The first 10 meter Keck telescope has been fully scheduled for astronomy since early 1994. Commissioning the initial three instruments and optimizing the primary mirror performance had occupied most of the previous year. Subsequently, considerable effort has been spent on mechanically and electrically improving the dome and shutter. Despite these problems, the percentage of time lost from astronomy to faults has been reasonably low. Although there remains much room for improvement, e.g. in more carefully controlling the dome thermal conditions and maintaining better alignment of the telescope optics in routine operation, the optical performance has been very encouraging. The median image FWHM reported with scientific instruments has been about 0.7 arcsec and there are firm indications that the site, the optics, and the instrumental seeing will allow us to reach a median of about 0.5 arcsec.

The second of two 10-meter telescopes comprising the W. M. Keck Observatory is nearing completion. Functionally, the Keck II telescope is a twin of Keck I, but in detail, many improvements have been made. Observatory and scientific instrument budgets are presented for the two telescopes. A new software system was developed for Keck II using EPICS-based architecture. Computer architecture for Keck II was also completely changed from the Keck I design using VMS and VAX computers to UNIX and SUN computers. The new telescope is completely assembled on the site on Mauna Kea, Hawaii. Design, construction, and testing of the Keck II telescope has taken significantly less time due to the experience and tools developed for the first telescope. An adaptive optics system is currently being developed for Keck II. Preliminary design of this system is complete and the system is expected to be commissioned in 1998. Configuration of the twin 10-meter telescopes was designed to allow combining of the optical beams from the two telescopes and to add smaller satellite telescopes for interferometry. Plans for this phase are being developed in detail.

Construction of the Gemini Telescopes is underway. The first mirror blank has been completed, concrete piers on both Mauna Kea and Cerro Pachon have been poured, fabrication of the telescope structures has started and the erection of the first enclosures have begun. In this paper, we give a progress report on the Gemini 8-M Telescopes Project. In addition, we highlight some of the unique scientific characteristics of the Gemini Telescopes including our approach to image quality performance, the use infrared wavefront sensors, the development of silver coatings and our 'adaptable observing' system. The scientific performance of these telescopes will be heavily dependent on atmospheric conditions, Gemini will be allocating at least 50% of its observing time to 'queue scheduled observing.' The implications of adopting this novel observing mode on the design of the control system and on the telescope operations model are briefly discussed.

Subaru telescope is a Japanese 8.2-m optical-IR telescope under construction on the 4,130-m altitude summit of Mauna Kea, Hawaii. The project is in the sixth financial year of the construction plan of nine years. The construction is now coming into the assembling phase, and we are starting to get some realistic qualitative estimation of the performance which will be achieved by Subaru telescope. We report here the general status of the construction, and part of such recent perspective. Some more details of individual parts of construction are described by other authors.

The Hobby-Eberly Telescope, nearing completion at McDonald Observatory in west Texas is an optical Arecibo-type telescope utilizing an 11-meter primary mirror and a 9.2-meter effective aperture. Innovative approaches have been employed to provide this large modern telescope at a total cost of $13.5 million. A joint project of the University of Texas, The Pennsylvania State University, Stanford University, the University of Munich, and the University of Goettingen, the telescope will be completed in mid 1997. First light is expected in mid 1996.

Operated by the Multiple Mirror Telescope Observatory (MMTO), the multiple mirror telescope (MMT) is funded jointly by the Smithsonian Institution (SAO) and the University of Arizona (UA). The two organizations equally share observing time on the telescope. The MMT was dedicated in May 1979, and is located on the summit of Mt. Hopkins (at an altitude of 2.6 km), 64 km south of Tucson, Arizona, at the Smithsonian Institution's Fred Lawrence Whipple Observatory (FLWO). As a result of advances in the technology at the Steward Observatory Mirror Laboratory for the casting of large and fast borosilicate honeycomb astronomical primary mirrors, in 1987 it was decided to convert the MMT from its six 1.8 m mirror array (effective aperture of 4.5 m) to a single 6.5 m diameter primary mirror telescope. This conversion will more than double the light gathering capacity, and will by design, increase the angular field of view by a factor of 15. Because the site is already developed and the existing building and mount will be used with some modification, the conversion will be accomplished for only about $20 million. During 1995, several major technical milestones were reached: (1) the existing building was modified, (2) the major steel telescope structures were fabricated, and (3) the mirror blank was diamond wheel ground (generated). All major mechanical hardware required to affect the conversion is now nearly in hand. Once the primary mirror is polished and lab-tested on its support system, the six-mirror MMT will be taken out of service and the conversion process begun. We anticipate that a 6 - 12 month period will be required to rebuild the telescope, install its optics and achieve f/9 first light, now projected to occur in early 1998. The f/5.4 and f/15 implementation will then follow. We provide a qualitative and brief update of project progress.

Construction is underway on two 6.5 meter telescopes for Las Campanas Observatory in Chile. This paper gives the present status of the Magellan Project.

The large binocular telescope (LBT) project have evolved from concepts first proposed in 1985. The present partners involved in the design and construction of this 2 by 8.4 meter binocular telescope are the University of Arizona, Italy represented by the Osservatorio Astrofisico di Arcetri and the Research Corporation based in Tucson, Arizona. These three partners have committed sufficient funds to build the enclosure and the telescope populated with a single 8.4 meter optical train -- approximately 40 million dollars (1989). Based on this commitment, design and construction activities are now moving forward. Additional partners are being sought. The next mirror to be cast at the Steward Observatory Mirror Lab in the fall of 1996 will be the first borosilicate honeycomb primary for LBT. The baseline optical configuration of LBT includes wide field Cassegrain secondaries with optical foci above the primaries to provide a corrected one degree field at F/4. The infrared F/15 secondaries are a Gregorian design to allow maximum flexibility for adaptive optics. The F/15 secondaries are undersized to provide a low thermal background focal plane which is unvignetted over a 4 arcminute diameter field-of-view. The interferometric focus combining the light from the two 8.4 meter primaries will reimage two folded Gregorian focal planes to a central location. The telescope elevation structure accommodates swing arms which allow rapid interchange of the various secondary and tertiary mirrors. Maximum stiffness and minimal thermal disturbance continue to be important drivers for the detailed design of the telescope. The telescope structure accommodates installation of a vacuum bell jar for aluminizing the primary mirrors in-situ on the telescope. The detailed design of the telescope structure will be completed in 1996 by ADS Italia (Lecco) and European Industrial Engineering (Mestre). The final enclosure design is now in progress at M3 Engineering (Tucson), EIE and ADS Italia. Construction activities on the Emerald Peak (Mt. Graham) site have resumed in the summer of 1996.

A project to build a large telescope at the Spanish Observatorio del Roque de los Muchachos (ORM), in the island of La Palma, has been underway for several years. Spanish astronomy has progressed steadily to the point where gaining access to large telescope time is key to its continued and future growth. Also the technological situation within Spain is such that building a large telescope is regarded as a scientific and advanced technology endeavor worth investing in. The Gran Telescopio Canarias (GTC) project has thus been granted approval in February this year, with an allocation of 51% of the budget. Current activities of the telescope project are mainly concentrated in the conceptual design and analysis of the telescope optics, including plans for the adaptive optics, its mechanical structure and enclosure, and an active campaign of site testing. We are also carrying out detailed programs aimed at producing and testing prototype models for sensors, actuators and the control system for aligning the primary mirror segments. In the following sections, the current status of the activities being carried out by the telescope project team are described.

The Vatican Advanced Technology Telescope incorporates a fast (f/1.0) borosilicate honeycomb primary mirror and an f/0.9 secondary in an aplanatic Gregorian optical configuration. We provide a brief technical and performance overview by describing the optical layout, the primary and secondary mirror systems, and the telescope drive and control system. Results from a high resolution wavefront sensor and a current wide-field image taken at the f/9 focus demonstrates the overall fine performance of the telescope.

The Gemini 8 meter telescopes performance with active and adaptive optics as a system is given. The telescopes are being designed to deliver near diffraction limited images at infrared wavelengths to the focal plane. This is achieved with a combination of innovative telescope design, a fully active control system and a natural guide star adaptive optics (AO) system for the Mauna Kea Telescope. The predicted delivered performance while under full active control is given at 2.2 microns. The top level AO system error budget is presented including the effects of instrumentation. The Gemini telescopes have been designed from the outset to be fully active; from control of the primary mirror surfaces and positioning of the secondary to ventilation of the enclosure by control over the ventilation gates. Descriptions of the concepts used in the various subsystems have been published previously. Here, we emphasize the system level interactions between the Gemini adaptive optics system and the telescope and instruments. This includes a performance summary of how the telescope operates with and without AO. First, the current system concept is outlined, which includes wavefront sensors/guiders in the following areas: (1) acquisition and guiding system, peripheral wavefront sensors; (2) scientific instruments, on instrument wavefront sensors; (3) adaptive optics system, facility wavefront sensor. The system performance depends upon the interactions of these three key sensor areas. For non-AO use, both peripheral and on instrument wavefront sensor may be used to support fast and slow guiding and active control of the telescope alignment and wavefront. For AO use, combinations of all three types of wavefront sensors may be used for adaptive atmospheric compensation in addition to the functions listed above. The system is designed to quickly change between modes of operation (AO to non-AO and back) under remote control.

Each of the 12 Nasmyth and Cassegrain foci of the ESO very large telescope (VLT) will be equipped with an 'adapter/rotator' which provides the mechanical interface for the science instruments and several key functions for the control of the telescope, namely a CCD sensor for acquisition and guiding, and a separate CCD sensor as wavefront sensor for the active optics control system. This paper describes the origins and concept for the VLT adapter/rotators, and the principal design drivers and constraints.

The mechanics of the secondary unit are located behind the secondary mirror as seen from the telescope focus. The secondary is undersized and defines the pupil of the telescope. There is a focusing and centering drive for slow adjustments of the secondary mirror. In addition, the secondary mirror has a fast beam steering mechanism for chopping and rapid guiding to remove atmospheric wavefront tilt during observations. The specified square wave chopping frequency is 5 Hz with a duty cycle larger than 80%. To achieve a high bandwidth, the secondary mirror is manufactured of light-weighted beryllium coated with nickel. The beam steering mechanism has a counter-vibrating mass to compensate for dynamic forces and moments. The chopping mechanism has been successfully tested. The code of the digital control used during the tests was generated using Matlab real time toolbox. The servos were implemented on a digital signal processor card equipped with a TMS 320C40. To compensate for resonances inside the bandwidth of the servos, a special filter is applied in the velocity loop. The design of the secondary unit is now completed and fabrication and assembly have begun.

The very large telescope (VLT) of ESO will be placed on Cerro Paranal in the Atacama desert in northern Chile. This site provides excellent conditions for astronomical observations. However, it is likely that important seismic activities occur. The telescope structure and its components have to resist the largest earthquakes expected during their lifetime. Therefore, design specifications and structural analyses have to take into account loads caused by such earthquakes. The present contribution shows some concepts and techniques in the assessment of earthquake resistant telescope design by the finite element method (FEM). After establishing the general design criteria and the geological and geotechnical characteristics of the site location, the seismic action can be defined. A description of various representations of the seismic action and the procedure to define the commonly used response spectrum are presented in more detail. A brief description of the response spectrum analysis method and of the result evaluation procedure follows. Additionally, some calculation concepts for parts of the entire telescope structure under seismic loads are provided. Finally, a response spectrum analysis of the entire VLT structure performed at ESO is presented to show a practical application of the analysis method and evaluation procedure mentioned above.

Active and adaptive structures, also commonly called 'smart' structures, combine in one integrated system various functions such as load carrying and structural function, mechanical (cinematic) functions, sensing, control and actuating. Originally developed for high accuracy opto-mechanical applications, CSEM's technology of flexure structures and flexible mechanisms is particularly suited to solve many structural and mechanical issues found in such active/adaptive mechanisms. The paper illustrates some recent flexure structures developments at CSEM and outlines the comprehensive know-how involved in this technology. This comprises in particular the elaboration of optimal design guidelines, related to the geometry, kinematics and dynamics issues (for instance, the minimization of spurious high frequency effects), the evaluation and predictability of all performance quantities relevant to the utilization of flexure structures in space (reliability, fatigue, static and dynamic modeling, etc.). material issues and manufacturing procedures.

The 8 m coating unit for the VLT mirrors is designed for the deposition of high reflective, homogeneous aluminum coatings. For the process of the film deposition the sputter technology is utilized. The design of the following major subsystems is completed: the vacuum vessel and the vacuum generation system, the thin film deposition equipment and the glow discharge cleaning device, the substrate support and rotation system as well as the supporting framework and the auxiliary equipment. Manufacturing of the coating unit has started. The pre- assembly and testing activities, which will take place prior to the shipment to the site, are defined. This paper describes the design features and the major performance requirements of the 8 m coating unit. The performance of the sputter source design has been verified in a qualification test. The deposition rate, the film thickness and reflectance, as well as the film purity have been measured. The test set-up and the results of the qualification tests of the selected magnetron type are presented and discussed.

The 'two-degree field' (2dF) project at the Anglo-Australian Observatory (AAO) gives the 3.9 m Anglo-Australian telescope (AAT) a field of view two degrees in diameter at the prime focus, equipped with 400 optical fibers for multi-object spectroscopy. The basic components of 2dF are the corrector lens optics, the robot which positions the fibers and a pair of spectrographs. All these are mounted on a new 'top end ring' so that the whole assembly can be easily put on and off the telescope. Here we give an update on the status of 2dF, highlighting features which have changed or been developed since earlier reports.

