The large variety of optical instruments causes problems with economic design and production. Different skills are needed in each section of the optical industry and this complicates the training of operators and their supervisors. There is little similarity between machines and equipment needed for making precision camera systems, microscopes, astronomical and terrestrial telescopes, or various types of scientific instrument. Electronic computers facilitate lens design and tray tracing, to optimise tolerances, but the demand for higher optical performance has not made production easier. Prototype manufacture and small batch production will always be a problem and, in the future, more effort will be necessary to attract sufficient people to this essential craft. Production engineers responsible for process development must understand the techniques and underlying theory so more consideration should be given to their education and training.

The paper reviews the factors limiting the performance of ground based optical systems, and suggests a form of specification which overcomes many of the difficulties inherent in methods currently used. The main problems are the need to use diffraction based calculations of image quality and to ensure that the criteria adopted describe performance adequately over the large wavelength range now used in ground based astronomy. The paper considers the effects of atmospheric turbulence, aperture diffraction, detector properties and system wavefront errors. Scatter by the optical system is treated separately from other optical characteristics. Scatter by surface roughness (including coatings) and spider diffraction all affect optical performance and the effects on image quality are considered.

Aircraft Head Up Display (HUD) systems provide the pilot with essential flight, navigation, and in the case of military aircraft, weapon aiming information. The information, which is generated on a crt display, is collimated by an optical system and projected such that it is superimposed on the pilot's view of the outside world. The pilot is thus able to fly the aircraft 'head up' thereby reducing workload and enhancing his weapon aiming capability. The influence of the HUD engineer, with his awareness of the HUD performance requirements and constraints imposed by the cockpit and cockpit-pilot relationship, on the optical designer in his choice of optical system design is discussed in this paper. Examples showing how optical systems may be folded, extended and shaped to suit a particular aircraft installation in order to improve field of view performance are considered.

The successful development of various electro-optical systems is highly dependent on precise electronic circuit design which must account for possible parameter drift in the various piece parts. The utilization of a comprehensive computer analysis program (SYSCAP) provides the electro-optical system designer and electro-optical management organization with a well-structured tool for a comprehensive system analysis'. As a result, the techniques described in this paper can be readily used by the electro-optical design community. An improved version of the SYSCAP computer program (version 2.5) is presented which inncludes the following new advances: (1) the introduction of a standard macro library that permits call-up of proven mathematical models for system modeling and simulation, (2) the introduction of improved semiconductor models for bipolar junction transistors and p-n junctions, (3) multifunction modeling capability to link signals with very high speed electronic circuit models, (4) high resolution computer graphics (both interactive and batch process) for display and permanent records, and (5) compatibility and interface with ad-vanced engineering work stations. This 2.5* version of the present SYSCAP 2 computer analysis program will be available for use through the Control Data Corporation world-wide Cybernet system in 1983*. This paper provides an overview of SYSCAP modeling and simulation capabilities.

In recent years spatial-frequency-based criteria (Optical Transfer Function and figures of merit based on it) have become widely used to define and describe imaging performance of optical systems. Over the same period electro-optical raster display systems have become ever more widely used. In applying spatial frequency criteria to raster display systems it is important to acknowledge the details of the image sensing mechanisms in the human visual system. When this is done it is found that there are serious limitations to the general applicability of spatial-frequency-based criteria. The paper summarises some properties of the human visual system which are important in assessing the raster display/ observer interface and outlines a mathematical model which attempts to define information transfer at this interface. Results of a highly controlled experiment aimed at confirming some of the interface characteristics are presented. The implications on such quality measures as the widely used thermal imager performance measure MRTD (minimum resolvable temperature difference) are discussed.

Referring to edge enhancement effect obtained with detuned interference filters already reported, a model is discussed where the source is made of a narrow annular aperture in the front focal plane of a lens, producing a narrow spatial frequency band illumination with circular symmetry in the back focal plane. Such a plane is made to coincide with the object plane of a 4-f optical processor. In the Fourier plane of the processor a circular aperture is placed, to perform low-pass spatial frequency filtering whose cut-off matches the spatial frequency of the illuminating configuration. In reasonable operating conditions the overall processing turns out to be described by a directional Hilbert transform of the object transparency along radial lines of the Fourier plane. Computation and experimental evidence are presented for the case of a simple object as a slit, showing up edge enhancement and background attenuation according to the image intensity distribution which consists of two equally energetic terms, the one associated to the geometrical image and the other to the sharp edge effect.

