[time-nuts] Fast than light neutrino

[Hal Murray]

javier.serrano.pareja at gmail.com said:
>> A fiber-based time-transfer would be nice complementary as it would provide
>> an independent timing path.
> Any ideas on how to proceed? This is unknown territory for me. 

You can get a lot of good ideas from the radio astronomers.  It's been 
discussed here in the past, but I don't know what terms to use when searching 
the archives.  I think it was mostly pointers to their papers.  They were 
interested is much shorter distances.  I think it was 10-20 km.


The idea is to send a signal in both directions over the same fiber.  If it's 
the same fiber, the transit times are likely to be the same in both 
directions.  If you send a pulse out and back, you can assume the time the 
pulse arrived at the far end was half the round trip time after it left the 
start.


Whatever you do, it will require a lot of cooperation from the people who own 
the fibers.

Telecommunications fibers can go roughly 100 km between repeaters.  Fibers 
don't go in straight lines so you will probably need 10-20 repeaters.  Maybe 
more.

Normal repeaters don't work in both directions.  You will have to use pairs 
of fibers and swap them to measure the difference in length.

You may have to automate that step.  Logically, what you want is a box 
containing:
  4 fiber connectors
  2 repeaters (one in each direction)
  several switches that can swap pairs of fibers.
  an ethernet port and CPU to flip the switches

Each switch would have 2 fibers "in" and 2 "out".  They would either go 
straight through or cross over.  (You can build that from 4 single-pole 
double-throw switches.)  I think you need 4 of those switches: two for the 
fibers going in each direction, and two to swap the repeaters.

Maybe the delays through the repeaters are consistent enough that you don't 
have to swap them.

In the way old days of FDDI, before they invented hubs, they were working on 
mechanical switches to bypass a workstation when it was powered off.  I 
wonder if that technology is still around.


One thing to watch out for...  The telephone companies are used to swapping 
fibers without telling customers.


If you had a setup like that, would you want to use it for all your tests, or 
just once to verify your GPS based timing?

For a one-shot run, it might be possible to use people to swap the fibers.  
Assume the repeaters don't need to be measured.  Assume it takes 1 minute to 
measure the round trip time and 1 minute to swap fibers.  With 20 repeaters, 
that's less than an hour.



-- 
These are my opinions, not necessarily my employer's.  I hate spam.