The Designer's Guide Community Forum https://designers-guide.org/forum/YaBB.pl Measurements >> Phase Noise and Jitter Measurements >> Query on Philosophy behind obtaining the Phase Noise spectrum from T-domain https://designers-guide.org/forum/YaBB.pl?num=1486878574 Message started by subtr on Feb 11th, 2017, 9:49pm

 Title: Query on Philosophy behind obtaining the Phase Noise spectrum from T-domain Post by subtr on Feb 11th, 2017, 9:49pm Hi,We know that the jitter of an oscillator accumulates because of its feeding back of an accumulated random phase due to any kind of noise. Let us take a sinusoidal oscillator for simplicity. The device noise which is stationary in nature affects the oscillator based on the time of injection governed by the impulse sensitivity function. There are ways to probe the phase in time domain. Assuming I take huge number of points so that my accuracy till some low frequency(Say 10KHz) is good enough. Theoretically I need to take infinite number of points to capture till DC. Now I break down the measurement into two cases :Case 1 - Infinite Cycles & 1 sample per cycle : Take each cycle's zero crossing. Compare it with Kth cycle of an ideal sine wave without jitter. We know that this difference would continue to grow without bounds because oscillator remembers everything. But we are sampling one point per cycle which will also show an accumulating behaviour as well. In fact we are looking at a point where the ISF is the same at every cycle Q1 . Why zero crossing and not some other crossing where ISF is non zero?Q2 . Am I averaging/Low pass filtering the effect of ISF? OR it doesn't matter? Q3 . Now I take PSD of the (sampled points minus ideal points) normalized to the sampling frequency. Does this PSD represent the exact phase noise spectrum?Case 2 - 1 Cycle & Infinite points : Now if I just take one cycle and take all points, I still get inifinite number of points. Say I compensate for the effect of already defined ISF at each point and now compare it with an ideal single cycle. ISF is not a dirac delta.Q4 . Do I have complete statistical data to represent the oscillator's phase noise spectrum?Q5 . If Yes, how do I get the phase noise spectrum from this data?Q6 . If No, what part of the phase noise Frequency spectrum do I have in hand?