Hi all  :)

In order to correlate PLL's phase noise with a measured one from E5052, I need to simulate the phase noise of an output logic divider with SpectreRF. Unfortunately, the logic divider biased from a noisy power would introduce much AM noise, which cannot be simulated with PSS+PNOISE by setting 'noise type=sources'. As such, I simulated its phase noise by 2 ways:
1) Setting 'noise type=modulated', I got a so-called PM noise (dBc/Hz) from it.
2) Setting 'noise type=jitter', I got a voltage noise spectrum (tdnoise) at the threshold voltage. Mutiplying this result by 2*pi*fo and dividing it by the slope at the threshold voltage, a phase noise curve is achieved.

By comparing the PM curve and the calculated one, there exists a large deviation between them and their shape is not unique either.
My questions are:
1) Which one of the 2 curves is the phase noise which could be compared with the measured result? Or I have made a mistake?
2) Does the PM noise could be used to describe the phase noise of a logic divider? Or it is only valid for oscillator?

Any reply is appreciated
Thanks!

Since you get only the normalized voltage noise of divider with the PNoise analysis, you have calculate yourself the phase noise with the equation (35) in the paper of "Predicting the Phase Noise of PLL-Based Frequency Synthesizers ".

But what make me confused is that, the result of normarlized voltage noise is very close to the phase noise. Who can explain that?