From: "Vladimir Galogaza" 
To: "ATM shore" 
Reply-To: "Vladimir Galogaza" 


>At 19:43 2003-05-11 -0700, dan otto wrote:

>What I did was to make a mask with  1/4 in. diameter holes every 1/2 inch
out from the center on an 8 inch f7.5 mirror.  I covered up all holes but
two at the >same radius for each measurement. One on the left side and
one on the right side. I then moved the camera toward or away from the
mirror until the images of the >slit merged into one image.  I can
measure the position of the stage to 1/1000 of an inch much like in the
Foucault test.
>Now the bad news. I could not measure where the image merged accurately
enough.  The image was about 60 pixels wide and looked merged within a range
of >positions of up to plus and minus 15  thou.  The Foucault Knife is much
more accurate than that.
>Any ideas how to make that more accurate.

James Burrows answered:
>Such small holes will give large diffraction patches, particularly with a
slow mirror.  As Suiter says, p. 284, relative to the Hartmann test, the
spots have to be >"identifiable, resolved, and not so large that
their positions are uncertain" - requirements on any test at the
focus.  You're sort of caught between the devil and the >deep blue sea
here; larger mask holes will produce smaller diffraction patches, but then
you're uncertain of the representative zone radius choice.


Why not Hartmanise this test modification. Instead of looking for the
longitudinal crossover position, register tester position for certain
lateral separation of
two spots in front of the crossover than find tester position beyond
crossing point where spots exchanged sides and are at the same or different
mutual lateral distance as at the first position. Than simple calculation
will determine crossover point hopefully better than by looking for
crossover point of two beams. If computerized, program can find centroids
of the diffraction spots, their mutual distance and  the same on the other
side and compute crossover point. Maybe better than by eyeballing.

Vladimir.

--- BBBS/NT v4.01 Flag-4