Jitter cacluation from phase noise??

I just found above pile of my questions which approved lately.

I really appreciate Frank's reply again.

Thank you.

 Hi to all,

 It seems I had the same questions as chadong.

Frank, you said to use PM simulation on autonomous circuits, but for an oscillator to determine Jitter shouldn't be used FM instead ? You also say that for autonomous circuits, only k-cycle jitter is meaningful, should be cycle-to-cycle jitter ?

 Thanks in advance.

Best regards,

Pedro

PM jitter simulation is more precise for signals that will trigger a threshold because you can specify a threshold level in the simulation setup. Jee is not a useful measurement for free-running oscillators because it is infinite. You can use both Jc and Jcc, depending on what is more meaningful for your application.

 Hi Frank,

 Thanks for the reply. So for an oscillator, that usually has the noise concentrated around the carrier, FM should be a better choice, don't you think so ?

 And one more question, what is your opinion about phase-noise/jitter performance in weak inversion when compared to strong inversion ?

Best regards,

 Pedro

No, I don't think so. Regarding your second question, this is a forum about how to set up simulations with Cadence tools. For circuit design questions, you could try for example http://www.designers-guide.org/Forum/.

Hi Frank,

Thanks again for your patience.

1 -I just would like to know how to get accurate values for jitter, because the change quite a lot if I touch in the integration limits.I did as you said 1 Hz - fo/2.

2 - My plot of phase noise is relative to the 1st harmonic. So should I measure this relating to the carrier or to the first harmonic ?

3 - I don't know why, but my 1st harmonic is so small (like 2 or 3 dB), that you can't compare it with the rest, I suppose.

4 - This kind of analysis is supposed to be done in dBc/Hz, that is the plot should not be linearized, correct ?

Best regards,

Pedro

Please take a detailed look at http://www.designers-guide.org/Forum/YaBB.pl?num=1224609785, this should answer most of your questions regarding jitter simulations. Sampled noise is periodic with the pss fundamental (and symmetric to half of the pss fundamental), so the harmonic you specify should make no difference.

Hi Frank,

Thanks a lot for your answers.

I am sorry but I have one more question for you:

-  When I simulate Phase-Noise (PSS+Pnoise) and then I simulate Jitter, if I plot Phase-noise on both cases they are completely different. To plot the value of jitter I plot phase-noise first, but this plot doesn't match the one generated by phase-noise analysis. What is the reason for this ? Which one is correct ?

Thanks a lot.

Best regards,

Pedro

You don't tell very precisely what your simulation settings are in both cases, but when you "simulate jitter", you probably have sampled noise, whereas when you "simulate phase noise", you probably have unsampled noise. Both are correct of course, they simply represent something different. Please see http://www.designers-guide.org/Analysis/sc-filters.pdf for details.

 Hi,

 Ok I will giveyou some parameters.

To simulate Jitter I have set PSS to shooting mode, 80 MHz (beat frequency), 10 harmonics, conservative simulation.

For PNOISE, relative harmonic to 1, the sweep from 1 - 100 MHz, 5 points per decade, 30 sidebands,Jitter as noise type and FM.\

I have selected then Pnoise Jitter and plotted the phase noise.

For Phase Noise I think is the same, but insted o jitter is selected sources.

 Thanks,

Pedro