[time-nuts] Cesium C Field Set?

Brooke Clarke brooke at pacific.net
Fri Feb 11 19:50:58 EST 2005

Hi John:

Well if your talking digits to the right of 1 then the SR620 has 12 
digits, i.e. the LSD is 1 pico second.

There's a procedure in  appendix B of the PRS10 Rubidium Source manual 
that allows making a time interval measurement on very stable sources to 
parts in E-12 in one second.  This is done by using the 10 MHz standard 
and reference signals as the start and stop inputs to a time interval 
measurement and then adding a 1 kHz external trigger arming input and 
averaging 1,000 measurements each second.  The result is a time interval 
measurement accurate to 1 pico second every second.

Have Fun,

PS It's been about 8 hours since I changed the GPS settings to 4 highest 
and the elevation mask to 30 degrees and so far the peak to peak 
wandering has been about 50 ns.  Much better than before, but time will 

PPS I have the 2100T set for 3200 GRIs time constant.  The manual 
indicates that this is desirable for stable sources.  The stability is 
now showing E13  4.9  and the absolute time difference is -0.03 micro 
seconds  But 0.03E-6 / (3200 * .099400) = 9.43E-11 not 4.9E-13 so I 
still do not know how the 2100x receivers are calculating the stability.

John Ackermann N8UR wrote:

> Bill Hawkins wrote:
>> We could get a little more quantitative about "impact" because there is
>> a relationship between the number of digits displayed and the number of
>> digits of accuracy. If you have an eight digit interval counter then
>> 10e-9 is enough accuracy. You can do that with the crystal in some
>> counters if it has been calibrated, but your external standard certainly
>> has enough accuracy.
> Good point, but my 5370 counter has 11 digits of time resolution (as, 
> I think, does Brooke's SR620), so the exponents start to get pretty 
> small...
> John

More information about the time-nuts mailing list