The Basement Mechanic's Guide to Testing Perpetual Motion Machines
There seems to be a growing interest in perpetual motion machine invention as a hobby. Basement tinkerers build overbalanced wheels, magnet motors, and cyclic hydraulic devices, pursuing the holy grail of achieving a device with a power efficiency greater than one (over-unity, they call it). Too often their ingenuity of design is not matched by equal ingenuity at testing the device, which is absolutely necessary to determine whether they are onto something. It's also necessary to convince any skeptic that your machine is worth serious attention. Furthermore, in science we don't take any hypothesis seriously unless it is in principle falsifiable.
Putting it more bluntly, anyone who has a perpetual motion idea is obligated to state without equivocation or obfuscation what experiment could be performed that might show that it is a flawed idea. If he cannot do that, he cannot expect anyone to take it seriously.
1. Testing is as important as inventing.So you've built a prototype perpetual motion machine? It seems promising, but how can you test its operation and its output/input efficiency? You need a way to compare its efficiency with that of ordinary machines, and to compare improved versions of your PMM with earlier versions, to see whether you are making progress in the right direction. You wouldn't want to waste time re-inventing the square wheel.
Many pitfalls await the unwary experimenter. Any machine novel enough to be PMM may have characteristics that make measurements difficult, and which can even fool ordinary measuring instruments.
You've blown your budget on expensive magnets, low-friction bearings and construction parts, and have little left over for sophisticated measuring equipment. Fortunately there are many inexpensive ways to test your inventions that are good enough for your needs, and, far more important, are not subject to large indeterminate error.
2. Purely Mechanical Devices.
One might have thought that overbalanced wheels, the first kind of perpetual motion machine, were a relic of history. But it only takes a casual search of the internet to discover that the idea is still very much alive. In spite of a long record of past failures, some inventors still hold out hope that some new combination of mechanisms and improvements will be discovered that will at last achieve the desired performance, in defiance of the known laws of physics.
So there sits your invention, a wheel, perhaps, with shifting weights in the form of hammers, levered weights, rolling balls, springs, and other embellishments. You give it a spin and it continues turning for quite a long time before stopping. You are encouraged by this. Of course it eventually stops turning because of that pesky friction, air drag, and other dissipative processes. But friction can always be reduced, improving performance. Is the idea worth improving and refining? After all, the Wright brothers' first "flight" didn't get very far off the ground.
I assume your wheel doesn't start itself, or you wouldn't have to give it an initial push. That push represents input work—a form of energy. After that push the wheel displays kinetic energy of rotation. Some of that energy is slowly lost in friction. The bottom line is this: does the output energy (including that lost through the processes of friction) exceed the input energy?
I also assume your machine slows continually and always comes to a stop, or you wouldn't be reading this, you'd be reaping the rewards of your genius. That tells you that it isn't yet producing more energy than that dissipated by friction and other processes.
Unfortunately the thermal energy (and sound energy) resulting from friction is difficult to measure. Perhaps we should reformulate the problem.
What you really want to know is whether your wheel's special features (those shifting weights, for example) actually improve the performance over an identical wheel that does not allow the weights to shift. If so, then the shifting weights really are improving performance. If not, then you are fooling yourself in thinking they were a good idea.
Suppose you gave the wheel an initial push, putting in a known amount of energy. Measure how long it takes to come to a stop. Then secure those shifting weights so they can't move, and do it again. [Use string, wire, or duct tape to secure them in place.] If those moving weights were really improving performance, the wheel should now come to a stop sooner than before. If the wheel with the weights immobile spins longer, then clearly the moving weights weren't helping.
You could imagine a pair of identical wheels side by side, one with the weights immobilized. Drive both shafts with a friction drive on a small motor, so they reach the same rpm. Remove the driving motor. Now see which one stops first.
Can we apply this same principle without building an identical wheel? If the shifting weights are improving performance they must be supplying energy during wheel rotation. This extra energy compensates for some of the lost energy due to friction. The same should be happening during the initial acceleration. The "higher performing" wheel should require less energy to get it to up a certain rpm than an ordinary wheel.
Method 1. Spring driven.Obtain a suitable spring to give the wheel an initial spin. The spring from a mouse or rat trap is ideal. Attach it in such a way that the wheel may be given a reproducible initial position, then release it to power the wheel and, finally, disengage itself to allow the wheel to continue to turn without impediment. School students in science classes often use this method to power model cars for various competitions.
