[MF3D.FORUM:1693] Optical Precision

[Tom Hubin <thubin@xxxxxxxxx>]

Hello Brian,

Brian Reynolds wrote:

> I'd like to talk about precision in optics.  Although it seems to go
> against common sense, you can get high degrees of precision in optics
> through very crude means.

True but you also need time and patience. I think the precision Don is
referring to is just the machining of mounts and assembly. Give or take
1mm is probably good enough and that is easy for any machinist. So it is
just a matter of proper layout and patience enough to go ahead and build
it. No optical precision required beyond purchasing appropriate mirrors.
   
> The process of grinding an optical surface is about as crude as you
> can get.  Traditionally one piece of glass is placed on a sturdy
> support (often a 55 gallon barrel), some water mixed with grit is
> sprayed on the glass and another piece of glass is pressed against it
> and pushed across it as you walk around the barrel.  Allowing the edge
> of one blank to pass over the edge of the other blank gives you a
> perfect sphere (concave on one, convex on the other).  After getting
> your sphere a careful amount of polishing gives a parabola.  If you're
> careful about not allowing the edges to cross you get a flat surface
> (which is what we want for mirrors in front of cameras).

Here is a little more on the manual process for making optical flats.
The surface on the bottom tends to become convex and the surface on the
top tends to become concave. I think that this is due to gravity as the
center of the top surface passes over the edge of the bottom surface. So
if you want a long focal length sphere just work the pair until perfect.
If you go too far then flip them over and continue. Some of the
curvature will come out as they tend toward flat and then reverse roles.

If your goal is flat surfaces then use 3 surfaces and work 2 at a time.
Two surfaces ground together will always form a convex/concave spherical
pair. By rotating through the set of 3 surfaces you end up with a flat
surface since that is the only spherical surface that can mate any 2 of
the 3. The other nice benefit is that you end up with 3 optical flats.

Tom Hubin
thubin@xxxxxxxxx