Conventional astronomical telescope makes use of a Ritchey- Chretien 2-mirror telescope, with a limited FOV except in the case of use of complex field corrector inducing spectral range limitations. For the future, large imaging telescope could offer main scientific advantages, like: (1) obtain a 3D description of the content of a large volume of the universe; (2) galaxy content and morphology; (3) galaxy red shift; (4) dark matter distribution; (5) absolute length scale. The present paper proposes for this large imaging telescope mission, the use of a Korsh, 3-mirror telescope thats characteristics are: (1) size approximately equals 2.5 meter; (2) field is greater than or equal to 1.5 degrees (but with no light on the optical axis); (3) image quality is less than or equal to 0.3 feet; (4) multispectral capabilities: from 0.35 micrometer up to 2.5 micrometer and more if needed; (5) real exit pupil with flat mirror; (6) good focal plane accessibility allowing multiple instrumentations: turret rotation of the previous flat mirror can be used, with several fixed instruments.

We describe the procedure and the results of the final optimization of the elevation structure of the large binocular telescope. The aim of such a process was to achieve the highest structural rigidity with the lowest total mass, taking into account the basic geometry and center-of-gravity constraints. Based on the method of the 'modal strains,' the optimization process took into account the first three resonant frequencies at seven equally spaced zenith angles, for two different restraint cases. The free parameters of the optimization were cross sections of the truss elements and thicknesses of the membrane elements. The telescope structure was divided into many groups. For each group, the mass was varied in proportion to the maximum strain experienced in the above mentioned 42 modes of vibration.

We present the final design of the alt/az structure of the large binocular telescope. As a final report of the structural performances of the telescope, this paper describes how the azimuth platform and the primary mirror cells have been modeled. Furthermore, a definition of the simulation of the various structural interfaces is given. Finally, the static and dynamic responses at various zenith angles are reported.

The ESO 8-meter telescopes, known as the VLT, utilize new technology in the design to achieve high image quality. Although previous telescopes have employed extensive component analysis and modeling, certain aspects of telescope performance are only captured by system level modeling. Therefore, to better understand design features on system performance, an end-to-end simulation model was developed jointly by ESO and Ball Aerospace. This required combining the diverse disciplines of structural dynamics, optics, atmospherics, control systems, signal processing, and random process modeling for disturbance generation into a single time simulation model. The end purpose of such at tool is to help in verifying requirements and performance for different VLT subsystems on one hand and to help understanding potential problems that could occur during the commissioning phase on the other hand. This paper describes the architecture and contents of the model, and gives some fist simulation results.

The analysis of more complicated, higher performance telescopes, instruments and telescope/instrument systems has improved through advances in the techniques of stray light analysis. These include the use of more sophisticated opto- mechanical modeling and improved assembly of stray light models. Improvements have also been made in creating more realistic scatter models for diverse surfaces, including contaminated surfaces. The large number of systems analyzed in the last few years has also led to an improved understanding of code strengths and limitations.

The main structure is defined as the telescope mechanical structure including the drives, the encoder system, the hydrostatic bearing system and all those subsystems which make the system self standing safe and testable as an electromechanical system, with the exclusion of the velocity and position control loops. The main structure is now almost completely assembled in Milan. The tests of the main mechanical performances (dynamic and static) have been carried out to gather information at the earliest possible stage of the assembling activities. The drives and the hydrostatic bearing system have been tested concerning their functionality. This paper aims to summarize concisely the results of the tests, to compare them to the design calculations and to show some possible design changes which could improve the performances of the telescope.

The Optikzentrum Northrhine-Westfalia has developed and constructed a number of LIDAR telescopes with mirror diameters in the range from 0.46 to 1.80 m. The paper describes shortly the features of the instruments and some auxiliary equipment.

Instrument design is, for the most part, a battle against errors and costs. Passive methods of error damping are in many cases effective and inexpensive. This paper shows examples of error minimization in our design of telescopes, instrumentation and evaluation instruments.

Safety, like quality and reliability, has to be designed into a product and respected during all project phases from the concept definition to the operation and maintenance phases. The VLT approach towards occupational safety and health and equipment safety starts with the definition of realistic safety requirements and applicability of ECC directives and national laws of the ESO Member States. The approach continues with preliminary safety analyses during the early project phases, with hazard analysis and safety verifications during the developmental phases, the training for safe operation, maintenance, and later material disposal. System safety is an integral part of the VLT project.

The Gemini 8-M telescopes project is an international partnership of six countries to build and operate two 8-meter telescopes, one on Mauna Kea in Hawaii and one on Cerro Pachon in Chile. The construction phase of the project has demanding scientific requirements, a fixed budget that is tight, and an aggressive schedule. The work is distributed internationally between the Gemini Project Office in Tucson, Arizona, organizations in the partner countries, and industrial contractors. The project organization and management procedures used to cope with this challenging situation are described. Plans are now being formulated for management of Gemini in the operational phase. The organization proposed to operate Gemini in a cost effective manner is also described.

The Gemini 8-m telescopes project is faced with the challenges of demanding scientific requirements, fixed budgets, and aggressive schedules. In addition Gemini is an international project and the majority of the work itself is being done by groups within the partner countries. Gemini presents a unique managerial challenge in that not only the detailed design and fabrication of the mechanical subassemblies will be contracted out to the international community but also the software, control systems and instrumentation. In order to meet requirements on budget and on schedule Gemini has had to develop a model for how to do distributed design, manage a distributed effort and adopt a development methodology tailored to its specific circumstances. This paper details the engineering practices put in place to ensure quality, the software and hardware standards adopted, and the management techniques and tools used to promote success. In addition this paper describes some of the lessons learned and, most importantly, what the author would change if this were repeated.

This paper gives an overview of the present status of the Telescopio Nazionale Galileo (TNG), a 3.5 m, active-optics telescope for the Italian community, whose initial characteristics were derived from those of the ESO NTT. For a more detailed description see e.g. in Barbieri et al. (1994). Its site is the Observatorio del Roque de los Muchachos, in the Canary Island of La Palma, on the West side of the mountain, at an altitude of about 2360 m. Construction and erection activities, started in 1993, are nearing completion. The telescope structure has been installed inside the rotating dome. The three main mirrors have also been transported to the mountain and aluminized in the WHT plant. They will shortly be installed in the telescope. Three major subsystems still undergo intensive activity in Italy, namely testing of the M2 and M3 units, testing of the operational version of the control hardware and software, and construction of the two rotator adapters for the Nasmyth foci. First light instruments are also being built. Arm A of the telescope is reserved for the imaging section, composed by a visual camera, a near-IR camera, plus a common adaptive optics module. On the other arm B, a faint object spectrograph with long slit, atmospheric dispersion corrector, multiobject and imaging capabilities, will be mounted. A fixed high resolution spectrograph with optical derotation is also being designed. Attention has already been given to the archive of the data. It is planned to have first light before the end of 1996, and to start regular scientific operations after an adequate period of debugging and commissioning.

The 3.8 m UK infrared telescope (UKIRT) is currently the focus of an upgrades program to improve its imaging performance, ideally to approach its diffraction limit in the near-IR at 2.2 micrometer, with FWHM approximately 0.'12. This program is now in its late stages. All the new systems have been designed, most have been manufacture and many have been installed. A new top end carries an adaptive tip-tilt secondary mirror with active precision alignment, which, with low-order active control of the primary mirror, should provide the desired intrinsic optical performance. The adaptive tip- tilt system will correct image motion from telescope vibrations and drive errors and from atmospheric wavefront tilt; delivered images are expected regularly to be less than 0.'5 over wide fields, and within a factor 2 or so of the diffraction limit, at least inside an isoplanatic patch of order an arcmin radius. To reduce facility seeing the primary mirror has been equipped with a ventilation system and will receive a 5 kW cooling system; the dome is being equipped with sixteen closable apertures to permit natural wind flushing, which can be assisted by the building air handling system in low winds. It is hoped that facility seeing -- excluding boundary layer effects -- will be imperceptible during approximately 85% of observable time. The upgraded UKIRT should be well capable of exploiting fully the very best conditions on Mauna Kea.

The Astronomical Instruments Group of Carl Zeiss Jena has supplied 13 RCC telescopes with 1 m aperture within the last 35 years. The last one of this series was procured by DARA together with a 12.5 m observatory dome from Zeiss to be used at the new optical ground station of ESA in Tenerife in 1991. The fully computer controlled astronomical telescope, however, will be utilized due to its versatile design for a rather diversified range of unique applications, such as optical communication via modulated laser beams to satellites and check-out of space debris. It has been equipped with a number of additional instrumentation, such as a 4 k by 4 k pixel CCD camera cooled with liquid nitrogen and a focal plane optical bench with highly sophisticated transmit and receive optics and high power lasers.

A study on the evaluations of the active optics system performance for the primary mirror of the Gemini telescopes project was conducted. Finite element analysis was employed to analyze the optical surface figures of the primary mirror. Four distinct influence matrices (generic 3-point kinematic system, combined unit case, 3-zonal multiple constraint system, and high frequency localized force set) were established based on the unit load cases at each of 120 active support locations and the restraint boundary conditions in the mirror model. A least square algorithm was developed. This scheme is able to accommodate the design constraints in the active optics systems. The active optics performances were evaluated by analyzing their capabilities in compensating the optical surface figures. For each active system, the level of the calculated active forces and the surface residual errors were examined. The results indicated that their performances are excellent and are in good agreement for most sampling surface figures. The results from the active optics performances based on the generic, the combined, and the 3- zone cases were practically identical for all the object surfaces. The active optics system based on the high frequency localized force set tends to predict a higher surface RMS error and requires a bigger force set. This effect became apparent when this system was applied to correct the surfaces from the natural mode analyses.

To minimize the wind buffeting effect on the primary mirror figure, the Gemini primary mirror cell is designed to provide additional mirror stiffness by coupling the mirror to the cell structure through a six-zone hydraulic support system. Therefore the cell structure is designed as though it were a light weight mirror for minimum top surface distortion. This paper describes the design requirements, the design features, and the detail predicted performance of this cell structure, particularly the effects on the primary mirror figure. As the cell structure supports the primary mirror with a six-zone hydraulic system, the mirror is coupled to the cell structure with three degree of freedom overconstraints. These overconstraints induce the possible distortion on the mirror figure due to the cell deformation. This paper presents a solution to eliminate this effect by supporting the mirror cell on the telescope structure through four bipods. The locations of the bipods are so arranged that the cell deformation will not distort the mirror figure as the telescope rotates from zenith to horizon pointing.

The mirror-cells of the LBT (large binocular telescope) 8.4 m honeycomb borosilicate primary mirrors have to meet various requirements in addition to providing support to the mirrors and to the Gregorian instrumentation. The mirror-cells are directly connected to the main telescope structure and have a structural function themselves in order to supply a very high stiffness boundary to the position actuators (hardpoints) of the primary mirrors. The cells also must guarantee an overall strength to make up the bottom part of the vacuum shell, whose top part is the bell-jar for the mirror aluminizing. Each mirror cell has to hold several components inside: 160 pneumatic actuators for the active optics of the mirror, the thermal control system and its 252 air ejectors, and 6 position actuators. A further requirement for the mirror cell design is also to provide access for the maintenance of all the above sub-systems. In this report we summarize the main mirror-cells functions, their final design and briefly describe how we met all the specifications.

The primary mirror cell of the very large telescope supports the primary mirror, the tertiary tower and mirror, and the Cassegrain instrumentation. Stringent requirements have been set to achieve the desired image quality, flexibility of use, and the necessary mirror safety. This paper describes the most important requirements set on the system and some of the design solutions which were chosen.

We review the present status of liquid mirror telescopes. Interferometric tests of liquid mirrors (the largest one having a diameter of 2.5 meters) show excellent optical qualities. The basic technology is now sufficiently reliable that it can be put to work. Indeed, a handful of liquid mirrors have now been built that are used for scientific work. A 3.7-m diameter LMT is presently being built in the new Laval upgraded testing facilities. Construction of the mirror can be followed on the Web site: http://astrosun.phy.ulaval.ca/lmt/lmt-home.html. Finally we address the issue of the field accessible to LMTs equipped with novel optical correctors. Optical design work, and some exploratory laboratory work, indicate that a single LMT should be able to access, with excellent images, small regions anywhere inside fields as large as 45 degrees.

By combining the excellent intrinsic thermo-mechanical properties of the SiC (silicon carbide) with a structural design based on a sandwich structure composed of two SiC face sheets CVD (chemical vapor deposition) deposited on a foam core of the same material, it is possible to manufacture very light and stiff primary mirrors for telescopes to be operated in space or on ground. The paper presents an analysis of the mechanical properties of this structure, it proposes a preliminary design of a segmented mirror and support pattern for space or for on ground telescopes, and it points out the important saving in weight which is possible with respect to other approaches used to manufacture lightweight mirrors.

Within the preliminary developments related to the Gran Telescopio Canarias, a test rig is being designed for the active control of the support system of the segmented primary mirror. The construction of this test rig will be divided in two phases: the first one will basically consist in the implementation and testing of a displacement sensor prototype as well as an actuator prototype. Phase 2 will consist in the development and characterization of the whole test rig, including two segment simulators and a number of displacement sensors and actuators. Also, a non contact optical system for the test rig behavior verification will be constructed during phase 2. This paper presents the conceptual design adopted for the active control system proposed for the telescope and a brief description of the development program, including the requirements of the displacement sensors and actuators. We intend the test rig not only for testing the active control system components, but also for checking different control strategies.