Stray light rejection characteristics of an optical system need to be carefully evaluated early in the design stage. Efforts to reduce the impacts of stray light can completely alter the optical configuration. The ability to produce acceptable surfaces of a given size is influenced by the required scattering characteristics of the optical elements. Sunshade requirements and associated baffle design also may often determine packaging dimensions. The usefulness of reimaging systems to control diffraction and thermal radiation is discussed. A brief review of the major tools used to determine stray light rejection characteristics of an optical system is presented. Examples illustrate the need to consider stray light characteristics at the first design phase for optical systems. The increased cost in time and complexity to modify a design originally based only on optical imagery is also shown in examples.

For several years the hand-held programmable calculator has been known to be an attractive tool for solving some relatively simple problems in optical computation, such as general ray tracing, first order optics and optical system analysis. The range of application can however, be extended far beyond this level and includes the semi-automatic design of unsophisticated optical imaging systems such as the air-spaced doublet and the Cooke triplet, for which examples are presented. In view of the limitations in the number of program steps, it is mandatory to implement such programs with utmost economy of use of these program steps and to utilize the inherent properties of these optical systems in order to obtain convergence in the automatic design. improvement part of the programs. While the most appropriate application may be in supporting the teaching of elementary lens design, these calculator programs can be quite helpful if relatively simple optical systems have to be designed without as access to a large computer facility and associated software.

An optical design program has been developed. It contains four main sections: Section A: small analysis, which contains first and third order aberrations and a tracing of a few selected rays to give exact ray aberrations. Section B: semi automatic optimization routine, which enables the user to bring the aberrations computed in Section A to a target value. The user can choose between different optimization algorithms: least squares, damped least squares and Glatzel's adaptive correction method. All methods are guided by a newly defined steering function which takes the non-linearity problems into consideration. The QR-Matrix-Decomposition method is used to compute the solution of the problem. With this method the efficacy of the parameters or of the different aberrations can also be analysed. Linear dependences between parameters and between aberrations can be detected and eliminated so that the damping factor can be used to handle non-linearities. The same can be done by factorized solutions. Section C: analysis of optical systems, again including first and third order aberrations, ray aberrations and a geometrical modulation transfer function. Section D: service programs including individual lens analysis for mounting and manifacturing purposes, zoom systems analysis, addition and subtraction of a surface, lens scaling, data printing on cassette recorder, reading data from the recorder, refractive indices for "normal" glasses, ray height analysis, etc.

The very fast evolution of the Hardware and the software brings the optical designer to choice betwen two attitudes. 1) To use the services of a specialized company which is continusly devoloping optical programs. 2) To write its own programs and improve them according to the needs. Theory and experience have to help themselves to realize an harmonious balance in order to get product improvements through programs improvements. CERCO has choosen the second alternative.

Since the 1980 Mills College conference, there have been significant advances in personal computer hardware, with increased speed and memory size. Software developments are affected by these advances; this paper is concerned with both hardware and software developments, and with expected future trends.

We have developed a new optical design program that we call Super-Oslo, which contains a superset of the capabilities found in our earlier OSLO programs. The new program was developed using the Hewlett-Packard 9000 Model 20 desktop computer, and also runs in the Hewlett-Packard 200-Series computers. The program employs an enhanced damped-least-squares optimization algorithm with a merit function that may be based on aberration coefficients, ray data, or user-computed data. A major feature of the new program is its incorporation of an interactive computational/control language.

Wavelength division multiplex systems can be studied with ray tracing and numerical convolution methods. Some considerations to form a model for microcomputer-aided design of an "all-glass"-system based on a GRIN rod lens and a prism grating are presented. Calculated system parameters agree with measurements. A realized experimental demultiplexer indicates insertion losses between 0.9 and 2 dB.

A newest optical calculation program has been completed on the Canon BX-1 desktop computer. Some of skills which are utilized in this program are described, especially, the program segmentations are emphasized.

