I used Behzad Razavi's book to design a fully differential telescopic op-amp [screenshot attached]. Then I follow the same steps to design my proposed opamp, but I am not getting the desired result. Can someone point out where I'm doing mistakes?

For your reference, I have attached hand calculation and simulation screenshots. Thanks

My design specifications are:

  • Supply voltage: Vdd=5V
  • Output voltage swing: 5V
  • Power ≤ 5mW
  • Gain: 500
  • UnCox= 151uA/V2
  • UpCox: 26uA/V2
  • Vthn=0.84V
  • Vthp: |1V|

enter image description here enter image description here enter image description here enter image description here enter image description here


2 Answers 2


You can't (easily) test an opamp like that in an open-loop configuration; you need feedback circuits to set the common-mode VOUT range and the (small) differential input signal.

Once you simulate with that configuration, you can extract the gain and other parameters from (VOUT+ - VOUT-)/(VIN+ - VIN-) etc.


A couple of things I can see right away.

  1. Im9=.7mA. Ivb1=Ivb2=Ivb3=.1mA I assume those are drain currents. The input devices should both be .7mA/2 = .35mA

  2. I don't why you are using full scale square waves as input to test. They will swing non linearly to the rail whether your gain is correct or not. You should try starting out with very small sin signals (like 10mvpp) to see if your circuit is working.

  3. check that your DC vout are properly centered for bias (i.e. VDD/2 is always a good bet). From your simulations it looks like it is hitting a rail, meaning it probably is not.

  4. I agree with jp314 that you should really have a common mode feedback to set your output dc and test with AC signal. You will easily see your AC is probably 0 or negative and not what you expect without proper bias. If you are trying to test with transient signals, you should have some feedback to control the gain. But you may have even more problems getting to that level without first addressing your bias design.

This shows you can set the bias without using cmfb. Problem on biasing of the differential folded cascode with ideal and non-ideal CMFB


Your Answer

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

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