Hi Ken,

I believe that I found an error in equation (33) of your paper "Predicting the Phase Noise and Jitter of PLL-Based Frequency Synthesizers" (Verison 1). In my opinion, each S_phi_k in the sum would have to be multiplied by T_k^2 and the sum would have to be divided by T_K^2.

Let me explain: In my opinion, what you really want to add up here is timing jitter. Phase noise as defined by equation (32) is timing jitter multiplied by the angular frequency of the fundamental. Thus, for equal amounts of timing jitter, you will get higher phase noise for higher signal frequencies. In order to convert the phase noise back to timing jitter, you would have to divide it by the angular frequency of the signal at that stage or multiply it by T_k/(2*pi). (I am talking here about noise "amplitude", the square root of noise power.) The factor 2*pi of course cancels out with the one in omega_K=(2*pi)/T_K that you then use to convert the timing jitter back to phase noise.

Best regards,

Frank