3

Optical communication with (single-mode) fibers is widespread. Pretty much the complete internet backbone is based on it.

However, I am wondering a bit how the light is coupled into the fibers. Coupling light into a single-mode fiber is a difficult task because the light must be focused very precisely on the fiber core. Everyone who has worked with single-mode fibers in an optical lab knows that 1. Usually, coupling light into a fiber takes at least one hour, even if specifically designed fiber-couplers 2 are used.

Also, if you order fiber-coupled lasers, often the sales engineer does not guarantee that the fiber coupling still works nicely after delivery due to slight vibrations during transport.

How can you deal with this issue if you setup optical communication networks? I can hardly imagine that the technicians sit for hours in the basement beside the modem to adjust the fiber coupling. Do they use a very bright light source so that low coupling efficiency is not an issue? Do they use special connectors with tolerances in the nm range?

cakelover
  • 453
  • 4
  • 7
  • *"Usually, coupling light into a fiber takes at least one hour, even if specifically designed fiber-couplers 2 are used."* I haven't used those models, but the very similar equivalents from Oz Optics only take a few minutes to align for single mode work, as should the free-space variety [based on a lens and fibre mount](https://www.thorlabs.com/newgrouppage9.cfm?objectgroup_id=809) - once you're in practice at least – Chris H Mar 30 '23 at 13:06
  • The cable spec appears to be IEC 60793-2:1992; the usual connectors for pre-made patch cords just clip into a SFP transceiver. I believe for larger networks technicians splice on fiber by polishing the ends and melting it together, which is not an instant process but doesn't seem to be too difficult either. – pjc50 Mar 30 '23 at 15:13

2 Answers2

2

The basic strategy is to avoid free space entirely, since as you've noted freespace coupling is tedious and prone to misalignment. So you start with a laser source bonded to a fiber during alignment. Then you put a connector on that fiber using a ferrel or other connector that precisely aligns the fiber in the center.

From there onwards you can keep joining fibers together one after another by directly touching the cores. With good polishing, losses can be on the order of 10-20%. If index matching fluid is used (to suppress scattering), losses can be even lower. Since everything is mechanically attached, alignment can be very robust against shaking.

user1850479
  • 14,842
  • 1
  • 21
  • 43
0

So I guess the coupling between modem and communication fiber is probably similarly solved as in this application (handheld laser source with single-mode fiber connector): https://www.thorlabs.com/newgrouppage9.cfm?objectgroup_id=5929

These handheld laser sources have (afaik) a mating sleeve 1 built into their housing. I guess, the laser is internally coupled to a fiber and the internal fiber is then connected to the inner side of the mating sleeve. The mating sleeve is a mechanical connector which makes it possible to position the fiber cores of inner and outer fiber very precise to each other. Interestingly, with this setup very high coupling efficiency (> 85 %) can be achieved.

The next question would be, how the internal laser is coupled to the internal fiber. But as far as I know, there are some good techniques if the connection should stay fixed and does not need to be released. Also, at telecommunication wavelength (~1550 nm), coupling is mechanically probably also easier than at visible wavelengths.

cakelover
  • 453
  • 4
  • 7
  • 2
    I used to work on devices that coupled single-mode fibre at high efficiency by a manual alignment process in a jig, followed by UV-curing glue. Volumes there were a few per year, and about 10× that for the multimode equivalents that I had more to do with. In volume production this could easily be automated as you only need tiny movement over 3 degrees of freedom. Then coupling between single-mode fibres is a matter of very good engineering tolerances on the connectors – Chris H Mar 30 '23 at 13:03