The Multiple Mirror Telescope has now been in regular operation as an astronomical telescope for over a year. We report on the experience gained with the operation of this relatively complex astronomical facility. We also describe a number of changes which have been and which are being made to improve the performance of the MMT.

Manufacture of the 4.2 metre William Herschel Telescope by Grubb Parsons of Newcastle and by Marconi Radar Systems of Leicester, England, will be completed by the end of 1983. It will be installed on the island of La Palma during 1985 and will become operational in 1986. Although the design of this telescope is conventional in many ways, it does incorporate a number of special features facilitated or made possible by its altazimuth mounting. The paper describes some of these features.

The philosophy and preliminary concepts for the design of a 3-to 4-meter class telescope are presented. A clear set of goals for the telescope is motivated by the need for a workhorse instrument to complement the major new ground and space based astronomical instruments in planning or nearing completion. Primary emphasis in the design is accorded to instrument mountings so that every instrument is kept operational and easily deployable at any time. New developments in optical, mechanical, and control system technologies are being exploited throughout the telescope. Many of these innovations are described.

We are considering to have a series of telescopes in this century. As a goal, it is thought that a new technology telescope with a effective diameter over 10-m will be built at the best site with high altitude. As the first step, design works for the 3-m alt-az telescope are started by some helps of various observatories in U.S.A and Europe. This telescope will have a 3-m primary mirror with honeycomb structure and short focal length (F/2). We expect that these efforts together with others are essential to building of very large telescopes and will make them with substantial cost reduction.

The conceptual design of the Multi-Mirror 18m aperture telescope is now well advanced. The experience gained on the design of the 4.2m Altazimuth telescope now in the final stages of manufacture at Grubb Parsons of Newcastle-upon-Tyne has enabled the design team at the royal Greenwich Observatory to include in the MMT concept a number of features which influence the overall cost of the Astronomical Installation. These features are described and compared to the conventional design of telescope

The proposed 15m NNTT offers a gain in collecting area over a 4m telescope of 14 and a gain in speed of up to 200. In this paper the types of problem where the gain in limiting flux is inversely proportional to area (D2-problems) and those in which the gain goes only as the diameter (D-problems) are identified. The performance of the NNTT is compared to a 4m telescope and to the Space Telescope in three configurations. The NNTT is shown to have a small advantage over the Space Telescope for spectroscopy of point sources even at low resolutions. For work on sources of diameter one arcsecond or greater the gains are ≈ a factor of 15. Even for the direct imaging of point sources the NNTT is only a little slower than the Space Telescope when there is no confusion. Three areas of astronomical investigation, studies of star formation, the Galactic halo and high redshift galaxies, are used to illustrate the impact a 15m telescope would make scientifically. These astronomical problems make substantial demands on the telescope performance and instrumentation. If these can be met the NNTT will lead to considerable advancement in the understanding of current problems and, given the order of magnitude improvement in infrared performance and spectroscopic capability, is likely to lead to the discovery of new phenomena.

A brief background and description is given of the plan to build a 15-meter telescope. The two design concepts currently under consideration are presented and contrasted. A choice will be made in the near future.

Optical science and technology concepts for a large deployable reflector for far-infrared and submillimeter astronomy from above the earth's atmosphere are discussed. Requirements given at the Asilomar Conference are reviewed. The technical challenges of this large-aperture (about 20-meter) telescope, which will be diffraction limited in the infrared, are highlighted in a brief discussion of one particular configuration.

A balloon-borne three-meter telescope for far-infrared and submillimeter astronomy has been proposed jointly by the University of Arizona, the Smithsonian Astrophysical Observatory, and the University of Chicago. The purpose of this project is to provide a facility for photometry, spectroscopy, and imaging in the spectral region 30 micrometers to 1 millimeter, which is largely inaccessible with ground-based telescopes. The three-meter telescope will provide a much needed gain in sensitivity and spatial resolution compared with present approximately one-meter sized balloon and aircraft telescopes. The telescope is to be a Cassegrain design with an angular resolution diffraction limited to a wavelength of 30 microns. It will be supported on a three axis, gyroscopically-stabilized system with a pointing stability of one arcsecond rms. The overall weight of telescope and gondola is expected to be approximately 2800 kg, assuming a lightweight mirror formed as a welded structure of Pyrex or fused silica. We are also studying the possibility of using a carbon fiber reinforced plastic sandwich panel for the primary and secondary mirrors and carbon fiber reinforced plastic members for the telescope structure. The intended operation is approximately five 8 to 10 hour flights per year carrying two instruments at a time.

