Automatic alignment is a computationally intensive task that slows some manufacturing processes which require it. In order to improve alignment speed we have designed a sensor that combines the imaging operation with the computation of the alignment error. The sensor is specially designed to work with the surveyor's mark in a two step operation. Coarse alignment is performed by operating as a four quadrant sensor and fine positioning is achieved using edge detection. Photodiodes are used for imaging and integrated amplifiers convert the photocurrents into an alignment error signal which is independent of illumination level. A version of this sensor has been fabricated and preliminary tests show that it functions as expected.

This paper describes both an analytical method and an autocollimation microscope for enhanced measuring the position of the center of curvature for a single optical surface in a lens assembly. Because the test is done in a non-destructive manner, it is possible to measure the position of the image of the center of curvature, but not the position of the center of curvature. The developed mathematical model determines the position of the centers of curvature in the lens assembly on the basis of measuring results. It takes into consideration the real position of all surfaces placed between test surface and microscope. When the measurement is complete for all surfaces, the optical axis of the assembly is determined as an average line through all of these measured centers. The decentration and tip of the optical elements can then be controlled with regard to this optical axis. Experiments verify the developed analytical model and software. The accuracy of the measurement is better than 10 microns.

Development of better, larger, and more dense focal plane arrays has stimulated the design of lens systems of larger aperture, longer focal lengths, and nearly diffraction limited performance over larger fields of view. As performance requirements, and the number of optical elements increase, the process of aligning lens elements during assembly rapidly becomes a critical issue. Tolerances on spacing, centration, and tilt become very challenging, as do the requirements on stability over extreme ambient excursions. This paper is intended to give a top level, general treatment of the subject, while describing certain techniques that incorporate precision alignment and measurement of critical lens parameters during assembly. These methods also allow monitoring during the accompanying bonding and curing cycles used throughout the build process. Examples serve to illustrate some of these techniques, and results are presented for specific cases. It is shown that centering, as evidenced by axial runout, can be set to the ten microinches level with a few microinches accuracy and maintained to about 50 microinches through the assembly and bonding process. With the use of computer data logging and analysis techniques, this can be extended to below five microinches.

A 20 power (20X) all-reflective microlithography objective has been fabricated for use in the soft x-ray region at a wavelength of 13 nm. The design uses the Schwarzschild configuration where two spherical mirrors form a point image from a point object. The centers of curvature of the two mirrors in such a system are coincident. However, to increase field of view, fifth- order spherical must be balanced by third-order spherical. This is accomplished by separating the two curvature centers longitudinally. Lateral separations quickly introduce coma into the wavefront. Holding the curvature center positions rigidly in place relative to the object and image positions is required for maintaining the wavefront quality. Suitable x-ray sources are not common; therefore, the rigidity must also be viewed as a level of ruggedness suitable for transportation from the assembly facility to the test facility. In this paper we share the techniques that we have used to satisfy the requirements of MTF, wavefront quality, ease of alignment, and ruggedness.

Designs for phased-array imaging telescopes covering a wide field of view (0.25 degrees) with satisfaction of all optical phasing conditions have been developed. Important concerns regarding the implementation of these telescopes include misalignment types and tolerances, and the complexity of active alignment systems needed to correct the misalignments. In this paper a phased-array telescope point design is briefly described. Possible misalignments in the array configuration are defined, and functional forms are given. A technique is introduced for including array misalignments in the wavefront aberration polynomial used to describe image quality in the final array focal plane. This polynomial is then used to show to what extent the misalignment-induced subtelescope aberrations defocus and coma may be corrected using only adjustments in the array configuration. Application to the point design shows that defocus and coma may be corrected, by the addition of piston and tilt terms, by factors equal to the ratio between the Seidel aberration coefficient and the corresponding Zernike polynomial coefficient. It is shown that correction possibilities result in looser subtelescope alignment tolerances and in the simplification of active alignment systems for individual subtelescopes.

