essele wrote:Microswitches just seem inaccurate to me (fine for limits though), so I was assuming slotted optical sensors would be good although there is the risk of dust interfering, and I can't seem to find any repeatability figures.
I thought so too, but decided to experiment and have stayed with them. While I don't have any figures to report, I set X and Y home, bored a hole there with a 1mm endmill into the table. After cutting a few projects I can lower the same bit right into that hole without hitting the sides. I used the inexpensive microswitches with an integrated plastic lever. I suspect the repeatability accuracy of the microswitches might be even greater than the rigidity of the machine allows. Grab opposite corners of your worktable and see how much you can wiggle it. As an example, I used this setup with the two-sided machining of the Walthers PPK pistol model I posted a while back. It required very tight tolerances to come out without a noticeable offset between halves. [Edit: actually I can't be sure that I re-homed between halves, so that example isn't necessarily a good one
I guess the main point I wanted to make is to encourage experimentation. When I do I'm more often than not blown away by the machine's capabilities.