A NEW HOMOPOLAR MOTOR

Y. H. KU’ AND AHMAD KAMAL’

SYNOPSIS

A new homopolar motor has been designed and constructed based on the principle
of the Faraday disk. The magnetic circuit consists of two halves made from Armco
iron. The rotor disk is made from a low carbon steel plate. Current is fed to the disk
radially through 8 brushes of copper graphite, while 8 other brushes on both sides of
the shaft form the output terminal. With 2 volts input voltage and 100 amperes
input current, the motor delivers approximately l/6 h.p. at 5800 r-pm. with 50 per cent
efficiency. Performance curves and a transient response oscillogram are presented.

INTRODUCTION

Since Faraday’s time, the principle of the homopolar generator has

been well known. The Faraday disk is sometimes called the Lorenz
apparatus, as Lorenz used it for the absolute determination of the ohm
(1).2 As Liwschitz remarked (2) : “Theoretically it is possible to pro-
duce direct current directly without resorting to the use of a commuta-
tor.” Practically, homopolar generators have been built and used
recently as tachometers (3).
During World War II, there was an attempt in Germany to make a
homopolar motor, and a fragmentary description was given by Boning
(4). The principle of this design was explained by Professor G. W. 0.
Howe (5) in November of 1952.
It is the purpose of this paper to describe the design and performance
of a new homopolar motor. Except for the Boning and Howe refer-
ences, there are no recent reports on homopolar motors.

THE HOMOPOLAR MOTOR DESCRIBED BY BiiNING

From ref. 4, we find that this design of a homopolar motor as de-

scribed by Boning consists simply of two iron halves A and B (Fig. 1)
forming the magnetic circuit and at the same time the moving part.
The energizing coil is made in the form of a spiral which is fixed and
thus forms the stationary part. The current is fed from one end as
shown by the arrow, transmitted through the contactor D to the iron,
and collected at E, which also forms the output terminal. The explana-
tion of the principle of this motor is given in ref. 5.

1Moore School of Electrical Engineering, University of Pennsylvania, Philadelphia, Pa.

* The boldface numbers in parentheses refer to the references appended to this paper.
 Y. H. Ku AND AHMAD KAM.41, [J. F. I.

FIG. 1. Homopolar motor described by B6ning.

DESIGN OF THE NEW HOMOPOLAR MOTOR

A number of altcrnativ-c designs have been made for a new homo-

polar motor. Since the moving part of the motor described by Boning
consists of the magnetic circuit, there is difficulty in running at high
speed with good efficiency. We have kept the magnetic circuit station-
ary and used Armco iron to get a flux density of 20,000 gauss (see Fig.
10). The rotor is in the form of a Faraday disk made from a low carbon
steel plate of i-in. thickness and annealed.
The principle of the Faraday disk as a generator is well known. Let
the flux lines perpendicular to the disk be represented by 9 and the flux

FIG. 2. Faraday disk as a hotnopolar motor.

July, 1954.1 A NEW HOMOPOLARMOTOR 9

density be denoted by B. By the relation E = Blv, we have, for Al =

Ax in Fig. 2, and v = wx,

AE=BvAx=Btix(Ax) (1)

where v = peripheral velocity at a point x from the center of the disk,

w = angular velocity of the disk = 2702, and n = rps.
The total electromotive force between the axis and the periphery of
a disk with radius R is then3

E=Jo B wxdx = Bw xdx k $BoR?

J0 (2)

for constant flux density and constant speed of the disk. Since the area
of the disk is equal to rR2, and w = 2?r (rps.) = 2nn,

E=qin (3)

where 4 is in maxwells, n in revolutions per second, and E in abvolts.

For E in volts, the right hand side of Eq. 3 has to be multiplied by a
factor lo-*.
Instead of supplying mechanical power to the disk and making it
rotate at a speed n, we can supply a current I between the periphery
and the axis of the disk. By the relation F = BZI, we have, for AZ =
Ax in Fig. 2,
AF=BIAx (4)

where F = force in dynes, B = flux density in gausses, I = current in

abamperes, and x = distance in centimeters.
The torque is given by T = F x, or

AT = B Ix (Ax). (5)
Integrating, for constant B and constant I,

T = lRBIxdx = BILRxdx = $BIR* (6)

or
T = cr
2?r (7)

where T = torque in dyne-centimeters.

3Strictly speaking, the lower limit of the integral should be I, the radius of the shaft of the
disk or rotor.
IO Y. H. Ku AND AHMAD KAMAL [J. F. I.

