I have to perform stability analysis on an op-amp (AD8671) which is being used as an op-amp in the circuit attached. I am using LTSpice to perform this. My approach is to give a 1V AC small signal at the input and see the frequency response at the 'testpoint' which should, if I am correct, provide me with the gain and phase of the transfer function. For this, I am trying to adjust the conditions so that I am able to create critical points, with (Gain > 1) and (Phase Shift = 180 degrees). I also need to check the effect of the snubber circuit attached to the node, as theoretically, with improper values of the same, the response should be unstable.

However, as of yet, I am unable to effectively see this, primarily because the gain is always <-160 dB as shown in attached image, in both open loop and closed loop conditions. Also, if I try to perform a transient analysis, the simulation is jittery, often getting stuck after a point and simulating at xyz nanoseconds/sec, which is obnoxiously slow.

My questions are:

  1. Is my approach correct?
  2. Is my circuit and script implementation, correct?
  3. If the answer to either of the above is no, what should I do/change?
  4. How do I speed up the transient response? (if possible)

Attached are relevant images of circuit and its frequency response. Any help would be appreciated :)

reference schematic

ltspice circuit

.ac directive's output

New contributor
Riyan Dahiya is a new contributor to this site. Take care in asking for clarification, commenting, and answering. Check out our Code of Conduct.
  • 1
    \$\begingroup\$ Before you get further you first need to pick a FET model. The default model has non practical parameters such as a zero threshold voltage. Right click the PMOS and choose something in the list that has an RDSon and VDSmax that’s close to what you’re trying to use in real life. \$\endgroup\$
    – Ste Kulov
    Commented Jul 10 at 1:30
  • 1
    \$\begingroup\$ The component values in the LTspice schematic are not OK yet. With Vset=4V there are already 5V from a total of only Vreg=9V lost across the 500 Ohm sense resistor R2/3. The theoretical 10mA of constant current would generate 10V across your 1k load which is not possible if you have only 4V headroom left or in general from only a 9V supply. The MOSFET with the huge 10k gate resistor is causing a major delay inside the feedback loop of a fast op amp. I don't see the current source being stable or how the Zobel network is going to improve on that. Is this a circuit that works IRL? \$\endgroup\$
    – Raonoke
    Commented Jul 11 at 8:48


Your Answer

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