The Max-Planck-Institut air Radioastronomie and Steward Observatory are jointly planning the establishment of a submillimeter observatory on Mount Lemmon, Arizona. The telescope will have a diameter of 10 meter and enable diffraction-limited operation at a shortest wavelength of 350 μm. Thus the reflector surface accuracy will be about 15 μm and the pointing accuracy 1". This paper describes the design of the telescope, which is being carried out in cooperation with design teams in German industry. The reflector support structure will be made of carbon fiber reinforced plastic (CFRP) to minimize thermal deformations. The reflector consists of composite panels, made of aluminium honeycomb core and CFRP skins. They will be replicated from molds, made of pyrex, which will be ground in the required paraboloidal shape by optical shop techniques.

The University of Arizona and the Max-Planck-Institut fur Radioastronomie are collaborating to construct a sub-millimeter wavelength radio telescope facility at the summit of Mt. Lemmon (2791 m above sea level) near Tucson, Arizona. We have designed a corotating building to protect the 10 m diameter Sub-Millimeter Telescope (SMT) against storm damage, to provide large instrumentation rooms at the Nasmyth foci, and to minimize degradation of the reflector profile accuracy and pointing errors caused by wind forces and solar radiation.

We discuss a four element non-redundant array telescope-interferometer for ground use. The elements are 8 meter mirrors, and the maximum array spacing and two element spacing are 75m and 108m respectively. The array may be used as three separate telescopes, one 11.3m and two 8m for work not requiring highest angular resolution. We discuss the problems of making speckle measures to high enough precision for synthetic images to be produced. We conclude by showing that the high resolution presents opportunities to make types of observation that are neither possible with VLBA nor NNTT.

By adjustment of the optical pathlengths of the MMT telescopes, it is possible to make the MMT into a phased array with a 680 cm baseline. We will describe experiments in speckle interferometry and spectroscopy which have been done this way using 2 and 3 of the MMT telescopes. We will describe planned adjustments of the optical configuration of the MMT to achieve coherent operation over a large field of view with all six telescopes phased simultaneously.

Large boule or spherically mounted telescopes are mechanically feasible and economical up to mirror diameters of at least 5 metres and focal ratios of two or less. An optical array consisting of nine such telescopes would provide a telescope of 15 metres equivalent aperture for all astronomical purposes, including direct imaging since it is now feasible to do this by aperture synthesis methods. The resolving power of such an array used in the interferometric mode would be equal to that obtained by the VLBA at radio wavelengths, and would thus provide information in the optical region not obtainable by monolithic or multiple mirror telescopes of the same equivalent aperture. A mechanical and optical design of a 5-metre individual boule-type telescope is presented with emphasis on the minimum weight structure of the outer shell. The major economy achieved by means of boule telescope design is due of course to the absence of the conventional dome and building, and a comparison is made with a conventional 5-metre telescope both in size and cost, and also with the cost of the 15-metre equivalent array and other feasible designs. It will be shown that the single boule-type telescope, at least up to 5 metres aperture, costs an order of magnitude less than conventional telescopes.

Parameters are given for a corrector lens design which will provide a 0.5 degree field of good definition from 134nm to 280nm at the Ritchey-Chretien focus of the F/2-F/15 Starlab 1-metre telescope. The optimization includes the window of the detector which is aspherized on the outside but not on the inside where the photocathode is deposited. The 130 mm diameter detector is preceded by two lenses, the first of which is convex-plano. Because this element has a flat surface it can be interchanged with other elements to either compensate for the chromatic aberration of filters (by making the lens thinner), and/or to provide dispersion for multi-object slitless spectroscopy. In the latter case, the interchanged element is wedged and has a course transmission grating deposited on the flat side. This combination of grating and lens is referred to as a grens.

