You don’t give your config, but as using the shared memory driver works, do you 
have a prefer peer active when you are using the atom driver? If there isn’t 
one, that will cause the ref clock to be falseticker’d?. Your atom jitter isn’t 
the issue.


"Ceux qui sont prêts à abandonner une liberté essentielle pour obtenir une 
petite et provisoire sécurité, ne méritent ni liberté ni sécurité."
Benjimin Franklin

> Le 9 mars 2015 à 10:52, folkert <folk...@vanheusden.com> a écrit :
> 
> Hi,
> 
> I have a raspberry pi with a gps connected to it.
> It runs linux kernel 3.18.5+ and ntp version 1:4.2.6.p5+dfsg-2.
> When I use my own rpi_gpio_ntp which is a userspace program, I get
> something like:
> 
>     remote           refid      st t when poll reach   delay   offset jitter
> ==============================================================================
> *127.127.28.1    .PPS.            0 l    8   16  377    0.000   -0.010 0.008
> 
> *127.127.28.1    .PPS.            0 l   11   16  377    0.000    0.005 0.012
> 
> *127.127.28.1    .PPS.            0 l   12   16  377    0.000    0.005 0.012
> 
> *127.127.28.1    .PPS.            0 l    1   16  377    0.000    0.001 0.011
> 
> (I omitted the peers)
> So the jitter is around 8us. Not bad for userspace.
> 
> Linux has this PPS interface for a while so now that it is easy
> configurable in config.txt on the rpi, I thought I give that a try. The
> strange thing now is that using that I get very bad jitter. So bad that
> ntp discards it as a false ticker!
> 
>     remote           refid      st t when poll reach   delay   offset jitter
> ==============================================================================
> x127.127.22.1    .PPS.            0 l    2   16  377    0.000    0.161 0.001
> 
> x127.127.22.1    .PPS.            0 l    7   16  377    0.000    0.148 0.006
> 
> x127.127.22.1    .PPS.            0 l   13   16  377    0.000    0.147 0.007
> 
> x127.127.22.1    .PPS.            0 l    2   16  377    0.000    0.146 0.006
> 
> This is odd: the pps interface should work way better due to lower
> latency etc.
> 
> What could be going wrong here?
> 
> The one at the top (28.1) uses GPS and the one at the bottom uses a
> combined Glonass/GPS receiver. Note that I also swapped tried using pps
> with 28.1 and rpi_gpio_ntp on 22.1 and then 28.1 fails and 22.1 is fine.
> 
> 
> Folkert van Heusden
> 
> -- 
> Curious about the inner workings of your car? Then check O2OO: it'll
> tell you all that there is to know about your car's engine!
> http://www.vanheusden.com/O2OO/
> ----------------------------------------------------------------------
> Phone: +31-6-41278122, PGP-key: 1F28D8AE, www.vanheusden.com
> _______________________________________________
> questions mailing list
> questions@lists.ntp.org
> http://lists.ntp.org/listinfo/questions
_______________________________________________
questions mailing list
questions@lists.ntp.org
http://lists.ntp.org/listinfo/questions

Reply via email to