This is an Optical Engineer's look at the engineering and testing concerns for a collimator requiring a 5 arcsecond bore-sight to its mechanical mounting datums. The collimator is of the off axis Newtonian type with a fairly fast f/number of 3.75. The primary aim of this paper is the special concerns of aligning a collimator very precisely to a mechanical datum, and doing it quickly. The optical engineering and test aspects of the alignment process are presented with an emphasis on information applicable to other designs. Metrology methods are all based on the LUPI or Laser Unequal Path Interferometer monitoring the Aspheric Mirror Alignment. The control of the bore-sight to the mechanical datums is reliant on a fixture system designed and built by the author to make volume production as painless as possible. The paper addresses how the fixture design can be used to help the alignment rather than specifics of this exact system design.

A non-contact approach for the alignment of optical elements and systems is described which allows the operator to clearly distinguish tilt and decentration errors of the individual optical elements. Advantages over alternate methods are discussed, and several examples of the use of the alignment tool are presented with experimental results.

The Strategic Defense Initiative Organization (SDIO) has supported a series of programs designed to automate the alignment, calibration, and control processes required for High Energy Laser (HEL) systems. HEL systems, whether based in space or on mobile platforms, must be aligned autonomously in order to meet mission requirements. Since these systems are generally large and lightweight, they are mechanically flexible. This flexibility results in launch or flight disturbances to alignment, as well as variations to alignment during long-term operation. This paper describes the alignment requirements for HEL systems, using the Space Based Chemical Laser (SBL) as the basis. Much of the discussion applies to any HEL system. The background of the alignment programs is described, and the goals of the group of programs are summarized. The alignment technologies described in this paper may find use in a wide variety of other applications requiring automated control of optical systems.

A broadly applicable methodology for automated alignment of complex optical systems has been developed. Investigations using numerical simulations have demonstrated that this methodology is robust and viable as an approach for aligning the resonator optics of the ALPHA high energy laser and the wide-field of view three-mirror beam expander developed under the Advanced Beam Control System (ABCS) program. The complexity of both systems makes them difficult and time consuming to align manually and also makes them prime candidates for automated alignment. The approach is based on the downhill simplex method in multidimensions, a general and very robust mathematical technique for finding the optimum of a function of many independent variables. Our confidence in the simplex algorithm's applicability to optical alignment has been enhanced by its success in these numerical simulations, in experiments, and also in additional simulations of other optical systems. We conclude that an autoalignment approach based on the simplex algorithm is an excellent choice for a wide range of optical systems of arbitrary complexity.

Experimental testing of the Downhill Simplex optimization algorithm is described in application to alignment of optical systems. A bench top experiment was designed to test the algorithm in the presence of real noise effects. A methodology of optimum design of merit functions for the algorithm is described and test data presented.

Techniques for autonomous alignment of beam control systems have been developed for a number of technologies including laser resonators, telescopes, diagnostic systems and components. These techniques have been designed to perform remote autonomous alignment on space based optical systems. An approach building upon these techniques has been developed for autonomous figure control of multi-actuator deformable mirrors. The algorithm used is a multidimensional minimization algorithm based upon the downhill simplex method of Nelder and Mead. This method is well-known for its extreme robustness and its broad applicability to a large variety of problems. In this paper results from numerical studies are presented that demonstrate the ability of this method in performing figure control of a deformable mirror. It is shown that the number of iterations required for the algorithm to converge to a practical solution grows linearly with the number of degrees of freedom. Near optimal numeric solutions are attainable, however, the required number of iterations increases non-linearly.

