[ntp:questions] GPS Stratum 1 and PPS
Miguel Gonçalves
mail at miguelgoncalves.com
Thu Jul 28 08:59:38 UTC 2011
Hi there!
I have been quite a time nut for a while but recently I bought a GPS 18 LVC
from Garmin and wired it up according to one of the many tutorials around
the Internet. I have a static IP and if this works out OK I am thinking
about opening my NTP port to the world. :-)
I am using FreeBSD 7.4 with a PPS_SYNC kernel and the GPS is connected to
COM1.
I am getting this
tick# ntpq -p
remote refid st t when poll reach delay offset
jitter
==============================================================================
*GPS_NMEA(1) .GPS. 0 l 11 16 377 0.000 0.007
0.004
+rustime01.rus.u .PPS. 1 u 38 64 377 59.228 1.422
1.749
-ptbtime2.ptb.de .PTB. 1 u 27 64 377 61.864 -2.420
0.922
+canon.inria.fr .GPSi. 1 u 37 64 377 44.454 1.346
0.728
and I think it's quite good but the problem is that my NTP daemon is
hammering the other servers on the configuration file. As I don't want to
get banned from using these servers I would like some opinions regarding my
config file:
tick# cat /etc/ntp.conf
server 127.127.20.1 mode 2 prefer minpoll 4
fudge 127.127.20.1 flag1 1 flag3 1 refid GPS
server rustime01.rus.uni-stuttgart.de
server ptbtime2.ptb.de
server canon.inria.fr
statsdir /var/log/ntp/
filegen clockstats file clockstats type day enable
filegen peerstats file peerstats type day enable
filegen loopstats file loopstats type day enable
statistics clockstats peerstats loopstats
logfile /var/log/ntp/messages
driftfile /etc/ntp.drift
Any ideas?
I have another machine that is physically close to this one I and got the
PPS signal from the GPS connected to the DCD line of the COM port also. I
intend to discipline the local clock with the PPS. I am numbering the
seconds using stratum 1 servers. Here's the configuration:
tock# cat /etc/ntp.conf
server 127.127.22.1 minpoll 4
fudge 127.127.22.1 flag3 1 stratum 0
server rustime01.rus.uni-stuttgart.de prefer
server ptbtime2.ptb.de prefer
server canon.inria.fr prefer
statsdir /var/log/ntp/
filegen clockstats file clockstats type day enable
filegen peerstats file peerstats type day enable
filegen loopstats file loopstats type day enable
statistics clockstats peerstats loopstats
logfile /var/log/ntp/messages
driftfile /etc/ntp.drift
The idea behind this is that if the server with the GPS is turned off for
maintenance I will still have the other machine getting the PPS and the
seconds from public NTP servers. Is this configuration correct?
By the way, on this second machine I get
tock# ntpq -p
remote refid st t when poll reach delay offset
jitter
==============================================================================
oPPS(1) .PPS. 0 l 14 16 377 0.000 -0.004
0.002
+rustime01.rus.u .PPS. 1 u 175 1024 377 59.726 1.577
0.143
+ptbtime2.ptb.de .PTB. 1 u 170 1024 377 62.099 -1.806
0.165
+canon.inria.fr .GPSi. 1 u 493 1024 37 44.637 1.385
0.794
This one is polling the Internet servers every 1024 seconds. Why doesn't the
other?
One final question... Is 1ms difference from stratum 1 close servers (36ms)
normal? For me it's quite good. Can it be better?
stratum 1, precision -18, leap 00, trust 000
refid [PPS], delay 0.03641, dispersion 0.00078
transmitted 4, in filter 4
reference time: d1dba30e.893d3d9b Thu, Jul 28 2011 9:57:50.536
originate timestamp: d1dba34e.8d92e2b2 Thu, Jul 28 2011 9:58:54.553
transmit timestamp: d1dba34e.8a966c79 Thu, Jul 28 2011 9:58:54.541
filter delay: 0.03818 0.03641 0.03738 0.04140
0.00000 0.00000 0.00000 0.00000
filter offset: 0.002177 0.001189 0.001666 0.003551
0.000000 0.000000 0.000000 0.000000
delay 0.03641, dispersion 0.00078
offset 0.001189
28 Jul 09:58:54 ntpdate[4779]: adjust time server 193.136.250.246 offset
0.001189 sec
Sorry for the large amount of questions but any help would be appreciated.
Thanks!
Cheers,
Miguel
More information about the questions
mailing list