Testing a SMOT

by Donald E. Simanek

What is a SMOT?

Greg Watson's SMOT (Simple Magnetic Overunity Toy) (1997) has attracted the attention of people who want to believe that perpetual motion is possible. The word "overunity" in its name suggests that it has an efficiency greater than one. In fact, some claim 113% for this device, but their method of determining this is suspect. It's a simple device, easily built with readily available square ceramic "refrigerator magnets".

The diagram illustrates the arrangement of components. Ceramic magnets are epoxied or otherwise fastened to an iron strip, with all their N poles (red) facing the same direction. Two such arrays are made, and fastened to a wooden board so that they form an angle. Bisecting that angle is an aluminum track on which a steel ball (purple) can roll freely.

As one would expect, when the ball is placed at point A, and then released, it will move with accelerating speed toward point B, toward the region of stronger field. It will overshoot past the magnets, and there it is allowed to fall, under the action of gravity, to a lower level.

Now the whole apparatus can be tilted, so that B is higher than A. If the tilt is not too great, the ball will roll uphill from A to B, then drop to a lower level.

Our drawing shows the magnet arrays farther apart at the left than the right. The original SMOT design had the arrays parallel, but tilted so that the left end was lower than the right end, compared to the track. The important feature is that the field experienced by the ball be increasingly greater as it moves from A to B.

The non-ferrous track is necessary to prevent the ball from being attracted to and moving toward one or the other of the magnet arrays. One can imagine other ways to constrain the ball to move along the centerline between those arrays.

To a physicist, nothing surprising is going on here. But some people, misled by careless and incomplete measurements of velocities, are claiming that the device gives the ball up to 113% more energy after the drop than it had at point A. If that were so, they could use this device as the basis for a continuously-running perpetual motion or over-unity machine.

Perhaps what's surprising to some people is that the ball starts from rest at A and attains speed by the time it reaches B. They assume that the ball gains energy from "stored energy" in the magnets. But in fact, at the starting point, the ball already had stored potential energy due to the work done when the ball was placed in position at A. (It won't move to that position by itself.) Any excess energy at point B is simply that small amount of energy it acquired as it was moved into position at the starting point. One can notice this, as the ball must be held at the starting point against the small force of the field pulling it forward. The ball gains kinetic energy as it moves from weaker field at A to stronger field at B, but when moving past the end of the magnet array, it experiences a retarding force from the magnets that would slow it down if it continued on a level path. It would slow to a stop if the level path is long enough and nothing else acts upon it.

The perpetual motionists suppose that the loss of energy was entirely due to friction of the ball rolling on the track. They fuss about rubbing graphite on the track to lower the friction. Friction is difficult to measure accurately, so it's a convenient excuse why the performance of the machine is not up to their expectations.

The perpetual motionists also try to arrange a string of SMOTS in series. They talk about "closing the loop" with ramps to route the exiting ball back to the starting point. They wonder why none of these actually work. Not once has a ball completed one full cycle.

In fact, some engage in premature speculation that there might be some "regaguing of the field" as the ball drops at B" that there might be energy gained from "zero point energy of the quantum field" and all sorts of similar pseuscientific mumbo-jumbo. They produce lots of talk, but they scrupulously avoid doing any definitive experimental tests that might demolish their belief that something unusual is going on here.

Nothing surprising is happening. Yet this device can serve as a simple example of how to properly measure the performance of this device, also applicable to testing many other types of alleged perpetual wheels.

Testing a SMOT

The PM enthusiasts claim that the SMOT somehow increases the energy of a ball moving through it. That's the claim we must test. Friction can never be completely eliminated in any mechanical system. So we arrange a way to move a ball through the smot, then do the same motion through the same space without the smot in place. We then compare the performance with and without the SMOT. If conventional physics is applicable, the performance should be no better (and probably somewhat worse) with the SMOT than without it.

Imagine a steel ball on a rigid wooden bar, rotating so as to move through the SMOT. Better yet, have a ball on each end of the bar, for balance, and for twice as much "SMOT boost" per cycle. Oh, why not have a wheel of spokes with balls on the end of each? Why bother with steel balls; just have steel or iron wheel spokes, whose ends pass through the smot. The figure illustrates this refined version.

Now as each spoke enters the SMOT, is it "sucked through", speeding up? Alas, it slows down on exit. But how much? And is it simply due to friction in the wheel? Is there any additional energy gain from passing through the SMOT? None.