It is assumed that the powers and movements of the components in the zoom system are known. One position of the zoom is designated the reference position, and the primary (third order) aberrations, including the spherical aberration and longitudinal chromatic aberrations of the pupil, of each component are used as independent variables. For the reference position and for a series of other zoom positions, paraxial marginal and pupil (principal) rays are traced. Initially these rays are traced assuming each component to consist of thin lenses in contact: subsequently they are traced through designed components. Using the data from these rays, the primary aberrations of each component in any of the other zoom positions are expressed as linear combinations of their values in the reference position. Special cases can arise, such as when the pupil in some other zoom position coincides with the object position in the reference position. These special cases lead to the need for four separate sets of coefficients, and a simple initial test indicates which set is to be used. In cases where a given component acts nearly as a field lens, the primary aberrations of the pupil for the reference position are used as independent variables in place of those of the image. The total primary aberrations of the whole system in the selected zoom positions are set equal to zero, or are given target values, and the resulting set of equations is solved by the method of damped least squares, using weightings for the overall aberration residuals and damping factors for the independent variables. For the initial, thin-lens, design there are only three independent variables (the spherical aberration, coma, and longitudinal chromatic aberration). The values of these needed for stable correction of the primary aberrations are determined, and the different components are then designed. The resulting thick-lens components will usually show greater variations of the aberrations on zooming than those predicted by the thin-lens solution. The formulae employed, which are perfectly general, are then used to obtain changes in the values of the primary aberrations of each component in the reference position (possibly with target-values for the overall primary aberrations in each zoom position if ray-tracing has revealed the higher order aberrations present) prior to re-design of the components. The technique is also useful in studying ways of improving an existing design in an economical way; for example, by imposing the condition that only one component may be subjected to re-design.

A computer program for the optimization and evaluation of optical systems, successfully used since 1978, is the object of the present article. Although it is quite simple when used for designing routine optical systems, the considerable amount of its flexibility makes it particularly suitable in handling complex optical systems. As a matter of fact, it can be applied in handling a variety of cases: cylindric and aspheric lenses, tilts, decentrations, afocal systems, multi-configuration mode systems. In order to accelerate convergence, a wide range of options concerning computing methods are provided. In fact, the choice of the mathematical method, of the way of computing the damping factor and of the way of handling boundary conditions is readily available. The use of solves, such as the optimization on best focus or on paraxial focus, reduces the number of variables, often accelerates convergence and, in some cases, prevents stagnation. The capability of reducing sensitivity to manufacturing errors, besides the automatic choice of glasses for any couple of wavelengths achromatization, makes it very effective in cost reduction. Moreover, weights on field, wavelength and aperture and many constraints among different elements and/or configurations can be assigned. System evaluation includes, among other features, diffraction and geometrical MTF and semiautomatic Statistical manufacturing tolerances. Moreover a wide variety of graphical outputs is provided: spot diagrams, MTF plots, layout plots, transverse ray aberrations, glass maps and manufacturing drawings. Results obtained with the present program are reported in the paper.

During the preparation of the paper describing the design and manufacturing of micro-electronic lenses it was observed that the wave aberration coefficients in the sagittal fan did not converge exactly to their axial values, when the field was asymptotically approaching the axis. Further work, which involved the calculation of partial derivatives of wave aberration with respect to curvatures and thicknesses along selected rays, also gave systematic errors when compared to the asymptotic values obtained from finite diff-erence calculations. The formulae used for calculating the wave aberrations were those introduced by H.H. Hopkins and their approximations are described in the above references. However, the design of an x-ray telescope for 5O Å wavelength had forced the authors to reexamine the basic ideas.

The advantages of a disc file based data management system for optical design are outlined, and the file structure of a program package written for the HP1000 computer explained. The sacrifice in computation speed involved in disc processing is redeemed by the efficiency of the whole design process and the response of the system to an interactive user. The use of currently available microcode routines can improve OTF and raytrace computation speeds.

The effect of altering the magnitude of a stabilizing factor used in the PSD lens design optimization technique is discussed. Analysis of this effect leads to a method of accounting for the influence of the inhomogeneous second partial derivatives. The improved convergence rate of this method will be demonstrated using examples obtained with the SYNOPSYS lens design program.

