So I have a high voltage pulse circuit. I have planned it to be around 90kv with the capacitor being 1.3uF. It is to be discharged through a spark gap. Before I test it, I of course want to save the capacitors life on maximum reverse voltage. The specs list the maximum reverse voltage (to keep the working life of the capacitor) to be 30% (of a listed rating of 100kv).

I am concerned about ringing after discharge and am not exactly sure how much reverse voltage there may be, but I would like to keep this value as low as possible. I have considered using diodes and resistors to dampen the ringing, but I am trying to keep the cost as low as possible.

Given this, my question is about a possible second spark gap as a scheme to keep the cost down.


simulate this circuit – Schematic created using CircuitLab

The idea is that the secondary capacitor will discharge through the secondary spark gap when the main pulse capacitor reaches about 10kv or so, thus limiting the reverse voltage on the main capacitor and therefore limiting the bounce back ringing on it so that it doesn't ever reach the 30% reverse voltage rating.

Now bear with me because I'm not super familiar with high voltage pulses and how they relate to the exact timing and dynamics of the current through the plasma channel of a spark gap, but I'm thinking that maybe one of two things might happen. Either the plasma channel will continue to conduct as the voltage on the main pulse capacitor reaches the point where the plasma channel on the secondary spark gap opens up, just draining the secondary capacitor as if it were parallel with the main pulse capacitor resulting in not much of an effect (if any), OR a miracle in timing might happen so that somehow the plasma channel of the primary spark gap ceases and the secondary capacitor and spark gap puts just enough charge on the primary pulse capacitor that it never really drops that far below 0V thus dampening the ringing and reverse voltage. (Or perhaps something else might also happen that I hadn't considered.)

So what do you think? Might this be a cheap way of not using HV resistors in series with HV diodes (I would have no idea how to estimate what current rating I would need for those diodes, or how much dampening effect I could financially afford in the diodes with a minimized resistance)? Or am I just wasting my time and am going to have to drop some dough on these HV components in order to save on the life of my pulse capacitor?

(Also, any other ideas on how to cheaply dampen this oscillation are appreciated.)

Thanks in advance and let me know if I need to be clearer on anything.

  • \$\begingroup\$ Ok. I think that I might have just answered my own question. Tell me if this is correct. The parasitic inductance of my capacitor is about 50 nH. I just need to find the series parasitic inductance and resistance and then add in series inductance and resistance until I reach the critically damped state. Correct? If it is, then I have another question. I'm trying to discharge the capacitor as fast as possible without having more than a 30% reverse voltage. how would I go about maximizing the discharge rate to find the best damping Factor which would be somewhere slightly below 1? \$\endgroup\$ – user198606 Apr 6 '19 at 7:04
  • 1
    \$\begingroup\$ The best dampening (zeta = 1, I think) is sqrt( L / C). Or some near value. Go examine the transient dampening plots. \$\endgroup\$ – analogsystemsrf Apr 6 '19 at 8:52
  • \$\begingroup\$ So I simulated this with a 1.3 uF, .1 ohm, 50 nH RLC circuit and it looks like the max reverse voltage is about -30 kilovolts, which is about 30% Max reverse voltage, at around 900 nanoseconds. Does that seem right? \$\endgroup\$ – user198606 Apr 6 '19 at 8:57
  • 1
    \$\begingroup\$ You might want to look into pulse forming networks. (PFN) \$\endgroup\$ – D Duck Apr 6 '19 at 11:13
  • \$\begingroup\$ I was kind of thinking the plots might look similar also. Thanks for your opinion. \$\endgroup\$ – user198606 Apr 12 '19 at 15:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.