Remove the SMOT, and let's study the behavior of the wheel itself. Put some sort of spring at the wheel axle, arranged so that it can be "wound" a fixed and precise amount, to then deliver that amount of stored energy to the wheel. The wheel will speed up as the spring releases its energy, then the spring disengages without disturbing the wheel's motion. We then time how long it takes the wheel to come to a stop.

Now do the same experiment with the SMOT. If the wheel is actually gaining energy from the SMOT, it will run for a longer time before stopping. The wheel friction is the same in either case. Dissapointingly, the wheel now stops sooner.

Of course, two trials are not enough. We need to do a number of repeated measurements under the same conditions without the SMOT in place, to see how consistent and reliable is the wheel and the friction. For light loads such as this, an unlubricated bearing is best. These repeated runs will establish what ± error limits to place on the time measurements.

With the SMOT in place, a similar set of repeated measurements is made. The error limits are determined. If the performance in the two cases differs by an amount greater than the error limits, and the difference favors the SMOT, then we might have something going on that is worth further investigation. But if the difference is within the error limits, or if the difference favors the case without the SMOT, then the fantastic claims of the SMOT are not supported.

The "black box" labeled "SMOT" in the diagram may be replaced by any magical device that is claimed to boost or increase energy of material objects passing through it. Therefore this serves as an example and guide for testing many other allegedly over-unity devices.

Other possible tests.

One may easily make some other experimental tests.
  1. Use a single steel rod fastened to an axle. Let it swing through the smot. Measure the initial height and final height of swing. Do the same without the SMOT.
  2. Turn the wheel described above so that its axle is vertical, and the spokes move in a horizontal plane. Turn the SMOT 90° also. Now gravity is eliminated from the problem. Carry out the same sort of comparative measurements.

Historical precedents.

Finally, we note that the seduction of this idea is the same one that was seen in the earliest historical magnetic perpetual motion proposals. Consider the classic design shown in the figure. A magnetic lodestone is fixed at the top of the post. An iron ball is supposedly drawn up the straight ramp by its attraction to the magnet. But it encounters a hole, where gravity causes it fall to the lower, curved ramp, back to the starting point where it again moves up the straight ramp. A minor engineering problem must be solved at the bottom, to allow the transition from curved ramp to straight ramp, but that's solvable. The trouble is that the motion down the curved ramp isn't simply under the influence of gravity, but also is affected by the attractive force of the lodestone at the top. It took scientists a long while to realize that magnetic fields are conservative, and that the work done by the magnetic field on a body moving around any closed loop in a magnetic field is zero. This also applies to SMOTs.

Whenever I speak to an audience of physicists, and show this picture, it always generates laughter, from those who have seen it before and those who haven't. They immediately recognize it as an absurd idea even before they carry out any analysis of it. In fact, they are unlikely to bother to do any analysis, because they know the general laws of field theory and know what they allow and what they don't. They are entirely justified in this reaction. Their response to the SMOT is the same, for the same good reasons.

Questions from readers.

Is it possible to get an array of SMOTS in series, adjusted so that a ball would go completely around the loop?

This is the "closed-loop" problem. No one claims to have achieved that yet. Those who have tried say that they can't get rid of that pesky friction.

In the absence of all friction and other dissipative processes would cyclic performance be possible?

Usually inventors begin by linking several ramps in a straight line. Remember I said that the ball already has energy Eo when placed in the starting position. That was the energy it gained as your hand moved it into the field at the starting position. The ball gains kinetic energy K going through the magnet array, and loses that same amount on exit, leaving it with energy Eo as it enters the next ramp. So, in the absence of friction, you might expect the ball would retain energy Eo at the entry to each successive ramp. That would seem to violate no physics, for a frictionless wheel would cycle forever as well. But you couldn't get any more energy out of it than that small Eo, and removing that energy would stop the ball's motion. A perpetually moving wheel does not violate any physics, but you can't get out of it any more work than put into it, and in extracting that work, you stop the wheel.

A cyclic loop device of this kind presents another problem: rolling resistance. A loop path necessarily requires the rolling ball to turn corners, or go around curves. That involves a change of momentum of the ball. In fact, even when we imagined chaining two ramps in series, there was a change in the direction of the ball's velocity as the ball went from the bottom of one ramp and then up the next, and the reason for that change is a force from the ramp acting perpendicular to the ball's velocity. This causes slight deformation of the ball and/or the track that results in a force acting on the ball opposite to its velocity (tangential to the track). This does negative work on the ball (removing energy from it). This is technically called "rolling resistance". Rolling resistance acts much like friction, but has entirely different causes. It results from the inevitable elasticity of materials. Nothing in the universe is perfectly rigid. See my physics demos page Even when the ball changes direction from one ramp to the next, rolling resistance degrades performance and removes energy from the moving ball, irreversably. This is worse than when a ball rolling along a straight track, for the normal force (perpendicular to the track) is greater at curves and corners, and so is the tangential force due to rolling resistance.