The most effective environment for doing lens design continues to evolve as new computer hardware and software tools become available. Important recent hardware developments include: Low-cost but powerful interactive multi-user 32 bit computers with virtual memory that are totally software-compatible with prior larger and more expensive members of the family. A rapidly growing variety of graphics devices for both hard-copy and screen graphics, including many with color capability. In addition, with optical design software readily accessible in many forms, optical design has become a part-time activity for a large number of engineers instead of being restricted to a small number of full-time specialists. A designer interface that is friendly for the part-time user while remaining efficient for the full-time designer is thus becoming more important as well as more practical. Along with these developments, software tools in other scientific and engineering disciplines are proliferating. Thus, the optical designer is less and less unique in his use of computer-aided techniques and faces the challenge and opportunity of efficiently communicating his designs to other computer-aided-design (CAD), computer-aided-manufacturing (CAM), structural, thermal, and mechanical software tools. This paper will address the impact of these developments on the current and future directions of the CODE VTM optical design software package, its implementation, and the resulting lens design environment.

Convergence properties of power series expansions in Seidel aberration theory can be studied from exact mathematical descriptions of wavefronts refracted by spherical surfaces. A generalization of Huygens' principle is suggested in order to study composite systems. Legendre transformation allows us to relate the solutions of eikonal equation to Hamilton characteristic functions.

The air gap between adjacent lenses can be regarded as an "air lens" imbedded in glass. In many situations these air lenses, in glass, have glass lens equivalents, in air. The resulting new designs, when certain of the original air lenses are removed, often have distinctly better aberration correction, or other features that are desirable, such as a different system length. This transformation is easiest when the air lens that is to be removed has a nearly concentric meniscus shape. The glass lens equivalent then has about the same shape, and is located outside the two lenses bounding the air gap. The latter disappears and the two adjacent lenses combine into a single lens. Examples of this transformation, and the resulting benefits, are shown for an air-spaced aplanatic doublet, a Cooke triplet, and a Houghton catadioptric system. This point of view is very useful for generating new design variations from well-known types.

Utilizing correspondences between actual residual aberrations and third order aberration coefficients, lens types and surface characteristics, which relate to final optical performance are analyzed with the latter in preliminary lens designing.

Instead of following the historical path of its development, the modern Double Gauss lens in its many forms is synthesized from the ground up. Each step is analyzed in simple terms, mainly by considering the contributions to the Petzvalsum correction by means of bending, spacing, and index difference, and by evaluating some characteristic aberrations. This pro-cess causes even a complex and difficult to understand design as the modern Double Gauss lens to reveal its inner workings. The relevant historical events are inserted at the appropiate points. It is of interest to note how far the historical development departs from an a posteriori synthesis. It has been argued that Gauss's contribution has been no more than his name. This issue is briefly addressed. The reasons for the tremendous flexibility of the Double Gauss lens becomes apparent from the synthesis and is illustrated. It is remarkable how the Double Gauss lens maintains its characteristics throughout all its forms, even to the point that its optimum performance in terms of the characteristic aberrations can be predicted once the aperture and fieldangle are choosen. A simple formula to this effect is presented and its accuracy tested by some typical examples from the patent literature.

The design and the imaging characteristics of wide-angle lenses are described, with reference to back focus determination. To define the characteristics of the lenses presented, an evaluation is made of the most conspicuous aberrations when wide field, angles are imaged. In addition, the mass and the volume of wide-angle lenses are dealt with critically.

Real time optical design can be done by microcomputer which integrated the optical design program with optical bench and lens sets together. The surface distance can be input to the computer by digit micrometer which connected to the lens holder in the bench. Lens datum input can be done simply by key in its code number which are connected to the lens library subroutine. With high speed program, the aberration, spot diagram, focal length, etc. can be check with the bench accessories such as the interferometer and the test chart directly. High speed optimization program and the servomotor to drive the lens holder to suitable positions are also worked.

The BOLEX ASPHERON first, and now the CETO PANSPHERON has been built for the focal length reduction of a basic positive lens.

Wide-field correction is achieved in reflective versions of the Schmidt and Schmidt-Cassegrain systems by the use of a fold mirror serving also as a corrector. The deformation of the corrector surface is divided into two functions one-with and one with-out rotational symmetry. The bulk of the deformation is defined by the rotationally symmetric part while the residual, which is of a much smaller magnitude, is nonrotationally symmetric. This separation of corrective polynomials facilitates fabrication of the elements. The larger rotationally symmetric part can be diamond turned on a two-axis machine under numerical control, whereas the small residual deformation can be hand figured to an overall system null or to a computer-generated holographic reference in an inter-ferometer. Diffraction-limited performance in the infrared region is achievable to an extent determined by tilt angle of the fold mirror and the f-number.

