The manufacturing achievement of Ritchey with the 40-inch RC telescope should be judged from the fact that this was the first time in 300 years that a primary had been manufactured with a form other than the parabola prescribed by Descartes, apart from his own experimental RC telescope of half the aperture completed in France in 1927. Ritchey had to rely largely on zonal testing at the centre of curvature since an autocollimation test against a plane mirror at the focus was no longer an automatic null test.
In the 1930's, two Schwarzschild telescopes were manufactured in the United States, a 24-inch for the University of Indiana and a 12-inch for Brown University [5.17]. The problems of obstruction and length made the Schwarzschild form less attractive than the RC form.
Apart from Ritchey's and Pease's ideas for giant telescopes, Hale himself was determined to advance further than the 100-inch Mt. Wilson telescope. Although he was attracted to the visionary plans of Pease for a 300-inch telescope, he finally settled for a 200-inch [5.2] [5.3]. This was a wise decision, as it already represented a step of a factor of two in diameter. The whole history of the reflector showed that even this represented a high risk. The fundamental issue, as always, was the nature of the primary mirror. This decision (on its structure) was taken in 1928, the site on Mt. Palomar was decided in 1934. We shall confine ourselves here to a brief account of the optical aspects of this remarkable telescope. Apart from excellent accounts by King [5.2] and Riekher [5.3], an overall review was given by Bowen [5.23]. General accounts are given by Woodbury [5.24] and Wright [5.25].
Hale and his colleagues, reinforced by the experience of Ritchey, realised that a massive glass blank of the 100-inch type would not only weigh about 40 tons but would lead to insoluble thermal problems - a nine year cooling period of a massive Pyrex blank was estimated! Many materials were considered, the first choice being fused quartz and the second Pyrex. Pyrex was made by Corning for ovenwear and had normally an expansion about one third of that of plate glass, while quartz was some 6 times as favourable as Pyrex (see RTO II, Chap. 3). Between 1928 and 1931, Elihu Thomson of G.E.C. pursued the development of fused quartz and made two 60-inch blanks. Because of the high cost of the programme and the limited size achieved, fused quartz was abandoned in favour of Pyrex. It was not until 20 years later that the solution of welding together segments of fused quartz led to the use of this material in large blanks.
Pease had made a design for a lightweighted blank with ribs and cylindrical bores for the support system. This reduced the weight of the 200-inch (5 m) blank by a factor of 2 to about 20 t, the worked face being only 12 cm thick (Fig. 5.18). The contract was given to Corning in 1931. A form of Pyrex was used with 80.5% SiO2 content [5.3] giving an expansion coefficient only about 3 to 4 times that of fused quartz [5.2] instead of the normal 5 to 6 times. Several intermediate blanks were cast, including a 120-inch finally used for the Lick 120-inch reflector. The first 200-inch blank was cast in March 1934, but
was considered unacceptable because mould cores broke loose and remained in the blank. The cores were then cooled and fixed with chrome-nickel steel attachments for the second casting in December 1934. This was successful, as was the cooling and annealing over an 8-month period. In March 1936, this second blank was transported by rail to Los Angeles. Hale died in 1939 and did not see the completion of the project. But the successful completion of the blank was the most fundamental and critical step, justifying his confidence in the entire project.
For the optical work, starting in 1936, much of Ritchey's technology was taken over. The basic test in achieving the parabolic form was again a zonal test of the differences of radii of curvature. Aspherising was a much greater task than with Ritchey's 100-inch primary because the relative aperture was f/3.3, easily the fastest large mirror ever figured at that time. The difference between sphere and parabola according to Eq. (5.1) is 0.136 mm. This was too large to be removed by polishing alone; alternate fine grinding and polishing for testing was performed, a slow and laborious process directed by Anderson, Brown and Hendrix. The second world war caused a further delay and the mirror was not finished until 1947.