Method 2. Weight driven.We can also utilize a method often used in freshman physics laboratories. Provide the initial acceleration to the wheel with a weight falling a certain distance. The wheel probably has an axle. Put a peg in the axle, and fasten a string on that peg. Wind up the string around the axle. Hang a weight on the other end of the string. Now tinker a bit, so that the weight is large enough to give a brisk rpm to the wheel as the weight "falls" to the floor, but takes a reasonably long time to do so, perhaps at least 5 or 10 seconds.
The comparison.Whichever method you use, make sure the wheel gets moving fast enough that the wheel's performance enhancing mechanisms are actually operating. Now time the weight's fall with a good stopwatch in two cases: (1) with all your performance enhancing mechanisms working and (2) with those mechanisms disabled.
Be sure that the weight of the entire wheel is the same in both cases and that nothing alters the bearing friction.
For a specific example, see Testing a SMOT.
If those mechanisms are really helping the motion, compensating somewhat for friction, then the time of fall with those mechanisms working should be smaller than when the mechanisms are disabled. If it's the other way around, you'd better rethink the whole idea, for those mechanisms are degrading performance, wasting energy rather than producing energy.
3. Dynamometers.I have never heard a perpetual motion machine inventor mention dynamometers or deProny brakes. These devices are standard methods used even today for measuring power output of machines. But then, I shouldn't be surprised, for the perpetual motion scene is populated by people ignorant of mechanics, engineering, mathematics and physics.
When I was a college student in the 1950s the Freshman physics course included a laboratory experiment that used a de Prony Brake dynamometer. Physics students today spend their lab time using fancy electronic measuring devices and may never learn traditional basic laboratory skills. This fosters a "black box" mentality that trusts someone else to provide the high-tech measuring devices, and someone else to calibrate and repair them, without ever inquiring what's in them and why they work as they do.
The principle of the brake is simple. A leather strap passes over a wheel that is driven by the machine, its ends held fixed, with a simple method for measuring belt tension (spring balance or suspended weight). You also need a tachometer to measure the wheel's angular speed (rpm). Then, using elementary physics:
P = τ ω
P = F v
P is the power in watts
The belt tension is adjusted until the wheel turns at its intended operating speed under load. The tachometer measures the shaft angular velocity. The beauty of this method is that it measures the power output of a machine under realistic load, at the speed the machine is expected to maintain if it were doing useful work. I've seen allegedly over-unity machines being run without load, which tells us nothing useful about the machine's work output or efficiency.
Anyone can build a dynamomter testing rig using inexpensive materials. Of course it isn't of any use if your PPM device doesn't maintain its motion for long enough to measure its rpm. Maybe that's why we never hear of this method being used to measure power output of a perpetual motion device.
4. Reactionless drives and "internal propulsion".A rather esoteric error has been made by inventors who weighed a spinning wheel or vibrating device, and found that it seemed to weigh less when running than when stationary.
It's that old nemesis of perpetual motion machine inventorsfriction. Back in the heyday of spiritualism, some gullible people were taken in by the classic "five finger lift" in which five persons lift another, using only one finger each. The trick depends on human physiology. When five people try to lift a rather cumbersome and non-rigid object like a person, their efforts are uncoordinated. But when they are told to take a deep breath, count to three and then lift in unison, the task is easy. But spiritualists thought it was due to some sort of spiritual or psychic "energy". Spiritualists did tests with an industrial mechanical scale large enough for six people to stand on. They found that the weight registered by the scales after the lift was slightly less than that before the lift. Any first-year physics or chemistry student used to working with mechanical balance scales would know the reason. Even the best scales have residual mechanical friction and hysteresis. So when balanced, then moved off balance, the scale doesn't return exactly to its original balance point. Experienced workers compensate for this by tapping the scale's housing till the truer balance is achieved. When the five finger lift was done on the industrial scale, the act of lifing the person suddenly unbalanced the scale toward a larger weight, and when the lift was complete, the scale slowly returned toward its balance point, but stopped short of that and did not register the correct weight. This is just one of the things that can mislead inexperienced users of mechanical scales.
Balance scales and even electronic balances can also be fooled by vibrations, due to mechanical "stiction" (the "stick and slip" phenomenon of friction). The scale itself is affected by nonlinear phenomena in its mechanism, and these can often display resonance peaks, dependent on the frequency of the vibrations. So a running motor on a balance scale may indeed seem to weigh less when it's running. That has fooled many people, and is one of the reasons for the strange results when Norman Dean demonstrated his "reactionless" drive of the early 1960s.
