I think if I take a closer look and figure out how many degrees between steps, and draw that out for the distance between mirror and workpiece - I may not need 2 micron accuracy. I could possibly get this done using a high resolution optical encoder, if the math works out.