[ntp:questions] PCI add on oscillator

Mauro Fiacco Mauro.Fiacco at ipaccess.com
Mon May 16 12:45:39 UTC 2005

Following this question, what would be the effect of having a fairly
stable on board oscillator (e.g. <<1ppm in a few days)? 

My understanding is that the NTP PLL assumes that offset measurements
(the output of the Peer Selection and Clock Filter algorithms) are
always more accurate than the oscillator own readings, which may be so
for standard PC oscillators.

How does the NTP algorithm cope with a situation where the stability of
the local oscillator can be ensured for prolonged period of time, while
the network is affected by high jitter and asymmetry?



On Wed, Apr 27, 2005 at 10:41:06PM -0400, Joshua Coombs wrote:
> Date: Wed, 27 Apr 2005 22:41:06 -0400
> From: Joshua Coombs <jcoombs at gwi.net>
> To: questions at lists.ntp.isc.org
> Subject: [ntp:questions] PCI add on oscillator

> Given standard pc clocks are less than ideal, with wide drift, etc, 
> does anyone make a good quality oscillator as an add in, say via PCI? 
> Ideally something NTP could discipline, so one could build a PC that 
> could serve reliable time for more than 5 mins if external time 
> sources became unavailable.  I believe I read an article by PHK where 
> he built one for test purposes, the discussion of 500ns accuracy got 
> me thinking about it again.

> Joshua Coombs 

> _______________________________________________
> questions mailing list
> questions at lists.ntp.isc.org
> https://lists.ntp.isc.org/mailman/listinfo/questions

Dr Mauro Fiacco 

ip.access Ltd                      
URL: www.ipaccess.com  

More information about the questions mailing list