This is very cool, it’s nice to see someone has replicated the delta stage already, despite the sparse instructions! It’s worth mentioning that an older version of the delta stage was used in OptiJ so it has been used in microscopy applications before, though not by me.
The difference between this and say a delta-bot printer is that our range of motion is fairly small compared to the lever length. That means that we can approximate the kinematics with a constant matrix, so if you call your motors [a, b, c] then you can map:
x -> [-cos(30), cos(30), 0]
y -> [cos(60), cos(60), -1]
z -> [1,1,1]
It won’t be perfectly linear, but the deviation from linearity shouldn’t be much worse than the out-of-plane motion in the microscope, provided the mechanism bends only at the flexure hinges (which is a decent if not perfect approximation).
There was a design for microswitches in the feet, done by a previous masters student in Cambridge, which will be lurking in the repository somewhere! That could be fruitfully repurposed for this, I think - though if you build an inverted microscope around it I guess you’d want to have the endstops at the top rather than the bottom, so you can home it without crashing into the sample…