Does PNOISE sim of a sample & hold circuit include clock jitter effect?

Peter,

That's only the case if you are worried about large signal effects of the noise, which usually is not necessary. The noise in the system will have both AM and PM effects. Put another way (in a simple case), noise will cause small-signal variation in the position of an edge as well as small-signal amplitude variation. When worrying about jitter (or phase noise), we don't really care about the effect on the amplitude (if it were an oscillator, usually the AM noise is very small anyway because the oscillator naturally is amplitude limited; in some kind of switching system, you normally only care about the noise at a signal transition because it alters the decision point - anywhere else makes no difference) - so the various modes in pnoise allow you to investigate this. The jitter mode utilizes tdnoise automatically - it just places a sample at a threshold crossing, and then in the post-processing computes the noise as a jitter metric (e.g. in s/Hz^2 rather than V/Hz^2 or integrated equivalents).

So if you're considering the noise (or jitter) as a small signal phenomenon, you don't have to worry about the alteration to the Periodic Steady State.

For more details, see <MMSIMinstDir>/tools/spectre/examples/SpectreRF_workshop/JitterAN.pdf - also look at  Predicting the phase noise and jitter of PLL-based frequency synthesizers or the papers on cyclostationary noise here.

Regards,

Andrew.

Peter,

That's only the case if you are worried about large signal effects of the noise, which usually is not necessary. The noise in the system will have both AM and PM effects. Put another way (in a simple case), noise will cause small-signal variation in the position of an edge as well as small-signal amplitude variation. When worrying about jitter (or phase noise), we don't really care about the effect on the amplitude (if it were an oscillator, usually the AM noise is very small anyway because the oscillator naturally is amplitude limited; in some kind of switching system, you normally only care about the noise at a signal transition because it alters the decision point - anywhere else makes no difference) - so the various modes in pnoise allow you to investigate this. The jitter mode utilizes tdnoise automatically - it just places a sample at a threshold crossing, and then in the post-processing computes the noise as a jitter metric (e.g. in s/Hz^2 rather than V/Hz^2 or integrated equivalents).

So if you're considering the noise (or jitter) as a small signal phenomenon, you don't have to worry about the alteration to the Periodic Steady State.

For more details, see <MMSIMinstDir>/tools/spectre/examples/SpectreRF_workshop/JitterAN.pdf - also look at  Predicting the phase noise and jitter of PLL-based frequency synthesizers or the papers on cyclostationary noise here.

Regards,

Andrew.