Re: [chrony-users] How to use a "PPS" with freq higher than 1Hz?

William G. Unruh __| Canadian Institute for|____ Tel: +1(604)822-3273
Physics&Astronomy _|___ Advanced Research _|____ Fax: +1(604)822-5324
UBC, Vancouver,BC _|_ Program in Cosmology |____ unruh@xxxxxxxxxxxxxx
Canada V6T 1Z1 ____|____ and Gravity ______|_ www.theory.physics.ubc.ca/

On Tue, 30 Mar 2021, Vincenzo Miceli wrote:

> [CAUTION: Non-UBC Email] Thanks Miroslav,
>
> I was able to get the PPS to start by removing the "lock GPS" and using a public NTP. I guess the NMEA
> coming from the GPS has too much jitter then? I don't fully understand though as the NMEA is always

There is no indication in the pulses that they are associated with 0,
.2, .4,... seconds so all chrony can do is to assume that the gps signal is
assocatied with the "closest" pulse. But the gps signal wanders all over the
place. the GPS receiver only puts out the GPS signal when it is not terribly
busy, and since it comes at a slow rate and is a long string, the lengthof the
string also makes a difference. Associating it with the nearest pulse is fine
at once per sec ( sinc ethe GPS ususlly comes through about 100-200 sec after
the "seconds" mark) for for an accuracy of .2 sec, it is not.

> coming once a second and at higher frequencies (I would think) the chrony algorithm has to deal with more
> samples closer together and can still sample the GPS once a second and compare that with the 10 pulses

It does not sample the GPS. The GPS comes in when the receiver wants to send
it and that is always quite late.

> per second, maybe with the PPS-GPS separation still working off the 200msec limit... I'm sure there are

10 pulses per 2 sec you said.

> good statistical reasons though 🙂🙂

There is a reason why conductors have different wand waves throughout a bar of
music, so musicians know which is the first beat of the bar. There is nothing
like that for PPS. All beats are the same. And GPS is horrible at telling you
which that first beat was.

Anyway, why do you want 5 PPS per second? It does very little to improve the
accuracy of the timekeeping, especially since you are going to use it to no
better accuracy than ms, or I suspect, seconds anyway.

>
> Cheers,
>
> Enzo
>
> _________________________________________________________________________________________________________
> From: Miroslav Lichvar <mlichvar@xxxxxxxxxx>
> Sent: Tuesday 30 March 2021 13:58
> To: chrony-users@xxxxxxxxxxxxxxxxxxxx <chrony-users@xxxxxxxxxxxxxxxxxxxx>
> Subject: Re: [chrony-users] How to use a "PPS" with freq higher than 1Hz?
> On Tue, Mar 30, 2021 at 12:10:33PM +0000, Vincenzo Miceli wrote:
> > refclock PPS /dev/pps0 refid PPS lock GPS precision 1e-9 poll 3 offset 0.0 rate 5 dpoll -3 prefer
> >
> > but monitoring with "watch chronyc sources -v" shows no readings from PPS
>
> A higher PPS rate requires a more accurate and stable source for
> "locking". For a 5Hz PPS it needs to fit within 40ms. Your screenshot
> shows "+/- 100ms". That's too much. You will need to set a smaller
> delay for the GPS source.
>
> If that GPS source is not stable enough, you might need to disable the
> "lock" option, or use an NTP server, in order to start the PPS refclock.
>
> --
> Miroslav Lichvar
>
>
> --
> To unsubscribe email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
> with "unsubscribe" in the subject.
> For help email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
> with "help" in the subject.
> Trouble? Email listmaster@xxxxxxxxxxxxxxxxxxxx.
>
>
>