Reduction in GPS timing error due to lack of SA

During the SA turn-off event of 1 May 2000 I happened to be collecting long-term 1 PPS data between an HP 58503A Time & Frequency Receiver and a Kvarz CH1-76 Passive Hydrogen Maser.

The following phase plot is a 24 hour excerpt from that data.

Here is a close-up of the same data plotted with one hour per division. The change in SA appears to occur just after 0400 UTC.

This data was collected with an HP 53132A time interval counter (PHM start, GPS stop) using one minute averages.

Before/after comparison of 1-day run

I have been using an HP 58503A GPS Time & Frequency receiver to accurately measure the frequency of an H-Maser. The goal is to determine the frequency accurately enough so that I can adjust the 5 MHz synthesizer output to within 1e-14 and then follow the performance over weeks or months to measure the frequency drift rate (expected to be less than 1e-15/day).

Earlier, with S/A, the typical standard deviation for one day of 1 PPS data from the 58503A was 20 ns:

Now, without S/A, the typical standard deviation is 6.3 ns, more than 3x better:

Presumably all the visible variations in this plot are due to remaining errors in GPS and/or the performance of the disciplined oscillator within the 58503A.

Do you have
Return to home page

Page created with vi -- last updated Wed May 03 19:31:39 UTC 2000