The main final results in terms of stresses and optical performances are reported for the large binocular telescope (LBT) primary mirrors. The two borosilicate LBT primary mirrors f/1.14 have 8.4 diameter and are produced at the Steward Observatory Mirror Lab (SOML). They are honeycomb shaped in order to achieve light weight, short thermal constant and high stiffness. The back plate is flat and the upper is paraboloid shaped. Each elementary cell has, in the lower plate, one circular hole permitting the ventilation of cell itself. The material used is the borosilicate Ohara E6. Different supporting systems have been analyzed from the mirror casting to the operative conditions, i.e.: supporting system during the cooling of the casting phase; supporting system for the handling after the casting phase and before the optical surface grinding and polishing; supporting system for the handling after the optical surface polishing and for maintenance; passive support system in non-operative condition; supporting system in operative condition. The stress checks carried out show that the values of the maximum principal tensile stresses are below 0.7 MPa for long times and/or stresses affecting large volumes, and are below 1.05 MPa for short times and small volumes. Optical performances in operative condition respect the specification.

Large convex aspheres are notoriously difficult to fabricate because of the tremendous cost and difficulty of making accurate measurements of the optical surfaces. The new 6.5- and 8-m-class telescopes require convex secondary mirrors that are larger, more aspheric, and more accurately figured than those for existing telescopes. Two powerful measurement techniques have been implemented at the Mirror Lab and demonstrated to be accurate and economical. The polished surfaces are interferometrically measured using holographic test plates. This measurement technique uses full-aperture test plates with computer-generated holograms (CGH) fabricated onto spherical reference surfaces. When supported a few millimeters from the secondary and properly illuminated with laser light, an interference pattern is formed that shows the secondary surface errors. The hologram consists of annular rings of metal drawn onto the curved test plate surface using a custom-built writing machine. This test has been implemented for secondaries up to 1.15-m diameter, with 4 nm rms surface measurement accuracy. In addition to this test, a swing arm profilometer was built to measure the rough surface during aspherization and loose abrasive grinding. The machine uses simple motions and high quality components to achieve 50 nm rms measurement accuracy over 1.8-m mirrors.

The 6.5 m upgrade of the multiple mirror telescope (MMT) will include a number of new secondary mirrors. For first light, there will be an f/9 Cassegrain secondary manufactured from a 1.0 m diameter Hextek borosilicate honeycomb meniscus blank. This f-ratio is designed to match that of the present MMT to allow the use of existing instrumentation for first light. This will be followed by the wide field f/5 secondary combined with a refractive corrector which includes an atmospheric dispersion corrector (ADC) to give a 1 degree unvignetted Cassegrain field. The f/5 mirror is made from a 1.7 m diameter lightweighted machined Zerodur blank. Two f/15 0.64 m diameter secondaries are being designed. The first of these is an adaptive secondary consisting of a thin 2 mm thick shell faceplate with 300 voice coil actuators and associated capacitive displacement sensors. A chopping f/15 secondary is planned using a rigid lightweight blank such as silicon carbide. The 8.4 m large binocular telescope (LBT) will have two secondaries for each of the two primaries. For first light, 0.87 m diameter f/15 Gregorian adaptive secondaries are planned. These concave mirrors will use the same thin shell faceplate, voice coil actuator and capacitive sensor technology currently being developed for the MMT f/15 adaptive secondary. A pair of 1.25 m diameter f/4 Cassegrain secondaries will be built next. These will be used together with refractive corrector optics to give a 1 degree field. These mirrors are being polished and tested at the Steward Observatory Mirror Laboratory (SOML) using the recently completed Secondary Fabrication and Test Facility. Stressed lap polishing is used to achieve the fast, highly aspheric surfaces and testing is done with the computer generated hologram (CGH) test plate technique. Each of these secondaries requires a support system, five axis actuation and thermal environmental control. The in-house development of this number of secondaries enables an integrated design approach. As much as possible of the development, design and hardware costs will be shared between secondaries. This paper describes the designs which are being developed for the support, actuation and thermal control for each of these secondaries.

The primary mirrors of the ESO 8-m class very large telescopes are actively supported, thin Zerodur menisci, 8 - .2-m diameter. The mirror blanks are produced by SCHOTT; the optical figuring, manufacturing and assembling of interfaces and auxiliary equipment are done by REOSC. Three mirror blanks have already been delivered by SCHOTT to REOSC. In November 1995 the project met a critical and very successful milestone, with the completion and testing of the first finished VLT primary mirror at REOSC. Specifications, manufacturing and above all testing methodology are addressed, and the final results are detailed. Optical performance at telescope level is assessed as well.

The Subaru 8.3 m mirror blank was sent in the summer of 1994 from Corning to Contraves US near Pittsburgh for polishing and figuring. It is expected to be delivered to the summit of Mauna Kea by the end of 1997. Hole boring on rear side is under process and the turnover of the mirror is planned for this autumn. Fabrication of the cell and actuators as well as that of the mirror control system have been finished in Japan. Performance tests using the actual mirror cell and a steel dummy mirror have shown the Subaru mirror support system satisfies the technical specification set forth in the design phase.

The Steward Observatory Mirror Lab is in the process of fabricating the 6.5 m mirror for the conversion of the multiple mirror telescope (MMT) to a single primary mirror. For this purpose the lab has developed a versatile polishing system built around the stressed lap polishing tool. The system must produce an f/1.25 parabolic surface with an accuracy corresponding to 0.09 arcsecond FWHM seeing and 1.5% scattering loss at 500 nm wavelength.

We describe the project management procedure applied to the manufacture of 8.2 m Zerodur blanks. Such a complex, challenging task needs detailed organization from planning via implementation to realization. Some examples of inspection and handling procedures are presented in detail. The excellent performance of all four 8.2 m Zerodur blanks is demonstrated with some characteristic properties like geometrical dimensions, internal quality and coefficient of thermal expansion. The delivery of the fourth blank in 1996 represents the culmination of a decade long development and production project.

It seems to be a fundamental law of nature that optical surfaces become dirty. On-site contamination has been recorded at the ESO La Silla Observatory and at the VLT site over a period of six years. Measured data are presented, and the efforts made at ESO since 1990 to define suitable on-line monitoring and preventive maintenance are detailed. In-situ cleaning techniques, existing equipment and procedures are reviewed. Emphasis is put on the carbon-dioxide snowflake cleaning technique and the integrated cleaning device of the 3.5 m NTT telescope is described. The preliminary cleaning and protection test conducted on the first finished 8 m mirror at the optical manufacturer's site is presented as well, and plans for the in-situ cleaning of the VLT mirrors are explained.

This study investigates the feasibility of a fast, wide-field imaging telescope using a mosaic CCD detector system. A 3- mirror design is proposed to obtain good image quality over a 2 degree field-of-view at an aperture of 2.5 m. The layout is shown to be practical with current technologies, leading to a modern telescope with potentially very good seeing properties.

The three-mirror Rumsey telescope is a flat field anastigmat. Compared to the curved field Schmidt telescope, its compactness and full achromaticity are particularly interesting features for potential CCD mosaics. In counterpart the difficulties involved in aspherizing the three mirrors explain the fact that no such telescope is presently being built for larger sizes than 30-cm primaries. This paper provides solutions to this problem via active optics methods; it considers the case where the tertiary is linked in a continuous geometry to the primary, both mirrors being aspherized from a same sphere. The optical and elasticity designs of a 1-meter telescope are presented. The active aspherization can be achieved as well by stress polishing as in situ at the telescope.

The future large telescope for the Observatorio del Roque de los Muchachos will have a 10 m diameter segmented primary made of hexagonal elements (similar to the Keck I primary). The telescope optical configuration is being selected to achieve excellent image quality and low emissivity out to the mid-IR. The design includes provision for two exchangeable secondary mirrors, one of which is to be an adaptive secondary. In order to perform an independent optical calibration of the segmented primary and the adaptive secondary mirrors, the telescope with the adaptive secondary will be of the Gregorian type, whereas with the conventional secondary it will have a Ritchey- Chretien configuration. Current work to define both telescope configurations and to make them compatible is presented in this poster.

The invention relates to a mirror system with two mirrors, comprising a concentrating reflector and a diffusing reflector fitted on the same optical axis, an image field and a detector. The diffusing reflector has a central drilling. A diaphragm is arranged in such a way that the point of intersection of the plane of the diaphragm and the optical axis forms the center of the curved image field of the mirror system. At the edge of the diffusing reflector there is a central diffuse light shutter. The concentrating reflector may optionally have a central drilling. The mirror system may be used as a telescope system. Prior art two-mirror systems like the Ritchey-Chretien system do not correct astigmatism, image field curvature and distortion. According to the invention, spherical aberration, coma, astigmatism and distortion are corrected by double reflection from the concentrating reflector and the substantially hyperbolic shaping of both mirrors.

The invention relates to a mirror system with two mirrors and four reflections, comprising a concentrating reflector and a diffusing reflector fitted on the same optical axis, an image field and a detector. The concentrating reflector has a central drilling. The concentrating reflector reflects the light to the outer part of the diffusing reflector, from where the light is reflected on the concentrating reflector again, that reflects the light on the central part of the diffusing reflector, that in turn reflects the light through the central drilling to the image plane. Prior art two-mirror systems like the Ritchey-Chretien system do not correct astigmatism, image field curvature and distortion. According to the invention, spherical aberration, coma, and astigmatism are corrected by double reflection at both mirrors, whereby the inner part of the diffusing reflector is elliptical, spherical or ellipsoidal shaped. That depends from the axial radius of curvature of the diffusing mirror in relation to the distance between both mirrors. Hence -- this shape depends from the overall focal length of the mirror system about the paraxial focal length of the concentrating reflector.

The Cambridge optical aperture synthesis telescope (COAST) is a four element interferometer which measures visibility amplitudes and closure-phases. It produced its first images in 1995 and is now in a complete form, very similar to the original conception. In this paper we discuss the design and current status of the interferometer.

The Cambridge optical aperture synthesis telescope (COAST) has now been developed to the point where stellar images with a resolution of 20 mas can be produced in a routine manner. Based upon our experiences in the design and commissioning of COAST this paper discusses the possible design of a next generation interferometer.

The interferometric mode of the ESO very large telescope (VLT) permits coherent combination of stellar light beams collected by four telescopes with 8 m diameter and by several auxiliary telescopes of the 2 m class. While the position of the 8 m telescopes is fixed, auxiliary telescopes can be moved on rails, and can operate from 30 stations distributed on the top of the observatory site for efficient UV coverage. Coherent beam combination can be achieved with the 8 m telescopes alone, with the auxiliary telescopes alone, or with any combination, up to eight telescopes in total. A distinct feature of the interferometric mode is the high sensitivity due to the 8 m pupil of the main telescopes, with the potential for adaptive optics compensation in the near- infrared spectral regime. The VLT interferometer is conceived as an evolutionary program where a significant fraction of the interferometer's functionality is initially funded, and more capability may be added later while experience is gained and further funding becomes available. The scientific program is now defined by a team which consists of a VLTI scientist at ESO and fifteen astronomers from the VLT community. ESO has recently decided to resume the construction of the VLTI which was delayed in December 1993, in order to achieve first interferometric fringes with two of the 8 m telescopes around the year 2000, and routine operation with 2 m auxiliary telescopes from 2003 onwards. This paper presents an overview of the recent evolution of the project and its future development.

In 1991 the Astrophysics Division of NASA's Office of Space Science convened the Space Interferometry Science Working Group to consider in more detail the science goals of a space interferometer mission to do wide-angle astrometry at optical wavelengths. In addition, the working group considered the merits of alternative mission concepts for achieving those goals. We describe the current state of the adopted mission concept, and review the candidate astrometric science program. In addition to the main goal of precision astrometry, the concept interferometer has a limited capability for high- resolution imaging using rotational aperture synthesis. A phase A start on this mission has been made in 1996, an launch is planned for 2003.

Planets with mass similar to Jupiter's are now known to orbit nearby stars. Are there also planets like Earth? If so, their thermal emission should be directly detectable, and thermal spectra could identify the strong features of carbon dioxide, water and ozone at the levels seen in Earth. But the very close angular separation (approximately 0.1 arcsec) and huge brightness difference (approximately 10⁷) between a star and such a planet present a technical challenge. Space interferometry could in principle solve both problems, by using destructive interference to cancel out the stellar emission, and aperture synthesis to recover high angular resolution images. We show how these two functions conflict, and point to a new interferometer design which allows them to be reconciled. One key technical challenge is to combine beams with strictly controlled amplitude and achromatic phase inversions, so as to cancel the stellar disc flux by a factor of a million. We show how refractive elements analogous to an achromatic lens can be used for this purpose.

ISLA will be an astronomical observatory, operating at the upper limit of our planet Earth atmosphere, offering space like observing conditions in most aspects. ISLA can be maintained easily, modified if required, and operated for very extended periods without polluting the stratosphere. ISLA's two 4-m and four 2-m telescopes will operate diffraction limited from 0.3 micrometer, over the infrared, far-infrared to the sub-mm spectral range. ISLA's individual telescopes will permit imaging with 20 milli-arcsec spatial resolution in the optical. ISLA's telescopes can be combined to form an interferometer with a maximum baseline of 200 m with nearly complete coverage of the u, v plane. Interferometric resolution will be of the order of 20 micro-arcsec for the optical. ISLA will thus offer significantly better spatial resolution than the intercontinental VLBI radio interferometers. ISLA's efficiency will be many orders of magnitude better than comparable ground-based telescopes. The light collecting power of ISLA's interferometric telescopes are orders of magnitudes greater than the space interferometers under discussion. ISLA, being an aviation project and not a space project, can be realized in the typical time scale for the development of aircraft: about 5 years. ISLA's cost for the whole observatory, including its movable ground station, etc., will be of the order of a typical medium size space mission. ISLA's lifetime will be in excess of many decades, as it can easily be maintained, modernized, repaired and improved. ISLA will become the origin of a new astronomical international organization with many participating countries in Europe, etc. ISLA's telescopes will be of the greatest importance to all astronomical fields, as it will permit to study much fainter, much more distant objects with microscopic spatial resolution in wavelength regions inaccessible from ground. ISLA's many telescopes permit easily simultaneous observations at many wavelengths for rapidly varying objects, from symbiotic stars to QSOs.