Dean's machine could sit on a floor or table, and slowly crawl along, or so he claimed. It could even move up a very slight incline. So "with a little more work" he hoped to make one that could be turned on its side, and levitate up away from the floor. That never happened.
What is happening is due to "stiction", the "stick and slip" phenomenon of friction. The coefficient of static friction is generally larger than the coefficient of sliding friction. A standard physics demonstration, the tablecloth pull, uses objects arranged on a tablecloth. When the cloth is pulled slowly, the objects move along with it, since the maximum value of static friction has not been reached and sliding doesn't initiate. But when the cloth is yanked quickly, the threshold force is quickly exceeded, the friction coefficient drops to a lower value and the cloth comes away leaving the objects on the table nearly where they were originally. The whole thing happens before the objects on the cloth can move very far. This is called the "table cloth effect", and it obeys classical physics completely.
Yet, apparently unaware of these well-known physics principles and engineering applications, people even today are still inventing, patenting and proclaiming "internal propulsion engines" and "reactionless drives". Many claim that they are violating Newton's third law. Sometimes they even invent new theoretical physics to account for their imagined violations of physics. Among these are Robert L. Cook's inertial propulsion system US patent 4,238,968, Dr. Gennady Shipov's universal propulsion system, 6,347,766, and James Woodward's theoretical proposal of a reactionless propulsion system, US patent 5,280,864.
Jerry Pournelle has a good account of Dean's device, and makes a suggestion for testing such things. Invariably inventors test them on surfaces or rails, leading one to suspect that friction is doing the dirty work. I haven't ever seen one tested on an air suspension table, to reduce that possibility. Also, anything that rotates may have a "fan effect" and move by pushing against air. So Jerry suggests reducing the friction by suspending the device from wires. Also, one should do the experiment in a vacuum. Actually, it would be sufficient to enclose all moving parts in a box that allowed no communication between the internal air and the external air. If, under these conditions, the device swings to one side when running and returns to center when stopped, then, Jerry says, he might get interested. But inventors do not do this, or anything close to it. Hmm...
Even with this arrangement, self-deception can occur, as in Henry Bull's impulse engine of 1935. You can read about it in Popular Science Monthly, Jan 1935, p. 27: Harry W. Bull: Reaction Motor. His device was in an enclosed box, and suspended from wires as a pendulum. Inside the box two weights were driven by electromagnets, one weight making an inelatic impact with a spring, the other making a nearly elastic metal-to-metal impact. When running, the box containing the device moved to the side. Why? Due to the asymmetric motion inside the box, the center of mass of the box and its contents shifts relative to the box. But the center of mass must still remain where it was before (relative to the laboratory). So the box moves aside, while its center of mass stays put. Newton's laws were working properly, as they always do.
I can't help wondering why these inventors who circulate video clips of their devices moving along a flat surface don't take it another step. Let the device move around a circular track (perhaps slightly "banked") and make a perpetual motion machine. Then they might be motivated to ask whether their "physics defiant" device is also violating conservation of energy, which they could test by measuring the power input to the rotating weights of their device. I confidently predict that energy conservation will be found to be working properly.
5. Magnet Motors.Current serious designs for PMM are seldom purely mechanical. That approach has been fruitless for so many centuries that many inventors conclude there's nothing new there to be discovered. Nearly all of those devices were variations of overbalanced wheels.
Magnet-motors are all the rage these days. But even folks who have had physics and engineering courses are often deficient in understanding of magnetic fields and magnetic materials.
Motors and generators are electromagnetic machines, and electrical engineers have had long experience with them. Magnets and wires moving in a magnetic field continually radiate electromagnetic fields, and these can cause interference with nearby electronic equipment such as telephone, radio and television sets. Such interfering radiation can propagate directly as fields, or can be conducted over power lines. Commercial motors, designed to minimize these problems, are housed within partially shielding enclosures, and often have capacitive-inductive filters to reduce interference with other equipment.
The experimenter's PMM magnet motor, however, is usually exposed and unshielded, without any attention given to designing it to minimize electromagnetic radiation. The output wave form is far from sinusoidal, and induction effects may modify the input wave form as well. These wave forms contain abrupt discontinuities and even sharp pulses and spikes, which simple electrical meters can't respond to properly. The radiated fields may affect the meter's circuitry directly. Also, the output likely has a considerable phase angle between current and voltage. For these reasons, simple electrical meters can give false readings.
