Hi all,

I am doing a Phase-Noise on a Pierce XTAL oscillator. I did it on Cadence using PSS + PNOISE. I have exported the data to Matlab and now I want to convert this data into Jitter. How can I integrate the Phase noise curve ? Please bear in mind that I am using log units, so dBc/Hz vs frequency.

Thanks in advance.

Best regards,
Pedro Cardoso

Hi,

Thanks for your answer, but I have already done that and the result is not what I was expecting.
In this case I used the cslculator, and the function 'integ' to integrate. Once yy the units of the are dBc/Hz and for xx axis Hz (frequency), should I do any manipulation of the data prior to integration ? That is, should I convert to linear units the plot, and only then calculate the integral ? As the phase noise is given in dBc/Hz, I guess I sould calculate the power of the carrier and add it to the yy axis.

Does this make any sense ?

Best regards,
Pedro

Quote:
which is the PHASE JITTER / Long Term Jitter.

If you want Period Jitter, then scale with an appripriate weighing function & integrate.
As to what weighing function you wanna use, there have been several metrics from different theoretical physicists lk Demir, Zanchi, etc. which are basically related to sinc function.[FT of a rectangular time window]

About Cadence's analog artist calculator, it provides the values for long term as well as short term period jitter correctly provided you are feeding the correct data.

--M

Than you very much to all of you.

Cheers,
Pedro

Dear all,

I have still some problems with integration. I want to calculate Jcc jitter. So I am doing a PNOISE + PSS simulation. But what gives me some doubts about the integration limit is that the Jcc varies a lot, due to the harminc's presence (I think so). So my question is, can I integrate till one fraction of fo/2 ? Which range would give me the most accurate result ? I think that at fo/2 I am couting the noise of the harmonic. Is this correct ?

Best regards,
Pedro