The microscopic examination of a planar sample in an oblique manner is proved to be feasible, indeed easy, using in conjunction with a microscope a simple attachment consisting of a unity magnification system and a diffraction grating. The illumination must be mono-chromatic. The unity magnification system focuses the sample onto the grating. The grating is chosen such that the first diffracted order corresponding to the incidence used in the arrangement is perpendicular to the grating. The grating is observed with the microscope in the conventional manner. The principle, properties, capabilities and limitations of this novel and versatile optical arrangement are explained and reviewed.

Basic optical properties of dielectric thin films for interference applications and of metallic optical coatings are reviewed. Some design considerations of how to use thin films best in optical systems are given, and some aspects of thin film production technology relevant to the optical designer and the optician are addressed. The necessity of proper specifications, inclusive of test methods, is emphasized.

The aspherical surface manufacturing process improvements let suppose equivalent improvements for non-rotationnaly symmetry optical surface. For this we introduce in our optical design program such surfaces considered as very useful for some applications . We are interested in surfaces with two symmetry planes represented as a power series expressed as follow : X=Σ a_mn Y^2m Z²ⁿ m+n ≤ 3. These surfaces can be tilted or decentered for any direction. The nine parameters used to characterize the surface, the five eccentricity parameters can be used as optimization parameters. The influences of each parameter for each ray are expressed with the exact partial derivatives according to the traditionnal CERCO method . Two exemples ares described - helmet mounted wide angle CRT display . - Off axis catoptric system for 10° field .

A unique mirror surface profile has been developed to produce annular, radial, or circular images of uniform intensity from incident radiation of non-uniform illumination. The design principles are ideally suited for industrial high power laser applications as well as for concentrator/integrator mirrors in the Abbe type condenser systems. The reflective surface is a series of segmented conical flats tangent to a discontinuous base aspheric. Each conical segment of specified width redirects the input light to a common image. The net contribution from all of the segments is a highly uniform intensity integrated image. The design procedure is represented in three stages, (1) optimizing the curvature and aspheric coefficients of the base curve; (2) calculating the point of tangency and width of each conical segment; (3) shifting the coordinate system to obtain the required annular or radial image dimensions. To facilitate the design, a semi-automatic computer program was written since commercially available optical design software does not handle discontinuous surfaces. Three examples of mirror designs will be presented.

A computer program has been developed to analytically describe the first and third order properties of systems with large tilts and decentrations. The program describes the imagery through analytic formulae, in both the pupil and field variables, rather than by interpretation of ray trace data at sample points in the field. The first order properties of both the image and the pupil are described in terms of tilt, decentration, magnification, keystone distortion, and anamorphic distortion parameters. In computing these parameters, it is important to take into account the transferred components of the parameters, which are due to the first-order properties of the previous surfaces. The third order properties are computed by representing the surface aberration contributions as vector fields and summing them vectorially. The third order properties are described in terms of a coma field with a node displaced from the center of the image, a binodal astigmatism field, a trinodal distortion field, and a focal surface that is curved and displaced from the Gaussian image in both the transverse and longitudinal directions. The program is applied to several systems, and the results are compared to ray trace data. The interaction between large first order aberrations of intermediate images and pupils, and the third order properties of the image is discussed. Applications of the vector aberration field technique toward an automated design procedure are considered.

An ultra-small convex glass-rod lens with a diameter of 0.5mm was developed, and a coupling efficiency of 21% between laser diode and single mode fiber was attained. This value is considerably higher than that of commercially available graded-index lenses. Also, a "hybrid lens" combining lens and prism was developed and showed good result for image transmission.

In this paper we present a set of trade study analysis relations that will enable the user to estimate performance of advanced technology cooled laser mirrors. The effects of design parameters frequently proposed for use in advanced technology mirrors, such as very dissimilar materials and light-weighting, are included. The analysis tool is primarily intended to enable one to rapidly and quantitatively rank advanced technology mirror concepts as part of performing benefit versus cost, schedule, and risk trades. Other uses include generating preliminary distortion inputs for optical codes, and obtaining insight into what material properties are desirable for different parts of the mirror. The approach used to derive the method is to separate the total distortion into a number of components using simple, but carefully contrived finite element analyses. Individual scaling laws are derived for each component. The end result is a tool that correctly predicts the performance of many advanced technology mirrors because high technology mirrors frequently involve design parameters that this approach includes accurately.

