I need to measure current through a microcontroller at high side, with about ~1 µA resolution, ranging from 1µA to 25mA. The result is sampled by a microcontroller and sent to a computer or an on board display. The sampling speed I need is about 500kHz.

The problem I am facing is that the linearity of the gain when the current drawn is in the one digit µA area is quite unlinear. Say my amplifier circuit (consisting of simple op-amps) has an amplification of 800, the actual amplification is around 1000 at 1-3µA, and closer to 800 at 9-10µA.

I guess this is due to the power drawn by the operational amplifiers themselves in some way, so I tried to preamplify the signal with LTC6102 first and then feed this into my opamp circuits (one circuit per range of current), but still I struggle with non linear amplification even from the LTC circuit.

I have been in contact with Linear, and they propose a buck boost converter with integrated coloumb counter, but this is I2C and cannot give me the speed I need for sampling.

Are there any examples of circuits that can actually measure these tiny currents while still maintaining the sampling speed? I don't have the schematics right now, but can provide later. I just had to ask right away cause I cannot get this problem out of my head.

  • \$\begingroup\$ There are opamps with nA leakage, why do you need anything more than 1k resistor and a buffer? \$\endgroup\$ – Gregory Kornblum Oct 13 '15 at 19:42
  • \$\begingroup\$ In fact, i would use ad8222 with high gain. \$\endgroup\$ – Gregory Kornblum Oct 13 '15 at 19:51
  • \$\begingroup\$ @GregoryKornblum I should have been more specific. I only have a 5V source, and the ad8222 need at least +/- 2.3 V. Can I put the negative input to ground and still maintain a good result from this chip? \$\endgroup\$ – chwi Oct 14 '15 at 19:29
  • \$\begingroup\$ Yes, it will work just fine. But on remember, it's not rail to rail neither on input nor on output. Which means the range is a little narrow on both sides. On the other hand, you will have perfect differential measurement with very high gain. \$\endgroup\$ – Gregory Kornblum Oct 14 '15 at 19:33

The µCurrent Gold claims a bandwidth in excess of 300kHz. To get higher than that you'd have to either characterize/bin the individual op amps used or replace them with an even more expensive part.

  • \$\begingroup\$ I've actually thought about copying this but haven't gotten to it yet. Will definitely try it out, thanks. \$\endgroup\$ – chwi Oct 13 '15 at 20:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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