Hey smtopps, I’m the 1678 alum who designed this system, you can thank @davepowers for referring me to this thread.
I had almost managed to forget that atrocious block of a markforged part in there…
Anyway @RoboChair was mostly spot on with his explanation, I just wanted to expand a little bit. The drum was the primary tensioning system. It captured the spectra line running down from the top of the elevator, and it was connected via hex shaft to a 1/2 ratchet we zip tied to the back of the 3D printed block. Once we had the parts dialed in it made tensioning extremely easy, all you did was grab a wrench and turn the shaft until it was tight enough. That being said, we had a lot of issues with the pulley rounding out or the cable crushing the printed material, this was definitely a case of 3D printed part abuse.
As Devin mentioned, the spring was for the most part unrelated to the drum. We put the spring on the return cable running down to the bottom of the elevator, so it never had to do any lifting. Exactly as @wgorgen did on their robot, we tied some cable connecting the top and bottom of the spring with enough slack that the spring could extend about an inch, then the cable would become tight and stop additional extension, to ensure we weren’t going to break the spring.
@Skyehawk iirc the spring would frequently reach maximum extension, but you’re right it most likely was not due to actual stretch in the dynema. I would guess it’s more related to slop in the entire system.