I'm currently building a filter circuit that is required to remove an DC offset from the raw signal and then amplify the gain. The rest of the circuit (not shown) is a 2-stage LPF with extra gain.

My problem is that when I add a HPF (fc = 0.002Hz) (or any HPF for that matter) at the output after the buffer, I get an offset of -10V. I've tried other op-amp models (LM741, LM324, etc) and the offset occurs but of a much lower value depending on the model. What opamp specification should I look at in the datasheet to determine why this is occurring?

The opamp I'm using is the THS4012. The input is at 100Hz and is shown as the yellow signal.

The green and red signal are the respective outputs after the buffer.

enter image description here

  • \$\begingroup\$ Try measuring the DC voltage across one of the input resistors (3.36M). The op-amp specs you will want to be looking at are: input current, input offset current. \$\endgroup\$ – Dwayne Reid Oct 1 '19 at 3:41

Dwayne is correct. Because you want a very low frequency filter, the resistors and capacitors have to be very large. All real op amps have currents that flow into or out of their input terminals. The large resistors cause this current to create a DC offset on the op amps. You want to use an op amp with a CMOS, or perhaps JFET input stage, so the input current is small. Lowering the resistor value is good, too, although then the capacitors get even bigger. It is possible that adding a resistor of 3.6 meg in series with the (-) terminal of the op amp would help, but that is not always the case, and the big resistors add noise.
In summary, you have chosen a very inappropriate op amp for this application: it has large input currents because it uses bipolar transistors (it is designed for high bandwidth.) Choose an op amp with the adjective "precision" in its marketing description, this will be more appropriate for low frequency work.

  • \$\begingroup\$ Okay, I've looked into this more in detail and it seems that the THS4012 has bias current is bad for my application of adding a passive high pass (with a low cutoff). Thanks for your help and Dwayne. \$\endgroup\$ – kurcio Oct 1 '19 at 6:40
  • \$\begingroup\$ I do have one other question. The other two stages I mentioned in my original post is a 2-stage LPF. That is infact a 2-stage FDA (THS4131 model) that has LPF with multiple feedback topology. When I connected the HPF before the FDA, the DC offset was not as strong, despite it having the same input bias current. Am I right in assuming it is because of the extra passive components in making up the LPF that this decreases the resistance as seen by the opamp inputs, and so the dc offset is not as high, coupled with the fact that is is an FDA so some of the offset error cancels each other out? \$\endgroup\$ – kurcio Oct 1 '19 at 6:43

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.