hi

i have a question about using pnoisetype=source on a driven circuit.

i read on the menu that the pnoise noisetype=source simulation reports noise averaged over a period of the input signal(Question: can you comment if this is true? ). and i wanted to test it.

i have a driven circuit being driven at 200MHz and use the pnoise noisetype=source, the pnoise plot will give a certain plot. (case1)

if all remain the same while i lower the driving frequency to 100MHz and use the pnoise noisetype=source for simulation(case2), the pnoise plot is 6dB lower.

this seem to confirm that the same noise is spreaded over twice as long a period(Question: does this sound correct to you?)

however, when i loaded the noise summary for case1 and case2, they are 2.7 times difference
(Question: while i expected them to be the same, what do you think here?)

i then run case1 and case2 with noisetype=jitter expecting to find the same Jee, but case1 gives 455fs rms and case2 gives 322fs rms.

i think a driven circuit should give the same Jee independent of frequency( assuming it's operating properly)  

i appreciate any comment you may have. thanks.


Frank

You might not take into account all noise contributions. Try setting the maxsideband parameter to a large value (see also section 2.2.1 of http://www.designers-guide.org/Analysis/sc-filters.pdf) or use the fullspectrum option (only available in APS) together with a suitable value of the maxacfreq parameter.

I would expect roughly the same noise and jitter for both of your cases (if the average noise is not dominated by the noise during the signal transition and if the signal slope at the transition is the same in both cases). However, if you use the same value of the maxsideband parameter, the maximum frequency of the noise you take into account in case 2 is only half the value of that in case 1.

You might also want to ask your question at http://www.cadence.com/community/forums/33.aspx or contact Cadence support if this does not solve your problem.