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 relation between VCO & loop divider output
https://designers-guide.org/forum/YaBB.pl?num=1211983294

Message started by keikei on May 28^th, 2008, 7:01am

Title: Jitter relation between VCO & loop divider output
Post by keikei on May 28^th, 2008, 7:01am

In a PLL locked to a reference frequency, what is the relation between the jitter of VCO output ('Jvco') and the jitter of the signal fedback to PFD input after the loop divider ('Jdiv') ? assuming the div ratio is N.
In a certatin paper the author gives Jvco = N * Jdiv. But according to some early discussion here, most people agree that Jvco = Jdiv. So I am totally confused.
Is the effect of divider on jitter dependent on its location, i.e., in loop or out of loop?

Title: Re: Jitter relation between VCO & loop divider out
Post by Stefan on May 28^th, 2008, 7:08am

It's just some basic stochastic equation
the standard deviation of N statistical indepedent events, each having a standard deviation of J1 is
JN=sqrt(N)*J1.

So neither one of the formulas you spoke about is correct.
And please don't talk about "certain papers", quoting them is more efficient here :)

Title: Re: Jitter relation between VCO & loop divider out
Post by keikei on May 28^th, 2008, 8:39am

yeah, I should have specified the paper name:
Time-to-voltage converter for on-chip jitter measurement, IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 52, NO. 6, DECEMBER 2003 by Tian Xia and Jien-Chung Lo

Title: Re: Jitter relation between VCO & loop divider out
Post by Visjnoe on Jun 6^th, 2008, 11:19am

Dear,

Jvco = Jdiv (for jitter components within the PLL bandwidth).

The formula with the sqrt(N) relationship does not hold. It only applies for calculating the jitter of a divider chain constituted out of N unit elements (e.g. div 2). When the jitter of a unit element Junit is known, the total jitter of the chain (divider) is given by sqrt(N)xJunit.

The phase noise before and after the divider has a 1/N^2 relationship. Since the period becomes N times longer after the division,
the jitter is the same.

Regards

Peter