[time-nuts] Comparing GPS receivers

Tom Van Baak tvb at leapsecond.com
Sun Jan 9 19:24:10 EST 2005


Bill ,

>
> Has anyone compared the raw 1 PPS signals from the receivers,
> without the sawtooth in the disciplined oscillator?

>
> What other GPS receivers have been compared?

I compared the outputs of four GPS receivers to see
how stable their 1 PPS output was. Take a look:

http://www.leapsecond.com/pages/gps4/cns-1pps.gif

The dark blue line is an 8-channel Oncore VP, showing
its characteristic +/- 52 ns peak-to-peak sawtooth jitter.

The pink line is the newer Motorola M12+ receiver; see
a third less jitter.

The yellow line is an M12+ with sawtooth correction
applied.

The fine light blue line in the center is an HP 58503B,
the 8 channel, bench-top version of the Z3801A.

While the HP GPSDO clearly has less jitter (see how
narrow the line is), in this post-S/A era, it actually has
slightly more mid-term wander than the raw M12+ with
sawtooth correction.

The above plot shows about 3 hours (10,000 seconds);
however the plot looks the same if it goes out to a day,
or a week, etc.

>
> Has anyone used the raw 1 PPS to run a clock display for a
> year or more? That seems like an adequate integration period.

GPS follows UTC(USNO) and UTC(USNO) follows
UTC so -- at the level we're dealing with here -- there
is no long-term error in a GPS signal to speak of.

The behavior of a GPS receiver is mostly revealed
within a few hours; I don't think a longer integration
time buys you much. I mean, if your GPS receiver
1 PPS wanders by 50 ns over a few hours it will also
show wandering by 50 ns over a day, or a month,
or year.

/tvb






More information about the time-nuts mailing list