IWQ— —-—- —w—'

 

 

DESIGN AND
CONSTRUCTION OF A

LABORATORY ELECTROMAGN ET

Thesis for the Degree of M. S.
MICHIGAN STATE COLLEGE

Arthur Jerome Luck
I950

 

 

W w IIIIIIII IIIIIIIIIIIIIIII IIIIIIIIIIIIIIIIIII

3 1293 01772 I35:

This is to certifg that the
thesis entitled

"Design and Construction of a Laboratory
Electromagnet."

presented by
Arthur Jerome Luck

has been accepted towards fulfillment
of the requirements for

MoSo degree in le31C8

QMJA/Irjr. C 70:"..7 r

Major professoru

 

Date may 15) 1950

0-169

 

 

 

 

 

LIBRARY

Michigan State

University

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1B8 cJCIRC/DateDue.p65-p.14

 

DESIGN.LND CONSTRUCTION OF A

LABOEA TOM ELE CTBCMAGNET

by

Arthur Jerome Luck

A Thesis

Submitted to the Graduate School of Nichigan
State College of Agriculture and Applied
Science in partial fulfilment of the
requirements for the degree of

MASTER 01' SC IENCII

Department of Phyiicl

1950

ACKNOWLEDGEM‘!‘

the writer wishes to express his sincere thanks to

Dr. 3. D. Spence. for suggesting this project. end for
his help in its execution. Also. to Mr. Charles Kingston.
who porter-ed.nuch of the work of construction. end who

supplied e nunber of suggestions releting to the mechanical

design.

mic/MA

{)71, ”(1
«7:9 ‘~- ~~.T‘r.r a8

II.

III.

IV.

CONTENTS

Introduction

Design of the Electromagnet

Construction of the Electromagnet

Operation of the Ilectromegnet

Bibl i ogrephy

Page

12

17

21

I . INTRODUCIION

The design of electronagnets follows general principles
which are widely known; the manner of application of these prin-
ciples is determined by the purpose for which the electronsgnet is
to be used. The electronagnet described in this thesis was intended
for use as a general laboratory magnet. providing fields of moderate:
intensity. in a fairly large volume.

lundamental approaches to the problen of mgnet design
have been nade at intervals spanning a considerable period of time.
lwing1 has considered the case of what is generally thought of as
the typical electronagnet. The field is produced in the space between
the ends of cylindrical pole pieces. and the magnetisation is axial.
It has been shown that. in this case, the maximum flux density attain-
able is of the order of 19.000 genes. Ihis represents the contribution
to the flux density due to the sagnetisation of the core material. so
that greater values night be realised by increases in the ugnetising
field component.

:l'abrya has shown that the magnetic field intensity produced

by an air-core coil is given by:

A
H=G§ar

l. Ewing, J .A. “Magnetic Induction in Iron and other Metals" p. 1146

2. Iabry. C. Production De Champs flagnetiques Intenses Au Ito”!
De Bobines Sans l'er. Journal De Physique 2. p. 129. 1910

where W: power dissipated

/’

resistivity of winding material

a conductor-to-winding-space ratio

a factor determined by the geometry of the coil.

A.
C}
c»

cross-sectional area of the core

Bitter3 has shown that. under the conditions of greatest
efficiency. and with an optimm (non-uniform) current distribution.
the constant '6' reaches a limiting value of 0.272. In the matter
of maximum efficiency in the utilisation of iron. it is shownII that
the direction of magnetisation should be not axial. but in such a
way that:

sinzoc : or
where 0L is the angle between the direction of magnetisation and
the coil axis. and r is the distance from the center of the coil.
This implies an equatorial sort of distribution of the iron. so
that. instead of raking up a core for the magnetizing coil. it
makes up a shell enclesing the coil. When iron is used in this tanner.
it is shown that the maximum contribution of the iron to the flux
density is of the order of 68.000 gauss. rather than the 19,000

suggested by lwing.

