CEO m

Fig. 4.39. Basic design (schematic) for an FR with intermediate image for a field of 0.9° diameter at the Cassegrain (RC) focus of the 3.5 m MPIA f/3 to f/8 telescope [4.15]. mpR = 1/2.67

in Fig. 4.40. The monochromatic image suffers from appreciable coma which may be correctable with some design modification. Fundamental, however, is the large chromatic difference of coma. The elements of the FR must be disposed over a large axial distance, so that some of them are still a long way from the pupil. The lens thicknesses cause problems as with the system of Fig. 4.37. A further problem is the strong singlet field lens which produces a pupil image with large chromatic aberration. This means that the principal rays of different wavelengths traverse the elements of the FR at different heights, a fatal situation for higher order chromatic aberrations.

We may conclude that, for an FR replacing a PF facility with a 1° field diameter, lens systems, even with an intermediate image, have no chance of fulfilling the quality requirement a) above with regard to the spectral range. If this is to be achieved, the basic power of the FR must come from mirrors, lens elements being only quasi-afocal correctors. Mirrors solve the chromatic problems associated with the power at once, but lead to obstruction problems. In [4.15] a number of mirror solutions were investigated, all using the PF of a supplementary Schmidt-based system. The basic design arrangement is shown in Fig. 4.41. To minimise obstruction problems of the detector at the PF of the Schmidt mirror, a higher reduction factor of mFR = 1/4.71 had to be used (f/8 to f/1.7). The optical correction would have been better with the previous reduction of f/8 to f/3. Figure 4.41 shows a 2-element corrector in front of the image. In some systems, this was simply a normal corrector for the Cassegrain focus; in others, it was effectively part of the FR and gave an uncorrected intermediate image. The systems considered were a singlet field lens with: a simple Schmidt system with shifted pupil; the same

Fig. 4.40. Spot-diagrams for the FR system of Fig. 4.39 [4.15]. The circle is 0.98 arcsec

Fig. 4.41. Basic focal reducer geometry using a Schmidt-based mirror system for the 3.5 m MPIA telescope. Reduction is f/8 to f/1.7 (mfr = 1/4.71) [4.15]

with a doublet corrector (uncorrected); a Bouwers-Maksutov system with a doublet corrector (corrected); a Hawkins-Linfoot system with a doublet corrector (corrected); the same with a doublet corrector (uncorrected); an extra field lens with a Baker-type system using 2 menisci and 1 aspheric plate and doublet corrector (uncorrected). Finally, the field lens (or lenses) were replaced by a field mirror combined with the same Baker-type system with doublet corrector (uncorrected) in order to establish the importance of the chromatic aberrations of the pupil. All of these telescope systems were discussed in some detail in § 3.6.4, the relevant literature references in Chap. 3 being [3.32] [3.38] [3.50] [3.54] [3.56] [3.57]. All results are for curved Schmidt fields. Here we shall reproduce two results, the best system with one field lens and doublet corrector (corrected intermediate image) and the best system with 2 field lenses and doublet corrector (uncorrected intermediate image). Spot-diagrams of these systems are shown in Figs. 4.42 and 4.43 respectively, in each case for a field of 1° diameter and a circle of 0.5 arcsec. In both of these designs, the doublet corrector effectively removes the astigmatism of the telescope system. In the first system, it is a normal corrector correcting the intermediate image; in the second system, it produces an optimum balance of aberrations feeding into the FR.

In the Hawkins-Linfoot design, the chromatic aberration of the concentric meniscus is corrected by a weak afocal doublet in the transferred pupil, with an aspheric to correct the zonal spherical aberration. The secondary spectrum of the afocal doublet corrector gives lateral chromatic aberration curvature of the upper spot-diagram line (Fig. 4.42) and also the additive effect of spherochromatism and secondary spectrum focus error at 1014 nm.

