I'm tasked with replacing a relay with a semiconductor switch in a high voltage, high current, short pulse duration application (all relative, I concede). Basically I'm charging a capacitor up to ~1kV and dumping it through a low-impedance load. The di/dt requirements are pretty gnarly: ~400A with a ~300ns pulse width. Due to other constraints it has to be a high-side switch, too.


simulate this circuit – Schematic created using CircuitLab

I originally tried using a high-power SCR, but all viable devices I've found have a di/dt limit of 100~150A/us, whereas I'm in need of about 2500A/us.


simulate this circuit

My next plan of attack is using an IGBT with a gate-drive transformer, though I'll have to make sure to keep the leakage inductance low.


simulate this circuit

My question, though, is if anyone has another approach to consider? The isolation is a bit of a PITA but that's the route I'll take if it's the most practical.

EDIT: we were using these voltage-controlled solidtrons which worked perfectly (in a 5-lead TO-247-ish package) but now this is no longer an option.

  • \$\begingroup\$ Your third diagram appears to lose the plot unless you plan on using a very big heatsink for the MOSFET. Plus, your 2nd diagram will not be able to turn the pulse off after 300 ns - it will turn off when the current falls below the holding limit for the SCR. \$\endgroup\$
    – Andy aka
    Aug 10, 2017 at 17:18
  • 1
    \$\begingroup\$ Have you studied Ware, A. A. , Phil. Trans., A, 243, 197 (1951). Kurchatov, I. V. , J. Nuclear Energy, 4, 193 (1957) or his later paper in 1961 called "High current gas discharges?" Or considered any kind of triggered gas discharge module? \$\endgroup\$
    – jonk
    Aug 10, 2017 at 17:25
  • \$\begingroup\$ @Andyaka the pulses are so short, and there's so much time between them, that we've been able to get away with the third circuit (also it's a beefy IGBT, not MOSFET, but there's no IGBT symbol available). For the second circuit there's enough inductance in the loop for the current to swing negative and turn off the SCR -- the pulse just gets squashed wide and short due to the SCR's low di/dt. \$\endgroup\$ Aug 10, 2017 at 18:43
  • \$\begingroup\$ @jonk I feel like I can get away with a simple yet clever design using more conventional semiconductors, hence my fishing for advice in that realm. But I'll certainly investigate this. \$\endgroup\$ Aug 10, 2017 at 18:51
  • \$\begingroup\$ I also just added a link to the current device's datasheet. I've never heard of a SolidTRON before (I assume it's proprietary) -- just wondering if there aren't other similar devices floating around that could do the job. \$\endgroup\$ Aug 10, 2017 at 18:51

2 Answers 2


I'll leave the semi-conductor circuitry to the experts on this topic. I'm just a hobbyist in electronics and this is way out of my hobbyist league. I can suggest an alternative device you may want to consider. If nothing else, perhaps it may encourage you to contact a few companies providing such devices and have a conversation with them about existing products that may fit your application area.

So below will be a general discussion of the idea. (I know that commercial gas discharge products today can handle upwards of \$20\:\textrm{kA}\$.) These products are especially indicated when you are talking about fast and high current and high voltage. So that is why I'm bringing this up.

Like everything, physics matters. A triggering delay may be very predictable (the hot-cathode with priming electrode comes to mind) or it can be a bit variable. Also, the speed of the discharge can be fast (several design parameters involved here) or it can be slower (long, narrow tube with outside metal band as an electrode, requiring the discharge to spread throughout a long, narrow tube.) So it's important that you know exactly what you need here: trigger delay time, trigger delay predictability, allowable pulse duration, etc.

A thyratron is a gas discharge tube that can be triggered. The solid state equivalent is, of course, the thyristor (SCR, et al.) The reason this came immediately to mind is because physicists have been researching high current, high voltage pulses for quite some time. A lot has been published since about 1950, or so. These were (I can't speak to now) particularly used when there was a need for extremely precise, repeatable, and narrow triggering pulses (such as was required for so-called "nuclear triggers" used with early plutonium bombs, for example.) So lots of money spent here, of course.

The physics of gas discharge requires at least 2 spatial and one time dimension of PDEs coupled to at least 6-dim ODEs to apprehend well. (That doesn't include radiation transport and atomic interactions.) A simpler version will use global rate equations that assume spatially averaged densities for the charged particles and neutral atoms and molecules, but then it still needs to deal with excited and metastable states. There is a modest, intermediate text by Lieberman and Lichtenberg called "Principles of Plasma Discharges and Materials Processing." (From John Wiley and Sons.) It has a good treatment on the fundamentals of discharges and those simplified global rate models I mentioned, plus something on collisions and DC and RF discharges.

Thyratrons in both cold- and hot- cathode varieties. They are still quite important; most especially whenever you see the combination of two or more of: (a) high-voltage; and, (b) high current; and, (c) fast-switching. Your application seems almost an obvious fit to me.

With cold-cathode thyratrons, you apply a voltage above their extinction voltage but below their strike voltage. In your case, you'd need to find one where \$1000\:\textrm{V}\$ is bracketed by the extinction and strike voltages. (They can easily be designed for such a voltage, if you can consider getting one designed.) Once that voltage is applied (easy for your case), a pulse is then applied to a trigger electrode that causes some local ionization which will then spread throughout the tube quite quickly.

A hot-cathode thyratron is a triode (but with substantial gas) where a negative grid bias is what holds them off from triggering. These may include a priming electrode to increase the speed of the trigger and to make for more robust and predictable triggering parameters, as well.

Triggering electrodes can be inside or outside the tube. The typical camera flash will use a small band of metal around the outside of the tube, for example. (Those are usually high-pressure Xenon tubes.)

There is no reason I can think of that you cannot make all this work with semiconductor devices. Your current requirements are within reach; your timing is within reach; your voltage requirements are within reach. If treated individually. But there are parasitics involved and your requirements may dig pretty deeply into the physics of semiconductor devices.

So I just wanted to offer an alternative area to check on, as well, since you opened the doors.

  • \$\begingroup\$ I feel ya -- I'm just looking for a semiconductor solution this time. I bet a thyratron would sound great in a guitar amp, though :) \$\endgroup\$ Aug 10, 2017 at 19:28
  • \$\begingroup\$ @calcium3000 Hehe. It would be just "NUCLEAR" to listen to!! \$\endgroup\$
    – jonk
    Aug 10, 2017 at 19:36
  • \$\begingroup\$ You can use a strobe tube. \$\endgroup\$
    – Andy aka
    Aug 10, 2017 at 19:47

I had a need for high voltage high current fast switching devices for use in high resolution ultrasonic imaging. 1500volts, 50A, <10ns. This was back in the early 1990's and I saw references to a device called an Austen Switch, an optical triggered semiconductor device apparently developed by team at Uni of Austen, Texas with picosec. di/dt. When my technical researcher made enquires about this device, she was told that the device was restricted and no information was available. Turns out it was developed for nuclear bomb triggers and at that time the issue was very sensitive because of foreign governments wanting to acquire this technology. I was also seeking ultra fast discharge capacitors for the same ultrasonic development, which was also a sensitive component. (I ultimately went for avalanche bjt's introduced by Ferranti at that time, ZTX415 ). Never did find out if the Austen switch made a commercial product debut.


Your Answer

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

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