| Homopolar Generator
The homopolar generator would be an ideal high current, low voltage
the production of Browns gas or hydrogen to run an engine or do welding
Released January 6, 1988
DePalma Energy Corporation,
1060 Channel Drive,
Santa Barbara, CA 93108
Two years ago the DePalma Energy Corporation initiated the construction
large N machine which was scaled from smaller machines which were used
verify the validity of direct extraction of electrical energy from space
The N-1 machine was to produce a high enough output voltage so as to
efficiently convertible to high voltage alternating current. Target
parameters for the machine were 12 v.d .c output at 10,000 amperes for
r.p.m. rotational speed.
The rotor is a key component of the N-1 generator, and into it goes
the time and effort in constructing the machine. Figure (1) shows the
completed rotor suspended from a crane. Two N-machines are connected
electrically in series on a common shaft. Magnetic polarity is NS-SN
rotation of the unit causes the voltages to be additive.
Electrical connections are made to the outer edges via liquid metal,
mercury wetted contacts as shown in figure (2). Some of the considerations
incorporated into the design are: The idea of series, two machines on
same shaft, is a good one because with magnetic polarity indicated only
identical mercury contacts are needed.
The drawback is the necessity of separating the two N machines far
each other so as to reduce the demagnetizing effect of one on the other.
further simplification is the use of only two support bearings for two
machines, providing the center section can be made stiff enough to place
vibrational problems above the planned operating speed of the machine.
The 9-1/2 inch separation of the two magnet stacks resulted in a 20%
in the field strength of the combination. Initial computer simulations
field strength inside 8-3/8 inch thick, 13 inch diameter magnet stacks
NdFeB magnet material indicated 9000 gauss. Actual measurements gave
of the expected figure, 6750 gauss.
Together with the 20% loss due to the proximity of the magnets the
field strength is 5400 gauss in the center of the stacks. We have found
relaxation in the strength of NdFeB magnets in open flux path operation
when the diameter to thickness ratio is what would normally considered
a safe 2:1.
The achievement of a field strength of 6750 gauss uniformly distributed
diameter of a 13 inch circle with permanent magnets without a closed
is impressive considering the size of the water cooled electromagnet
produce the same result. Nevertheless it was felt that the flux inside
machine could be increased by making the machine partially self-exciting.
This was accomplished by subdividing the 14 inch diameter by 3/8 inch
copper disc in the center of each magnet stack into two parallel windings
two turns each. This is done in the manner of Tesla, (Reference 1).
Since the machine is operating on the slope of the demagnetization
the magnets a small increment of magnetic induction will result in a
disproportionate increase of the magnetic flux of the magnets. With
sufficient current drawn from the machine the output impedance will
and may become negative.
At present the measured internal resistance of the machine is 75 micro-ohms
800 amperes and the voltage output is 1.216 volts per 1000 r.p.m. Mercury
contacts must be amalgamated by hand before assembly and seem to improve
running time. Mercury builds a surface layer of oxide in contact with
atmospheric oxygen. Oxide sludge can be eliminated by operating the
in an inert helium atmosphere and hermetic sealing of the unit.
The N-1 generator incorporates both of these features which also prevent
leakage of mercury or mercury vapor into the environment. The rotor
constructed on a BeCu shaft 2 inches in diameter and 49 inches long.
used for strength and good electrical conductivity.
The center section between the magnet stacks is stiffened by a 6 inch
1/2 inch wall aluminum cylinder which encloses the central axle of 2
diameter BeCu. The two central shafts are screwed together in a coupling
which when the magnet stacks are rotated one relative to the other,
and places the aluminum cylinder in a state of compression. The stressed
center section is stiffer than if the connection was a solid 6 inch
cylinder. All components are anodized and teflon coated so that electrical
conduction takes place along the designed path.
The magnet stacks themselves are constructed of 200 pieces each made
NdFeB hexagons of dimension 2 inches (across the flats) by 1 inch thick,
magnetized. Each magnet weighs about 1 pound. The magnets which repel
other intensely are assembled and glued together with proprietary adhesives.
The assembled magnet stack is wound with 1/4 inch of graphite fiber
followed by 1/4 inch of epoxy-fibreglass. As constructed the rotor should
capable of 10,000 r.p.m. without damage. Further development would result
a rotor capable of standing 20,000 r.p.m. without flying apart. The
contacts operate totally satisfactorily but are not entirely leakproof.
