BNC overvoltage protector?

[Circlotron]

I want to use a high voltage probe to test the open cct output of various automotive ignition coils. Most would be in the 35kV max range. I can find a probe that goes to 40kV but occasionally I would need it to go as high as 50kV. I Don't want to risk my nice new scope in case the probe flashes over internally, so is there such a thing as an overvoltage protector with say a pair of end-to-end zeners in a nice little BNC female to male? I could always get a T-piece and put the zeners on a connector on the branch but if there s already something available that would be nice.

[rastro]

Isn't optical isolators typically used to protect test equipment from higher voltages. 

[johansen]

a google search for inline bnc surge suppressor and similar search terms brings up a variety of products that range in price from one left on Amazon for 3$ to the same junk for 32$ plus shipping.. and similar looking products for twice the price.

I would evaluate exactly what you need to do before buying something when you can make your own from the 400 volt MOVs found in a handful of products commonly thrown out in the nearest dumpster. or just buy some 150 volt MOV's for 2$.

[alm]

I would expect any BNC surge suppressor to be designed for 50 ohm or 75 ohm connections. Connecting 50/75 ohm hardware to a high-Z scope probe designed for 1 Mohm input impedance will affect the signal and may increase the capacitance of the input beyond the compensation range of the probe.

Consider the failure mode of the HV probe. The probe presumably attenuates the signal by 1000x, so even 50 kV would only be 50 V at the scope input, a level that most scopes should be able to handle safely. The problem with exceeding the probes voltage limit is that the attenuator of the probe may arc or short. In this failure mode the full 50 kV would reach the scope input. So anything designed to handle failure of the probe should be able to handle whatever voltage you expect at the input. And I wouldn't trust the BNC cable or hardware to handle 50 kV either, so better keep that far away from anything you don't want it to arc to, like your body or other electronics.

[c4757p]

Personally, I'd find a cheap-ass piece of crap 20 MHz scope and keep that around for things like this. Then who cares if you break it? And if you do decide you care, you can probably fix it.

[alm]

If you go that route then I would recommend to get a scope with a tube (valve) front-end rather than transistors. Tubes are less sensitive to high voltages than FETs, although I wouldn't count on either to handle 50 kV.

[c4757p]

Or just throw some clamp diodes into the frontend, if you're not incredibly concerned with bandwidth and accuracy. Besides, you're not going to find a tube scope nearly as cheap as a solid-state one, and you're screwed if the tubes go tits up.

[Circlotron]

Perhaps a gas discharge tube voltage arrestor might be a good choice.
http://au.mouser.com/ProductDetail/Littelfuse/CG90L/?qs=sGAEpiMZZMsdWQmBKEIh8oMcc%252bFReu0rJCMQg2J7WxE%3d

[johansen]

There are typically 3 or 6 gas discharge tubes on the PCB on the back of most CRTs, they are about the size of a 1/4th watt resistor.

I would expect them to arc over at about 200 volts.
My 1000:1 tek P6015 oscope probe has a neon light bulb inside the probe for this purpose.

[hofig]

Solutions like MOVs and zeners can have problems with too much leakage and capasitance. I've used gas discharge tubes. Depending on the tube and probe you may run out of adjustment range of probe though.