Hello everybody,
I would like to simulate the phase noise of a divider.
To have an idea of the noise at the ouput of the divider I first tried to
do a pss + pnoise simulation. Then I divide the ouput signal given from the pnoise simulation by the rms value of my output.
My signal is a square of 1,2v differential (so 2.4v pp), so the result I get  in dBc is
noise=20 log10(out/Vrms).
According to a recent message I saw on "Estimate pk-pJitter using eye diagram in Spectre" the noise I get should be higher than the noise I simulate with the Pnoise strobed analysis.

That's not the case.
in a simple test circuit
the noise with standard pnoise is -164.5dB
with the strobed Pnoise analysis the output noise I get is -147dB


to get this noise in dBc I apply the previous formula.
with standard Pnoise I have -162dBc
for the strobed analysis....
I don't know how to do ...
does somebody have an idea?

Thank you a lot.

Hello Ken,
in fact, I read a reply saying that most of the time , the noise not involved directly in the jitter was dominating the measure of the standard Pnoise simulation.

here is the post :
"If you use standard PNoise rather than the strobed PNoise,
then you are including noise that occurs at times other than
when the signal is crossing its threshold. Noise produced when
the output is all the way high or all the way low does not contribute
to jitter, but often dominates the noise that standard PNoise measures."


In my case, the noise is higher with the strobed analysis...
I suppose I should take the -147dB and not the -162dB...(bad news...)
I also doing a DLL in parralel.
I measured the noise with a standard Pnoise analysis. In that case, I suppose it is a little bit different as the AM noise is also quite important. Am I wrong?
I read a post about noise simulation with DLL and the guy was using the same method as for the divider, I mean the strobe Pnoise analysis.
What is your feeling  about this ?

Thank you a lot for your time.
Have a nice day
                                   Horizon

Hello Ken,
Thank you a lot for your answer. I will use the strobe Pnoise analysis for my dividers.
I still don't understand something...sorry...

You wrote
> If you were to plot the noise power as a function of time during >a cycle, you would see that the noise would peak at the >transitions. So if you were to periodically sample the noise >during the transition, you would be sampling at the peak, and >the values in the resulting noise sequence would be large.

>If instead, you sampled the noise when the signal was either >high or low, the noise power in the sampled sequence would be >lower because there is less noise at that time.

So if the standard Pnoise analysis gives the average of the noise for one cycle.
if there is a peak at a specific point and then something lower, the average of the noise should be lower than the noise given by the strobed Pnoise analysis.
In the simulation it's the case.

but  you wrote
>However, if you look at the continuous time noise, you have to >consider the total noise power over the cycle. Here the noise at >the transition may contribute only a small amount to the total >noise power because the transition only lasts for a short period >of time. The power contributed while the output is stable (either >high or low) might be much greater even though its amplitude is >smaller because the output is stable for most of the cycle.

I thought the standard Pnoise analysis output was the average of the noise on one cycle ? Am I wrong ?
In the previous text I copied from your post, I understand that the standard Pnoise  output is giving the sum of the total noise and not the average.Is it the case ?
If yes, I would like to get the phase noise of my divider in dBc.
What I did is :
I took the value given by the strobed analysis and I divide it by the power of my carier.

with standard Pnoise I get -162dBc
with the strobed Pnoise I get  -144.5dBc
The is an huge difference, but I suppose I should take into account the -144.5dBc.

Than you again a lot for your time.
                       Horizon

Horizon,
     When referring to power, the total power and the average power are the same thing. They both represent the total energy delivered over the length of a cycle divided by the length of the cycle.

-Ken

Hi Everyone,

I am a new user to the STROBED PNOISE function. The posts on this thread are very helpful. The problem I've been having with this is simply getting pnoise to simulate the noise at the right points. In the pnoise form, I specifically add the threshold crossing points but when I plot the results, those specified points never get added. The timeindex for each point always starts at 0 (and so on) but never include my points. Does anyone know why that is?

Thank you in advance for any responses.

Frank, thank you very much for your helpful response. We'll update our software and try again.

-Andrew.