As part of my studies, I am designing a measurement setup for measuring leakage current of SMD capacitors in the ranges up to pico Amps.

At the moment I am in research periond and later I will have access to some high end measurement devices such as a picoammeter, a SMU and a low leakage relay switch matrix.

What I have in mind at the moment is soldering 10 SMD capacitors on a cartridge and connect each capacitor to the SMU and picoammeter through the switch matrix

One of the concerns, as I approached here, Is thermocouple effect. If I know that my measurement performance should be in nano and pico Amps range, would this make a huge issue? I am asking this because when I go to PCB design phase, I would have to consider using connectors and PCB routes of same material (Copper).

Thermocouple Effect

For example, considering the connection between the test fixture and SMU, picoammeter and relay matrix should be done using coaxial or triaxial (the later can be the best option, but I do not have at the moment any idea of how economic will it be, considering the co to tri converters and theire own price).

So my question, in short, is that should I be worried about this effect having in mind reaching picoamp measurement?

Please let me know if I am not clear enough with the question.


2 Answers 2


The thermocouple effect should be no issue to you. It is very small as Jason already pointed out, but more importantly it causes a voltage offset. Your system is measuring current. A small voltage offset in series with the current source should be irrelevant.

As Jason also said, you do need to think about leakage current carefully. A guard trace around the high impedance node will be necessary. You may also have to use reed relays so that you can place the magnetic part yourself to insure extra low leakage from the magnetic drive circuit to the contacts. Check the datasheet of the relay you plan to use. It probably either doesn't say or doesn't promise leakage to the coil connections that low.


How far apart are the relay contacts? The only reason you'd have to worry about thermocouple-induced voltages is if there is a significant temperature difference between the two; intentional thermocouple coefficients are in the range of microvolts per degree C.

I would be much more worried about leakage currents on the surface of the PCB. If you're measuring picoamps, you need to keep your traces very short and may need to use guard rings and/or have slots milled into the board to keep leakage currents from flowing between your sensitive circuit nodes and other sources of voltage.

  • \$\begingroup\$ Thanks for reply. This is the datasheet from the relay matrix I am going to use. Unfortunately it does not mention the distance between the relay contacts. But it is a decent one! keithley.com/data?asset=427 it is keithley 7158 current scanner card \$\endgroup\$
    – Dumbo
    Oct 3, 2011 at 12:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.