Two aspheric reflecting elements can correct exactly the spherical aberration and coma of a spherical primary mirror. The remaining off-axis aberrations depend mainly on the sizes of the secondary and tertiary mirrors. Two solutions are presented in which the secondary mirror is 1 meter in diameter. The first is for a 5-meter f/0.8 primary suitable for use as an optical Arecibo (fixed primary) telescope. The second is for a fully steerable telescope with a 12-meter f/1.33 primary. In both cases, 1 arcsecond images are formed at the edge of a field 4 arcminutes in diameter. The collimation tolerances for the corrector assemblies are stringent in comparison to most conventional telescopes, but modest in comparison to the precision required to maintain a segmented primary mirror.

An equation similar to the Abbe sine condition is derived for a Wolter type II telescope. This equation and the sine condition are then combined to produce a so called generalized sine condition. Using the law of reflection, Fermat's principle, the generalized sine condition, and simple geometry the surface equations for a Wolter type II telescope and an equivalent Wolter-Schwarzschild telescope are calculated. The performances of the telescopes are compared in terms of rms blur circle radius at the gaussian focal plane and at best focus.

With the continuing trend to use very fast primary mirrors in telescopes, there will be a need for optical designs for magnifying correctors. Such correctors will serve to give an aberration-corrected field of view at a larger f-number and scale than for the primary mirror alone. One solution to this is to use a system with a small secondary mirror. The final focus is closer to the prime focus than the usual Cassegrain position. One form of this is described together with a new two-lens corrector.

The 30 m Millimeter Radiotelescope (MRT) will have a beamwidth of less than 10" when operated at a wavelength of 1.2 mm. It is an open air telescope located on a mountain ridge in southern Spain. There the instrument is exposed to severe environmental influences, especially wind and temperature changes. The pointing and tracking accuracy required is of the order of a few seconds of arc. Simulations have shown that these specifications cannot be met with conventional servo design. An improvement of performance can be obtained applying the modern concept of state control. The state controller needs the acquisition of data about position and velocity at several points of the instrument in real time. Additionally the state controller requires repetition rates of a few milliseconds, in which the data have to be read, converted and the servo algorithms calculated. This is performed by microprocessors operating in CAMAC crates close to the driving system. The whole system is controlled by two identical computers. One usually controls the antenna and data acquisition from the receivers, the other can be used for data analysis or as a backup controller. All receivers are connected via CAMAC, too. Special software tools have been developed for the use of this system. They allow an easy access to the variety of different process control items needed to drive the complete system. They have proved to be a powerful aid in developing the process control hardware and in the installation phase. They will be used for the communication between the operators and the system. In the development of the system the possibility of expanding the system to remote observing has been kept open.

The use of optical telescopes for infrared and radioastronomy, where observations can also be made in daytime, requires reproducible positioning. Exceptionally good tracking is demanded particularly for radioastronomy, when the observation range extends to the millimeter- and submillimeter wavelengths. For radio telescopes with a small antenna beamwidth (below 1 arcmin) position errors should not exceed a few arcsec. This accuracy can only be achieved when the antenna position (angle) and velocity as well as the motor position and velocity are measured. The measurement of magnitudes in the arcsec region can be done with commercially available incremental encoders. For these encoders electronic circuits have been developed to guarantee the required resolution. The design of the electronic circuits also provide possibilities to correct encoder errors and interface to the host computer. The installation of the measurement electronics at the 30 meter MRT is finished, and tests have been started. These show that the required accuracy can be achieved.

A system capable of satisfying the accuracy and stability requirements dictated by Shuttle-borne payloads utilizing large optics has been under joint NASA/Sperry development. This device, denoted the Annular Suspension and Pointing System, employs a unique combination of conventional gimbals and magnetic bearing actuators, thereby providing for the "complete" isolation of the payload from its external environment, as well as for extremely accurate and stable pointing (≈0.01 arcseconds). This effort has been pursued through the fabrication and laboratory evaluation of engineering model hardware. Results from these tests have been instrumental in generating high fidelity computer simulations of this load isolation and precision pointing system, and in permitting confident predictions of the system's on-orbit performance. Applicability of this system to the Solar Optical Telescope mission has been examined using the computer simulation. The worst case pointing error predicted for this payload while subjected to vernier reaction control system thruster firings and crew motions aboard Shuttle was approximately 0.006 arcseconds.