3. Bitter. 1'. “Design of Powerful llectrosngnets"
Review of Scientific Instruments 7 (u79) 1936

It.‘ Bitter. I. "Design of Powerful llectromagnets'
Review of Scientific Instruments 8 (318) 1937

II. DIIIGN OF THE ELECTROMAGNET

the air-core electromagnet has manifest advantages. In
addition to low cost and simplicity'of construction. it possesses
excellent stability with respect to current changes. Also. the
field intensity is directly’proportional to the current. making
determination of operating conditions easy. From the information
given above. it is a simple matter to determine the power requirements
for an air-core electromagnet having the desired characteristics.

Assuming a core area of 10 sq. cn.. a space factor of
0.75. an optimistic salue of 0.2 for G. and windings of annealed
capper. the power required. in watts. is:

V : 0.57310.3 Hz

so that. if a field intensity of 10,000 oersteds is desired. the
power required is of the order of 50 kilowatts. A.power input of
this magnitude would involve serious cooling problems. Channels for
a coolant would be required. reducing the space factor of the coil.
So it seems that this power estimate is probably'well below*what
would actually be required.

It is apparent that. although the air-core electromagnet
is appropriate to the production of intense fields in relatively
small volumes. the magnet of moderately low power and large dimensions
must depend.primarily upon the magnetization of iron for its flux
density.

If we consider rebry's equation for an air-core coil as

applying to a coil which is to nagnetise an iron core. it is apparent

that. since the field intensity varies as the square root of the
power input. the maximum power efficiency will be at the smallest
power input. Moreover. cooling problems become troublesome if the
power input is large.

A third factor which is of importance in the design of an
electrongnet is stability with respect to variation in current. A
variation in the magnetizing current should produce a minimum variation
in the flux density produced by the magnet. If possible. this condition
should be achieved without the use of complicated control equipment .
i'hese considerations all point to the use of iron in a condition of
maximum saturation. A relatively weak magnetizing field will then
suffice to produce the required flux density; the cooling problem will
be simple to solve. and. when the iron linking up the magnetic structure
is well past the knee of its ngnetization curve. the stability of the
magnet will approach that of an air-core coil.

The immediate application of the electrongnet seemed to
be in connection with research in radio-frequency and microwave
spectroscopy. Discussion with individuals working in these fields.
and reports of work in progress. indicated that a maximum available
flux density in the region of 10.000 gauss should be entirely adequate.
his value is well within the limit of 19.000 gauss stated by Ewing
for axial magnetisation. so that a conventional electromagnet design
seemed appropriate.

An important difficulty in using sagnetie fields for resonance

experiments: lies in the stringent requirement for uniformity and constancy

of the field. both in time and in space. The time variation may

be minimized by the use of iron in the saturation region. Unfortun-
ately. highly saturated pole faces produce quite non-homogeneous
fields. nomfi has shown that the homogeneity of the field may be
improved by the shaping of pole faces. The pole-face shape for
maximum homogeneity - a truncated cone with semi-vertical angle of
39°lli' - differs from that for maximum field intensity. But the
maximum intensity attainable by the use of this sort of pole-face
is still well above that required here.

Since the form of pole—face required for the greatest
ultimate usefulness was uncertain. it was decided to build the
magnet with simple flat pole—faces. to be capped at some later
date with pole pieces to meet specific requirements. In order to
provide adequate homogeneity of the field between these faces. the
diameter of the cores was made as large as conditions would permit.
A six inch diameter was chosen. this being the largest size which
could be handled and machined without great difficulty.

The utilisation of iron in an electrongnet represents a
compromise between theory and practice. The distribution suggested
by Bitter tends toward an 'ironclad' structure. Certain mechanical
difficulties attend the use of iron in this fashion. me most
important of these is the matter of access to the workim space of

the ngnet. This problem has been solved. for powerful electromaets.

5. lwing. J. t; 'Itagnetic Induction in Iron and Other Metals”
p. l

in satisfactory fashion by Dreyfuss, in the Uppsalla electromagnetse

and by Bitter. in the Arthur D. Little Electromagnet7.

In the present case. where extreme efficiency in the use
of iron is of no great importance, and where flexibility of arrange-
ment seems of the utmost importance, it was decided to use an extremely
open type of construction.

there appeared also to be a strong possibility that flux
densities much less than 10.000 gauss might be required on occasion.
So it was decided that some arrangement for producing saturation of
a large part of the magnetic circuit. at lower values of flux density.
should be provided.

