Building A Magnetic Levitating Quadcopter

hover Three days ago on October 21, 2014 it was announced to the world the Back to the Future hoverboard was real. It’s a Kickstarter, of course, and it’s trending towards a $5 Million dollar payday for the creator.  Surprisingly for a project with this much marketing genius, it’s a real, existing device and there’s even a patent. From the patent, we’re able to glean a few details of how this hoverboard/magnetic levitation device works, and in our post on the initial coverage, we said we’d be giving away some goodies to the first person who can clone this magnetic levitation device and put it up on

[jellmeister] just won the prize. It’s somewhat cheating, as he’s had his prototype hoverboard working in July, and demoed a more advanced ‘upside-down quadcopter’ device at the Brighton Mini Maker Faire in September. Good on ‘ya [jelly]. You’re getting a gift card for the hackaday store.

hoverLike the Kickstarter hoverboard, [jelly] is using an array of magnets rotating in a frame above a non-ferrous metal. For the initial test, eight neodymium magnets were arranged in a frame, suspended over 3/4″ aluminum plate, and spun up with a drill. With just this simple test, [jelly] was able to achieve 2kg of lift at 1cm and 1kg of lift at 1 inch of separation. This test also provided some valuable insight on what the magnets do to the aluminum or copper; the 3kg aluminum plate was nearly spinning, meaning if this device were to be used on small plates, counter-rotating pairs of magnetic lifters would need to be used.

The test rig then advanced to two pairs of rotors with standard hobby brushless motors, but stability was a problem; the magnetic rotors provided enough lift, but it would quickly fall over. To solve this problem, [jellmeister] took a standard quadcopter configuration, replaced the props with magnetic rotors, and successfully hovered it above a sheet of aluminum at the Brighton Maker Faire.

Since [jellmeister] has actually built one of these magnetically levitating hoverboards, he has a lot more data about how they work than an embargoed press release. The magnetic rotor hoverboard will work on aluminum as well as copper, but [jell] suspects the Kickstarter hoverboard may be operating right at the edge of its performance, necessitating the more efficient copper half pipe. The thickness of the non-ferrous plate also makes a difference, with better performance found using thicker plates. No, you bojo, hoverboards don’t work on salt water, even if you have pow-ah.

So there ‘ya go. That’s how you build a freakin’ hoverboard. [jellmeister]’s design is a little crude and using a Halbach array for the magnetic rotors should improve efficiency. Using a 3D printed rotor design is a stroke of genius, and we’ll expect a few more quad-magnetic-levitating-things to hit the tip line in short order.

Demos of [jellmeister]’s work below.

Oh. These things need a name. I humbly submit the term ‘Bojo’ to refer to any device that levitates though rotating magnets and eddy currents.

81 thoughts on “Building A Magnetic Levitating Quadcopter

    1. Nope. You’re probably thinking about the Curie temperature, where a ferromagnetic material loses its magnetic properties. For iron, this is something like 800°C.

      The lifter works on the principle of induced eddy currents creating an opposing magnetic field. The metal therefore doesn’t need to be magnetic – indeed, if it was magnetic (e.g. iron), the ferromagnetic attraction would be far greater than any repulsive force created by eddy currents. While heating of the plate may be a bit undesirable, I don’t see it should affect the lifting performance. (Well, in theory. Resistance increases with temperature, so the lifting _may_ decrease with increasing temperature, but you’d need to get really hot before that became an issue).

      1. neodymiums lose most of their magnetic properties around 400F, I know this because I discovered you can solder to that nickel coating. With my Weller, I had to use the lowest temp possible and it still was demagnetizing N48 magnets.

      2. That’s equivocating… Besides, it is theory that I remember, and apparently I do remember it correctly.

        Perhaps the engineering of it comes down to rider weight vs. plate thickness. It would be interesting to see the final dollar figure of a skate park like this.

      3. In most materials, its resistance goes up as it heats up. The eddy currents in the plate the device is pushing away from will heat the plate. The heat will increase the resistance, the increased resistance will reduce the (eddy) current flowing in that area, and that reduced current will cause reduced (lift) force. So I think the claim of reduced lift is valid.

        1. Assuming lifting off an Al plate, using the data on page 1134, I estimate about a 1.25% loss of lift per degree C of temp rise in the plate. I did this estimatation by looking at the slope around 288K (room temp), and assuming that the field from the spinning mags would remain constant, that the current would still be the product of the field scaled by the material’s resistivity, and that the reduced current would be directly proportional to the change in force.

    2. Just wanted to point out some support for the idea that the levitation effect will work well into a temperature range where you have much bigger problems with your skate park:

      1. The entire skate surface would act like a giant heatsink. Unless you have the park jammed full of people in a stationary grid lock, the heating effect for each tile is momentarily. Can’t say the same for the motors, power circuits and batteries etc on the board though and they will have power losses.

    1. Why? It’s a gift card? It cannot be exchanged for money and only works in this store? How is the taxman gonna know that tshirt he got using a giftcard was obtained using currency that was not taxed?

      Do you count the cash your grandma gives you for Christmas as taxable? Do you list all the freebies you get at shows and conventions as taxable income?