A solar telescope, which consists of an open steel framework, is under construction. The telescope will operate without a dome in order to improve the local seeing. The telescope drives should be stable against the fluctuating wind forces. Calculations indicate that conventional gear drives show too much torsional and bending flexibility in the shaft of the pinion. In addition, the stability requires line contact between the teeth of meshing gears under a relatively low load, which requires an extremely good alignment of the gears. A new gear design is presented in which the pinion of the first gear stage and a pair of wheels of the second stage form a single block, which can rotate about a self-aligning spherical roller bearing. The line contacts between the teeth of the first and second gear stage, in combination with the spherical roller bearing, form a statically determinate system for this single block. In this way line contact between the gear stages is guaranteed and the single-block construction minimizes the torsional deflection. A shaft through the spherical roller bearing would bend too strongly. This problem is solved by a special shaft design, which incorporates the bearing in the shaft construction. The described design may be of interest for future large telescopes because it reduces the telescope vibrations caused by wind buffeting.

This paper describes the Site Evaluatibn program currently in progress for the 15-Meter National New Technology Telescope. The data obtained in this program will supersede or augment earlier tests in the critical areas of sub-arcsecond seeing evaluation and suitability for infrared astronomy. Synoptic comparisons will include sites representative of two, possibly three, distinct high altitude environments: 1) Mauna Kea in Hawaii, surrounded by a large body of water, 2) the Pinaleno Mountains in Arizona, surrounded by a low-level southwestern desert, and 3) possibly a site in the coastal region of Chile. Concurrent observations with identical equipment are planned over a two year interval.

The summit of Mauna Kea on the island of Hawaii is widely recognized as one of the best ground-based sites in the world for optical and infrared astronomy and, as such, it is a potential location for at least two of the advanced technology telescopes now in the planning stages. Because of the importance of further understanding the site characteristics for undeveloped areas near the summit of Mauna Kea, the University of Hawaii's Institute for Astronomy, which is responsible within the state for the site, has undertaken a comprehensive site survey program to (1) study the orographic properties of the mountaintop and (2) to relate these properties to the image quality as observed through a large telescope. This site survey program will be closely coordinated with the U.S. National New Technology Telescope site survey. In our survey, we will be particularly interested in simultaneous measurements throughout a two-year period of wind speed, wind direction, and microthermal turbulence over a grid of test towers placed within the summit area. With such data--and simultaneous monitoring of image quality at the summit of Mauna Kea--we hope to identify the best potential locations for large telescopes on Mauna Kea and, in turn, to effect the most responsible use of this international resource.

For superior land based astronomy locations such as Mauna Kea, Cerro-Tololo and Mt. Graham, all precautions must be taken to preserve the high quality of the seeing. In addition to the floor cooling and well insulated domes, it is essential to extract heat generated within the telescope. The critical locations are hydrostatic pads, the cassegrain station, prime focus cage, and telescope's tube cross head.

When it started operation, the 3.9 m AAT suffered from poor matching of air temperatures inside and outside the dome. There was little natural ventilation and the forced ventilation was limited, being originally designed only to supplement active control of the dome air temperature (which had been cut from the installation to save funds). Obviously, improvements were essential and reliable measurements of local seeing were needed to decide the best way to proceed. Comparison of routinely observed seeing with inside/outside temperature differences indicated local degradation of about 0.5 arc sec diameter for each 1°C. Fast response thermometers showed that the worst air temperature fluctuations were generally confined within about 5 m from the dome aperture. A Hartmann technique, devised to allow direct comparison of the seeing of the 3.9 m telescope with that of a small telescope, mounted either inside or outside the dome, confirmed the magnitude of the dome seeing. After a careful review of heat flows within the building, the forced ventilation was substantially upgraded and improved seeing has resulted. However, a variant of the Hartmann technique, using the 3.9 m telescope in autocollimation, has shown that local degradation is still often 0.5 arc sec diameter or worse. Work to quantify the remaining causes of seeing degradation will continue, including the development of a system for routine microthermal analysis. The optical and microthermal techniques used at the AAT are widely applicable to the seeing studies vital for the large telescopes of the future.

