Fullscreen HTML5Med Res (??MB)HTML4

The Keck observatory biggest optical telescop Recently commissioned on the Hawaiian island of Mauna Kea is the world’s biggest ever optical telescope. At 10 metres in diameter, it is a great deal larger the previous biggest, the Russian 6-metre reflector. This is the story of the Keck Tele­scope. Part 1: By BOB SYMES The Hawaiian Islands, a group of eight main and about 130 smaller volcanic islands, are spread across approximately 2600km in the Pacific Ocean and they rise some 5,500 metres from the floor of the central Pacific Basin. The highest shield mountain of this chain, Mauna Kea, rises a 4,205 metres above sea level, thus leading to the claim that it is the highest mountain on earth, from base to summit. The altitude, combined with the islands’ remoteness from major centres of air pollution, and the prevailing NE trade winds, which combine to keep the weather relatively constant and the air clear and dry, were major considerations in choosing the site for an observatory complex. A prime observing site can more than double the efficiency of any telescope. This is a most important consideration for any large telescope where returns in scientific knowledge need to be balanced against the huge costs involved. After a world-wide survey of possible sites in 1963 by Gerard Kuiner, Mauna Kea stood out as the best place in the northern hemisphere for nighttime observation. The dry air at this altitude, where more than 90% of the atmospheric moisture is below the instruments, is critical for infrared observations, water vapour being the primary attenuator of radia­tion in this part of the spectrum. 4  Silicon Chip Furthermore, Hawaii has a relatively small population and industry is minimal. This leads to low light and industrial pollution. The island also has strict regulations affecting light pollution with particular emphasis on maintaining astronomical quality of the night sky. The summit of Mauna Kea is usually above the inversion layer at night. The layer of clouds that often form below the summit on the windward side of the island as a result act as a further trap for light coming from Hilo 30km away. Gases and aerosols emanating from the occasionally active Kilauea volcano which can affect spec­ tro­ graphic investigations are simi­ larly trapped by the inversion layer. The stability of the air above the inversion layer provides exceptional optical resolution. The ultimate limitation to reso­ lution on earth-based telescopes is air stability, which invari­ ably reduces the theoretical resolution of the instruments them­ selves. On Mauna Kea, sub arc-second “seeing” is normal, and on nights of good air stability, resolutions of better than 0.5 arc-seconds are possible. As a result of these considerations, the Mauna Kea Observa­ t ory was founded in 1967, in affiliation with the University of Hawaii. It has the distinction of being the highest observatory in the world. While conditions at the summit are conducive to astronomi­cal obser- y – the world’s pe Taken under starlight, this photograph of the Keck Observatory, shows the enormous scale of the mosaic telescope which has 36 hexagonal mirror segments kept in alignment by computer control. Note the man standing at one side of the dome opening. Each mirror segment weighs 400 kilograms, giving a total mass of glass of 14.4 tonnes. The total moving mass of the tele­scope is 270 tonnes. vations, they are not quite so good to the astronomers and technicians who operate the facilities. At times the weather can be severe, it is always cold, and oxygen deficiency may be a serious problem for some. For this reason, people intending to work at the summit need to acclimatise at a mid-level facility at Hale Pohaku (9300 feet – 2800m) which was constructed in 1982. The University of Hawaii’s Institute for Astronomy at the Manoa Campus in Honolulu leases the land above 12000 feet (3650m) from the state of Hawaii, and has dedicated it as a Science Reserve. In turn, the university provides site facilities for other observatories who wish to erect telescopes on the summit. Currently, there are eight telescopes in operation on Mauna Kea plus one in the commissioning phase – the W. M. Keck tele­scope, the subject of this article. Neglecting the atmospheric restrictions referred to above, the angular resolution of a telescope mirror (or lens) depends solely on its diameter and the wavelength being investigated. When the angular separation of two stars is very small, it might be imagined that by merely using enough magnification, the stars would resolve into two distinct images. Because of diffraction effects within the optics however, the image of each object is not a point