HV Optocouplers

[bullet308]

Hi:

Could somebody tell me something about optocouplers? More specifically, are there any optocouplers out there that could trigger on, say a momentary 400-1400v 10uA surge on the input side and survive?

[ivan747]

I don't think many will survive those conditions if they are reverse biased. Maybe a filter can dissipate the surges and a high breakdown voltage diode could help to protect it. After all, the impedance of the surge is too high to destroy a filter.

[Psi]

I think you'd have to use a fet/transistor to take the 1400V signal and switch on the optocouplers led.
But i'm guessing you want to avoid a powersupply for the input side.

[bullet308]

Allow me to add some details (I was in a hurry when I threw the question up earlier).

The "surges" in question are not incidental or environmental, but an integral part of the devices function, the device being a Geiger counter.

For those that are not familiar with how Geiger counters work, there are essentially two sides to the electronics package: the high voltage side which loads the Geiger-Mueller tube at 400-1400v and a low voltage side that detects when the high voltage side fires upon the tube being penetrated by a ray or particle of ionizing radiation. The low voltage side then counts it, averages it, turns it into a LED blink or a  "click" in a speaker, whatever you want it to.

Usually, there are some HV diodes and other circuitry involved in keeping the high voltage side from leaking over and blowing up the low voltage side, but a common failure mode in some Geiger counters is for that circuitry to break down, resulting in blown transistors and such,

I was beginning to think in terms of using an optocoupler for that purpose as a way of making the HV and LV sides completely isolated from one another. If there is zero electrical interface between the two, that cant happen any more, can it?

Anyway, that is what I was thinking. Thanks for the feedback.

[mikeselectricstuff]

I have seen optocouplers with a neon as the emitter used in telecom applications, not sure of they still make them.
However I don't see that it should be much of a problem in a geiger application, as the grounds can be the same between the HV and LV, and protecting a low-level signal from HV just needs a high value resistor and zener.
However you do need to pay attention to details like the voltage rating of the resistor, so in practice you'd use a few in series.