I’ve recently completed the first stage of modifying the mirror cell to convert it to use 6 point supports instead of 3 points.
By modelling deformations of the 350mm mirror in PLOP, I get the following figures for the current 3 points:
Support radius = 123mm
P-V: 1.65 x 10-4mm (~1/5λ at 500nm)
RMS: 3.81 x 10-5mm (~1/13λ)
For a mirror with a zygo report of 1/12λ P-V, this is less than impressive!
By using equally spaced 6 point supports, these figures are:
Support radius = 110mm
P-V: 1.07 x 10-5mm (~1/46λ)
RMS: 2.37 x 10-6mm (~1/210λ!)
So it can be shown that the deformations are potentially over an order of magnitude better using 6 points. Interestingly, allowing PLOP to run for a 9-point cell for the same mirror gives a slightly worse result than the 6 point, and though the 9 point is much better than a 3 point it’s harder to adapt this cell into.
My modification adds three “U” shaped supports that attach to the existing cell parts (being attached loosely via an M8 bolt to allow them to rock a little). There are two nylon mushroom head bolts that screw into these parts which the mirror will sit upon (these are the black cross heads in the image).
I think I’ll most likely need some mechanism to stop the bars rotating (but this might just be a bit of tape), and I’ll need new lateral nylon grub screws as 2 of the three current ones are well and truly stuck!