how to get the jitter of a pll

[yxh12321]

hi all:
would anyone teach me a more accurate method to get pll's jitter?
I used to get jitter through eyediagram in cadence or cmoscope. but I find a problem .through this way,I get peak-to-peak jitter which I think is not very accurate to show the performance of a pll. because peak-to-peak jitter is pretty related to the time scale. I have read some paper,find RMS jitter is more reasonable. so would someone kind-hearted do me a faver?I would be very appreciate!!!

BTW,compared with phase noise ,jitter is better to show the performance of a pll? or the other one ?

 

[biff44]

If you have access to something to measure the phase noise (like a good spectrum analyzer), the phase noise measurement tells you more. For instance, you can deduce the PLL control loop parameters, and see stability issues in the phase noise data.

I think people using digital methods (dsp, adc) tend to think in terms of jitter, while people using modulation (communications systems, radar systems) tend to think in terms of phase noise.

 

[yxh12321]

If you have access to something to measure the phase noise (like a good spectrum analyzer), the phase noise measurement tells you more. For instance, you can deduce the PLL control loop parameters, and see stability issues in the phase noise data.

I think people using digital methods (dsp, adc) tend to think in terms of jitter, while people using modulation (communications systems, radar systems) tend to think in terms of phase noise.

hi biff44, thank you very much for reply.
I follow my teacher to study PLL which is used for digital methods. My teacher let me find a more accurate way (than peak-to-peak jitter)to judge a PLL with hspice.
So would you teach me more about how to measure RMS jitter with hspice?
And someone says hspicerf can get me the noise of VCO, PFD+CP+iref+LPF,devider, then put these noise in Matlab , and use the transfer function to compute the RMS jitter . I got the noise of VCO in hspicerf, but I can't get the noise of PFD+CP+iref+LPF .It confused me for a long time . please help me . I would be very appreciate

 

[AdvaRes]

Hi yxh12321,
Simulate your pll phase noise and by integrating, you can calculate the rms jitter

 

[haykding]

Hi, you shoul simulate each block in PLL and after sum all together . Or you can write VerilogA code and measure jitter (PLL SIM).

 

[yxh12321]

hi AdvaRes,
thank you very much for your reply
I don't know how to simulate the noise of PFD+CP+LPF with hspicerf ,could you tell me more details ? or you can give me some paper about this.
I would be very appreciate

---------- Post added at 10:30 ---------- Previous post was at 10:27 ----------

hi fubsyman,
thank you very much
and my problem is I don't know how to simulate the noise of PFD+CP+LPF with hspicerf. Could you help me again?

 

[haykding]

hi fubsyman,
thank you very much
and my problem is I don't know how to simulate the noise of PFD+CP+LPF with hspicerf. Could you help me again?

Let's distinguish two type of jitter : RANDOM and DETERMINISTIC . So what type of jitter you should measure ? If deterministic , I can say this : you should create testbench where this three devices will be together connected, and despite this use sin voltage source on supply , do tran simulation, and measure jitter with Cosmos Scope .

 

[yxh12321]

hi fubsyman,
thank you for your reply
I don't quite understand you . And my problem is how to get the RMS jitter of a PLL, or how to get the noise of PFD+CP+LPF.Could you tell me more details?

 

[mark_lhm]

in this CPPSIM tool , this part PFD+CP+LPF (DETECTOR) require the dBc/Hz ( X/Y ) but not the V^2/Hz (dB) , I am confused how to transfer the simulation output noise to phase noise ???

---------- Post added at 04:25 ---------- Previous post was at 04:22 ----------

the vco PN is easy to get !
you can connect PFD two input port together using the same input pulse (the locking pulse frequency normally)
and then you can check the LPF output port , to see the output noise !
that is the way , you can get the PFD+CP+LPF output noise !

---------- Post added at 04:28 ---------- Previous post was at 04:25 ----------

you can use the SpectreRF tool by using the PSS plus Phasenoise function to simulate !
after you get the VCO and Detector (PFD+CP+LPF) phase noise (dBC/Hz)
you can use the CPPSIM (BY MIT) to get the RMS jitter in some frequency range !

---------- Post added at 04:29 ---------- Previous post was at 04:28 ----------

Normally you should set the LBW and operating FREF frequency and M value in this tool!

 

[yxh12321]

hi mark_lhm,
I am very appreciate for your detailed reply!

"in this CPPSIM tool , this part PFD+CP+LPF (DETECTOR) require the dBc/Hz ( X/Y ) but not the V^2/Hz (dB) , I am confused how to transfer the simulation output noise to phase noise ??? "
This problem is the same as I encountered when I use Hspicerf .It has confused me for a long time . The unit of Y-axes is V^2/Hz !! But it is dBc/Hz in a paper .I followed the paper and I don't know why they are different. Could you please help me again?

 

[mark_lhm]

Firstly , I use the PSS plus PhaseNoise to check the output noise of the PFD+CP+LPF , and I can get the (dB)
And then I choose the output item is "phase noise " but not the "output noise "
As you know the VCO phase noise output can get the dBc/Hz
But as I comment , I am confused whether the Detector should used the phase noise item ???
It is my appreciation if anybody can give any good suggestion for us here !