Astronomical seeing has been studied by measuring image size and motion using relatively small (~0.5 m) telescopes. These techniques have been successfully extended to telescopes in the 2-4 m class. For site survey investigations, it is desirable to use a compact telescope elevated such that seeing measurements are free of image degradation due to ground effects. The smaller instrument suffers from image degradation due to diffraction under the testing conditions of greatest interest, namely, sub-arcsecond seeing. Although image motion measurements can be extended to conditions of excellent seeing, it has not been demonstrated that small telescope images are representative of large telescope images with regards to image size or motion. In an attempt to understand image scaling laws between large and small telescopes, we have studied image motion and profiles for telescopes from 0.6 to 4.0 m in diameter under .seeing conditions extending from sub-arcsecond to several arcseconds. A model is presented in which image motion data taken with small telescopes can be used to predict image profiles that would be measured with a 15 m class telescope. We also consider the signal to noise aspects of measuring seeing by way of image motion and small telescopes.

As part of a program leading to the production of 8m honeycomb mirrors, we have recently made two 1.8m blanks. These have honeycomb sandwich form, with hexagonal honeycomb ribs sandwiched between front and back plates. Each is cast in one piece from borosilicate glass, using techniques that can be extended to larger sizes. Pieces of the glass are melted together in a circular container made of hard ceramic tiles, held together against hydrostatic pressure by bands of nickel alloy. Voids in the glass to give the honeycomb structure are formed by hexagonal blocks of ceramic fiber, held down against flotation with silicon carbide bolts. Liquid glass runs over the blocks to form the face sheet, and under the blocks, which are spaced above the base tiles, to form a back sheet with holes. After the casting has been annealed and cooled, the base tiles are unbolted and the ceramic fiber blocks removed from the glass honeycomb by water blasting. Both blanks are of high quality, free from cracks and voids, and with an adequately low bubble content. The second and better blank, made of Ohara's E6 glass, is now to be figured to high precision, 0.25 arcsecond images, and is to be tested for an extended period in the Multiple Mirror Telescope.

A facility is to be built to make 8m diameter glass honeycomb mirror blanks by casting, in the same way that has been demonstrated with 1.8m blanks. The only major difference is that the larger furnace will he rotated on a turntable so as to preform the deep parabolic surface needed for f/2 mirrors. This paper explores the tolerances in glass homogeneity, thermal control and support of the blank necessary to meet the stringent imaging requirements of telescopes in the best ground based sites. In round numbers, homogenity in expansion coefficient of 10-8/°C and thermal equilibriation to 0.1°C are required. Laboratory measurements show that both can be met by a ventilated honeycomb of borosilicate or similar glass. Adequate resistance to wind pressure and buffeting can be achieved by an axial support that responds to pressure on the three defining points. The total time of about 6 weeks for melting, annealing and cooling of the blanks is set by the time constant of around 8 hours for the glass honeycomb to follow the furnace temperature.

The SIRTF optics are intended for operation around 10°K. Construction and evaluation of the telescope assembly at room temperature would be most convenient, but require that optical performance changes (other than focal shifts) on cooling remain within the assigned system tolerances. "Diffraction limited" performance for 2-μm radiation is a desirable goal for SIRTF. With support from the Space Technology Branch, Space Science Division, NASA Ames Research Center, Itek Optical Systems has completed measurements of the optical figure of a 0.65-meter, lightweight, fused silica mirror at temperatures in the 13 to 17°K range. Test requirements and preparations were reviewed at the 26th Symposium. Test results, and our evaluation of their significance are presented here.

