Also to be able to go down the elevator must have a bigger ballast than the main sub proportionally but that's easier to fix.

The game seems to calculate the ratio [total volume/empty volume] to determine the Y speed of a submarine.
I never tested the max dive speed but the speed when everything is empty is +18km/h whatever the size of the sub or its shape.

To keep the elevator working in any condition make sure that you have a room filled with water in your main sub to reduce its maximum Y speed (up).
My advice it to use that room full of water for artifact containment ; )
You will have to manually set your nav terminal zeroing as a side effect.

I think I may have figured out the problem. The test track is in perpetual motion (sinking or floating) when tested in the editor. If you enclose it with its own ballast and have it maintain position, then the elevator seems to work fine. Essentially, the elevator car cannot overcome the momentum of the track moving relative to it. Will continue to fiddle with it, but it seems like the elevator will only work reliably on a static structure like a colony. Alternatively, I suppose it could be used on a docked submarine for offloading of supplies from the bottom of the ship to the colony (which is what I hope to use it for!). May be difficult / impossible to implement on a moving sub without an engine that functions in the Y axis.

At first glance it would seem that you could just have the sub maintain position to use the elevator, but it looks like the ballast of the elevator car changes the buoyancy of the entire sub, so it's kind of like chasing its own tail.

Quick update - if I spawn the elevator car without the track and use a nav terminal to track velocity, it looks like it has positive buoyancy when empty and negative buoyancy when full, just as it should. This makes me think that the problem is with the elevator shaft. I can't tell whether the car is becoming stuck inside the shaft or something else preventing it from freely moving up and down.

Great ideas, thank you for the response! Very cool design - well done.

@Muffin Man this is because the room for the ballast is not actually 0.5 nuetral. It actually ususally needs some water still present to become nuetral if my memory serves me right... this is because there is three hull segments that hold water but only one gets targeted as the hull to calculate boyancy. Its sort of weird.

tbf I havent worked on this in a while. If you want to fiddle around with it be my guest. There is a signal function before the pump that serves to initate nuetral ballast level after docking. Mess with that. You can also always go into the editor and change the hull size for pump flow.

For some reason it doesn't want to become positively buoyant even when the ballast is completely empty. I can't figure out why.