The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Measurements >> Phase Noise and Jitter Measurements >> Jitter vs Phase Noise
https://designers-guide.org/forum/YaBB.pl?num=1302793605

Message started by mtp3 on Apr 14^th, 2011, 8:06am

Title: Jitter vs Phase Noise
Post by mtp3 on Apr 14^th, 2011, 8:06am

Hi!

I am new to the forum so please bear with me, I am simulating an inverter based ring oscillator (5stages) and when I measure the phase noise and then compute the jitter (Jc) for the same RO as described on the forum.

The phase noise values are positive around the oscillation frequency but the phase noise at 1Mhz offset from the oscillation frequency is -55dBc/Hz which is not a very good value While when I compute the jitter the Jc comes out to be around 140fs(Integration limit 0 to fo/2) where fo is the oscillation frequency and is around 11.9Ghz.

Now my question is, the phase noise value is relatively poor as compared to the values one finds in research papers but the jitter value is very small in the fs (femto seconds) range but it is usually in ps range even in the best of designs.

I would be grateful if some one could point me in the right direction. Thanks in advance :-/

Title: Re: Jitter vs Phase Noise
Post by rajkumar palwai on Apr 19^th, 2011, 2:29am

Hi
The jitter number that you will get by integrating the total phase noise power and dividing it by the 'w0' (fund frequency in rad/sec) is edge jitter but not the Cycle jitter, Jc.

Basically the phase noise gives u the total integrating noise on a particular VCO node rising edge and falling edges. But while calculating the cycle jitter, Jc, u will see the time difference variation between two succesive edges, where most of the low freq noise is correlated and only high frequency noise differs and shows up.

If you see 2 succesive edges of a VCO node, all the noise that has been integrated on the first edge is also there on second edge. But whatever high freq noise (f>1/T) that came during the particulate time period interval shows up only on 2nd edge. Only this extra noise is used in calculating the Cycle Jitter.

So, u can assume that u need to do some sort of differentiation on the phase noise and then integrate the resultant power to get the Cycle jitter power.

For exact formula to do that diferentiation, see the equation (1-18) in the following artilcle
www.sabreindia.net/cadence/application-note.pdf

Regards
rajkumar

Title: Re: Jitter vs Phase Noise
Post by rajkumar palwai on Apr 19^th, 2011, 3:03am

If you want to see the edge jitter number (the number which you find in the research papers), then u need to integrate the total phase noise power and divide it by 'w0'(fund freq in rad/sec).

Spectre pss+pnoise does not do this integration. You can probably extract those phase noise values in a text file and do the integration in MATLAB. Dont forget to change the phase noise units from dBc to volts^2. This conversion is easy and equal to 10^(L(f)/10), wher L(f) is phase noise in dBc/Hz

-regards
rajkumar

Title: Re: Jitter vs Phase Noise
Post by sheldon on Apr 20^th, 2011, 6:11am

mtp3,

Use PNOISE jitter option to calculate the jitter. There is an app note
in the documentation that explains how to use it.

Best Regards,

Sheldon