For ground-based large diameter lightweight Pyrex mirrors, internal ventilation is a necessity in order to minimize the thermal distortions. However, the demand on the ventilation system does not appear to be significant. Measurement on a sample blank showed significant effects on the blank's thermal time constant and temperature distribution. Compared with the unventilated case, a mild air flow (17 feet per minute) reduces the thermal time constant from 3 hours to 1 hour. A similar kind of flow (21 feet per minute) reduces the internal temperature difference from 2.4°C to 0.3°C, as well as reduces the temperature difference between the blank and the ambient air from 3.2°C to 0.5°C.

Several different types of passive support systems, suitable for a thin meniscus type mirror in a 3-metre optical telescope, were investigated. A thin F/2 primary mirror with a diameter-to-thickness ratio of 25, as compared to the usual 6 to 8 ratio, is proposed. In order to minimize astigmatism in the mirror figure a support system for a thin mirror must exhibit very low friction or hysteresis effects. Comparisons on the advantages and disadvantages of using pneumatic supports, counterweighted lever supports and hydraulic supports for the axial support mode were made. For a more detailed assessment of the performance of a counterweighted lever system, utilizing commercial flexural pivots instead of ball bearings, full size prototype support units of the axial and radial type were built and tested. Friction and hysteresis losses were observed to be 0.02% for the axial support unit and 0.07% for the radial support unit. These values are low and they are consistent with the requirements of mirror support systems of high quality thin optics.

The techniques of modal control analysis derived by Creedon and Lindgren [1], an error analysis method proposed by Howell and Creedon [2], and various methods of finite element analysis are applied to a meniscus mirror of unusually large aspect ratio. The analyses predict the figure error for a set of characteristic error disturbances and correction strategies. Several arrangements of piston actuators are shown to he near optimum relative to the disturbance inputs considered, and in each case the related plots of modal coefficients, rms surface error, and the Howell and Creedon error analysis demonstrate the efficacy of the thin meniscus approach. In addition, the engineering approach to achieving a stable optical surface in this way is shown. Decomposing the distorted optical surface into modal terms gives it a natural order and allows the analyst to demonstrate that for some very harsh mechanical loads, relatively few modes are the chief contributors to the wavefront error. Therefore, by designing the active system for the control of the most significant modes, the mechanical engineering task becomes manageable.

A method of monitoring the figure of the primary mirror of the University of Texas 7.6-m telescope is under study, in which slope errors are determined by reflection of laser beams from small flat mirrors cemented to the back of the thin meniscus primary mirror. The method requires calibration against a stellar image wavefront sensor but does not require continuous presence of the stellar image for operation. The method will be described, together with an error analysis. An optimum distribution of test points will be shown based on finite element and modal structural analyses of the mirror blank. A discussion of possible operational errors attributable to the laser system, based on similar experience at the University of Arizona - Smithsonian MMT Observatory, will be given.

A new interferometric method for optical testing has been developed. The test arrangement is similar to that of the Hartmann method using a screen with a rectangular grid of holes. The detection takes place close to the focus where the Hartmann images are partly overlapping and interfering. The positions of the interference maxima contain information on the wavefront errors of the optical system. As the size of the interference maxima is considerably smaller than the size of the Hartmann images obtained with the same screen, the accuracy of the position measurements is respectively better. Also much smaller hole spacing can be used. An application of the method could be wavefront error sensing in active optical systems. A CCD camera could ideally be used as detector for the measurement of the wavefront errors and for generating the error signals to the active optical system.

The Cassegrain secondary of the CFHT telescope is bent in its cell in order to remove residual spherical aberration. Bending forces are generated by pneumatic pressure and vacuum. Their resultant is kept equal to zero in order to "float" the mirror in its cell and avoid support reaction.

The stressed mirror polishing method is being used to produce two thin two-meter diameter off-axis parabola mirrors. These mirrors are prototypes for a large segmented mirror telescope and the initial objective is to achieve 0.03 micron (RMS) surface accuracy. The fabrication set-up will be described along with improvement modifications that have proven to be necessary. Polishing results to date will be presented.

