Can small signal analysis be used for large signal operation

[katrin]

small signal vs large signal

This question has confused me for a long time.

I want to do some theoretical analysis for a CMOS amplifier, just like to derive some formulas of its input impedance, gain .....with Gm, r0 and so on.

But I noticed that normally such kind of analysis with transconductance Gm, output resistance r0 are defined for small signal models. But my circuit is actually working in large signal range, because the input signal is not small signal any more.
On the other hand, the circuit analysis I've seen were all done in small signal way, for large signal operation people just simulate it.

So I am wondering for such case could I still apply small signal for circuit analysis?
If not, how can I analyse the large signal operation of circuit in a theoretical way?

 

[LvW]

small signal analysis

This question has confused me for a long time.
I want to do some theoretical analysis for a CMOS amplifier, just like to derive some formulas of its input impedance, gain .....with Gm, r0 and so on.
But I noticed that normally such kind of analysis with transconductance Gm, output resistance r0 are defined for small signal models. But my circuit is actually working in large signal range, because the input signal is not small signal any more.
On the other hand, the circuit analysis I've seen were all done in small signal way, for large signal operation people just simulate it.
So I am wondering for such case could I still apply small signal for circuit analysis?
If not, how can I analyse the large signal operation of circuit in a theoretical way?

Hi KATRIN,

the so called small-signal analysis is a linear analysis, that means it assumes that (around a fixed dc bias point) the circuit can be regarded as sufficient linear. Therefore, all the mentioned parameters - gain, transconductance, Rin, Rout,... - also are linear small signal parameters. As soon as the signals are so large that the assumption of linearity is not true anymore, you cannot speak about voltage gain, output impedance,.....
In this case, it is important what you are interested in: waveforms, distortion, amplitudes, power consumption,....
Mathematical calculation of these nonlinear characteristic parameters is problematic and rather involved. Best method: Measurement resp. simulation.

 

[erikl]

small and large signal analysis

how can I analyse the large signal operation of circuit in a theoretical way?

Mathematical calculation of these nonlinear characteristic parameters is problematic and rather involved...

Very true. And that's exactly what SPICE and other SPICE-like simulators are doing. Their mathematical methods and its usage are public, and can be understood and comprehended (also not so easily sometimes, though ;-) ). Actually, the math behind this type of simulation describes - and calculates the solutions of - the theory which KATRIN may be looking for.

 

[leo_o2]

small signal analysis vs large signal analysis

Small signal analysis can't be used for large signal.
Large signal can be also analyzed with characteristic like slew rate.

 

[sreevyas]

large signal analysis

con't be used, one operating point is in linear region other is saturation region

--sree

 

[hyc]

Re: Can small signal analysis be used for large signal opera

I still think SPICE (or SPICE alike) by all means is to verirfy what you think W/L should be, and not to design W/L for you. So if there is no way to mathematically analyze the large signal behavior, are we forced to be a SPICE monkey, meaning are we shooting at the dart randomly and hopefullyl by chance we will get it right eventually, or if there is indeed a somekind of analytical way to estimate before submitting the job to SPICE for verification, for example, as leo_o2 pointed out, the slew rate is a large signal behavior but can be formulated.

 

[tdy]

dc shows all operating point, ac shows circut behaviour (trend) on one desired operating point. transient is the real word.