Hi,
I am designing a limiting amplifier, and tried to simulate jitter using pnoise analysis. My set up is like this:
--beat frequency : 4G
--output amplitude 400mVp-p (differential)
--input amplitude : from 2mV-400mVp-p (differential) (ideal sine wave from psin)
--pnoise setting : sweeptype = absolute, noisetype=jitter,threshold value=0, crossing direction=all, sweep from 1G->10G
Jitter(Jcc) results : 1) vin=2mV --> Jcc[Second][k=1]@(3G,5G)Hz:rms = 18p
                           2) vin=4mV --> 9.05p
                           3) vin=200mV --> 0.263p
                           4) vin=400mV --> 0.191p
So, jitter results seems to be strongly dependent on input amplitude ?? Is it right ??
I also run transient analysis with different input amplitude(from 2mV-->400mV), and check phase variation of output voltage(zero crossing)
-- on rising edge : delta ~ 2.75ps
-- on falling edge : delta ~ 0.91ps (Input Freq is 4GHz)
I am not sure about the relationship between jitter number using pnoise analysis and AM->PM number using transient analysis as above?
Could someone help me to clarify this.
Thanks alot.
ha
P.S: when running pnoise analysis from 1G to 10GHz, in the output log window I saw pnoise sweeping from 1G --> 10G TWICE !! Why is that??

Dear Stefan,
I'm still confused. So, pnoise(jitter) does not take into account the effect of changing input level by itself,  unless you change it and simulate again??
When varying input from 2mV to 400mV, in transient analysis, phase error is about 2.75ps while pnoise(jitter) difference is 18ps-0.191ps ~ 18ps
Is there any relationship between those numbers ??
thanks a lot,
ha