[time-nuts] NTP, PPS and < 10 us offsets

[Chris Albertson]

Long intervals are a good thing.

What NTP does is discipline the rate of the clock.   So lets say you
wanted to adjust the rate of some mechanical clock.  You first set it
as good as you can then wait an hour and see if it gained or lost
time. then you move the fast/slow lever on the back of the clock.
After a few hours the clock is pletty good and you have to wait a full
day to see an error.  Then as you zero in on the exact rate setting it
takes a week to see any error.

This is kind of what NTP is doing if you want to correct out tiny
errors in the clock rate you have to wait a long time

That said I think there is a "maxpoll" parameter you can add o the
/etc/conf file.

But why?

Maybe not you, but I think some people think "Oh my, NTP is not
checking the time very often.  It must not be accurate."   That is
thinking backwards, you need the long poll interval to see small
errors in rate.   But this assumes a stable local clock.  NTP balances
this.



On Tue, May 14, 2013 at 5:30 PM, Miguel Barbosa Gonçalves <m at mbg.pt> wrote:
> Hi!
>
> I am trying to figure out why NTP takes so long to react to oscillator
> changes... I want it to track the PPS from a Trimble Acutime as closely
> as possible.
>
> When checking ntptime on startup I see that the NTP daemon is using
> PLL and starts the adjustment interval at 8 s. It then increases to
> 16, 32, 64, 128 and stops at 256.
>
> I read in the documentation that this limit should be 128 but is now 256.
>
> --- start from http://www.ntp.org/ntpfaq/NTP-s-config-adv.htm ---
> intervals 57 says that there were 57 calibration intervals. When PPS
> pulses are arriving, this number should increase. Each frequency
> adjustment requires a good calibration interval. The length of the
> current calibration interval can be found as interval 128 s (128
> seconds is the default maximum length). Remaining numbers count
> abnormal conditions as explained below.
> --- end ---
>
> How can I make it stop earlier? At 64 s for example?
>
> Any ideas?
>
> TIA,
> Miguel
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.



-- 

Chris Albertson
Redondo Beach, California