source, but a so-called “Airy disc” whose diameter is 1.1 λ/D radians, or 2.27 x 105λ/D arc-seconds (D being the diameter of the objective lens/mirror in centimetres). If the two discs substantially overlap, any increase in magnification merely gives a larger blur of light, but does not result in separation of the images. The stars will be just re­solved, however, when their Airy discs touch; July 1993  5 light that we see left that object so much earlier in the history of the universe. But at those vast distances, the light reaching the earth is extremely feeble and the apparent size of the object is extremely small. So unless an instrument can be built that can gather as much of the available light as possible, and of sufficient angular resolution to show details of structure etc, little information can be gleaned from these objects. Mirror problems This model of the Keck telescope again shows the enor­mous size of the main mirror. It is much larger than most domes­tic swimming pools and with a focal ratio of f/1.75 (focal length divided by the diameter) it is deeply concave. ie, when the centre to centre distance is equal to the diameter of the disc. Since D is the denominator, by increasing the diameter of the primary mirror/ lens, the diameter of the Airy disc will be proportionate­ly smaller, hence the resolution will increase. Last century, the noted British amateur astronomer W. R. Dawes, working with close double stars, gave an empirical limit for the resolution of a telescope in arc-seconds as 11.5/D (the “Dawes Limit”). Strictly speaking, this figure is wavelength dependent, and refers to visible light of 5 x 105cm. Since it is a rule of thumb rather than an exact physical formu­la, the difference across the visible spectrum is marginal, and can be neglected. 6  Silicon Chip Wavelength does become important however, when calculating the resolution in the infrared spectrum. It follows therefore, where resolution is a factor, that the bigger a lens or mirror can be made, the better. The same goes for light gathering, although in this case it is surface area that is important rather than diameter. The two are not necessarily related. Doubling the diameter of a circular mirror gives four times the light-gathering power, and a doubling in resolution. Since researchers are forever trying to look further back in time, this increase in light-gathering power becomes of great importance. The further away an object is, the further back in time we can look, since the The simple solution is to make bigger monolithic mirrors or lenses. But the problems associated with them ultimately become insurmountable. Lenses supported only around their circumferenc­es sag under their own weight. Once the sag becomes apprec­ i­ able, image-quality deteriorates to a point where it becomes unusable. Thus it is unlikely that large lenses will ever again be used for astronomical work, although the existing ones still perform admirably. The largest of them, the 40-inch (1m) telescope at Yerkes Observatory at William Bay in Wisconsin, built by that most famous of telescope builders, Alvin Clark, and dedicated in 1897, is likely to remain forever the greatest of refracting telescopes. Larger mirrors are easier to design and build, since they can be supported from the rear, and since only one critical surface has to be figured to high accuracy, as opposed to the four (or sometimes six) surfaces that need to be ground and polished for an achromatic objective lens. In addition, flaws such as bubbles, inclusions and striae in the glass of a mirror are acceptable, whereas they would be intolerable in a lens system. Nevertheless, massive engineering prob­lems remain. The larger a mirror becomes, the thicker it needs to be to avoid flexure and hence the heavier it becomes. The mounting becomes bigger and heavier, along with the cost, and finally there is reached a point at which further gains are no longer feasible. There is the additional problem that the more mass of glass there is, the longer it takes to reach thermal equilibrium, and during this time, image quality suffers due to local distortions in the mirror. As new materials and techniques became available, the boundary of what was feasible was pushed further 8MM VIDEO CASSETES These 120-minute 8mm metal oxide video cassettes were recorded on once for a commercial application and then bulk erased. They are in new condition but don’t have the record protect tabs fitted. The hole in the upper right corner will have to be taped over. $9 Ea. or 5 for $38 LARGE NIGHT VIEWERS One of a kind! A very large complete viewer for long range observation. Based on a 3-stage fibre optically coupled 40mm first generation image