One of the major fields of application of speckle-interferometric and speckle-imaging techniques is the photometry of astronomical objects exhibiting structure smaller than the seeing limit. The accuracy of the photometry depends critically on the accuracy to which the modulation transfer function (MTF), that describes the atmospheric-telescopic attenuation of the Fourier amplitudes of the object under consideration, is known. The estimation of the effective MTF is especially difficult when no known reference object is within the field of view.

A method is presented that allows to estimate the effective MTF from the observation of arbitrary structure with the use of FRIED-KORFF theory. The ratio of the squared modulus of the average Fourier transform and the average power spectrum serves as an estimator for the FRIED parameter r0. To a first approximation, this ratio is independent from the observed object. Additionally, the behaviour of the ratio in regions beyond the seeing limit in the Fourier plane may be analyzed to obtain an estimate of the speckle interferometry signal-to-noise ratio. The basic concept of the ratio method will be described, its accuracy will be discussed. First results on the application of the ratio method to observations of solar granulation will be presented.

Adaptive optics is a technique for correcting atmospheric wavefront disturbances to yield diffraction limited imaging. It is a technique whose advantages are most apparent in the 2-5µ spectral region, where wavefront corrections are derived from study of visible objects. Graphs are presented to show performance needs of adaptive optics systems.

The astronomical seeing at the CFHT is excellent. It now averages <FWHM> = 0.9" and is continuously improving as a result of seeing conservation measures in the telescope building. Our studies suggest that the average natural seeing on Mauna Kea should be ~ 0.6". Very high (<0.2") resolution imaging at the CFHT, through adaptive optics, appears feasible.

1. Introduction. A popular - often textbook supported- view is that the diffraction limit of large telescopes is out of reach from ground-based observatories.

We summarize on Fig.1 the scope of angular resolving power, in the range lOOnm to 10. nm, as it appears today.

A quick look reveals, expressed in absolute angular terms, a number of stimulating points :

(i) recent advances in speckle techniques allow to reach the diffraction limit of large ground-based telescopes, notwithstanding the seeing effects. Initially devised to measure double bright stars separation in the visible, these techniques have been extended to weak visible stars (mv ~ 14 to 16), to infrared sources (l-5µm), to full image reconstruction, and even to diffraction limited images obtained with optics suffering from residual aberrations (Faint Object Camera on the Space Telescope);

Synthetic aperture telescopes are more and more considered as a promising scheme for the next generation of Very Large Telescopes : the basic idea is to combine the beams of two or more independent telescopes. The question immediately raising when considering such a technique is the pathlength stability and the amount of pathlength difference that can be tolerated, assuming a reasonable bandwidth, to preserve coherent addition of the wavefronts and hence, high spatial frequencies observations.

To that respect the I2T (two telescopes interferometer) in operation at CERGA constitutes a remarkable test-bench for studies of mechanical, optical and atmospheric constraints that are to be faced when dealing with amplitude interferometry between several telescopes.

The coude beams from a pair of 1.5 meter telescopes are recombined to achieve coherent aperture synthesis. The spherical telescopes move along a 70-meter baseline, expandable to 300 meters. For the VLT, a grid of railway tracks with 2-D carriages meets the requirements for incoherent and coherent observing.

Steps in the design and construction of a coherent telescope array have been described previously (Labeyrie, 1982, 1984). In 1975, it was first verified that stellar light from two telescopes could be recombined coherently. In 1978, it was decided to develop spherical mounts for 1.5-meter telescopes. Among the expected features were: 1. coude focus with only one flat mirror; 2. wheels for mobility along railway tracks; 3. structural stiffness for interferometric stability, also implying vibration amplitudes inferior to 0.1 micrometer at each reflective surface. This last and most critical requirement implies materials which have to be stiff and also damped.

We describe recent progress with interferometric imaging using the Multiple Mirror Telescope (MMT). All six telescopes can now be phased over a wide field of view simultaneously resulting in a (u, v) plane coverage corresponding to that of a 686 cm aperture telescope. We describe the open-loop phasing control of the MMT for gravitational changes and we describe a concept of an internal cophasing/coalignment system for MMT type telescopes.