Both acrylic plastic and glass are common materials for hyperbaric optical windows. Although glass continues to be used occasionally for small windows, virtually all large viewports are made of acrylic. It is easy to uderstand the wide use of acrylic when comparing design properties of this plastic with those of glass, and glass windows are relatively more difficult to fabricate and use. in addition there are published guides for the design and fabrication of acrylic windows to be used in the hyperbaric environment of hydrospace. Although these procedures for fabricating the acrylic windows are somewhat involved, the results are extremely reliable. Acrylic viewports are now fabricated to very large sizes for manned observation or optical quality instrumen tation as illustrated by the numerous acrylic submersible vehicle hulls for hu, an occupancy currently in operation and a 3600 large optical window recently developed for the Walt Disney Circle Vision under-water camera housing.

Aberration corrected holographic diffraction gratings have become the dispersive element of choice in many spectroscopic instruments due in large part to the fact that they are often the only optical element required. These gratings are used in various configurations including monochromator, flat field spectrograph and monograph (scanning spectrograph). This paper shows how these configurations can be optimized for best results for a given problem.

A computer program has been written which permits us to do aberration calculations on holographic optical elements based on ray tracing. In this article we first examine the diffraction phenomena which leads to a simple method for defining light ray deviation. Because most optical systems with holographic optical elements have no circular symmetry it was necessary to give appropriated definitions for the Seidel-aberrations. The structure of the computer program is given and a practical example is worked out.

The diffraction efficiency of a grating depends critically on the groove profile, which in the case of holographic gratings is usually more or less sinusiodal. In order to obtain high efficiency it is thus of great importance to have accurate control of the groove depth during the fabrication process. Usually the groove depth is determined by adjusting the exposure, and then developing the substrate for a fixed time. In this work, a technique is studied where also the development time may be varied in order to obtain the desired groove depth. After interrupting the development, the groove depth of the photo-resist grating, without metallic coating, is determined by measuring the diffraction efficiency in tranmission and comparing it with theoretical efficiencies. Good accuracy is achieved since, in the groove depth range where the efficiency of the final reflection grating has a maximum, the transmission efficiency of the photoresist grating increases rapidly with increasing groove depth. A great advantage is achieved by using a HeNe laser for the efficiency measurements. Since the resist is insensitive to red light, the measurements do not affect the sinusoidal solubility distribution, and the development can be continued to increase the groove depth, if desired.

A new technique concerning the production of blaze holographic gratings by using Fourier synthetic method is proposed. The description of the holo-grating ( 1200 G/mm, blaze angle 8_o 38', blaze wavelength 2500 Å ) with synthetic groove profile is given and the comparison with the theoretical result is conducted. The calculation of the phase-shift of reflection light wave is also given. An experimental setup which is simple in construction, convenient in adjustment and high precision in control is put forward.

Certain types of holographic optical element have been used as collimators and combiner mirrors in Head Up Displays (HUDs) and their unique properties have made possible a number of optical configurations. The elements considered are reflection holograms of point sources with the fringes roughly parallel to the surface. A particular element may be required to behave both as an efficient reflector and as a good transmitter at adjacent angles. Tight control over the fringe spacing and the degree of refractive index modulation is essential especially in systems requiring a combination of holographic elements.

A gradient index capability has been implemented in CODE VTM, a comprehensive lens design and analysis program. The Sharma' algorithm is used for raytracing. This algorithm is very efficient for determining the position of a ray but is less efficient for calculating the optical path. A series expansion of the optical path integral was implemented which improves the convergence of the optical path calculation. The terms in this series can be used to determine an optimum step length for the numerical solution of the differential equation. Several types of gradients may be analyzed including radial, axial, spherical and combinations of these. Examples will be given.

An achromatized Fourier processor working with spatially coherent but temporally incoherent light is described. A specific configuration is proposed wherein two zones plates combined with an achromatic objective permits to cancel the chromatic scale factor as broad-band illuminaaon is used in the diffraction process. Experimental results demonstrate the ability of the system to work over the visible spectrum.

