[ntp:questions] Getting NTP to correct only the clock skew

David Woolley david at djwhome.demon.co.uk
Sun Apr 22 09:00:31 UTC 2007

In article <4628d626$0$16192$426a74cc at news.free.fr>,
Spoon <devnull at localhost.com> wrote:
> David Woolley wrote:

> > Incidentally, I suspect he has a requirement to maintain timing to a
> > higher standard than the final consumer.  The final consumer may well
> > be just crystal stabilised, so he may be trying to improve that by a
> > factor of about 10.

> I don't understand this statement. Could you explain what you meant?

Although you can control the effective clock rate of the retiming PC, you
cannot control the clock rate of the receiving card.  One would expect them 
to have similar uncontrolled frequency errors.  Therefore it only makes sense
to lock the sending and receiving PCs together if you want their frequencies
to be more accurate than the receiving PCI card's.

I'm not really sure how much this buys you.  Maybe, if everything was at the
worst end of its range, you'll half the rate of underruns or overruns.

More information about the questions mailing list