intensifier, with a low light 200mm objec­tive mirror lens. Designed for tripod mounting. Probably the highest gain-resolution night viewer ever made. ONE ONLY at an incredible price of: $3990 BINOCULAR EHT POWER SUPPLY This low current EHT power supply was originally used to power the IR binoculars advertised elsewhere in this listing. It is powered by a single 1.5V “C” cell and produces a negative voltage output of approximately 12kV. Can be used for powering prefocussed IR tubes etc. $20 IR BINOCULARS High quality helmet mount, ex-military binocular viewer. Self-powered by one 1.5V “C” size battery. Focus adjustable from 1 metre to infinity. Requires IR illumination. Original carry case provided. Limited stocks, ON SPECIAL AT: $500 IR FILTERS A high quality military grade, deep infrared filter. Used to filter the IR spectrum from medium-high powered spotlights. Its glass construction makes it capable of withstanding high temper­atures. Approx. 130mm diameter and 6mm thick. For use with IR viewers and IR responsive CCD cameras: ON SPECIAL $45 12V OPERATED LASERS WITH KIT SUPPLY Save by making your own laser inverter kit. This combination includes a new HeNe visible red laser tube and one of our 12V Universal Laser Power Supply MkIII kits. This inverter is easy to construct as the transformer is assembled. The supply powers HeNe tubes with powers of 0.2-15mW. $130 with 1mW TUBE $180 with 5mW TUBE $280 with 10mW TUBE MAINS OPERATED LASER Supplied with a new visible red HeNe laser tube with its matching encapsulated (240V) supply. $179 with 1mW TUBE $240 with 5mW TUBE $390 with 10mW TUBE GREEN LASER HEADS We have a limited quantity of some brand new 2mW+ laser heads that produce a brillant green output beam. Because of the relative response of the human eye, these appear about as bright as 5-8mW red helium neon tubes. Approximately 500mm long by 40mm diameter, with very low divergence. Priced at a small fraction of their real value $599 A 12V universal laser inverter kit is provided for free with each head. ARGON HEADS These low-voltage air-cooled Argon lon Laser Heads are priced according to their hours of operation. They produce a bright BLUE BEAM (488nm) and a power output in the 10-100mW range. Depends on the tube current. The head includes power meter circuitry, and starting circuitry. We provide a simple circuit for the supply and can provide some of the major components for this supply. Limited supplies at a fraction of their real cost. $450-$800 ARGON OPTIC SETS If you intend to make an Argon laser tube, the most expen­sive parts you will need are the two mirrors contained in this ARGON LASER OPTIC SET. Includes one high reflector and one output coupler at a fraction of their real value. LIMITED SUPPLY $200 for the two Argon LASER mirrors. LASER POINTER Improve and enhance all your presentations. Not a kit but a complete commercial 5mW/670nm pen sized pointer at ONLY: $149 LARGE LENSES Two pairs of these new precision ground AR coated lenses were originally used to make up one large symmetrical lens for use in IBM equipment. Made in Japan by TOMINON. The larger lens has a diameter of 80mm and weighs 0.5kg. Experimenters delight at only: $15 for the pair. EHT GENERATOR KIT A low cost EHT generator kit for experimenting with HT-EHT voltages: DANGER – HIGH VOLTAGE! The kit also doubles as a very inexpensive power supply for laser tubes: See EL-CHEAPO LASER. Powered from a 12V DC supply, the EHT generator delivers a pulsed DC output with peak output voltage of approximately 11kV. By adding a capacitor (.001uF/15kV $4), the kit will deliver an 11kV DC output. By using two of the lower voltage taps available on the transformer, it is possible to obtain other voltages: 400V and 1300V by simply adding a suitable diode and a capacitor: 200mA - 3kV diode and 0.01uF 5kV capacitor: $3 extra for the pair. Possible uses include EHT experiments, replacement supplies in servicing (Old radios/CRO’s), plasma balls etc. The EHT generator kit now includes the PCB and is priced at a low: $23 LED DISPLAYS National Seminconductor 7-segment common cathode 12 digit multiplexed LED displays with 12 decimal points. Overall size is 60 x 18mm and pinout diagram is provided. 