Comment on Optical fiber coupling to photonics chip

Optical fiber coupling to photonics chip by Wojciech Lewoczko-Adamczyk at https://www.youtube.com/watch?v=rHcfemtfP2s

Suggest you use the whole screen for the graphs, equations, diagrams. It is too small to read easily and I do not need to see your face while you are looking at the screen. Such are called “screen videos”, where you and the viewer are looking at the whole of something together. All your links are to generic pages, not to your work and these topics. The links are too vague and waste time tracking down this interesting topic.

Some very nice methods, materials and components. But you are right that there is a mismatch in size and methods.

My immediate impression from the mismatch would be to put an array of the chips on the whole face of the fiber, map the channel, and find optimal paths (primarily timings and correlations). If you have lots of fiber, use it. Don’t make the fiber more expensive to fit one chip, just put lots of chips on existing infrastructure. It means more sensitive detectors in 2D arrays, but that is not too hard. But you can’t hand if off to a graduate student. It becomes a statistical communications channel, not forcing small pieces to work together.  Putting N^2 channels through an existing fiber is a fun project, and challenging enough to keep your interest.  Just like disk drives and memory chips, you map the bad sectors, the bad and good channels.  Some would call it “lucky imaging”.  It will work for copper as well.

Richard Collins, Director, The Internet Foundation