That is very cool too, great work!

I have always seen Jones’ machines being illusions with lots and lots of misdirection. Later models way more than the original. I believe that a lot of the publicity surrounding his machines is purely PR. Like the statement that only one person has gotten close to the secret in 40+ years, I find it almost impossible to believe. In the BBC video when Martyn Poliakoff opens the letter and says he still doesn't know how it works…the letter likely instructs him to be vague and keep the mystery alive. He says in a later video that he was disappointed when he found out the secret. This is the video:

https://youtu.be/muruba8ORGA

I accidentally stumbled on a video of someone spinning a fidget spinner with magnets and em coil and a reed switch…and I realized the magnets are 120 degrees apart. Jones has said he deviced several methods for his machines… though on all 5 I have seen, there’s boxes 120 degrees apart on the rim. The fidget video is how I got thinking of a way to make one. Doesn't need to run for 2 years. If it runs for a day when I turn it on, its enough of a curiosity for me! Ok, I wouldnt object a longer run but not having fidled with electronics since my teens, don't wanna jump into too deep water.

My other fascination is black holes…that I wont try to build…

Very cool, Mr. Jenkins! I have long admired your work. (I never built one of your candles but love the thinking behind them.)

mvmagic: I, too, doubt the authenticity of the claim that "only one person has gotten close to the secret." But that sounds way more romantic than the likely truth. It seems to walk a fine line between entertainment and deception.

I would still say that it doesn't make much difference whether you use a reed switch or a hall effect sensor, except that a reed switch is easier.
I would agree that it's worth doing this with an Arduino, and learning a little bit of coding to do it. The main difficulty, as I see it, is making sure that the wheel rotates at a constant, even speed. To be sure of that you need to be able to time the revolutions, and deliver exactly the right burst of power to the electromagnet, according to the speed the wheel is going. You might even need more than one sensor. Those things are much easier to arrange with a microprocessor than with hardware alone.
I would drive the coil with a mosfet. And don't forget to use a diode with the coil to protect the circuitry.

Possibly slightly off topic, but years and years ago I saw a similar device possibly in the Science Museum in London. A bicycle wheel in a glass case turning with no apparent power source. I came to the conclusion that it was powered by the electric lights that illuminated it. The heat from the lights expanding the spokes of the wheel on one side, causing it to become slightly off centre so that the weight of the wheel was slightly more on one side than the other causing it to turn. It may be that the spokes were made of something that expands more than the regular steel ones.