The Designer's Guide Community
Forum
Welcome, Guest. Please Login or Register. Please follow the Forum guidelines.
Mar 29th, 2024, 5:25am
Pages: 1
Send Topic Print
Phase noise of a driven circuit from Spectre's frequency-domain jitter analysis (Read 1601 times)
raczko
New Member
*
Offline



Posts: 4

Phase noise of a driven circuit from Spectre's frequency-domain jitter analysis
Feb 29th, 2012, 7:05am
 
Hi,

I have a question with a long history of analysis Smiley

You have a driven circuit (say, buffer) and would like to know its phase noise contribution to the system.
Using Spectre, you should use the 'jitter' engine and not the analysis provided for autonomous circuits ('sources'). However, this analysis (as name suggests) returns only jitter, without phase noise information.

Of course, with a single number for jitter it is not possible to recover
the phase noise curve, however, Spectre can return frequency-dependent jitter (in the Jee form).

If we have frequency-dependent jitter (already in Unit-Intervals), wouldn't this be equivalent to phase noise? What would be the correct formula for this jitter-phase noise conversion? Would the shape (and values) be true (e.g. 1/f) ?


I am looking forward to your comments.

Back to top
 
 
View Profile   IP Logged
Frank Wiedmann
Community Fellow
*****
Offline



Posts: 677
Munich, Germany
Re: Phase noise of a driven circuit from Spectre's frequency-domain jitter analysis
Reply #1 - Mar 1st, 2012, 5:51am
 
You get from time deviation to phase deviation by multiplying with 2πf. You should be aware, however, that this is not phase noise according to the usual definition because it includes contributions from higher frequencies due to the aliasing caused by the sampling operation (see http://www.designers-guide.org/Analysis/sc-filters.pdf and http://www.designers-guide.org/Forum/YaBB.pl?num=1224609785 for details).

If you are interested in the traditional (narrowband) phase noise, you should use modulated pnoise instead of jitter analysis. Which one is relevant for you depends on your application.
Back to top
 
 
View Profile WWW   IP Logged
raczko
New Member
*
Offline



Posts: 4

Re: Phase noise of a driven circuit from Spectre's frequency-domain jitter analysis
Reply #2 - Mar 7th, 2012, 3:59am
 
Dear Frank,
Thanks for your help.

In the thread that you've linked, you advise using 'jitter', time-domain analysis to get proper results for driven circuits. Why would now the modulated analysis be preferred?
I understand that in case of 'sources' method, the time-averaged noise results will degrade the actual numbers (where only the noise AT "zero-crossing" is important). How would the modulated method be better in that? Isn't the 'sources' method a subset of the 'modulated' method?

Thanks again, looking forward to be hearing from you.
Back to top
 
 
View Profile   IP Logged
Frank Wiedmann
Community Fellow
*****
Offline



Posts: 677
Munich, Germany
Re: Phase noise of a driven circuit from Spectre's frequency-domain jitter analysis
Reply #3 - Mar 8th, 2012, 6:42am
 
I did not say that modulated analysis is to be preferred. If modulated or jitter analysis is relevant for you depends entirely on your application.

Jitter analysis is relevant if a threshold crossing of the signal triggers some action. Modulated pnoise analysis is relevant if the signal is filtered so that only the sidebands of the fundamental frequency have an effect.

"Sources" is not a subset of "modulated". It's the time-averaged noise like you would see it on a spectrum analyzer.

For some more information, see http://support.cadence.com/wps/mypoc/cos?uri=deeplinkmin:ViewSolution;solutionNu... (and make sure to take a look at http://support.cadence.com/wps/mypoc/cos?uri=deeplinkmin:ViewSolution;solutionNu...).
Back to top
 
 
View Profile WWW   IP Logged
raczko
New Member
*
Offline



Posts: 4

Re: Phase noise of a driven circuit from Spectre's frequency-domain jitter analysis
Reply #4 - Mar 9th, 2012, 1:02am
 
Dear Frank,

Thank you for your answer. Please accept my apologies, by no means have I wanted to offend you.

The application is very simple. It is a matter of computing the phase noise degradation due to insertion of a circuit in a driven environment. We can assume that the signal source is noiseless and has a known rising time (slope). We can also assume that we have a chain of these buffers, so that the loading is equal for each stage and switching happens always at the same voltage levels.

'Sources' - being time-averaged is out of the question.
'Jitter' - seems most reasonable, but does not offer PN as result and can suffer from noise aliasing (as you described).
'Modulated' - offers PM results, but how is it in fact computed if no 'switching level' is given (as in jitter).

From what I observed in simulation, the 'modulated' method gives most reasonable results and when the circuit is altered, the results change according to expectations.

Thank you for the links to cadence support documentation. That gives a bit of clarification, however, do you perhaps know a source of more detailed description of these analyses?

Thanks again.
Back to top
 
 
View Profile   IP Logged
Frank Wiedmann
Community Fellow
*****
Offline



Posts: 677
Munich, Germany
Re: Phase noise of a driven circuit from Spectre's frequency-domain jitter analysis
Reply #5 - Mar 9th, 2012, 2:15am
 
No offense was taken.

As I said, what is relevant for you depends on how you will use the signal at the output of your circuit that you are examining.

Jitter pnoise analysis does not "suffer" from aliasing. Aliasing is a real-world effect that is inherent to sampling.

Modulated pnoise analysis works by simulating the noise in both sidebands of the fundamental frequency and the correlation between the noise in both sidebands. You can take a look at the netlist to see the simulations that will be done for this. ADE then uses the results of these simulations to compute PM and AM noise.

Searching the Cadence support site will often help to find information on these subjects. However, the most detailed description of sampled noise simulation is probably still http://www.designers-guide.org/Analysis/sc-filters.pdf. Searching this forum will also often help.
Back to top
 
« Last Edit: Mar 9th, 2012, 4:19am by Frank Wiedmann »  
View Profile WWW   IP Logged
Pages: 1
Send Topic Print
Copyright 2002-2024 Designer’s Guide Consulting, Inc. Designer’s Guide® is a registered trademark of Designer’s Guide Consulting, Inc. All rights reserved. Send comments or questions to editor@designers-guide.org. Consider submitting a paper or model.