Here's a checklist of pitfalls, and suggestions for finding them and ensuring they don't affect your measurements.
One inventor's machine had a battery to start it, and then, the inventor claimed, the machine's own power recharged the battery. It wasn't hard to show that the battery never was being recharged, but was supplying energy all the time. A small value resistor was wired in series with the battery (connected at a battery terminal) and a simple microvoltmeter connected across the resistor to continually monitor the voltage across the resistor. The voltage never changed sign. Therefore the current never changed direction. Therefore the battery supplied energy the whole time, and was never being charged. Case closed.
6. Electrical measuring devices and circuits.One quick way to see if your electrical meters are lying to you is to compare measurements made with one set of meters of a particular type (say analog meters) with measurements made with a different type (say electronic meters). Do thiese measurements on the actual machine. If the results are different, you know that one or the other type of meter (or both) is deceiving you. Do the readings change as the meters are placed in different positions and orientations, or if the connecting wires are shifted in position? If so, be suspicious that the shielding or grounding is not adequate.
Older meters often have an analog electromagetic movement driving a needle moving across a calibrated scale. Digital meters generally have no moving parts. I have seen situations where analog meters, even expensive ones, can be directly affected by electromagnetic radiation from a circuit, especially a circuit with unshielded components, or with magnets. The fields act directly on the electromechanical movement of the meter, and can produce readings that bear no accurate relation to what you are trying to measure.
Electromechanical meter movements may be dependent on the orientation of the meter with respect to the circuit being tested. Their readings may also be dependent on the physical orientation of the meter with respect to gravity&emdash;whether it is upright or lying on its back. In fact, I recall once using very expensive name-brand precision meter that was gravity-dependent, and the instruction manual clearly said that it should only be used lying on its back, not upright. Sometimes it pays to read the instruction manual, even before you buy. It was an example of the fact that "precision" and "accuracy" are not synonyms.
And don't neglect the connecting wires from meter to circuit. These can form an "induction" loop that acts as an antenna for electromagnetic AC radiation. Using shielded and properly grounded wires, or even a twisted pair, can help avoid this.
Consult a good book on electrical measurements and you will discover that proper measurements often demand considerable attention to grounding and shielding of all components of the circuit and the measuring instruments.
Internal resistance of meters can cause problems. Most modern voltmeters have very high impedance inputs (several megohms), which largely avoid this, but ammeters are usually not ideal "zero resistance" devices. It's better to measure current with a high quality voltmeter connected across a precision resistor. That resistor should not be one that has resistance wire wound in a simple coil. (One can obtain resistors with wire wound non-inductively, specifically constructed to avoid such errors.)
Terminals and junctions in your circuit can cause problems, especially at junctions between two different metals. These can cause thermoelectric potentials that are temperature-dependent. Surface oxidation should be scrupulously cleaned from terminals and wires with liquid metal cleaner, then thoroughly washed with distilled water.
And don't forget that pesky phase angle. If there's a phase shift between voltage and current (and there usually is in these devices) then separate meter measurement of voltage and current is a waste of time unless you know the size of that phase angle. An oscilloscope may be necessary to determine the phase angle.
But why not avoid the possible problems of voltmeters and ammeters entirely? The definition of electrical power for both AC, DC and even for pathologically complicated waveforms is based on their heating effects. Using the output electrical power to heat a resistor automaticaly bypasses the phase angle problem. If you power an incandescent light bulb with AC, and another identical lamp with DC, and still another with some horribly complicated wave form, even with a phase angle, then when the lamps burn equally brightly, they are receiving the very same power. More precisely, when they are burning at the same temperature in identical environments, they are receiving (and radiating) the same power.
What could be easier (and cheaper) than using incandescent lamps to compare electrical machine performance by measuring lamp brigtness by photometric methods? You still must use some care and good judgment in designing and carrying out the procedure.
When there's a large difference between the output of the two machines being compared, your eye can easily tell which bulb burns brightest. For smaller differences, a photographer's light-meter at a fixed distance from the bulb may be sufficient. When quantitative results are desired, the test (comparison) lamp can be driven by an adjustable DC source with an ordinary voltmeter and ammeter to measure the electrical power to the lamp. The power source is adjusted until the test lamp's brightness and color exactly matches that of the same lamp driven by the machine.
Of course, when making such comparisons, the input power to the PMM must be controlled and measured also.
