The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Measurements >> Phase Noise and Jitter Measurements >> phase noise simulations for buffer/dividers
https://designers-guide.org/forum/YaBB.pl?num=1386865117

Message started by Ryan Cheung on Dec 12th, 2013, 8:18am

Title: phase noise simulations for buffer/dividers
Post by Ryan Cheung on Dec 12th, 2013, 8:18am

Hi all,

Recently, we had a meeting with cadence guys about how to simulate the phase noise for a buffer/divider. They suggest to use PM to do the simulation. But years ago, when I began to simulate the buf/div phase noise, I got much info saying the right method should be the jitter analysis and I’ve used the jitter method for years to now. This noise is a strobed noise rather than the time averaged noise, so the jitter method should be the right one. But they still think we should use source/PM for the simulation because the E5052 just measures the phase rather than jitter. So I’m much confused. Would you pls tell which one is the right one?

Thanks a lot :D

Title: Re: phase noise simulations for buffer/dividers
Post by Ken Kundert on Dec 12th, 2013, 11:54am

Generally you choose the noise post-processing options to match the way the noise will be observed. Presumably in the circuit the output of the buffer will be observed by some thresholding circuit such as a mixer or digital gate. In this case, jitter is the most interesting thing to observe because that is what the observing circuit is sensitive. Noise that occurs away from the threshold will not effect the observing circuit.

If you are trying to match a measurement with a spectrum analyzer, then it becomes the observing circuit and you need to match it. I took a quick look at the E5052, which appears to be an Agilent signal source analyzer. It claims to be able to measure phase noise and jitter, but it is not clear to me how it is doing it. Jitter assumes a threshold, but I did not see a thresholding circuit in their block diagram. Measuring phase noise generally requires a phase detector, but I did not see that either.

So, to summarize, you choose the style of noise post processing to match the characteristics of the observing circuit. You can think of the block being driven by the block being measured as the observing circuit, or if you are trying to match a measurement, you think of the measurement equipment as the observing circuit. Jitter is probably more appropriate for the former, and it is unclear what is best for the latter without a better understanding of the test equipment.

-Ken

Title: Re: phase noise simulations for buffer/dividers
Post by Ryan Cheung on Dec 12th, 2013, 5:25pm

Hi Ken,

Thanks for the reply. allow me get more info on equipment first and talk later.

-ryan

The Designer's Guide Community Forum » Powered by YaBB 2.2.2!
YaBB © 2000-2008. All Rights Reserved.