Completed the mirror arm to match my calculated optical path.  3 grams, without the mirrors and it prints flat with no supports.  In openSCAD, hull() is your friend.  

Minus the original mind bending problem of modeling the optical measurements from extended and having to design/print this flat, hull() actually makes complicated shapes like this relatively painless.  I eventually figured out I could just rotate the optics model.  Duh..

The flat bits extending from the cylinders mate with matching flats on the head giving it a positive stop for positioning.  I originally added screws here for adjustment, but I didn't like how that turned out and deleted them. The stops are not clearanced so I can shave them down a bit to fine tune the stop angle.

I printed it out and the slot for the larger mirror I made was too thin.   The slot angle is critical and grabbing a pretty small area on the mirror, so I can't really trim the plastic it to fit and expect good results.  

I corrected the slot size in the model.  Printing another.