All-mechanical magnetic levitation
with neodymium supermagnets
1995 William Beaty

NOTE: this is an adult-level project. It requires a machine shop in order to build it, and the finished project is extremely dangerous. For simple and safe, try simple maglev train.

While working on science museum exhibits in 1990 I came up with the above idea: it is known that a spinning metal disk will lift and fling a strong magnet. Therefore, metal rods with opposite spin will lift a magnet but WON'T fling it sideways. It works! I used "sched-80" heavy wall copper tubes about 1-3/8" diameter, 12" long, with 1/4" wall thickness. I hammered aluminum plugs into the tubes, carved shaft-tips with a lathe, built endblocks and bearings, spun them with an AC/DC motor, and managed to levitate a stack of two 3/4" diameter neodymium magnets. The spinning tubes must move at about 5000 RPM before the magnet starts floating.

The tubes I ordered from ALASKAN COPPER AND BRASS, Seattle WA, 1-800-552-7661. The exact size isn't too important, but the tube should have a VERY thick wall, 3/16" to 1/4" thick if possible. My copper tubes were about 1-1/4 OD and 12" long, about $25 each back then [ P. Ledlie has found a second source for tubes: "Copper and Brass Sales" at 503-254-2600, more expensive though. ] Alaskan Copper has a stock size tube of 1" ID, 1-5/16" OD which should work.

I used an old motor I already had, a Dayton AC/DC 115V 1/15HP 5000RPM unit. I believe that these are sold in Grainger catalog under universal AC/DC type, 1/15 horsepower. The drive belt was a large O-ring, I put a 1/2" pulley on the motor shaft, and moved the copper tube bearing endblocks so the tubes pinched the O-ring/belt slightly (so when the motor spun one tube, the other tube also touched the belt and spun.) Note that the arrows in the above drawing are wrong, the magnet floats MUCH higher if the tubes spin the other way instead. Spin them so they "blow" the magnet upward rather than "sucking" it downwards.

If you wish to reproduce the above device, be aware that you'll probably need access to a machine shop. To connect bearings to the tubes, I made tapered aluminum plugs on a lathe, chilled them and hammered them into the ends of the tubes, carved them into small shaft stubs which fit my ball bearing races, then carved some plastic end blocks using drillpress and bandsaw.

WARNING WARNING WARNING
Note that when many pounds of copper tube is spinning at 5000RPM, it is EXTREMELY DANGEROUS. If you bump your fingers against it, it could EAT YOUR ENTIRE ARM INSTANTLY. Or if the sharp edges of a magnet get caught, the magnet could explode and the pieces fly out like bullets. It's FAR more dangerous than any power tool. For safety I made a close-fitting plexiglas box to cover the tubes: I heated a sheet of plexi on a greased aluminum foil sheet in an oven, molded it over the top of the tubes while wearing heavy gloves, then carved it square and used it as the top plate in a plexiglas box. The magnets only hover about 1/4" above the spinning tubes, so if you build a cover, it must be close-fitting without touching. The cover isn't required for initial testing, just for operation around curious fingers.

Try to avoid using steel for anything except the ball bearings, otherwise the hovering magnet will be sucked to the end of the tubes. If you decide to make the tubes lots shorter than 12", be aware that the magnet might become unstable and drift towards the steel ball bearings. This might be fixable: put a small piece of iron below the tubes, close enough to attract the magnet to remain centered, but not so close that it pulls the magnet down into the tubes.

Other notes: neodymium-iron-boron (NIB) cylinder magnets are required. Ceramic or alnico magnets are too weak, although they might hover if the RPMs of the tubes could be doubled or tripled. Samarium-Cobalt magnets work, but are too expensive. One source for rare-earth NIB magnets is Force Field, www.wondermagnet.com. I know that their 3/4" dia, 3/8" thick NIB disk works well. I haven't tried the small cylinders, but they PROBABLY work OK. See my NEODEMO file for more fun things to do with NIB supermagnets. Heed the safety warning!

How does it work?

Brief explanation. When a magnet rapidly approaches a copper plate, the changing field from the magnet causes the electron-sea of the copper to swirl in a loop-shaped path. All metals, even non-magnetic ones, are full of electron-fluid, and when magnetic fields are moved through them, they apply a pumping force to the electrons. In physics-speak, the moving magnet induces an electric current. This circular current itself acts like a powerful electromagnet, and creates a magnetic field of its own. This field repels the incoming magnet. (Search on "Lenz's law" for more about this.) As a result, magnets repel all metals, especially the good conductors like copper and aluminum. However, the repulsion force only exists briefly after the magnet approaches the metal. The electrical resistance of the metal rapidly slows the circulating current. As a result, when you bring a magnet near a piece of nonmagnetic metal, the magnet and metal repel each other, but only for a fraction of a second. Drop a magnet onto a copper plate, and the magnet is slowed slightly, but does not hover. But if the magnet could keep moving, or if the metal plate was spinning fast, new regions of metal would cause the current to renew itself as they approach, and the repulsion force would not die away.

Another way to make a magnet hover: use a superconductor. When a magnet approaches a superconductor plate, it induces a circle of moving charge. Since the superconductor has no resistance, the current will never be slowed, the repelling fields will not die away, and the magnet will hover. But superconductors require super-cold liquid nitrogen for their operation. They are also fairly expensive: $10 to $20 for a 1" disk.

If we could just MOVE our magnet suddenly sideways across the copper, a new pattern of current would be induced, and the magnet would be repelled again. If we could move the magnet rapidly across the metal plate, it would lift up and fly. One form of "Maglev Train" uses this effect, it contains powerful magnets in its lower surface, and "flys" over a thick aluminum trough.

The magnet-roller device turn the train idea upside-down. The "train" stays still and the track moves fast.

All-mechanical magnetic levitation with neodymium supermagnets 1995 William Beaty

How does it work?

LINKS

All-mechanical magnetic levitation
with neodymium supermagnets
1995 William Beaty