In the Baker 2 meniscus - plate design, there is a meniscus on each side of the pupil and an aspheric plate in the pupil. All three elements are of quartz, enabling chromatic correction with no secondary spectrum. The field lens is split into two since this system is very sensitive to pupil aberration. The secondary spectrum of the transverse colour originates in the 2-glass corrector in front of the intermediate image. This could be suppressed with a single glass type with somewhat greater spacing.

With further design improvements, the system of Fig. 4.43 is capable of 0.5 arcsec performance over the whole field and wavelength range. For a field of 0.7° diameter, the quality is appreciably better. In this sense, the quality aim a) above has been attained. Although a single glass was not used in the above design, it would be possible to do this, eliminating secondary spectrum effects and giving optimum UV-transmission. The weak point is requirement g), that not too many optical elements be required. The system has a total of 2 field lenses, 4 corrector lenses (including 2 of a normal Cassegrain field corrector), an aspheric plate and a spherical mirror. By comparison, the equivalent PF corrector uses 3 lenses with spherical surfaces and suffers less from obstruction problems of the detector.

Fig. 4.42. Spot-diagrams for a focal reducer f/8 to f/1.7 designed for the 3.5 m MPIA telescope for a field diameter of 1o. Doublet corrector (corrected intermediate image), one field lens and a Hawkins-Linfoot camera. Circle = 0.50 arcsec = 14 ^m. Image radius 1031 mm [4.15]

Fig. 4.43. Spot-diagrams for a focal reducer f/8 to f/1.7 designed for the 3.5 m MPIA telescope for a field diameter of 1°. Doublet corrector (uncorrected intermediate image), 2 field lenses and Baker-type camera with 2 menisci and one plate. Circle = 0.50 arcsec = 14 ^m. Plot field for axial spot-diagrams = 10 ^m. Image radius = 1052 mm [4.15]

Fig. 4.43. Spot-diagrams for a focal reducer f/8 to f/1.7 designed for the 3.5 m MPIA telescope for a field diameter of 1°. Doublet corrector (uncorrected intermediate image), 2 field lenses and Baker-type camera with 2 menisci and one plate. Circle = 0.50 arcsec = 14 ^m. Plot field for axial spot-diagrams = 10 ^m. Image radius = 1052 mm [4.15]

It is clear that FR with large angular fields at the Cassegrain focus can only compete with PF correctors at considerable extra cost and effort in the number of optical elements involved and their arrangement with the detector. Concerning the latter, it should be added that a Schmidt-Cassegrain camera solution would also be quite feasible.

4.5.3 Other Cassegrain focal reducers.

For more modest fields and requirements, many proposals have been made. In 1956 Meinel [4.68] published results of an f/2 Cassegrain camera constructed with a field lens at the focus of the 82 inch McDonald telescope (field diameter 20 arcmin) and a standard Leitz Summicron photographic objective placed behind a Tessar type collimator with a filter between them. The pupil is imaged to the correct point in the Summicron. According to Meinel, the huge gain in speed with the camera was attained without image quality loss beyond the normal seeing limit.

An example of successful improvisation under time pressure is the Stockholm FR described by Jorsater [4.69]. This converts a field of 43 mm diameter (5.4 arcmin) at the f/11 RC focus of the 2.5 m Nordic Optical Telescope (NOT) to an f/3.5 image. The field lens is a spectacle lens of 70 mm diameter and f' = 600 mm. The FR is an f/2.5 Konica camera objective. Images of FWHM of 1 arcsec are produced, giving 2 pixel sampling on a CCD detector. An important property is excellent freedom from ghost images.

More ambitious FR using mirrors in various forms have been proposed by Boulesteix et al. [4.70] for Fabry-Perot interferometry, and by Geyer and Nelles [4.71].

