|
The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl Design >> RF Design >> Difference between SP and PSP analysis https://designers-guide.org/forum/YaBB.pl?num=1157833342 Message started by jobless on Sep 9th, 2006, 1:22pm |
|
Title: Difference between SP and PSP analysis Post by jobless on Sep 9th, 2006, 1:22pm Hi, Could someone please help me understand better the difference between SP and PSP analyses. I am unable to completely understand when they say in SpectreRF manual that "PSP is a small-signal analysis similar to the conventional SP analysis, except, the circuit is first linearized about a periodically time-varying operating point as opposed to a simple DC operating point." I am unable to understand what is linearization about a periodic waveform and what is a periodically time-varying operating point. Since the waveform in the steady-state is a large signal thing, the outcome of PSP or SP such as Z11 are all time varying inside a period. Do you average it out? And therefore, depending upon the precise moment in a period, which is chosen for the linearization, the outcome of PSP and SP will vary. My question is which point in a steady-state time period is chosen for this purpose. Say, I have an oscillator circuits (traditional LC-VCO) and I perform both SP and PSP analysis across the two oscillating nodes,that is,across the LC Tank. For this I put the psin port across the Tank with very large input resistance to avoid changing the Q of the tank. Is this method correct? Regards -J |
|
Title: Re: Difference between SP and PSP analysis Post by ACWWong on Sep 9th, 2006, 4:02pm jobless wrote on Sep 9th, 2006, 1:22pm:
SP is small signal s-parameter analysis calculated around a DC operating point. That's it. The DC operating point is a linear-time-Invariant operating point, and does not allow frequency translation. PSP is small signal s-paramter analysis calculated around a PSS operating point. The PSS run is a period steady state, which is a periodically linear-time-Varying operating point which allows for frequency translation. jobless wrote on Sep 9th, 2006, 1:22pm:
So SP has nothing to do with PSS. The SP analysis result will depend on the DC solution (which is done automatically, but can be forced follow a given DC operating point/nodeSet if you set the options up for this). PSP uses the full PSS solution as its operating point. So the PSP result will depend on the PSS result. jobless wrote on Sep 9th, 2006, 1:22pm:
If you want to measure the differential small signal negative resistance, then a SP analysis is ok. You can use any value for the port impedance you like, it'll make no difference to the SP result (other than scaling the smith chart, as it doesn't change the DC operating point). The simulation results will show the S11 being "outside" the smith chart meaning unstability/oscillation. Plotting something like Re(ZM) will give the value of the negative resistance. Do remember the result is only a small signal result around the DC solution. One could offset the tank nodes with a real DC offset and redo the SP analysis to show the small signal negative resistance at some "snapshot" point in the swing. If you are attempting to use PSS/PSP to do the same thing, then I recommend you simulate the circuit "driven" by the port (i.e. set to sine and given the ocsillation frequency and voltage swing you want that would normally ensue in free running transient/ ocsillator mode PSS) to get the PSS, then do the PSP sweep. hope this helps with your understanding and best of luck with your circuit. aw PS. Please also note that using a balun and two ports (differential and common-mode) is a more sound way to measure differential circuits. see http://www.designers-guide.org/Analysis/diff-sp.pdf http://www.designers-guide.org/Analysis/diff.pdf |
|
Title: Re: Difference between SP and PSP analysis Post by jobless on Sep 11th, 2006, 10:41am Dear Wong Thank you very much for your detailed reply. It was certainly helpful. Another question that I failed to mention in the mail was the comparison of outputs of SP and PSP and their physical interpretations. Consider mag(Z11), the peak of PSP, Z-parameter analysis. It is much more sharp and the peak value is about 100-200 times more than that of SP. Can the process of linearizing about a periodically operating point than just using DC operating point make such drastic changes. As we know that voltage swing of a traditional VCO is roughly Idc*Rp(4/pi*Idc*Rp to be precise). One can calculate the Rp(Tank impedance at resonance) using SP or PSP analysis. Also the Q of the LC tank can be estimated as Wo/BW (W0 as freq of osc and BW as bandwidth of the tank impedance function). If we were to use PSP(hoping that it will be more accurate due to linear-time-variant analysis), the Q value will of the order of thousands. Also the Rp, Z11 at the resonance, is around 100-120Kohms. The output of SP is definitely quite closer to the reality. In this light, how do I interpret the Z-parameter results of PSP analysis. Please help me. Regards -J |
|
Title: Re: Difference between SP and PSP analysis Post by ACWWong on Sep 12th, 2006, 2:54am jobless wrote on Sep 11th, 2006, 10:41am:
No. If you are doing a negative resistance oscillator, then the swing is not simple ohms law, or even with a 4/pi square wave modifier.... I suggest Tom Lee's book "the design of CMOS radio frequency Integrated circuits" or equivalent. Tank losses do come into play in limiting the swing, but so does change of negative R due to swing, finite slew, L*di/dt etc. If you are adding the port across the output of the LC-VCO tank, then you will hope to see a net negative resistance in SP (on a balanced DC operating point). The PSP will deviate from the SP depending on the swing because of things like non-linear capactiances etc. but errors of 100 ~ 200 times means something is either going on in your circuit or your PSP setup is wrong. cheers aw |
|
The Designer's Guide Community Forum » Powered by YaBB 2.2.2! YaBB © 2000-2008. All Rights Reserved. |