[ntp:questions] Re: Bought a GPS receiver--What next?

John R. Ackermann N8UR jra at febo.com
Tue Nov 2 13:30:57 UTC 2004



Richard B. Gilbert wrote:
> David Schwartz wrote:
> 
>> "Richard B. Gilbert" <rgilbert88 at comcast.net> wrote in message 
>>
>>    One millisecond is pretty poor PPS accuracy for a GPS receiver 
>> intended for timing applications. One microsecond is typical.
>>
>>    DS
> Evidently I don't write any better than Motorola did!!  The leading edge 
> of the PPS is exact!  In the Oncore M12 it is specified to be within 50 
> nanoseconds of the UTC second!

Coincidentally, I'm using the 1pps output from an M12T right now to 
calibrate an HP 5065A Rubidium frequency standard (comparing 1pps 
outputs via a time interval counter with 20ps resolution to determine 
the frequency offset of the Rb).

I was hoping I'd be able to post a graph of the performance right now, 
but I'm having a problem with format conversion -- the analytic software 
outputs a 3MB bmp file and my tools won't successfully convert it to 
something more web-friendly.  I'll work on that problem tonight.

But the short story is that over 12 hours, and after removing the Rb's 
frequency offset (about -1.5x10e-12), the "noise" on the 1pps output 
from the GPS is +/-60ns, with a small number of peaks that extend a few 
nanoseconds further -- the absolute maximum deviation over the period 
was about +/-75ns, with only a couple of samples hitting that extreme. 
Averaging over 100 seconds reduces the noise to the +/-15ns range.  Oh, 
and I'm not using the sawtooth correction capability of the M12T -- I'm 
just looking at the raw PPS output.

I'll try again tonight to generate a plot and will post the URL here.

John



More information about the questions mailing list