2.50 Ea. or 5 for $10 BATTERIES Brand new industrial grade PANASONIC 12V-6.5AHr sealed gel batteries at a reduced price.Yes, 6.5 AHr batteries for use in alarms, solar lighting systems, etc. Dimensions: 100 x 954 x 65mm. Weight of one battery is 2.2kG. The SPECIAL price? $38 PIR DETECTORS What are the expensive parts in a passive movement dector as per EA May 89? A high quality dual element PIR sensor, plus a fresnel lens, plus a white filter. We include these and a copy of PIR movement detector circuit diagram for: $9 MASTHEAD AMPLIFIER KIT Based on an IC with 20dB of gain, a bandwidth of 2GHz and a noise figure of 2.8dB, this amplifier kit outperforms most other similar ICs and is priced at a fraction of their cost. The cost of the complete kit of parts for the masthead amplifier PCB and components and the power and signal combiner PCB and components is AN INCREDIBLE: $18 For more information see a novel and extremely popular antenna design which employs this amplifier: MIRACLE TV ANTENNA - EA May 1992: Box, balun, and wire for this antenna: $5 extra SODIUM VAPOUR LAMPS Brand new 140W low pressure sodium vapour lamps. Overall length 520mm, 65mm diameter, GEC type SO1/H. We supply data for a very similar lamp (135W). CLEARANCE AT: lenses: two plastic and one glass. The basis of a high quality magnifier, or projection system? Experimenters’ delight! $30 CRYSTAL OSCILLATOR MODULES These small TTL Quartz Crystal Oscillators are hermetically sealed. Similar to units used in computers. Operate from 5V and draw approximately 30mA. TTL logic level clock output. Available in 4MHz, 4.032MHz, 5.0688MHz, 20MHz, 20.2752MHz, 24.74MHz, 40MHz and 50MHz. $7 Ea. or 5 for $25 FLUORESCENT BACKLIGHT These are new units supplied in their original packing. They were an option for backlighting Citizen LCD colour TVs. The screen glows a brilliant white colour when the unit is powered by a 6V battery. Draws approximately 50mA. The screen and the in­verter PCB can be separated. Effective screen size is 38 x 50mm. $12 MAINS FILTER BARGAIN For two displays - one yellow green and one silver grey. SOME DIFFERENT COMPONENTS 1000pF/15kV disc ceramic capacitors ..............$5 20kV PIV - 5mA Av/1A Pk fast diodes .........$1.50 3kV PIV - 300mA / 30A Pk fast diodes ........... 60c 0.01uF /5kV disc ceramic capacitors ...........$1.80 680pF / 3kV disc ceramic capacitors .............. 30c Who said that power MOSFETS are expensive?? MTP3055 N-channel MOSFETS as used in many SC projects ............................$2 Ea. or 10 for $15 MTP2955 P-channel MOSFETS (complementary to MTP3055) ..........................$2 Ea. or 10 for $15 BUZ11 N-channel MOSFETS $3 Ea. or 10 for $25 Brief DATA and application sheet for above MOSFETS free with any of their purchases (ask) Flexible DECIMAL KEYPADS with PCB connectors to suit ...........................................................$1.50 1-inch CRO TUBES with basic X-Y monitor circuit CLEARANCE <at>..............................................$20 Schottky Barrier diodes 30V PIV - 1A/25A Pk. 45c 100 LED BARGRAPH DISPLAY Note that we also have some IEC extension leads that are two metres long at $4 Ea. Yes 100 LEDs plus IC control circuitry, all surface mounted on a long strip of PCB. SIMPLE - a 4-bit binary code selects which one out of the 10 LED groups will be on, whilst another 4-bit binary code selects which one of each group of 10 LEDs will be ON. Latching inputs are also provided. We include a circuit and a connecting diagram. VERY LIMITED QUANTITY WEATHER TRANSMITTERS FM TRANSMITTER KIT - MKll A complete mains filter employing two inductors and three capacitors fitted in a shielded metal IEC socket. We include a 40 joule varistor with each filter. $5 These brand new units were originally intended to monitor weather conditions at high altitudes: attached to balloons. Contain a transmitter (12GHz?) humidity sensor, temperature sensor, barometric altitude sensor, and a 24V battery which is activated by submersing in water. The precision all mechanical altitude sensor appears similar to a barometer and has a mechani­cal encoder and is supplied with calibration chart. Great for experimentation. $16 Ea. SOLAR CHARGER Use it to charge and or maintain batteries on BOATS, for solar LIGHTING, solar powered ELECTRIC FENCES etc. Make your own 12V 4 Watt solar panel. We provide four 6V 1-Watt solar panels with terminating clips, and a PCB and components kit for a 12V battery charging regulator and a three LED charging indicator: see March 93 SC. Incredible value! $42 6.5Ahr. PANASONIC gel Battery $35, ELECTRIC FENCE PCB and all onboard components kit $40. See SC April 93. $7Ea. This low cost FM transmitter features pre-emphasis, high audio sensitivity as it can easily pick up normal conversation in a large room, a range of well over 100 metres, etc. It also has excellent frequency stability. The resultant frequency shift due to waving the antenna away and close to a human body and/or changing the supply voltage by +/-1V at 9V will not produce more than 30kHz deviation at 100MHz! That represents a frequency deviation of less than 0.03%, which simply means that the fre­quency stays within the tuned position on the receiver. Specifications: