Pushed beyond customary limits or novel combinations of technologies
- Active Optics
- Large Metal Blanks
- Shack-Hartmann wavefront sensors
- Real-time processors
- Time Reference System
- Data Archive Systems
- Virtual Observatories
- Cryogenic bearings
- Thermally controlled cabinets
The concept of actively controlling the form of meniscus mirrors was developed by ESO for the NTT. One of the main drivers was to break the classical cost power law for large telescopes, but even at First-Light the NTT demonstrated image quality almost never seen before on previous ground-based telescopes.
Since the NTT, essentially all large optical telescopes world-wide use Active Optics. This is even true for Honeycomb borosilicate glass mirrors which, before the NTT, were seen as the main rival to thin meniscus mirrors pioneered by ESO.
During the development of the NTT, ESO carried out a development programme to investigate various types of metals for possible application to the NTT primary mirror, and potentially the VLT as well. Although metal mirror telescopes had been tried in the past, most of these were not very successful because of a lack of a proper understanding of the material properties and manufacturing processes.
ESO carried out evaluations of various types of steel and aluminium materials, manufacturing processes, annealing methods and coatings, including long-term optical tests on a series of polished mirrors. These evaluations showed that, with the right choice of material and manufacturing process, metal mirrors would be a satisfactory and cost effective solution, at least for mirrors up to 4 metres in diameter.
The fact that the NTT did not, in the end, have a metal mirror was due largely to schedule and other constraints rather than technical feasibility.
The development of Shack-Hartmann wavefront sensors was a key technological development that allowed Active Optics to succeed in practice. It combined the optical device proposed by Roland Shack at the University of Arizona in 1970 with CCD detectors, and allowed the optical alignment and shape of the main optics of the NTT and, later, the VLT to be verified and corrected in real time. A key optical component of the Shack-Hartmann wavefront sensor is the lenslet array which contains 400 lenslets, each 1.0 mm and, eventually, 0.5 mm across. Here ESO worked with the Paul Scherrer Institut in Switzerland to develop the master and Jobin-Yvon France to manufacture the final copies.
Several commercial systems based on Shack-Hartmann wavefront sensors have been developed for optical testing, for example from Imagine Optic in France (for optical testing), Zeiss in Germany (for eye surgery), Spot Optics in Italy (for optical alignment and testing) and BfI Optilas in Germany for laser beam profile measurement.
Shack-Hartmann wavefront sensors are presently in wide-spread use in adaptive optics systems. Here, ESO has again worked together with several firms to achieve state-of-the-art systems by developing the technological and manufacturing know-how of industry. For example, an ESO development contract for lenslet arrays placed with the Finnish-Swiss firm Heptagon, allowed the firm to reach specifications that went far beyond anything that had been achieved previously by the firm.
An on-going ESO development programme is applying a large number of standard commercial Linux PCs (presently 44) in a Beowulf cluster for fast parallel processing of wavefront errors. The processors are completely standard and therefore cheap, but use the latest and fastest commercial technology. This is seen as a highly promising approach towards achieving the processing power needed for future multi-conjugate Adaptive Optics systems.
A parallel development, which also has implications for future laser guide-star systems, is aimed at developing a 589 nm fibre laser system. This is seen as the long term solution to laser guide-star systems, especially when multiple lasers are required for multi-conjugate adaptive optics correction, due to their intrinsic simplicity and robustness.
ESO, together with the Lawrence Livermore Laboratory with whom ESO has a Partnership Agreement, are working on two different approaches to realising such a fibre laser. ESO's approaching is to use a frequency doubled Raman laser, and has placed a contract with IPF Technology in the UK to develop an ESO design for prototype testing. Although this work is still on-going, it has already produced two ESO patent applications (see below).
The measurement of time has always been inextricably linked to astronomy. In the past, many major observatories, including La Silla, relied on Caesium clocks as time references. These devices were cumbersome, expensive, and had to be regularly transported to national time standards for recalibration.
