The Large Millimeter Telescope LMT

With the construction of the HHT we had reached a new record in the size of a radio telescope, if one takes the shortest operational wavelength as the unit of length. This is illustrated in Fig. 7.1, where the "precision" of the reflector is plotted against diameter. Taking the minimum wavelength as 16 e, we see that the most accurate telescopes reach a resolution of 5 - 10 arcseconds at the shortest wavelength. In view of the atmospheric absorption at frequencies above 1 THz it seems barely worthwhile to push for increased surface accuracy to 10 mm or less. Even at the best terrestrial sites, now under development, like Chajnantor at 5000 - 5500 m altitude in northern Chile or the South pole, the windows between 1 and 2 THz will "open up" only on rare occasions. But on a good site with long-term adequate transmissivity up to 400 GHz, one might consider operating a telescope with a surface accuracy comparable to that of the MRT, but with a significantly larger diameter.

This idea was developed during 1994 in discussions between astronomers from the "Instituto nacional de Astrofísica, Óptica y Electrónica" (INAOE) in Tonantzintla, Mexico and the University of Massachusetts (UMass) in Amherst, MA, USA. Both were looking towards expansion of their activities. UMass had operated for almost 20 years a 14 m millimeter telescope nearby on a mediocre site and there appeared room for a new large astronomy project in Mexico. The two institutes proposed to jointly build and operate a large mm-telescope on a very high mountain in Mexico.

This resulted in the LMT/GTM-Project (Large Millimeter Telescope / Gran Telescopio Milimétrico) with a goal to design and construct a 50 m diameter Cassegrain antenna on the 4600 m high Cerro la Negra in central Mexico. The scientific desire was to observe to a wavelength of 1 mm, perhaps 0.8 mm, and hence the primary specification was set to 75 mm reflector surface rms error, with a goal of 70 mm, and a pointing accuracy and stability of 1 arcsecond, goal 0.6 arcsecond. These are formidable specifications, which have barely been met with the MRT, which with its diameter of 30 m has just over one third of the reflector area of the LMT.

The first conceptual design for the telescope placed the antenna inside a fixed radome with a skin transparent to millimeter waves. This was a natural way (the 14 m telescope was also enclosed in a radome) to avoid the influences of wind and the extreme weather conditions directly on the antenna. Thus the antenna could be light-weight, because survival conditions are essentially moved to the radome. But there are serious issues with other aspects of the operation. As we have noted before, the thermal equilibrium of the telescope structure is of the utmost importance to reach the specifications. This is difficult to realise in a closed space with transparent walls. Strong stratification of the air will occur, leading to large temperature gradients over the height of the structure, unless the air is mixed by large blowers inside the radome. This would be tantamount to placing the antenna constantly in a rather strong wind. Then there is the need for the radome to survive the extreme conditions of storm and possibly snow and ice. This amounts to putting the material saved for the antenna into the support structure of the radome. Finally, and most severely, any currently available radome fabric with sufficient strength causes significant loss to the received radiation at short mm-wavelengths. For a discussion of these aspects, see Baars (1983).

Eventually, it was decided to consider also an "exposed" design. After all, the MRT in Spain had been operating successfully for about 10 years in extreme conditions similar to those on Cerro la Negra. Thus in September 1997 a request for quotation was issued, which gave bidders the option of a radome enclosed or exposed solution. Three offers were received with highly different characteristics. One was a relatively lightweight antenna in a radome of 60 m diameter, the others were exposed telescopes. A highly optimised homologous structure with a BUS completely realised in CFRP was posed against a homologous steel structure with an "active" reflector surface consisting of 180 sectors each controlled in position by motorised adjusters. The choice went to the last design, the main aspects of which we summarise now.

The LMT design originates with MAN Technologie and shows strong similarities with the 30 m MRT. A cross-section of the telescope is shown in Fig. 7.16. There are however a number of differences and original design features which deserve description. These originate in the awareness that a structure of this size and performance cannot be built passively for an economic price, if at all. It will be necessary to introduce "active" features in the design to satisfy the major specifications of surface and pointing accuracy. In the LMT these are:

- actuator system for the correction of the reflector surface contour,

- sensor system for the main structural deformations, mainly in the mount,

- optical positioning system for the secondary reflector.

