[time-nuts] GPS for use with NTP server?

[Hal Murray]

> I've downloaded and read some of the GPS user manuals.  The new ones are 12
> channel and older ones are 5 or 8.    I guess with a surveyed location you
> could get time from just one satellite.  But you'd need more if you wanted
> to take account of the atmosphere.  I don't know if receivers do this. 

You need more than just a surveyed location.  You also need firmware that 
will use the surveyed location.


> But my question was about NTP and if NTP on a typical computer could make
> use of 1PPS that is down in the nanoseconds 

A microsecond or few is generally great for NTP.

One of the main considerations is temperature of the local crystal which can 
fluctuate due either to CPU load or environmental changes.
  http://www.megapathdsl.net/~hmurray/ntp/slope.gif
Each color is 24 hours.

My suggestion would be to start with the lowest cost solution and see if you 
can detect any problems that might be associated with the lack of quality on 
the PPS signal.  I doubt if you will find anything.



-- 
These are my opinions, not necessarily my employer's.  I hate spam.