The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Simulators >> RF Simulators >> PSS vs PAC
https://designers-guide.org/forum/YaBB.pl?num=1285625422

Message started by andyt on Sep 27^th, 2010, 3:10pm

Title: PSS vs PAC
Post by andyt on Sep 27^th, 2010, 3:10pm

If I run a PSS analysis with one large (FLO) tone and one small signal tone (FRF) and then plot H1 (harmonic mag at FLO-FRF) and the 3'rd order intermod product, both vs PRF (RF signal power) should I get a curve similar to the corresponding PAC curves? It seems that I should provided the small signal is sufficiently small.

Andy T

Title: Re: PSS vs PAC
Post by andyt on Sep 28^th, 2010, 10:01am

Hi Pancho,

Let me rephrase the problem. If I run a PSS analysis with one large (FLO = 1G) tone and one small signal tone (FRF = .8G) and then record the magnitude of the harmonic at the frequency slot F= FLO-FRF=1.0-.8=.2G. If I then repeat this type of run for different magnitudes of the small signal tone (PRF), I can make a plot of the value of the first order harmonic at 0.2G vs PRF. That plot is, in principle, the same as the similar plot created in a typical PAC plot of the first order magnitude vs. PRF. My question is "Should these curves be the same if the small signal magnitude in the PSS runs is sufficiently small?" The same question would apply to the 3'rd order intermod plot.

Thanks for your help.

Andy T

Title: Re: PSS vs PAC
Post by Ken Kundert on Sep 28^th, 2010, 4:48pm

The words you are using have very specific meanings that you may not understand.

A small signal to the simulator is by definition 'sufficiently small', meaning that the simulator always linearizes the circuit before applying that signal, and so the computed output is always proportional to the specified input.

Small-signals are only present in small signal analyses, such as AC and PAC.

Every tone in a PSS analysis is considered by the analysis to be large.

So, given this terminology, I am having a very hard time interpreting your question. Are you asking a question about a PSS analysis, where all tones would be considered large, or are you asking about a combination PSS/PAC analysis were you apply only the LO in the PSS analysis and apply the RF in the PAC analysis?

-Ken

Title: Re: PSS vs PAC
Post by andyt on Sep 29^th, 2010, 11:27am

Hi Ken,

It is possible I am misusing the terms as I do not have an RF simulator, so let me rephrase it somewhat.

Suppose I do two runs:

Run#1:
I run an ordinary PSS analysis. In this analysis I have one large signal, say 1GHz at 0 dBm applied to the LO port of a mixer and one small amplitude signal (say .5GHz at -30dBm) applied to the RF port. A Fourier analysis yields the value of the signal at the IF port and I record its dBm value at the mixer conversion frequency (1-.5 = .5Ghz). I then increment the magnitude of the RF signal by +2dBm and repeat until I reach PRF=0 dbM. I take the results and I plot the dBm value of the IF signal vs the dBm value of the RF signal. I get a straight line with a slope of roughly 1.0, at least until compression sets in.

Run #2:
I run another PSS with only the large signal applied and then a PAC analysis. I plot the signal at the conversion frequency vs the RF power.

My question is: Shouldn’t I expect the curves from the two runs to be similar? Isn’t this what PAC does so efficiently?

One final question. I assume PAC is not really AC at all, but a type of transient analysis run in which the circuit equations are somehow linearized about the “periodic operating point” of the necessary PSS run. If this is so, what difference does the PAC amplitude make except to scale the results?

Are there any papers you can recommend on the theoretical basis of PSS and especially on PAC?

Thanks

Andy T
:)

Title: Re: PSS vs PAC
Post by Ken Kundert on Sep 29^th, 2010, 3:04pm

Okay, I understand what you are asking. The two situations are the same as long as the RF signal remains small enough so that it does not cause the circuit to respond in a nonlinear fashion and so long as the RF signal is large enough so that it is not lost in the noise floor of the Fourier analysis. So what you will see when you are simulating with the pure PSS approach is that at very low input signal levels the output signal may get lost in the noise, but if it does then at some point as you increase the input amplitude the output will start increasing in proportion to the input until you start compressing the RF input, at which point the slope of the output will drop. PAC will give results that are the same as the middle region without the starting region (numerical noise) or the ending region (compression). And it will do so faster and more accurately that PSS analysis.

This is described in:
http://www.designers-guide.org/Analysis/intercept-point.pdf

You can get background into PSS and PAC analyses by reading:
http://www.designers-guide.org/Analysis/rf-sim.pdf

-Ken

Title: Re: PSS vs PAC
Post by andyt on Sep 29^th, 2010, 3:29pm

Hi Ken,

Thank you. Your description makes perfect sense.

Andy T. ;D