The magnetic material most often used in the magnetic
structures of magnets of this type is Armco iron. a highly purified
soft iron. Althom its performance at moderate values of saturation
is excellent. it can be seen from the curves in fig. 1 that. at high
flux densities, cast steel and an 1020 steel equal and even surpass
pure iron in maxi-1m permeability. So it was decided to construct
the magnet frame. “in so far as possible. of an 1020 steel.

Determination of the characteristics of an iron-core
electrongnet under widely varying conditions is rather difficult.

Two factors are responsible for this difficulty. One is the variation
6. shout-en and Burge: I'rhe Uppsala Electromagnet" Journal of
Scientific Instruments. and of Physics in Industry
26. 10 (331) Oct. 1949

7. Described in a brochure published by Arthur D. Little, Inc. in
February. 1950

 

 

020 steel

—— pure Tact lrdn ,
steel

D

tuneoled I
not annealed

‘dIAA

 

 

5
o

 

 

: kllomsswells per sq. in.
0|
<3

 

B
O

 

 

 

 

 

 

0 IO

20

ICC 200 300

H: ampere turns per inch

MAGNETIZATION CURVES

Fige

I

Raters: Electromagnetic Devlces .

in the permeability of the core material over a range of operating
conditions. This variation is complicated in character, and cannot
be expressed analytically. The other factor is what might be termed
the Iflux leakage'. Rotors8 suggests an approximate method of deter-
mination of this factor. The basis of this method is the breaking
down of the total leakage into geometrically simple paths. The
leakage through each path is evaluated or approximated. and the total
leakage taken as a summation of the individual parts.

.A convenient way in which the flux leakage may'be expressed
is by means of.a I'leakage coefficient". This is a ratio of the overall
permeance of the air gap. including all leakage paths. to the permeance

of the useful path. The permeance of the useful path is:
2..

fl
P's-173A;
The permeance due to the fringing about the edges of the

pole pieces is:

pi e. Lea/u ("u + 3h)

The permeance due to the path between the external (cylindrical)

surfaces of the pole pieces is:
4(flz‘nl)
"3

The symbols above relate to those shown in Fig. II.

P3 "z/“r; In

8. Rotors: I'llectromagnetic Devices" p. 139

 

 

 

 

 

 

 

 

 

 

Fige

FLUX PATHS

a fourth leakage path exists between the cylindrical surface
of the core. and the outer frame. But, since the outer frame to be
used is open in character, the leakage along this path is neglected.

The leakage coefficient is:

 

P P 7' 40‘1““)
v: “*5“ =. 1.5—5- ”35.23“??-
‘ .

The effects of flux leakage may be further interpreted by computing
an 'dquivalent area' A' of the air-gap. This may be considered to
be the area which would be required to produce the total permeance,
assuming the flux to be uniformly distributed between two parallel
pole faces, separated by the actual air gap length. from this
description. it follows that the equivalent area is given by:

A‘= 1) A
where a represents the actual area of the pole face. Irom this
equivalent area may be computed the total flux required to produce

a particular flux density in the useful air gap:
I
Q) = but == VBA

Since '\) is a function of g. the air gap length, the total flux

requirement varies over a wide range, depending upon the air gap

length. rig. III shows the variation of A' with g. for the desigi used.
Once a value for the total flux has been obtained. the

required cross-sectional area of the magnetic structure may be determined

from the magnetization curves of the haterials used. In order to produce

 

soo . /

 

400

 

 

200
IOO /

 

 

 

 

 

 

etteetlve eree se.em.

mt ‘ t
elr sap length la.
EFFECTIVE AREA OF POLE FACE

L __ mom ____h._____

 

Fig. III

homogeneity of the flux between the pole faces, the material in
that locality should remain at a low state of saturation. On the
other hand. in order to produce stability with respect to current
variation. the magnetic structure in general should be completely
saturated. An.obvious method of producing this situation is to
make the pole faces of some material with a much higher value of
saturation flux density than the rest of the structure.

In order to evaluate the performance of tentative designs
under various conditions, a graphical method was adopted. .Ls a
starting point, a typical magnetization curve for an ironsand-air
structure was used. This gives the ampere—turns requirement for a
specified total flux. as a sum of the excitation required for the
iron part of the structure. plus that for the air'portion of the
structure. The total permeance of the air gap. including the leakage
paths. is used here.

