PSS, PNOISE and Transient Noise

Pedro,

To try to answer your questions:

1. You can either import the results into Matlab using the "spectre toolbox" for Matlab (covered in the documentation), or in ViVA you can use File->Save from the results browser to save into Matlab form.
2. Presumably you could compare them in Matlab?
3. Transient noise is including the same as a small signal noise source. If using noisefmax only, you get white noise at noisefmax across the whole spectrum. If you specify noisefmin you get coloured noise (i.e. flicker noise) from noisefmin upwards. Yes, the noise is up and down converted. Bear in mind that you get the signal containing both the signal and the noise.
4. PSS+PNOISE are completely independent from transient noise. PSS+PNOISE will be a much more efficient way of simulating the noise in a periodic circuit - and probably more accurate - unless the circuit has a large signal response to the noise.
5. It depends! Very hard to know without looking at the circuit. Why do you want to use transient noise rather than PSS+PNOISE anyway?

Regards,

Andrew.

 Hello,

Thanks for your answers.

Relating to point 4, so if I choose either transient noise analysis or transient analysis without checking the transient noise analysis, the result of the Phase-Noise/Jitter calculations will be the same on the two cases, right ?

Relating to your question in point 5, I was using transient noise to calculate the value of jitter in Matlab (both Jcc and Peak-to-Peak) and than to draw an histogram.

I have as well calculated the Phase-noise under Cadence (as described in the manual), and I have used another approach integrating the waveform (being the units linear). The value of RMS Jitter that I have obtained is completely different from the calculated one and the one given by the Cadence simulation

Regarding the circuit, is a basic Crystal Pierce oscillator (three-point oscillator), being the crystal modeled by an RLC series.

 Thanks a lot,

Pedro

Pedro,

The results of the pnoise will not be affected by whether you are running transien tnoise or not.

Note that integrating the result to obtain the jitter is not necessarily correct - because you need to be aware of the fact that there is a log relationship, plus the fact that if you're doing multi-cycle jitter you can only integrate over the band which you've simulated in the pnoise analysis. I suggest you take a look at the Jitter app note that is in <MMSIMinstDir>/tools/spectre/examples/SpectreRF_workshop . I don't know how you've calculated it or what you're comparing with.

Regards,

Andrew.

 Hi Andrew,

 Thanks a lot. I will take a look at the Jitter app.

 Best regards,

Pedro Cardoso

I have the same problem with you. The integral noise obtained by matlab is different from that obtained by cadence. Do you know why now?thank you

community.cadence.com/.../jitter-simulation-pss-pnoise-and-matlab-get-different-results