The expected progress of diffraction limited imaging methods and the apparition of new super resolution techniques like differential speckle interferometry would justify the construction of a 15 m class telescope dedicated to diffraction limited observations in order to fulfil the potential of high angular resolution astrophysics of 15 m class instruments, but the construction of such a telescope is conceivable only if its cost is much smaller than the cost of the equivalent all purposes VLT. In this paper we suggest that a telescope with a long and thin rectangular primary ( 16 m X.4m say ) , able to rotate around the optical axis to ensure a full coverage of the frequency plane, would do almost as well than a conventional 16 m aperture telescope for high angular resolution astronomy for a cost substancially reduced. The performances of such a Large Slit Aperture Telescope ( LSAT ) for classical and differential speckle interferometry are examined and the releases on the optical and mechanical constraints allowed by the dedication of the instrument to speckle techniques are discussed.

I-INTRODUCTION The speckle interferometry method has been introduced in 1970 by A. Labeyrie who showed in the visible dcmain that it was possible to reach the limiting spatial resolution of large telescopes. From 1977, the method is extended to the near infrared between 2 and 5ym (Lena, 1977; Wade and Selby, 1978) ,7 years later infrared speckle systems are operating on several large telescopes (ESO, KPNO, AAT, UKIRT, CFHT... ).

The speckle interferometry method has allowed to better understanding the optical properties of the atmosphere, especially owing to the work of F.Roddier (Roddier, 1981). It has raised new problems like the phase restitution of the object spectrum and has largely contributed to the development of image reconstruction methods.

High-resolution speckle imaging is one of the most fascinating possibilities of a Very Large Telescope (VLT). Various speckle methods can yield a resolution of about 0.01" with a 8-m to 10-m VLT. As exiting as the resolution is the limiting magnitude of the speckle methods. The limiting magnitude is extremely seeingdependent. The following limiting magnitudes are possible with speckle interferometry, speckle masking and speckle spectroscopy:

4" seeing: limiting magnitude 15

2" seeing: limiting magnitude 17.5

1" seeing: limiting magnitude 20 (!)

Two eight-metre mirrors on an alt-azimuth mount are considered as a Future Large Telescope. The combined collecting area is ten times that of a four-metre class telescope, and the main modes of use are for optical spectroscopy and infrared imaging. In one square arc second, at 10µm wavelength, the telescope with adaptive auxiliary optics will yield 30 to 100 fully resolved image elements (not speckles). The mechanical arrangement proposed to give path length stability is to link the two tube centre sections along a horizontal axis and to use a total of four altitude bearings. A combined focus is necessary for the highest angular resolution, but other work such as spectroscopy can be carried out either at a combined focus or using separate instrumentation for the two eight-metre apertures. Twin prime and Cassegrain foci and horizontal Nasmyth platforms are intended for such pairs of spectrographs and other instruments, depending on their weight and field requirements. A pair of prime focus spectrographs with f/0.8 cameras is one appropriate system. (The horizontal Nasmyth platforms to accommodate larger spectrographs are a feature of this telescope which is not feasible in other MMT-type configurations). Another significant capability is that for a coud6-type light path from either aperture to the vertical axis. This can be achieved with four reflections (including the primary) for some applications. In addition to the conventional uses of the coudg light path, there is the possibility of using it to produce interferometric links to other telescopes. These links may be important as a future extension of the instrument.

The performance of an optical telescope can be improved considerably by opening the shutter only at moments of good seeing. The resulting instantaneous highresolution images must be corrected for shift before adding them all together. The optimum aperture diameter for which this technique works well is 1 meter or less, depending on atmospheric conditions. Image stabilisation not only improves the resolution and thus the point-source sensitivity, but it is also expected to improve the performance of speckle interferometry and optical aperture synthesis. This makes its implementation on large telescopes desirable. While it is certainly possible to do this by treating a large filled aperture like a set of independent subapertures, a much more logical approach would be to build an array of 1-meter telescopes, preferably mounted in a single large frame so that it can be pointed in all directions without extensive pathlength compensation. Such a “Many Mirror Telescope” has many advantages, and is feasible because image stabilisation helps to solve the beam-combining problem.

Power law relationships between cost and aperture of optical telescopes are shown to be approximations to polynomial expressions. These polynomials, which are derived for telescopes of traditional and cost-reduced design, have implications for the cost-effective design of an optical array telescope.

