About S22 parameter of PA in transmitters?

HI,guys,I met a confused problem in designing PA. The impedence of antenna load is 50 ohm, and simulated S22 of the PA is closed to zero, but the output power is accepted(more than 20dBm). I think if the output power is ok, it will be OK even if S22 of the PA is bad. Also, bad S22 is even helpful to PA beacuse it rejects the signal coming from the antenna. Is what i said right or not? Hope for your advices!

NO. If your S22 is bad, that means it's not standard 50R load, but the antenna is designed as input port is 50ohms, so the true power input to the antenna is less than 20dBm, and the antenna performance is not so good as designed.

For Power Amplifiers, s-parameters are not used.Altough, 20dBm is not sucha high power but optimum load impedance is more convenient to match the power amplifier to the antenna.
S-parameters are valid for a small signal operating point only.That's why, measured s-parameters are not very meaningful..

I think, if the antenna is 50-ohm perfectly matched, it does get the 20dBm power; if unmatched, the radiated power depends on the phase of the antenna and PA's S11 as the return wave from the antenna will be reflected back to the antenna. Normally A low S22 can benefit a PA. I guess: 1. maybe the S22 you measured is not a true S22 as high non linearity involved such as a C-class amplify without exciting. 2. The transistor has a high output power margin. It can be powerful enough even far away from the optimize working point.

Thank you all! Sorry that I think i have missed something for making the problem clear. Here S22 is the return loss of the amplifier which has included the matching networks, and the impedence of antenna load is 50 ohm. The output power is got by PSS simulation, which is more than 20dBm. I have no problem with that S-parameters are valid for a small signal operating point only. So I feel that PSS simulation should be more convictive, which means that I'd like to preserve my point above. How's your point? Thanks!

Are you operating the PA in small signal conditions to feed antenna?

No. The input power is about -10dBm to -5dBm. So PSS simulation is more suited for PA than S parameters. I think so.

Depends on your PA size...
And could you translate PSS? I am not familiar with this acronym. Thanks

PSS simulation is Periodic Steady-State simulation. I wonder if it is similiar to HB simulation. Anyone can tell me the difference between these two methods? Thanks.

HB is a one of analysis method for "PSS".

Generally, "PSS" is implemented by two methods, HB or Shooting-Newton.

So if you use a term of "PSS", you have to make clear what analysis method you use as "PSS".

Which do you use as "PSS", HB-PSS or Shooting-Newton-PSS ?

The Designer's Guide Community Forum - Very curious about optimum impedance from load-pull in PA designs

Usually we use Shooting-Newton-PSS in Cadence and HB in ADS. Both HB-PSS and Shooting-Newton-PSS work with large signals, right? But I do not know the difference between the two methods. It would be wonderful if someone can tell me the difference, haha. Thanks anyway!

Not correct.
Currently HB-PSS is more standard than Shooting-Newton-PSS even in Cadence Spectre.
The Designer's Guide Community Forum - Transmission-line oscillator: problem with PSS
Tip of the Week: Guidelines for simulating oscillators - phase noise simulations - RF Design - Cadence Community

Right.

See http://www.designers-guide.org/Analysis/rf-sim.pdf

Thank you! I think i should spend some time to read the reference! Thanks again! haha