In September 1995 the Cambridge optical aperture synthesis telescope (COAST) became the first optical interferometer to produce an image of a stellar source from phase-closure and visibility amplitude measurements. These observations demonstrated for the first time the feasibility of operating long-baseline optical/near-infrared interferometers for high dynamic range high-resolution imaging. Here we present these and subsequent observations made with COAST and describe the methods used to analyze such data.

Access to the precision of the Stellar Reference System has to be improved for the coming years and decades. With new intermediate angle measurements (a few degrees), a dual-object IR interferometer can bring a decisive contribution to the Tycho Reference Catalogue. The instrument should be highly efficient, in terms of sensitivity and automatic star captation. Two possible beam combination schemes are discussed in this respect: a correlator interferometer and a Michelson stellar interferometer.

In the last few years, we have undertaken a number of studies and experiments to assess the impact of various environmental factors on the performance of the ESO very large telescope interferometer (VLTI). The investigated topics include atmospheric turbulence, wind loads on the telescope structure, vibrations created by equipment, natural thermal variation, thermal load from electronics, natural and man-made seismic noise, as well as acoustic noise. A first report of this activity was given in a previous paper. This paper presents the final results obtained in 1995. The main outcome is the very good confidence that the VLT 8 m telescope and the infrastructure design is adequate for interferometric use at optical wavelengths down to the visible.

The optical path difference model (OPD model) determines where to position the delay lines in order to compensate for on-axis delay as seen from an astronomical object of known coordinates. This model is equivalent to a pointing model but applied to the interferometric delay. The objective is to reduce the time to search for fringes and to improve the delay lines blind tracking accuracy. This aspect is of prime importance not only when considering the overall operational efficiency of the interferometer but also its ability to quickly observe a set of program objects even after relocation of the auxiliary telescopes. The optical path difference model is based on a precise knowledge of the interferometer configuration by including a set of calibration measurements. This paper describes the main characteristics of the model and includes the results of a simulation developed to fit telescope axis misalignments which contribute to optical path difference errors.

Due to the small separation (6 m) between its two 8.4 m mirrors and to the fact that they are on a common mount, the LBT interferometer operates in a way that is very different from other large telescope interferometers and more similar to that of a telescope with an elongated pupil. Important features of this optical configuration are the simple resulting PSF, the wide coherent field, the high optical efficiency and the low emissivity of the combined focus. In order to exploit fully these characteristics the telescope was designed in a way that makes possible the use of an advanced adaptive optics system, based on adaptive secondary mirrors and artificial reference stars. This AO system is presently being developed and the first results obtained with a laboratory prototype of the adaptive secondary mirror are very encouraging. Optics for beam combination that provides a coherent field larger than the isoplanatic field has been designed for various wavelengths between visible and mid IR. The entire system was evaluated more in detail in the near IR range, consider to be of the highest priority because in this range LBT should be capable of achieving at the same time an extremely high sensitivity and an angular resolution of about 0.01 arcsec.

The 10-meter class Hobby-Eberly telescope (HET), now nearing completion, provides technology for optical Arecibo-type telescopes which can be extrapolated to even larger apertures. Utilizing a fixed elevation angle and a spherical segmented primary mirror provides cost effective and pragmatic solutions to mirror mounting and fabrication. Arecibo-type tracking implies a greatly reduced tracking mass and no change to the gravity vector for the primary mirror. Such a telescope can address 70 percent of the available sky and exhibit optical quality easily sufficient for effective spectroscopy and photometry. The extremely large telescope takes advantage of several key engineering approaches demonstrated by the HET project to achieve a cost comparable to similarly-sized radio rather than optical telescopes. These engineering approaches include: bolted pre-manufactured primary mirror truss, factory manufactured geodesic enclosure dome, air bearing rotation of primary mirror, tracker, and dome systems directly on concrete piers, and tracking via a hexapod system. Current estimates put the cost of the ELT at $200 million for a 25-meter aperture utilizing a 33-meter primary mirror array. Construction of the ELT would provide the astronomy community with an optical telescope nearly an order of magnitude larger than even the largest telescopes in operation or under construction today.

A 25 m four mirror live optics telescope is studied. M1 is spherical with 141 segments and f/0.96. M1 is re-imaged onto M4, also with 141 segments. Image FWHM is less than 0.10 arc sec over greater than 20 arc min. A horseshoe solution with a simple azimuth platform is applied. M1 segments are supported by a fine meniscus form truss structure, tied to the horseshoes by a coarser mesh. A FEM with 10⁴ dof was developed and applied. Live optics control M1 and M4 segments (the latter with potentially high bandwidth), M1/M4 segment balancing and servos. Correction signals in tilt, coma and defocus are traced. A correlation tracker and a laser guide star system are included. Low and high wind speed regimes are studied. An end-to-end simulation model is developed, based on modal representation of our FEM. Image quality dependence on wind load is studied from segment piston and tilt deflections. Eigenmodes are recorded. Using wind time series, we study dynamic effects and image quality resulting from the 141 segment spots. Automatic segment control at a bandwidth of only 1 Hz gives excellent image quality. We foresee to reach a bandwidth greater than 50 Hz, securing a system partly adaptive, with effects of atmospheric wave front tilt removed through M4 segment tilting at high frequency. Further progress include optimization of mechanical design and end-to-end simulation model, wind tunnel testing and studies of wave front sensor, correlation tracker and instruments. A fully adaptive system is tentatively studied as is coherent operation at IR wavelengths.

The tremendous growth in the building of large 8 m and 10 m telescopes, which give substantial gains in sensitivity over the current 4 m telescopes, presents a significant challenge to the builder of a future 21st Centrum groundbased telescope. To try and explore the possible scientific motivations that may drive a future groundbased facility, I have chosen a current observational project whose completion is beyond the capabilities of our new generation of telescopes. By examining what is required of a groundbased facility to undertake spectroscopy on the majority of the objects in the Hubble deep field (HDF), it becomes apparent this project will need a very large imaging infrared array (VLIA) or a 50 m telescope. The main conclusion of this comparison is that any groundbased facility capable of undertaking this project is likely to cost at least one billion dollars. The choice between the two differing approaches should therefore be driven by the scientific aspirations of the 21st century community of astronomers. Superficially, the 'scientific edge' probably belongs to the VLIA facility, with its ability to probe structures at infrared wavelengths down to the milli-arcsecond scale. The more profound issue is whether it is time for groundbased astronomers to begin looking to space for the placement of their next 21st Century telescope.

The basic design and an analysis of the performance possibilities of a 25 m class optical telescope are presented here. The configuration consists of a 28 m segmented spherical primary M1 followed by three highly aspherical corrective mirrors M2, M3 and M4 which also deviates from Cartesian shape. The construction is axially folded. The combination M1- M2 forms a focus close to a coupling aperture in M4, and the combination M3-M4 relays this focus to the final focus behind M1 and M3. The combination M2-M3 images M1 on the segmented M4 to be used for correction of wavefront errors induced by M1 form errors arising from gravitational sag and windbuffeting. Several types of aspherical figuring of M2, M3 and M4 all resulting in a field performance better than characterized by a rms spot radius smaller than 0.1 arcseconds within a full FOV of 21 arcminutes are presented.

Segments alignment is one of the fundamental factors affecting the shape of the PSF of a segmented mirror telescope. Tilt and piston of each segment must be very well and uniformly adjusted in relation with the rest of the segments. The Hartmann-Shack sensor is very efficient in detecting the local wavefront tilt, but the task to take measurements of piston using this sensor is annoying enough specially when the wavefronts are affected by atmospheric turbulence. We show, with numerical simulations, that curvature sensing is sensitive enough in order to detect piston errors in the segmented mirrors, even in the presence of the atmospheric turbulence. It would permit the taking of measurements of piston with natural stars during observation time.

The efficiency of new generation ground based optical telescopes with large segmented mirrors will depend largely on the solution of the problems in the way of the development of methods and hardware for adaptive mirror control. A methodological approach to conceptual design is given here as well as description and results of the tests of the segmented primary mirror angular position control system, which has been basically developed during construction of the AST-1200 experimental telescope with seven segments primary mirror.

Based on system analysis of technical and operational requirements and on actual trends in large optical telescope development a concept and an alternate design are proposed for telescope load-bearing case representing the first attempt to integrate functions of different elements of telescope consisting in making physically integral tube, mounting and enclosure which has not been met in the world practice of telescope construction.

The prices of more than 60 telescopes of the most diverse types and sizes are analyzed according to telescope aperture. Actualized prices in 1995 were calculated taking inflation into account, as well as fluctuations in the DM exchange rate. We discuss the dependence of prices on factors like size, mass, optical system, number of construction parts (similarity laws), as well as on special technical requirements. The relationship between telescope aperture and costs shows interesting differences between the classical astronomical telescopes and the new, large telescope projects. The number of repeated construction parts and the number of realized instruments with the same design is of great importance. Cost reductions are best achieved in the conception and design phase of telescope development.

The VLT enclosures main functions are to protect the telescopes during operational as well as non-operational phases from any adverse weather conditions and to provide optimal conditions for observation. An adequate design of a ventilation and wind protection system is important for the performance of the enclosures with respect to the minimization of the corresponding seeing effects. The VLT enclosures are equipped with ventilation doors on the azimuth platform level, with louvers on the rotating part and with a windscreen at the observing slit. Extensive qualification tests of the louvers and windscreen mechanical assemblies have been performed during the enclosures development phase. This paper gives an overview over the general layout of the enclosures and the major subsystems, summarizes the main functional specifications and gives the main results and conclusions of the functional performance tests. Presently the first enclosure erection is nearing its completion and pre- commissioning of all systems will commence. The status of the site erection of the enclosures is presented and the planning for the next phases of the erection is presented.

This paper describes the operational model for control of thermal and wind flow environment inside ESO's very large telescope (VLT). The purpose of this model is to manage the local seeing in an optimal way using the systems available in the VLT. The technique applied for optimal temperature control is first discussed. Given a meteorological forecast for the ambient temperature of the following night, a model-based procedure can determine the optimal air cooling temperature set point and mirror back plate cooling time histories in order to minimize a performance index related to seeing. Using temperature sensors in the VLT, this procedure will, in practice, constantly be repeated throughout day and night in a fixed horizon predictive control manner. Results of extensive wind channel measurements are then presented. The average root-mean-square (rms) value of turbulent pressure fluctuations on the primary mirror has been measured for various opening configurations and wind directions. Based on a correlation between wavefront aberration an the average rms pressure, the optimal ventilation configuration can be chosen, with the additional constraint of maximizing enclosure flushing. Measured values for wind velocities on the primary mirror can be used in the thermal model for evaluating forced convective exchanges. The integrated model, coupling aerodynamical and thermal optimization, is then presented, an its implementation for the VLT is discussed, as well as further ideas for improvement.

The Gemini Project plans to implement thermal control of the primary mirror using two distinct systems. The first system consists of temperature controlled radiation panels behind the primary mirror that are used to control the bulk temperature of the primary. The second, a surface heating system, will be used to adjust the optical surface temperature by passing an electrical current through the reflective coating. Combining these two systems allows the optical surface temperature of large monolithic mirrors to follow nighttime ambient temperature fluctuations, minimizing mirror seeing effects. To aid in the design of a surface heating system for the Gemini 8-m primary mirror, a development program was initiated. As part of this program, analysis techniques were developed and prototype systems using mirrors up to 1-m were fabricated and tested. This paper reviews the progress and results of this surface heating development program to date.

The preservation of the excellent image quality of the 'Roque de los Muchachos' Observatory (ORM) is one of the most important design criteria for the enclosure of the future large telescope. Additionally, the cost of a large telescope and its instrumentation is so high that it would be regrettable if its own enclosure deteriorates the telescope performance for astronomical observations. For these reasons, the IAC and CEANI are conducting an aero- and thermo-dynamical study based on numerical simulations, attempting to minimize the seeing degradation due to factors within the control of the designers. This paper first describes the general concept of the enclosure we are planning for the future 'Gran Telescopio Canarias' (GTC) and second the set of three dimensional simulations we are conducting to evaluate the effect on seeing of four different enclosure topologies located at two possible sites.

The control of the telescope thermal environment at the 3.8-m United Kingdom Infrared Telescope (UKIRT) is based on the requirements that dome seeing should not degrade the image quality by more than 0.05 arcsec (FWHM) and that mirror seeing should be reduced to negligible proportions. After quantifying steady state and transient heat flow around and through the building, we set out on a program to meet these requirements. Major telescope enclosure upgrades to address dome seeing include natural dome ventilation with 16 apertures in the base of the dome and for near still-air nights, forced-air ventilation via the plant room exhaust system. To address mirror seeing, we are in the process of installing a day-time mirror cooling system that can drive and/or keep the primary mirror between 0 degrees Celsius and 2.5 degrees Celsius colder than the predicted night-time local dome air temperature. Nevertheless, during the night, if the primary mirror is warmer than the local dome air, a flushing system is available to blow away warm convective air cells as they form. This paper describes design considerations of the natural dome ventilation system (DVS), the hardware of the primary mirror cooling and flushing system and the performance of the mirror flushing system on a dummy mirror segment.

