Stability for small signal and large signal?

[jordan76]

Hi

What is the difference between small signal stability and large signal stability?
How to achieve large signal stability by compensation?
Is there any theory behind large signal stability?

Any inputs are welcome, thanks!

regards,
jordan76

 

[jiaming_lee]

Hi

What is the difference between small signal stability and large signal stability?
How to achieve large signal stability by compensation?
Is there any theory behind large signal stability?

Any inputs are welcome, thanks!

regards,
jordan76

examing small signal stablility is straightforward: just apply an AC simulation to check phase margin and gain margin.

However, AC sim assumes the unchaged DC OP (this is not true in real application), so it can not garentee the total 100% stablility. So, the large signal check is needed. Normally, connect the opamp into feedback mode as application does, then give DC and a very small sine wave to simulate the small signal; check the output. if no ringing, the opamp is stable in large signal. (monte carlo analysis would help to check the mismatch caused unstablility)

 

[jordan76]

examing small signal stablility is straightforward: just apply an AC simulation to check phase margin and gain margin.

However, AC sim assumes the unchaged DC OP (this is not true in real application), so it can not garentee the total 100% stablility. So, the large signal check is needed. Normally, connect the opamp into feedback mode as application does, then give DC and a very small sine wave to simulate the small signal; check the output. if no ringing, the opamp is stable in large signal. (monte carlo analysis would help to check the mismatch caused unstablility)

Hi jiaming_lee,

First of all, thanks for your reply!
Second, I have still two questions about large signal stability issue.
Q1: What do you mean by applying DC and a very small sine wave to simulate the small signal?
My guessDo it mean doing transient analysis by applying addition of DC(large) and a very small sine wave(how small should it be)?)

Q2: Can any insights be given from the results of the above analysis? For example, any simple equations possible?

regards,
jordan76

 

[jiaming_lee]

examing small signal stablility is straightforward: just apply an AC simulation to check phase margin and gain margin.

However, AC sim assumes the unchaged DC OP (this is not true in real application), so it can not garentee the total 100% stablility. So, the large signal check is needed. Normally, connect the opamp into feedback mode as application does, then give DC and a very small sine wave to simulate the small signal; check the output. if no ringing, the opamp is stable in large signal. (monte carlo analysis would help to check the mismatch caused unstablility)

Hi jiaming_lee,

First of all, thanks for your reply!
Second, I have still two questions about large signal stability issue.
Q1: What do you mean by applying DC and a very small sine wave to simulate the small signal?
My guessDo it mean doing transient analysis by applying addition of DC(large) and a very small sine wave(how small should it be)?)

Q2: Can any insights be given from the results of the above analysis? For example, any simple equations possible?

regards,
jordan76

Hi there:

Q1: you are right that apply DC at your input for simple bias-- for example the half supply, then add a small sine wave inseries with the input, then do the tran; this is just like you do the AC analysis--a DC OP and an AC source. However, the small sine wave should be small enough, so that the any of your transistos in the OPA would not go out of active region.

Q2: If your OPA is stable, the output should also be a sine wave with proper amplitude as you imagin; if the OPA does not stable, the output should touch the boundry of your pwer supply quickly after several cycles.

BTW: another way to exame large signal stability is:
put your OPA into your application connection---always negative feedback. then give a step at your input; check the output. If you find the rings---overshot are fewer than 5, the OPA can be stable in large signal.

 

[extraord]

but when the OPA is 2 inputs 2 output fully differential, and just directly wire connected feedback, how can I add the input signal without break the loop
to simulate the settling time?

Thanks

 

[jordan76]

HI jiaming_lee,

Thanks for your reply! I will click the helped me button for you soon

However, if the large signal stability is not met, is there any means(hopefully not too complex) to help me make design changes (adjust W/L etc) to improve it?


Hi extraord,

In your case of fully differential OP, the input signal may be a common mode DC signal in series with a very small sine wave.

regards,
jordan76