In October, 1985, new encoder readout electronics were installed on the MMT to read the existing absolute encoders. With the new electronics, the angular resolution of the encoder system has been increased from 24 bits (0.077 arcseconds) to 26 bits (0.019 arcseconds) and the amplitudes of the raw uncorrected errors related to the period of the InductosynsTM* have been reduced from a few arcseconds to the order of +/- 0.5 arcseconds or better.

The Multiple Mirror Telescope exhibits good tracking performance and excellent pointing accuracy. Considerable engineering effort has been expended over the past four years to investigate methods for improving accuracy, and to identify and understand the sources of the remaining errors. We can achieve tracking errors as low as 0.5 arcseconds peak-to-peak without autoguiding, and RMS pointing errors better than 1 arcsecond, but changes of approach may be needed to reach our tracking performance goals.

So named because it resides on top of the focal plane instrument, the MMT top box is a two-level optical bench with many remotely controlled features which can be configured for experimental use of the MMT. It houses the relay optics and television cameras for focusing, coaligning and guiding the six telescopes of the MMT, as well as the integrating sphere, filter wheel, Hartmann mask wheel, and relay optics of the comparison source for instrumental calibration. It also provides a rigid mounting platform for research instruments as well as special fixtures for telescope tests and collimation. In accommodating these tasks the top box incorporates several novel features which may interest designers of other large telescopes.

The Multiple Mirror Telescope uses six 1.8 m diameter folded Cassegrain telescopes to achieve the collecting area equivalent to that of a single 4.5 m telescope. However, for some applications, this capability is realized only if the images are brought to a common focus and coaligned in such a way that the overlapping images appear to result from a single primary. The method of adjustment used at the MMT to coalign the six images does not involve the primaries or the tertiaries; instead, the cassegrain secondary mirrors are tilted in orthogonal axes. Since the SPIE Conference on Advanced Technology Optical Telescopes in March 1982, the MMT has rebuilt the optical configuration used to sense image positions at the focal plane (the top box) so that image stacking, automatic guiding, and star acquisition can be made fully operational procedures. This report describes the coalignment system and the results of tests to date.

A 10 micron imaging camera has been used to identify subtle and not-so-subtle thermal pollution of the near telescope environment at several large telescopes, and can also be used to monitor progress in eliminating such pollution. This paper reports some highlights of data taken at the Multiple Mirror Telescope (MMT), the Kitt Peak National Observatory (KPNO), and Cerro Tololo Inter-American Observatory (CTIO) 4 meter telescopes, and Canada-France-Hawaii (CFHT) 3.6 meter telescope on Mauna Kea, the McDonald Observatory 2.7 and 2.1 telescopes and the 5 meter telescope at Mt. Palomar. This portable television equipment was purchased by the MMT with funding from the National Science Foundation and can be made available with some trained technical assistance.

All six MMT telescopes can now be optically cophased simultaneously over a wide field of view yielding coherent coverage of the complete Fourier transform plane corresponding to that of a pupilmasked telescope of 6.86m aperture. Open-loop phasing control compensates for flexure-induced path-length changes due to variable gravitational loading as a function of elevation. The system has been used to produce diffraction limited images and differential images of Alpha Orionis using narrow-band (1.2A) filters centered both on Hydrogen-alpha and on a similar bandpass out of the absorption line. Corresponding wide (100A) and narrow-band images of Gamma and Epsilon Orionis show the expected result for unresolved sources at the diffraction limited resolution of the fully-phased MMT.

The polarizations of multiple mirror telescopes are calculated using Mueller calculus. It is found that the Multiple Mirror Telescope (MMT) produces a constant depolarization that is a function of wavelength and independent of sky position. The efficiency and crosstalk are modelled and experimentally verified. The two and four mirror new generation telescopes are found to produce sinusoidal depolarization for which an accurate interpretation of the incident Stokes vector requires inverse matrix calculations. Finally, the depolarization of f/1 paraboloids is calculated and found to be less than 0.1% at 3000 A.

We describe the performance of the MMT as an astronomical telescope and the performance of its instrumentation. We present statistics relating to the number of publications resulting from data collected with the MMT, telescope and instrument down time, and site related statistics. We also briefly discuss plans for new and continuing development projects.

A complete preliminary optical design and tolerance analysis for the 15-Meter National New Technology Telescope (NNTT) is presented. Some designs of instruments being planned will also be discussed. The optical designs have been optimized to the performance goals that will take special advantage of the multiple mirror telescope concept, but within the practical limitation of the foreseeable technology trends.

The NNTT is being designed as an array of four telescopes, each 7.5 meters in primary diameter, on a common mounting, phased in the infrared, and able to produce 0.25 arcsecond images. Rapid changeover from combined beam to individual telescope operation is required. To accomplish this, a mechanical configuration employing exchangeable top-end optics modules has been designed. Early results of structural analysis are presented. Structural provisions for driving the telescope are discussed. To meet imaging goals, we conclude that careful structural design plus active position control of the optics is necessary and feasible.

Materials for large primary mirrors should have low density, high elastic modulus, low coefficient of thermal expansion, dimensional stability over time, favorable polishing characteristics and affordable cost. The mirror's structural shape can contribute to high stiffness, low weight and a short thermal time constant. These characteristics can be provided by cast honeycomb borosilicate glass mirrors, but the coefficient of thermal expansion is high enough to require thermal control to achieve images at the 1/4-arcsecond level. This paper describes a series of tests performed on a 1.8-meter cast honeycomb mirror with an integral ventilation system for temperature control. Surface distortions have been related to the pattern of temperatures present, and the effect of ventilation on the thermal time constant has been explored. The results are extrapolated to larger size mirrors using finite-element analysis.