Possible sites for an astronomical observatory have been explored. Maritime stable air is expected to reach Pico de Paul (1640 m) almost perfectly unaltered, 4 km windward from the sea, providing there is extremely good seeing. About 62% of the nights in summer and 41% of the nights in winter are expected to be clear. Seeing measurements have been started.

We discuss a device for real time compensation of image quality deterioration induced by atmospheric turbulence. The device will permit ground based observations with very high image resolution. We propose an instrument with two channels. One is an ordinary image detection channel, while the other uses a Hartmann-Shack wavefront detector to measure image degradation. This information is obtained in the form of a set of lenslet focus shifts, each corresponding to the local tilt of the wavefront. Through modeling, the entire wavefront is reconstructed. Consequently, we can estimate the optical transfer function and its corresponding point spread function. Through convolution techniques, the distorted image can subsequently be restored. Thus, image correction is performed in software, eliminating the need for expensive live optics designs. Due to the nature of atmospheric turbulence, detection and correction have to be made with 50 - 100 frames per second. This implies a need for very high computing capacity. A study of the mathematical operations involved has been made with special emphasis on implementation in the hardware architecture known as radar video image processor (RVIP). This hardware utilizes a high degree of parallelism. Results available show that RVIP together with complementary units provide the necessary high-speed computing capacity. The detection system in both channels must meet very high demands. We mention high quantum efficiency, fast readout at low noise levels and a wide spectral range. A preliminary investigation evaluates suitable detectors. ICCDs are so far the most promising candidates.

In order to encourage adequate dome ventilation to reduce or eliminate dome seeing at the 3.8 m United Kingdom Infrared Telescope (UKIRT), a dome ventilation system (DVS) was designed to be installed in the lower dome skirt. The modifications to the dome for the new DVS apertures consisted of installing a reinforcing frame containing an insulated rollup door and adjustable louvers. This paper describes the finite element structural analysis of the reinforcing frame, the detailed design of the frame hardware, the design of the programmable language control (PLC) system for controlling the opening and closing of the rollup doors, and the fabrication and installation of a prototype frame assembly. To date, a prototype assembly has been installed that confirms the design, and fifteen production assemblies are currently under fabrication for installation by September 1996.

Efficiency of image selection technique using high-speed shutter controlled by wave-front sensor or seeing monitor are estimated. This estimate shows that we can obtain moderately improved images with acceptable loss-time, for example FWHM equals 0.54 arcsec image can be obtained using 20% of observing time when seeing size is 0.62 arcsec.

This article summarizes the seeing effects of the atmospheric turbulence found in the immediate vicinity of the telescope, namely that generated by the enclosure and the telescope itself. The main contributions from the telescope and the enclosure are elucidated and quantitative engineering parameterizations applicable to an actual telescope and observatory design are provided.

Beginning in 1975 Sternberg Astronomical Institute of Moscow University (SAI) developed a search of places with the best astroclimate in Middle Asia. Mount Maidanak (150 km to south from Samarkand) was chosen after investigation of the meteorological conditions, temperature fluctuations and seeing quality by astroclimatical expeditions in a different city testing for Moscow University Observatory. Having an isolated summit Maidanak has good astroclimatical parameters: 2000 clean observational hours/year, median seeing about 0.7 arcsec (Artamonov et al. 1987, Bugaenko et al. 1992). At the end of 1992 SAI mainly finished the construction of Maidanak Observatory with a 1.5 meter RC telescope, but in 1993 the development of the observatory was stopped after nationalization by Uzbekistan. At present Sternberg Astronomical Institute and Tashkent Astronomical Institute (new owner of the observatory) continue to work in joint observations and try to create International Maidanak Observatory.

The new generation of ground-based telescopes offer unique observational opportunities for astronomy. By adopting the paradigm that observations must be matched to conditions it will be possible to use an 8 m telescope to couple the spatial resolution of the Hubble space telescope with a least 10 times the collecting area. To effectively exploit these characteristics will require a considerable degree of pre- planning and prediction of environmental and atmospheric conditions combined with the ability to dynamically schedule observations. In this paper we describe the approach being taken by the Gemini telescopes to implement this new observing mode which is essential to realize Gemini's ambitious science requirements.

The singular way of scheduling at major observatories has favored certain types of astronomy. This has led to the discrimination of certain types of observations which could not be accommodated. It is the goal of future operations to open possibilities for some additional types of observational astronomy. Together with improved observatory operating the new observing modes can provide significant progress in the acquisition of astronomical data. Specifically, the capability to optimally match the schedule to observations should prove a major advantage of service observing. The VLT data flow project is designed to accommodate these new observing modes together with conventional observing. Simulations and first experience with the ESO NTT will prepare and refine the concepts and procedures foreseen for the VLT. The acceptance of service observing by the astronomical community is of critical importance for this new operational mode to succeed.

The basic objective of modern observatories is to globally maximize their efficiency and ensure a high, constant and predictable data quality. These challenges can only be met if the scientific operation of such facilities, from the submission of observing programs to the archiving of all information, is carried out in a consistent and well controlled manner. The size, complexity and long operational lifetime of such systems make it difficult to predict and control their behavior with the necessary accuracy. Moreover they are subject to changes and are cumbersome to maintain. We present in this paper an object-oriented end-to-end operations model which describes the flow of science data associated with the operation of the VLT. The analysis model helped us to get a clear understanding of the problem domain. We were able in the design phase to partition the system into subsystems, each of them being allocated to a team for detailed design and implementation. Each of these subsystems is addressed in this paper. Prototypes will be implemented in the near future and tested on the new technology telescope (NTT). They will allow us to clarify the astronomical requirements and check the new operational concepts introduced to meet the ambitious goals of the VLT.

The Instituto de Astrofisica is finalizing the design for a 10 m class optical/infrared telescope at the Observatory Roques de los Muchachos, La Palma. It is widely recognized that the traditional observing modes are inefficient so it is intended to explore other alternatives to maximize the scientific return from the telescope and to provide the widest range of observing modes. This paper outlines the impact of the proposed observation philosophy on the Spanish 10 m telescope.

This article briefly surveys the life and scientific work of Tycho Brahe (1546 - 1601), one of the greatest astronomers of all times. He successfully designed and constructed the most advanced astronomical observatory of his time on the island of Hven and during twenty years he effectively directed what is considered to be the first, modern research institute. He inaugurated a new era of observational astronomy and emphasized the need to determine instrumental errors, just before this field of natural science was revolutionized by the invention of the telescope. He laid the observational basis for Kepler's investigations of the planetary motions which eventually served as a cornerstone for Newton's description of the universal force of gravitation. Tycho's challenges were similar to those of his colleagues of our days and it is of some interest to understand the environment in which he operated and the keys to his success.

This contribution, The History of Optical Theory of Reflecting Telescopes and Implications for Future Projects, is a shortened form of the Karl Schwarzschild lecture given in Bochum in September 1993. Some material has been added from an invited paper given in Padua in December 1992. For a full account, with figures and tables, the reader is referred to these two papers.

Interferometric arrays of many large telescopes will follow the current precursor interferometers. A few dozen telescopes, equipped with adaptive optics for intra and inter-aperture phasing, mobile on a 1 - 10 km terrestrial platform, can provide snapshot images having 10^-4 to 10^-5 arc-second resolution. On visible objects as faint as m_v equals 25, blind phasing is achievable with the help of laser guide stars on each telescope. The corresponding science is particularly rich and relevant to current issues in stellar physics and cosmology. Following the completion and test of a prototype 1.5 meter telescope, specifically designed for a 27- element interferometric array, larger component telescopes of 8 to 10 m may become buildable at a sufficiently low cost for affordable arrays. A preliminary design concept is presented. In space, arrays of free-flying telescopes currently studied by NASA and ESA, can in principle provide a better limiting magnitude and longer baselines, reaching perhaps 100 km. The current pace of space projects however makes it likely that large ground-based interferometers will be in use before space equivalents.

Adaptive optics are used to enhance the capability of large telescopes by actively compensating for the distorting effects of the atmosphere. This paper discusses the advances in technology and theory which led to the development of adaptive optics systems on the world's large telescopes.

We present the preliminary results of a feasibility study performed by a team of scientists and engineers from NASA, academia and industrial concerns. The candidate concept is a deployable 8 meter diameter telescope optimized for the near infrared region (2 - 5 microns), but with instruments capable of observing from the visible all the way to 30 microns. The observatory is radiatively cooled to about 30 K and would be launched on an Atlas II-AS to the Lagrange point L2.

We are fortunate to live in a new heroic period of astronomy. In the second half of this century our understanding of the universe has undergone an exciting and profound change. This has been brought about by a number of factors: the development of physics, the discovery of astronomy, the extension of observations to all wavelengths from radio to x rays and finally the development of computers. These new findings and tools have permitted us to elaborate new theories and models of the universe as a whole. In my own mind I see three great themes for the next century of astronomy. The first is the quest for physics. The second is the quest for the origins. The third is what I could call the quest for living space. To pursue these themes we study the universe in the entire electromagnetic spectrum of wavelengths with ever larger telescopes and ever more refined detectors and instruments as we heard at this conference. The new facilities are producing and will produce ever larger quantities of data in such amounts that the information cannot be received, calibrated, analyzed and even understood in traditional ways.

This ia a progress report; we define the requirements and details of the latest design of the Gemini Adaptive Optics System (GAOS). The specifications flow from those astronomy programs needing AO for which Gemini will be best remembered. Based on a critical list of astronomical projects and ignoring those likely to be completed earlier by other ground or space telescopes, the astronomers set requirements for image quality, corrected field of view and sky coverage probability. In median seeing (r_o equals 25 cm) we require the signal to noise ratio with GAOS be double that achieved without it, using only tip/tilt/fast-focus of the secondary mirror. Our design is unique because the specifications are for end-to-end image quality delivered to the detector of an instrument while maximizing sky coverage. The error budget includes both telescope errors and instrument effects, with only about one third of the total for the residual uncorrected atmospheric errors, traditionally the only ones considered in papers on adaptive optics

Herewith the ESO VLT adaptive optics program is introduced. This program includes several activities and projects such as the ADONIS project, COMIC and SHARP II cameras, a data reduction package dedicated to AO observations, the low noise fast readout CCD camera development for wavefront sensor, the quad cell avalanche photo diodes development. One of the main projects is currently the Nasmyth adaptive optics system (NAOS). The high level requirements and conceptual design of NAOS for Unit Telescope 1 of the very large telescope project are described. An introduction to the laser guide star program at ESO is also presented.

The large telescope planned for the Observatorio del Roque de los Muchachos will include an adaptive optics (AO) facility, designed to provide diffraction-limited images in the near infrared. The telescope itself is being designed to allow optimal performance of the AO system. The optical errors of the segmented primary mirror will be specified such that, in closed loop, they do not significantly reduce the Strehl ratio of the final image. In order to achieve high throughput and low emissivity, it is hoped to carry out the adaptive correction by making the telescope secondary mirror adaptive. Turbulence and windspeed profiles measured at La Palma are being used to predict AO performance, as well as to investigate the increase in isoplanatic angle with conjugate height of the deformable mirror, and the focal anisoplanatic errors resulting from the use of laser guide stars.

The recent discoveries of Jupiter-mass planets around nearby stars by measurement of stellar reaction motion (Mayor et al., 1995; Marcy and Butler, 1996) may be viewed as the beginning of a new era for ground-based astronomy. The next step is to obtain direct images of giant planets around nearby stars. In this paper, we show that this goal can be met by using adaptive optics (AO) on the new large telescopes with very smooth primary mirrors. Detailed simulations of an advanced AO system show that a Jupiter twin at 10 pc can be detected at 5 standard deviations above the residual halo noise in a single night of observation. With Gatewood's recent discovery (Gatewood, 1996) of a Jupiter mass planet at 2.5 AU orbiting Lalande 21185, there is now a perfect target for the first application of the new technique. This and other nearby stars will be imaged in a survey planned for the new single-mirror 6.5 m MMT and its twin Magellan telescope in Chile.

Switching-on an adaptive optics loop, the system usually rapidly locks the target in the correction mode. When the sky background contribution is not negligible, the SNR may depend upon the correction status. While in an already closed loop, the SNR could be high enough to maintain the situation of good correction by the adopt system, during the switching-on phase, this condition could be far from true, because it is possible that the SNR isn't enough to lock the loop, starting from the open loop situation. We discuss this problem (called the bootstrap problem for adaptive optics) with particular attention on some filtering system working in the described situation and tuned to solve the problem.

The adaptive optics instrument adaptor for the 3.6 m Canada- France-Hawaii telescope (CFHT) is currently in the commissioning phase. The heart of the system is a 19 electrode bimorph mirror (1:6:12), used with a 19 sub-aperture, curvature wave-front sensor and a separate tip-tilt re-imaging mirror. The performance evaluated in the laboratory and on the sky are presented: the adaptive optics control system provides a 100 Hz servo bandwidth with modal control capabilities. We report astronomical images with median Strehl ratio of 20 (at 1.25 micrometer) to 60% (at 2.2 micrometer), with a FWHM of 0.1 arcsec and a sensitivity allowing image quality improvement with guide stars as faint as m_R equals 17.

