[time-nuts] Jitter Definition

Magnus Danielson magnus at rubidium.dyndns.org
Fri Apr 18 13:54:12 EDT 2014


On 04/17/2014 04:54 AM, HagaaarTheHorrible wrote:
> Hello there,
> I tried searching the archives (and google, IEEE, NIST, ITU), but didn't really find a satisfying answer, so I thought I'd ask directly.
> In short:
> Is there any kind of standard definition for Jitter which is commonly accepted?

Yes. You want to read ITU-T Rec. G.810.
In short, it takes phase deviations (or Time Error if you so choose) and 
separates the variations into jitter and wander. The separation is given 
at 10 Hz, so phase deviations faster than 10 Hz is "jitter" and those 
below is "wander".

This 10 Hz limit is arbitrary and it is sometimes (eh, most times) 
overused. The real meaning for this limit is that this is the typical 
bandwidth of an SDH Equipment Clock PLL, so deviations below 10 Hz is 
accumulated through a chain, and comes from the reference side of the 
PLL while deviations faster is suppressed by the PLL so only the jitter 
originating from the locked oscillator will be dominant on the output.

> I (think I) understood Jitter and phase noise by now, yet I need to give some references in my bachelor's thesis, so I'm looking for a definition. So far I haven't found a real definition of the different "types" (RMS,p2p,c2c,...) and components(RJ,DJ) of Jitter, but I guess there must be some kind of accepted standard!?
> If anyone could point me to some "official sources" which are "accepted in the industry", I'd be very grateful.

Do look into ITU-T Rec. G.810, 811-813, 823-825.

If you need more, let me know.


More information about the time-nuts mailing list