The NNTT will be a multiple mirror telescope consisting of an array of four 7.5-meter telescopes to be coaligned and cophased by means of an internal optical metering device. As a result, the NNTT acts like a single telescope with a 15-meter collecting diameter and a 21-meter resolution diameter. This paper describes the coalignment/cophasing system.

Here we review work done in the past pertaining to enclosures and their effects on the telescopes they are designed to protect. From this experience we adopt what are considered benefits and try to avoid what are considered disadvantages in the protective enclosure designs, and apply the results to the National New Technology Telescope enclosure. We also present the results of one contracted study, done by Merz and McLellan of England. This study includes a preliminary design and a realistic cost estimate.

The characterization of atmospheric turbulence at and above the two candidate sites for the NNTT 15-meter telescope has resulted in the development of specialized techniques. The equipment used to measure astronomical seeing, microthermals, water vapor and temperature is discussed along with sample data and calibration results. By using instruments whose altitude coverages overlap, it has been possible to "bookkeep" quantitatively all of the sources of image degradation, especially near the ground.

As part of the NOAO National New Technology Telescope Site Evaluation Project, on-site measurements of thermal infrared sky irradiance, atmospheric turbulence, optical image quality, and meteorological conditions were begun in 1984 on Mauna Kea, Hawaii, and Mt. Graham, Arizona. These synoptic measurements will be used to assess the comparative astronomical performance levels of these sites and will provide the basis for a site recommendation matched to the specific requirements of the 15-meter NNTT. This overview describes the observation and analysis procedures employed for this site intercomparison.

Very large telescope projects, due to their increased expected performances and cost, have prompted several campaigns to compare existing and potential observatory sites throughout the world.ln the frame of the ESO-VLT studies, an extensive geographical survey of Northern Chile led to the selection of a few representative summits for comparison with the observatory of La Silla. Standard meteorological towers, acoustic radar, Rawinsondes, microthermal sensors and optical seeing monitors will be collecting experimental data in the years to come. After a brief review of the relations between atmospheric and optical seeing parameters, the particular case of Northern Chile is described with the help of data collected up to naw..Then follows a technical description of the instrumentation developed by ESO including the differential Hartmann open air monitor.

In order to be efficient, new astronomical methods, like high angular resolution and speckle interferometry require subarc second seeing conditions and long speckle boiling time. It is now well known that there is a close relationship between the seeing and the integrated vertical profile of atmospheric turbulence, from the focus plane of the telescope up to the stratosphere. The measurement of the spread function of a star image does not tell anything about the respective contributions of the inside of the dome, the boundary layer and the free atmosphere. Disturbances coming from the first two slabs can be avoided by : removing the heat sources, cooling the floor, building the observatory above the inversion layer, simulating the site in a wind tunnel... But there is no a priori knowledge about the free atmosphere turbulence. In the literature, such profiles are too scarce and sparse to foresee a clima-tology. In this paper, it will be shown that there is a strong experimental correlation between a good seeing and a low wind speed at the tropopause level. The study rely on already published sesonal variations of the seeing in La Silla, Chile, and Hawaii, and on atlas of climatology of the atmospheric circulation at 200 millibars level. It seems that subarc second seeing conditions require tropopause wind speeds lower than 20ms-1. A quick look at the wind behavior let us hope very good image quality above sites located in the northern part of Chile, near Peru. A fortunate consequence of our assumption is that slow tropopause wind selected sites will likely have a long speckle time life.

The turbulent surface boundary layer has been monitored within the summit area of Mauna Kea on the island of Hawaii. Measurements were made using three meteorological towers instrumented with microthermal sensors at three levels, ambient temperature sensors at two levels, and anemometers at two levels. Spectra of the temperature fluctuations were also measured. The spatial and temporal variations in microthermal activity associated with ground turbulence are quantified and explained. From these findings, the best locations and ground heights for telescope facilities are determined. The relative importance of turbulence near the ground in the degradation of image quality is deduced.

The average seeing at the University of Hawaii 2.2 m telescope from 10 August 1985 to 9 January 1986 was 0.95 ± 0.02 arcsec full width at half-maximum at a wavelength of 0.70 pm. We believe that the dome makes substantial contributions to this average because the see-ing degrades with increasing temperature difference between a point near the primary mirror and the outside air. There is no significant dependence on wavelength (over the range 0.63-0.82 μm), wind direction, or time (over the interval covered to date). The data reported in this paper are the baseline against which we will assess the results of a long-term program of dome thermal balance improvements.

The scientific need for building telescopes with much higher spatial resolution than can be achieved with existing or planned instruments is well known. Michelson-type interferometers with baselines from a few tens of meters up to several kilometers have been proposed. While the larger baselines will undoubtedly be constructed in space, consideration of ground-based sites for baselines up to a few hundred meters will be important because of the relative costs. To measure fringes in the absence of optical path variations due to atmospheric turbulence, the interferometer baseline, B, must be stable.

We present the fabrication methods, testing and evaluation techniques, and performance results for the mirrors for the Extreme Ultraviolet Explorer (EUVE). Our finest mirror produced to date has a measured half energy width of 8"at optical wavelengths. With a polished nickel surface, the telescope throughput was 35% at 44A and 60% at 256A. The surface roughness is 20 A rms.

