

si 
*-8 



§ 



5- 



196 



INTRODUCTION TO M.K.S. UNITS 



It will be noted thai the use of the weber gives Faraday's law 
in the form 

e = — d<P;dt [V] 

without the factor of 10~ 8 required for this law in the older 
mixture of C.G.S. e.m. unit and practical unit (maxwell and 
volt). Rationalization leads to simplification in the m.m.f. 
relations, whereby the m.m.f. F = I instead of (MttI, and 
H = III instead of O-^nljl : it further simplifies the energy 
density to hBH [J, m 3 ] in place of BHI&tt [ergs/cm. 3 ]. The unit 
relations are summarized in Tables 1 and 2. 

The unfamiliarity of rationalized M.K.S. magnetic units is 
their only disadvantage. The maxwell, gauss and oersted must 
be replaced by the weber, weber per square metre and ampere 
(-turn) per metre. With the flux and flux density the changes are 
merely in the position of a decimal point, as the weber is 10 8 max- 
wells and the weber per square metre is 10 4 gauss. The latter is 
actually more convenient, as the flux densities set up in air-gaps 
for engineering purposes will frequently be found to have values 
of the order of unity. Tn the case of magnetizing force, the 
awkward factor 4tt occurs in converting H values from oersteds 
to ampere (-turns) per metre. However, users of B'H magnetic 
data frequently use a '" rationalized " version with B in gauss 
and H in ampere-turns per centimetre : in this case the conversion 
is again a simple decimal matter, as 1 A/m is 10~ 2 A/cm. 

Rationalization also affects the absolute permeability \i = BjH 
of a magnetic material. But if the magnetic properties of free 
space are clearly recognized as in the expression /z = n r fx , the 
relative permeability /x r remains unchanged in its familiar form 
and value. 



WIGAN 
CENTRAL 
LIBRARY 



INDEX 



A.C. bridge iron -loss tost or. 35 
A.C. potentiometer tester, 36" 
Ageing magnetic. 42 
Alcomax II, 3, 31. 90. 100, 101, 102, 

H)4. 109. 110. 120. 139 
Alcomax III. IV, 140 
■ Alfer," 187 
Alloying, 21 
Alloying 

iron and nickel. 20 

iron and silicon, 28 
Alni, 96, 98, 99, 109, 110 

AlnicM. 96, 98, 99, 101, 104, 109, 110 

Alnico D.G., 104 

A I perm . 79 

Alternators, turbo-, 26 

Anisotropic alloys, permanent magnet, 

99 
Anisotropic sheet, 26 
Anisotropy, 71 

silicon-iron. IT 
Anka. 174 
Annealing, 24 

magnetic silicon-iron, 49 

transformer -grade silicon-iron after 
cutting, 54 
Anomalous loss, 7 
Apparent eddy current loss, 6 
Armatures, 80. I 2 I 

instrument, 82 

of rotating machines, losses in, 3 
Armco iron, 4.">, NO 

magnetization curves, 42, 78 
Audio -frequency power transformers, 

81 

B.s.l. test pane], 124 
Ballistic galvanometer, 127 
Barkhausen jumps, 13 
Basic M.K.S. units, 190 
Becker's theory, 21 
Binary alloys, 44 
Blocfa wall, 19 
Bohr magneton, 7 

Brass, containing ferromagnetic im- 
purities, 23 

C cores. 71, 81, 83 

( I.G.8. unit system, 191 

( almaloy, 177 

Campbell iron-loss tester, 35 

Carbide-bearing permanent magnet 

steels, 96, 108 



C'arhonyl-iron powder, 166 
< 'airier-telephony, 94 

( last irons, non-magnet ic, 173 

Cathode-ray tubes. 94. 140 

Chokes, 81, 82 

Chronovar, 188 

chuck, magnetic, 80, 141. 142 

( 'hurdicr tester, 34 

Clockspring toroids, 71, 81 

Clutches, magnetic. 80 

Cobalt -iron alloys, magnetost rictive 
effect of, 1ST 

Cobalt . magnet izat ion curves for single 
crystals of, 14 

Coercivity. 2<>. 21 

Cold-reduced steel for transformers, 26 

( 'olumax, 104 

Columnar crystal magnets, 103 

Coma Hoy. 10a. 106 

Comol. 1 05 

Constant -permeability alloys, nickel- 
iron, 68 

Copper-nickel-iron alloys, 22 

Cores, 80 

Crystal, energy stored in, 17 

Crystal, single strip, 50 

Crystal structures. 13 

Crystallographic directions, magneti- 
zation curves in different, 14 

( Yystals. 

