[time-nuts] A readl atomic wristwatch
lists at rtty.us
Tue Oct 1 19:28:28 EDT 2013
…. also don't forget that it's got a magnetic field susceptibility. Rotate it in the earth's magnetic field and you get something very similar to a 2 G tip on a crystal oscillator. That alone is enough to nuke the average error budget for 1,000 years.
Of course there's also that pesky and oh so obvious 0.2 wander in the second hand. Can't see at all how a *real* Time Nut could put up with that. :)
It is indeed a marvelous device. Packing all that into a tiny package wasn't easy at all.
On Oct 1, 2013, at 6:59 PM, SAIDJACK at aol.com wrote:
> there is only one problem: there is just no way they can claim only 1
> second error in a 1000 years unless they also have a GPS receiver for
> calibration in there, which kind of mutes the point as that has been done using wwvb
> etc. Let's do the math:
> 1/ (1000 * 365 * 24 *3600) = 3.171E-011 average error required over 1000
> The CSAC has a thermal spec of +/-0.5ppb for -20C to +70C, meaning 5.6E-012
> per Degree C sensitivity.
> So the temperature would have to stay stable within a couple degrees C,
> hardly possible in a wrist watch.
> Also, initial CSAC aging is 0.3ppb per month, about 100x worse than they
> claimed 1000 year per second accuracy. It gets better over time, but still
> around a ppb per year or so of aging is expected, far off from the 0.032ppb
> They did a great job integrating that together, and its novel, but the
> marketing department is off by many orders in magnitude in their accuracy
> I know CSAC applications that would be very happy to get around 0.1 second
> error per year consistently without external re-calibration in a stable
> In a message dated 10/1/2013 15:32:27 Pacific Daylight Time,
> marks at non-stop.com.au writes:
> That's the first time I have seen a practical explanation and working
> example of a CASC in operation.
> Can I say "Awesome"?
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
More information about the time-nuts