Grazing incidence telescopes are required for extreme ultraviolet astronomy (100 < < 1000 A) when observing at wavelengths shortward of 500A, and provide the highest throughput over the entire bandpass. Grazing incidence telescopes of conventional design have a substantial soft x-ray response as well as an extreme ultraviolet response. However, the available bandpass filters designed to transmit radiation longward of 400Å also transmit soft x-rays, which compromises the quality of the extreme ultraviolet data. We describe a grazing incidence telescope which is designed to suppress the soft x-ray throughput. This telescope incorporates a Wolter Schwarzschild Type II mirror with large graze angles. It retains all the desirable features of an extreme ultraviolet photometric survey telescope (high throughput, wide field of view, compactness) and, in addition, has no soft x-ray response. A telescope of this design will be flown on the Extreme Ultraviolet Explorer mission to make a survey of the sky at extreme ultraviolet wavelengths longer than 400Å.

We describe an Astrometric Telescope Facility (ATF) which is intended to be operated on the Space Station in the 1990's. The main purpose of the telescope will be to do long-term, highly accurate relative astrometry of nearby stars to detect gravitational perturbations by companions with masses as small as Neptune. This will require astrometry accurate to 10 microarcseconds, at least 100 times better than that currently being done at ground-based observatories, and the operation of a facility in space for many years. Because the NASA Space Station is expected to be a permanent and continuously occupied base in low earth orbit, the system described here involves an articulated telescope mount at one of the upper points of attachment on the Space Station structure. The astrometric technique developed by Gatewood et al. (1980) will be used, in which the relative positions of star images in the focal plane of the telescope are indicated by the relative phases of the modulations of star brightnesses introduced by translating a Ronchi ruling across the focal plane at uniform speed. This technique has been proven to be capable of precise astrometry from Allegheny Observatory and when freed from the effects of the earth's atmosphere should provide the required accuracy. Technical issues under consideration include the damping of vibrations on the Space Station, required accuracy of fine guiding, optical configuration, metric accuracy of the Ronchi ruling, and choice of detectors.

Site and configuration options for a successor to the Hubble Space Telescope are discussed and one candidate is presented. The telescope is a traditional Cassegrain with a 10-meter diameter monolithic primary, adaptive secondary, and passively cooled optics. Wavelength coverage is from the far ultraviolet to the near-infrared. The observatory is located in the geosynchronous orbit to minimize environmental constraints and increase observing efficiency.

A three-mirror telescope derived from the Paul corrector is described. It differs from the original Paul design in several respects. (1) The third mirror is located behind the primary mirror instead of in front of it. (2) The telescope is made off-axis so that there is no central obstruction, thus avoiding the extension and asymmetry of the diffraction pattern caused by the spiders holding an on-axis secondary mirror. (3) Baffling is not a problem as it is with the usual Paul design. The focal surface is flat where a moving ronchi grating is located. This is the first element in the astrometric analyzer. A real image of the pupil is produced behind the focus. This is helpful in the design of relay optics (not described) which reimage the grating onto a CCD.

The W.M. Keck Observatory and Ten Meter Telescope is presently being constructed on the summit of Mauna Kea, Hawaii. We briefly describe the status of the project.

The Keck Observatory Ten Meter Telescope has a Ritchey-Chretien optical design which maximizes the field of view for an f/15 focus at about 20 arcminutes. An f/25 infrared focus provides a 9 arcminute field. The optical design and the optical tolerances will provide 0.5 arcsecond images (80% enclosed energy diameter) over these fields of view. This is equivalent to 0.25 arcsecond FWHM seeing conditions. Instrumentation is designed to allow for rapid, semi-automatic instrument exchanges. At major foci, the instruments will be contained in interchangeable modules having a common weight and center of gravity that attach kinematically to the telescope.

This paper describes of the basic concepts and current status of the ESO NTT Project. The NTT is an optical telescope with a primary mirror of 3.5 meters and two Nasmyth foci. It will operate at ESO's La Silla Observatory, and is planned for installation in 1988. The project combines a number of innovative technical solutions.

The ESO VLT base line concept is a fixed array of four 8 m telescopes; other concepts based on the same optical configuration for the unit telescopes are however still considered. Among other possibilities, the two main options for the primary mirror blank are either low expansion glass ceramic or highly structured steel. Limited tests have been done with both solutions. The unit telescopes have a classical Alt-Azimuth architecture but are designed so as to operate in the open air without a surrounding building. Their proper functioning relies heavily on the active control of the optical elements. The building is intended to provide only a day time protection. Several types of low cost enclosures have been investigated. A partial protection against the prevailing strong winds is also under consideration.

A new concept for a Very Large Telescope is presented. A filled aperture of about 12 m diameter is realized by a central monolith surrounded by a ring of sectors. It represents a pure two-mirror Ritchey-Chretien optics. The optics is supported by a yoke mount of extreme rigidity. It allows easy access to a very large focal area.

The Japanese National Large Telescope (JNLT) proposed is an optical infrared telescope with an aperture of 7.5m, promoted by the Japanese astronomers community. The systematic conceptual study of the JNLT was started in 1984 at the Tokyo Astronomical Observatory (TAO) by its JNLT working group, of which the present authors are members. The study included optical designs, comparisons among different types of main mirror, mechanical structures, driving and guiding systems, control systems, and dome structure among others. We select some particular aspects in this report which may deserve to be discussed at the SPIE meeting. Although the details of individual topics will be published by specific contributors separately, those who are interested in each topic should contact the JNLT working group at TAO.