          1. I think part of your message got cut off, but you’re right I’m not licensed to give tax advice.

            fprintf(m, “With fond wishes for a speedy recovery,\n”);
            fprintf(m, ” Elvis\n”);

        1. Do you make sure to declare all your imports to customs, so they can be taxed? Or are you happy when your order from abroad comes through with no extra taxation?

          Let’s face it, we don’t pay silly taxes if we don’t have to. Mum’s the word.

        2. I understand if you win the lottery it’s supposed to be taxed, but is it still valid if you win a gift-certificate? The gift-cert is not real money and cannot (easily) be exchanged for money?

  1. What about one with another set of magnets that are static balanced against the rotating magnets force enabling it to traverse walls and perhaps ceilings.
    Although the undulater style arrangement of the magnets might be necessary for the greater mass.

        1. The French Metric is sort of convenient for science – not so much now that we don’t use slide rules. The Imperial and one of its offspring, the U.S. metric, are scaled to human use and single syllable unit names that are very distinct and avoid miss-communication when spoken. Also 10 is commensurate with 2 and 5 and a French Metric scale can not be made by folding. 12 is commensurate with 2, 3, 4, and 6 and a scale or ruler, including fractional units, can be made by folding. Just sayin.

          1. Great moment in Spinal Tap. They confuse ‘ and ” but would never confuse “foot” and “inch”. In the French system we have the same problem with “m” and “M”.

      1. Well, the three countries using exclusively the imperial system of measurements are the United States, Myanmar and Liberia. Not exactly enlightened company.

        “We all” are already using the metric system, and it’d be nice if you’d join in.

    1. I use metric for everything I do as in science it really is a must. I do however use inches to describe approximately 25mm as it is quicker, is a useful scale most people (in the UK at least) can visualise and for me at least sounds less precise than 25mm. I use inch to mean 20 – 30mm as opposed to 25mm exactly.
      An interesting aside is the reason the metric system didn’t find its way into the former British empire immediately. When Napoleon invited the word to come and see the new metric system, Britain had just sank his fleet and so were not invited!

  2. What are the legal implications of this, considering it was patented. And is the patent even valid, as it relies on a “discovery” and/or direct application of a basic physics quirk of nature, instead of being an actual invention (I am obviously not a patent lawyer)

    1. Everything you built yourself is in respect of patents fine (not for weapen though), moreover if he has proof that he was making the stuff (in public) before the patent was filed it will affect the patent.

    2. He’s fine in respect to the patent- as long as he doesn’t sell it, he’s fine and dandy.

      Also, you can patent an application of a thing, you don’t need to patent the thing- take medicines for instance, they’re not patenting the chemical, they’re patenting using the chemical as medicine.

      1. Not true. You cannot make anything that infringes a patent, even if you don’t sell it. However, if they go after him he can probably successfully argue that he did his before they announced their patent. The rest of us, however, have been put on notice and cannot make one of these hoverboards. We’re free to make some other hoverboard that doesn’t violate the patent, however, and if the patent only mentions rotating magnets then we may have an out: oscillating magnets might work, too.

        1. Arguably, if the previous hoverboard device was built and displayed before the patent was first filed (the priority date), then it is publically disclosed and the patent is worthless. One would argue “No, I’m not building a Hondo hoverboard, I’m building a Jellyboard!” and you are pretty much untouchable. Of course, you want to be very very sure that the prior art really was prior, otherwise you are setting yourself up for some unhappiness.

        2. A patent is not, nor has it ever been, an enforceable restriction on non-sale production. A patent exists to both protect the recapture of development costs and protect profits from ideas for specified period(s) of time. The gov’t can’t hold a gun to your head and force you to purchase a certain widget rather than making it yourself… patents are reactive, not proactive. Given the number of patents presently outstanding, if they did restrict production in an absolute sense, no one would be allowed to do anything (though I suppose eating, breathing, and screwing are in the public domain at this point).

        3. You’re wrong Rick. Anyone can make exact copies of an invention disclosed in a patent. They just can’t make a profit on it in the country(s) the patent(s) are active. After the patent expires, or if the inventor fails to pay the maintenance fees, then all bets are off.
          A patent in one country only protects the inventor in that country. A US patent will not stop a Chinese manufacturer from mass producing and selling the invention in all countries.

    3. I can answer this one, but I don’t offer this as advice, just my opinion… I have no issues because:
      1. The patent is an application. The A1 at the end means it is currently unenforceable as it is not granted.
      2. The only published version I have found so far is US and I am in the UK.
      3. It is unlikely that it would be deemed that I am impacting their ability to make money
      4. Most of their claims are likely to be rejected as they are either not novel over publicly disclosed prior art or lack an inventive step. I am aware of another group on the Web who someone pointed me at when they saw what I was doing who made a mini hover board back in 2011, 2 years before Hendo’s priority date. Their version appears better than ours!
      Not much to fear:)

  3. wouldn’t an electronically switched electromagnetic halbach array be better since all you need is a changing field? I’m thinking along the lines of a switched mode power supply, or do electromagnets have far too weak a field compared to whatever rare earth magnets are in this thing. It seems like the biggest limit is the fact you have to rotate so much mass as fast as you can and the fact permanent magnets only give so much field per unit mass, so electromagnets would solve both problems.