In this paper we will first of all recall the general aspects of a WDM device before discussing devices that use active components or passive components, based on filters (dichroic or Fabry Perot) and based on dispersive elements (prisms or gratings). We will then conclude with a description of a 4-channel optical WDM device.

The mastering of an industrial process to economically manufacture high precision aspherics is of utmost importance to obtain high performance infra-red systems, in the 3-5 and 10-13 micrometers ranges. The research and results of such a process is presented. Starting with a close spherical surface, material is being brought by vacuum evaporation. Between the rotating lens and the source, one or several masks of specific shapes determine the aspherical curvature profile. Up to 50 micrometers of material can thus be added to a sphere.

The effects of reducing the transmittance and MTF of the optics in a thermal imager on the subjective performance have been shown, as has an instrument for measuring the full aperture transmittance spectra of IR lenses. MTF degradations for large (ca. 2510mm diameter) monocrystalline germanium blanks have been described.

There are indications that we are entering an era wherein economics will play an increasing role in the optical design and production process. Economics has always been a factor in the competition between commercial ventures in the product arena. Now, we may begin to see competition between different technologies for the scarce resources of the society, including money. A proper design approach begins with a thorough examination and refinement of the requirements from the top down. The interrelationships of the various components must be properly understood and balanced. The specifications must be clear, complete, and realistic. Improper or incomplete system design can cause an extensive waste of resources. The detail optical design to meet the performance requirements has sometimes been the only part of the process that the designer has considered his own responsibility. The final optimization should also consider economic related factors: the cost of tolerances, the available tools test plates, materials, and test equipment. In the preliminary design stage, he should have decided which alignment and test methods are most appropriate to the system. The distribution of tolerances in an optical/mechanical system is a frequently neglected opportunity to reduce cost. We have reported previously on our work in this area, and expand further on it in the context of this paper. The designer now has an opportunity to generate better designs at a lower cost that are more economical to produce. The watchword for the 1980's may become the one found in the assembly automation industry: "more, better, for less".

The developments in the field of lasers and laser optics have placed increasingly demanding requirements on optical components designed for use in state of the art laser systems. Specifications for these components are often imposed without considering the requirements of the fabricator and can result in excessive costs and lengthened delivery times. The dichotomy of the specifying user and the component fabricator needs to be explored and resolved industry wide.

The requirements for testing the optical systems associated with advanced EO sensors and displays have become more and more demanding over recent years. Using conventional visual optical test techniques, optical contractors would be faced with prolonged testing programs to ensure compliance of the latest generations of optical equipments with the relevant specifications. The advent of the mini-computer and the development of signal processing techniques has opened up a new vista to those involved in optics testing. The paper will describe examples of how optical requirements have been interpreted and the approaches that have been adopted to proving optical performance.

The main features of this new laboratory instrument are: - measurement of the lens with the optical axis in the vertical position (as in practical use) - angular measurements by means of photoelectric sensors, rather than by visual observation. Thus, geometric calibration can also be carried out in the near infra-red range of the spectrum - automation of the measurement procedures, including spectral adaptations. The characteristics of the system are described, some technical data are given, and typical results are shown and compared with other methods. The problem of photoelectric orientation is also discussed.

The Modulation Transfer Function (MTF) imaging properties of an aerial reconnaissance lens have been tested using a recently developed, highly automated OTF measuring system. Combining sophisticated software with knife-scanning technology, the instrument is the most advanced OTF system ever built. Based upon user entered lens-under-test properties and specified tolerance requirements, the measuring system automatically determined the image analyzer focusing increment, number of sampled points per image scan, scan velocity and signal filtering through the use of various algorithms and computational techniques. Using a novel coarse and fine focusing routine, the instrument determined the plane of best focus for both sagittal and tangential information. As a final summary report, it automatically quantified and displayed, both alphanumerically and graphically, the "best focus" OTF, thru-focus OTF, astigmatism and field flatness data. A discussion follows of this new system's attributes and profound impact on classical lens testing techniques.