Superimposed upon this is a second set of curves. represent-
ing the fluxzdensity produced for a given total flux. These are a
family of straight lines. one for each air gap length. The members
of the family vary in slope because of the variation of effective
area of the pole face for different values of gap length.

This combination of curves constitutes a useful monograph,
with which it is possible to determine, for a specified design. both
the excitation requirements for some particular performance, and the
performance which may be expected for a given excitation. The set

of curves for the design adopted is shown in fig. IT.

 

B: kllogouss~ h h

59.4? IRA/1‘s 21072

 

 

 

 

 

 

 

 

  
 

-’
v“ o.

  

 

 

 

 

 

 

-_ 25 SO 75 IOO
NI: thousands of ampere turns

CALCULATED PERFORMANCE
A: all four filler rods In place
8: no tiller rods

elr gap lengths: ------ 0.5 Inch

l.O Inch
— - — - 2.0lnch
L... Ml- .

 

 

 

Fige IV

10

l‘or example. it may be desired to find the ampere-turns
required for a flux density of 10.000 gauss. at a gap length of
1 inch. The procedure is as “follows: On the D-§ curves. follow
the l-inch line to its intersection with the 10.000 gauss line.
Then move horizontally in either direction to the intersection with
the appropriate iron mgnetisation line. from that point downward
along a line parallel with the slope for a l—inch gap ”length to
the NI scale. finding the excitation required to be some 35,000 ampere
turns. The operations may. of course. be carried on in the reverse
direction. depending upon the information supplied. A great utility
of this sort of representation is that it provides a comprehensive
and compact diagram of the limits of performance of a given design.

The problem of cooling the magnetising windings is a grave
one. in smgnets of high power dissipation. In order to preserve a
reasonable space factor. liquid cooling is generally resorted to.
later cooling presupposes an adequate supply of pure. soft water:
insulation problems nevertheless appear. in alternative is oil
cooling. This presents no insulation problems. but generally requires
the use of heat exchangers.

Where the power involved is small. air cooling seems to
be an entirely adequate method. Shawg describes an electromagnet
comparable with that discussed here. which was successfully air cooled.
9. Shaw. 1. J. “Design and Construction of an Electromagnet for

Investigation of the Magnetic Properties of items and
Molecules“. Review of Scientific Instruments. 2(611) 1931

11

Although the use of air as a coolant requires that a considerable
portion of the winding space he devoted to air passages. in a magnet
of relatively low power. the sacrifice in space factor may be more
than balanced by the increase in convenience of operation. A liquid:-
cooled magnet requires a considerable time to reach temperature
equilibrium. due to the relatively large heat capacity of the coolant.
In an air-cooled magnet. temperature equilibrium is reached quickly.

The eventual design evolved into an open H-type form. this
construction is superior to a U-shaped structure in that the forces
acting upon the poles do not cause misalignment of the pole faces.

In addition. flux leakage is minimised. The longitudinal frame
members are tubular. themselves presenting a rather small cross-
sectional area. Therefor this portion of the structure. which
makes up a considerable portion of the total path length. my be
highly saturated at relatively small values of total flux. l'or
larger values of flux. filler rods may be inserted into the tubes.
providing the necessary increased cross-sectional area.

a further advantage of the Open design is the extreme
accessibility to the working space of the magnet. Access is possible
from the tap and bottom. as well as both sides. The tubular frame
members may be utilized as supports for equipment to be used in

connection with the magnet. An assembly drawing of the magnet is

shown in Pig. V.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

12

III. CONSTRUCTION OF THE MAGNET

In order to decrease the amount of heavy machine work
required. stock sizes of standard materials were used. in so far
as was possible. Sal 1020 steel. which was used for the major portion
of the frame. is available in a large variety of sizes in round. cold-
drawn‘bars. These are drawn through dies. so that the accuracy of
dimension and the surface finish are comparable with that of a ground
cylinder. Sinkinch diameter stock of this type was used for the cores
of the magnet; these required only a facing-off Operation.

