[time-nuts] A question about GPS pps quality.

[John Ackermann N8UR]

Poul-Henning Kamp said the following on 07/26/2007 11:46 AM:

> In message <46A8C00D.50603 at kudra.com>, Robert Sexton writes:
>> ); SAEximRunCond expanded to false
> 
>> 1.  If I interface one of these to my closet-based server, do I really 
>> have a Stratum-1 source? 1us accuracy doesn't sound all that great, but 
>> I'm nut so sure that the server can take advantage of a pulse with 
>> greater accuracy.
> 
> Stratum-1 - certainly.  1us accuracy ?  Highly unlikely.
> 
> The clock crystal in your computer is far too temperature sensitive
> for that.
> 
>> 2.  Would I get significantly better performance if I shelled out $75 
>> for one of the new trimble timing GPS modules, or found an old oncore 
>> timing module on ebay?
> 
> No.
> 
>> I know that the oscillator on my old AMD based UNIX machine isn't all 
>> that great, but I'll leave that question for another time.
> 
> That is where you should spend your money first.
> 
> Any GPS with a PPS output is sufficient to what you can do with a
> normal clock crystal.

Poul-Henning speaks with more practical NTP experience than I do, so I'd
give greater weight to his answer than to mine...

John