Re: ATM Focault Testing

[Dwight Elvey <elvey@hal.com>]

"Meeks, Robert" <RFMeeks@epicsolutions.com> wrote:
>   I understand the theory fine, but what I can't envisage is how you
> can see the mirror?  Doesn't the knife edge block your view?  Especially if
> the knife edge is right at focus, it would have to move laterally quite a
> bit to intersect the light and seems like it would definately block your
> eyes view of the mirror.

Hi Bob
 The resolving knife edge does in fact partly block the view of
the mirror. Take a straight edge and bring it close to your
eye. You'll see the edge get fuzzy. This is because it is out of
focus but you can see past the edge because part of the light
goes past the edge and hits the lens of your eye while part
is blocked. None the less, all the light that just grazes
the edge still gets into your eye but it looks spread because
it is out of focus. This grazing light is from different
angles of the image you see.
 Now image what the mirror is doing. It
takes the light from the source edge and focuses the source edge
right on the resolving edge ( assuming it is adjusted at the
correct point).  The light from the source illuminates
the entire mirror and each part of the mirror contributes
to each part of the image formed at the resolving edge. Now let just
the finest sliver of light go past the resolving edge. This sliver also
contains light from the entire mirror but passing the edge at slightly
different angles ( the light cone ). This is because, with a perfect
mirror, the sliver has light from the entire mirror. When you place
your eye close to the resolving edge, you see the illuminated mirror,
not the focused source. Both the resolving edge and the image of
the source are out of focus in front of your eye. Of course,
this is for a spherical mirror with both source and resolving edge
at the center of curvature. If things are not exactly right with
the mirror, you'll see places on the mirror that are dark or overly bright.
This is because the light isn't being reflected such that it
just grazes the edge, it either is blocked by the edge or
misses it completely. This of course only works for a spherical
surfaced mirror and what we want is a parabolic surfaced mirror.
This brings us to zones. When looking at zones, we ignore all
but a small area of the mirror on either side. In this way,
we can find the center of curvature for each part of the mirror.
This can be compared to calculated values to determine the quality
of the mirror.
 Placing the light source and resolving edge slightly on either
side of the mirrors axis, causes a little distortion but it
is small enough to not cause too much trouble. In the slit-less
tester, they are usually top and bottom. This causes an apparent
astigmatism of the mirror. The top and the bottom focus slightly
different than the area across the center. In the more traditional
tester, the source and resolving edge are offset sideways a little.
This cause a little bit of coma. In both cases this isn't enough
to damage the value of the test.

> 
> I guess the quantifying question is:  How do you intersect the reflected
> light from your light source without intersecting your view of the mirror so
> as to analyze the shadows?

We just let the smallest of sliver of light pass the resolving
edge. It is a vary fine lateral adjustment. When everything is
adjusted correctly, we can even see heat waves between the tester
and the mirror.
Dwight