The rectangular end plates were cast of low-carbon steel.
carbon content specified less than 0.20 percent. This material
produced a clean casting. Distortion in casting was small. and
the metal was easily worked. The tubular longitudinal frame members
were made of standard-weight welded steel pipe. of h inch nominal
sise. ‘Assembly of the pipes to the frame was by means of threaded
collars. which were made from standard steel pipe couplings. The
holes in the end plates through which the pipes pass were given a
rather extreme taper - about 7 degrees. Great care was taken in
preparing the pattern. so that a hole was produced in the casting
which required little in the way of finishing. The tapers were
ground slightly. and the collars individually fitted to the holes.
using a stop nut on the inner face of the end plate. and the tapered
locking collar on the outside. a Joint was produced which appears
mechanically strong. and yet permits adJustment of the alignment of
the end plates by screwing the two collars up or down the threads on

the pipes.

13

The only major machining operation needed on the cast
steel and plates was the boring of the hole to receive the sixrinch
diameter cores. The end.plates were mounted face-to-face. in the
same orientation that would be used in the final assembly. and the
two bered simultaneously on a horizontal boring mill. A clearance
of 0.005 inch was allowed in the operation. Ewinggo states that
the reluctance of a smoothly machined Joint is equivalent to that
of an.air-gap several thousandths of an inch long. even when the
surfaces are in firm contact. So it was felt that there would be
no object in specifying an extremely small clearance.

. The internal surfaces of the welded steel pipe were feund
to hold to consistent dimension. Specifications for standard It inch
welded steel pipe state an internal diameter of n.026 inches.
However. the swaging effect of threading the ends ofOthe pipe.
reduced the inside diameter of the threaded portions by about

.015 inch. making a free but close fit on a h inch cold-drawn steel
bar. The filler rods. to be inserted into the pipes. were made up

of lengths of 3 inch cold-drawn steel bar. fitted at the ends with
four inch lengths of M inch diameter steel bar. Thus. good contact
is maintained at the ends. yet the bars can be inserted and withdrawn
easily. No difficulty'was encountered in fitting the filler rods.
except in the case of one of the four pipes. in uhich a projecting

weld was found near one end. One of the four pipes had the threads

10. Ewing. Ind. I'Magnetic Induction in Iron and Other Metals"
p. 288

IN.

on one end extended for’a length of about ten inches. This makes
it possible to manipulate the bar in such a way that the locking
collars and stop nuts may be removed. and the pipe slid out without
disassembling the entire frame. This permits easy access to the
coils. in case of some need for adjustment or servicing.

be central solid steel cores are supported at the outer
ends by the holes bored in the cast steel end plates. The inner ends
are supported in steel bushings. which were machined from a 6-inch
standard steel coupling and two short lengths of extra-strong steel
pipe. These bushings are in turn held in place by aluminum clamping
discs. which also act as end-plates for the windings. The discs
rest on Jack screws. bearing on the two lower longitudinal frame
members. The alignment of the pole faces may be adjusted by means
of these screws. The aluminum discs are clamped in the horizontal
direction by four brass rods passing back through the cast steel and
plates.

Control of the air-gap length is by means of two large
lead screws. bearing handwheels at their outer ends. and passing
through threaded plates back of the end plates. Since the force
acting upon these screws may'be of the order of five tone. the
mechanism must be quite strong. The lead screws were machined from
three-inch diameter steel bar. with a pitch diameter of two inches.
The length of thread engagement in the end plate is two inches.

It appeared that” the weakest portion of this structure would be

15

found in the screws which hold the cap on the outer end of the core.
where the lead screw is attached. Tor this reason. eight hardened
steel cap screws of the highest quality were used in each end. The
lead screw pitch is ten threads to the inch. providing a smooth
control of the positions of the pole faces. The range of adjust-
ment permits a gap length of eight inches; wide enough so that the
cores may be capped with pale pieces of special design and material.
Welded steel pipe. of 6 inch nominal sise. was used to
carry the windings. Since the inside diameter of this pipe is
about 6.060 inches. adequate clearance was provided for the sixpinch
cores. The pipes were welded to the end plates. concentric with
the cores. The weld provides sufficient strength to support the
windings. so that the windings may be placed in.pesition and
removed with no difficulty. These pipes also act as stops for.
the bushings which support the inner ends of the cores. Since
the pipes and bushings are made of low carbon steel. they also
form part of the magnetic structure. augmenting slightly the cross-
sectional areas of the cores for a large portion of their lengths.
The coils are wound in.pancake style. of .500 inch‘by
.OMO inch copper ribbon. The insulation consists of a coating of
heavy Iormex. covered with a.wrapping of Tor-ex impregnated glass
fiber. This insulation is capable of withstanding much higher
temperatures and voltages than are expected in normal operation;
the current carrying capacity of the wire is considerably in excess

