GPS Timekeeping Performance

Here are some representative data plots showing GPS timing from a Motorola M12+ Timing receiver compared to an HP 5065A Rubidium frequency standard. The standard had an offset of about -1.69x10-12.

First, here is the raw data collected for just under 24 hours:

After removing the HP 5065A's frequency `offset, we get this:

Zooming in on 20 seconds' worth of data shows the timing sawtooth that is inherent in the Motorola GPS design; this averages out over time:

This plot isn't really fair to the Oncore, as it can output a correction to account for the sawtooth value. Unfortunately, that requires use of a computer to process the data, while here I'm just feeding the 1pps output into a time interval counter with no simple way to apply the correction.

Here are a couple of other zoomed-in plots showing some interesting irregularities in the data. Make of them what you will.

Finally, here's a plot of the Allen Deviation (showing the stability of the signal over various averaging times):