Apart from the introduction of low expansion borosilicate glass (Pyrex), the 1930's saw another major revolution in the technology of the reflecting telescope: vacuum evaporation of aluminium to replace chemical silvering. Apart from much improved reflectivity in the UV, the protection by translucent oxide gave a coating much more robust against tarnish (see RTO II, Chap. 6). A special coating plant for Al was built in Palomar. The primary was first tested in the telescope by Anderson in December 1947. Lengthy adjustments of the complex support system followed. This support system comprised 36 supports, combining both axial and lateral functions, inset into the cylindrical bores. The mirror had been left with a "turned-up edge", a trend often encouraged by opticians to avoid the dreaded "turned-down edge", requiring repolishing of the whole surface for its correction. It was hoped that the turned-up edge would compensate in function in the cell. This was not the case. In May 1949 the primary was removed and the edge zone retouched by hand, only 9 hours work but spread over 6 months because of careful testing. In November 1949 it was declared finished and the telescope started operation in 1950.
Figure 5.19 shows one of R.W. Porter's famous drawings of the Palomar telescope. The horseshoe equatorial mounting was a direct modification of the 100-inch cradle mount to permit access to the polar region. The aperture was large enough, for the first time, to include a prime focus cage for the observer, rather than a Newton focus. As discussed in Chap. 4, the PF field was limited to about 12 mm diameter by field coma and this was extended by factors up to about 12 by the Ross correctors giving final apertures from f/3.6 to f/6.0. The Cassegrain system was f/16, giving m2 = -4.85, quite typical of present day telescopes. The coude focus (f/30) is reached either by one plane
mirror of variable rotation angle sending the beam down the polar axis; or via three plane mirrors if the telescope is observing near the pole [5.3]. An important feature for the optical quality was the so-called Serrurier truss, a tube design due to M. Serrurier which compensated lateral decentering coma by equal lateral sags of the tube at the secondary and primary. This is discussed further in RTO II, Chap. 3 in connection with active optics. The success was measured by the result that, for all positions of the tube, the focus of the primary did not vary from its mean position by more than 25 ^m [5.2].
In 1950, one of the first photographs of a galaxy using the Ross field corrector was published [5.26]. It showed NGC 147 clearly resolved, the smallest star image being slightly more than 0.5 arcsec in diameter, confirming the excellent quality of the primary and the potential of the telescope. The 100-inch could only resolve this "nebula" with red-sensitive plates. Riekher [5.3] quotes the workshop tested quality of the Palomar primary as 68% of the geometrical optical energy within 0.5 arcsec and 95% within 1.0 arcsec, which agrees well with the above result, bearing in mind the addition of the Ross corrector.
From the point of view of telescope optics, the following features of the 200-inch were particularly notable compared with its 100-inch predecessor:
- The primary material (low expansion Pyrex)
- The primary structure (lightweighted)
- The primary relative aperture (f/3.3 instead of f/5.1)
- The primary support (36 supports performing axial and radial functions - see RTO II, Chap. 3)
- The Al reflecting coat
- The use of field correctors in the PF
- A very large dome and building including all facilities
The next section will show what a remarkable advance it was, bearing in mind that all of these features were determined in the 1930's.
5.3 Reflectors after the 200-inch Palomar Telescope up to about 1980
As always, after a major advance, there followed a period of consolidation, prolonged in this case by the aftermath of the second world war.
