Finding gain of stage 1 in 2 stage opamp using stb analysis

Dear Amogh,

The stability analysis will not provide any information on the gain of the first stage of your multi-stage amplifier. The gain of the first stage will be impacted by the feedback action and potentially by any compensation you have included. You might try to infer it if the gain of the second stage is well known over some frequency range.

Shawn

Thank you Shawn. I am mostly interested in DC gain contribution by first stage and second stage. Is there a way to find out from stb sim? One way I can think of is breaking the loop with an inductor and running ac simulations. Not sure if there is a better way.

Dear Amogh,

AmoghD said:

I am mostly interested in DC gain contribution by first stage and second stage. Is there a way to find out from stb sim?

Not that I am aware of.

AmoghD said:

One way I can think of is breaking the loop with an inductor and running ac simulations. Not sure if there is a better way.

I don't know the specific circuit topology you are using - is your feedback signal AC only? In other words, will the DC operating point be stable if you break the loop? If so, you might consider the use of an AC resistor in lieu of an inductor when you break the loop. Since you will need a large inductance value, this can cause simulator/spectre issues. The analogLib AC resistor allows you to set a 0 DC value and very large AC value.

I can think of one other possibility, but it is more complex. Hopefully, this might provide some insight.

Shawn

Stability analysis (stb) definitely can't do this, and it's not what it's intended for.

One other possibility though is to do an AC simulation of the normal closed loop configuration, and given that ac analysis gives you the signal level at various nodes in the circuit in response to your AC input, you measure the gain as the ratio of the signal level at two different nodes (or the output of the first stage relative to the differential input stage). So something like this:

dB20((VF("/stageout") / (VF("/vinp") - VF("/vinm"))))

Regards,

Andrew.

Dear Andrew,

Andrew Beckett said:

Stability analysis (stb) definitely can't do this, and it's not what it's intended for.

Thank you for confirming my comment (not to mention taking your time to read my response and responding)!

Andrew Beckett said:

One other possibility though is to do an AC simulation of the normal closed loop configuration,

The reason I did not suggest this was that even in an AC based closed loop simulation, the gain of the first stage will be influenced by the feedback action. I was considering that  Amogh may be interested in the gain of the first stage without the impact of the feedback (since the feedback action can impact its load impedance and DC operating point).  This was further suggested to me by his attempt to isolate the output of the first stage with an intervening AC blocking inductor value. If I was incorrect in this assumption, your suggestion is quite reasonable.

I just wanted to add a comment to thank you and to clarify my comments to Amogh to prevent any confusion on his part.

Shawn

Thank you Shawn and Andrew for your responses. My goal was to see how much dc gain the first stage provides in the overall loop gain. I think feedback action matters for this. So Andrew's suggestion will mostly work. 

Shawn,

I'm not sure simulating the circuit open loop is an enormously useful thing to do - setting the operating condition closed loop and opening it this way. Since you never operate the circuit like this, it's questionable whether it tells you anything very useful (at least I'm not convinced it's useful). So even though the feedback loop has an influence, you're going to have this problem anyway, so trying to discount the effect doesn't seem that helpful to me.

Anyway, hopefully between us we've given Amogh some options!

Andrew.

Dear Andrew,

Andrew Beckett said:

Anyway, hopefully between us we've given Amogh some options!

Yes - that is all that truly matters - I agree!

To provide you a bit of background regarding my comments...I have occasionally found it necessary to run a set of open loop AC analyses for a closed loop system to debug start-up issues. Specifically, for those circuits whose start-up may result in multiple states of operation, I have used open loop AC analyses (often combined with a transient AC analysis) to verify the performance/robustness of a start-up circuit to place the circuit in the proper operating state.

Shawn