Hi > On Sep 2, 2018, at 7:17 AM, Magnus Danielson <mag...@rubidium.dyndns.org> > wrote: > > Hi Gerhard, > > I see that this became a separate thread. > > On 09/02/2018 01:38 AM, Gerhard Hoffmann wrote: >> >> >> Am 01.09.2018 um 20:40 schrieb Magnus Danielson: >>> One should first know that there is a lot of papers now on frequency >>> transfer over fiber. The stability achieved on the best ones so far >>> greatly below that of the optical clocks that they want to compare. >> >> Please, in a nutshell: what are the worst offenders: >> >> - tranceivers (mechanical, temp, other misfeatures) >> - cables ( bending, temp, mechanical stress) >> - others? >> >> In the case of transceivers: are there desirable modifications >> that would alleviate the problems? > > Acoustical sensitivity, low frequency changes. > > For optical clocks and frequency transfer, just the vibration from > traffic and other activity causes disturbances which disturbes the group > delay. What is done for these links is to actively compensate then using > a return path and closing the loop with a controller, very much like a > PLL. The length of the loop limits the bandwidth and hence how high up > the compensation can be done, so for longer stretches, this needs to be > repeated. They have now built links from PTB to SYRTE and NPL. > > Temperature shifts are slower, but also compensated though the active loop.
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”. Bob > > Close proximity to strong power-currents have also been shown to cause > modulations, so separate from power-cables if you can. > > Remember that the end nodes have very stable clocks, so their effects > can be taken out of the equation. For other setups, such as telco > operation, that's a completely different ballgame. > > Cheers, > Magnus > > _______________________________________________ > time-nuts mailing list -- time-nuts@lists.febo.com > To unsubscribe, go to > http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com > and follow the instructions there. _______________________________________________ time-nuts mailing list -- time-nuts@lists.febo.com To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com and follow the instructions there.