[time-nuts] Thunderbolt accuracy...??
magnus at rubidium.dyndns.org
Sat Dec 20 18:13:43 UTC 2008
Michael Baker skrev:
> Hello, TimeNutters--
> John Miles, KE5FX wrote:
> > Well, sure, it's more accurate than the undisciplined
> > OCXO in the counter, that's the idea behind the Thunderbolt. :)
> n3izn at aol.com wrote:
> > I just wanted to ask the group if the
> > Thunderbolt would be more accurate than the internal reference? I
> > want to think it is but my link to the thunderbolt spec sheet is
> > no longer valid.
> Some time ago, I took my T-bolt over to the Metrology and
> Standards Lab at the University of Florida and set it up
> to run overnight and let it do its full survey process.
> In spite of the fact that the antenna was in a rather poor
> location it locked up quickly and seemed to run flawlessly.
> The monitoring was only intended to look at the frequency
> accuracy of the 10 MHz output-- other timing characteristics
> so dear to Time-Nuts hearts were ignored.
> After its overnight warm-up and survey process was done,
> we found (over a period of 48 hours) that the 10MHz output
> was never worse than 1.0 X E-12 and was generally better than
> that by a considerable margin about 90% of the time.
You should have logged the reported frequency error and compared with
the logged error. An average over a longer period would probably go deeper.
But it does show the potential for getting pretty good accuracy for a
small lab at a reasnoble cost. You should monitor the condition of the
receiver to ensure you are locked, and preferably log the frequency
error to be able to see any deviations.
More information about the time-nuts