DIGITAL COMPUTER AIDS IN OPTICAL SYSTEMS DESIGN

• Document Type: CREST
• Collection: NGA Records (Formerly NIMA)
• Document Number (FOIA) /ESDN (CREST): CIA-RDP78B04770A001500060086-7
• Release Decision: RIFPUB
• Original Classification: K
• Document Page Count: 4
• Document Creation Date: December 28, 2016
• Document Release Date: September 26, 2005
• Sequence Number: 86
• Content Type: MAGAZINE

Reprint from:
Ad&
Approved For Release 2005/11/21 CIA-R DP78BO477
OA001500060086-7
data processing division ? port chester, new york
NGA Rev1 DrCYdlfletERelease 2005/11/21 CIA-RDP78BO477OA001500060086-7
 Alk
DIGITAL pJLO o I yif5/Aib?-RDP78BO477OA001500060086-7
IN OPTICAL SYSTEMS DESIGN
by EUGENE THORBURN
Optical Engineer
Pacific Optical Corporation
Common to the successful solving of practically all design
problems in the development of today's highly complex
physical systems is the mass of computations that must be
continually processed as the work progresses.
How well this load of computation is managed may well
be the determining factor in the profit or loss aspects of a
new development; at the least, it strongly influences the
degree of precision attained in the final product.
The formulation of the basic approach, the selection
of the appropriate theories and concepts, the garnering
of the necessary data, and the establishment of the correct
design procedures, these are all matters within the tech-
nical control of the design engineer.
But the rapid and accurate evaluation of mathematics
that represent the system's performance is not. For this,
the design engineer today is dependent on the extent and
appropriateness of the computing and calculating equip-
ment that is available to service him at the right time
and at the right place for maximum speed in the hand-
ling of this mathematical load.
Mathematical evaluations on what any system will
accomplish when it is built is called "proving out" the
design philosophy. And unless continuation of the design
approach can be maintained by frequent mathematical
evaluations accomplished quickly, it will either delay final
production of the system or waste a great deal of the
design group's time - a serious cost penalty in today's
Diagram (above) illustrates parameters involved in mathematical
representation of path of light ray through optical systems. System
performance is measured in terms of aberrations which are devia=
Lions of actual image from image derived assuming perfect lenses.
era of "profit-squeezing." Design complexity, with its
associated volume of mathematical problem solving and
performance prediction, is especially prevalent in the field
of advanced optics.
In optical system development, the mathematics em-
ployed actually predict the performance of the lenses and
lens systems in terms of the deviations of light rays pass-
ing through the system from those optical paths that would
give the desired object-image relationship.
These actual deviations from the theoretically required
paths are imposed by the physical limitations of optical
materials. Also, there exists always some finite difference
between the nature of the point-source world and physical
reality.
This relationship frequently imposes severe problems
and involved calculations in optical design because the
product of these deviations is the difference between the
object as considered by the optical system and the image
actually produced by the system.
The difference between the image and the object are
grouped under such nomenclature as focus, resolution,
depth of field, and other representative optical perform-
ance parameters.
No one has ever constructed a perfect lens system. This
is a physical impossibility. Fortunately though, it is possi-
ble to accurately predict what degree of imperfection any
given system will have.
But this process of predicting lens system performance
is highly complex, and its solutions require not only long
but also tedious mathematical calculations. The extenf
of the mathematical labor borders on the fantastic and
can frequently price a system right out of the realm of
practicality, certainly remove it from a competitive cost
range.
The length and complexity of the mathematics required
can rapidly be appreciated by considering that, prior to
the advent of electronic computing methods, a competent
lens designer often spent two or more years developing
and perfecting a lens system of average complexity. Nov,
with the rapid advances in the field of optics, manual
calculations by a designer could take a lifetime.
The majority of the designer's calculating time is not
necessarily spent in formulating the basic system design.
In most instances, he has to crank through the arithmetic
involved in determining the effects of various adjustments
in the component characteristics of the overall system. This
means that, whenever he substitutes one component for
a more suitable lens part, he has to recalculate these
changes and how they effect the overall system.
optical firms and computers
Before electro-mechanical desk calculators first came
along, the designer would spend years calculating any
given project. The new calculators reduced the figure
to months of computing time. But even this considerable
Approved For Release 2005/11/21 : CIA-RDP78BO477OA001500060086-7
 Approved For Release 2005/11/21 : CIA-RDP78BO4770A001500060086-7
n_e&ction left the overall cost in time too heavily balanced
in fi vor of essential computing time.