10 meter telescope, 250 Angstrom optics ( 1/20 wave ), ultra thin 5cm thick mirror, or 32 multiple mirrors, compact alta-azimuth mount. Includes new configuration and novel concept for optics supports. which extends the capabilities of the instrument and markedly decreases the cost of fabrication and installation. The completely engineered details include on-site production and finishing of the mirror blank, using the telescope drives and bearings. Design of the foundation and the observatory housing is incorporated in the general arrangement, which also considers "enviornmental" operation of telescope with storage housing. The mirror system and mount structure is particularlly designed to weather the effects of wind and temperature. Extensive use is made of compensating design features, enhanced by computer control. Available foci include prime, Cassegrain, Nasmyth, and coude', with a maximum of four mirrors being required.

The Mark III stellar interferometer is a phase coherent long baseline telescope Michelson stellar interferometer that has been under construction since 1984. The instrument is designed primarily for accurate wide angle astrometry at the milliarcsecond level. It will use two color measurements to reduce the error due to atmospheric turbulence and a 20-meter baseline, 2.5 milliarcsecond resolution, that is laser monitored to achieve high precision. Initial operation of the Mark III is scheduled for late 1986. The Mark III interferometer uses a number of active subsystems to improve the ease of operation of the stellar interferometer. The siderostat pointing system uses laser interferometers for open loop control and an image correlation star tracker for closed loop control. The laser system, in addition to pointing the instrument, will be used to measure the errors in the gears as well as thermal drift. An optical delay line is used to equalize the path lengths in the interferometer. The delay line is programable in 50Å steps with 100A accuracy over a peak to peak range of 20 meters (1 part in 109) with maximum slew rate of 60 cm/sec. A Kalman filter phase estimator will be developed for the stellar fringe tracker. In addition to astrometry, we are planning to build additions to the instrument for optical aperture synthesis. Operation in the phased imaging mode for bright objects as well as the amplitude mode for photon starved objects are eventually planned with a multispectral, multi-r° aperture fringe detector and high speed, real time signal processor.

Often it is desirable to have a large field of view when using telescope arrays for interferometric imaging. I examine for different types of arrays what the limitation of the field of view is and what can he done to optimize it.

The array concept for ESO's Very Large Telescope (VLT) project opens the possibility of observations with very high angular resolution. The long baseline provides a resolution span from approx. 0.5 marcsec in the blue to approx. 30 marcsec at 20 um wavelength. Realistic target for the interferometric operation is the IR range from 5 to 20 pm and later a gradual expansion to shorter wavelengths. In the interferometric mode the gain of the 8 meter single apertures of the VLT is only given if adaptive optics is applied for a real-time partial or full phase compensation of the degradations due to atmospheric turbulence,

Intensity interferometry measures the square of the absolute value of the normalized coherence function. The phase of the coherence function is lost. The possibility of deriving the phase function by using higher-order intensity correlations from the signal collected by a large array of mirrors is discussed. The array is not phased and each mirror is considered to be of the intensity interferometer type, i.e., it only concentrates light onto a detector without regard to the imaging quality. Knowledge of the coherence function yields the two-dimensional intensity distribution of self-luminous or intrinsically luminous objects.

The goal of this study is to investigate the heterodyne interferometer as the basis of an optical receiving array capable of synthesizing source images to extremely high angular resolution, either from a terrestrial site or from space. It is hoped that an angular resolution of one milliarcsecond can be achieved in the presence of the terrestrial atmosphere and that images can be produced of quality comparable with those now obtained with modern synthesis radio telescopes, that is, dynamic range approximately 40 db and field of view of one arcsecond. Such a system, if its sensitivity were sufficient, would find manifold applications in astronomy, in space surveillance, and in remote sensing of the earth and planets from space platforms. Despite the negative evaluations of heterodyne interferometry as a useful astronomical technique at optical wavelengths, based on the perception that insufficient bandwidths would be obtainable, it appears useful to reexamine the potential of this method in the light of recent technological developments. Optical heterodyne methods have been developed to a very workable state of perfection for application to remote sensing problems in atmospheric research and for applications in fiber optical communications. An attractive feature of the heterodyne technique is that, once the optical or infrared signals have been converted to "intermediate" frequencies, all the successful data-handling methods of radio interferometry can be applied. It is believed that the bandwidth limitation can be overcome by modern methods of electronic integration and data processing, and that the heterodyne method may well provide solutions to the problems of atmospheric inhomogeneity and mechanical precision that have long plagued high-resolution optical astronomy.

Two high-precision and movable telescopes of 65 inch aperture are being constructed for long baseline infrared spatial interferometry (1). The telescopes are of novel design and emphasize the control of optical pathlengths and of telescope positions to a precision of about one micron. This will allow the relative phase of interference fringes to be determined over rather large angles of the sky and over some period of time, so that high precision astrometry as well as aperture synthesis can be done at wavelengths near 10 microns.

Active optics, as envisaged for the ESO NTT, is based on four principles: 1. The linear relationship between the applied support force and the deformation of the mirror resting on a force based support system. 2. The orthogonality of the corrected aberrations. 3. The restriction to correct only those aberrations for which small forces are necessary, i.e. long wavelength aberrations as e.g. astigmatism and spherical aberration, which are also those modes induced by errors in the system. 4. The use of precalibrated forces for these corrections. Using the Shack-Hartmann test method for the analysis of the wavefront, we have confirmed these principles in an experiment with a lm diameter spherical mirror with a thickness of 18,9 mm and 78 supports.

