0
\$\begingroup\$

I am working on a photodiode transimpedance amplifier based on an OPA2374 and decided to simulate it in TINA TI.

However, several unexpected issues arose.

  1. AOL and GBW product vary significantly with input common-mode voltage and supply voltage
  2. Degraded stability when photodiode generates a DC current

Here I do a sweep of the input DC offset (so as to not saturate at the negative power supply) and plot the open loop gain of the op-amp.

This is the schematic I used: enter image description here

And these are the results: enter image description here

Similar results are obtained when sweeping the negative voltage supply: enter image description here

I must stress that all the ranges I swept through are well within the input common-mode range and supply voltage ranges for the OPA2374.

Also, as a further sanity check, I redid the circuit with all the test fixtures and the same op-amp macromodel in LTspice. The results and behavior were identical.

However, when changing just the op-amp to one already present in the LTspice device library, I could not observe any variation when doing sweeps of either DC offset or supply voltage.

What's happening here? Is this a case of a misbehaving op-amp model or something else?

\$\endgroup\$
3

1 Answer 1

1
\$\begingroup\$

I haven't looked at the datasheet, but I wouldn't be surprised if the loop gain changes due a different common-mode input voltage. They will define how hard you drive the differential input stage (the overdrive voltage) and this can change the small signal parameters of your input stage, such as the transconductance.

As far as I know there exists a "Boyle" macromodel for op-amps which typically includes an actual transistor input-stage, so changes in common-mode voltage can be reflected on the simulations.

The input stage is a great contributor to all the open-loop voltage gain, as such, it also dominates the noise performance of the op-amp.

The question is, does this variation in the low-frequency open-loop gain be tolerated by your design? Does a 20dB decrease of gain at those low frequencies important to you?

EDIT: On a second thought, perhaps 20dB might be a bit too much of a variation for an op-amp. Maybe there's something not properly model. However, I think you should be able to account for this in your design.

\$\endgroup\$
11
  • \$\begingroup\$ Yeah, I don't know. 20dB difference is a lot. \$\endgroup\$ Feb 14 at 10:14
  • \$\begingroup\$ And I also found this issue being raised before here: e2e.ti.com/support/amplifiers-group/amplifiers/f/… \$\endgroup\$ Feb 14 at 10:15
  • \$\begingroup\$ I think this difference might affect the stability, the feedback capacitor is dimensioned for a certain GBW product. \$\endgroup\$ Feb 14 at 10:16
  • \$\begingroup\$ @Virgil_Tibbs but use it in your actual design. Does it make a difference in your final design? If you can frequency compensate it, you can make it good enough to make it stable in both situations. Is the distortion coming out of the amplifier good enough for you in both situations? \$\endgroup\$
    – Designalog
    Feb 14 at 10:20
  • \$\begingroup\$ @Virgil_Tibbs you should also consider whether that common-mode variation is expected in your system. Where's that voltage coming from? \$\endgroup\$
    – Designalog
    Feb 14 at 10:24

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.