[time-nuts] Frequency over fiber (was WWV and legal issues)
kb8tq at n1k.org
Sun Sep 2 10:49:47 EDT 2018
Tidal effects can be very “non linear” as you approach a coast line. Lots of change
over a fairly short distance. If indeed the world decides to put in a global PTTI fiber
system, all of this would get into the mix on some links. It appears that the existing
technology would handle the issues.
Of course there’s still some guy named Bob running that back hoe without checking
for buried lines …..*That* we could test for … likely no need to run the experiment. :)
> On Sep 2, 2018, at 10:07 AM, Tom Van Baak <tvb at leapsecond.com> wrote:
>> I suspect there’s a longer list of “slow” environmental effects that are also taken
>> care of with the compensation setup. One would guess that crossing a active
>> fault line would be “interesting”.
> Yes, here's a back of the envelope calculation for you:
> - the Pacific Northwest moves on the order of 10 cm per year 
> - 1 meter of time is 1/299792458 = 3.3 ns
> - 10 cm/year is 3.3 ns / 86400 / 365 = 1e-17 df/f
> - the best laboratory optical clocks are down to that level of stability 
> On the other hand, in the real world you'd have to convince me that you've found two national timing labs with 1) state-of-the-art optical clocks, 2) which operate as phase (time) standards instead of as frequency standards, 3) or run continuously for a year (instead of a few times per week), 4) are connected by stabilized fiber, 5) that cross plate boundaries moving anywhere near as much as 10 cm/year, and 6) the optical time nuts running the clocks don't already factor geodetic effects like this into their clock comparisons...
> Unfortunately I won't be able to measure this. Even if John Miles (who also lives near Seattle) and I find optical clocks on eBay some day, and we find a way to run 30 miles of fiber between us without anyone noticing, we are both on the same tectonic plate so the drift cancels out. Note that lunar/solar tidal effects would be common mode to us as well.
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.
More information about the time-nuts