We describe the general principles behind a polychromatic adaptive optics prograff astronomy which was started at NOAO recently. In this program the atmospheric wavef distortions are measured at visible wavelengths (700nm) using an astronomical object vicinity of the infrared object of interest. The resulting wavefront corrections ar applied to an infrared imaging system which utilizes a two-dimensional detector arra

However pertect an adaptive optical system can he, it win never tuiiy correct tne image. We discuss here the effects of amplitude errors due to stellar scintillation, the effects of chromatic errors due to both refraction and diffraction in a two-wavelength system, and the effects of non-isoplanicity. All these errors are directly related to the height of turbulence layers. An expression is derived for the scale height of turbulence which should be considered as an important parameter in selecting new sites for astronomica observations. The performances of adaptive optical systems are usually described in terms of Strehl ratios. We present here the results of computations of the whole transfer function for lon exposure compensated images. It is shown that the point-spread function for a partially compensated image generally consists of an Airy disk surrounded with a halo. The ratio of the energy in the Airy disk over the energy in the halo is independent of the telescope aperture and provides a better measure of the quality of the compensation.

In the NOAO IR Adaptive Optics Program, we have elected to develop a system to use the visual light from an object to sense the wavefront errors and generate the corrections re-quired to give diffraction-limited imaging in the near-IR (2-10 microns). We have evaluated the performance of a variety of possible sensor systems, computing both their ideal performance and their expected performance with available detectors. A major consideration in this evaluation is the ability of the sensor to measure mean wavefront tilts of the visual wavefront over subapertures corresponding to the (larger) IR wavefront scale lengths. We have chosen to use a Hartmann-Shack sensor with red-sensitive image intensifiers and a Reticon detector.

The infrared emission of astronomical sources is buried in a large background due to the thermal emission of the telescope. The real time deformations of an adaptive flexible mirror are likely to produce a spurious modulation of this background, adding noise to the signal. We estimate here the amount of noise introduced by such a mirror.

An optical Very Large Array is proposed for short-term ground-based observing. With 15 apertures arranged non-redundantly along a circle, phase closure techniques can reconstruct high-resolution images in the snapshot mode. In one of the two design options considered, delay lines are avoided by displacing the telescopes during observation. Telescope designs involve replicated optics and spherical mounts. We also mention steps in the development of a floating array in space: a small telescope with electrostatic mount is to be tested aboard a Saliout spaceship in 1986 to qualify the TRIO technology.

Since 1903 when good instrumented data first became available over 100 great earthquakes of Richter Magnitude 8.0 or greater have occurred in the world. Many of the high seismicity zones in the world are in locations of large astronomical observatories. These include Chile, California, Hawaii and Japan. When building new observatories it is necessary to provide seismic protection in case an earthquake does occur. It is also important to reevaluate existing facilities to determine if they are adequately protected according to the latest engineering practice. This paper relates the locations of the world's major observatories, particularly those in the United States, to the earthquake hazards in these areas. Instances of damage to instruments during earthquakes will be numerated. Criteria for designing facilities to resist earthquake loads will be summarized, and certain protection techniques will be presented.

The current design of large telescopes calls on innovative technology to achieve high performance while keeping cost at an acceptable level.

Innovative designs of enclosures are being studied for the generation of large telescopes which are presently being developed, essentially in order to keep costs from increasing unacceptably with the size of the telescopes. These studies and the generally positive experience with MMT-type buildings, largely open to the wind during observation times, are confirming the trend toward a radical change of phylosophy in the concept for telescope enclosures, which recognises that in many cases the open air environment is more favourable to optimum telescope operation than the nominally stable and controlled environment inside a classical dome. This approach has been fully taken in the preliminary design of the ESO Very Large Telescope array which sees the telescopes operated essentially in theAfree atmosphere, being only protected against strong winds by a semi-permeable wind screen'. The aim of the preliminary studies was to achieve a comprehensive view of the different aspects of the open air environment and their influence on the design of the telescope and its performance. The paper describes some of these studies where we have tried, sometimes at the price of simplifying approximations that sacrifize quantitative accuracy, to identify the real phenomena and the parameters that most influence them, evidencing the orders of magnitude of the physical quantities involved.

The ESO Very Large Telescope consists of an array of four telescopes operated in the free atmosphere. The telescopes are protected against strong winds by a semi-permeable windshield oriented normal to the prevailing winds. The paper describes the studies performed for the concept of the wind-shield, and in particular mathematical simulations with a two-dimensional finite-element model for fluid flow were undertaken. This model incorporates the complete Navier-Stokes equations along with thermal stratification of the atmosphere; a turbulence parametric scheme has been adapted to close the equations. The model enables detailed studies to be carried out on the influence of wind-shield efficiency for telescope operation when a certain number of parameters are modified (for example porosity of wind-shield, height of platform above the ground, etc.). It appears that telescope operation is only partially sensitive to changes in structural parameters; the study, however, has underlined several problems which were unexpected. In particular, strong vertical velocities directed upwards at the upper part of the telescope, as well as rapid flow beneath the platform itself, may lead to certain constraints on maximum telescope operation. Two design concepts for the windshield are described, which are presently envisaged : a fixed structure with louvers, and a completely removable structure. The structure will be some 24 m high, stretching the width of the platform (approximately 160 m).

