[ntp:questions] Re: RMS error calculation

rtxo gnu at wraith.sf.ca.us
Tue Oct 28 16:05:49 UTC 2003


Piotr Trojanek wrote:
> Hi!
> 
> I would like to estimate error that clock will have after period of
> free-running. I have found some Allan's papers about this, but as usual,
> they are quite difficult for me to understand:(
> 
> Now I belive, that offset predictino error can be computed from Allan's
> deviation plots. Could you tell me do I follow the right way to find the
> answer?

Hello Piotr--

The ppm figure ntp produces, assuming constant temperature, should allow
a simple linear extrapolation over time. Of course, temperature fluctuates
and produces aging and other badness in rtxo's, so the extrapolation will
not be accurate over, say, 6 months. But it should give a starting point.

The ntp executive brief has a formula:


  http://www.eecis.udel.edu/%7emills/exec.html

...that includes some other terms, that might help you generate an estimate.

Good luck.

../Steven





More information about the questions mailing list