preferred orientation of, 25 
silicon-iron, 26 

Cunico I and II. 105, 106 

Cunife alloys, 105, 106 

Curie point, I I 

I Mii-ent transformers, 82 



D.C. applications, materials for, 44 
Demagnetization curve, 21 
Demagnetizing permanent magnet, 129 
I design, permanent magnet, 112 
Diffusion -hardening alloys, permanent. 

magnet, 96, 109 
Dispersion-hardened alloys, cooling in 

magnet ic field, 31 
Domain boundaries, is 

examination of, 15 
Domain dimensions, 1 6 
Domain-oriented alloys, nickel-iron, 66 
Domain theory, 7 
Domains, ferromagnetic, 11 
Dynamo iron, 24. 29 

silicon-iron, 53, 80 



IK 



198 INDEX 



Eddy current loss, apparent, 6 
Eddy current losses. 3 

in thick and thin plates. 5 
Electricity meter magnets, 13] 
Electrolytic-iron powder, 167 
Elinvar, 188 
Energy loss, 2 

Energy, magnet or ryst a I line. IS 
Energy stored in crystal. 17 
Epstein square tester, 34 



Faraday homopolar generator, I 24 
Fcrrite permanent magnets. 

application of, 1 68 

composition and manufacture, I 60 

theory of, 151 
Forrites, magnetically -soft, 85 

applications. !I4 

Curie point, 86 

electrical and magnetic properties, 85 

machining, 92 

mechanical properties, 90 

permeability, 86 

shape of pressings, 92 
Ferromagnetic domains. 11 
Ferromagnotie theory, 1 
Ferroxcube, 85 
Field, molecular, IS 
Filter coils, 82 
Flake-iron powders, 167 
Flux density, I 
Fluxmeter, 126 
Focusing magnets for cathode -ray 

tubes, 140 
Forgings of high-permeability alloy, 74 



Gadolinium, 10 

Gaussmeter, 126 

Generator, Faraday homopolar, 1 24 

Generator rotors, 80 

Generators, 80, 81, 121, 132 

Cram- and domain-oriented alloys, 

nickel-iron, 66 
Grain-orientation, silicon-iron, 47 
Grain-oriented alloys, nickel-iron, 64 
Gramophone pick-ups, 82 



H.C.R. metal, 82 
Hametag process, 168 
Heat-treatmont 

in hydrogen atmosphere, 24 

in magnetic field, 30 
Heusler alloys, 1 

High initial permeability alloys, 69, 81 
High-permeability powder cores, 165 
High -resistivity alloys, nickel-iron, 64 
Hoskins alloy 567, 177 



Hot -rolled non-oriented alloys, 51 
Hy.omax. 100, 101, 104. 140 
Hvdrogon atmosphere, heat -treatment 

in, 24 
Hvdrogen-reduced iron powder, 167 
Hynico, 98, 99 

Hysteresis loop, measurement of, 32 
Hysteresis loss, 6 

silicon-iron. 24 
Hysteresis, magnetic, 2 



Impi hities, effect of, 24, 41 

Incremental permeability in silicon- 
iron, effect of polarizing field. 64 

Inductive switching action, 82 

Initial relative permeability, 20 

Instrument armatures, 82 

Instrument magnets, 132 

Insulating nickel-iron laminations, 74 

Intensity of magnetization, 1 

Intermediate frequency transformers, 
94 

Interna] strains, 24 

Internal stresses, 20 

Iron, 40 

Iron-aluminium alloys, 78, 79 

Iron-aluminium-silicon powder alloys, 
169 

Iron nnd nickel, alloying, 29 

Iron and silicon, alloying, 28 

Iron-cobalt alloys, 75 

Iron domain, 13 

Iron, effect of impurities in, 41 

Iron-loss measurement, 34 

Iron, magnetization curves for single 
crystals of, 14 

Iron powder magnets, 23, 144 

iron powders, i •><> 

Isoperm nickel-iron, 69 



Jae metal, 177 



Laminations, nickel-iron. 71 
Langevin- Weiss theory, 8 
Lifting magnets. I 2 1 
Lloyd- Fisher tester, 34 
Loss, 

anomalous, 7 

eddy current, 3 

energy, 2 

factors, 161 

in armatures of rotating machines, 3 

rotational hysteresis, 3 

total iron, 6 
Loudspeaker magnets, 137 
Low-carbon steel, 42, 43, 45, 80 



INDEX 



199 



M.K.S. units, 190 

Machine frames, 80 

Magnadur, 150 

Magnetically-soft ferrites (we Ferrites) 