If I is expressed in amperes, 4 in maxwells, and T in ounce-inches,

Eq. 7 becomes,
T = 2.255 X 10e7 (b I oz. in. (8)
In the present case the actual disk diameter was made 4.22 in. or
10.72 cm. For the radius R equal to 5.36 cm., the area of the disk is
90 sq. cm. Assuming an effective area of 80 sq. cm., and a flux density
of 20,000 gausses (lines per sq. cm.), the total flux C$is 1,600,OOOmax-
wells or 1.6 megalines. The current supplied to the disk was limited to
100 amperes. So by Eq. 8,

T = 2.255 X 10-T (1.6 X 106) X 100

= 36 oz. in.

For a homopolar motor, the relation I;: = C#I

n given by Eq. 3 may be
considered as representing the back emf. So, for 4 = 1.6 X 106 and
n = 100 rps. (assumed), E = 1.6 X lo8 abvolts or E = 1.6 volts.4
Since the applied voltage to the disk or the rotor must be greater than
the back emf. by the IR drop, where R represents the combined resist-
ances of the rotor and the brushes, V, the input or rotor voltage, will
be around 2 volts.
The order of magnitude of the disk resistance can be evaluated as
follows :
R

RD= .k?$)= l-
2?rb
Inx
v
(9)

where b = thickness of the disk = 0.122 in. = 0.31 cm., R = radius of

the disk = 5.36 cm., r = radius of the shaft of the disk = &r in. =
0.795 cm., and p = resistivity of the steel = 17.4 microhm-cm. (as-
sumed). Substituting these numerical values in Eq. 9, we get,

R = 17.4 x 10-G In 5.36

D
27r (0.31) 0.795
= 8.95 X 10m6(1.9066) = 1.71 X 10-B ohm.

This value is negligible relative to R B, the resistance of the brushes.

DESCRIPTIONOF THE ROMOPOLAR MOTOR

The assembly drawing of the motor is shown in Fig. 3, with details

of the magnetic circuit shown in Fig. 4, and details of the rotor (disk)
shown in Fig. 5. The actual photograph of the motor is shown in Fig.
6. Two views of the disassembled motor are shown in Fig. 7.
The magnetic circuit consists of two halves made of Armco iron (see
Figs. 3 and 4). The iron has a carbon content of 0.015 per cent, silicon
0.003 per cent, and total impurities 0.07 per cent. It has a saturation
4This is equivalent to E = 16 millivolts/rps.
July, 1954.1 A NEW HOMOPOLAR MOTOR II

ARMCO IRON

“X” 1 "Y"

SECTION ‘Y’-“T” SECTION ‘X”-“X’

FIG. 3. Design of a new homopolar motor.

“Y”
“X.-t
r’ I

0.37

” Y”
HALF
SECTION’X”-‘X”

FIG. 4. Details of the magnetic circuit.

12 Y. H. Ku AND AHMAD KAMAL [J. F. I.

flux density of 21,000 lines per sq. cm. It was annealed by the manu-
facturer before being machined.
The rotor is in the form of a disk (see Figs. 3 and 5) made from a low
carbon steel plate of g-in. thickness, and annealed. The disk was
shrunk to a $-in. shaft, and both the shaft and the disk were ground to
the final size. AAsshown in Fig. 5, the disk was finally ground to a

FIG. 5. Details of the rotor.

FIG. 6. General view of the motor.

thickness of 0.122 in., a little less than g in. The shaft rotates through
two bearings imbedded in the two halves of the magnetic circuit. The
air gap is 0.010 in.
The current was fed to the disk radially through 8 brushes of copper
graphite (86 per cent Cu) which are specified to have very low coefficient
of friction, extremely low contact drop and very low specific resistance.
Beryllium copper springs were used at a pressure of 2 psi. .4s shown in
July, 1954.1 A NEW HOMOPOLAR MOTOR I3

the right side of Fig. 3, these 8 input terminal brushes are connected to-
gether by a copper strip. There are 8 other identical brushes arranged
on both sides of the shaft and connected also in parallel to form the
output terminal.
The field is established by two coils imbedded in the circular space
made in the two halves of the magnetic circuit (see Fig. 3). The coils
were made of 916 turns of A.W.G. 18 double glass insulated wire and
have a total resistance of 8.5 ohms at 20” C. Theoretically it was
estimated that with 6 amperes field excitation we can get about 1.4 X
lo6 flux lines to pass through the disk. Experimentally, we find that
with 5 amperes field excitation, we can get 4 equal to more than 3.6 X

FIG. 7. Motor disassembled.

lo6 flux lines corresponding to a flux density B of more than 20,000

gausses (lines per sq. cm.). The experimental magnetization curve is
shown in Fig. 10.
The total weight of the machine is about 75 lb.
DIAGRAM OF EXPERIMENTALCONNECTIONS

The diagram of connections for the experiment is shown in Fig. 8.