Very interesting work was done for the Texas 7.6 m telescope project with a view to replacing the PF by a Cassegrain FR with more modest fields than the 1° discussed above. Meinel et al. [4.72] described a system consisting of a field lens working with a 4-mirror FR to convert the f/13.5 Nasmyth focus to f/3.0, the field covered being 8 arcmin. Figure 4.44 shows the arrangement. The system consists of a modified "Bowen-type" camera using an inverted Cassegrain to increase the aperture, combined with a Gregory to avoid a refractive field flattener near the image. The exit pupil of the telescope is imaged on to the first, convex, mirror of the inverted Cassegrain, which in turn images it on to its Gregory secondary half. The mirrors are respectively spherical, nearly spherical, hyperbolic and elliptical. The great advantage of the system is its compact nature. The authors show spot-diagrams over the 8 arcmin diameter field which are substantially within 0.2 arcsec. There is some chromatic variation due to the field lens. For the modest field transmitted, there is no vignetting and only moderate central obstruction.

MacFarlane [4.73] has also proposed a number of Nasmyth FR designs for the same Texas 7.6 m telescope project. Most of them are based on the INCA (Inverted Cassegrain) type discussed by Rosin [4.74]. These are basically similar to the Meinel proposal of Fig. 4.44 but use only two reflections instead of

Fig. 4.44. Focal reducer designed by Meinel et al. [4.72] for the Texas 7.6 m telescope project. The f/13.5 Nasmyth focus is converted to f/3.0 over a field of 8 arcmin diameter

Fig. 4.44. Focal reducer designed by Meinel et al. [4.72] for the Texas 7.6 m telescope project. The f/13.5 Nasmyth focus is converted to f/3.0 over a field of 8 arcmin diameter

4 - see Fig. 4.45. Rosin gave a solution with spherical mirrors, corrected for spherical aberration if the magnification is correct. MacFarlane gives such a solution with mirrors of equal curvature and a magnification of 1/3.73. If both mirrors are made aspheric, a flat-field anastigmat corrected for all four aberrations is possible. In fact, this system was effectively invented by Schwarzschild (Ref. [3.1] in Chap. 3) for a parallel incident beam, as shown in Fig. 3.8, and abandoned as useless for a normal telescope. In the MacFar-lane design, it was unclear whether a modification could retain anastigmatic imagery and also compensate the additive field curvature of the Cassegrain telescope and positive field lens. He gave a reduction from f/13.0 to f/6.0, using a configuration with the Cassegrain image and field lens slightly to the right of the concave mirror. A reduction to f/3.0 gave a much better performance over a field of 10 arcmin diameter, the spot-diagrams being within 0.2 arcsec. But no solution could approach the field performance of PF cor-

Fig. 4.45. INCA (Inverted Cassegrain) focal reducer proposed by MacFarlane [4.73]

rectors, confirming the conclusion of Wilson [4.15] that this is only possible with more complex systems.

MacFarlane [4.73] also considered the advantage of a field mirror to avoid the problems of chromatic aberrations of the pupil, which Wilson [4.15] had investigated. MacFarlane confirmed a significant advantage and proposed a practical field mirror arrangement as in Fig. 4.46. The Nasmyth M3 is inclined to the axis at about 40° instead of 45°. The concave field mirror is then inclined at about 5° to the axis, which should be acceptable as it is very near the image. The system of Fig. 4.46 shows a simple Schmidt camera as FR, but more complex systems would be possible. The Petzval sum of a concave field mirror is opposite to that of a positive field lens. This would be advantageous with an INCA type FR but Wilson [4.15] pointed out that a Schmidt type FR usually dominates the total field curvature and a field mirror adds further to its effect. So a field mirror is more interesting for INCA or straight Schmidt-Cassegrain FR solutions.

