Ed Phillips' Magnetic Levitation

A discussion of magnetic levitation began on July 31, 1999, on the SciClub mailing list with this message from list owner William Beaty (billb@eskimo.com):

Hey, SCIENCE NEWS has a brief article on a new form of magnetic levitation. See the 7/24/1999 issue, where a neodymium magnet is levitated between finger and thumb!

It's done by lifting a small, powerful permanent magnet using a much larger magnet far above. In the article the authors used a huge electromagnet 8 ft. above the little neo. magnet. They also mention that a "handheld" version exists, so a huge magnet is probably not a
requirement. But this doesn't produce stable levitation. So, they place a piece of weakly diamagnetic material above the small magnet, and this repels it downwards. Gravity pulls down, the strong attraction of the huge magnet pulls up, and the weak repulsion from the diamagnet pushes down. With a low-gradient field from a large magnet, the strange
repulsion profile of the diamagnet forms an energy well, and the magnet hangs in space. Even the weak repulsion of human fingers is enough to stabilize the hovering magnet!

I tried this by lifting a 1/2" neo. disk magnet using a large stack of similar magnets, but unfortunately the temperature of the little disk-magnet messes everything up. If the magnet is warmed by my hands, it attracts less strongly, and if it cools a bit, it flies upwards immediately.

The article mentions that the researchers used graphite plates as the diamagnetic stabilizer. A plate of Bismuth metal should work even better.   


I've just (August 21, 1999) made one of these devices and did find that Bismuth worked a bit better than graphite -- at least the graphite that I could find.


Ed Phillips (evp@pacbell.net) responded the same day:  I have made several of these "levitators", using graphite disks, a little RS "neodymium-iron pill", and one of the RS ceramic ring magnets as the main magnet. Can supply details if anyone is interested. Not much work and works quite well.   I have had one sitting on top of my desk for a couple of years now. Got hooked on this (and home-made "levitrons") by Lee Heflinger, W6WZV, who is one of the authors cited in the Science New paper.

I tried bismuth and found that it didn't 't work as well as (some of) the graphite I have. The two graphite disks are out of a 1" diameter rod, cut off with a saw while holding the rod in the chuck of my lathe.   Finished them flatter with sandpaper, although I'm sure that wasn't
necessary. For those who haven't experimented with diamagnetism, it is interesting and very simple to investigate. I have found the most quantitative results to be obtained by hanging a piece as a torsion pendulum, and measuring the angular displacement as I approach the end with a small magnet. Also works if you just hang a lump on a thread and "push" it with a magnet brought up close. I have a number of different samples of graphite, some of which are quite diamagnetic and some of which aren't. The best graphite is substantially more responsive than the bismuth I bought (at great cost) from McMaster Carr.

If anyone is interested in the levitrons, I have some information on them too. Latest activity there is spinning a top (made of a RS ceramic ring) in a vacuum. The top is spun with the fingers and then the lid of the bell jar lowered over it and the valves to the vacuum system opened. The longest run I've had so far is just over 50 hours, during which time the speed drops from the highest stable value of around 1740 RPM to the lowest, almost exactly 1000. I would warn anyone who tries these experiments that it is easy to get "hooked" and spend far more time than may be appropriate.


On August 1, Ed responded to requests for more information with the following:

Guys: Please excuse this awkward way of sending this stuff.  Hope it works, and that there isn't a rule here about sending large notes! John and Pat, IF this works you'll be getting this twice but that's better than not getting it at all.

The attached pictures and simplified drawing give the details of a static levitation  demonstrator identical in principle, but easier to build, than the one in the center of Martin Simon's Page (http://www.physics.ucla.edu/marty/diamag/index.html).  Martin has a couple of good link's to the Dutch site of Andre Geim. You will find pictures of the diamagnetically-levitated frog there, plus lots of other interesting stuff. He also has a link to his own articles on "Spin-stabilized magnetic levitation" which explains the Levitron is great detail, and includes some interesting history of the subject.   It's worth an hour of your time to run this stuff down and, just possibly, you too might get hooked!

The magnets are from Radio Shack, and the upper one is fitted with a wooden hub and #8-32 adjusting screw which is screwed through a threaded hole in the upper plate.  The upper magnet is covered with a plastic pill bottle lid, to make it easier to pick off the little magnet if it flies up there during adjustment.  The actual spacings in the model shown can be estimated from the distance between the top of the top support and the top of the base, which is 3 inches.   The graphite disks were cut from 1" round bar stock because I happened to have it.  (Messy, messy, messy! Still haven't gotten my lathe chuck really clean!)   Thickness was arbitrary. Other shapes could be trimmed as convenient as NONE of the dimensions are critical except the spacing between the disks and the spacing between the upper magnet and the center of the disks. 

To provide damping (NOT necessary) 0.006" copper sheet was glued to the inside of the graphite disks with cyanoacrilic adhesive.   I used this copper because I had it, and it was thin and easy to trim with scissors.  I suspect that aluminum, even heavy kitchen foil, would work equally well.  

All construction details can be modified to fit the convenience of the builder, and any materials on hand should work.  All screws should be brass or non-magnetic stainless except for the upper adjusting screw.  The two screws which support the upper graphite plate have nuts to adjust the height above the base.  Once the thing was built the graphite plates were spaced about 3/16", the upper magnet was installed at about the spacing shown, and the lower magnet slid in between the graphite plates with the correct side up to cause the two magnets to attract each other.  The upper magnet is lowered until it causes the bottom one to lift off the graphite plate, which with the 3/16" plate spacing will cause it to fly up to the bottom of the top plate. 

The adjustment process consists of closing the gap between the plates and readjusting the top magnet spacing until the bottom magnet floats in the middle of the gap.  (If the gap is too large it may have two stable positions, one near the top plate and one near the bottom.) That's all there is to it, except for finding the graphite.  Mine was "in stock", but perhaps someone on this list can help. 

Note that not all graphite is equally diamagnetic, or even diamagnetic at all, so it will be necessary to test whatever you acquire by suspending it from a long thread, approaching it with a strong magnet, and watching the deflection from vertical. Bismuth metal should, in principle, work better but the only source I found (McMaster Carr) was VERY expensive, at around $25 per pound for metal chunks which must be melted and cast into the desired shape, and the stuff I got didn't work any better than the graphite I had. Good luck and enjoy!


If anyone builds one of the things, Ed would appreciate hearing from them.

I just learned of a site that has an excellent discussion of this topic with detailed instructions for building a floating magnet.