Five universities are constructing a 3.5 m telescope facility on Apache Point (2800 m above sea level) in southern New Mexico. The telescope enclosure design, in addition to providing protection for the telescope includes thermal control features intended to minimize local seeing effects. The components of the thermal design of the telescope enclosure include minimization of heat and cold production, removal of observer support functions to a detached operations building, control of air flow through the telescope chamber, minimization of thermal mass which is thermally coupled to exposed surfaces, forced ventilation of instruments, telescope and enclosure base, emissivity control of surfaces coupled to the sky, and use of resistive heaters for condensation control in the telescope chamber.

We have calculated the deflection of the surface of a 3.5 meter diameter borosilicate mirror using the finite element method. The mirror is a 0.46 m thick honeycomb structure with 25 mm thick face plates and 13 mm ribs. The cell spacing is 0.192 m and is regular except near the inner and outer perimeters. Axial support for the mirror will be provided by 48 air pistons. The air pressure in the pistons will be regulated to balance the axial component of the mirror weight. The axial air pistons are located near nodes in the rib pattern of the honeycomb thereby facilitating access to the interior of the mirror for ventilation and transverse support. A simple wind loading case was examined and the resulting deflections are shown to be small. The effects of thickness variations in fabrication of the mirror are presented and are small for the magnitudes of the variations expected. Several thermal load cases are described.

We describe a system of pneumatic piston mirror supports for use in an altitude over azimuth telescope which react to gravity and wind loading. A pressure controller provides dynamic compensation of variable wind loading and changes in the gravity loading as a result of altitude angle changes. An active air circulation system which ventilates every honeycomb cell can be implemented without interference from the mirror supports. The system can be expanded in principle to accomodate 8 m honeycomb mirrors.

The Apache Point Observatory 3.5 meter altitude-azimuth telescope features lightweight honeycomb optics, a fast f/1.75 primary figure, multiple on-line instrument capability and a control system designed for efficient remote operation. Friction coupled rollers drive the axes and couple the encoders. We have concentrated particularly on reducing local seeing effects by controlling heat in the vicinity of the telescope. Measures include mass reduction, emissivity control, insulation and forced ventilation.

Six telescope truss geometries are evaluated for both wind and gravity deflection stiffness. The existence of simultaneous solutions for gravity and wind deflections is demonstrated. A baseline design for the Texas 7.6-meter telescope with an F/1.4 Angel lightweight borosilicate primary mirror is shown.

In the ESO Very Large Telescope base-line concept four mechanically identical 8-meter telescopes are combined to form a fixed array. Several structural designs for the 8-meter telescope, based on the classical alt-azimuth mounting concept and considering the environmental conditions for a domeless telescope, have been worked out. In a design optimization process weight and cost of the telescope structure have been minimized under static and dynamic constraints, e.g. a first eigenfrequency of the telescope structure greater than 7 Hz. Whereever possible an open truss design was preferred to a closed box structure. The different design concepts are described and the results of a static and dynamic structural analysis, i.e. the deflections due to gravity and 'wind loading as well as the fundamental eigenfrequencies and mode shapes, are presented. Finally an attempt was made to predict the dynamic performance limits of large telescope structures.

A novel method of telescope pointing has been developed based on He-Ne laser interferometric measurements of mirror position. We have applied this technique to a Pfund-type telescope, which consists of a fixed parabolic mirror illuminated by a flat mirror which rotates in altitude and azimuth. If the parabolic mirror is stationary the angle of pointing depends only on the change of orientation of the flat mirror with respect to the parabolic mirror. The relative angles of mirrors with radius of 1 meter are measured to a precision of ~.05 arcsec. Fluctuations in the index of refraction of the atmosphere between the mirrors are the primary source of this limit; however, changes in pointing position involving only a gradient in the index of refraction perpendicular to the optic axis are largely compensated by this pointing technique. Conversion of interference fringe counts to precise angle of pointing involves solutions of equations of modest complexity, but is easily handled by a small computer.

The telescope control system software for the Isaac Newton and Jacobus Kapteyn telescopes on the Roque de los Muchachos, La Palma has now been commissioned. This paper summarises the software developments for these systems and describes aspects of the the design philosophy which is being adopted for controlling the 4.2m William Herschel telescope.

This report presents a design for an inexpensive and highly reliable dome position encoder which has few moving parts and which eliminates the need for a Hechanical engagement to the dome. This hybrid incremental/absolute optical encoder is now in use on the dome of the 1-meter Nickel Telescope at Lick Observatory, and will soon be installed on the dome of the Observatory's Shane 3-meter Telescope. This report discusses the costs and important points of the construction, installation, operation, and maintenance of the encoder. It also explores the feasibility of using this encoder on the domes of large telescopes such as the Keck Observatory 10-meter.

The optics within the Hubble Space Telescope consist essentially of a primary and secondary mirror. Measurements from three wavefront sensors are telemetered to the ground where a series of computations are made to determine whether the telescope is performing within specifications. If not, optimum secondary mirror realignment parameters are computed based on the aberrations present in the wavefront. The mirror support structure is held at three places by six links. Position adjustment is made by independently extending or withdrawing each of the links. These six actuators are small angular motion stepping motor drives with eccentric output shafts. Rotation of the shaft produces a displacement of the link as a function of the actuator shaft angle. The angle is determined from the output of a potentiometer coupled to the actuator output shaft or by counting the number of steps taken since launch. This paper presents a mathematical analysis of the secondary mirror control structure. It describes a geometrical model which permits a calculation of all six actuator positions given the position of the secondary mirror, and the calculation of the secondary mirror position, given the angular position of the six secondary mirror actuators. A series of tests for validating the model is described. In these tests mirror positions predicted by the model were compared with actual positions measured by test instrumentation. These tests demonstrate conclusively that the mirror can be accurately positioned within require-ments in a few iterations.

