Perpetual Motion Machines Don’t Work

So what’s the trick behind this thing in the video?

The answer is below the fold.  I’m sure the physicists and everyone smarter than me will figure it out by watching.


In the base is a battery powered magnet.  As the ball descends, the magnet increases the ball’s velocity, allowing it to make the jump.  Without the magnet, it would fall.

In a separate video, a magnetic compass was placed on the base, below the ramp.  The needle stayed fixed while the ball was on the platform.  As the ball rolled past the compass and entered the magnetic field, the needle oscillated wildly.


  1. OverlappingMagisteria says

    Just having a magnet in the base is not enough. The magnet would have to power off when the ball reaches he bottom. Otherwise, the ball would slow down as it goes up the ramp. So there would need to be a sensor that turns the magnet off at the right time.

    In the YouTube comments someone mentioned that it is essentially a rail gun. I think that is a more plausible explanation. There is an electric current flowing through the wire that the ball rolls along. Electric currents produce a magnetic field. The ball is metal, so it can carry a current from one rail to another. The current through the ball interacts with the magnetic field from the rails which accelerates it.

    Its really cool toy, even if its not a perpetual motion machine!

    • Rob Grigjanis says

      OM, yeah, a railgun for sure.

      Another doohickey which looks like perpetual motion.

      The magnets at either end of the battery also act as conductors which complete a circuit with current going through the coil from one magnet to the other. The resulting solenoid magnetic field exerts a force on the magnets, which have to be oppositely oriented to have the force on each magnet in the same direction.

  2. blf says

    For at least the USA Patent Office, a working prototype of any “perpetual motion” “invention”-claim MUST accompany the application. This is (now (last I checked)) the only patent-application which must be accompanied by a working model.

  3. wrpinpnw says

    Doesn’t need to be a railgun, just an electromagnet. The metal ball on the rails completes the circuit and turns the magnet on. The descending contact path is longer than the ascending path, so the magnet spends more time accelerating its descent than decelerating its return.

    Takes a bit of tuning to keep the net acceleration near zero, but that’s true of anything that emulates perpetual motion.

  4. Peter B says

    Rail gun? No. Magnet Yes. But how is the magnet triggered?

    The ramp appears to be made of one continuous piece of bent wire. I was first thinking of resistance change when the conductive ball first touched the wire ramp. But I think I saw something better. There appears to be a small guide under the catch basin – ostensibly to guide the ball onto the track.

    My guess is that the catch basin support structure is electrically connected to the small guide under the catch basin. The brief guide, ball, and ramp connection starts the magnet sequence. After perhaps a delay of several milliseconds the hidden coil below the catch basic is energized for just long enough to power the ball through its trip up and back to the catch basin.

    • Rob Grigjanis says

      Peter B, I now think you’re right that it’s an electromagnet, but I think the trigger is just the presence of the ball on the track. It closes a circuit, resulting in current flowing across the ball, and through the concealed solenoid.

      The solenoid is oriented vertically inside the base, just under the lowest point of the rail. As the ball hits the bottom of its arc, the magnetic field is at a maximum, oriented vertically, and the current flowing across the ball interacts with this field to give the ball a sharp push in the direction of its motion.

      Since the field drops off sharply with distance, it’s the closest approach of the ball to the electromagnet which largely determines its motion.

      Of course, I could well be wrong yet again…;-)

    • Peter B says

      The sentence: The brief guide, ball, and ramp connection starts the magnet sequence.

      Would have been better written: The brief guide+ball+ramp connection starts the magnet sequence.

      (I was assuming a microcontroller doing the timing.)

  5. OverlappingMagisteria says

    wrpinpnw: The two rails connect in the front, so using the ball to complete the circuit wouldn’t work – the circuit is already complete.

    I think Peter B might be onto something with a switch at the top and the magnet is just timed to shut off – no sensor required. But a magnets follow the inverse square law, so it would exert a relatively weak force on the ball most of the time and only get strong as the ball gets close to the bottom. This means the timing to shut off the magnet would have to be pretty precise: too early and it it doesn’t get as much of a push. Too late, and it gets a large pull backwards and slows it down. Not impossible to get the timing right since its a pretty controlled track making it predictable… so maybe!

    The rail gun idea could be technically feasible but I admit that I’m not sure what amount of current would be required and what is feasible for a battery. Rail gun weapons need millions of amperes, but that’s to get from 0 to bullet speed. This just needs a little boost. It would also provide a consistent boost the whole way.

  6. wrpinpnw says

    Ah, I missed that — joys of a small screen. I took a closer look and you’re right. That also complicates the railgun hypothesis: IIRC those also depend on current flowing through the projectile, so the rails are usually not connected at the end. The ball will short the circuit as it rolls, which will divert at least a little current, but it would be a good trick to get enough current through it for substantial acceleration over a couple of inches.

    Probably easier to detect the change in resistance when the ball hits the rails and use that to start a timer, as you said.

    I was also wrong about it being tricky to calibrate. Any amount of acceleration that puts the ball back in the bowl will do: the shape of the bowl bleeds off any excess energy, and returns the ball to an initial velocity near zero.

  7. OverlappingMagisteria says

    Only if the connection is conducting.

    I did consider that perhaps the connection between the rails was something non-conducting. But it looks to be made of the same material – just one wire bent around. So perhaps the creator took a lot of time to blend a nonconducting material into a conducting one… but why bother? If the rails need to be separate from each other, it would be far easier to just have 2 separate rails with no fake connection. Having them appear connected does not really anything, so why waste the time building it that way?

  8. lochaber says

    well, if you open the vid through youtube, and click on some of the follow-up vids, there are a couple others showing very similar (but not quite identical, the wood grain patterns on the base plate are slightly different) gadgets, with various views and experiments. One shows what looks like a USB-C charging base that connects to two of the metal feet on the bottom. One shows a compass that seems to track the ball bearing as it travels, and one shows the person turning the gadget on/off while the bearing is at the low point. I was too lazy to do much more than click on a couple related youtube vids.

    If nothing else, the person creating these is a pretty skilled woodworker, as the base appears solid/uncut at a casual inspection, yet there is clearly some magnetic battery-powered bit in the wooden base.

    Maybe over-speculation, but if they can do that good of a job of obscuring the woodwork on the base, I wouldn’t be surprised if they also used a tiny tube of metallic-appearing but non-conductive something for the rails, so that it could appear to be a completed circuit, but not. Or maybe it’s just some magnet/electricity stuff, idunno, I only took intro physics, and skipped most of that, I was a pretty bad student…¯\_(ツ)_/¯

Leave a Reply