A photograph of the setup is shown in Fig. 9.
In Fig. 8, the field or stator circuit is supplied by a 110~. d-c. source.
For & = 8.5 ohms, a voltage Vr equal to 51 volts will give 1, equal to 6
amperes. The rotor or disk voltage Ir, is maintained at 2 volts, ob-
tained from an 18~. d-c. source.
The homopolar motor was connected to an electric dynamometer to
measure the output torque.
The speed of the motor was measured by mounting a small aluminum
I4 Y. H. Ku AND AHMAD KAMAL [J.F. I.

wheel on the shaft on which two small magnets were fixed diametrically
opposite as shown in Fig. 8. A small pickup coil was fixed to the frame.
The rotation of the two magnets induced an emf. in the coil which was
connected to the horizontal input of an oscillograph. ;i voltage wave
from an oscillator was applied to the vertical input of the oscillograph.
The frequency of the calibrated oscillator was adjusted until a stationary
pattern was obtained. Thus the frequency of the rotation of the motor
was determined.

IELD CIRCUIT

ELECTRIC
DYNAMOMETER

ROTOR CIRCUIT

OSCILLATOR
FIG. 8. Diagram of connections.

FIG. 9. Experimental setup.

July, 1954.1 A NEW HOMOPOLARMOTOR I5
PERFORMANCE CURVES

The magnetization curve (Fig. 10) was first obtained by driving the
machine as a generator at constant speed. The output voltage was
measured by a high impedance vacuum tube voltmeter for various field
excitation currents. The flux was then calculated. The test was re-
peated at different speeds as a check. From Fig. 10, it is seen that for
I, = 4 amperes, the flux through the disk is 1.6 megalines. For a homo-
polar generator, since E = n 4 10e8 volts, for n = 1 rps., E = 16 milli-
volts/rps. The homopolar generator reported in ref. 3 gives a calcu-
lated value of 1.0 millivolt/rps. and a measured value of 1.7 mv/rps.

0 I 2 3 4 5 6
FIELD CURRENT (AMPERES)
FIG. 10. Magnetization curve.

The saturated flux density is given as between 10,000 and 11,000 gausses.
By doubling the flux density, a modified design would give E = 3.4
mv/rps. Since the total flux is proportional to the area of the disk,
doubling the radius of the disk would increase the generated voltage by
4 times, giving E = 13.6 mv/rps., which is comparable to our value
E = 16 mv/rps.
As torque depends on the product of flux C#I
and the current I through
16 U. H. Ku AND AHMAD KAMAL [J.F. I.

the rotor or I, according to Eq. 7, the relation between torque and the
rotor current is a straight line, for a given value of 4, depending on I,.
Four such lines are plotted in Fig. 11. The torque varies from about
10 oz. in. to 30 oz. in. as the rotor current increases from 40 amp. to
100 amp.
The torque-speed characteristic curves are given in Fig. 12. They
are similar to the torque-speed curves of a d-c. shunt machine. For
constant I, and hence constant 4, the back emf. E is proportional to the
speed n, according to Eq. 3. Now, for constant 4, the torque is pro-

T
:U

1
30- -i

0 ,’
,’
,<

IO ,'
<
&
/ Z

0,
30 ' i0 60 70 60 90 1
ROTOR CURRENT (AMPERES)
FIG. 11. Torque vs. rotor current I,.

portional to the rotor current I,, which is equal to (V-,?:);I?, where 1’

represents the rotor voltage, or Ir,, and R the combined disk and brush
resistances. Except for possible variations of the value of I?, at differ-
ent speeds, the torque-speed characteristics are essentially straight lines
as shown.
For constant 4, the speed n is proportional to the back emf. E, which
is equal to (V-IR). So, at no load, when the IR drop is small, n is
practically proportional to the rotor voltage V,. The relation between
speed and rotor voltage is shown in Fig. 13.
July, 1954.1 A NEW HOMOPOLAMMOTOR I7

(H3NI 33NnO) 3f-tbt101

18 Y. H. Ku AND AHMAD KAMAL [J. F. I.

In order to find the relation between the rotor voltage and the rotor
current at different temperatures, experimental results are plotted in
Fig. 14. The solid line curves (A) and (B) show the V-lcharacteristic
of the rotor at 65” F. and 150” F., respectively. The dash and dotted
line curve shows the same relation at 150” F. with half of the brushes
removed. It is seen that with all the brushes in, the rotor voltage drop
is between 0.4 and 0.45 volt, for I, = 100 amp. For I, = 42 amp.,
Vr = 0.34, and hence the resistance R = 0.008 ohm. At 150” F., the
resistance varies from 0.012 to 0.004 ohm. Lqt 6.5” F., the resistance
varies from 0.010 to 0.0045 ohm. Under actual running conditions, the
brush resistance should be higher.