Tom Napier (see references below) even suggests measuring electrical power by causing it to heat an insulated container of water and measuring how long it takes for the water to come to a boil. Don't scoff. Sometimes the simplest methods are the best.
7. The power transfer theorem.A principle well known to electrical experimenters is the power transfer theorem. Strangely, it is almost never mentioned by inventors of magnet motors. Any electrical device that has output current also has a characteristic internal impedance within the device itself. In a DC device, that impedance is simply its resistance, so let's speak in terms of this simpler DC example. The power transfer theorem is easily proven from Ohm's law and the electrical definition of power. The result is this: The maximum power is delivered to the load when the load resistance is equal to the source resistance. Therefore the power delivered to the load depends on the load's resistance.
Now that output current that passes through the load is also passing through the electrical generator's internal resistance. When these two resistances are equal half the power of the generator is delivered to the external load, half is dissipated (wastefully, and mostly as heat) in the generator itself. Fortunately maximum power efficiency isn't the same as maximum power transfer, and one can achieve generator efficiencies greater than 90% by making the motor resistance (or impedance, in the AC case) as small as possible. Similar considerations apply to electric motors.
One mistake amateurs often make in testing is to measure the device's input when the device is attached to a load, then use a different load resistance when measuring the output. One must be careful that the meters used do not alter the effective output load. A mistake of this sort could even make your PMM's performace seem worse than it actually is!
8. Testing the Bhaskara wheel.The Bhaskara wheel, the oldest known perpetual motion machine proposal, was assumed to operate by a principle of "overbalance". (Villard's wheel with seven articulated hammers was another example.) The liquid in the vials on one side of the wheel is always farther from the axle, therefore apparently favoring downward motion of that side. This notion is false, and can easily be tested.
Spin a bicycle wheel and see how long it takes to stop. Then attach the vials half-filled with liquid to its rim at an angle, as in the Bhaskara wheel. Give the wheel a spin and time how long it spins. It will come to a stop much sooner than the simple bicycle wheel. Now give it a spin on the opposite direction. It will take about the same time to stop no matter which direction you start it spinning. This is clear evidence that the wheel is not overbalanced to favor motion in one direction.
You could even put the bottles of water in a freezer, and freeze them solid. Use sturdy bottles that won't burst. Then quickly attach them to the wheel, and give it a spin. It should spin much longer than when the water was liquid, demonstrating that the motion of the water in the vials wasn't sustaining motion, but was actually slowing the wheel.
Nit-pickers may object that the initial spin given to the wheel might not be the same in all cases. But the results are so dramatically different with the water bottles and without them, that variations in the initial push don't matter much. One could easily improve the procedure by standardizing the initial push in any of the ways I described above
9. Subtler gotchas.Once while doing a research project requiring extremely sensitive low-current measurements, a strange effect was seen. Of course we had made all efforts to properly ground everything to a common ground with heavy grounding straps, and to shield sensitive components. We were working in a laboratory two floors below ground level. But we'd see a residual noise on the signal during the day, but not at night. The spurious signal was barely seen on an oscilloscope, but was real. At first we suspected electromagnetic interference from office equipment, computers, or printers, which might not be operating after working hours. But the interference signal was constant, not intermittent. Further observation showed that it began approximately at sunset and ceased at sunrise. Yet our lab had no windows and received no sunlight whatsoever. Finally, after seeing this on our oscilloscopes for many weeks, I decided it might be useful to listen to it. So I fed the signal to an audio amplifier, and I could barely distinguish garbled voices and music. It was interference from a local radio station. But why the sunset/sunrise effect? It was an AM station on one of the FCC-designated "clear channel" frequencies, which is required to cut its power from sunset till sunrise, to provide a clear channel for a high powered "clear channel station" operating on that same frequency, to give that station broader coverage at night.
10. Divide and conquer.Perhaps your invention is so complex that the thought of building one for testing seems daunting, and could break your budget. Inventors often face this problem. One approach is to test just the crucial parts of the idea, one at a time.
Look carefully at your design and try to identify what subsystem is crucial to its supposedly superior performance. What is in it that is special, unique, never before employed in such devices?
Can you model just that crucial subsystem, in a way that is inexpensive to build and simple to test? What we want to do is to test not the whole device, but to test the innovative concept that the device depends upon.
At the very least, considerations of this sort will help you to focus on the essentials, and help you to understand your invention better. In so doing you may discover a better, simpler, or cheaper way to do it, or you may discover that the clever idea you hoped would make it perform spectacularly wasn't quite clever enough.