What is Strobed Periodic Noise simulation in SpectreRF

The tdnoise option (at the bottom of the pnoise form, where it says noise type - default is "sources") is the strobed periodic noise analysis.

Note that for jitter you can also pick the jitter noise type; this uses tdnoise (for a driven circuit, or for an oscillator if you pick PM jitter rather than FM jitter) by identifying the points to sample the noise based on a threshold crossing.

Regards,

Andrew.

 Dear Andrew,

Thanks a lot for your reply.

I understood the modulated and jitter option.

But what is tdnoise option ? How it is differnt from the other two?

Can you please throw some light on it.

Kind Regards,

You could run cdnshelp and search for  tdnoise ...

At a high level, it inserts an ideal sampler at the output of the circuit which samples the noise at the time instants you specify; you can either give specific time points during the period, or ask it to output every nth point (using skipcount) or specify the total number of time points you want. At each of these time instants, it does a pnoise. This allows you to see how the noise varies throughout the period, rather than computing the time-averaged noise over the period (which is what normal "sources" mode does). If you know the noise at a switching transient, and also know the slope of the signal at the same time, the jitter can be computed - this is what jitter mode does for you.

Regards,

Andrew.

 Dear Andrew,

Thanks a lot for your reply.

After reading through spectreRF(latest manual), I have following confusions:-

1:-

If I enter total number of timepoints let's say 4 and the PSS total timepoints=400, then tdnoise takes which timepoints ?

2:- For a stand alone Oscillator relative harmonioc is set to 1, for doing pnoise analysis using Modulated, Jitter and tdnoise.

But for driven circuit, the relative harmonic is set to 0 ( while sweeptype is kept relative) for Modulated,Jitter and tdnoise.

This I am not able to understand. Why this differnce from Oscillator.

What I am thinking is:- Let's say Modulated PM analysis for driven Circuit. The Phase noise is due to  fc+fm & fc-fm components.

Where fc is the center/output  frequency ( PSS beat freq) of the  driven circuit & fm is onepoint of the sweep frequncy.

So the output noise frequency is calculated with respect to ( relative) to fc. So relative harmonic shoould be 1 NOT 0.

But then why relative harm=0 set in the document for driven circuit.

Kind Regards,

1. It spreads the 4 timepoints evenly over the period. 
2. With an ideal sampler when using pmjitter or tdnoise, all noise gets folded into the 0th harmonic. So you might as well use that. It actually doesn't really matter which you pick, because they'll all end up being copies of each other - but I'd pick the 0th. With the "modulated" PM analysis, you're simulating two pnoise sweeps, one above and one below the carrier and then it combines the results smartly to give you the PM and AM components - so it has to be done around the 1st harmonic.

 For answer to second query, I am not able to understand the first sentence of it.

I think what you are trying to say is for driven circuit (let's say a simple buffer is driven by a square clock of frequency fc), the relative harmonic No. entry doesn't matter. i.e. if you enter 0 or 1 or 2 or anything else doesn't matter for  jitter and tdnoise analysis.

But for modulated PM analysis it should be 1.

Am I right on this ?

Kind Regards, 

Yes.

Andrew.