With the start of commercial operation of the GPS system around 1990, ESO decided to base the VLT time standard on this technology. The ESO Time Reference System consists of a commercial GPS receiver and an ESO-developed time encoding system, fibre-optic distribution system and local crystal frequency standards in each VME Local Control Unit. The system allows the hundreds of VLT real-time processors to be precisely synchronised to absolute UT time at the sub-micro-second level. It also allows any particular sub-system to be used "off-line" using the local reference oscillator. This is important both for reliability, in case the master clock is unavailable for any reason, and for laboratory testing.
As well as providing accurate absolute time to all processes and at any location within the observatory, the ESO Time Reference System allows any number of independent processes, for example secondary mirror chopping, data readout, and shutter opening, to be synchronised without the need of special cable links, etc. In principle, this could even be done over intercontinental distances.
Data archiving is a fundamental aspect of modern astronomy. ESO telescopes collect data at unprecedented rates. In 2003, for example, the amount of science data collected from the ESO Observatories was about 10 Terabytes per year, and this is expected to rise to 250 Terabytes per year in 2013 when VST, VISTA and ALMA are in operation. This increase is considerably higher than the rate of increase in computing power that has to process it. Another problem facing data handling is the need to transport very large amounts of data between the Observatories and the archive in Garching, as well as between the archive and the end users. This considerably exceeds the capacity of current WAN connections.
The ESO Next Generation Archiving System (NGAST) has been developed to overcome these problems by using commercial computer hardware and standards in a novel way. NGAST is built around a RAID array of standard modularised hard disc drives. At the present time, these media offer the highest data storage density, as well as being robust, compact and easy to transport. A key element in NGAST architecture is the modularisation of the storage system. Each bank of disc drives has its own processor so that, as the storage capacity increases, the processing power increases at the same rate. Thus, the processing time, even for operations affecting the complete archive, remains constant irrespective of the eventual size of the archive. Moreover, the intrinsic modularisation allows for easy extensions and upgrades to take advantage of progress in storage and processing technology.
ESO has been a major player in the rapidly developing field of virtual observatories and leads the 6-institute European AVO (Astronomical Virtual Observatory) consortium. Astronomical data archives have existed for a long time. The main problem with these in the past has been that the data have often not been stored in a way that allows people other than the original observers to make full use of them. Also, exchanging data between one archive to another was usually difficult, if at all possible. To overcome this, standard processes and data formats have been developed that allow astronomers to retrieve data from almost any science archive irrespective of which telescope the data came or where the astronomer is located.
The key to Virtual Observatories lies in defining internationally accepted standards for access and data distribution. Within AVO, ESO has been working on the VOTable format for the exchange of tabular data that is based on XML. This already has wide acceptance within the community and should be available for the archives of the AVO members at the beginning of 2005. The target within Europe is to have an open access VO covering all European data centres by the end of 2007.
Complex infrared instruments require mechanical functions that work very reliably under vacuum and at cryogenic temperatures as well as room temperatures, and exhibit good thermal conductivity. Moreover, standard lubricants cannot be used in a vacuum. Unlike a 'warm' instrument in which mechanical functions can usually be replaced in an hour or so, even the smallest intervention inside a cryogenically cooled instrument requires a cycle of typically 4-5 days to warm up the instrument, repair it, re-evacuate it and cool it down again. No standard mechanical bearings will work properly under these conditions. To overcome this problem, ESO has developed the technology for cryogenic ball-bearings to enable bearings up to 200 mm in diameter to be produced that can operate reliably under these difficult conditions and without lubricants. This technology involves the selection of materials for all parts, for example the use of ruby balls, the surface treatment of metallic parts and spacer design.
Heat generation near telescopes has long been recognised as a major contributor to the degradation of optical quality. Unfortunately, HI-TEC telescopes like the VLT use ever more active control systems to maintain optical alignment and image quality instead of the massive structures used in the past. This proliferation of electronics in and around the telescopes has necessitated some innovative designs for cabinet cooling. The problem is not simply to remove the heat of the electronics but to maintain the surface temperature of the cabinet within about 1 degree of the ambient temperature so that the effects of thermal convection currents are negligible. As the electronic cabinets are generally severely space limited, the used of external thermal insulation is excluded. Several designs have been developed at ESO for both large and small cabinets that actively control the external cabinet temperature to follow that of the ambient air. Careful control of the internal air flow is also mandatory to avoid external hotspots.