What is the difference between Jcc, Jc, and Jee in the jitter simulation?

[squirrelcn]

In the latest version of cadence, there is a jitter performance named "Jee".
What is the difference between Jcc, Jc, and Jee, in the jitter simulation?
Thanks a lot!:!:

 

[Jeekoky]

Re: Jitter Issue

Jee is the edge-to-edge jitter. It is the variation in the delay between a triggering event and a response event. When measuring edge-to-edge jitter, a clean jitter-free input is assumed. Edge-to-edge jitter assumes an input signal, and so is only defined for driven systems. It is an input-referred jitter metric, meaning that the jitter measurement is referenced to a point on a noise-free input signal, so the reference point is fixed. No such signal exists in autonomous systems.

Jc is referred to as k-cycle jitter or long-term jitter. It is a measure of the uncertainty in the length of k cycles and has units of time. When k=1, the standard deviation of the length of a single period, is often referred to as the period jitter, and it denoted J, where J = Jc when k=1.

Jcc is cycle-to-cycle jitter. It is a metric designed to identify large adjacent cycle displacements. It is like edge-to-edge jitter in that it is a scalar jitter metric that does not contain information about the correlation in the jitter between distant transitions. However, it differs in that it is a measure of short-term jitter that is relatively insensitive to long-term jitter.

Added after 6 minutes:

From what i calculate. I found there may be some errors in c@dence Jc calculation. I test it base on the same phase noise and same integral frequency band. But the c@dence result is always about 2.5 times lower than what i calculated. And because c@dence just add this feature, it doesn't offer much informantion. So until now, i don't know why.