This paper presents the electronics, computing hardware, and computing software currently being built to provide real time modal control for a laser guide star adaptive optics system. This approach offers advantages in the control of unobserved modes, the elimination of unwanted modes (e.g. tip and tilt) and automatically handles the case of low resolution lens arrays. In our two step modal implementation, the input vector of gradients is first decomposed into Zernike polynomial modes by performing a least squares. The number of modes is assumed to be less than or equal to the number of actuators. The mode coefficients are then available for collection and analysis or for the application of modal weights. The control loop integrators are at this point in the algorithm. To calculate the DM drive signals, the mode coefficients are converted to the zonal signals via a matrix multiply. At closed loop bandwidths slightly below maximum, it will be possible to do the full two part multiply in real time. Thus the modal weights may be changed quickly without recalculating the full matrix. When the number of gradients measured is less than the number of actuators, the integration in the control loop will be done on the lower resolution grid to avoid growth of unobserved modes. These low resolution data will then be interpolated to yield the DM drive signals.

The first images of astronomical objects have been obtained with a telescope exploiting wavefront compensation with adaptive optics where the reference beacon was generated by laser excitation of mesospheric sodium. This was done using the FASTTRAC II low-order adaptive optics system at the multiple mirror telescope (MMT). FASTTRAC II is a prototype for a full-scale adaptive optics system under construction for the 6.5 m telescope that will replace the MMT in late 1997. The 6.5 m system is designed to provide correction to the diffraction limit of resolution in the near infrared (1 - 5 micrometer) with high Strehl ratio and excellent sky coverage. This paper describes the new system and its expected performance in view of the achieved performance of FASTTRAC II.

The new 6.5 m single mirror multiple mirror telescope (MMT) will be equipped with adaptive optics capabilities to enhance high resolution infrared astronomy. Before we build the 64 cm diameter adaptive secondary, we fabricated a smaller prototype mirror. The adaptive secondary uses voice coil force actuators with an average spacing of 30 mm. Surrounding each actuator is an analog capacitor position sensor operating in a digital closed loop at 10 kHz. This allows the force actuators to be controlled as if they were position actuators. The adaptive secondary configuration and performance test results are presented, followed by the changes to be incorporated into the next curved shell prototype.

The possibility of improving the image quality of the Italian Telescopio Nazionale Galileo (TNG) by means of real-time on- line image selection by using a fast shutter is studied. The tip-tilt signal is checked against a user defined threshold to trigger on rejection of those images which would significantly degrade the long-exposure point spread function because of jitter broadening. The dependency of the achievable improvements on the rejection threshold, on the shutter opening and closing times, and on the atmosphere conditions (r_o) is analyzed. The pros and cons of a ferroelectric liquid crystal shutter are discussed.

The status report of the adaptive optics module for the 3.5 m TNG telescope is briefly given together with the description of three important subsystems: the off-axis tracking capability of the wavefront sensor; the CCD to be used for the wavefront sensor and the modal filtering approach to be implemented in the module.

We have previously proposed the use of commercially available actuators and displacement sensors for use in the control of large deformable secondary mirrors. We have identified the magneto-strictive (MS) actuator as a promising candidate based on manufacturer's specifications for stroke, power consumption, and service life. We have identified both capacitive and eddy-current displacement sensors as possible choices for completing the required control loop around the actuators. The purpose of the tests described here was to characterize the performance of a MS actuator in terms of hysteresis, linearity, power consumption, heat dissipation, and frequency response, and to confirm the manufacturer's specifications for longevity. The purpose of the sensor testing was to compare their performance in terms of frequency response and packaging constraints. Results of the testing program are presented.

The components used in curvature wavefront sensors are described with respect to mounting and adjustment requirements. In particular, the techniques used in the construction of the wavefront sensor employed in the adaptive optics bonnette at the Canada-France-Hawaii telescope are described in detail. Descriptions include methods of mounting components, design and adjustment of optics and of fiber-optic feeds for avalanche photo-diodes, and ways of monitoring wavefront sensor alignment and performance during operation.

The authors describe characteristics and key results of the optical design for adaptive optics at the Galileo telescope. Two off axis parabolae are used to image the pupil on the deformable mirror, and in order to keep to a minimum the number of reflections, one of them is used as tip-tilt mirror. The several constraints, both mechanical and optical, which led to the definitive design are reported (pupil position, effective focal ratio, off axis performances, size of the module . . .). Because tip-tilt is realized through tilting of a non-flat mirror care has been given to the deterioration and distortion of the image for a given range of misalignment of the off axis parabola.

A new concept for a device able to simulate evolving atmospherically distorted wavefronts for the adaptive optics system of the Italian facility Telescopio Nazionale Galileo (TNG) is presented. The goodness of performances of an adaptive optics system depends upon the testing accuracy of its components made under different seeing conditions. Moreover, in order to not loose good seeing nights, the possibility to perform tests during day-light can become strongly important. Introducing a simulator able to generate the image of an astronomical object deformed by atmosphere can help to solve many such problems. In order to create the wanted distortion we use phase changing plate (PCP) screens. A system of moving screens and adjustable zooms provides to change, over a wide range, some fundamental parameters of the atmosphere turbulence like fired parameter r_o, Greenwood frequency f_G and the isoplanatic angle (theta) _o.

In the framework of the adaptive optics for TNG the mechanical, optical and engineering aspects of the laser projection system for the TNG are briefly shown. The laser beam is projected through the elevation axis on the Nasmith foci usually devoted for spectroscopic purposes and otherwise unused during imaging scientific activities. The optical design of the projector is also outlined. It is foreseen that in the very first phase a Rayleigh laser will be operated at TNG, later replaced by a mesospheric sodium LGS generator.

The recovery of the absolute tip-tilt of a laser guide star (LGS) is a crucial one in the laser assisted adaptive optics programs, especially in the visible wavelengths where the natural sky coverage is poor. A number of techniques have been proposed to achieve this goal: the multicolor LGS by Foy, the use of auxiliary telescopes, the use of auxiliary laser projector, the use of the propagation delay, by our group. In this paper two more possible solutions are presented. These techniques are partly speculative or require a strong technological development and are here reported in the hope that the dissemination of new and unconventional ideas on this problem will lead to some efficient solution.

The system overview and the current status of an adaptive optics system for the Cassegrain focus of Subaru 8.2 m telescope under construction atop Mauna Kea is presented. The system is composed of a wavefront curvature sensor with 36 elements photon-counting APD modules and a 36-element bimorph deformable mirror. We aim to get the Strehl ratio of greater than 0.6 at the K band (2.2 micron) using natural guide stars as wavefront reference under the average seeing condition (approximately 0.45 arcsec) at Mauna Kea. It is scheduled to be in operation in 1998. Expected performance, especially the sky coverage when employing natural guide stars are also presented. currently we are testing prototype system with basically identical specifications as those of the final system. We present here the optical system, deformable mirror, wavefront sensor, control system of the final system, and simple introduction and experimental results of the prototype system.

ThermoTrex Corporation has designed and built a prototype of the fast steering mirror to be used for image motion control in the TNG adaptive optics system. The principal characteristic of this mirror is the use of voice coil actuators whose positions are controlled with closed loops based on capacitive sensors. Here we report the main features of the mirror assembly and laboratory measurements done to characterize the mirror behavior. The Bode diagram of the mirror is reported and discussed.

The Keck II telescope control system is a completely new design from the Keck I system. The entire software project, design, implementation and commissioning, has taken 16 months from inception in January 1995 to completion of all major subsystems in April 1996. First light took place in January 1996, and scientific quality images were taken in April 1996. Full operations are planned for October 1996. This paper focuses on the management and software engineering practices and the software tools that made it possible to complete a project of this magnitude with the limited time and resources at our disposal. Areas of potential software sharing and collaboration with other observatories are also discussed.

The experimental physics and industrial control system (EPICS) originated in the high energy physics community and has been used for several years to control accelerators. It is now in use or soon to be in use at several observatories around the world. In 1995, it was decided that Keck II telescope would have a new EPICS-based control system rather than use a copy of the Keck I system. This decision was made because it was felt that EPICS provided a superior software infrastructure to that developed for Keck I, and that it would scale well to encompass adaptive optics and eventual use of the two telescopes for interferometry. The new control system was developed throughout 1995 and the early part of 1996, leading to first light in January 1996, making it the first fully EPICS-controlled telescope control system in the world. This paper describes how EPICS has been used to implement the control system, including a detailed discussion of the axes control, pointing and timing system, and of how they interact with each other.

The servo design and model of the W. M. Keck telescopes autoguider is presented. Telescope servo models often do not include the guider loop and therefore do not take advantage of traditional control analysis and test techniques to improve performance. Guide camera dynamics, computational and transport lags, and compensation networks are discussed. A means of measuring the actual frequency response characteristics of the guide loop is presented and the results are compared to those predicted by the model. Guide performance as a function of integration time is illustrated. An improved compensation network is developed and its performance examined.

The ESO VLT control software consists of all the software, which will be used to directly control the VLT Observatory and associated instrumentation. This is now in the implementation phase, performed to a large extent by ESO staff in the VLT software group. Consortia of Institutes responsible for some ESO instruments and contractors, who implement some of the telescope subsystems, are also involved. The main foundation body of the VLT control software, called VLT common software, is basically complete in its main functions. It has a size of about 500 K lines of code. This software is used in all the developments for the VLT telescopes and instruments and is distributed by ESO to all the collaborating consortia and contractors. The key components of the telescopes control software (TCS) have also been implemented. They make use of the VLT common software and have been field-tested in a first version in December 1995 at the ESO New Technology Telescope (NTT) in Chile. The NTT is being upgraded in parallel to the VLT development, using the same software. At the end of this conference a scheduled period of 6 months of site tests with the VLT main structure is going to start in Milan, Italy. This will allow us to perform hardware specific tests on this software. The Alt/Az axes control and hydraulics bearing subsystems are also part of the tests, which involve a set-up of two workstations and three VME/VxWorks based controllers. In parallel the first enclosure, including its software, is going to be accepted at the VLT site in Chile. This will mark the beginning of the control system implementation at the Paranal site. This paper gives an overview of the VLT control software, and describes its main components and characteristics.

A considerable risk is involved in developing a large complex control system, with a long lead time to integration and commissioning. To reduce this risk for VLT, ESO decided in late 1993, to make use of NTT as a testbed for the VLT control system. By upgrading the system using VLT components, e.g. standardized VME I/O boards, VLT common software, and whenever possible VLT application software, these components are tested in real-life conditions well before the start of VLT commissioning. In order to minimize disturbance to scientific operation, the strategy has been to field test subsystem by subsystem during short test periods, and afterwards restoring the original system. Each of these field tests allows an early verification of VLT control architecture and a possibility to correct direction of development. The project culminates in July 1996, when the telescope will be shut down for scientific operation and the complete control system replaced and recommissioned. The new system will then provide a VLT identical high level interface, allowing continued prototyping of operational procedures, scheduling, and science operations. The paper reports on the status of the NTT upgrade project with emphasis on how VLT has and will gain from this prototyping activity.

The design of the software for the Gemini 8 m telescopes is nearly complete. Great care has been taken to develop a system with the flexibility to support astronomy into the next century without disassociating itself from the current methods of observing. The goal has been to design a system that supports the complexities involved in high performance telescope operation while providing an interface that is easy to operate in all modes from classical observing to modern queue-scheduled approaches. The resulting design has been crafted to augment the skills of observers and system operators and has produced a development strategy intended to encourage the interaction of scientists and developers.

The Gemini telescope control system (TCS) uses an integrated pointing/tracking scheme, based on the 'virtual telescope' (VT) concept, but extended to encompass autoguiding and control of the chopping secondary (M2). The TCS treats each autoguider probe as a separate VT, with a different target from the main telescope (namely a guidestar instead of a science source), different color, flexure and so on. The constraint that both the guider VT and the main-telescope VT share one mount (Az,El) enables the (x,y) of the guidestar image to be predicted. Comparisons between the predicted and measured (x,y) positions enable the guiding system to generate rapid adjustments to M2 tip/tilt, taking out windshake as well as other tracking errors. The TCS monitors M2 tip/tilt and adjusts the mount pointing as required, preventing unwanted buildup of M2 offset. Independent control of M2, to accomplish rapid scans while smoothly tracking the (Az,El) mount for example, is provided by having separate 'mount' and 'source' VTs. In this case, the constraint that there is only one mount (Az, El) enables the M2 tip/tilt to be determined which will place the target image in the required place. Chopping is done by implementing the different chop states as different VTs all running in parallel. The TCS is oblivious to the use being made of the different M2 tip/tilts that the different chop states call for, that being left to the various subsystems. The Gemini plan calls for five VTs (two permanent guide detectors and one on-instrument, plus the mount and target VTs), and all but the mount VT support three chop states. Operating at rate of 20 Hz, this means that the entire pointing calculation has to be performed 260 times per second, and some care has been taken to organize the computations efficiently.

Tracking errors of any kind will degrade the image quality of a telescope. By using a CCD camera to record image motion during blind tracking it is possible to analyze the resulting data and identify the origin of these errors. The part of the image motion which is caused by seeing effects has been analyzed in numerous papers and is not discussed in this report, which will instead concentrate on telescope and servo errors. A strategy for analyzing and identifying these errors is outlined and suggestions for further analysis are given.

The design of the Subaru observation control system is reviewed. The software of this system is so called high-level and controls both telescope and instruments through their controller. The aim of Subaru observation control system is to utilize telescope time most effectively by scheduling observation from remote sites. The design of data acquisition and simple data analysis subsystems are extensively reviewed.

The overall updated plan for constructing 7 scientific instruments and 3 baseline programs for the 8 m Subaru telescope is shown. Somewhat detailed descriptions are given further for projects to develop large format CCDs, faint object camera and spectrograph (FOCAS), high dispersion spectrograph (HDS), Subaru prime focus camera (Suprime-Cam), and Cassegrain adaptive optics system (AO).