A monocentric axicon camera has been designed for use in high precision photogrammetry with steel ball or point source targets. An example has been constructed and assembled successfully in a prototype camera body. The axicon patterns which it records remain sharp and symmetrical for all object distances from 300 mm to infinity, over an angular field 40o in diameter. Theoretically the distortion of this camera should be zero, and first measurements indicate that it does not exceed 100 nm across 50 mm plate-width.

In order to pilot and visually monitor the Canadarm operating on the shuttle, a zoom lens with a minimum wide angle of 98° field of view was needed. The design of the chosen solution is presented. There are 22 lenses arranged in 4 groups allowing mechanical compen-sation. The range of focal lengths is 8.2 to 25 mm ; the aperture is f/3.6. The weight is 1.6 kilos.

There are zoom system applications in which transverse image displacement caused by centring errors can be disturbing. By choosing an optimum reference point (adjustment of reference mark) image displacement is minimized. Formulae are given which permit residual image displacement to be already calculated from the design with thin lenses. Examples show that different types of zoom system differ markedly in sensitivity to this effect of centring errors.

The paper is intended to familiarise optical designers with mechanically and optically compensated zoom lenses which operate in the thermal infrared waveband and especially the 8-13 micrometer waveband. Apart from illustrating various infrared zoom lenses a description and performance analysis is given for two final designs.

A low attitude infrared earth sensor for 3-axis stabilised satellite needs wide angle field of view of the order of 120°. Since the requirement of the optics is only to image the earth as a sharp circular disc, it can be achieved by a conical lens which can be fabricated by generating an axially symmetric cone into a near hemispherical lens. The angle of the cone is same as the earth's subtended angle. The design of the lens was verified by fabricating a lens for the visible region.

Reflecting and catadioptric lenses have been used in astronomical telescopes for many years. They have recently been widely used in large aperture and man portable image intensified night vision equipment. The afocal telescope used with scanning infra red systems operating in the 8-12μ waveband generally use refracting telephoto objectives. These can be configured either as a single field of view lens, as part of a dual or multi field of view switchable system or as the basis for a mechanically or optically compensated zoom system. However, for physically large high magnification telescopes, catadioptric systems can offer advantages over refractors. Two types of catadioptric lens are described, the first (x7.5) utilising a full aperture and the second (x30) a sub-aperture germanium element to correct spherical aberration. The two telescopes were designed using an 11 mm. diameter exit pupil and a scanner space field of view of 60°. Only spherical lens surfaces were used.

A 200-600 mm zoom system is proposed, which has been designed for the 0.6-1.1 micrometers wavelength band for target acquisition and aiming. The zooming capability is of high interest to avoid loosing the target as it is moving ; but in order to maintain a precise aiming line, the mechanical movements of the lens groups must not create a boresight shift in excess of 100 microradians. The system is also equipped with a wide dynamic iris, allowing illumination to vary with a ratio of 10₄.

A low cost scanning optical system for an antitank seeker head operating in the 8 to 12 pm wavelength region is described. The overall system consists of a spherical dome and a very simple two aspherical elements telescope. The objective provides a high resolution over a 1° by 2° degrees FOV at a relatively large aperture and remains extremely lightweight and compact. This device is included on a two axis gyroscope to allow the line of sight deviation and the inertial stabilization of the gimbal objective. Considerations which direct material selection are presented. They concern : - transmission and absorption coefficients in the spectral domain - thermal and spectral variation of the refractive index - mechanical and thermal properties to sustain the meteorological and missile speed effect. The reflective IR imager is made of two aspheric primary and secondary mirrors which provide - thermal insensitivity - easy aligment procedure in the visible domain - low production cost by means of replication The innovative optical design program principles are presented and the sensitivity to high order asphericities analysed. Optical blur performances of this device are compared to those of a paraboloidal primary - plane secondary one. The manufacturing process for both the master and the replica are reviewed along with the assembly tests. The 55 mm wide - 50 g weight scanning mirror is shown and its main characteristics discussed. Maximum emphasis is put on the technological solutions choosen for its subminiature motorcoil and its position detector. Both "black and white" and "false coloured" images of an armoured vehicule are shown as recorded by the IR seeker during a real time tracking sequence. A ten micrometer hybrid focal plane detector has been developed jointly by SAT and LIR. This technology promises to offer advantages for advanced missile homing heads. Images ob-tained with this direct coupled Cadmium Mercury Telluride photovoltaic detector and CCD multiplexer can be seen