It is widely assumed oscillators are cyclostationary (This appears in Ken's paper "An Introduction to cyclostationary noise" as well.)

Demir's famous paper however claims noisy oscillators are stationary. This makes a huge difference in modeling phase noise, so which one do you believe is correct?   ???

Thanks....Sohrab Samadian, UCLA

Thank you Ken, that was enlightening...

So in a typical RF circuit noise generated by oscillator is cyclostationary and time-average power spectral density can not be used to calculate noise power at osc. output.

Then, if there is a cyclostationary noise source present in an oscillator, how does one calculate output noise power to see it's effect on phase noise? It's not time average of PSD, any clue how to approach this?

I love this website!! In America nothing is free... seems like not in here though... anyway, thanks for your time

Sohrab,
The time average PSD combines both the AM and PM components of the oscillator noise. Generally in the range that is of greatest interest, the noise produced by an oscillator is almost purely phase noise. In this case, the time average PSD is the phase noise.

-Ken