Starting from an optimised structural design (homology) these active systems can reduce the rms surface error and improve the pointing accuracy by about a factor 5 under the defined environmental conditions. To reach our goal we have thus added electrical, mechanical and optical elements to the purely structural design of the telescope; this procedure is called mechatronics (Karcher, 1999, 2006).

A structure similar to that of the MRT, but enlarged to 50 m diameter, will show deviations from pure homology about three times as large, because the deformations scale with the square of the diameter. Thus we must count with a surface deviation on the basis of gravity only of the order of 200 mm. The second major problem is posed by the pointing specification. It is unlikely that a 50 m diameter "sail" in the operational wind of 10 m/s can be held stable to one arcsecond without some fast, active correction mechanism. Also, the deformations due to temperature gradients must be controlled or corrected which is twice as hard as in the case of the MRT (see Sec.7.2). These facts force us to employ the mechatronics aspects in the design.

Fig. 7.16. Cross-section through the LMT. The green concrete foundation carries the antenna on 4 bogies and a bearing atop the central concrete pier. The alidade (gray) and elevation section (purple) are made of steel, as are the BUS (brown) and quadripod. Note the man in the basement for scale. (Drawing LMT Project)

Fig. 7.16. Cross-section through the LMT. The green concrete foundation carries the antenna on 4 bogies and a bearing atop the central concrete pier. The alidade (gray) and elevation section (purple) are made of steel, as are the BUS (brown) and quadripod. Note the man in the basement for scale. (Drawing LMT Project)

But there are more mundane problems in going from 30 to 50 m diameter. For instance, it was determined that the single, large azimuth roller bearing, used in the MRT, could not be realised. In the LMT alidade the vertical load is carried by 4 bogies, each with 4 wheels, onto the rail track with a diameter of 25 m. The lateral loads are carried by a pintle bearing which is placed on top of a central concrete pillar. This minimises turn-over moments and contributes significantly to reducing pointing errors originating in the alidade. In the elevation yoke the support for the reflector backup structure (BUS) is realised by the four corner points of the ballast arms, quite similar to the solution presented before for the WSRT antennas. The BUS is a homologous space-frame structure. The calculated gravitational deformations in zenith and horizon position are illustrated in Fig. 7.17. These are more than a factor ten larger than allowed in the specification and will be compensated by the active surface. The accuracy of current finite element analysis programs is such that this correction can be performed reliably. The stiffness of the BUS is sufficient to keep wind induced deformations of the reflector at an acceptable level without the need to correct these in real time. Such a correction would require a fast, real-time measurement of the deformations or the detailed wind pressure over the surface from which the deformations could be calculated. The 100-m diameter Green Bank Telescope has provisions for such a measurement but it has not yet been fully implemented.

As was mentioned earlier, a serious source of deformation is caused by temperature differences in the structure. The solutions chosen for the LMT are quite similar to those of the MRT. All steel parts are covered by thermal insulation and the air in the BUS volume is constantly circulated by fans to avoid stratification. There remains the problem of the vastly different thermal time constant between the space-frame of

Horizontal position 415 Mm rms Zenith position 316 pm rms

Fig. 7.17. Computed reflector deformations of the LMT from gravity for the zenith and horizon positions. The values 300-400 mm are so much above the specification that active control of the surface will be necessary.

Horizontal position 415 Mm rms Zenith position 316 pm rms

Fig. 7.17. Computed reflector deformations of the LMT from gravity for the zenith and horizon positions. The values 300-400 mm are so much above the specification that active control of the surface will be necessary.

the BUS and the heavy thick members of the alidade. This necessarily leads to a slowly changing temperature difference between BUS and the alidade and a concomitant large-scale deformation of the BUS. This behaviour is very similar to that found in the MRT. The presence of remotely controlled surface adjustment enables us to correct for these deformations on the basis of calculations using measured temperature fields in the structure. We described this procedure in the discussion of the MRT.