A small amount of mercury is lost in the operation of the machine which
amounts to about 1/2 c.c. per minute. A continuous flow system can be
provided to recycle and clean the mercury.
As can be seen from figure (2) the edge of the copper disc protrudes
distance slightly less than 1/2 inch into a slightly larger channel
cut into a
copper bussbar which encircles the magnet stack. The actual point of
with the mercury is made at a radius determined by exploration with
effect magnetometer. There is a point, somewhat above the surface of
magnet stack where the fringing field produced by the presence of the
disc is exactly canceled by the flux linkages proceeding over the outer
surface of magnet from the north to south poles. Thus the mercury sliding
contact takes place in a region of zero flux.
The presence of strong magnetic flux around the rotor, and its considerable
weight, 800 lbs., make it a very difficult piece to handle mechanically.
movement around a machine shop for the various operations, every motion
be thought out carefully. After the work is finished all the tools have
demagnetized, including large lathes and milling machines.
These difficulties have to be counterbalanced by the easy accessibility
liquid metal sliding contacts and the higher voltages obtainable from
large radius magnet stacks. Since voltage increases as the square of
machine radius, the loss in flux from non-closure of the flux path is
than compensated by the increased voltage output. i.e., if a 6 inch
central axle in a closed path machine were operated at say 2 x 6750
13.5 KG then the voltage obtained would be only 1/2 of what would be
with half the flux (6750 gauss) at twice the radius (12 inches).
The other overriding consideration is that no closed flux path is possible
a machine operating with super-conducting magnets since the attainable
magnetic strength, 90 KG, will saturate known magnetic conductors. Figure
shows the partially assembled machine, figure (4) shows the completed
machine under test.
A loadbank capable of absorbing 100 KW @12 v.d.c. is in the background.
are preliminary for two reasons. Firstly we are operating at only a
fraction of power output for which the machine is capable, and second
only possible to estimate the electro-mechanical efficiency of the drive
A second more definitive round of testing will take place when we have
installed an in-line direct torque sensor between the drive motor and
generator. The exact comparisons can be made between direct horsepower
mechanical input and the electrical output.
Electrical testing of motors and generators is a carefully specified
and has been followed in accord with well established engineering practice.
The results will be presented here.
Reference to the testing of d.c. machines: "Principles of Direct-Current
Machines", Alexander S. Langsdorf, M.M.E., 5th ed., McGraw-Hill
Co., Inc., 1940.
The drive motor used for our experiments is a d.c. machine originally
supply a constant 30 v.d.c. at up to 500 amperes in aircraft service
Operating speed was 4000-8000 r.p.m. and output voltage was regulated
control of the field excitation, nominally 24 v.d.c. @ 12 amps.
The machine has six poles and six interpoles excited by a combination
armature and field currents. The voltage picked off the commutator by
sets of brushes passes through the interpole winding which adds 1/2
drop for each 100 amperes of armature current. The interpoles are used
prevent commutation losses in electrical machines operated at high speed.
The presence of the interpoles reduces the size of main poles by about
with a consequent loss in efficiency. The complete subject is dealt
with in a
Langsdorf's book. In addition to the IR drop in the interpole winding
is a one volt drop assigned to each carbon brush in accord with AIEE
Electromechanical efficiency of d.c. machines studied hits a maximum
85% in the center of the operating range. Efficiency falls rapidly at
speeds and decreases much less rapidly at speeds above the maximum efficiency
point. (See Langsdorf p. 525)
The average of five runs on January 5, 1988 are presented. Three of
were at no load condition and two were under load. Operating speed for
tests was exactly 2600 rpm. Measurements were very consistent differing
from run to run.
machine speed : 2600 r.p.m.
internal resistance @ 800
amps : 75 micro-ohms
load resistance : .003875 ohms
volts/1000 r.p.m.: 1.216
Derived calculations, two loaded runs, three unloaded runs.
motor power input loaded: 5030 watts
unloaded : 2383
N-1 electrical output : 2480
(1) straight electrical efficiency (no corrections) = 93.8%
(2) electrical efficiency corr. for brush and interpole winding drop
(3) electrical eff. corrected assuming motor electromagnetic efficiency
(4) efficiency assuming motor eff. of .4 = 271%
[note: What is being measured is electrical efficiency. Examination
figures shows loaded input is supplying electrical output and no-load
mechanical losses (expressed electrically). On this basis a corrected
loaded input (case 2) of 4440 watts expresses a no-load mechanical loss
of 2155 watts and an electrical output of 2480 watts (simultaneously).