A novel neural net approach has successfully solved the time consuming practical problem of aligning the many optical elements used in the resonator of high power chemical lasers. Moreover, because the neural net can achieve optimal performance in only 2 - 4 steps, as compared with 50 for other techniques, the important ability to effect real time control is gained. This represents a significant experimental breakthrough because of the difficulty previously associated with this alignment process. Use of either near or far field image information produces excellent performance. The method is very robust in the presence of noise. For cases where the initial misalignment falls outside the regime encompassed by the training set, a hybrid approach utilizing an advanced conventional method can bring the optical system within the capture range of the neural net. This reported use of a neural net to rapidly convert imagery information into high precision control information is of broad applicability to optical, acoustic, or electromagnetic alignment, positioning, and control problems.

Soon after first light, it was discovered that the Hubble Space Telescope (HST) could not be brought to its optimum design focus. In a brief historical review, this paper tells how the spherical aberration was uncovered and quantified by independent review teams throughout the first year of observatory operation.

One of NASA's planned tasks during the first servicing mission to the Hubble Space Telescope in December 1993 is to correct the well known vision problem of the telescope due to an incorrect fabrication of the primary mirror. An exhaustive study of solutions to this problem resulted in a recommendation to place dime sized pairs of mirrors into the beam paths of five instrument channels to correct the spherical aberration attendant to the primary mirror. The name of the mechanism designed to carry these correction optics into the focal plane region of the telescope is COSTAR (correction optics for the space telescope axial replacement). For COSTAR to successfully deploy, four articulating arms carrying the correction optics into the crowded focal plane volume of the telescope must physically clear another opto-mechanical device sharing this space, the Wide-Field Planetary Camera (WF/PC). This paper describes the application of 3-dimensional computer graphics in a through-the-window virtual reality environment to simulate and visualize the planned deployment of COSTAR. Several computer generated animation sequences are shown that verify mechanical clearance of COSTAR's arms with respect to WF/PC.

Correcting the spherical aberration of the Hubble Space Telescope (HST) requires precise optical alignment and stability. To assure that the required alignment can be achieved and maintained on-orbit, the pickoff mirror and three of the four fold mirrors of the second generation Wide Field and Planetary Camera (WFPC-2) have been made actively controllable in tip and tilt. The Pickoff Mirror Mechanism (POMM) and the Articulating Fold Mirrors (AFMs) are commanded from the ground to their required positions once the WFPC-2 is installed in the HST. The POMM is a set-and-forget device that utilizes stepper motors, while the AFMs are maintained in position by the continuous application of control voltages to electrostrictive ceramic actuators. This paper describes the assembly level testing and calibration of the AFMs, and the development of a software tool that generates the commands for adjusting the positions of the POMM and AFMs to achieve system level optical alignment. Our experience with the POMM and AFMs through system level calibration and testing of the WFPC-2 instrument is described.

The second generation Wide-Field/Planetary Camera (WF/PC-II) for the Hubble Space Telescope (HST) was modeled to access the impact of manufacturing, alignment, and environmental tolerances on performance. This analysis showed that the lateral registration of the image of the optical telescope assembly (OTA) pupil to the surface providing the spherical correction must be aligned and maintained through launch to 50 microns; WF/PC-I was an order of magnitude less sensitive. Inherited WF/PC-I hardware was subjected to new WF/PC- II environmental tests. As a result WF/PC-II was reconfigured to ensure on-orbit performance: the focus mechanism was removed to increase stability through launch and on- orbit, and four formerly fixed mirrors were actuated, to provide capability for on-orbit pupil alignment. This paper traces the evolution of the WF/PC-II error budget from its WF/PC-I beginnings to the current configuration. This information should be of general interest to designers of future HST instruments.

End-to-end tests of the second generation Wide Field and Planetary Camera for the Hubble Space Telescope were performed with an optical stimulus that accurately simulates the optical configuration of the aberrated Hubble Space Telescope. This paper describes the optical design of the stimulus and the tooling used to control its alignment and validate is performance.

This paper describes the use of a Hartmann-type pupil mask and CCD camera to perform wavefront and focus validation of the Hubble Space Telescope simulator before and during environmental testing of the second-generation Wide Field and Planetary Camera. The method yields a focus accuracy at F/24 of about +/- 100 micrometers even in the presence of 3.5 waves of surface spherical aberration. The method avoids the introduction of potentially imperfect auxiliary optical tooling (e.g., null correctors).