G
3 0.5
z
- 0.4
$
2 0.3
G
0.2
---HALF BRUSHES REMOVED

0.1

0
0 IO 20 30 40 50 60 70 80 90 100
ROTOR CURRENT (AMPERES)
FIG. 14. Voltage-ampere characteristic of rotor and brushes.

OUTPUTAND EFFICIENCY

The measured output in fractional horsepower is plotted against the

rotor current 1, in Fig. 15 as a straight line. For I, = 70 amp., the out-
put is 0.10 h.p. For I, = 100 amp., the output is 0.15 (or about 6) h.p.
The efficiency curves for various values of I, are also plotted in
Fig. 15, with excitations corresponding to I, = 1 amp., 2 amp., 3 amp.,
or 4 amp. As the total input includes the field input and the rotor in-
put, the efficiency curve corresponding to I, = 1 amp. is found to be
higher than the other curves corresponding to I, = 2 amp., 3 amp., or
4 amp.
Taking the D curve for example, let I, = 50 amp. For 1, = 1 amp.,
the field input is equal to VJ, = I,*R, = (1)2 (10.5) = 10.5 watts,
since Rf at running temperature is 10.5 ohms. Since the rotor voltage is
July, 1954.1 A NEW HOMOPOLARMOTOR 19

maintained at 2 volts, the rotor input is equal to VJ, = (2) (SO) = 100
watts. The total input is therefore (100 + 10.5) = 110.5 watts.
From Fig. 11, we find that for If = 1 amp. and I, = 50 amp., the torque
is equal to 11.7 oz. in. From Fig. 12, the corresponding speed is found
to be 6060 rpm. The power output is given by

2?m T 2d6060) (11.7) = 0 0704

H.P. =
12 x 16 x 33,000 = 12(16) (33,000) ’

where n is in rpm. and T in oz. in. This corresponds to an output of

52.5 watts. The efficiency is therefore equal to 47.5 per cent.

30 40 60 60 70 60 90 100
ROTOR CURRENT (AMPERES)
FIG. 15. Output and efficiency curves.

Take I, = 100 amp. and If = 1 amp. The field input is 10.5 watts.
The rotor input is now increased to 200 watts. The total input is then
210.5 watts. The torque corresponding to I, = 100 amp. and 1, = 1
amp. is found from Fig. 11 as 26.4 oz. in. From Fig. 12, the correspond-
ing speed is 5800 rpm. So the power output is 0.1515 h.p. or 113 watts.
The efficiency is therefore equal to 53.7 per cent.
 TRANSIENT RESPONSE

The oscillogram for transient response is shown in Fig. 16. The

first (top) curve shows the variation of speed with time. The second
curve shows the transient rotor current. The third curve shows a steady

FIG. 16. Oscillogram for transient response.

field current of 3 amp. The next two horizontal lines are the zero lines
of rotor current and field current, respectively. The next curve shows
the transient rotor voltage with the zero line below (bottom curve).

Acknowledgment
The authors are indebted to Drs. J. G. Brainerd and C. N. Weygandt
for advice and encouragement. Many thanks are ‘due?0 Mr. Miles E.
Nelson and Mr. John H. Pedley for their help inlmaking the model
machine at the Moore School Workshop. Great appreciation is ex-
tended to the Stackpole Carbon Company and the Flower Manufactur-
ing Company for supplying brushes and holders.
EEFEEENCES
(1) G. P. HARNWIXLL,“Principles of Electricity and Electra-Magnetism,” New York, McGraw-
Hill Book Co., Inc., 1949, pp. 348, 429.
(2) M. LIWSCHITZ-GARIK, “Electric Machinery,” Vol. I, New York, D. \,.a11Nostrand Co.,
Inc., 1946, p. 77.
(3) C. C. JOHNSON,“A Homopolar Tachometer,” I.R.E. Proceedings, February 1952, pp. 158-
60.
(4) P. BONING, “Homopolar Motor,” E.T.Z., February 1, 1952, p. 76.
(5) G. W. 0. HOWE, “A New Motor Desigu,” Wireless Engineering, November 1952, pp. l-3.
(6) S. W. GLASS, “The Measurement of the Frictional Characteristics of Brushes,” Iron and
.St~cl Engineer, July, 1937, p. 39.