The Digital Piezo Transducer (DPT) is a piezo electric actuator incorporating a capacitance micrometer as an integrated position sensor. This internal micrometer has a position-al resolution of much better than a nanometre and 10 hysteresis. The DPT is operated in a closed servo loop and is capable of sub-nanometre resolution, repeatability and stability. Its response time is 0.2 msecs and it has a maximum range of 50μm.

The usual procedure in finite element structural analysis is to produce a printed or graphical report of strain and/or stress for an object. In the situation where one must derive control information for the correction of static (or very low bandwidth) loads, as may be the case for large astronomical mirrors, it is often impossible to obtain data about the exact loading of the mirror structure. Strain gauges may be placed at the collimating points, to provide information for the adjustment of support forces, but they are not sufficient to sense the mirror's figure. The finite element method not only provides a means of building a working model for an arbitrary mirror structure but also supplies an interpolative approximation to the optically significant strain that is very straightforward, computationally efficient, and particularly useful for the inevitable circumstance of having incomplete data. In addition, the measurement data may be directly related to rotational degrees of freedom in the finite element model, allowing the design of meas-urement hardware to measure local surface tilt rather than optical path displacement. This procedure has the potential of providing a means of correcting the mirror's support system "off-line", that is, between relatively infrequent calibrations using data from a star image analyzer. The various standard finite element codes now commonly used provide a convenient means of constructing the stiffness matrix for a model of a large complex mirror structure. To facilitate development of the efficient iterative use of a finite element interpolation procedure, the stiffness matrix may be removed from the general finite element analysis program, partitioned, and solved with a least squares routine to provide an optical surface error estimate. Looping on this procedure and testing for convergence finally produces the correction desired. Understanding the limits of accuracy for this procedure, based on the kind and number of measurement points and the number of actuators, is crucial to the design of a figure control system.

This paper describes an alignment technique using an optical design program and an interferogram analysis code. The method is demonstrated on a two-mirror telescope. Instead of minimizing a merit function, which is normally composed of the designed system aberrations, a "demerit" function is constructed from the coefficients of the measured aberration function of an assembled (misaligned) system. The interferogram analysis program fits the aberration function to a Zernike polynomial expansion. The optimization routine of the optical design program is instructed to reproduce these coefficients using only the alignment parameters of the secondary mirror (tilt, decenter and air space) as degrees of freedom. At the final stage, the program produces the actual misalignment parameters. In the experiments described here, real interferograms were not used, but simulated systems were analyzed.

We propose to use a Shack-Hartmann camera to align and phase the thirty-six segments of the ten meter primary mirror of the Keck Telescope. Tilt errors are determined by employing the Shack-Hartmann device in a manner exactly analogous to that used in mirror testing and in adaptive optics. Piston errors are determined (with the same device) by exploiting diffraction effects near the intersegment edges. These two tests are shown to have sensitivities that are comparable to each other and that are sufficient to achieve the design goals of the telescope. We describe how the tilt test can be extended to give information both about the individual segment figures and about the global mirror parameters. Some of the complications and potential systematic effects associated with these tests are discussed.

A wave front corrector inserted at the image of a surface producing an error can remove the wave front error over a reasonable field of view. The relationship of magnitude of error, constants of a given configuration, and the residual wave front error is explored. We present the results of the production of a sample wave front error surface by diamond turning with the Large Optical Diamond Turning Machine (LODTM). The cost consequences of this new ability to upgrade a submillimeter-quality telescope to a high-acuity optical telescope are shown.

Two new lens-prism correctors are designed for secondary foci, of which the diameter of the primary is 7.5 meter, the f-ratio of the primary is 2 and the final f-ratio is about 5.4. The field of view of the R-C system is 46 and that of the quasi R-C system is 1°. These correctors can be used for the correction of both the aberrations and the atmospheric disper-sion. They consist of two cemented lenses. Each cemented lens is made of two different kinds of glasses, of which the indexes of refraction are near and the dispersions are different. The cemented surface is tilted. The two cemented lenses can rotate oppositly around the optical axis for correcting the atmospheric dispersion at different zenith distance. All of surfaces of these correctors are spherical. Spot diagrams of two systems are given for three relative orientations of two cemented lenses, each at nine positions in the field of view, which show both the monochromatic image quality and the effect of correcting the atmospheric dispersion are excellent.

Interferometric modification of the Hartmann test method has been used for observatory tests of the optics of the Danish 1.5 m telescope on La Silla. Some results of the tests and our analysing methods are presented. We use the measured wavefront aberrations to compute the point source diffraction tmage. The computing method can also be used for theoretical analysis, e. g. for investigating the effects of the mirror deflections due to the supporting system.

As part of a design study for a balloon-borne 3-m telescope for far-infrared and submillimeter astronomy, it has been necessary to test and evaluate the state-of-the-art for ultra lightweight primary mirror material candidates. In addition to being ultra lightweight (I0kg/m2), the primary mirror of this telescope must operate at -50°C with 300 diffraction limited performance. Evaluating the performance of mirror candidates has required cooling them in a low temperature test chamber while remotely monitoring the surface figure changes at several different temperatures. Some of the test panels were constructed with surfaces not smooth enough to provide recognizable 10p fringes suitable for analysis by standard Fringe techniques. To deal with this problem, we have successfully developed an analysis technique and software for a microcomputer which reliably generates surface maps from pathological interferograms. The programs operate rapidly and can provide a variety of outputs such as surface map averages, differences, Zernike polymonial coefficients, and RMS residuals during the tests. The system can also compensate for camera and test optics field distortions.

