The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Mixed-Signal Design >> Jitter In CTDSM
https://designers-guide.org/forum/YaBB.pl?num=1218692010

Message started by viks_iisc on Aug 13^th, 2008, 10:33pm

Title: Jitter In CTDSM
Post by viks_iisc on Aug 13^th, 2008, 10:33pm

Hello friends, Its realy great to find a good site for analog designers !! :)

Well, i am working on CTDSM, and was trying to make sure that my CTDSM will nt be degraded enough by jitter.

For that right now, i am using SIMULINK. The problem is that to model the jittery clock, i need to generate pink noise (1/f) noise. and i am unable to generate it in MATLAB or SIMULINK. If anybody has already worked on it. Please help me out !! :)

Thanks !!

Title: Re: Jitter In CTDSM
Post by thechopper on Aug 14^th, 2008, 6:40pm

Hi,

Have you tried integrating white noise? That should be simple to do.

Regards
Tosei

Title: Re: Jitter In CTDSM
Post by vivkr on Aug 14^th, 2008, 10:59pm

HdrChopper wrote on Aug 14^th, 2008, 6:40pm:

Hi,

Have you tried integrating white noise? That should be simple to do.

Regards
Tosei

Hi Tosei,

Be careful !!! Integrating white noise does not give 1/f noise. It gives 1/f^2 noise. However, I think this post is in the wrong forum. It ought to be in the behavioral modelling forum. I remember having seen some discussion long ago on generating pink noise there, but can't find the thread.

Regards
Vivek

Title: Re: Jitter In CTDSM
Post by Ken Kundert on Aug 14^th, 2008, 11:40pm

You can use a fracpole to generate flicker noise. Just drive a fracpole with white noise and it will color it pink. Spectre has a fracpole built in. Or you can use the FracPole suite on http://www.designers-guide.org/Modeling/.

-Ken

Title: Re: Jitter In CTDSM
Post by thechopper on Aug 15^th, 2008, 7:53pm

vivkr wrote on Aug 14^th, 2008, 10:59pm:

Be careful !!! Integrating white noise does not give 1/f noise. It gives 1/f^2 noise.

Hi Vivek,

I think you are right. My mistake. Thanks for clarifying.

Regards
Tosei

Title: Re: Jitter In CTDSM
Post by ywguo on Sep 6^th, 2008, 6:36am

Hi viks_iisc,

If you use simulink, there is a toolbox named SDtoolbox. SDtoolbox is developed by A. Fornasari, P. Malcovati and available at http://www.mathworks.com/matlabcentral/fileexchange/loadFile.do?objectId=7589&objectType=FILE.

It has simulink model for jittered sinewave and colored noise. It is easy to use although I don't understand the algorithm. :)

Quote:

The problem is that to model the jittery clock, i need to generate pink noise (1/f) noise.

Furthermore, I don't think you need to generate pink noise (1/f) for modeling the jittered clock. Perhaps it is good enough to assume a gaussian distribution random jittered clock.

Best regards,
Yawei

Title: Re: Jitter In CTDSM
Post by thechopper on Sep 7^th, 2008, 5:44pm

ywguo wrote on Sep 6^th, 2008, 6:36am:

Furthermore, I don't think you need to generate pink noise (1/f) for modeling the jittered clock. Perhaps it is good enough to assume a gaussian distribution random jittered clock.

Hi Yawei,

Correct me if I´m wrong but I think that 1/f noise can also be gaussian distributed, so both features are not necessarily contradictory.

Regards
Tosei

Title: Re: Jitter In CTDSM
Post by ywguo on Sep 9^th, 2008, 2:56am

Hi Tosei,

You put forward a good question. Intuitively, I thought 1/f noise is not Gaussian distributed. However, it has not solid base and proper reasoning. So I agree with you now. It is possible to generate a 1/f noise whose amplitude is Gaussian distributed.

Thank you.
Yawei