[time-nuts] Question regarding 1PPS timing accuracy and GPSDO

Hal Murray hmurray at megapathdsl.net
Thu Jan 13 08:33:17 UTC 2011

> Suppose I want to make a GPSDO using the 1PPS output of generic GPS device
> XYZ, what kind of accuracy should I be looking for?

> I understand the concept of "smaller error is better", but I lack the
> practical experience and would like to get a handle on the actual numbers.

> When a GPS devide XYZ datasheet quotes a "timing accuracy" on the 1PPS
> output of lets say 1 us RMS, would it be useful? I.e would /you/ consider
> using it? What if it was 100 ns RMS on the 1pps? Or 10 ns RMS? Or ... At
> roughly what number would you consider the 1pps as a candidate for your

I think you have to dig deeper into what the datasheet means.  GPS has good 
long term stability but not so good short term stability.  If the device you 
are trying to lock to GPS has good long term stability, you can average over 
a long time period to filter out the short term errors from the GPS.

I think you want to look at the ADEV curves for your target device and your 
GPS setup.  Someplace out to the right, the curve for GPS will cross over and 
go below the curve for your local oscillator.  That's where you want to set 
your filter time constant.

If that's good enough for you, then sure, use the PPS signal.

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

More information about the time-nuts mailing list