timeaverage and sample(jitter)

Pnoise timeaverage will not produce the same results as pnoise sampled(jitter) for a driven circuit. 

Noise Type = timeaverage measures the average noise at all time points in the PSS fundamental (1 cycle).

Noise Type = sampled is for circuits with sharp transition edges.  Sampled pnoise measures the noise power at a specific event (time point on the waveform).

• Timeaverage phase noise and edge phase noise (sampled pnoise) evaluate two different types of noise, thus they are not expected to be the same.
• Timeaverage Phase noise evaluates the average phase noise along the waveform, while Sampled edge phase noise calculates the phase noise at a specific time point defined using the jitter event.
• However, pnoise=timeaverage phase noise can be equal to pnoise=sampled Edge phase noise under specific assumptions

• • The output signal has very small and negligible AM noise and additive Phase noise.
  • The output large signal is a pure sinusoidal waveform.
  • The above assumptions are typically valid in oscillators with pure sine wave and negligible AM noise (often, AM is small when looking at near-carrier phase noise)

best regards,

Tawna

I know sampled jitter and timeaverage are different, so I add a limiter(comparator in ahdlLib) when I run timeaverage, AM noise hall be eliminated. The result shows AM noise level is way much smaller than than PM noise(50dB difference). 

From this moment, I think pnoise=timeaverage phase noise can be equal to pnoise=sampled Edge phase noise. Does the output signal have to be pure sine wave?

If this is a driven circuit, timeaverage phase noise will not equal sampled(jitter) edge phase noise.

The only time you will see ttimeaverage phase noise = sampled(jitter) edge phase noise is when you have a sinusoidal oscillator and negligible AM noise.

I filed a Case with Customer Support on your behalf.  Has the Application Engineer gotten back with you?  (Spoiler alert:  They're going to tell you the same thing. :-) )

best regards,

Tawna

What if I need to check the phase noise spectrum around 1st harmonic with both rising edge and falling edge?

For example, I am working on a LO circuit.

Hi Tawna,

My email is niu.yaoqi@3peakic.com.cn, so AE can reach me out.

Regards,

Lu Lv 

Hi Lu Lv, 

I've escalated your Case so an AE should be reaching out to you in the next 24 hours.

best regards,

Tawna

I think a clock with jittery rising edge could be modeled as below. Sampled(jitter) pnoise analysis would yield n(t).
sin(wt + n(t) (cos(wt)+1)/2)
When n(t)<<1, it can be approximated as below.
sin(wt + n(t)/2) + n(t)/4 + cos(2wt) n(t)/4
Time average pnoise analysis around the fundamental would yield n(t)/2. The remaining is at baseband and around the second harmonic.