Magnetic ageing, 42, 53 

Magnetic amplifier, 61, 82 

Magnetic annealing, 49 

Magnetic change point, 1 1 

Magnetic chucks, 121 

Magnetic compensating alloys, 175 

Magnetic field, cooling in, 31 

Magnetic field strength. 1 

Magnetic irons and steels, commer- 
cially-available, 44 

Magnetic memory -storago devices, 82 

Magnetic powder cores, 159 

Magnetic -pulse generators, 82 

Magnetic quantities, 195 

Magnetic recording, 179 

Magnetic shielding, 82 

Magnetization curve, 1,9, 16 

Magnetization, intensity of, 1 

Magnetization process, 13 

Magnetizing permanent magnets, 122 

Magneto, 132 

Magnotocrystallino energy, IS 

Magnetostriction, 19, 20 

Magnetostrictive applications, 184, 187 
materials, 185 
transducers, 187 

Martensitic stools, 21 

Measurement 

hysteresis loop, 32 
iron -loss, 34 
magnetization curve, 32 

Mei-himical-contact rectifiers, 82 

Micropowder permanent magnets, 144 

Mild steel magnetization curves, 42 

Miller indices, 14 

Miniature transformer cores, 62 

Molecular field, 9, 18 

Molvbdenum-permallov powders, 168 

Motors, 80. 81. 182 

Mu metal. 29, 79, 81 

Mutenip, 177 



X.M.C. steel, 174 

X.M.H.G. alloy, 177 

Nickel-coppor-iron alloy powders, 168 

Nickel domain, 13 

Nickel-iron alloys, 21, 25, 30. 63 

applications of, 29, 81 

commercially-available, 71 

constant -permeability, 68 

domain-oriented, 66 

grain- and domain-oriented, 66 

grain-oriented, 64 

high initial permeability, 69, 81 

high-resistivity, 64, 82 



Nickel-iron alloys, 

high saturation-density, 82 

Isoperm, 69 

laminated assemblies of. 83 

magnetostrictive effect of, 186 

manufacture. 71 

Permivar, 69 

powders, 168 

rectangular-loop, 82 

Supermalloy, 68 

temperature-sensitive, 64 

tornary and higher ordor, 68 
Nickel-iron laminations. 71 

insulating, 74 
Nickel-iron non-laminated parts, 74 
Nickel, magnetization curves for single 

crystals of. 14 
Nickel, magnetostrictive effect in, 185 
Ni -Resist, 173, 174 
Ni-Span " C," 188 
Nomag, 173, 174 
Non -magnetic cast irons, 173 
Non-magnetic steels, 171 



Orientation, preferred, 21 



Permalloy C, 81, 82, 125 
Permanent columnar crystal magnets, 

103 
Permanent magnet alloys 

anisotropic, 99 

diffusion-hardening. 96, 98. 100,109 

precipitation-hardening, 105, 109 
Permanent magnet 

applications, 131 

carbide-boaring steels, 108 

demagnetizing, 129 

design, 1 1 2 

design, physical and mechanical 
aspects, 109 

efficient utilization on recoil, 119 

ferrites, 150 

handling and storage, li".i 

materials, 22, 95 

non-static working conditions. I 2 1 

steels, 97, 107 
Permanent magnets 

application of ferrite, 155 

for radio, 137 

magnetizing, 122 

micropower, 144 

precipitation-hardening. 1 06 

properties, 148 

sintered, 104 

special-purpose tests, 128 

stabilizing, 118 

testing, 124 



200 



INDEX 



Permeability, I, 20, 160 

initial relative. 20 
Permeance lines, 116 
Permivar, 69 
Pole pieces, 80, 83 

Power and dial ribution transformers, 81 
Powered iron magnets, 24. 14-1 
Powder core materials, 166, 169 
Powder coiv- 

applications and uses of, 163 

high-permeability, 165 

historical development. 169 
Precipitation-hardening permanent 
magnet alloys, 1 0.~». 109 

Preferred orientation, 27 

of crystals, 25 



Rao ah permanent magnets, 142 
Radiometnl. 82 

Kadio, permanent matmets lor, 137 
Rationalized M.K.S. magnet icuiuis, 193 

Reactors, 80 

Reco 2A, 98. 99 

Recoil lines, 117 

Recoil energy contours, 120 

Relays, 82 

high-speed, 82 
Remalloy, 105 
Retentivity, 23 
Reversible permeability, 117 
Rhometal, 82 
Rotational hystorosis loss, 3 



Satiuaiji.k reactor, 81, 82 
Saturation value, 1 

variation of. 30 
Sheet floater, 14 
Silicon content, effect of, 45, 53 
Silicon -iron, 25, 28, 4". 