One of the earliest projects was also the biggest of all: the USSR 6 m telescope in the Caucasus mountains (Mt. Pasthukhov, Zelenchuk). The optical concept essentially goes back to Maksutov in 1952, who proposed a 6 m parabolic primary working at f/4.0. Maksutov considered a steeper primary would give unacceptable problems of aspherising. Now Eq. (5.1) shows that the aspherisation required is a linear function of aperture and an inverse cube function of the f/no; therefore that required for an f/4 mirror of 6 m aperture is only about two thirds of that for an f/3.3 mirror of 5 m aperture. Viewed like this, since the successful results of the Palomar 5 m telescope were known in 1952, the optical layout at f/4 must be seen as cautious. More important was the nature of the 6 m blank. Like Couder in France [5.27], Maksutov had been experimenting with metal mirrors since the thirties [5.3]. Speculum (bronze) is too heavy to be of modern interest, but aluminium, stainless steel and beryllium, among other possibilities, are extremely interesting because of their better thermal conductivity than glass. These possibilities are discussed in detail in RTO II, Chap. 3. Maksutov pleaded for a metal mirror for the 6 m. However, the choice finally was a low expansion glass similar to Pyrex. A meniscus form was chosen with a thickness of 650 mm, without lightweighting. The glass volume was thus three times higher than for the Palomar telescope with correspondingly longer cooling period. The blank manufacture was performed between 1963 and 1968, the weight being 42.7 tons. The optical work was finished in 1974. Careful test results, both with an interferometer and by Hartmann testing, gave 62% geometrical energy concentration in 0.5 arcsec and 91% in 1.0 arcsec, values almost as good as Palomar. The first tests in the telescope were performed in 1975. Since the demands on optical quality had been made more severe for other telescopes of this period compared with the Palomar 5 m (a specification from the thirties), a second borosilicate blank was cast and worked, to higher quality. This was inserted in the telescope in 1979. However, the thermal inertia of a massive glass blank of this size is so large that this set the practical limit of quality achievable. It was intended to replace this second borosilicate blank by a blank in glass ceramic (Russian "Astro-Sitall"), a zero expansion material which had become available about 1970 in the western world (see RTO II, Chap. 3). Apparently, it has not proved possible to cast such a huge glass mass in this material without breakage, which is confirmed by the experience of Schott in Germany (RTO II, Chap. 3). The solution would be a much thinner blank controlled actively (RTO II, Chap. 3), but the 6 m telescope is a concept of an earlier period. I visited the telescope in 1984 and was given information on the quality in a very open and friendly spirit. The three aspects limiting the optical quality are apparently:
- The thermal inertia of the mirror
- The thermal conditions in the dome
- The site, for which the seeing was less good than hoped
The average image quality was quoted as better than 21 arcsec and the best about 11 arcsec.
From an optical point of view, the 6 m telescope (Fig. 5.20) is less advanced than the 5 m Palomar telescope. But from the point of view of mechanics and controls it was epoch-making. This was above all due to B. K. Ioannisiani, the most determined proponent of the Alt-Az mounting [5.3]. The control system to solve the 2-axis tracking problem was the work of N. Michelson and O. Melnikov. This bold decision was a great success and a strange contrast
with the western world where, in spite of a fully developed industrial base in electronics, the courage to take the logical step back to the compact and gravity-symmetrical Alt-Az mount was lacking for a whole generation. For this reason alone, the 6 m will go into telescope history as a great telescope. Furthermore, we must remember that the somewhat mediocre optical quality is compensated by the huge size.
Another, smaller telescope of the post-Palomar epoch should be mentioned: the 3 m Lick reflector. This was commenced in 1946 on the basis of the availability of a 3 m Pyrex blank originally cast in the Palomar 5 m blank development programme (see § 5.2). Originally, it had been intended for an autocollimation test flat, but this was abandoned. The thickness of this blank was such that a relative aperture no faster than f/5 could be made, giving a basic optical geometry and tube format no shorter than those of the Mt. Wilson 60-inch and 100-inch telescopes. The telescope is just big enough for a PF cage. A notable aspect of the optics of this telescope is that the first really complete scientific Hartmann test of a large reflector in function was performed by Mayall and Vasilevskis [5.28] (see RTO II, Chap. 2), following the forerunner work of Bowen [5.29] on the 200-inch Palomar telescope. In their most favourable test at the PF, the final quality of the primary was given by the geometrical energy concentrations 70%, 95% and 97% in 0.34, 0.68 and 1.35 arcsec diameter respectively. This set a manufacturing quality standard of d80 — 0.40 arcsec for the 80% energy concentration which was applied for a whole class of succeeding telescopes.
