Apologies if this is considered a necropost.
I wanted to update you and others who may be interested in our experience. We are and have been using optical fibers on our UAS and UGV. I have 10G fiber that can bend in a 1" radius by spec, but in practice we have tested it to 0.3" without issue (though we do not use or recommend that). Media converters are not as large as you’d think, unless you’re referring to the consumer “boxes” for Ethernet/RJ45 to fiber or similar. We use generic SFP+ modules for our high data rate applications and smaller transceivers elsewhere.
I don’t endorse a particular type or brand, but I’ll link some for reference:
- Broadcom/Avago HFBR-14xxZ and HFBR-24xxZ Series PDF
- Moog Protokraft’s transceivers, such as this LC type, and this D-SUB style one
- These are useful for their ease of termination for low-speed applications
Regarding cost/difficulty of termination, it isn’t that bad. Even from a marked up, retail outlet like amazon (useful for small orders) you can get the no-epoxy pre-polished quick SM SC connectors for $0.70/ea, SM LC for $2.50/ea, or MM LC (anaerobic glue) for $2.50/ea. Splices are similar in price. All the tools needed for that type of connector are available from amazon, ebay etc for $50-70. For the lengths involved in UAVs, the low loss of fusion splices are just not necessary. Compare to roughly 25 cents for a 8pin JST-GH and 8 three cent pins, and a 30 cent board header, it’s roughly between 50 cents to two dollars more per connector and $5 (and up) more for the board part. Termination of a single fiber with a quick connector is faster than terminating 8 crimps and installing them in the connector housing/shell.
@Kdub you are correct for the smaller UAS this will not be practical. For people who need the noise immunity, electrical isolation, or other benefits, it might be the right choice. On a large UAS with high-data rate sensors, it might be your only option. I’m not a die-hard fiber advocate, but I want people to know it’s an option, and you don’t have to jump feet first into the jaws of fusion splicing and 10G to access it. For example I made use of a setup using cheap ($1 to $5) 200kbaud transceivers and arduinos for high voltage (~3kV DC) isolation.
~edit for link correction, last time it was pointing at an anaerobic glue connector, that while faster than epoxy, was not the quick <1 min ones I was attempting to reference