I have students building a coilgun. At first they used a manual switch, despite my suspicion that it would not switch on cleanly. After looking on the oscilloscope at the trace of a one shot, we can see that the switch is causing a series of on events ramping up due to bouncing. However, it also appears that the time constant of the system is too big.

We want to use an SCR or something equivalent to switch power on very quickly. However, because the pulse is too wide, it often pulling the projectile backwards. We therefore also want to turn the power off quickly.

SCRs cannot be shut off via gate current, so what is a good approach? We can think of several. Advice?

  1. Increase voltage and decrease capacitance. This maintains energy while reducing the time constant, but we are already at 160V, and I am loath to let high school students boost voltage to multiples of that.

  2. Use power transistors rated for high voltage ~400V in parallel to achieve a high current rating

  3. Some other device we have not heard of? Any suggestions appreciated.

  • \$\begingroup\$ What current do you want to achieve? \$\endgroup\$
    – pjc50
    Nov 23, 2016 at 19:43
  • \$\begingroup\$ Is just current enough? There are very cool MOSFETs. What is the system? I would imagine series of coils, not just one. \$\endgroup\$
    – user76844
    Nov 23, 2016 at 19:45
  • \$\begingroup\$ Well eventually, more than one coil. First we are trying to get one right, and then of course we would build more. We have MOSFETS rated for 400V@3A continuous. We could probably go 10A instantaneous. So we could theoretically put 10 in parallel if we had to. If you can recommend great MOSFETs please do, but realize this is a high school budget. Also the new MOSFETs tend to be surface mount. That's not impossible but a lot harder for them to deal with. \$\endgroup\$
    – Dov
    Nov 23, 2016 at 19:54
  • \$\begingroup\$ The issue with turning inductors on and off quickly is always v = LdI/dt. You're not going to be able to establish the current quickly without lots of volts to push it in, and you're not going to be able to turn it off quickly without dealing with the resulting voltage somehow or other. I doubt capacitance is much of an issue. \$\endgroup\$
    – user1844
    Nov 23, 2016 at 20:24
  • 2
    \$\begingroup\$ @carloc, you are being unreasonable. They love it, they are learning a lot. They are not trying to kill anyone. en.wikipedia.org/wiki/Mass_driver \$\endgroup\$
    – Dov
    Nov 24, 2016 at 14:14

2 Answers 2


The real problem with coil guns is not in turning on the magnetic field, but in turning it off. That is therefore what you should be concentrating on and designing around.

The coil looks like a inductor to the circuit. When you first apply a voltage to the coil, the current, and therefore the magnetic field, build up linearly in time. The projectile starts out stationery, then accelerates as the magnetic field gets strong enough to overcome gravity and static friction. That takes time. You therefore have a relatively "long" time to get the field built up initially.

The reverse is true when shutting down the field. Ideally, you want the field to be strong until the projectile passes the point where the field is pulling it back instead of propelling it forward. This is the point when the projectile is moving fastest, so there is a much shorter time window to get it right.

Unfortunately, the inductance of the coil makes it impossible to shut off the current instantaneously. The rate at which current in a coil decreases is proportional to the reverse voltage applied. To stop the current instantly would require infinite voltage. The speed of the transistor, or whatever you use as a switch, isn't so much the issue as the reverse voltage it needs to withstand.

Since you are driving the coil with 160 V, whether the switch drops a volt or two doesn't matter much. You can use either MOSFETs or BJTs. I would seriously consider BJTs in a real production design, but considering your circumstances I'd go with MOSFETs. One reason is that they parallel better without your students having to get into some details that may be over their heads.

Use N channel power MOSFETs in parallel as a large low side switch. Drive each with a separate off the shelf gate driver. You'll need a 12 V or so supply for those. The inputs of the gate drivers can be tied together, which gives you a single logic signal to turn the switch on and off.

So far you have the 160 V supply, the coil, and the FETs acting as a single switch all in series. That works for turning on the coil, but you have a big problem turning it off. The FETs can switch quite fast, but when they do, the coil will produce a large inverse voltage. If you don't do anything about that, the voltage will go as large as it takes to force current thru the FETs in the short term anyway, which will damage the FETs.

You need to give this coil kickback current a place to go, while still letting the voltage go as high as you can handle, but not higher. The higher you can let the reverse voltage go, the faster the coil current will stop, which means the longer you can keep it on without pulling the projectile back at the end.

400 V FETs are reasonably available, so I'd plan the kickback voltage to be a little less than that. Let's say we aim for 350 V to leave some margin. Since the coil starts at 160 V, that means the reverse voltage on the coil can only be 190 V, so let's say 200 V.

One way to achieve that is to put a diode with resistor in series across the coil. The diode is oriented so that it conducts when the kickback voltage is present. The resistor is sized so that 200 V is across it when the peak coil current is put thru it.

There are many other details to discuss. But, this is getting long already so I'll quit here.

  • 1
    \$\begingroup\$ Lots of words, but no schematic... \$\endgroup\$
    – EM Fields
    Nov 29, 2016 at 0:18
  • \$\begingroup\$ First, a schematic would help. Second, any way to switch in a circuit to let the energy in the first coil start driving a second coil as it decays? That would be one way to dissipate the energy and use it for something beneficial. \$\endgroup\$
    – Dov
    Nov 29, 2016 at 1:36
  • 1
    \$\begingroup\$ @EMFi: Yes, this answer is about the general concepts. The great thing about volunteer work is you can pick the part you feel like doing. In any case, the criticism is a bit disingenuous from someone who hasn't contributed anything useful at all. If you think there should be a schematic, go post one yourself instead of complaining that others didn't use their free time to post one. \$\endgroup\$ Nov 29, 2016 at 11:50
  • \$\begingroup\$ @Dov: Yes, having a sequence of coils and taking the energy from one to start powering the next is a valid concept. However, I'd get a single coil working first. Multiple coils is a more advanced design. It will benefit from the experience of knowing how to drive a single coil effectively. \$\endgroup\$ Nov 29, 2016 at 11:53

Another approach reduces time constant to give an impulse current so short that it dies away before the projectile gets very far. To do that, you'll need far fewer turns than hundreds - perhaps as few as ten (depending on if/how you use iron core elements). You may not use solid, soft iron core in the coil but you may use insulated thin strips, or a bundle of iron rods. Use large gauge coil wire, and connect with very heavy wire, or wire braid with a short path through switch to capacitor bank. You'll need a very robust switch for this, perhaps a big breaker-box switch, and perhaps a few banked in parallel.
The inductor kick-back voltage is not a problem here if you charge your capacitor bank, then disconnect the charger before throwing the switch for launch - the coil current dies far too quickly. Recognize that the huge pulse current abuses many capacitors that are not designed for this kind of service.


simulate this circuit – Schematic created using CircuitLab Excluding electronic switches makes this more approachable to students, and seeing the large wires, switches might show that huge currents (even if momentary) are involved. Also, the heat dissipated by the charging resistor demonstrates that work must be done to charge C1. However, remember to open the charging switch before tripping the launch switch. This setup managed to launch an aluminum ring at truly dangerous velocities.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.