Return to the INE Main Page

Date: Thu, 11 Apr 1996 17:05:15 -0700
Mime-Version: 1.0
To: (INE)
From: (Evan Soule)
Subject: A Report


By Roger Hastings PhD

Transcribed By George W. Dahlberg P.E.

[Reformatted and spell checked by Patrick Bailey, INE, 16 Apr 96]

I do not intend to recapitulate the theory presented in Newman's book, but rather to briefly provide my interpretation of his ideas. Newman began studying electricity and magnetism in the mid 1960's. He has a mechanical background, and was looking for a mechanical description of electromagnetic fields. That is, he assumed that there must be a mechanical interaction between, for example, two magnets. He could not find such a description in any book, and decided that he would have to provide his own explanation. He came to the conclusion that if electromagnetic fields consisted of tiny spinning particles moving at the speed of light along the field lines, then he could explain all standard electromagnetic phenomena through the interaction of spinning particles. Since the spinning particles interact in the same way as gyroscopes, he called the particles gyroscopic particles. In my opinion, such spinning particles do provide a qualitative description of electromagnetic phenomena, and his model is useful in understanding complex electrical situations (note that without a pictorial model one must rely solely upon mathematical equations which can become extremely complex).

Given that electromagnetic fields consist of matter in motion, or kinetic energy, Joe decided that it should be possible to tap this kinetic energy. He likes to say: "How long did man sit next to a stream before he invented the paddle wheel?" Joe built a variety of unusual devices to tap the kinetic energy in electromagnetic fields before he arrived at his present motor design. He likes to point out that both Maxwell and Faraday, the pioneers of electromagnetism, believed that the fields consisted of matter in motion. This is stated in no uncertain terms in Maxwell's book "A Dynamical Theory of the Electromagnetic Field". In fact, Maxwell used a dynamical model to derive his famous equations. This fact has all but been lost in current books on electromagnetic theory. The quantity which Maxwell called "electromagnetic momentum" is now referred to as the "vector potential".

Going further, Joe realized that when a magnetic field is created, its gyroscopic particles must come from the atoms of the materials which created the field. Thus he decided that all matter must consist of the same gyroscopic particles. For example, when a voltage is applied to a wire, Newman pictures gyroscopic particles (which I will call gyrotons for short) moving down the wire at the speed of light. These gyrotons line up the electrons in the wire. The electrons themselves consist of a swirling mass of gyrotrons, and their matter fields combine when lined up to form the magnetic lines of force circulating around the wire. In this process, the wire has literally lost some of its mass to the magnetic field, and this is accounted for by Einstein's equation of energy equals mass times the square of the speed of light. According to Einstein, every conversion of energy involves a corresponding conversion of matter. According to Newman, this may be interpreted as an exchange of gyrotrons. For example, if two atoms combine to give off light, the atoms would weight slightly less after the reaction than before. According to Newman, the atoms have combined and given off some of their gyrotrons in the form of light. Thus Einstein's equation is interpreted as a matter of counting gyrotrons. These particles cannot be created or destroyed in Newman's theory, and they always move at the speed of light.

My interpretation of Newman's original idea for his motor is as follows. As a thought experiment, suppose one made a coil consisting of 186,000 miles of wire. An electrical field would require one second to travel the length of the wire, or in Newman's language, it would take one second for gyrotons inserted at one end of the wire to reach the other end. Now suppose that the polarity of the applied voltage was switched before the one second has elapsed, and this polarity switching was repeated with a period less than one second. Gyrotons would become trapped in the wire, as their number increased, so would the alignment of electrons and the number of gyrotons in the magnetic field increase. The intensified magnetic field could be used to do work on an external magnet, while the input current to the coil would be small or non-existent. Newman's motors contain up to 55 miles of wire, and the voltage is rapidly switched as the magnet rotates. He elaborates upon his theory in his book, and uses it to interpret a variety of physical phenomena.


