The Howard Johnson Magnet Motor

Primary references:—

 Science & Mechanics magazine, Spring 1980, p45, and US Patents 4,151,431 and 4,877,983.

On April 2, 1979, Howard Johnson obtained US Patent 4,151,431 for a permanent magnet motor. Further details of Johnson's inventions were reported in the Spring 1980 issue of the now defunct Science & Mechanics magazine. From my old, faded copy of it, here is a scan of its cover:—

Science & Mechanics magazine, Spring 1980

A discrepancy

Right at the outset, there is an obvious discrepancy on the cover page. A fifteen horsepower output converts to 11,185.5 watts. With say a 95% efficient belt drive, a 15-HP motor should still easily be able to drive a generator rated at 10,000 watts, i.e. twice the 5000 watts stated.

The Science & Mechanics report showed some of the devices that Johnson had built:—

Some of Johnson's experimental devices

The device shown at top right is apparently the one that convinced the US Patent Office to grant the first patent. However other experimenters (including myself) soon found that it is not too difficult to obtain continuous rotor movement in a device like this provided that the central magnet is hand held, as shown above. The manual effort required to hold the magnet more or less in the correct position means that the person holding it is unavoidably exerting energy, which is transmitted by magnetic forces to the rotor, causing it to turn. As soon as the central magnet is mounted — even spring-mounted — on some kind of stator, the device no longer works (in my experiments, at least).

Modelling

Nevertheless, more recently I decided to analyse a few examples of Johnson motors. I started with one based fairly closely on Figures 9 and 10 from the patent:—

Example 1

I also looked at a couple of variations:—

Example 2

Example 3

At first, I had difficulty obtaining clear-cut results from my analyses of these models. Generally, there were unbalanced forces and torques for all of them; but I was very skeptical about these — one reason being that they failed to meet some simple Newton's Third Law checks.

Ultimately I added a small ring around the three outer magnets as shown in this example. For whatever reason, in his patent Johnson preferred the outer magnets to be the rotor (which he called the armature). They are circumferentially magnetised, and the inner magnets are radially magnetised.  

This added ring meant the outer magnets could be modelled as a single object, with full symmetry about the z-axis (red). The small ring would make only a very minor difference to the device's performance. Now, to find any net energy output, it was only necessary to find whether this outer object had any net integral of torque through angle, over an operating cycle.

This approach gave consistent and credible results. With this approach I didn't find any net energy output in either of the Examples 1 or 3 above.

Conclusion

It is not worthwhile for me to spend further time on Johnson's work, at least for his US Patent 4,151,431 motor.

Showing magnetic flux lines for a 2D model of a Johnson motor
with a single armature magnet