tuning range: 88-101MHz, supply voltage 6-12V, current consumption <at>9V 3.5mA, pre-emphasis 50µs or 75µs, frequency response 40Hz to greater than 15kHz, S/N ratio greater than 60dB, sensitivity for full deviation 20mV, frequency stabil­ity (see notes) 0.03%, PCB dimensions 1-inch x 1.7inch. Construction is easy and no coil winding is necessary. The coil is preassembled in a shielded metal can. The double sided, solder masked and screened PCB also makes for easy construction. The kit includes a PCB and all the on-board components, an electret microphone, and a 9V battery clip: $11 Ea. or 3 for $30 LARGE LCD DISPLAY MODULE - HITACHI These are Hitachi LM215XB, 400 x 128 dot displays. Some are silver grey and some are yellow green reflective types. These were removed from unused laptop computers. We sold out of similar displays that were brand new at $39 each but are offering these units at about half price. VERY LIMITED STOCK. $40 OATLEY ELECTRONICS $15 Ea. PO Box 89, Oatley, NSW 2223 STEPPER MOTORS Phone (02) 579 4985. Fax (02) 570 7910 $12 MAJOR CARDS ACCEPTED WITH PHONE & FAX ORDERS These are brand new units. Main body has a diameter of 58mm and a height of 25mm. Will operate from 5V, has 7.5deg. steps, coil resistance of 6.6 ohms, and it is a 2-phase type. Six wires. ONLY: PROJECTION LENS Brand new large precison projection lens which was original­ly intended for big screen TV projection systems. Will project images at close proximity onto walls and screens and it has adjustable focussing. Main body has a diameter of 117mm and is 107mm long. The whole assembly can be easily unscrewed to obtain three very large P & P FOR MOST MIXED ORDERS AUSTRALIA: $6; NZ (Air Mail): $10 July 1993  7 and further back. The 200-inch (5m) Hale telescope on Mount Palomar in southern California would not have been possible without the development of Pyrex, a low expansion glass which allowed the 14.5-tonne mirror to reach thermal equilibrium in time for the astronomers to still have some dark hours in which to do their work! The development of air-conditioning and efficient insulat­ing materials also helped by keeping the inside of the dome and hence the mirror at a constant average night-time temperature; ie, cold. Nevertheless, Mount Palomar seems to be about the largest size telescope that can be made using conventional mirrors and equatorial mountings. Continuing the development of new techniques, the 6-metre BTA (Bolshoi Teleskop Azimutal’ny = Large Alta­ zimuth Telescope) on Mount Pastuk­ hov in southern Russia was the first large tele­scope to use an altazimuth mount instead of an equatorial, since the equatorial would have been too massive to control accurately, and too costly to build. But the advent of the altazimuth mount had to await comput­ers with sufficient power to control the continually changing position of the telescope, since the calculations to move each axis are far more complex than the requirements of an equatorial mount, where (more or less) the drive has only to be able to rotate the polar axis at the sidereal rate. Altazimuth telescopes have the additional complexity of field rotation 8  Silicon Chip during ob­ serva­ tion, a problem not encountered with equatorial mounts. Further computing and mechanical complexity is involved in resolving this problem. The type of glass used in a tele­ scope mirror has a great bearing on the ultimate size that can be produced. Ordinary borosilicate glass is easy to cast in large sizes and to stress relieve after casting, but thermal expansion is so great that it is unusable in this role. The development of Pyrex, in reducing thermal expansion to tolerable limits, enabled much larger mirrors to be contemplated, but casting an homogenous blank was far more difficult, and it had a tendency to crack when being stress-relieved, a process that often took months or even years of slow cooling. Fused quartz has been used successfully but the extreme difficulty of making large blanks has limited its use on very large telescopes, as has the development of new and better materials. As each new glass was developed, the rewards in temperature stability were greater, but so were the problems of manufacture. Cervit and its Soviet counterpart SITAL (used in the replacement mirror for the 6-metre telescope) were the first successful attempts to make a complex ceramic-glass mixture, where the coefficient of expansion of the ceramic almost exactly countered the opposite coefficient of expansion of the glass. This was taken a step further with the development of Zerodur by Schott of Mainz, Germany. After initial cast- ing, careful control of the subsequent stabilising/stress-relieving therm­al cycle results in a glass in which