Extremely small currents...

Thread Starter

Ambrose Balin

Joined Nov 25, 2007
I've been considering possible methods to measure extremely small currents (~pA). How is this done? Would a current-voltage converter do the trick, or would the output voltage also be too small to measure?

The reason I ask is because I am a beginning physics major in my first circuit analysis lab. In our lab we don't have the capacity to measure such currents, but I know that they are used in industry (electron tunneling microscopes, for example). Based on what we've done in class, it doesn't seem like an Op Amp could be used here, and yet the measurements are made. I'm just curious, I guess.


Joined Apr 20, 2004
I do it all the time. It takes good noise control and a high input resistance in the measuring circuit. I use OPA134's with input impedances on the order of 5.0 X 10 ^13, and input resistors of 10^9 ohms. After that, it's just a matter of applying enough gain to amplify the signal for analysis. 10^-12 * 10^9 = 10^-3. A voltage gain of 1000 gives 1 volt out. I use gains up to 10,000.

If you're after more sensitivity, there are electrometer grade op amps with input impedances on the order of 10^15 ohms.

Keeping the experiment in a Faraday cage helps a lot.


Joined Dec 25, 2006
The OPA134 has a Vos of 2mV worst case, and with a Av=1000, that is 2V. This is not including Vos vs. temperature and time.
The input bias current is 100pA worst case which is a little high in my opinion dealing with really small currents; e.g., <10pA.
With a 1 giga ohm feedback resistor that would give you about 4 uV of thermal noise which will be amplified a 1000 fold by the next gain stage.
Besides, the OPA134 is really meant for audio given its specifications.
I would look at the OPA129 with an input bias current of 250fA worst case; however, the Vos is higher. Damn those FET inputs.
In both cases, you are talking some serious design and not a simple one or two stage amplification. Excellent layout, good shielding of the PCB, guard rings around the inputs (teflon standoffs??), input bias cancellation techniques, and good old fashion offset trimming.:cool: