[time-nuts] GPS Controlled Rubidium

Brian Kirby kirbybq at bellsouth.net
Thu Apr 7 23:13:26 EDT 2005

I have been working on a project to GPS discipline an Efratom FRS-C 
rubidium oscillator via Brooks Shera’s GPS controller.

I had to build a DC amplifier to interface the 5x10-10 rubidium 
sensitivity to the 7.5x10-9 controller sensitivity. Basically a gain of 
X15 was needed. Brook’s controller has several PI filters and the data I 
will report is in the longest PI constant, running in mode 7.

A 1 second Allan variance was recorded as 1.71x10-10, with a Maximum 
Time Interval Error (MTIE) of 37 nS per 24 hours. 10 seconds was 
7.55x10-12, MTIE 28 nS. 100 seconds was 1.5x10-12 MTIE 29 nS. 1000 
seconds was 1.65x10-12, MTIE 28 nS. 10,000 seconds was 1.69x10-13 32 nS. 
I used another rubidium for reference when recording the 1 to 1000 
second measurements, and from 100 to 10,000 seconds, I used a Z3801A for 
reference. Sure wished I had a cesium to compare too! I used a Motorola 
M12+ GPS receiver as the 1 PPS source for the Shera controller.

Brook’s normal controller updates the rubidium’s EFC every 30 seconds. I 
believe that is too fast for the unit. He has programmed a new PIC for 
me that updates at a 120 second rate. Testing is in progress with the 
new unit and it will take about 30 to 45 days to run all the tests that 
I require for comparison.

More information about the time-nuts mailing list