A though the later developed electronic computers are
now frequently applied in the more complex segments of
gent ral industry, their use by optical firms has been rela-
tively rare.
This lack of utilization might possibly be explained by
the fact that, although optical manufacturers have had
very definite need for their own computers, the initial cost
of computing systems in relation to the potential sales
volume of optical systems was difficult to justify by cost-
conscious management.
On the other hand, a management team in the optical
industry, aware of all contributing phases of their indus-
try's problems, needs to balance a considerable first fi-
nancial outlay against direct savings in design labor costs
and indirect saving resulting from more rapid completion
of final deliveries of the systems.
In a great many cases it has been proven that the in-
troduction of fully electronic computers greatly reduced
the amount of time spent in optical system development
and changed the status of optical designers from arith-
metical monitors to creative engineers.
Under the regime of the electronic computer, calcula-
tions as well as evaluations of the complex mathematical
representations of optical system performance requires
not months, but minutes, and more often seconds.
To the casual observer the mathematics predicting the
performance of a single lens or the composite of lenses
in a multi-lens optical system might seem rather elemen-
tary in comparison to the highly complex forms in use
in today's technology. And, no optical designer would take
issue with this observation. However, it is not the degree
of mathematical sophistication involved, but the sheer
weight of the computational burden that has turned the
optical designer to the use of digital computers.
This mass of mathematical labor is the result of optical
design being more of an art than science. While the re-
lationships between the behavior of light rays and the
characteristics of various media are exactly bound by un-
equivocal equations, the utilization of these relationships
to produce high performance optical systems depends to
a major degree on the judgement, experience and patience
of the optical designer.
The various optical aberrations that constitute the de-
viation of the actual image produced from that produced
by It theoretically perfect lens system cannot be singled
out one by one and corrected without certain penalties
in other aspects of system performance. This interaction
between the several forms of aberrations require that op-
tical designers operate in a constant state of compromise
to arrive at the 'optimum' design.
This necessity for compromising advantages and weigh-
ing disadvantages sets the requirements for the ability of
the designer to follow very carefully the performance
tret ds of the system as shown by the computations. On
the basis of the calculations, complemented by his ex-
perience and design judgement, the designer must make
those interacting changes and adjustments in lens configu-
ration, material, and system concept that will eventually
result in satisfactory system performance.
This cut-and-try procedure is as old as optical system
design, but at least relief from the drudgery of the com-
putations has been provided by the digital computer.
The figure on page 42 is a functional diagram of the
relationship between light rays and optical surfaces that
form the basis of optical system design. This interaction
and its associated mathematics must be investigated at
each optical surface of the system and for a large enough
assortment of light rays from various portions of the ob-
ject to give a proper evaluation of the system performance.
Opening and surface equations would express the relation-
ships at the first surface. Transfer equations would relate
the results of the first surface effects of the second surface
where the surface equations must be applied again. Clos-
ing equations would establish the coordinates of the light
ray as it reaches its focal point at the end of the system in
terms of height above the optical axis and distance from
the theoretical focal plane.
The final coordinates of the ray and the intermediate
orientations are exact traces of the ray's path through the
system. The deviations of this path from the path pre-
dicted by assuming perfect lens performance are a meas-
ure of the optical system's performance.
The final design of a lens system requires the thorough
investigation of system performance by means of this
Royal McBee's LGP-30 desk-sized digital computer is now operating
in the design department of Pacific Optical Corporation. The com-
puter's flexibility and memory capacity make it well suited for the
field of optical systems design, according to Pacific officials.
Approved For Release 2005/11/21 : CIA-RDP78BO477OA001500060086-7
 Aft i
Approved For Release 2005/11/21 : CIA-RDP78BO477OA001500060086-7
exact ray trace method. As many as thirty"rays testing the
performance of various portions of the lens surfaces must
be carried surface-by-surface through the system. The lens
designer observes the system performance in terms of the
ray orientation and distribution, and makes adjustments
and modifications required to optimize the system.
