Measuring very high resistors

Sometimes condenser microphone capsules fail, because they get polluted and show a certain resistance between the diaphragm and the backplate. This leak resistance is very high, in the order of giga-ohms (I assume).
What I would like to do, is make a measuring device that is able to measure the resistance of a microphone capsule.
This is what I have in mind: use a FET OpAmp, for example a good-old TL071 as a voltage follwer. (Output connected to inverting input). The non inverting input to ground via a 1 Giga-ohm resistor. The output can go (through a resistor) to a measuring instrument. The 'resistor under test' would be connected between the + of the power supply (symmetrical power) and the non-inverting input of the OpAmp. In this case calibration could be like this: full scale = 1 G.Ohm, half scale = 2 G.ohm, 1/10 of the scale = 10 G.ohm. Would this work to measure the leak resistance in a capsule?
(During a quick test I have measured the 'resistance' of several screwdrivers  ;D , and some are better than others!)

from the ti data sheet  typical input bias current is 65 pico amps (200 pA max @25'C) with  5pA of offset  current (100pA max @25'C). So worst case at 25C of 200pA +/- 100pA.

Worst case over the full temperature range is 7nA +/-2nA  (note the C graded parts are worse with 7nA +/- 10 nA)

The circuit should work reasonably at room temperature. At worst case full temperature range not so much. 

The 1G resistor to ground will convert the input current to voltage before the test resistance is connected.

JR

PS: I borrowed a fluke insulation resistance meter to perform safety tests on a project and it only indicates up to 550M full scale @ 500V

Well, it might be that it isn't 'scientifically' correct, but as far as I can see it is good enough to get a rough indication.
It would also be possible to add an 'offset compensation', to null the output, without a device under test connected.
I don't need an exact value anyway. Just an indication if there is any significant resistive 'leak', or not.
The fact that I could get an indication of the insulation values of different screwdrivers (!) makes me think that this simple setup might work for what it is intended for.

I think that it is a very cleaver idea. Many of the tools we use may not be State of the Art, however they are a great indicator that helps or leads us to a solution to our quest for knowledge.

I think the hand oil on the screw is some of the leakage.

Duke

Audio1Man said:

I think that it is a very cleaver idea. Many of the tools we use may not be State of the Art, however they are a great indicator that helps or leads us to a solution to our quest for knowledge.

I think the hand oil on the screw is some of the leakage.

Duke

Click to expand...

3 terminals measurement should ignore any surface dirt reading.

JS

RuudNL said:

Well, it might be that it isn't 'scientifically' correct, but as far as I can see it is good enough to get a rough indication.
It would also be possible to add an 'offset compensation', to null the output, without a device under test connected.

Click to expand...

I pondered the possibility of using a dual or quad TL07x op amp, so you could perhaps get a reference current fron one of the other similar (same temperature) op amps in the package.  But your circuit with no resistor being sampled will reveal it's input current error. 

I don't need an exact value anyway. Just an indication if there is any significant resistive 'leak', or not.
The fact that I could get an indication of the insulation values of different screwdrivers (!) makes me think that this simple setup might work for what it is intended for.

Click to expand...

It should work OK at room temperature. I am making ASSumptions that the offset and bias current increase at high temperature (I do not know that for a fact). You could test for yourself by alternately heating the IC with an iron tip held nearby, or cooling it with some spray cold.

My only concern is that the worst case input currents could swamp out your measurements, but that erroneous condition will be apparent even with no resistor being tested.  A no /low leakage reading will be more trusted, than a hi leakage reading that might have op amp input current added in.

JR