In May of 1985 Joe Newman demonstrated his most recent motor prototype in Washington, D.C.. The motor consisted of a large coil wound as a solenoid, with a large magnet rotating within the bore of the solenoid. Power was supplied by a bank of six volt lantern batteries. The battery voltage was switched to the coil through a commutator mounted on the shaft of the rotating magnet. The commutator switched the polarity of the voltage across the coil each half cycle to keep a positive torque on the rotating magnet. In addition, the commutator was designed to break and remake the voltage contact about 30 times per cycle. Thus the voltage to the coil was pulsed. The speed of the magnet rotation was adjusted by covering up portions of the commutator so that pulsed voltage was applied for a fraction of a cycle. Two speeds were demonstrated: 12 R.P.M. for which 12 pulses occurred each revolution; and 120 rpm for which all commutator segments were firing. The slower speed was used to provide clear oscilloscope pictures of currents and voltages. The fast speed was used to demonstrate the potential power of the motor. Energy outputs consisted of incandescent bulbs in series with the batteries, fluorescent tubes across the coil, and a fan powered by a belt attached to the shaft of the rotor. Relevant motor parameters are given below:

       Coil weight    :  9000 lbs.
       Coil length    :  55 miles of copper wire
       Coil Inductance:  1,100 Henries measured by observing the current
                         rise time when a D.C. voltage was applied.
       Coil resistance:  770 Ohms
       Coil Height    :  about 4 ft.
       Coil Diameter  :  slightly over 4 ft. I.D.

       Magnet weight  :  700 lbs.
       Magnet Radius  :  2 feet
       Magnet geometry: cylinder rotating about its perpendicular axis
       Magnet  Moment of Inertia:  40 kg-sq.m. (M.K.S.) computed as one
                                   third mass times radius squared

       Battery Voltage:  590 volts under load
       Battery Type   :  Six volt Ray-O-Vac lantern batteries  connected
                         in series

A brief description of the measurements taken and distributed at the press conference follows. When the motor was rotating at 12 rpm, the average D.C. input current from the batteries was about 2 milli-amps, and the average battery input was then 1.2 Watts. The back current (flowing against the direction of battery current) was about -55 milli-amps, for an average charging power of -32 Watts. The forward and reverse current were clearly observable on the oscilloscope. It was noted that when the reverse current flowed, the battery voltage rose above its ambient value, verifying that the batteries were charging. The magnitude of the charging current was verified by heating water with a resistor connected in series with the batteries. A net charging power was the primary evidence used to show that the motor was generating energy internally, however output power was also observed. The 55 m-amp current flowing in the 770 ohm coil generates 2.3 Watts of heat, which is in excess of the input power. In addition, the lights were blinking brightly as the coil was switched.

The back current from the coil switched from zero to negative several amps in about 1 milli-second, and then decayed to zero in about 0.1 second. Given the coil inductance of 1100 henries, the switching voltages were several million volts. Curiously, the back current did not switch on smoothly, but increased in a staircase. Each step in the staircase corresponded to an extremely fast switching of current, with each increase in the current larger than the previous increase. The width of the stairs was about 100 micro-seconds, which for reference is about one third of the travel time of light through the 55 mile coil.

Mechanical losses in the rotor were measured as follows: The rotor was spun up by hand with the coil open circuited. An inductive pick-up loop was attached to a chart recorder to measure the rate of decay of the rotor. The energy stored in the rotor (one half the moment of inertia times the square of the angular velocity) was plotted as a function of time. The slope of this curve was measured at various times and gave the power loss in the rotor as a function of rotor speed. The result of these measurements is given in the following table:

            Rotor Speed       Power Dissipation    Power/(Speed Squared)
            radian/sec             Watts           Watts/(rad/sec)^2
                 4.0                6.3             0.39
                 3.7                5.8             0.42
                 3.3                5.0             0.46
                 3.0                3.5             0.39
                 2.1                2.0             0.45
                 1.7                1.2             0.42
                 1.2                0.7             0.47

The data is consistent with power loss proportional to the square of the angular speed, as would be expected at low speeds. When the rotor moves fast enough so that air resistance is important, the losses would begin to increase as the cube of the angular speed. Using power = 0.43 times the square of the angular speed will give a lower bound on mechanical power dissipation at all speeds. When the rotor is moving at 12 rpm, or 1.3 rad/sec, the mechanical loss is 0.7 Watts.

When the rotor was sped up to 120 rpm by allowing the commutator to fire on all segments, the results were quite dramatic. The lights were blinking rapidly and brightly, and the fan was turning rapidly. The back current spikes were about ten amps, and still increased in a staircase, with the width of the stairs still about 100 micro-seconds. Accurate measurements of the input current were not obtained at that time, however I will report measurements communicated to me by Mr. Newman. At a rotation rate of 200 rpm (corresponding to mechanical losses of at least 190 Watts), the input power was about 6 Watts. The back current in this test was about 0.5 amps, corresponding to heating in the coil of 190 Watts. As a final point of interest, note that the Q of his coil at 200 rpm is about 30. If his battery plus commutator is considered as an A.C. power source, then the impedance of the coil at 200 rpm is 23,000 henries, and the power factor is 0.03. In this light, the predicted input power at 700 volts is less than one Watt!


