Yiming
Guest
|
Ken,
This question might be dumb but I am somewhat confused. In the matlab script of your jitter modeling paper, you integrate the jitter using function cumsum() before you use psd() to obtain the power spectral density of phase noise. What is the theory behind the integration?
Assume the output voltage of the oscillator is x(t) = A*cos[wc*t + theta(t)] = A*cos(wc*t) - theta(t)*A*sin(wc*t), then the RMS power <x(t)^2> = A^2/2 + A^2/2 *<theta(t)^2>, where <theta(t)^2> = integral(Sphi(f), -inf, inf), and Sphi(f) is the phase noise PSD. Sphi(f) again equals to the Fourier transform of the auto-correlation function of phase noise <theta(t+tau)*theta(t)>. If the phase noise is white, it would results in Sphi(f) is white, too. ("an approach to consistent jitter modeling for various jitter aspects and measurement methods", Masashi Shimanouchi)
What am I missing here?
|