half of the mass is crypto­crystall­ine and half is a supercooled liquid – the socalled “ceramization” process. Again, the coefficients of ex­pan­sion of the two phases are equal but opposite and closely cancel each other out. The worst example of thermal problems in a large telescope came with the original 42-tonne pyrex-like primary of the BTA, where a change of no more than 20°C per day in glass temperature could be tolerated and still maintain a useable figure during night­-time observing runs. The next development was that of thin-mirror telescopes. Usually, the thickness:diameter ratio of the glass blank is between 1:6 to 1:8. As mentioned before, these larger mirrors become inordinately heavy and need to be supported by inor­dinately massive mounts, and the problem of pointing finesse and controllability as well as thermal equilibrium considerations again dictate limits. Thus was born experimentation and success­ ful implementation of thin-mirror technology, with thickness:diameter ratios of 1:10 to 1:25. These were made possi­ble by the rigidity of the newly developed glasses, and by cast­ing the blanks so that they tapered in thickness from the centre out, as well as having anti-flexure webs incorporated on the rear of the mirror. This went a great way to reducing the problems associated with weight, flexure and thermal equilibrium. ORDER FORM BACK ISSUES MONTH YEAR MONTH YEAR PR ICE EACH (includes p&p) Australi a $A7.00; NZ $A8.00 (airmail ); Elsewhere $A10 (airmail ). Buy 10 or more and get a 10% discount. Note: Nov 87-Aug 88; Oct 88-Mar 89; June 89; Aug 89; Dec 89; May 90; Aug 91; Feb 92; July 92; Sept 92; NovDec 92; & March 98 are sol d out. All other issues are currently i n stock. TOTAL $A B INDERS Pl ease send me _______ SILICON CHIP bi nder(s) at $A12.95 + $5.00 p&p each (Australi a only). N ot avail abl e elsewhere. Buy five and get them postage free. $A SUBSCRIPTIONS ❏ New subscription – month to start­­___________________________ ❏ Renewal – Sub. No._______________   ❏ Gift subscription ☞ RATES (please tick one) Australia Australia with binder(s)* NZ & PNG (airmail) Overseas surface mail 2 years (24 issues) 1 year (12 issues) ❏ $A84 ❏ $A42 ❏ $A105 ❏ $A53 ❏ $A130 ❏ $A65 ❏ $A130 ❏ $A65 ❏ $A240 Overseas airmail ❏ $A120 *1 binder with 1-year subscription; 2 binders with 2-year subscription GIFT SUBSCRIPTION DETAILS Month to start__________________ Message_____________________ _____________________________ _____________________________ Gift for: Name_________________________ (PLEASE PRINT) YOUR DETAILS Your Name_________________________________________________ (PLEASE PRINT) Address___________________________________________________ Address______________________ _____________________________ State__________Postcode_______ ______________________________________Postcode___________ Daytime Phone No.____________________Total Price $A __________ ❏ Cheque/Money Order ❏ Bankcard ❏ Visa Card ❏ Master Card 9am-5pm Mon-Fri. Please have your credit card details ready 10  Silicon Chip ______________________________ Card expiry date________/________ Card No. Phone (02) 9979 5644 Signature OR Fax (02) 9979 6503 Fax the coupon with your credit card details 24 hours 7 days a week Mail coupon to: OR Reply Paid 25 Silicon Chip Publications PO Box 139, Collaroy 2097 No postage stamp required in Australia As each of the above problems of large mirror making were more or less, successfully solved, even larger mirrors became feasible. But there remained one difficulty that couldn’t be reduced easily – that of the actual figuring and final polishing of the reflecting surface itself. It is generally agreed amongst optical engineers that doubling the diameter of a mirror makes it 10 times more difficult more difficult to grind. The amount of material to be removed is significantly greater, and the final zonal corrections are fraught with time-consuming difficulty. If one zone is high, it only has to be polished down to specification, but if it is low, the entire surface has to be polished down to accommodate the low spot. Even though we are speaking of microns, the work involved in polishing down a large mirror is massive. And always bearing in mind that not only does the final figure have to be good, but the focal point cannot be changed by any corrections or re-figuring, as by this abolise or hyper­bolise the surface by deepening the centre with a sub-diameter polishing lap. The first new technique is a computer controlled polishing engine, usually combined with a laser profilometer feeding back to the controller. It has the advantage of good accuracy and is much quicker and less labour intensive than manual polishing. Since large mirrors are so seldom