Shape variable mirrors in a large telescope are suggested so that by using least mirrors and the simplest correctors excellent image quality will be acquired at all focal stations for various works. The topic of application of mechanical methods to the field of optical design and testing is put forward. 2. Based on the ideas of Rosin, Angel et al., we propose to use small correctors instead of big ones in cases when a continuous field is not necessary, e.g., multi-object spectroscopy. 3. Segmented field focal reducers and focal reducers of conical optical fibers are introduced. Segmented field correctors are also proposed. 4. An improved wobbling mode of the secondary mirror is introduced for in-frared observations. 5. The optical configuration of a 7.5-meter telescope with above new ideas is briefly described.

The need to analyze the performance of the Solar Optical Telescope (SOT) prompted the development of a technique to merge the capabilities of structural and optical analysis. The result is the NASTRAN - ACCOS V interface. This paper describes the testing and application of this interface.

The results of a feasibility study on lightweight mirror blanks made of the glass ceramic Zerodur will be reported. This work was performed by Schott under contract to the European Southern Observatory (ESO). The goal of this study was to find the most economical way to produce a lightweighted blank of Zerodur with a diameter of 8m, a radius of curvature of 32m, a center thickness of 0.4m, and a total mass of 15,000 kg. There are three production processes which are technically feasible for the manufacture of lightweighted mirror blanks for earthbound astronomical application. The three different production methods result in different geometrical shapes of the honeycomb structure. The final decision as to which process will be used must take into consideration the differences in total costs and production time.

The ESO-VLT is based on the combination of four 8m telescopes so as to provide a 16 m equivalent aperture. Two main possibilities for the manufacture of the mirror blanks are considered: a zero expansion glass ceramic from Schott and a lightweight steel structure. Manufacturing possibilities for both options as well as European industrial potentialities for the optical figuring are being reviewed. Results of the analysis of a particular honeycomb structure for the steel mirror and experimental data obtained on several test mirrors are presented.

Large telescope primary mirrors cast in borosilicate glass hold an exciting potential for future astronomy. The Steward Observatory mirror laposaJpory has resurrected the casting technology and is heading the research effort for this type of primary ". The University of Texas large telescope group, after conferring with Steward Observatory and NOAO, has ventured into the structural analysis of some permissible structured mirror designs for the benefit of the national telescope effort as well is to consider this type of mirror as a candidate for use in an additional telescope at McDonald Observatory'. Durin;pe past three years, various designs have been proposed both by Texas researchers and the Arizona groupsJt . Because of the intricate rib structure inherent in the castings, finite element models are fairly time-consuming to construct, therefore the number of numerical models constructed tends to be few, and methods of simplifying the analysis can often help to get an overall picture of mirror flexure. In examining local flexure, however, because of the extremely small values, a fair amount of work using fine meshes and accuracy in model representation has proved valuable in identifying sane unique problems for large mirrors. Certain proposed patterns and methods of support exhibit an unexpected and troublesome amount of deflection near the outer edge under self-weight loading. In a computer environment originally designed for batch operation, we have constructed a crude computer-aided engineering system, wherein iterative application of finite element analyses is combined with a small amount of rib redesign at each iteration, and have managed to improve the flexure situation somewhat, but a superlative design that uses the entire surface to define the aperture has yet to be found. The flexure of the individual ribs near the mirror's edge suggests that a more direct loading of the outermost support points be achieved, placing more material in direct compression and having the least amount of material subject to bending. An important outcome is that serious consideration should generally be given to producing a casting oversized enough to obviate edge flexure problems, contrary to the traditional practice of considering the mechanical edge of the optic as the aperture definition.

If they are to achieve the best possible images, honeycomb borosilicate telescope mirrors must follow ambient temperature changes while maintaining low internal temperature gradients. This is best done by cooling or heating the internal, edge and back mirror surfaces at the same rate established by convection on the front surface. We propose to use directed jets of air at ambient temperature, arranged with more air flow on thicker sections so as to match cooling rates. A full scale glass thermal model of one honeycomb cell of an 8m mirror was built, along with a system to flow in air at controlled temperature. We find that the air jets realize high thermal coupling efficiency, and allow good control of internal gradients.' With the air cooling steadily at 0.25°C/hour, typical of nighttime cooling at good sites, internal gradients were < 0.1°C and the overall lag between air and glass temperature was 0.25°C. This performance, achieved with a flow rate of 6-10 liters/sec per cell, will ensure negligible image degradation from convection at the mirror surface (mirror seeing) or from thermal distortion of the mirror substrate.

Polishing a large mirror is difficult, requires a long time and is expensive. In replication technology, the large scale production of identical mirrors is possible through precise mold polishing, in very little time and at a low cost. We show here excellent reproduction of optical quality in replicating a 50 cm spherical mirror. With suitable resin systems, the final quality depends solely on the polishing of the mold. We plan to make larger size mirrors in the near future and investigate the feasibility of 3 meter replicated mirrors.

A 2-meter diameter off-axis parabola has been produced using the stressed mirror polishing technique. Surface accuracy obtained was better than 0.25μm (rms), somewhat short of the intended goal. Further improvement is limited by mechanical factors. Results are presented. No further work is planned. Some of the equipment will be loaned to the fabricator of mirrors for the Keck Ten-Meter Telescope.