The product works remarkably well for what it is doing and for the low cost of manufacture. It will lift quite a bit of weight. I have levitated an 8 oz orange with it, and it had no difficulty with that.
The patent application for the device is available online. US patent number 2007/0170798A1. It is tough to read though. It reads like it was translated from French and Chinese by someone who didn't understand the invention, which is probably just what happened.
Ernshaw's Theorem states that it is impossible to create a static magnetic field that can levitate an object stably on all axes. The configuration of permanent magnets in the Levitron Revolution holds the levitating magnet stable along the vertical axis and is also stable in tilt. But along the horizontal axes the floating magnet acts like it is balanced on a pin.
The Revolution corrects for this using (apparently) four electromagnets to nudge the floating magnet back into position. Position is sensed by hall effect magnetic sensors in the platform. The four LEDs in the corners indicate which electromagnets are active. When the LEDs are on the electromagnets are off. The idea is that there is no continuous power demanded by the electromagnets, just slight nudges to keep the floating magnet centered on the magnetic field created by the permanent magnets. In normal operation you can sometimes see a slight flicker in the LEDs, but mostly they just appear lit and the electromagnets are off.
The electromagnetic correction system works quite well and will normally keep objects levitating as long as power is applied. In fact, it works so well that if you are careful you can rotate the base around 360 degrees with the magnet stably floating off in space to the side of the base one minute and hovering underneath it the next. It looks like magic. This is quite a test of the control system, and if you aren't smooth in your motion the control system will be overwhelmed and the magnet will snap down onto the platform.
Limitations on the power of the control system electromagnets, which are quite weak compared to the permanent magnets providing the levitation, along with the design of the control circuitry which appears to be optimized for slight nudges to keep the magnet centered rather than large pushes to bring the magnet to initial stability make it challenging to get the magnet levitated in the first place. The levitating magnet has a strong tendency to snap to the platform unless swiftly and accurately positioned in the exact horizontal center.
Here are some hints I found helpful for getting the magnet initially positioned:
1. Put some of the packaging foam the base was packed with over the platform when positioning the magnet, then remove it after the magnet is floating. This will reduce the tendency of the magnet to slam hard into the platform, as the force increases exponentially the closer the levitating magnet gets to the platform. This is particularly important when levitating the globe, as the magnet in the bottom of the globe can come detached from impact with the platform.
2. Jiggle the magnet very slightly as you position it. It looks to me like the circuit is designed to be particularly sensitive to AC variations rather than DC levels. This makes sense as its normal mode of operation is quickly nudging the levitating magnet back to the balanced position, rather than keeping the electromagnets on at some continuous level to hold the levitating magnet in place.
3. You use the LED lights in the four corners to determine centering. You have to watch all four at once. When all four are lit the magnet is centered and you can gently let go. The center point is very, very tiny, virtually a pinpoint, and the region that the electromagnets can provide sufficient corrective force is scant larger. You have to have the magnet centered very close to this point for the electromagnets to have the ability to maintain stability.
4. Once locked-on the levitation platform should consume little power, but while you are centering the levitating magnet the power consumption can be unsustainably high. The manufacturer recommends taking no more than 10 seconds to position the magnet. If you don't succeed in that time move the magnet a foot above the platform, let the LEDs reset for 10 seconds and try again. If your unit overheats it is probably due to taking too long to get the magnet centered. There is supposed to be a thermal cutoff in the unit, but I wouldn't rely on that. Make your attempts brief and allow cooling time between them.
5. You need to make many brief attempts, not one long grinding attempt, to center the magnet on the platform.
6. There are several good videos on YouTube about positioning. Watching them will save you a lot of frustration.
The consensus in the videos is that it may take you a couple hours to get the magnet floating for the first time. But after some practice you will be able to do it reliably in ten seconds to a minute.
The inventor, Janick Simeray, has a website with interesting photos and information and has some videos on YouTube.