This gives an overall energy balance of 4440 watts input giving rise
watts total output (sum of elec. and mech.). This expresses an overall
efficiency for the N-1 unit of 104.5%. Without corrections for motor
efficiency 195 extra watts have been extracted from space.]
Set of assumptions (4) seems most likely since this figure falls into
with the average of power ratios reported for N machines by Trombly
Tewari (1.75 - 2.50), DePalma (original Sunburst machine 2.5 - 3), Kincheloe
(improved Sunburst 4.9).
For the calculations made on the DePalma Energy N-1 unit for the five
averaged the loaded drive motor current ranged from 194 to 197 amps.
motor voltage ranged from 25.5 to 26 volts respectively.
Unloaded drive motor current ranged from 89.5 to 96 amps voltage from
25.8 volts. Loaded current output was 800 amps. @ 3.1 v.d.c.
Unloaded d.c. output voltage ranged from 3.1 to 3.22, avg. 3.16. All
taken at constant speed of 2600 r.p.m.
Change in loading on the drive motor only changes armature current,
voltage remains essentially constant.
The next phase of measurements will refine the efficiency determination
direct measurement of generator driving torque x r.p.m. from an in-line
torque sensor. This installation will take some time however it was
useful to put out an initial report since the reported results are consistent
with the findings of other investigators.
Many workers are attempting to construct a self-running motor generator
combination operating on the principle of direct extraction of electrical
energy from space. The presently used combination uses a Faraday disc
excited from an N machine mounted on a common shaft.
Even if the electro-mechanical efficiency of the Faraday disc is state
art 96% the losses in the system may make it difficult to get the loop
over 100%. N machine output increases as the square of the voltage or
but mechanical losses consist of constant factors which are speed invariant
and mechanical N machine output must be high enough to overshadow all
constant and speed variable losses.
Examination and digestion of the parameters and figures shows it may
necessary to have electrical output of the N machine in the 30 - 40
range before a Faraday disc N machine combination could self-sustain.
machine such as we are using for a drive motor has too many losses to
considered for a self-running combination.
The self-running combination will probably require a drive motor with
back e.m.f. compared to the Faraday disc with the same mechanical output.
some years I have been advocating the use of a motor with such
characteristics. This is covered in reference (2).
A permanent magnet version of this machine has been built which operates
shows interesting characteristics. Back e.m.f. is generated in the Boning
motor from the N effect voltage generated in the central axle which
magnetized and rotating. This voltage subtracts from the applied voltage
which is causing the motor to rotate.
As can be appreciated from the geometry of the situation this back
be made 25% of that which would exist if a Faraday disc were used to
diameter as the Boning spiral in an inside out Faraday disc motor. It
clear that if a principle of energy extraction is operative as an electrical
generator, that same principle could be applied in the form of an electrical
motor with reduced back e.m.f.
The object to develop an N machine electrical generator of high power
capability and useful voltage output has been realized with the N-1
Further measurements will refine the understanding of this. Exact torque
measurement will remove uncertainties from the exact power gain of this
Future work on the rotor will attempt to achieve high dynamic balance
unit can be run at a designed speed without excessive vibration. Research
the self-running machine will continue with the study and further construction
of the Boning motor. As work continues additional reports will be issued.
The N-1 rotor unit is a manufactured item and is available from the
Bruce E. DePalma,
6 January 1988,
Santa Barbara, California
(1) Nikola Tesla; The Electrical Engineer, N.Y., Sept 2, 1891. Published
Nikola Tesla Museum, Beograd, Yugoslavia, 1956.
(2) Wireless Engineer, November 1952, Vol. 29, No. 350.
(3) by Paramahamsa Tewari, Dept. of Atomic Energy, Nuclear Power Board,
Bombay, India. A high quality videotape documenting the experiments
described in the paper will be available from DePalma Energy Corporation.
Format NTSC or PAL