The infrared instrumentation plan for the Subaru telescope is described. Four approved infrared instruments and one test observation system are now in the construction phase. They are coronagraph imager using adaptive optics (CIAO), cooled mid- infrared camera and spectrograph (COMICS), infrared camera and spectrograph (IRCS), OH-airglow suppressor spectrograph (OHS) and mid-infrared test observation system (MIRTOS). Their performance goals and construction schedules are summarized. The plan for procurement and evaluation of infrared arrays required by these instruments is briefly described.

Exploiting instrument platforms like the current generation of 8 - 10 m class telescopes represents a new era in instrument design, construction, handling, and use. Gemini's instruments are no exception to this revolution. For example, since at least 50% of Gemini's observing time will be queue scheduled, Cassegrain-mounted instruments will effectively remain on- line, ready to be called into service for typically months at a time with minimal delay to match observing programs with changing conditions. Furthermore, effective instrument emissivities of less than 1% will be needed to take advantage of the very low emissivity of the telescopes. Here we report on the technical status of the phase I instruments, describe attention being given to the total system performance of the telescopes and instruments, and list some of the considerations going into the phase II instrument program.

We briefly review the existing instruments at the first Keck telescope and their performance characteristics. These include the high resolution echelle spectrograph (HIRES), the low resolution imaging spectrograph (LRIS), and the near infrared camera (NIRC). Other Keck 1 instruments include the long wavelength spectrograph (LWS) and long wavelength imaging camera (LWIRC). The instruments currently being developed for the second Keck telescope are described and their expected performance characteristics are described. These include the deep imaging multi-object spectrograph (DEIMOS), the near infrared echelle spectrograph (NIRSPEC), the echelle spectrograph and imager (ESI), the diffraction limited near infrared camera (NIRC-2), and the ultraviolet side of the LRIS (LRIS-B). Keck 2 will also have a major new facility, an adaptive optics (AO) system. This system will deliver diffraction limited images in the 1 - 5 micron region and will be used in front of the NIRC-2. This AO system will contain a laser to generate an artificial sodium star, thus giving AO essentially full sky coverage. The AO system design and status is summarized. Keck Observatory is also planning an interferometer using the Keck 1 and Keck 2 telescopes, with a baseline of 85 m. We describe the plans and progress on this adaptive optics augmented infrared interferometer.

The FUEGOS multi-object area spectrograph will be one of the six instruments of the first generation for the very large telescope. It will be installed on one of the Nasmyth platforms of the telescope unit 3 at the beginning of the year 2001. The instrument includes a multi-object fiber linked spectrograph to observe simultaneously up 80 targets and a fiber anamorphoser for two-dimensional spectroscopy of extended objects. Description of the instrument, its performances and status of the project are reported in this paper.

The two Gemini multiple object spectrographs (GMOS) are being designed and built for use with the Gemini telescopes on Mauna Kea and Cerro Pachon starting in 1999 and 2000 respectively. They have four operating modes: imaging, long slit spectroscopy, aperture plate multiple object spectroscopy and area (or integral field) spectroscopy. The spectrograph uses refracting optics for both the collimator and camera and uses grating dispersion. The image quality delivered to the spectrograph is anticipated to be excellent and the design is driven by the need to retain this acuity over a large wavelength range and the full 5.5 arcminute field of view. The spectrograph optics are required to perform from 0.36 to 1.8 microns although it is likely that the northern and southern versions of GMOS will use coatings optimized for the red and blue respectively. A stringent flexure specification is imposed by the scientific requirement to measure velocities to high precision (1 - 2 km/s). Here we present an overview of the design concentrating on the optical and mechanical aspects.

This paper describes the design of DEIMOS -- a dual beam, off axis, multi object spectrograph of medium resolution, being designed for the Keck II telescope on Mauna Kea in Hawaii. The difficult and advanced scientific goals of the DEIMOS project have generated many challenging design requirements. The DEIMOS team at Lick Observatory has been responding to these challenges with new and unique concepts in instrument design and fabrication.

The ESI (echellette spectrograph and imager) is a multi-mode Cassegrain spectrograph currently funded and under construction at UCO/Lick Observatory for the Keck II telescope. The ESI instrument has three modes. The 170.0-mm collimated beam can be sent directly into the camera for imaging, through a prism disperser, or to an echellette grating with prism cross-dispersion. An all-refracting Epps camera and a single 2 K by 4 K detector are used for all three modes. The direct-imaging mode has a 2.0 multiplied by 8.0- arcmin field of view with 0.15-arcsec pixels. Filters may be placed either near the focal surface of the telescope or in the parallel beam, and the option of a future upgrade including a Fabry-Perot at the pupil image is available. The low-dispersion prism-only mode has a dispersion of 50 to 300 km/sec/pix, depending on wavelength, and this mode can be used with a 8.0-arcmin long slit or in a multi-slit mode with user- made slit-masks. The high-dispersion echellette mode gives the entire spectrum from 0.39 to 1.09 microns with a 20.0-arcsec slit length in a single exposure, with a dispersion of 9.6 to 12.8 km/sec/pix.

The high resolution optical spectrograph (HROS) for Gemini is currently within its conceptual design phase. The science requirements for this instrument demand spectral resolutions of 50,000 and 120,000 with entrance slits of 0.57 and 0.24 arcsec respectively. Amongst the current large telescope projects, HROS will be the only instrument of its class to be mounted at a Cassegrain station and this will pose considerable technical challenges which are described in this paper: HROS will be a spectrograph with unique characteristics, like prismatic cross-dispersion, immersed echelle grating and active compensation of flexure. HROS is expected to perform better than any other high resolution spectrograph with respect to throughput, resolution and simultaneous spectral coverage.

The need for atmospheric dispersion correction on large telescopes is well known. Therefore it was decided to implement atmospheric dispersion correctors for FORS, the focal reducer/spectrographs of the ESO very large telescope. The boundary conditions at the VLT Cassegrain foci excluded however all previously known ADC concepts and therefore we were forced to design a new one, the longitudinal atmospheric dispersion corrector (LADC) consisting of two thin prisms with variable distance. This design has several advantages compared to the 'classical concepts:' among others it avoids tilting the pupil axis and uses only one material (silica) which has a very high transmission over the operating wavelength range of FORS (330 - 1000 nm).

ISAAC is a 1 - 5 micrometer imager/spectrometer currently under construction at ESO and scheduled for installation at one of the Nasmyth foci of the first 8 m unit telescopes of the VLT in 1998. It comprises two cameras, optimized for the 1 - 2.5 micrometer and 2 - 5 micrometer wavelength ranges, which can be used to directly view the telescope focal plane or the intermediate spectrum formed by a grating spectrometer. The complete instrument is cryogenically cooled by means of a continuous flow liquid nitrogen circuit and two closed cycle coolers and is housed in an approximately 1.5 m diameter vacuum tank permanently attached to the Nasmyth adapter/rotator. The observing modes, instrument design, construction status and performance estimated using a recently developed software simulator of the instrument are described.

We are currently building a panoramic wide field near infrared imaging camera based on 4 Rockwell Hawaii HgCdTe 1024² detectors. The survey instrument will operate in the J and H bands and will be as scientifically versatile and as easy to use as a large format CCD camera. It is expected to be ready for astronomical use by late 1997. It will be particularly well-suited for surveys of star-forming regions, low mass stars, distant galaxies, clusters and QSOs. The camera will be commissioned at the prime focus of the 2.5 m Isaac Newton telescope, where the image scale is 0.45'/pixel, giving an effective field of view of 14.6 by 14.6 arc minutes. The field of view of this camera with 0.15' pixels is 5.1 by 5.1 arc minutes and is thus approximately 60 times larger than the current near-infrared imager on Keck (NIRC). When combined with a 4.0 m class telescope, the combination is approximately 10 times as powerful as the Keck 10.0 m, when the apertures are taken into account. The options for upgrading the camera into a wide field spectroscopic survey instrument are currently being investigated.

For the next generation of instruments which will be used at 8 m telescopes large format arrays are needed. Better image quality obtained by adaptive optics requires sampling to higher spatial frequencies. The large field of these instruments increases the demand for array formats as large as 1024 by 1024 and beyond. For this reason ESO is committed to the development of megapixel infrared detectors. In a multimode instrument covering the 1 to 5 micrometer spectral range a detector has to fulfill very different requirements. For high resolution spectroscopy low dark current and read noise are required. For broad band thermal imaging a high well capacity is needed to reduce the speed required to read out the array before it saturates. This paper gives a status report of ESO's activities related to large format arrays. An ultrafast data acquisition system has been developed to read out large format arrays. The performance goal is to achieve shot noise limited operation in the wavelength region of lambda equals 1 to 5 micrometer. The array controller is capable of handling the high data rates generated in the thermal infrared. The design of the controller was mainly driven by the requirement to read out the 32 parallel video channels of the SBRC 1024 by 1024 InSb detector in 50 msec. The array controller can also cope with the low read noise required for flux levels of less than 1 photon/sec. A new test camera for large format arrays has also been built. First test results obtained with the Rockwell 1024 by 1024 HgCdTe array are presented. The noise and dark current performance will be discussed with regard to OH line suppression. Read speed requirements will be defined for advanced readout techniques of image sharpening applying on chip tracking in the multiple nondestructive readout mode.

Reliable calibration of astronomical data from large telescopes is an essential factor in obtaining high sensitivity observations for both the astronomer at the telescope and the archive researcher. A dedicated calibration unit provides an efficient and predictable method of observing calibration frames. Such a facility is being designed for the Gemini telescopes. It is required to calibrate instruments with wavelengths from the UV to the infrared, covering a broad range of both spectral and spatial resolution. We present the design of this calibration unit and the predicted performance with the 'first-light' Gemini instruments.

MPE has developed 3D, a new type of a highly sensitive near- infrared integral field spectrometer. It has been designed to multiplex spectral as well as spatial information thus obtaining a full data cube in a single integration. At a spectral resolution between 1000 and 2000 and a field of view of 16 by 16 pixels, optimized for subarcsecond spatial resolution imaging spectroscopy, it has a much higher efficiency compared to conventional techniques. Outfitting one of the VLTs with a near-IR 3D type instrument will provide a powerful tool for diffraction-limited integral field spectroscopic research, in particular on faint high-z galaxies in the early universe. The basic design, recent upgrades as well as plans for a possible VLT-3D instrument are presented.

In addition to linking telescopes with grating spectrographs, optical fibers are now commonly used to smooth the photometric barycenter of the slit thus making possible the measurements of very small amplitude Doppler-Fizeau wavelength shifts of the spectral lines. However, optical fibers still transmit a fractional amount of guiding and seeing errors. To better the precision on the measurement of the barycenter of the spectral lines it has been suggested that using microlenses at both ends of the fiber link could further increase the level of scrambling of the input image on the fiber. Optical fibers coupled with microlenses take advantage of the large plate scale factor present on slow foci of large aperture telescope as for example, the multiobject high resolution spectrograph FUEGOS for the European very large telescope. This experiment compares the coupling of a conventional 'bare' optical fiber link with a fiber cable with microlenses at both ends. Light is coupled into both fibers in the optimum condition for focal ratio degradation reduction and to avoid fiber dependence, the same fiber has been used for both links. The set-up used for this feature is described. A description is given of the test bench set up to measure the residual photometric barycenter shifts through both links and of the experimental condition to achieve subpixel resolution down to a thousandth of a pixel. The measurements are done both on the slit and on a spectral line through a spectrograph. The spectral stability obtained with those two fiber links are discussed and have been finally compared with an optical fiber double scrambler.

Michelle is an all-reflective long-slit spectrometer and imager which is being built at the Royal Observatory Edinburgh for operation on both the UKIRT 4 meter and the Gemini 8 meter telescopes. The project is now well advanced with major optical and mechanical components in manufacture. In addition to summarizing the optical specification and opto-mechanical design of the spectrometer, we present the results of optical testing of the spectrometer collimator.

We describe the main features of the optical and mechanical design, and the architecture of the control system of SARG, the white pupil cross dispersed echelle spectrograph for the Italian Telescopio Nazionale Galileo (TNG) telescope. SARG is designed for the spectral range lambda equals 0.37 up to 0.9 micrometer, and for resolution from R equals 19,000 up to R equals 144,000. SARG uses an R4 echelle grating in quasi- Littrow mode; the beam size is 100 mm giving an RS product of RS equals 46,000 at order center. Cross-dispersion is provided by means of a selection of four grisms. A dioptric camera (F/5.05, R equals 144,000) images the cross dispersed spectra on a mosaic of two 2048 by 4096 EEV CCDs (pixel size: 13.5 micrometer). Expected peak efficiency is 0.17 at R equals 38,000, and greater than 0.10 over the whole range from lambda equals 0.4 to 0.9 micrometer. Confocal image slicers, modification of the Bowen-Walraven type designed by Diego, are foreseen for observations at R equals 76,000 (3 slices) and 144,000 (5 slices), allowing high efficiency even in fair seeing conditions. Minimum interorder separation is 8 arcsec. Further features of SARG include an absorbing cell for accurate radial velocities and a Lyot mask (located on an image of the entrance pupil before the slit) for spectrocoronographic observations. SARG is thermally controlled, in order to avoid deterioration of the optical performances.

GMOS is the general-purpose optical spectrograph for the Gemini telescope. It is to be mounted at the Cassegrain focus and has been designed from the outset to have open-loop active correction for the effects of instrumental flexure. This is done by moving the CCD detector mosaic in two axes to counter the image drift due to flexure. GMOS is also focused by moving the detector and the three axis mechanism to achieve these motions is described here. The mechanism also provides the necessary thermal isolation for the cooled CCD and is driven by stepping motors outside the CCD vacuum vessel.

