S-parameter and power gain relationship

[mohamis288]

if you consider the diagrams of RFICs, you find out that the "power gain-output power" or "output power-input power" diagrams have nonlinear relation. so is it correct to perform S-parameter simulation to obtain power gain using S21 parameter?
I mean, in S-parameter simulation, we do not define input power. so according to what parameters, ADS performs s-parameter simulation? in my opinion, it is better to give an input signal (pulsed RF, sine wave , ... ) with certain power and then measure output signal. in this way, we can calculate power gain associated with the specific input power. is it correct? I have attached the diagrams.

 

[BigBoss]

S-parameters are Small Signal Parameters and there is no Relationship Between those Graphs.
They have been obtained by driving with Large Signals and Measured at the Output as Powers.
You cannot use s-parameters for this purpose.

 

[mohamis288]

S-parameters are Small Signal Parameters and there is no Relationship Between those Graphs.
They have been obtained by driving with Large Signals and Measured at the Output as Powers.
You cannot use s-parameters for this purpose.

thank you @BigBoss
I got my answer but can you explain more about the application of S-parameters especially S21 ( which is famous as a factor for gain assessment) and S11 (which is famous as a factor for loss assessment) ? in both case, we do not have any information about input parameter, so why we use them as criteria for gain and loss?

 

[BigBoss]

Scattering parameters - Wikipedia

en.wikipedia.org

 

[volker@muehlhaus]

application of S-parameters especially S21 ( which is famous as a factor for gain assessment) and S11 (which is famous as a factor for loss assessment) ?

As BigBoss already mentioned, S-parameters are small signal parameters. So for small input power and linear operation, S21 is indeed the power gain of your amplifier. If you increase input power, you will see saturation effects and gain will decrease. This is what you see in the curves.