made, the computer polisher is usually made as a one-off special for that particular mirror and this adds substantially to the cost, speed of execution notwith­ standing. The second new technique is known as spin-casting. Glass is melted in an electrically heated mould and held for a time to soak so as to remove as many bubbles and other imperfections and inclusions as possible, and then spun whilst cooling to produce the required paraboloidal shape. The mould also incorporates a honeycomb base which creates a lightweight blank. The resulting curved blank dramatically reduces All this had to come together at the top of a windswept mountain where the air is so thin that the engineers & construction workers had to contend with dizziness, headaches, forgetfulness & dehydration. time the structural engineering side would be well on the way to designing and building the mounting, which by now cannot be changed. There is also a trend to design large tele­scopes with very fast optics, often less than f/2. Firstly, this gives the observer a much brighter image to work with, albeit at a reduced image scale. Also, the supporting structure can be much lighter because of the shorter tube involved and significant savings can be made in the design and construction of the dome. Two techniques are successfully used today to partially overcome the difficulty of grinding and polishing mirrors to the required shape. Virtually all telescope mirrors have a para­ boloidal or hyperboloidal cross-section and the traditional technique is to first grind it to a spherical surface, and then after testing for the sphere by traditional optical means, to par- the amount of material that has to be removed and hence the time to attain the final figure. Several large astronomical mirrors in the 6-metre to 8-metre range have been cast successfully with this method, although at least three (8.2- metre blanks for the European South­ern Observatory’s Very Large Tele­scope –VLT) have cracked and have been destroyed in the annealing stage. A final technique that had been discussed theoretically for years is that of stressed mirror polishing. In effect, the mirror blank is deliberately distorted to a predetermined shape and then polished to a spherical section by conventional methods. After final polishing, the distorting forces are removed, and the mirror takes up (hopefully!) the desired shape. The greatest proponent of this new method was Jerry Nelson of the University of California. In the late 1970s he proposed that large astronomical mirrors could be produced this way. He made a further proposal, one that was to have a great bearing on the design and building of modern tele­ scopes – that large mirrors be made of multiple segmented smaller mirrors rather than one large blank. The idea of segmented mirrors to avoid the weight problem and the increased complexity that accompanies figuring large single mirrors is not new, having been discussed by the third Earl of Ross in the mid 1800s. In the late 1940s, Horn-d’Aturo in Italy actually made a 61 hexagonal-segment mirror. This formed a 1.8-metre f/6 telescope that gave good images, although it was unsteerable. With the previously discussed years of telescope design, glass making and polishing technology, and adequate computing power, the stage was set for the development of the most ambi­tious optical device ever built, the 9.82-metre W. M. Keck tele­scope. The driving force behind the radical new telescope was Jerry Nelson. He spent a great deal of time convincing the pun­dits that such a project was feasible, since nothing on this scale had ever been tried before. From the start, the concept and design were revolutionary. New methods had to be devised to construct mirror segments, the warping harness, support struc­ t ure, actuators, and the computer programs that brought them all together. The segmented mirror design on such a large telescope was novel, and there was no previous experience at this scale to draw on. The mount would have to be rigid enough to keep the segments in exquisite alignment but light enough to gain from the benefit of such a design. The electronics to sense and correct misalign­ment had to be developed from scratch. Even the grinding and polishing of the mirror segments themselves were to use new and untried techniques. In all facets, innovative thinking and methods had to be employed. And all this had to come together at the top of a windswept mountain where the air is so thin that the engineers and construction workers had to contend with dizziness, forget­ fulness, headaches and dehydration, while solving the engineering problems that would be inevitable with such a massive SC undertaking. July 1993  11