I read one ton of ore yields 36 ounces of water.
You read incorrectly. We don't know today what the admixture of ice and regolith is in those craters. As an example, its possible that banks of ice are sitting against the crater walls.
My ability to read is intact, but thanks for showing a concern
Edit: I think the NASA report said 32 ounces actually, but I was being generous in providing a 36 ounce per ton water ore to mine at close to liquid hydrogen temps in total darkness.
I also read the bottom of a crater has a bunch of big rocks and such to deal with. I sill vote for collecting the "waste" water off the crew. If you need more water, bring up some dogs or something and collect the water they produce. You aren't going to catch me walking in the bottom of one of those craters. Way to many ways to die on that mission.
This is a job for a really cool robot. I can't figure out how to get it up and down the crater walls. Maybe a walking robot is needed. I think NASA has looked at walking robots in the past. I know they have a first class robot department in building 9 at JSC. The even have a working copy of that robot that came to life in a movie a few years ago.
If they can't build this robot, it can't be built.
I am thinking some type of "robot camp" on the bottom of the crater and a conveyor belt to bring the raw ore up out of the crater. This would be better than a robot carrying the ore up itself. Can ball bearings operate at the temps required? Probably, but the design of them needs to be done. Or do you heat any area with a bearing