annealing after catting, 54 

applications of, 80 

co Id -reduced, non-oriented, 57 

cold -reduced oriented. 57 

commercially available, 51 

crystals. 20 

dynamo grades, 80 

effect of frequency and flux density, 

54 
•_'! -ain-orientation, 47 
hysteresis loss, 24 
magnet ie ageing, 53 
magnetization characteristics, 53, 62 
magnetization curves, 78 
sheet. 3 

transformer sheet. 6 
use of cold reduction, 46 
variation in total loss with thickness, 

54 



Silver Fox. 174 

Single-crystal strip, 50 

Sintered permanent magnets, 104 

Small power transformers, 82 

Soft magnetic mat .-rials, 24, 37, 40, 43 

applications of, 80 

basic, 40 

ferrites. 85 
Stabilizing, permanent magnet. 118 

Staybrite, 174 
Steinmetz coefficient, 3 
Steinmetz empirical law. 3 

Steel, low-carbon-content. 45 

magnetization characteristics, 43 
Steels, tuartensitic. 21 

Steels, non-magnetic, 171 

Supermalloy. 2!). 68 

Swedish charcoal iron, 42, 80 



TBLOON R2799 alloy, 177 
Telephine diaphragms. S3 
Television receiver deflector coils, 166 
Temperature compensator. 3d. 177 
Temperature-sensitive alloys, nickel- 
iron, 64 
Ternary and higher order alloys of 

nickel-iron, 68 
Testing permanent, magnet, 124 
Thermoperm. 177 
Tieonal, 06, 100, 101, 102, 104, 109, 

110, 139 
Total iron loss, 6 

Transducers magnet ostrietive, 187 
Transformer sheet, 6, 53 
Transformers, 80, 81, 82 

and matching, 81 

audio-frequency power, 81 

cores of, 58 

current, 82 

intermediate frequency, 94 

miniature, 62 

power and distribution, 81 

small power, 82 
Turbo-alternators, 26 
" Two-state " characterist ics, 82 

Ultrasoxic applications, 187 

V-l'KKMKNpr-R. 83 

magnetization curves, 75 
Vectolite, 150 

Vibrating frequency standards, 188 
Vicalloy I, 106, 108 



W'au.ky permeameter, 32 
Yokes for machines, 57, 80 






"7*D7 



NEWNES 

ELECTRICAL ENGINEERING 

PROGRESS SERIES 

Consulting Editor: Prof. M. G. Say, Ph.D., M.Sc, M.I.E.E. 

Here is an important contribution to electrical 
engineering literature . . . six comprehensive 
authoritative books, each one covering its 
subject in an essentially practical fashion. The 
six titles in this series are : — 

ELECTRICAL EARTHING AND ACCIDENT PREVENTION 

Covers all that an electrical engineer needs to know about earthing practice, whether 
in power-system earthing, installations, or communications systems. Includes sections on 
the earthing of special installations, such as domestic, industrial, ship and mining installa- 
tions, and of different types of plant such as machine tools, portable appliances, welding 
equipment, and electromedical equipment. Specialist contributors throughout. 
Demy Svo. 2 $ s - ,iet 

ROTATING AMPLIFIERS 

Deals with the Amplidyne, Metadyne, Magnicon and Magnavolt and their use in 
control systems. In addition to the theoretical aspect there are separate chapters on the 
construction and lesting of the four commercial machines of this type, on the principles 
of control systems, and on the applications of rotary amplifiers. 
Demy 8vo. ]ls - 6d - net 

CATHODE-RAY TUBES 

A complete, authoritative account of current theory and practice for all types of 

cathode-ray tubes, including television, radar, instrument, camera, switch and storage 

tubes. Special sections on the flying-spot scanner, the monoscope, and oscilloscope 
technique are given. 

Demy Svo. 25s - nel 

MAGNETIC AMPLIFIERS AND SATURABLE REACTORS 

This important book clearly sets out the principles, construction and application of 
saturable reactors, magnetic amplifiers, and magnetic modulators. 
Demy 8vo. 2is - mU 

MAGNETIC ALLOYS AND FERRITES 

An up-to-date work covering theory, permanent materials, soft materials and miscel- 
laneous materials. A special point is the inclusion of practical information on both soft 
and hard ferrite materials 
Demy 8vo. 2is - net - 

CRYSTAL RECTIFIERS AND TRANSISTORS 

A comprehensive work on this subject. Of increasing interest to electrical engineers, 
particularly those who are concerned in any branch of communications, including 
telephony, radio, television and radar. 
Demy 8vo. , 2is - net - 



FROM ALL 
BOOKSELLERS 



Published by 

GEORGE NEWNES, LTD., 

TOWER HOUSE, SOUTHAMPTON STREET, 
LONDON, W.C.2 




MAGNETIC 
ALLOYS 

AND 
FERRITES 




Editor : 
M. C 



NEWNES 



i