[ntp:questions] Meinberg NTP monitor, silly question

Dave Hart davehart at gmail.com
Sun Dec 20 06:52:58 UTC 2009


On Dec 20, 00:32 UTC, G8KBV wrote:
> But what does 'o' mean?

It means the PPS peer, and when present effective trumps any * seen.

> The "Application Log" shows:-
> Using user-mode PPS timestamp for GPS_NMEA(2)  Not that I think it'll
> make any significant difference, but I thought I'd set all this up so
> that the PPS thing would run in Kernel mode, but obviously not.

So 127.127.20.2 (NMEA) is using the user-mode hack to get DCD
timestamps, which is only effective when you restrict the GPS to one
sentence per second.  However, 127.127.22.2 (atom/PPS) is using real
kernel-mode PPSAPI timestamps.  In this configuration, NMEA is
"numbering the seconds" with its sloppier timestamps, and once its
offset is under .4s, the PPS refclock takes over with its interrupt-
time timestamps.

> State   Remote          Refid           Stratum Type            When    Poll
>         Reach   Delay   Offset  Jitter
> *       127.127.20.2    GPS             0       Local clock     9       16      377
>         0.000   0.089   0.006  
> o       PPS(2)          PPS             0       Local clock     8       16      377
>         0.000   0.124   0.004

The jitter figures are nice and low for Windows.  I would hope the
offset to PPS(2) would eventually be trimmed to 25us or less, and even
with temperature swings you might hope to keep within 150-200us at
worst.

Cheers,
Dave Hart




More information about the questions mailing list