    1. Halbach arrays are created by creatively arranging permanent magnets to mostly cancel the field on one side of the magnet. They’re crazy common, and you probably have several already. Those flat fridge magnets? Ever wonder why they only stick to the fridge on one side?

      You can’t do that efficiently without a whole lot of electromagnets – and the problem with using an electromagnet to levitate as well is that you’ve now got losses in both the induced currents and in the original currents to generate the field.

      Really, though the magic in figuring out the best way to do this is in figuring out the most efficient field geometry. A Halbach array is ideal for concentrating the field on one side of a surface, but here you probably want it even more concentrated than that.

      1. You already have losses in the motor (electromagnets) from spining the magnets. What Quazi is suggesting is essentially taking the motor stator and using it to induce currents in the copper plate directly. Since it eliminates the intermediate rotor it may not be such a bad idea.

        1. Permanent magnets are used because of the size of air-gap relative to the device. It simply takes too much power to throw something like a 1/2 Tesla field across a 1cm air gap with an electromagnet of this size. Conversely, a fairly small (and cheap) volume of rare earth magnets can create that field. Scale it up to use on a car or train, and the situation reverses.

        2. The losses in the rotating assembly that *wouldn’t* be there in a ‘static’ assembly would only be the mechanical rotating losses, and those can be made arbitrarily small. The majority of the drag on the rotating assembly will be magnetic drag, and that’s there no matter what.

          A static electromagnet will have lots more electrical losses, since generating a field of that size will take a large amount of current – and possibly more annoyingly, *changing* that field will require shifting large amounts of current around.

          1. I haven’t worked out the math but it intuitively seems like the Lorenz force would induce the same kind of power draw from the motor as ideally switching a field would, and you’d be getting the same amount of electrical losses and then some to “change the field” induced by the permanent magnets by rotating them.

            I’m not an electrical engineer, but since the speed you switch the field has a very real effect on the dissipative power losses, i’d figure the faster you can change the field the better your efficiency will be, since that’s from what I understand the reason switched mode power supplies switching at very high frequency are much more efficient than transformer windings do equivalent inductive transfers using mains.

            These hover engines seem to be a clever arrangement to use permanent magnets on motors to induce a field through an air gap, so they’re essentially from what i understand an air core transformer which dump all their energy supplied to a motor straight into a plate, after all if this thing weren’t spinning it’d be doing nothing but lying on the ground. It seems like instead of in a roundabout way using power to transfer energy through a motor connected to magnets that have a very spread field with limited rotation speed to the floor, a very fast switching “magnetic lens” with a focal point some distance away could much more efficiently induce the field at a much more configurable tiny/large point on the ground being a weird variant of a switched mode transformer.

            After all the goal isn’t really to make a big strong magnetic field, as much as it is to as efficiently as possible power / induce a force to counteract weight. you could probably design a much more ideal arrangement to do this than a spinning ring of magnets, all you want is the force not the field.

    2. The motor would only need to draw power to compensate for the *mechanical* work done in rotating the magnets + usual losses. In the case of a rotating magnetic field, you’ll needed to provide the current to generate magnetic flux equivalent of the rare earth magnetic + your losses for copper (i^2*R), core and switch. You would need multiple electromagnets/phase windings to make a 2D rotating magnetic field.

      It probably works out to be easier to implement and use less power with off the shelf motors and magnets.

      1. Drop a neo magnet down a copper pipe. Its a typical demo in a physics class. There is a LOT of drag, like 5 seconds to drop a footer (foot or meter – choose magnet as needed). I would guess a pretty good load on the motors.

  4. How about making a rolling non magnetic set of round cylinders, similar to Pringles cans and using that as a base? Attached with strings to the board so that they are dragged along with it.

    1. the tanks in Leo Frankowski’s “A boy and his tank” did just that. The tracks were a series of conductive rods that were ejected from the tank at the front and picked up again at the back. On a conductive surface, the tanks could retract the chain of rods and use the surface itself for maglev.

        1. Mixing units? No way around that: pound/inch, kg/centimeter, kg/inch, pound/centimeter – any which way you have a mass unit mixing with a length unit. So they come from different regimes – who gives a crap! In all cases it *works*.

          It’s not like they start by calling it ‘2.5cm’ and later refer to that same measurement as ‘about an inch’

  5. I suggest the generic names of floaty things in eddy currents should be Chesterfield devices. So named because Eddy’s velvet paisley-covered Chesterfield sofa was found floating in the space-time continuum… Anything less would be uncivilized to non-panicked hitchhikers.

    1. Interesting future direction but not really feasible at scale currently… In plane conductivity is high and, assuming it behaves the same as pyrolytic graphite (which is basically lots of aligned graphene layers), you get a TINY extra helping hand from its diamagnetic repulsion:) – not that I’ve been playing levitating graphite sheets over magnets today at Derby Maker Faire or anything…

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s