Our idealization of zero friction and no dissipative processes is, of course, unrealistic, and was only used here to illustrate the fact that friction is not the only reason this device cannot have over-unity performance. We could also idealize by assuming that all parts of the system are perfectly rigid, so even rolling resistance is absent. But it still won't work.

Elsewhere on these pages we often began our analysis with a "zero friction assumption" to help us find the other more fundamental reasons why a device can not work. That is usually sufficient to expose the flaw in the inventor's idea. But this is one example of a device where we must look deeper into the science of materials. Even our frequent example of the spinning wheel on a frictionless axle will still not spin forever if it is in a gravitational field, for deformation of the axle and the hole in the center of the wheel give rise to rolling resistance that exerts force in opposition to the wheel's motion, even if there is no friction.

Perhaps the perpetual motion inventors should first work on devising a material of infinite hardness. That might require some very expensive machine tools to fabricate. This reminds me of the alchemists who sought the universal solvent that would dissolve anything. But first they would have to find a perfectly insoluble material to make a bottle to keep it in. And so on, ad infinitum. Science gets complicated when you bring infinity into the picture.

If the kinetic energy of the ball on exiting the SMOT is greater than the kinetic energy it had at rest at the starting point, where does that energy come from?

There is no excess energy. Part of the appeal of the SMOT is a tactile-visual illusion. We see the gain in speed up the ramp and say "Wow, where did that energy come from?" But compare a more familar example of a ball rolling down a ramp. We take the ball from the table and place it at the top of the ramp, it gains considerable speed rolling down and can then roll a great distance on a level surface. It is hard to realize that all that energy came from the act of carrying that ball from the table to the top of the ramp (from A to B). It seemed "effortless", since we are used to lifting and carrying small objects. But the energy is "there" (at point B), evidenced by the fact that you have to hold it in place at the top of the ramp, and as soon as you remove your hand from the ball, the ball "takes off" and begins to roll down the ramp. The potential energy is due to the position of the ball in the earth's gravitational field. The same is true in the SMOT, where the ball at the starting point has potential energy due to its position in the magnetic field. You can "feel" the pull on the ball due to the magnetic field. This is no different from holding a ball and then dropping it. It started at rest, but already had potential energy due to its height above the ground.

"Where was the energy when we placed the ball at the starting point?"

Where is the energy of the ball at the top of the ramp in the gravitational example? The energy is in the position of the ball with respect to all other things that exert force on it. It is a "configuration" energy due to geometry and the force laws. That's what potential energy is.

Those who think the SMOT has energy output/input greater than one have made a simple mistake. They take the input energy as simply its gravitational energy at the starting point. They neglect to include the potential energy of the ball due to its position in the magnetic field.


It's generally not difficult to test hypotheses about mechanical systems, and it doesn't require costly apparatus. One is sometimes tempted to think that all the pompous talk we hear from perpetual motion enthusiasts is all bluster. They seem to scrupulously avoid doing the simple and inexpensive tests that could conclusively demolish their preposterous claims.


Hans-Peter Gramatke has a picture of a modern smot on his web page Perpetuum Mobile—Actual Devices. He notes that the smot is not new, but goes back to 1922. Reference: Prachar, F.: Jak jsem hledal a nalezl perpetuum mobile (How I searched and found the Perpetuum Mobile). Prague, 1922.

An interesting web page exposes the errors made by people who calculate the coefficient of performance of the SMOT at 113%, a figure much quoted on web pages: SMOT. Unfortunately I don't read Russian, but a Google translation of the page is easily understandable. The bottom line is that in these careful experiments, the energy efficiency (correctly measured) of the SMOT was never greater than 80%.

Other patents for such devices, often called "magnetic conveyors": USA 1859764 (1932), 3,263,796 (1966), 4215330 (1977), 4074153 (1978), and 4215330 (1980).

Donald E. Simanek, June 2004. Minor revisions, Oct. 2017.

Return to front page.
Return to the top of this document.
Return to the Testing Perpetual Motion Machines.
Return to The Museum's Main Gallery.