A three-mirror telescope is described which covers a field 4° in diameter at f/1.6 with images smaller than 0.31 arcsec total spread (0.17 arcsec r.m.s.); an alternative version, also f/1.6, covers a field 2° in diameter with images smaller than 0.049 arcsec (0.023 arcsec r.m.s.). Both forms of this telescope are fully baffled against sky fog contamination, the illumination is uniform over the central 2° field and decreases by about 13 per cent at the edge of the 4° field, and the tube length is little more than the focal length.

With the aim to propose alternative solutions for Multi Mirror Telescopes, we present some ideas about a MMT for High Resolution Spectroscopy purposes. The alternative solution is intended both in technical conception and in economical costs.

This possible project consider the use of a system of 36 mirrors about 1.5 meter in diameter, collecting light independently each other and transmitting the light down to the slit spectrograph by optical fibers.

To reduce mirror costs and structural weight, we propose to develop the “replique” technique over light-weighted supports modelled and adjusted over a master convex mirror.

Different project design for a modular hive structure are considered.

The Anglo-Australian Telescope (AAT), though based on the design for the Kitt Peak and Cerro Tololo 4 metre telescopes, differs in many significant respects. E.g. its declination axis is offset from the polar axis and its polar axis structure is much stiffer. Also the computer was, from the outset, more closely integrated with the telescope drive and control than was previous practice. The sound structural design, good gearing, and accurate mirror supports, together with thorough computer calibration, allowed the AAT to set new standards in pointing by the time it was fully operational in 1975. Good pointing saves time in acquisition, greatly aids day time observing, and leads to good blind tracking. Currently, calibrations covering the whole observable sky give rms radial residuals very near 2 arc sec. Blind tracking errors are generally within ± 1 arc sec for 1 hour or more and natural frequency oscillations less than 0.1 arc sec peak to peak. Better tracking is, of course, available using the offset autoguiders, which employ image dissectors and function on stars to about 15m. Intensified SEC Videcon television cameras are used for acquisition and for guiding with light reflected from spectrograph slit jaws.

Initially, the AAT seeing suffered from poor matching of inside and outside air temperatures because the intended active cooling of dome air was cut from the installation to save funds, leaving an inadequate capacity for forced ventilation. Correlation of observed seeing with dome temperature excess indicated degradation of about 1/2 arc sec/°C. In 1980 the forced ventilation was upgraded to change the dome air in less than 4 minutes, with optional flow direction, upwards (exhausting out the observing aperture) or downwards. Measurements of seeing within the dome, but with the dome open, now show degradations typically 0.4 and 0.6 arc sec diameter with upwards and downwards ventilation respectively. Stellar seeing measurements, including routine observations in which the ventilation direction was alternated nightly, have also indicated superiority of the upwards flow. Taking advantage of this, filters can now be installed on the air supply to keep the optics cleaner.

A report is given on the site-testing programme for a Very Large Telescope in Chile. Data are now available from seven months of investigations. Comments are made on the places which have so far seemed most attractive. Meteorological data from constant monitoring of a coastal mountain, Paranal, are presented and discussed. These data indicate an average percentage of photometric nights of around 85. Atmospheric integrated water-vapour content seems very low in the winter but relatively high in the summer. Other meteorological parameters seem favourable for astronomical observations.

I have been asked to remark on the instrumentation which, at least from the viewpoint of the 1980's, seems most important for early implementation on a VLT in the 1990's. To avoid simply making a long list of many possibilities, each of which has some advantage over its competitors, I will divide the discussion into three parts: a summary of the principal observational advances made possible by a VLT in the 15-20 meter class; a list of what is arguably the resulting basic set of needed instruments; and a brief consideration of the telescope design features, technology development programs, and political/financial efforts required to make those instruments a reality.

Preliminary feasibility studies have been carried out in the framework of the ESO VLT project which have permitted the evaluation of several possible approaches to the instrument matching problem. This paper briefly discusses the detector long term perspective, recalls the main scaling laws and attempt to provide some quantitative analysis of spatial performance of instrumentation in a few typical cases. Field slicing technique is proposed as a potential solution to solve detector matching problems.

In this poster paper we shortly present recent developpment and preliminary results obtained with the focal reducer of the Canada French Hawai 3.6 meter telescope with new trends for deep imagery and multiaperture spectroscopy.