SPECIAL PROJECTS FLATS

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP67B00511R000100180027-3
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 8
• Document Creation Date: December 19, 2016
• Document Release Date: March 21, 2000
• Sequence Number: 27
• Publication Date: March 5, 1962
• Content Type: MEMO

Technical Memorandum
180027-3
March 5, 1962
Approved For ease 2006/04/13: CIA-RDP67B005111011
SPECIAL PROJECTS FLATS
The optical system for Special Projects was tested as a completely assembled
unit during December. Certain characteristics of its performance indicated
that some of the scanning flats were responsible for unacceptable amounts of
degradation. After disassembly of the unit, the flats were tested by a
scatter plate tecLnique illustrated in Figure 1.
Each flat, still mounted as it is to be used, was brought into a position so
that the optical path from the scatter plate to the flat was reflected verti-
cally downward to a spherical mirror. Thus, the scatter plate was positioned
at the center of curvature of the sphere as reflected by the flat, and the flat
was thus positioned at a 450 angle to the vertical corresponding roughly to its
normal position in use. The scatter plate was adjusted in focus to give as
nearly a straight central fringe as could be obtained. The adjustment, in
lateral position, was made to provide 6 to 8 fringes across the aperture.. A
mercury laps provided a source of light and the fringes were photographed by
a camera equipped with a Steinheil Quinar f/2.8 135mm lens. A yellow filter
was placed in front of the camera lens. Contours of the surface of each flat
were obtained by the steps illustrated'in Figure 2.
An enlargement of the negative was.made to provide a picture of the fringes
approximately 5x7" in size. By eye, small holes were pick-punched through
the center of each bright fringe,spaced random along the fringe, close enough
together to indicate the regions. of strong curvature. The print was then
overlaid with a piece of drafting paper on a light table and the small holes
were clearly visible. Three points are selected through which to pass the
reference plane; two of these must be along the same fringe and the third
should be separated by 4 or 5 fringes. The selection of these points will
determine the shape of the contours finally plotted, but with some skill
this method provides a reference plane very close to the optimum.
A straight line is drawn between the two points selected on the same fringe
and the perpendicular distance from this line to the third point divided
into the same number of equal spaces as there are fringes between the line
and the point. Lines parallel to the reference line are then drawn through
these points and additional points, with the same space external to the line
and the selected point. Since the flat is at 450 and is seen in reflection
twice, the space between two fringes represents t ?.17 displacement
of the surface. A small ruler is made up divided in such a way that the
spacing of the reference lines represents .3S4 X . The distance of each
 Approved For Pease 2006/04/13 : CIA-RDP67B0051190100180027-3
Special Projects Flats - 2 - March 5, 1962
pdnt pricked in the print from its corresponding straight line is measured
with this ecale and that number recorded next to the point. The contour
lines are then drawn by eye through the grid of numbers. Figure 3 shows the
contours of the four flats on the forward scanner.
It can be seen that flats 1, 2 and 4 are rather strongly astigmatic, while
#3 is saddle shaped. Number 4 is by for the worst, having some .0\ from
its lowest to its highest point and no juggling of the reference surface can
make a significant improvement in this value. Figure 3 is the best, having a
maximum deviation of only .25e\ , and it is felt that the subtraction of some
spherical power, which might actually be power in the flat or a small error in
focal position of the scatter plate, would greatly improve the appearance of
these fringes.
Flats # 1 and 2, having ranges of .3A and .4X1are primarily simple cylinders.
It is interesting to note that of these surfaces, only flat #3 gave images
approaching acceptable resolution.
Figure 4 shows similar contour plots for the four flats on the aft scanner.
Flat #2 appears to be much the best of the lot so far. Although the maximum
range of departure is still .2h , the majority of the surface area is well
within .1X. Flat #1 is strongly ast%matic. Flat #3 is almost pure saddle
with a range of .4\ , The axes of the saddle appear to go through the corners
on flat #3 and through the centers of the side on flat #4. Number 4 has an
extreme range of .4 X . So far then, we have two flats, #3 from the forward
scanner and #2 on the aft scanner, which might be good enough to use on the
next assembly.
Figure 5 shows flat #4 after it was removed from its mounting and, by comparison
of the before and after fringes, it is seen that the shape of the surface
has not been significantly changed. A slightly different set of fringes led
to a different choice of reference plant, but the over-all shape of the surface
remains unchanged. Flats R1, R2 and R3, also shown in Figure 5, are the so-called
"reworked" flats. Of these, R2 is much the best and begins to approximate the
quality which we require. Rl and R3,both with approximately .2 peak to peak
deviation, are of the same general quality as flat #3 and flat A2, and so we
should probably use those on the next assembly as well. It is hoped that the
technique which produced these reworked flats can be further improved to
produce results like R2.
Figure 6 shows two folding flats, neither of which have been used in existing
systems, and the folding flats in the camera are not available for this test.
Of these two, #9 is the better, although it would appear that both could be
used. It is to be remembered that two--inches is still to be cut from the
bottom of these flats and that, if this were done, a better reference plane could
be chosen. It would appear that both would then be within .1/\ peak to peak.
Approved For Release 20 R P67B00511 R000100180027-3
 Approved For Pease 2006/04/13 CIA-RDP67B0051190100180027-3
Special Projects Flats - 3 - March 5, 1962
Before closing, the following comments about the method are in order. It is
felt that this method is simple and straightforward and the reduction is a
routine matter, taking approximately one-half hour per set of contours. On
the other hand, the method probably does not provide an accuracy of better
than 1/20N as it now is employed. Locating the proper focus by eye is perhaps
not sufficient for accuracies better than this value and it is not possible
with this method to separate with certainty the spherical errors and astigmatic
errors in the flat under test. I do not believe that one would be misled by
the fact that a good flat would appear bad, but there is some danger that an
error interpreted as astigmatism might well be spherical and of the same
magnitude. In our present use, these two types of errors are equally destructive
of image quality. In the future, we would intend to use Fabry-Perot type
interference for the checking of the quality of these elements.
It is the conclusion of this investigation that the folding flat presently
in use in the forward camera should be retained since this complete assembly
has given good images. Folding flat #9 should be cut and coated and used for
the aft camera. The forward scanner should be provided with flat #3, presently
on it, and reworks Ri, R2 and R3. These should provide a nearly acceptable
forward scanner. Unfortunately, the aft scanner must remain the way it is,
with only A2 expected to produce good images. It is possible that new flats
will be available late in March, but we should proceed with the assumption
that they will not.
Approved For Release 2006 CIA.-RDP67BOO511R000100180027-3
 Approved For Foase 2006/04/13 : CIA-RDP67B00511 0100180027-3
SPh e,re,
F/,1t ur