Fortunately, the nature of optical aberrations is such
that they may be pitted against each other to achieve
overall improvements. That is, carefully chosen use of
certain aberrations in some lenses may result in a counter-
ing of the aberrations resulting from other lenses to the
end that the final overall system aberration is much smal-
ler than the individual lens contributions.
This portion of the design procedure places the greater
portion of responsibility for success on the designer's
mastery of the "art." His ability to recognize the nature of
the aberration, his knowledge of the most efficient correc-
tive action, and his appreciation of the effect of the cor-
rective action on various other system parameters, marks
the difference between success and failure of the design. It
is in this area of the design effort that the digital com-
puter, by furnishing the designer with rapid evaluations of
the effects of his design judgement, proves most valuable.
To permit rough estimates of system performance dur-
ing the preliminary design stages, optical designers em-
ploy approximations to the ray trace equations which pro-
vide reasonable evaluation of the third order aberrations
and overall system performance. The usual procedure is
for the designer to prepare, on the basis of past experience
and theoretical performance calculations, the complex of
lenses and optical surfaces he deems necessary to perform
the required optical task.
Once the basic system has been established the third
order aberrations are computed. At Pacific Optical, the
LGP-30, purchased from the Royal McBee Corp , has been
programmed to perform this series of computations. The
capability of the program is such that systems consisting
of as many as forty optical surfaces may be analyzed.
To use the computer, the designer feeds in the curva-
ture, thickness, and index of refraction associated with
each surface of the system. The output of the computer
consists of the following aberrations: spherical, coma, as-
tigmatism, distortion, transverse longitudinal color, trans-
verse oblique color, and Petzval curvature.
These values are printed out in terms of the contribu-
tions of each surface, and the total value of each form of
aberration is also printed. Plotting these values permits
the designer to re-evaluate the performance of the system
and begin the series of modifications that will lead to the
final design.
Previously, the majority of design work was done using
the third order aberrations, except for the very final sys-
tem modifications, since the amount of computation was
drastically reduced in comparison to the exact ray trace
procedure. However, the utilization of the LGP-30 has
permitted more frequent application of the ray trace tech-
nique. The entire ray trace procedure has been pro-
grammed on the LCP-30.
As in the programming of the third order aberrations,
forty optical surfaces may be considered, and the inputs
of curvature, thickness, and index of refraction associated
with the several surfaces are all that are required. Com-
puter output consists of Y, sin I, and sin U, at each sur-
face plus values of Y, U, hs, and X? at the focal plane.
Consideration of the capabilities of digital computers in
optical design problems have led to the concept of utiliz-
ing the computer as a means of accelerating the optimiza-
tion process. Under the proposed system the basic optical
system would be established and the corresponding sur
face data fed to the computer. A suitable criterion for
optimal system performance would be established as the
computer objectives. A program would be proposed per-
mitting the computer to make adjustments in the charac-
teristics of the surfaces on the basis of systematic trial and
error operations.
Pacific Optical Corporation is devoting considerable
effort in the development of such a computerized design
program. In fact, the anticipation of the long range neces-
sity for and advantages of such a program had considera-
ble weight in making the choice of computers be pur-
chased. The flexibility and storage capabilities of the LGP-
30 make it suited for application to these computing con-
cepts, according to Pacific Optical officials.
In considering any segment of our rapidly advancing
technology, no part can be isolated from the whole. Every
science today is being buffeted and shaped by the needs
and demands of other sciences.
Our recent leap into space with missiles and satellites
has loosed a flood of demands for more precise and elab-
orate optical systems for visual tracking, astranavigation.
The streamlining of industrial manufacturing is open-
ing a broad market for optical measuring techniques yield-
ing increased resolution in process control systems.
Television is impatiently awaiting improved camera
lenses, motion pictures are desparately searching for bet-
ter depth dimension effects, and the progress of aerial
photo reconnaissance and mapping in three dimensions is
hungry for improved equipment.
Nor is the matter entirely one of merely broadening
and refining the product Along with expanding applica-
tions has come a compacting of the time with which these
new demands for optical equipment must be satisfied.
On both of these counts, broadened application and
sharply constricted delivery schedules, the in-plant, read-
ily available computer has become inevitable if optical
systems manufacturers are to meet their responsibilities
in the years ahead.
Approved For Release 2005/11/21 : CIA-RDP78BO477OA001500060086-7