Reverse Voltage Protection

[FonziDaytona]

Hi everyone, I am new here.

I have a few defibrillator capacitors on their way to me that I want to use for can crushing and other fun discharge experiments. I can’t think of any affordable way to protect them from voltage reversal, so thought that you guys may have some ideas. To switch them, I am going to use a Jennings vacuum contractor (rp101f) rated for 30kv DC and 50kA peak. 

As an alternative, I could use some DC link capacitors that I am using for a high power ruby laser project.

[klugesmith]

How about a great big silicon diode in parallel with the capacitor?   Then if the LCR circuit is under-damped, so "discharge" current continues after capacitor voltage reaches zero, the current goes through diode instead of reverse-charging the capacitor.

Remember that in normal defibrillator service, the discharge time is measured in whole milliseconds and the maximum current is only around 100 amps.   That's still much higher than the design current of microwave oven capacitors.  You don't have to worry about over-stressing the device if a lethal electric shock happens.

[FonziDaytona]

If my calculations are correct, I could be seeing a max of 46kA (assuming low ESR). I haven’t been able to find any diodes with ratings like that for under $2k or so :-(

I may just accept that these capacitors will provide some fun and then give up the ghost. Any ideas as to how long they’ll last?

How are DC link capacitors under these circumstances?

[klugesmith]

>> If my calculations are correct, I could be seeing a max of 46kA (assuming low ESR).

You can determine the effective R and L of your pulse power circuit, non-violently, by treating the system as linear.
Charge capacitor to 22 volts instead of 2200 volts, then close your contactor. 
With ordinary oscilloscope, you can get the capacitor voltage waveform. 
Damping and voltage-reversal percentage will be obvious. 
Discharge current is easy to compute.
Then multiply by 100 to predict the current of a full energy shot.

Nonlinearities will happen if full energy makes wires explode or coil-and-can geometry change.  I think they will generally reduce the stresses on capacitor, compared to linear model.

If your can crusher RLC values are like mine, there's plenty of damping to begin with.
Is 20% voltage reversal going to be hard on the capacitors?  I guess they are not electrolytic.

If there were a clamp diode, as I proposed, I bet its peak current and I^2-T integral are a small fraction of those in the capacitor and load. Because of the damping.

[FonziDaytona]

Looks like I have a weekend project to go with my new ruby laser! I’ll get to it and report back.