The optical geometry of a whole generation of telescopes was effectively determined by a notable paper by Bowen in 1967 [5.30]. He considered the optical layout of a telescope at a time when the photographic plate, as the basic detector for nearly 100 years, was being rapidly replaced by photomultipliers for single objects or image-intensifier tubes of various sorts for small fields. The long refractors with f/15 to f/20 were admirably suited to visual observation if supplemented by a relatively cheap battery of eyepieces. The slow photographic plates at the beginning of the century forced the revolution in the speed of primaries because exposure times of galaxies and nebulae for speeds slower than about f/6 were unacceptably long. This dominated Ritchey's thinking and led to his first RC telescope with f/6.8 following his f/5 primaries for the 60- and 100-inch classical telescopes. Bowen's analysis rationalised the known fact that, while the rate of accumulation of quanta of a photographic plate depended only on the aperture, the limiting magnitude, as determined by the information capacity of the photographic plate, is primarily a function of the focal length. For the "unbaked" plates of the time, with an efficiency less than 1%, Bowen concluded that the focal ratio for direct imaging at the Cassegrain focus should be f/8 - f/10. For spectroscopy, which is less critical in the sense that the sky background is so diluted as to become negligible, such a focal ratio was favourable to pixel matching and echelle spectrographs. These f/nos were also suitable for photomultipliers. At the prime focus, he considered the focal length for direct imaging should be about 6-10 m, giving in 100-inch to 200-inch telescopes relative apertures between f/1 and f/4.
Out of this analysis emerged the concept of the Bowen-type telescopes with an optical geometry of about f/3,f/8,f/30
in the prime, Cassegrain and coude foci respectively. The magnification m2 = -8/3 gave reasonable obstruction ratios for normal image positions of the Cassegrain image behind the primary. From Eq. (2.86)
with b = b/f, b being positive and f negative (see Fig. 2.12). With b typically about -0.13, we have Ra — 0.31, giving with normal fields an obstruction from the secondary of the order of 0.35. A PF of f/3 was only 10% steeper than the 200-inch Palomar primary, but few opticians had experience of working large mirrors as steep as this and there was a general reluctance in the 1960's to go much steeper than Palomar.
The Bowen geometry was a big driver for the RC telescope since classical photography with large plates at the Cassegrain focus was considered an essential feature, requiring fields of 0.5° diameter or more. In view of the success of the Ross correctors at the PF of the Palomar 200-inch, interest in PF field correctors was strong and converged out of the work described in Chap. 4 on a general requirement of a 1° diameter field for classical photography. The photographic plate experienced a certain resurgence of interest with the introduction of "baking" plates in nitrogen or forming gas to drive out water and increase the quantum efficiency from less than 1% to possibly 4-5% at maximum. A number of optical forms emerged for a range of telescopes all having apertures in the class 3.5-4 m, thereby large enough to permit a PF cage. The solutions chosen ranged from a classical form with a parabolic primary through a strict RC form to various quasi-RC forms, usually with primaries somewhat more eccentric than that prescribed by the RC form. The motivations were essentially as follows:
a) Classical telescope with parabolic primary
The principal motivation was the use of the naked primary without corrector for direct imaging at the PF. The limitation of field coma is, of course, very severe at f/3. From (3.87) we have for the field coma of the parabolic primary
which can be written in the form
Converting into angular aberration from (3.198) gives with n' = —1 and upr1 expressed in arcsec l 3 1
With the Bowen prescription of N1 = —3.0, Eq. (5.4) gives a field coma of 1 arcsec for a field diameter of 96 arcsec. For a primary with D = 3.5 m working at this relative aperture, the corresponding linear field diameter is only 4-9 mm. Apart from the fact that a coma limit of 1 arcsec is generous by today's standards, such a linear field is very small for practical use and demands very accurate centering of a detector of similar size.
A second motivation often quoted is that the classical telescope is easier to manufacture. This matter is dealt with in RTO II, Chap. 1. With modern methods of manufacture there is a negligible advantage of a parabolic primary over an RC for the primary and only a modest advantage for the secondary: this argument is no longer valid.
So far as correctors are concerned, as was shown in Chap. 4, there is no advantage of the classical telescope, either in the prime or Cassegrain foci.
