Shoot Hard Drive Platters Skywards On The Power Of Magnetism

Project Hathor is an electromagnetic ring launcher that launches aluminium hard drive platters 45 feet skywards at the touch of a button. The hard work is done by a bank of capacitors which are charged to 2kV from a microwave oven transformer, before being discharged into a coil of wire on which the hard drive platter is sitting. The resulting burst of magnetic field induces a huge current in the platter, and that current in turn creates an opposing field which launches the ring into the air.

The launcher is the work of [Krux], at the Syn Shop hackerspace in Las Vegas, and he’s made a beautiful job of it. The capacitor bank has ten 3900uF 400V electrolytic capacitors wired as a single 1560uF 2kV capacitor, there are two 225W 2Kohm wire wound discharge resistors, and a beautifully designed home-made high voltage contactor featuring tungsten electrodes. The whole project has been carefully built into an acrylic case for safety, for as [Krux] points out, microwave oven transformers will kill you.

As well as the project web site, there is a YouTube playlist, an image gallery, and a GitHub repository containing all the project’s details. You can see the launcher in action in the video below, launching platters into the Nevada night right on cue.

Rather surprisingly this appears to be the first electromagnetic ring launcher we’ve featured here on Hackaday. We’ve featured its close cousin the rail gun many times though, from this insanely powerful one through this one that went through two iterations to this comparatively modestly sized model. Now that the ring launcher cat is out of the bag will [Krux] and friends see some competition? It would be nice to think so.

35 thoughts on “Shoot Hard Drive Platters Skywards On The Power Of Magnetism

  1. When hooking up capacitors in series, you want to place the same value resistor in parallel with each one. This keeps the potential voltage the same for each capacitor.

    With out the parallel resistors, manufacturing variations can cause one capacitor to have more voltage on it then the others. This can lead to capacitor failure from exceeding the rated voltage.

    The parallel resistors also have the added benefit of bleeding the voltage when off. ;)

      1. One time I have seen HV diodes built to be connected in series. They were in a mushroom-head shaped package, about 4-5cm in diameter and about 1,5cm height with threaded studs to be screwed together to a cascaded arrangement. 10 together looked a little like an insulator of a HV overhead line.
        Probably they had some controlled breakdown characteristic like Z-diodes, so they could be connected in series without balancing resistors.

      1. Just be careful, if you’re using cheap 10% resistors that you have measured within 1%, they may not stay that way:
        – they are usually carbon film which has a very large thermal coefficient.
        – they may degrade by several % over their lifetime. They should stay within 10%, but it’s enough to ruin your measured binning.

      2. Wasting your time. Electronics rarely requires such accuracy in passives, which is why they sell many more 20% and 10% passives than precision varieties, and if you need precision, just buy precision resistors that are guaranteed accuracy from the get go. More cost up front, but saving you time and as the other guy said, you may not have precision when things heat up….

    1. They probably will not blow at 401V, they are not Tantalum caps. Although its not the best design practice this is no high rel project anyway. I doubt that it sees more than 100 shots before just getting boring. It’s more like a scientific experiment.

    1. The voltage is inconsequential. It’s the current that does the work. Do your engineering properly and you can make this work on a car battery. And you don’t even need capacitors.

      A favourite demo in an old lab I was in used just an iron-core coil plugged into a variac, plugged into a regular 120V circuit. You could hover the plate or loop at low voltage out of the variac, or by just closing the switch directly to the 120V line, you could lob a ring of aluminum through the ceiling tile. No capacitors involved, and it didn’t pop the breaker.

        1. I said, do the engineering…
          That launcher in the video has a coil inductance of maybe 4 uH. Its resistance is in the ballpark of 2 milliohms. The capacitor bank he’s using has an equivalent series resistance of roughly 60 milliohms. The vast majority of the 3 kJ energy stored in those capacitors is dissipated in the capacitors themselves, not in the coil, not in the magnetic field, and certainly not in the projectile.

          Your turn: what inductance to you think is appropriate for a coil for this device? How many turns and what gauge of wire should it have been? Given the high internal resistance of the capacitors, is it better to put them all in parallel, or all in series?

