Allan variance for a disciplined clock

Allan variance for a disciplined clock

Post by Mauro.Fiac » Fri, 02 Jun 2006 18:55:01

Hello to All,

I understand (e.g. from Mills papers) that the Allan variance can be used to
characterise timing sources and clock drifts.

I was wondering about the significance of calculating the Allan variance
for a clock disciplined by NTP, while being affected by network jitter
(assuming high quality timing source).

Can the Allan variance be used to provide a quality measure of NTP in a
particular network? or lower order statistics, such as mean of the
frequency/offset errors are more apt to quantify NTP.

My aim is to provide a quantitative measure of NTP ability at
disciplining a particular oscillator, under different network jitter.
Such that a minimum network quality can be recommended in order to
achieve given accuracy (or in my case frequency stability).

Anybody has done any work in this direction?



Mauro Fiacco

ip.access Ltd

questions mailing list