The hardware construction of FORS1, the first of the two focal reducer/low dispersion spectrographs of the ESO very large telescope (VLT), is now finished. An extensive testing program is under way which will guarantee that the instrument is fully understood and well calibrated when it will be installed at the Cassegrain focus of the first unit telescope of the VLT in 1998. This program includes a full characterization of the optical system and the evaluation of the setting accuracies and reproducibility of the numerous electromechanical functions as well as testing the flexure compensation which will minimize image shift during telescope motions. Telescope and star simulators were specially built for this purpose in order to test the optical and mechanical behavior of the instrument on the 8 m-telescope. Acceptance tests of the optical performance and the subsystem tests of all electromechanical functions indicate an excellent quality, especially for the complex multi object spectroscopy unit, while the overall system tests are just starting.

For telescopes without internal calibration systems, screen flats are invaluable for calibration (Gililland, 1992; Gullixson, 1992; Massey and Jacoby, 1992; Tobin, 1993). An understanding of the surface optical properties of dome flat screens and their illumination systems is important in obtaining high quality calibration data over the widest possible spectral range. However, most current screens have a number of limitations. The range of their spectral reflectivity is limited and their bidirectional scatter properties are not Lambertian. The illuminating systems of calibration screens are not optimized as well. This paper explores some aspects of the optical properties of screens and their illumination systems.

The Astrophysical Institute Potsdam (AIP) has decided to develop an integral field spectrograph/spectrophotometer making the best possible use of 8 - 10 m telescopes with good image quality. The scientific motivation and driving requirements are outlined, along with a predesign study for the fiber-coupled spectrograph as part of the instrument. The fully-dioptric, high throughput fiber spectrograph is a self- contained module which may be of interest for other applications as well, e.g. multi-object fiber spectrographs.

The RUBIKON is a photon-counting multi-element detector system of the Digicon type, using a Spectracon with an S20- photocathode and a linear silicon diode array of 512 elements of 38 by 500 micrometer each, plus two 200 by 200 micrometer diodes at each end for adjusting, each diode having its own charge-coupling amplifier. The fast, low-noise electronic read-out system includes computer control of the spectrograph and data pre-reduction. Spectra obtained with the RUBIKON attached to the 61 cm-Bochum telescope demonstrate the high sensitivity and the large dynamic range.

The Michelle mid-infrared spectrometer and imager is being built for the UKIRT and Gemini telescopes on Mauna Kea. In common with many other instruments being built for the new generation of telescopes, the focal plane array detector used requires a large number of electrical connections. The Michelle instrument integrates the wiring in to the design of the Joule-Thompson cooler. Three thin-walled stainless steel tubes form part of the structure of the cooler. The output, bias and clock wiring pass down separate tubes. The wiring is made up of custom woven cables optimized for low capacitance (outputs) or controlled impedance (clocks). The wiring assemblies can be removed from the tubes without straining the wires in any way. The stainless-steel tubes also form part of a high integrity screen totally enclosing the array drive and data acquisition system. The only breaks in the screen are to allow light on to the array, power to the electronics, and fiber optic connections to the controller computing system.

The fabrication requirements of the Gemini multi-object spectrograph (GMOS) slit mask is discussed particularly in terms of the slit-to-slit position, slit geometry and the telescope operation. The demand for precision slit masks with high quality slits of width of less than quarter arcsecond and an allowable fabrication time of two hours required examination of innovative fabrication processes and mask materials. Different fabrication processes including high precision cutting processes, water-jet and laser machining systems are evaluated according to cost, speed and efficiency, and the findings are documented. Different candidate mask materials including low thermal expansion metals and novel materials such as graphite paper and carbon-fiber composite sheet, are evaluated according to their relevant mechanical and physical properties, and the findings are also documented. In addition to identifying that the most suitable mask material is unidirectional carbon fiber sheet and the corresponding fabrication process is a Nd:YAG laser machining system, the mask handling system for GMOS is described and methodology to minimize systematic fabrication errors is also proposed.

Some components of the photometric equipment, designed and made to use in high speed multichannel stellar photometry for the whole earth telescope (WET) and other astronomical programs, are described. Despite this special purpose, the described equipment could also be used for any kind of stellar photometry, especially for observations of variable stars. Small size and weight of the equipment make it suitable for observations in expeditionary conditions. An original solution of the step-motor feeding for the filter wheel rotation is presented, and the filter setting system FD 306 based on it is described and discussed.

The most innovative feature of the Gemini multiobject spectrographs (GMOS) is the capability for integral field spectroscopy. This will allow the Gemini telescopes to obtain spectra over a contiguous rectangular field of area 50 arcsec² with a sampling of 0.2 arcsec. The field will be reformatted into two long slits so that each element in the field is dispersed into a long spectrum containing up to 900 resolution elements at spectral resolutions up to 10,000. Background subtraction will be carried out via a separate field with identical optical characteristics. This will support a number of background-subtraction techniques including beam-switching. The integral field unit will be loaded into the focal plane in the same way as a slit mask to allow a rapid changeover between integral field and aperture spectroscopy. The design employs a combination of optical fibers and microlens arrays with enlarging fore-optics. The fibers give the desired reformatting ability to maximize the length of the spectrum while the microlenses provide both contiguous field coverage and optimal matching to the slow telescope and spectrograph optics. The integral field capability may be augmented and upgraded by adding different units. Of particular interest are options for finer spatial sampling (0.1 arcsec) and for operation in the near-infrared.

I propose a new design of integral field unit (IFU) for two- dimensional area spectroscopy. The unit acts as a coupler between the telescope and spectrograph to reformat a square or rectangular field into a long slit. In addition to providing good transmission, it avoids the geometrical losses inherent in previous designs based on fiber optics, lens array re- imaging and image slicing. The proposed design is a new type of image slicer in which the original two-dimensional image is sliced into narrow sub-images that are re-imaged side by side to form a long one-dimensional image at the spectrograph input. The new design is much more compact than previous designs making it easier to insert in the front of a spectrograph without any modification to the spectrograph support system. The design uses much smaller optics than previous designs. The small number of reflections -- 4 to 6 depending on the telescope focal ratio -- and the smaller instrument size, which simplify the cooling of the instrument, makes the design well suited for infrared spectroscopy. The application of this design to 8-m telescopes and its use with adaptive optics in the optical and infrared is discussed, particularly with respect to the multi-object spectrographs of the Gemini telescopes project. In this case, the proposed design gives a field of 8.3' by 11.4' along with a background field of 2.6' by 3.6' with a spatial resolution of 0.09' by 0.16' and roughly 1200 spectral resolution elements in each spectrum. The spatial resolution is well suited to images produced by the Gemini telescopes in tip/tilt mode (typically FWHM approximately 0.3' in 10th percentile seeing) and with higher order adaptive optics. This design is also applicable to 4-m telescopes with -- or without-- adaptive optics. The possibility of inserting such an instrument in front of existing spectrographs is discussed.

Gravity-induced flexure has been a long-standing challenge in Cassegrain spectrographs at 4-meter class telescopes; it is the more so at the scale of 8-meter telescopes. This is of particular concern for the Gemini high resolution optical spectrograph, which will be Cassegrain-mounted for its routine mode of operation. In this paper we address the general flexure problem, and how to solve it with the use of active optics. We also present the results of an experimental active flexure compensation system for the ISIS (intermediate- dispersion spectroscopic and imaging system) spectrograph on the 4.2 m William Herschel Telescope (WHT). This instrument, called ISAAC (ISIS spectrograph automatic active collimator), is based on the concept of active correction, where spectrum drifts, due to the spectrograph flexing under the effect of gravity, are compensated by the movement of an active optical element (in this case a fine steering tip-tilt collimator mirror). The experiment showed that active compensation can reduce flexure down to less than 3 micrometer over four hours of telescope motions, dramatically improving the spectrograph performance. The results of the experiment are used to discuss a flexure compensation system for the high resolution optical spectrograph (HROS) for the 8 m Gemini telescope.

Autofib-2 is a robotic fiber system for the prime focus of the William Herschel telescope capable of placing up to 150 fibers in the 1 degree focal plane of the telescope. The fibers are fed to a purpose built spectrograph (WYFFOS) mounted on one of the Nasmyth platforms. Autofib-2 and WYFFOS are now entering a common user phase as fully commissioned instruments. We describe the novel techniques used to achieve the high precision in fiber placement delivered by this instrument and the quality control procedures devised to measure and monitor instrument stability. The characterization of the distortions of focal plane delivered by the prime focus corrector of the telescope was a vital procedure during the commissioning. We describe the methods of measuring these distortions and discuss the limitations of the instrument, telescope and astrometry.

As telescope apertures increase, fundamental limits for doing efficient slit-based spectroscopy are being reached. We propose that lens arrays feeding fiber bundles be used in the focal plane of large telescopes to counteract these problems and in the process add the capability of integral field spectroscopy. A description of the first phase of the SPIRAL (segmented pupil/image reformatting array lenses) project is presented. This system will provide spatially mapped spectroscopy with high spectral resolution and high throughput. An alternative mode in which the lens array is used to segment the pupil rather than the sky is also described. Finally, we briefly discuss our future instrument development plans.

One of the main limitations to the study of faint objects in the near-infrared (1 to 2 micrometer) is the luminous, varying sky background from very narrow OH emission lines originating in the Earth's upper atmosphere. This source of background contributes 95% to 98% of the total sky counts in the J & H atmospheric windows. We present the optical layout of the Cambridge OH suppression instrument. COHSI is designed to deliver OH suppressed, R equals 500, spectroscopy for both J & H spectral bands simultaneously providing an integral field mode and a multi-object mode. COHSI also has an OH suppression imaging mode. A modular approach has been selected for COHSI with the instrument consisting of three components. The first section consists of simple re-imaging lenses and a lens array interfacing the telescope to a set of optical fibers. This decouples the design of COHSI's main components from the telescope allowing COHSI to be easily used with different telescopes and making it free from flexure problems. The second section of COHSI is the OH suppression 'filter' itself. The size of this section is significantly smaller than in other similarly planned instruments. The third and final module of COHSI is the cryogenic low-resolution imaging spectrograph.

For spectroscopic observations with a segmented mirror 28- meter telescope, an alternative to image slicing of a single composite image is to separate the 141 images and stack the unsliced images from the segments along the slit of the spectrograph where an array of micro optics redirects the beams from the segments to superimpose on the grating of a spectrograph. The focal length of the collimator can then be increased in proportion to the square root of the number of beams, resulting in a proportional increase in the slit width without loss of spectral resolution. An additional advantage is that the longer focus collimator introduces less aberration and can be used off-axis thus avoiding central obstruction near the grating. Following the collimator, an aspherized grating, developed by Gerard Lemaitre and manufactured by Jobin-Yvon, has the advantage of minimizing the number of optical surfaces in a spectrograph.

This is an initial survey putting together some facts and ideas relating to a possible generation of new instruments which could dominate optical astronomy. These instruments will be diffraction-limited spectrographs used on diffraction- limited telescopes. They will enable new research, given some new technical development. The slit area is fundamental. Two cases are distinguished -- a full Airy disk within a larger entrance aperture, and an Airy disk which is truncated by a slit. To demonstrate the feasibility of some relevant optical designs, designs are given for two large diffraction-limited spectrograph cameras. Their good imaging properties could enhance exiting spectrographs. Low resolution instruments may use optics of only a few millimeters aperture, leading to the suggestion of using a number of small spectrographs to map the isoplanatic patch at the full spatial resolution of a telescope.

A new application for the technique of holographic Fourier transform spectroscopy [Stroke and Funkhauser, Phys. Lett. 16, 272 (1965)] is presented. We propose placing a Michelson interferometer in the focal plane of a zenith-looking liquid mirror telescope [Borra et al., this conference]. A wedge whose apex runs perpendicular to the star drift direction is introduced between the two arms creating a variation in the star's apparent intensity due to interference as it drifts across the focal plane. The record of this intensity variation is the interferogram of the star, i.e. the Fourier transform of its spectrum. Imaging the mirrors of the Michelson onto a CCD array thus allows simultaneous spectral measurements of numerous sources. We present calculations indicating that a limiting magnitude of 18.6 may be reached with a spectral sampling of 4 A covering from 5000 A to 9000 A by summing 100 nights of observations.

The scientific objectives, design concept, and expected performance of a coronagraphic spectrometer that is planned for installation as a spectrographic mode of the coronagraphic imager with adaptive optics (CIAO) for the 8 m Subaru telescope are described.

It is suggested that multi-mode slab waveguide spectrographs, developed for wavelength division multiplexing (WDM) in optical communications, might have a role in astronomical instrumentation. A stack of such waveguide devices could form an effective instrument for multi-object spectroscopy. The performance of some existing devices is described, and their applicability to astronomy investigated. It is found to be limited. However, it is shown that only a modest improvement in resolution (by a factor of 2 - 4), together with a realization of the potentially very high optical efficiency, could yield an astronomically-useful device. The problems anticipated in developing such an instrument are outlined.

The main feature of large ten-meter telescope is usage of 84 primary mirror segments which are united in one reflecting system. This decision makes easier the problem of the primary mirror manufacturing but brings another task -- drives designing for every mirror segment moving with high precision (nanometer accuracy) which provides perfect coincidence of mirror segments constantly. There are suggested two variants of the drive construction based on magnetic-rheology liquid. This drive provides nanometer precision and millisecond quickaction.