The only significant advantage is for the coude with its high N, typically f/30 or longer. It should be recalled that the classical and RC solutions converge for an afocal system. It follows that the classical telescope is nearly aplanatic for coude foci with high m2 values, whereas RC solutions have considerable field coma. However, this would not normally be decisive as coude systems work with fixed spectroscopic equipment and small fields.
b) RC solution
This has the obvious advantage of being aplanatic in the Cassegrain focus, giving good field correction over about 0.5° diameter if field curvature is ignored and astigmatism is the limitation.
The over-correction of the primary is a help in the design of PF correctors (see Chap. 4), although the advantage is modest for modern systems with high values of m2.
The coude is less favourable for field coma, but this may not have much weight in the decision (see a) above).
c) Quasi-RC solutions
These were discussed in Chap. 4 and consist in optimizing the Cassegrain focus by departing from the RC form when using a corrector. Normally, the axial performance (spherical aberration) is negligibly affected by the corrector.
Such solutions given optimum performance at the Cassegrain focus with corrector, e.g. with a Gascoigne aspheric plate and field flattener. Without corrector, there is some field coma.
Since the primary asphericity is normally somewhat increased, these solutions are also more favourable for PF correctors.
The coude is slightly worse than the RC solution, but the difference is small.
The telescopes listed in Table 5.2 show examples of all these solutions. The successful manufacture of a number of RC or quasi-RC telescopes of modern design to high optical standards must be seen as a major consolidation of telescope optical technology following the Palomar 5 m, even though all these telescopes were smaller. Equally notable, probably even more so, was the application of zero-expansion or nearly zero-expansion glasses to the optics (RTO II, Chap. 3). In the 1950's and 1960's, the technology of fusing quartz segments (boules) was perfected, the first major blank being the 4 m blank for the Kitt Peak Mayall telescope, made by the General Electric Company. Normal fused quartz has an expansion coefficient of about 5-10-7, about one sixth that of borosilicate glass (Pyrex). Glass ceramic was invented in the 1950's at Corning, but since then Corning has concentrated on fused quartz blanks, above all in the form of ULE quartz (Ultra-Low Expansion) with an expansion coefficient of about 110-7, only about 1% of that of plate glass. Glass ceramic has a coefficient effectively zero, but is slightly temperature dependent. Originally made by Owens Illinois under the name Cervit, it is now made above all by Schott in Germany under the name Zerodur, also in Russia (Sitall) and China. These materials are discussed in detail in RTO II, Chap. 3.
Table 5.2 is very instructive in revealing the trends of the time. The most significant advances in the optics were the switch to very low or zero expansion glasses for the primaries and the successful manufacture of RC or quasi-RC mirror forms. The blanks are all massive as compared with the lightweighted type for Palomar. This was no longer significant for blank distortion from thermal effects, but has often proved disadvantageous from the effects on the local air due to the high thermal inertia of the primaries. The ability to manufacture RC systems successfully was essentially due to null (or compensation) testing introduced by Dall in 1947 [5.46] [5.47], although the basic idea was already published by Couder in 1927 [5.48] but its significance was not then understood. This will be treated in detail in RTO II, Chap. 1.
Also of significance for the optics is the mounting. The reluctance in the western world to follow the Alt-Az development initiated in the 1950's and 1960's for the Russian 6 m telescope has been mentioned above. The retention of equatorial mounts increased the volume and hence the air mass in the domes, thereby giving greater problems with dome seeing. The 4.2 m William Herschel Telescope (WHT) was the second large telescope to have an Alt-Az mount. It also pushed the f/no of the primary to f/2.5. This advance from f/3.3 of Palomar took 50 years.
The relative conservatism of this period was to some extent explained by a legitimate concentration of effort on electronic detectors. The culmination has been the CCD (Charge Coupled Device) which has increased the efficiency compared with unbaked plates by almost 2 orders of magnitude. This has
Was this article helpful?
Compared to film cameras, digital cameras are easy to use, fun and extremely versatile. Every day there’s more features being designed. Whether you have the cheapest model or a high end model, digital cameras can do an endless number of things. Let’s look at how to get the most out of your digital camera.