Making A Coil Gun Without Giant Caps

Whenever we see a coil gun project on the Internet, it seems to involve a bank of huge capacitors. [miroslavus] took a different approach with his gun–he wanted his project to be built without those monster caps.

It’s powered by quadcopter LiPo batteries, 2x 1400 MaH drone batteries wired up in series and triggering 21SWG copper coils that [miroslavus] created with the help of a custom 3D-printed winding rig he designed. The rigs have ridges to help you lay the coils down neatly, and they also have mounts for photodiodes, ensuring the gun knows when it’s loaded.

When triggered, the Arduino Nano activates a pair of IRF3205 MOSFETS with logic signals stepped up to 20V, shooting lengths of 7mm or 8mm steel rod. The gun isn’t exactly creating plasma discharges with its launches, but it’s a fascinating project nonetheless.

Check out the disposable camera coil gun project and the coil guns for newbies posts we previously ran.


26 thoughts on “Making A Coil Gun Without Giant Caps

  1. Are the capacitor really a problem? A small capacitor, of similar size as the used LiPo batteries should outperform the batteries in delivered peak energy. This does simplify the electronics since no capacitor charging circuit is needed.

    1. The problem is that the traditional coilgun, and the type that this guy is building, is trying to dump all the energy into the projectile in a very short pulse, and as you know the coil is an impedance that resists changes in current, so you need very high voltage to get it to react that fast.

      Hence why, even as some super/ultracapacitor might outperform the battery in theory, its voltage is simpy too low. This is a probelem even for the LiPO pack, so you need a different type of capacitor that can tolerate hundreds of volts and has a very low ESR, which means you have to use capacitors the size of soda cans to get any performance out of a coilgun.

        1. That sounds really inefficent since you want peak field strength between the rails. Wouldn’t an array of permanent magnets (or electromagnets) above and below the rails work better?

          1. You know those permanent magnet guns? The ones that are almost perpetual motion except nobody’s managed to “close the loop”? Of course they never will, but… anyway…

            They can fire a magnet with a fair bit of a ping. I wonder about one of those as a first stage for something like this. Think it’d help?

            Also the photodiodes I think are more likely to be used in timing rather than to check if it’s loaded (which is pretty obvious). You measure the moment the projectile is at whichever point, and use that to decide when to energise the next coil. There’s probably some maths involved but trial and error would be good enough.

  2. Fairly impressive for a coil-gun running directly from the batteries as normally the capacitors are to store a higher voltage to overcome the impedances of the coils/etc…

    Though what made me LOL was the mounting the MOSFETS directly to the wood: Wooden heatsink, wooden cool down.
    I guess the fets didn’t run long enough to heat up.

    1. I didn’t have any small heatsinks sitting around. But they don’t really get hot. Even when fired without projectile. Which basically turns the coil on for one second and then turns off the MOSFET was just warm to the touch while the coil was melting the plastic tube.

  3. Looks like a datasheet reading misunderstanding – the 20v he read is for the absolute maximum gate voltage (not the trigger level), the Vgs(threshold) is marked as 2 to 4v (okay, for a low Vds but it should be low as FETs have low resistance). It even lists in that same snapshot that with a Vgs of 10v, it can run continuously at 80A (110A at room temp) – at the datasheet I quickly googled, it showed a chart with the typical gate-source voltages for various currents through the drain when pulsed for short periods of time, like his application, and the Vgs levels were lower still (100A looks to be about 6v, 30A is around 5v).

  4. I wonder how much time it actually takes for the magnetic field to collapse when teh coil is switched off. When the projectile leaves the coil and triggers the photo transistor it may already be too late and the projectile will be slowed down by the magnetic field that is still present. Maybe it would be better to have the photo transistor at the ingress end of the coil to trigger when the projectile enters the coil, then have an adjustable time delay to turn the coil off again – this way the setup could be adjusted for optimal performance..

    1. With the flyback diode directly across the coil like he did, it’s going to take forever. The current decay depends on the voltage across the inductor, since U = L * dI/dt. Now only the forward voltage of the diode and the voltage drop across the internal resistance of the coil are the total voltage, resulting in a slow decay. For a faster decay, you should place a resistor or zener diode in series with the flyback diode, such that you nearly but not quite reach the MOSFETs maximum drain to source voltage. Alternatively, you could switch both sides of the coil (high side and low side), and clamp with diodes to the opposite sides of the battery. Each of those options will result in a much larger voltage across the coil when it’s turned off, and therefor a much faster current decay.

      1. Great suggestions. I’ll have to consider this as well. I like the adjustable delay for switching the MOSFET but it’s going to be pain in the butt with 6 coils. I’ll try out few methods.

  5. A wholly different approach is needed to get an effective coilgun running directly off of batteries. Instead of one or three coils, you need dozens of low-impedance coils that are switched rapidly in a wave sequence over distance to give the projectile some time to accelerate – a linear motor instead of a single impulse.

    Consider that a pellet gun will have a muzzle energy of about 10 J. Let’s further suppose you want to reach some appreciable speed, like 100 m/s so it’s better than just throwing a rock. To accelerate the projectile to full speed over the distance of a 2″ long coil takes about 1 milliseconds, and 10 Joules over 1 milliseconds is 10 Kilowatts. Far more than the battery pack or the electronics can supply.

    However, if you did the same thing over a 20″ linear motor, you’d have ten times more time to accelerate and your peak power drops to 1 kW which is perfectly doable with a small LiPo pack.

    1. Plus, you get much higher efficiency out of a linear motor than you do out of a pulsed coil.

      Reason being that the ferromagnetic projectile saturates fairly easily, and that tops off the pulling force you can apply to it. Worse still, the projectile gets permanently magnetized and you lose energy into re-arranging the domains. Add more current and you get diminishing returns, and your efficiency drops down to 0.1%. Even the best coilguns have efficiencies in the single percentage points.

      A linear motor in contrast has a copper or aluminium slug as a projectile, and it works by inducing an eddy current into it, which in theory has no limit except proximity and skin effects that apply at high currents and switching speeds. In any case, you can apply much higher field densities without so much loss, yet you don’t need it to be that high because you’re not trying to achieve a huge single impulse, so you achieve much more for much less.

    2. That is what I was thinking about, got 6mm ID 0.5m carbon tube, many IRF3205, 10 rolls of 0.1mm wire to make Litzt wire, coil should have 30-50 turn only 7mm long but 40-48 pieces, current aimed 200A-300A 20V.
      Would use 2400mAh, 130 peak C lipo, And an STM32… with many outputs, coupled to IR diodes and fiber optic cables to switch the FETs via IR diode – transistor. I have no clue what will happen to the 6x6mm Nd-magnets, will they keep up while accelerated.
      The bright side of doing this, if it fails as coil gun it still can be useful as linear drive…

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.