Since I am preparing this document on my home computer, it will be convenient to use the Basic computer language to write down formulas. The notation is * for multiply, / for divide, ^ for raising to a power, and I will use -dot to represent a derivative. Newton's second law of motion applied to Newman's rotor yields the following equation:

            MI*TH-dot-dot + G*TH-dot = K*I*SIN(TH)      [1]

       where     MI = rotor moment of inertia
                 TH = rotor angular position (radians)
                  G = rotor decay constant
                  K = torque coupling constant
                  I = coil current

In general the constant G may depend upon rotor speed, as when air resistance becomes important. The term on the right hand side of the equation represents the torque delivered to the rotor when current flows through the coil. A constant friction term was found through measurement to be small compared to the TH-dot term at reasonable speeds, but can be included in the "constant" G. The equation for the current in the coil is given by:

            L*I-dot + R*I = V(TH) - K*(TH-dot)*SIN(TH)      [2]

       where          L = coil inductance
                      I = coil current
                      R = coil resistance
                      V(TH) = voltage applied to coil by the
                              commutator which is a function
                              of the angle TH
                      K = rotor induction constant

In general, the resistance R is a function of voltage, particularly during commutator switching when the air resistance breaks down creating a spark.

Note that the constant K is the same in equations (1) and (2). This is required by energy conservation as discussed below. To examine energy considerations, multiply Equation (1) by TH-dot, and Equation (2) by I. Note that the last term in each equation is then identical if the K's are the same. Eliminating the last term between the two equations yields the instantaneous conservation law:

       I*V=R*I^2 + G*(TH-dot)^2 + .5*L*(I^2)-dot +.5*MI*((TH-dot)^2)-dot

If this equation is averaged over one cycle of the rotor, then the last two terms vanish when steady state conditions are reached (i.e. when the current and speed repeat their values at angular positions which are separated by 360 degrees). Denoting averages by < >, the above equation becomes:

            < IV > = < R*I^2 > + < G*(TH-dot)^2 >      [3]

This result is entirely general, independent of any dependencies of R and G on other quantities. The term on the left represents the input power. The first term on the right is the power dissipated in the coil, and the second term is the power delivered to the rotor. The efficiency, defined as power delivered to the rotor divided by input power is thus always less than one by Equation (3). This result does require, however, that the constants K in equation (1) and equation (2) are identical. If the constant K in equation (2) is smaller than the constant K appearing in equation (1), then it may be verified that the efficiency can mathematically be larger than unity.

What do the constants, K, mean? In the first equation, we have the torque delivered to the magnet, while in the second equation we have the back inductance or reaction of the magnet upon the coil. The equality of the constants is an expression of Newton's third law. How could the constants be unequal? Consider the sequence of events which occur during the firing of the commutator. First the contact breaks, and the magnetic field in the coil collapses, creating a huge forward spike of current through the coil and battery. This current spike provides an impulsive torque to the rotor. The rotor accelerates, and the acceleration produces a changing magnetic field which propagates through the coil, creating the back EMF. Suppose that the commutator contacts have separated sufficiently when the last event occurs to prevent the back current from flowing to the battery. Then the back reaction is effectively smaller than the forward impulsive torque on the rotor. This suggestion invokes the finite propagation time of the electromagnetic fields, which has not been included in Equations (1) and (2).

A continued mathematical modeling of the Newman motor should include the effects of finite propagation time, particularly in his extraordinary long coil of wire. I have solved Equations (1) and (2) numerically, and note that the solutions require finer and finer step size as the inductance, moment of inertia, and magnet strength are increased to large values. The solutions break down such that the motor "takes off" in the computer, and this may indicate instabilities, which could be mediated in practice by external perturbations. I am confident that Maxwell's equations , with the proper electro-mechanical coupling, can provide an explanation to the phenomena observed in the Newman device. The electro-mechanical coupling may be embedded in the Maxwell equations if a unified picture (such as Newman's picture of gyroscopic particles) is adopted.

Return to the INE Main Page