We come now to the important aspect of the pointing accuracy. For a telescope of this size with the best design solutions and using all earlier experience we cannot expect to obtain a pointing stability of better than 3-5 arcseconds in the specified operational wind of 10 m/s. So also here we shall need to introduce active elements in addition to the encoder-servo system. Karcher (1999) has suggested the term "flexible body compensation" (FBC). The temperature measurements leading to a reflector adjustment, just discussed are an example of such an FBC. For correcting the pointing errors we need sensors for the measurement of those structural deformations which lead to pointing change without being sensed by the angle encoders on the main axes. Those caused by gravity will be constant in time and only dependent on elevation angle. They are encompassed in the overall pointing model (see Sec. 5.5). The errors of a dynamical nature, caused by the wind, lead to bending mainly of the alidade. We remarked already that the reflector itself is sufficiently stiff to avoid significant deformation under wind influence. It acts as a "sail" and causes moments on the alidade leading to pointing errors. In the LMT a set of inclinometers, placed on top of the elevation bearings, senses the bending in elevation and cross-elevation direction. Together with a highly optimised servo-control system this brings the pointing errors within the specification. Simulations indicate that the FBC system reduces the wind induced pointing errors by about an order of magnitude to an rms error of 0.8 arcsecond in 10 m/s wind.

As a last item, the position of the secondary reflector is also monitored by a laser tracker system. This enables corrections for quadripod deformation. This is a rather straightforward extension of the actuator control of the subreflector, needed to keep it positioned in the best-fit focal point of the primary reflector. Position shifts due to temperature and wind influences are sensed by the laser tracker and correction signals are applied to the actuators.

Fig. 7.18. The panel unit of the LMT. The red-green subframe is connected to the BUS by 4 actuators (shown in blue on the green spars). The cyan rectangle with 4 blue beams is the isostatic structure providing axial support of the panel baseplate (light blue). The green top layer is the reflecting surface.

Fig. 7.18. The panel unit of the LMT. The red-green subframe is connected to the BUS by 4 actuators (shown in blue on the green spars). The cyan rectangle with 4 blue beams is the isostatic structure providing axial support of the panel baseplate (light blue). The green top layer is the reflecting surface.

We conclude the discussion of the LMT with a few remarks on the most important part of the telescope: the reflector surface. As with the MRT, it was decided to subdivide the surface in a relatively small number of panel units, which, in this case, will be supported on their four corners by the motorised adjusters. The reflecting skin of the units will be pre-adjusted in the shop to the prescribed paraboloid, so that on the telescope only the larger units need to be measured and adjusted. The units are about 5 x 2 m2, suitable for transportation and handling, leading to a total of 180 units and 720 adjusters. The error allowance of the panel units is 25 mm. This cannot be achieved by only a "hard" support of the unit corners; the center would sag by about 300 mm. This is solved by the introduction of an "isostatic" intermediate support, made of stainless steel, between the surface skin and the four-point support, as illustrated in Fig. 7.18. The intermediate support functions similar to the "whiffle tree" support used in optical telescopes, notably the 10 m diameter Keck telescopes on Hawaii (Nelson et al., 1983). The baseframe of aluminium is supported on 8 points by the isostatic structure and shows a deformation under gravity of only 10 mm. This baseplate will carry the surface, for which several options are available. One considered is a continuous CFRP sheet supported on some 200 adjusters and pre-adjusted in the shop. Another possibility is the use of a number (8) of separate surface plates. For the inner 3 of the 5 panel unit rings, the surface will indeed be composed of 8 subpanels per unit, fabricated in a new technology, developed by the company Media Lario of Italy for the European ALMA prototype antenna. Because historically this development came before its application to the LMT, we summarise this technology below in the section on the ALMA antennas. These panels can be seen in Fig. 7.19, where the inner rings are being installed.

Fig. 7.19. View of the LMT backup structure with the second and third panel ring being installed (Sep. 2006). The approximately 5x2 m2 panels each carry 8 surface plates. Note the yellow dust settled onto the surface at the 4600 m high site. (E. Mendez, INAOE)

Once operational the LMT will provide exciting new observing capabilities. At the time of writing the mechanical assembly of the telescope is in an advanced state (see Fig. 7.20). The project is plagued by cost overruns and the date of full operation is uncertain. Once completed the telescope will add a collecting area of about 2000 m2 for short mm-wavelength astronomy, a significant increase in the world's supply.

But, as is the case with all single, full aperture telescopes, the angular resolution, even at the level of 5-10", is still about an order of magnitude worse than that of optical telescopes. A significantly higher angular resolution can only be achieved by the application of interferometry and aperture synthesis. In the last section of this chapter we shall discuss some of the features of the antennas of a high resolution array for submillimeter wavelengths as short as 0.3 mm.

Fig. 7.20. The 50-m diameter LMT under construction at an altitude of 4600 m on Cerro la Negra, central Mexico in summer 2006. (E. Mendez, INAOE)
0 0

Post a comment