It works! (After a week of adjusting and testing) A NIB magnet with
some metal pipe is supported motionless, 1 inch below the coil. "..look Ma, no
The left photo shows the coil and earlier circuit board. The other two are of the neat side and the not-so-neat side. Alright, so I don't know how to do PCB's.
The coil will draw 10 A at 13V and is wound on an old solder bobbin. In operation, however, the draw is less than 2 A at 12 V. The drive circuit uses three LM324 quad op amps (of which 9 of the 12 are used) and a dual 555 timer (NE556). This drives a IRFP450 MOSFET rated at 500V 14A under the CPU cooler fan. There are two BYV29 500 V, 9 A, 60 nS diodes to absorb the back EMF when the coil switches off. The circuit comes from Rick Hoadley's excellent magnet site. I have made only minor modifications including reducing minimum pulse width from 4 to 1 uS and using the MOSFET instead of an IGBT.
This is a standard 200 W ATX computer power supply which is 7 years old. It supplies +12 V 8 A (yellow wire), -12 V (blue) and +5 V (red). It has been modified to include a power switch on the back panel and a divider across the 5 V line with a 24 ohm and a 10 ohm resistor in series. This allows some load for the 5 V line (? need 0.1 A) plus a 3.3 V reference for the orange lead which normally senses the computer board voltage of 3.3 V. I did blow out the first supply I used but no problems since this modification which also powers the LED through a 100 ohm resistor from the 3.3V line. I have tidied up the leads into one multicore cable terminating in a 25 pin plug.
In action with a fancy mount and fitted into a suspended PVC pipe with all the wires tidied up, it can support a dish with 110 g of sweets. (This was for a lunchtime demo). It worked OK but unfortunately I left it running inadvertently for some time and the electromagnet melted through the end cap. Ooops. Remarkably the Hall device and the electronics were unharmed.
Magnetic Levitation Mk II
Mounted on the coil are the two Hall effect devices (one end is shown on the right connected to shielded cable). There is a thermal cutout not shown (microwave ovens have lots of these rated at 145 - 160 degrees C). Also not well shown are a 0.22 uF 250 V polyester capacitor and a freewheeling BYV29 500 V, 9 A, 60 nS diode and heatsink to constrain the RF hash locally.
This shows the mounting inside a 12 inch (30 cm) acrylic sphere, with a high flow centrifugal fan supplying airflow through the PVC above it. Without the copper block beneath it, the magnet tends to oscillate and become unstable after perhaps 10 seconds or so. Despite my best efforts to counter this electronically, the most effective way to dampen it is to have a heavy aluminium or copper block beneath it about 1 inch (2-3 cm). This still gives an effective display and you can spin the NIB magnet or have it completely in your hand still levitated and resisting being moved away. Giving the magnet a simple spin with the fingers will keep it rotating for 1 1/2 hours limited by friction and eddy currents. It will also support an iron object such as a spanner without a magnet as above.
(Video 625k MPEG - run mouse over to play)
The control electronics are essentially the same but I now use two paralleled MOSFET's on a larger heat sink for extra power handling. The single control is for control of power and hence position.
The magnetic levitation demonstrated by Sylvia at the GDS. This (the mag-lev, not Sylvia) has been on constant hands on public display since November 2004 and no-one has pinched the magnet yet. The alarm that sounds as the magnet is pulled away helps....
Magnetic levitation of a coil 2006
The left photo shows the coil levitating above the copper plate. The middle photo shows the 284 g coil lifting a 217 g solder reel (total 500 g). Current was a bit higher and distance off the base was 1/4 inch compared with 1/2 inch before. The right photo with the clamp meter shows the current of 6.5 A.
Above, left photo shows voltage across coil of 100.6 V, middle photo shows current of 6.61 A (giving 665 VA) and the right photo shows true RMS power of 0.55 kW (550 W). I presume this is a power factor of 0.82.
Above, left photo shows double coils elevating with only 2 support wires plus the effect of a large iron mass beneath it (none as it is greater than the skin depth in copper of about 5 mm at 50 Hz). The right photo shows the setup of the two coils and true RMS power of 1.3 kW.
I did try levitating a MOT with the E section opened and facing down but really needed a better winding. Hence I tried a non cored coil.
Levitation with magnetic induction 2006
Above, left photo shows the cut down transformer. The middle photo shows an aluminium disc in place, and the right photo shows the disc being levitated. The central drinking straw is to stop the disc sliding off the field. It will spin freely.
Diamagnetic levitation 2004
The left photo shows the display setup with large bismuth plates and the right photo shows the closeup of the smaller but easier to see single bismuth cylinders.
The left photo above shows a 10 g bismuth cylinder showing a weight (with support block) of 38.96 g. The right photo shows that this reduces to 38.57 g in the electromagnet field. This is a 0.39 g or 3.9% reduction in the weight of the bismuth. It is not well optimized though. Hardly levitation as the field is not strong and also does not have a strong gradient.
The left photo above shows the 10 g bismuth cylinder with a weight (with support beam) of 24.38 g. The right photo shows that this reduces to 24.16 g in the NIB magnet field. This is a 0.22 g or 2.2 % reduction. With the same magnet one can elevate a thin sheet of pyrolytic graphite which is the most diamagnetic stuff around. Unfortunately I don't have any to show.
Meissner effect levitation on a track http://www.youtube.com/watch?v=CQ70tayKZh8
Simple Meissner levitation http://www.youtube.com/watch?v=pO2eDJBr50E
This page was last updated June 03, 2009