The University of Arizona has initiated the construction of a precision generator for larger diameter optical surfaces (large optical generator). This LOG machine will be capable of generating a rotationally symmetric primary mirror of up to 7.3 meters diameter, and segments of mirrors to 10 meters and larger diameters. Initially the machine will be used for generation of molds for a sub-millimeter telescope array, but is eventually intended for wide application mirrors of visual wavelength accuracy.

The design of the University of California Ten Meter Tele-scope (TMT) calls for a primary mirror composed of 36 hexago-nal segments. An active control system maintains the optical figure of the array in the face of gravitational, thermal, and wind perturbations. Three degrees of freedom of each segment, piston, and tilts about two axes, are controlled. An engineering demonstration of the active control system and the development of segment fabrication techniques are in progress and will be completed this year. We report below the current program status. Major goals include the development of optical testing methods and the evaluation of segment support, control actuators, and sensors, as well as the overall performance of the control system. The engineering demonstration will use a full-sized segment, 1.8m across and 0.075m thick. The fabrication of this very thin and flexible segment presents unusual challenges. The polishing support system has been developed. We have also developed test apparatus and procedures that provide an extremely accurate measurement of the radius of curvature. A circular mirror has been polished to a specified radius of curvature. Any warping that results from cutting the segment from this mirror will be established soon. The passive support of a segment is to be provided by two systems, one to carry axial loads and one to carry radial loads. The design of these systems is described. The TMT active control system is designed to use capacitive sensors to measure the relative orientations of the segments, and torque-motor driven displacement actuators to adjust segment position. The demonstration system uses 4 sensors and 3 actuators to control the tilt and piston of a single segment with respect to a reference mirror. The design of the system prototype is described.

The active control of segmented mirrors requires actuators to move the segments in response to perturbations. Each segment of the University of California Ten Meter Telescope has three of its six rigid-body degrees of freedom actively controlled; piston and tilt about two axes. The system design requires the actuator to carry a load that varies as the telescope moves from zenith to horizon. The maximum load is one third of the segment mass, about 150kg. The system design also needs actuator adjustment resolution less than 20nm over a range of 3mm with a 2µm/sec response rate. Actuators which satisfy these requirements have been designed, built, and tested. A torque motor turns a screw shaft whose axial load is taken by a roller thrust bearing. Simultaneously the screw drives a roller nut to position the mirror segment. The roller screw converts rotary to linear motion with nanometer smoothness over a large dynamic range. A stick-slip behavior in the thrust bearing makes the mechanical system non-linear for small motions. Each actuator has a microprocessor-controlled servo loop and the servo loop algorithm compensates for this non-linear behavior. The actuator design and servo loop algorithm are described and the results of servo loop performance tests are given.

The measurement of the relative orientations of mirror segments is the basis for the active control of a segmented mirror. We describe a system of capacitive displacement sensors which will be used to make these measurements in the University of California Ten Meter Telescope. The global sensing system design and the control algorithm are summarized. The detailed mechanical and electronic designs of a sensor are described along with the performance of prototype sensors under a variety of tests. The relationship between the sensor characteristics and the image quality of the segmented mirror is described. The sensor output is 1 millivolt per nanometer of displacement. This sensitivity and the small noise, drift, and thermal sensitivity are adequate to allow essentially continuous figure control to the precision required for optical and infrared astronomy.

Developments are reported which establish a basis for the precision alignment of the segmented optical elements of a large, multipanel telescope. The apparatus described in this paper employs a form of Young's double slit interferometer and a CCD photodetector array to obtain a measure of the step as well as the gap between adjacent mirror segments. A simple adjunct to this arrangement can also provide a means for measuring the relative angle between the normals at the edges of adjoining mirror segments. Despite the preliminary nature of the developments reported herein, the sensor is based on well-known principles and the critical criteria for establishing a viable sensor system have been established experimentally. The parameters of adjacent mirror segment step, gap, and relative angle can all be measured with but a single sensor.

Eighteen years of observing data for the largest telescopes at Kitt Peak National Observatory and six years of similar data at Cerro Tololo Inter-American Observatory are given in Tables 1 and 2. Precipitation data for the same period at KPNO are contained in Table 4. Some summaries of these data and a few conclusions useful for telescope design considerations are discussed.