Author Options:

Electromagnetic levitation via a grid of magnetic "pixels"? Let me clarify... Answered

 An idea popped into my head a few months ago whiles reading about superconductors and how they create a magnetic "mirror image" of the magnet levitating above them. I imagined a grid of small electromagnets arranged similar to the pixels of a computer screen, each one individually controlled by a computer. Some sort of sensing mechanism, of equal resolution to the grid, would detect the magnet's flux at any given point and send that to the software. The computer would model, in real-time, how it must adjust the magnetic grid to compensate for the momentum and weight of the unique magnet floating above the grid. Basically, the software would mirror the magnetic field to allow for perfect stability and even maybe some level of control by offsetting the stability slightly to make the magnet move across the table. So is this even possible? Does it already exist? Do I get a cookie?

*Target all weapons on that idea and FIRE! Shoot it down.*


Have any of you had success in your attempts to build this? I am hoping to make a similar system that uses an infrared camera (trashpicked from a wii remote) for feedback to control the electromagnets.

I have not actually made any real attempts to test this idea. I don't have the expertise/resources to pull it off. However, in my own thinking on this, I conceived of using sensors that were embedded next to each electromagnet that would detect the magnet mounted on the bottom of the hovering object. It would be a tight control feedback loop. I had not considered using a camera but maybe that is a more practical solution.

I suspect a lot of the low level stuff that quad/hex/octo-rotor drones do to stay stable is relevant here too, so maybe the hovering object can have accelerometers/gryos that send commands to the grid for self-stabilization.

Gyroscopes are a really cool idea, I never thought about that. I had also considered ultrasonic distance sensors but they are not very precise, and I think it would not give high enough "resolution" to be stable.

When I finish my project I'll make an instructable for it, so be on the lookout in coming months if you are interested. I might have to reduce the idea to 2d instead of 3D though for cost purposes.

I hope you see success!! Instead of ultrasonic distances sensors, try laser range finders. Might be higher resolution? I still think mini Hall Effect sensors, one for each 'magnetic pixel' of the grid, would provide the necessary feedback, as long as they have a high resolution (not just on/off).


Here (in above URL) is a grid of alternating magnets, an array of NSNSNS super magnets. A graphite disc is being moved across the array with light, possibly due to heating effects in the disc.

Hope this helps you get started on inventing an open source mag-lev personal transport device. (Hint: Put the array above and float it over the ground.)

Sounds like you're trying to reinvent the sort of linear induction motor and magnetic levitation combination that folks used to propose for monorail systems. The levitation is the hard part, though there are several Instructables that illustrate small-scale examples. Linear motors are actually fairly common these days; may of the sliding doors you see at stores and so on are powered by one. Putting them together is a "simple" matter of getting enough power and designing the right feedback loops to stabilize the beast.

 That is closer then I found on my own, but still not my exact idea. I have been trying to search out an existing example of what I have in mind but so far what you mentioned is the closest. I guess my idea is limited by how small and powerful an electromagnet can be engineered to be. I am thinking of an array, on a flat table, of small electromagnets (maybe less then one tenth of a centimeter), arranged in a grid of equal spacing so that an object that is magnetic (either by permanent or electromagnetic methods) could be floated or controlled using a servo feedback controlled by a computer. Thanks for the answer!

 Hi Fluxtesla,

I've tried to construct the idea that you have been describing with natural magnets a little over a year ago. Ive spent some time finding magnets from old speakers, found about 40 magnets and wanted to try the idea of "magnetic pixels" as you call it in order to levitate an object on the grid. After numerous trials unfortunately Ive failed to see the desired effect but I havent had the chance or the time to play with it and there isnt much flexibility with natural magnets. I know the idea must work. I swear Ive been dreaming about this idea for so long now and so glad that Ive found this web site by pure accident through Google. Constructing electromagnets especially of pixel size would be possible with MEMS technology.
Please reply as I would like to communicate with you regarding this, maybe we can help eachother out.

Chingiz Kh.

Hello sir, I am very sorry it took me an entire year (! how does time move like that?) to get back to you...this site never sent me an email letting me know I got a response and I proceeded to forget about it (I only found my own post again today as the first result in Google!). What you tried failed due to Earnshaw's theorem which comes along and ruins your fun by forcing simple geometric arraignments of magnets to be unstable when levitated. http://en.wikipedia.org/wiki/Magnetic_levitation

One way around Earnshaw's theorem is to use "servo feedback" mechanisms like those used in Maglev trains and suspension of magnets. However, those ideas are too limited and offer very little control. The idea you and I have is to use a computer and software and sensors arrayed on the table to control and change the magnetic fields of each little electromagnet rapidly in direct response to instability. I do wonder what the upper limits for small size and high field strength (at the same time) are. The object to be levitated would be another magnet and it's "field mirror image" would be approximated and generated by the whole apparatus in real time, providing smooth levitation.

Again, sorry for not writing back sooner but at least I am now in undergrad level EM physics right now so soon I will have at least a bit of expertise to help with.

Hi Flux,

This is Rezwan Khan. I too have similar idea for my dream project. 

I named these pixels as physical pixel which can be move through our command in a specific area..

They can change their color and place depends of given command. But, I am still in the research. I just completed my MCA and going to do Ph.D for this Research.

Presently working as a Software Engineer in R&D Team for US company from INDIA.

This new idea will be the breath of future generation. Wish u good luck

Rezwan Khan