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

Richard B. Gilbert rgilbert88 at comcast.net
Tue Apr 13 17:19:41 UTC 2010

unruh wrote:
> On 2010-04-13, Rob <nomail at example.com> wrote:
>> unruh <unruh at wormhole.physics.ubc.ca> wrote:
>>> On 2010-04-13, lhommedumatch <ludocluba at yahoo.com> wrote:
>>>> On 13 avr, 00:48, unruh <un... at wormhole.physics.ubc.ca> wrote:
>>>>> On 2010-04-12, lhommedumatch <ludocl... at yahoo.com> wrote:
>>>>>> I'm working on a oceanographic vessel
>>>>>> ntpgmtaceb is our reference clock that is synchronised with a gps at
>>>>>> the beginning of a cruise.
>>>>> Why not put a Garmin GPS 18 ?on board the ship and get gps time on the
>>>>> whole cruise?
>>>> That's our policy. I can't decide for that.
>>> If time is important then those who can decide need to be informed. If
>>> it is not important, why worry about 1ms. Esp as over the course of a
>>> cruise the time is apt to be out by seconds, not ms. 
>> In many applications, it is more important that all devices within
>> some environment agree upon the same standard time, than that this
>> standard time is within a small offset of true atomic time.
>> This has been discussed before here, but you don't seem to get it.
>> He may well want all his computers to be within a millisecond of the
>> same time reference, but may not at all care what that reference is.
>> There are many applications where this is a reasonable expectation.
> That may be what he wants, but it is not what he asked for. He stated
> that he was using an external source as his ntp server, and wanted his
> machine to be within a ms of it. That is a job admirably suited to a GPS
> receiver. He then stated that this was on an oceanographic vessel which
> was set at the beginning of a cruise. A gps receiver again would be
> ideally suited to giving the time throught the cruise. Unless the cruise
> is underwater, the gps signals would be available at all times during
> the cruise. 
> Now if what he wants is to sync all of the computers on board the vessel
> to a single time source, not matter how bad that time source is, then I
> agree, that suggesting a gps is largely irrelevant ( except it can be
> used to make his single source a good time source rather than a terrible
> one). There are a number of procedures for synching a large number of
> systems. But that was NOT what he asked about. You may well be a better
> mind reader than I am, however. 
>> Suggesting a Garmin GPS to be added to each and every configuration
>> brought up by someone is really not adding much.  When it is not there,
>> it is not there.  And the configuration may still be very useful.
> Of course. That is why it is important for people to ask questions based
> on what they need, rather than on what they think is the solution to
> their need. 

And that is why you may need to ask "What problem are you trying to 
solve?"  There is also a very good article:


More information about the questions mailing list