Ultra low leakage overvoltage protection (TVS,...) for picoampere measurements.

[Fabse77]

Hi everyone,

i'm building a low current measurement setup which can use voltages up to 2500 V.

So I have a high voltage precision lab supply, and an electrometer like a Keithley 617 or some other pA capable SMU (see attachment). Now I need to protect the Electrometer from the high voltage if the device under test breaks down. The Keithley for example allows a max input voltages of 250V.

At first I was thinking about Zener diodes or TVS, but they have all way to high leakage and because they are parallel to the electrometer they should be extremely low leakage.
There is a device from Tektronix, especially made for this purpose: https://de.tek.com/datasheet/model-2290-pm-200-protection-module-users-guide

with the promising spec: Leakage: < 10 pA at 200 V typical
but it costs around 1300 USD. And I wonder what they are using to achieve this value. The schematic on the case looks like it only uses a TVS diode.
Does anyone have an idea of a protection device for Vbd~200-250 V with such low leakage current? I think a gas discharge tube should work too, but datasheets usually only say insulation resistance > 10GOhm. How good are they really?
Thanks

[Gyro]

I did some experimentation with GDTs when I was building my simple Picoammeter. I found that standard 230V telecom GDTs had a leakage of several pA at 10V. The ceramic ones were better than the glass ones. A 1kV ceramic one came in at about 1pA...

https://www.eevblog.com/forum/projects/picoammeter-design/msg798921/#msg798921

I went with making an air spark gap in the end.


P.S. Welcome to the forum 

[sdouble]

yes, you have many option in the sub-10V range... but with 200V you are not in the same ballpark.

[David Hess]

You might be able to do it using collector-base junctions of high voltage transistors as low leakage high voltage diodes but a gas discharge tube or even several neon bulbs in series would be a sure thing.

[Fabse77]

 Thanks for the feedback.

I ordered some GDT from different manufacturers and will do some leakage current tests. I will share my results here when i'm done.

[Alex Nikitin]

If you only use the electrometer for current measurements, a high voltage capable resistor in series with the input and BAV199 dual diode to ground is all you need, providing the voltage drop on the resistor is acceptable in the current range used. It would be better still if your HV supply is sensibly current limited (say, at 0.1mA).

Cheers

Alex