of that required. The choice of this wire was determined by its

16

availability; however. the resulting rather low resistance of the
magnet reduces the power dissipation.

Each pancake consists of 70 turns; inside diameter of the
pancake 1. 7 inches. and the outer diameter 1. about 16 inches.
The winding operation.was performed under a tension of about 30
pounds. in.a rig using a variable speed motor driving a 70:1
reduction.worm gear. The combination of the worm gear on the
winding reel. and a friction.brake on the spool. made it possible
to stop winding at any'point without losing tension.

d.skeleton.type winding reel was used. making it possible
to remove each finished pancake from the winder. and fasten the
ends with no difficulty. The reel was then removed from the coil.
and winding of the next coil could be resumed while the previous
coil was being treated. Sodium silicate was originally used to
coat the pancakes. Although the results appeared good at first.
subsequent drying of the coating made the film too brittle. and
it tended to chip and peel. Clear pyroxylin lacquer was later
used as a binding material. with good results.

The inner turn of each coil was bent and flattened to
a right angle. and the inner end was brought out so that all con-
nections between coils could be made externally. A.tota1 of he
pancakes were wound. providing a total of 29h0 turns. The coils
are connected in series in two sections of 21 coils each. The
sections are brought out to separate terminals. so that they may
be used either in series or in parallel. The series resistance is

about h ohms; in parallel. the resistance is about 1 ohm.

17

The coils are mounted on the 6-inch steel pipes welded
to the end plates. The inner turns of the coils clear the surface
of the pipe by about a quarter inch; the coils are held concentric
with the pipe by means of small maple wedges spaced about the inner
diameters. The coils are held apart from one another axially. by
means of maple spacers cemented to the coils near their'peripheries.
The spaces thus produced. about one quarter inch wide. serve as
cooling passages for air. The surface available for cooling is
approximately 15,000 square inches - a highly adequate surface to
dissipate the 1% kilowatt or so involved in normal operation. without
resort to forced draft.

The weight of the electromagnet is approximately 3000 pounds.
To provide the necessary strength and rigidity. a frame of heavy
h inch steel channel was welded directly to the cast steel end
plates: this frame is proportioned so that the air gap is at a
convenient table-top height.

The weight of the magnet is distributed by the frame over
an area of actual floor contact of about MOO square inches. so that
no special provisions are needed to protect the floor surfaces.
However. a floor strength adequate to support a static load of

300 pounds per square foot is required.
IV OPERATION or THE ELECTROMAGNET

The electromagnet has been operated at several points in

its Operating range. using the motor-generator set installed for

18

general use in the Physics-Mathematics building. ‘A Fluxmeter was
used to determine the flux density in the air gap. With all four
of the filler rods removed. and an air-gap length of 1 inch.
saturation occurred in the region between 5000 and 6000 gauss.
with excitation of about 20.000 ampere turns. With all four of
the filler rods in place. saturation occurred at about 10.000
genes. with excitation of about 60.000 ampere turns. Increase of
excitation to 120.000 ampere turns produced a flux density of about
12.000 gauss. indicating that Operation in that range is quite stable.
The flux densities associated with saturation in the various
operating conditions appeared to be in good agreement with those
given by the calculated performance curves. However. the excitations
required were considerably higher than those anticipated. This is
probably due to the reluctance of the Joints in the magnetic structare.
which were not computed. The power actually required for operation
at the planned maximum flux densities is small enough so that no
new problems of supply. control. or heat dissipation have been
introduced.
Idlthough the power requirement for the maximum operation
of the electromagnet is modest. problems of power supply are encountered.
The current required for saturation. with series operation of the
coils. is of the order of 20 amperes. The motor~generator set which
is available for general use in the Physics-Mathematics building is
capable of supplying currents of several times that amount. at appro-

priate voltages. However. one difficulty lies in the fact that

19

this power supply is in fairly constant demand. and operates
at a fluctuating load. .1 second difficulty is that of control
current. In order to vary the flux density produced by the
magnet. it is necessary to vary the current supplied to the
coils. Using a remote set of generators. a large resistor
bank would be necessary to handle the total current. It was
decided that an entirely separate power supply for the magnet
would be needed. The possibilities considered were:

