The PPB parameter is its estimate of the error of the oscillator edge from where it needs to be. They specify it in mysterious parts-per-billion, but never really say billions of what. They do say positive values mean the 10MHz signal is running slow relative to GPS time.

I'll bet someone has some curves out there that could give a quick order of magnitude estimate... I have a poor quality chart here that shows an AT cut with about 40 ppm change for a 50 degree C change, and an SC changes about 5 ppm over the same span.

The estimators will filter themselves regardless when m has become a number of multiples, essentially when the filter bandwidth becomes large compared to the inverse of the tau. If the noise is benign white noise it should be kept at a low level, but work on flicker noise is where most effort should be spent. Cheers, Magnus

We should standardise some test for estimating the input contribution.

Thus, only for short-term stability may straight averaging work for estimation. Cheers, Magnus

That makes the estimation process pretty tough. Pay for them as if they were dead, feel good if you get a working one. Don't feel bad if you get out-bid... There are indeed things that a Cs can do that a GPS locked Rb can not. Running a Cs 24 hours a day 365 days a year for years and years can be pretty expensive.

I bet they'll drop estimating your bill too! Best, Will *********** REPLY SEPARATOR *********** On 6/27/2011 at 2:54 PM Brooke Clarke wrote: >Hi John: > >Yes. > >In S. California there was a big investigation about their accuracy.

Never under estimate the bureaucratic mind...

Delambre estimated 6377 for equatorial in 1810.

The FTS tubes in my 4050 have a Zeeman frequency of around 43KHz, which if my estimate is right is about 10KHz different from the HP Zeeman frequency of around 53KHz. And as I recall, the little HP submarine C-Beam has yet another Zeeman frequency, but I can't dig up its manual at the moment.