[ntp:questions] how to have offset < 1ms

unruh unruh at wormhole.physics.ubc.ca
Wed Apr 14 23:55:54 UTC 2010


On 2010-04-14, lhommedumatch <ludocluba at yahoo.com> wrote:
> On 14 avr, 21:15, Terje Mathisen <"terje.mathisen at tmsw.no"> wrote:
>> John Hasler wrote:
>> > Richard B. Gilbert writes:
>> >> With a closed network the options are limited. ?With a place to site a
>> >> GPS antenna, you can install a GPS clock that will be accurate to
>> >> within +/- 50 nanoseconds.
>>
>> > It occurs to me that the sort of fancy GPS that an oceanographic ship is
>> > sure to have on the bridge may already be capable of providing time
>> > service.
>>
>> The operative word here is _may_. Many (most ?) navigation GPS setups,
>> even the fancy RTK models, couldn't care less about providing an
>> external PPS signal.
>> :-(
>>
>> Terje
>> --
>> - <Terje.Mathisen at tmsw.no>
>> "almost all programming can be viewed as an exercise in caching"
>
> The ntpgmtaceb (this clock loses 1.10e-11second every second => ie 8.6
> 10e-7s each days) is providing zda+pps, I have worked on using PPS
> with LinuxPPS.
> But I need to patch the kernel. I'm using CentOS 5.4 => linux2.6.18.

Why do you need to patch the kernel? Just use shmpps or gpsd to deliver
the PPS time to ntpd to discipline the clock. Now, that may only give
you 1usec accuracy, instead of 1/2usec, but since discussion was about
ms, that should not be the problem


> I have succeeded to work with linux2.6.27 on a ubuntu PC but it's
> still experimental and can't be used for operationnal recording.
>
> You all have right, the problem is to have all the devices
> synchronized no matter if the time is not correct.
> But I'm just responsible for one device. So the problem is how can I
> be synchronized to ntpgmtaceb.

If that is your problem, then you HAVE to figur out what is going wrong
on your network. 
Use tcpdump with timestamps to try to see what is going on-- is the
delay asymmetric? Attach a laptop into the network near your computer
and near the ntpgmtaceb to see where the delay is happening. 



>
> Thanks for all this discussion, and I'll need to be at shore to read
> all this in a more convenient way.




More information about the questions mailing list