[time-nuts] ntpq output

[David J Taylor]

> The following is a few ntpq -p outputs. Just random samples over a week. 
> I seem to periodically have large offset.
>
> I'm open to suggest how to improve this if the numbers don't look 
> normal.
>
> Please trim  the raw data to keep the size of any replies as small as 
> possible.
>
>>      remote           refid      st t when poll reach   delay   offset 
>> jitter
[]
>> +235-69-67-68.st 130.159.196.117  3 u  190 1024  377   88.459   -7.813 
>> 4.419
>> *cheezum.mattnor 129.7.1.66       2 u  614 1024  377   63.319   -2.871 
>> 4.571
>> +rigel.jeffkapla 69.164.222.108   3 u   90 1024  377   88.959   -5.622 
>> 1.890
>> +w1-wdc.ipv4.got 10.0.77.54       4 u   86 1024  377   91.660   -5.236 
>> 3.399

Your delays look awful.  I would hope for 15-30 ms over the Internet. 
Either you have a poor connection, or your servers are a continent away. 
Can you find servers nearer (in network terms) to you?  Have you tried the 
pool servers for your region?  If you can't improve the delays, you 
/might/ get an improvement by polling more frequently, setting maxpoll=9 
for example - making the maximum interval 2^9  i.e. 512 seconds.

The obvious improvement is to get a startum-1 server yourself, which may 
not cost the earth - perhaps US $30-50 if you have somewhere with a 50% 
sky view.  I have notes about the Garmin GPS 18x LVC and Sure GPS board on 
my NTP page:

  http://www.satsignal.eu/ntp/index.html

Cheers,
David
-- 
SatSignal software - quality software written to your requirements
Web:  http://www.satsignal.eu
Email:  david-taylor at blueyonder.co.uk