To facilitate the accurate placement and alignment of the corrective optics space telescope axial replacement (COSTAR) structure, mechanisms, and optics, the COSTAR Alignment System (CAS) has been designed and assembled. It consists of a 20-foot optical bench, support structures for holding and aligning the COSTAR instrument at various stages of assembly, a focal plane target fixture (FPTF) providing an accurate reference to the as-built Hubble Space Telescope (HST) focal plane, two alignment translation stages with interchangeable alignment telescopes and alignment lasers, and a Zygo Mark IV interferometer with a reference sphere custom designed to allow accurate double-pass operation of the COSTAR correction optics. The system is used to align the fixed optical bench (FOB), the track, the deployable optical bench (DOB), the mechanisms, and the optics to ensure that the correction mirrors are all located in the required positions and orientations on-orbit after deployment. In this paper, the layout of the CAS is presented and the various alignment operations are listed along with the relevant alignment requirements. In addition, calibration of the necessary support structure elements and alignment aids is described, including the two-axis translation stages, the latch positions, the FPTF, and the COSTAR-mounted alignment cubes.

The corrective optics space telescope axial replacement (COSTAR) configuration contains mechanisms in each science instrument channel that allow for on-orbit correction for image plane focus and for lateral and axial mapping of the Hubble Space Telescope (HST) primary mirror onto the aspheric corrector mirrors. The optical alignment of the COSTAR optics is accomplished in two phases. In Phase I, the mirror bezel tilts and lateral positions are determined through the use of surrogate flat mirrors with the mechanism's positions held at the mid-range of their travel. The Phase I alignment is followed by Phase II interferometric optimization of all five optical channels. At the conclusion of the Phase I alignment, the optics are positioned accurately enough to allow simultaneous correction of most channels on orbit through the use of the mechanism compensation and telescope fine-pointing control. Individual mirror positions and orientations are determined through the use of alignment telescopes, theodolites, alignment lasers, and reference fiducials incorporated into the COSTAR Alignment System (CAS).

The purposes of the Phase II alignment are to coalign mirror pairs for the FOC and FOS channels and to set the compensation mechanisms of each channel to the optimum positions to allow the overall system performance to be determined and verified through use of the RAS/HOMS equipment without requiring adjustment of the mechanism compensators during testing. This alignment process is performed with the COSTAR instrument installed in the COSTAR Alignment System (CAS), using a well characterized interferometer system as the optical source. The interferometer uses a custom designed reference sphere with built-in spherical aberration to enable the highly aberrated two-mirror systems to be observed in double-pass.

The Refractive Aberration Simulator (RAS) produces an image field which matches the as- built Hubble Space Telescope (HST) image field to a very high degree at 632.8 nm. The instrument is used to provide HST-like aberrated image plane illumination to the COSTAR and FOC-STM instruments during the final performance verification operations. It is capable of providing up to 15 simultaneous inputs to the instruments or metrology equipment located at its image plane. It consists of 6 anti-reflection coated glass lenses, each requiring lateral positioning accuracy as fine as +/- 10 microns, axial positioning accuracy of +/- 140 microns, and angular positioning of +/- 12.7 arcseconds, a fiber coupled laser source system, removable aperture stops for providing obscured or unobscured wavefronts, and a three-axis positionable source plate with removable fiber couplers for measurement or alignment operations.

The results of efforts concerning the creation of a testing system of the wavefront deformations in the telescope optical highway based on the use of the holographic optical element (HOE) are presented. The paraxial analysis of the HOE recording and the work of the testing canal are given. The solution of the non-determinacy problem of the testing results is described. Mathematical modeling of the holographic testing scheme and its correlation with telescope performance are made in frames of geometrical and wave optics.