1. Storage batteries

2. Dryhdisc rectifiers

3. A,Motor-Generator set
The third of these possibilities was chosen. partly because of
availability. but mainly because the output of a generator is
easily controlled by varying its field excitation. In this way.
variation of the rather large magnet current may'be accomplished
by the variation of a small generator field current. In addition.
the slight ripple in.the generator output. being high in frequency.
is easily filtered.

The circuit connections for the Operation of the magnet
are shown in rig. VI. The generator is a compound.wound machine.
rated to have an output of 50 volts at NO amperes. Since the
voltage is rather low. the coil sections of the magnet are used
in.parallel. The shunt field is separately excited. probably
from the general set of storage batteries. Since the resistance

of the shunt field of the generator is about 25 ohms. the excitation

2O

requirement will be about 2 amperes at 50 volts. .A rheostat
capable of’handling this is placed in series with the shunt
winding.
Connected in series with.this shund field is the
secondary winding of a low voltage. high current transformer.
The resistance of this winding is about 1 ohm. This transformer
is for use in modulating the flux density of the electromagnet.
as is generally required for the observation of resonance phenomena.
Taps on the secondary. and a rheostat in series with the primary
winding of the transformer make it possible to vary the degree of
modulation; switches are provided to remove the transformer com-
pletely from the circuit. in cases where modulation is not required.
In order to indicate thermal equilibrium. a small ceramic
thermistor. forming one arm of a Wheatstone Bridge. is placed
within one of the air cooling passages. It is not anticipated
that heating will be a problem. since continuous operation at a
current of 20 amperes produced no perceptible rise in.the temperatures
of the coils. Currents up to ho amperes. with series Operation.were
maintained for approximately 10 minutes: only a slight increase in
temperature was noted.
At the time of writing. the power supply has not been
completed. .Lfter completion. it is planned to obtain calibration
curves for the various conditions of operation. using a nuclear

induction form of fluxmeter.

 

electromagnet

 

 

JU

 

 

 

 

    

A-

 

 

 

 

CIRCUIT

Fig. VI

DIAGRAM

 

A.O.

 

 

 

 

 

THE COMPLETED ELECTROMAGNET

l.

2.

3.

9.

BIBLIOGRAPHY

Bitter. I. “The Design of Powerful Electromagnets“
Review of Scientific Instruments. 1k p. R79: p.h82 1936
i. P- 318: 1937

Ewing. J. A. "Magnetic Induction in Iron and Other Metals“
D. VanNostrand 00.. New York. 3rd.Ed.. 1900

Fabry. 0. “Production de Champs Magnetiques Intenses au.Moyen
de Bobines Sans FerI Journal de Physique'g. p. 129; 1910

Kapitza. P. EA Method of Producing Strong Magnetic Iields'
Proc. Royal Society of London 1_05_. p. 691; 1921;

Hepkins. N.J. "Magnetic Field Strength.Meter Using the Proton

Magnetic Moment“ Review of Scientific Instruments g9.
p. ”01; 19h9

Boters. 3.0. “Electromagnetic Devices”
John Wiley and Sons. Inc. New York. 19hl

Shaw. 3. J. ”Design and Construction of an Electromagnet for
Investigation of the Magnetic PrOperties of.1toms and
Molecules“ Review of Scientific Instruments 1g. p.611;.
1931 .

Snellman. 0. and Burge. E.J. "The Uppsala Electromagnet"
Journal of Scientific Instruments and of Physics in

Industry 2g, p.331: 19h9

Wall. T. 1. “Applied Magnetism" D. Van Nostrand 00..
New York 1927.

HICHIGQN STQTE UNIV. LIBRQRIES

31293017721303