          1. I did a short simulation with 1560µF, 4µH/2mOhm and 2kV. With no ESR you get a peak current of 35kA and very weak damping of the oscillation. With 60mOhm ESR you get about 19kA and strong damping (about one half cycle) until the energy is dissipated completely. I did not know however, how I should take the “load” (Al Plate) into account.
            With a 4kV 390µF 240mOhm setup you get only 12kA peak and very strong damping of the oscillation.
            With 3*3 caps – 1200V, 3900µF, 24mOhm you get 23kA peak. So with a 4µH coil you could take the voltage lower, but not too much.

          2. So, to put this nice and simple, he’d be better with the caps in parallel, and a nice thick coil?

            D’you think rise time is important in this? Probably not as much as in a can crusher. I can see from simple Ohm’s law, that the coil ought to have the higher resistance to get most of the voltage, though of course that means less current. But surely that’d just mean the caps take a bit longer to discharge, but still discharge fully.

            Then I’m stuck. Go on, explain it, please!

          3. @Greenaum- You want the current peak and rise time to be as large / fast as possible in this application because you are trying to induce a current in the platter that opposes the field produced by the coil. So in can crushing or disc launching, the design goals are the same.

          1. Correct. The stored energy is what matters here. Period. If you increase the voltage, you get lower capacitance (for a given price or volume of capacitors). There’s no free lunch, and you can’t magically just turn the voltage knob without changing other things.

            It’s up to the designer to correctly engineer the system with the right resistance and inductance to transfer that stored energy to the projectile with reasonable efficiency. This example, though fun and spectacular and a nice looking build, is pitiful in its efficiency.

            You say “increase voltage”. OK, let’s not spend any more on capacitors, and just arrange all ten in series to make a 4kV bank, with the same stored energy. Now that bank has an internal resistance of 240 milliohms. You just dropped the amount of current you can deliver in HALF, despite doubling the voltage. Was that a smart engineering decision? You tell me…

          2. Hey Mark I just tested a new integral gap bowtie design. It initiates my binary with as little as 1J! I’ll send you some of them. No need for GDT, and the gap design is set for 1.3KV.

          3. Hey Dave and Brett. Yep, no GDT. I will send some to Dave and he can give you some of them Brett. You will understand how they work when you see them. For secondaries you will need more energy and keep the rise-time short.

      1. There was/is a piece of physics lab demo gear that is great to show this. It is a single turn of copper bar about 8mm square with the ends silver soldered to two metal plates. It is arranged so that the plates can slide into side by side beakers: one with hot water and one with cold water. This thermopile produces so much current that an iron cap can not be pulled loose from the single turn and iron core by 5 or 6 college football players.

        So, a turn or two has very low inductance and can take incredible currents. Check Youtube for electromagnetic coin shrinkers. The magnetic field strength is in proportion to the current and the number of turns. Fewer turns and bigger wire and he can get rid of some capacitors or put them in parallel to have the same energy.

  2. This was a very popular demo device in science fairs in the 1950’s and 60’s – the height of the science fair competitions. Back then they used an aluminum ring and a launching coil a foot or two long. In a show place like an athletic field house there was room to shoot them nice and high.

  3. Nice.
    But if it comes off slightly sideways, you have the equivalent of Odd Job in James Bond – dead guys with bits slicked off. Or if it falls on someone’s head, as opposed to their hard hat – more dead guys.
    Still a very nice demo, though.

    1. Considering our recently featured South African pancake machine as an example, we like pancakes here :)

      So that would make something of an impressive pancake maker. Put disc with batter on it on machine, supply disk with a lower strength oscillating field to heat it up and cook the pancake, then hit it with the kilovolts to toss the pancake. Voila! Bits of pancake everywhere!

  4. Here is my take. Spin the disc up to 15000 rpm. Have two aluminium channels horizontal and intercepting the edges of the disc. Insert a dielectric like acrylic between the aluminium channels. Apply aluminum and insulate the acrylic. Positively charge the disc. Make stages to propel the disc like a electromagnet gun. The disc should carry a fairly large magnetic field. This will be opposed by the aluminum channel. The charge on the channel forces the disc out when the spindle is removed. Help me out here this could be awesome.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.