It is most instructive to compare the proposal of MacFarlane of Fig. 4.46 with some of the telescope solutions proposed in § 3.6.5.3. The system of Sasian (Fig. 3.90) achieves pupil transfer by a cylindrical "Nasmyth" mirror instead of a field mirror. But the final f/no is conventional because of the basic geometry. The system of Wilson and Delabre (Fig. 3.89), on the other hand, produces the pupil transfer with a normal "Nasmyth" flat by using an extra mirror. The fourth powered mirror (Mi and M2 being spherical) is replaced by a Cassegrain pair, which is effectively an FR, giving f/2.8. This gives a field quality within 0.2 arcsec over a field diameter of 30 arcmin, compared with MacFarlane's aim of 10 arcmin diameter, and also has the advantage of a steep spherical primary and a spherical secondary. This shows the advantage of building the FR into the basic design of the telescope. The total number

Fig. 4.46. FR concept using a field mirror proposed by MacFarlane [4.73]

of elements is the same as in the MacFarlane proposal, assuming he uses a simple Schmidt without further complication.

The general conclusion on FR, considered as an addition to a normal telescope, still seems to be that they cannot replace a PF corrector giving a field of the order of 1° diameter. They are much more interesting for more limited fields. CCD detectors favour such applications. The commonest FR are simply a non-spectral mode in a general spectrograph.

5 Major telescopes from Lord Rosse to about 1980

5.1 Major telescopes in the speculum mirror epoch to 1865

In Chap. 1 a brief account of the development of the reflector up to William Herschel was given. From 1800 to about 1840, the pendulum of rivalry between the reflector and the refractor swung back in favour of the refractor through the epoch-making work of Fraunhofer in the development of optical glass and its systematic application to refractors such as the Dorpat refractor of aperture 24.4 cm. The fact that refractors of relatively modest size were able, in spite of the limitation of secondary spectrum, to compete with or even excel reflectors with apertures up to 1.22 m (Herschel) was a measure of the two great weaknesses of the reflectors of the time: the poor efficiency through low reflectivity and the problems of mechanical manipulation in such sizes.

Excellent accounts of developments in the nineteenth and early twentieth centuries are given, above all, by Danjon and Couder [5.1], King [5.2] and Riekher [5.3]. Here the intention is only to establish the essential aspects leading to the modern reflector. As pointed out in Chap. 1, the theoretical basis until Schwarzschild in 1905 [5.4] was still the "classical" telescope with a parabolic primary as laid down by Descartes and applied in Newton, Herschel, Gregory or Cassegrain forms.

The further development of the reflector after William Herschel was, above all, due to Lord Rosse (William Parsons) and William Lassell in Great Britain. In the 1830's, Rosse systematically investigated the problems of casting large mirrors in speculum metal. A major problem up to then was the danger of crystallisation through slow cooling. This had forced Herschel to use a very high copper content for his largest mirror with negative consequences for the reflectivity. Some workers tried to solve the crystallisation problem by rapid solidification on a cold iron plate. Rosse showed that this was an illusion as it softened the material and led to an inferior polish [5.3]. By ingenious technical compromises in the cooling procedure, he was able to produce blanks up to 0.9 m in 1839 and 1.82 m (60 inch) in 1842 using an optimum alloy for polishing (68.2% Cu and 31.8% Sn). In parallel, Rosse performed pioneer work in a concept which has now again become ultra-

Fig. 5.1. 90 cm light weighted, built-up blank made by Lord Rosse in 1839 (courtesy Rolf Riekher)

modern (see RTO II, Chap. 3), namely lightweighted built-up blanks [5.1] [5.3]. Fig. 5.1 is reproduced from Riekher [5.3] and shows the construction of a 90 cm blank made by Rosse. The ribs were of a Cu-Zn alloy with the same expansion coefficient as the speculum faceplate. The sectors were soldered together. Rosse could detect no difference in optical performance between the 90 cm massive cast blank and the built-up blank. However, he decided in favour of the massive cast approach for his 6-foot blank. Five blanks were cast, of which the first and last were polished, the other 3 breaking because of uneven temperature in cooling.

Rosse was also the first telescope maker to develop a polishing machine systematically.

Another important advance over the Herschel technology was the first use in a major telescope (1.82 m) of the whiffle-tree support concept invented by Thomas Grubb [5.5] (Fig. 5.2). Three plates support the mirror weight on universal joints at the centre of gravity of the 3 mirror sectors. These, in turn, each support 3 more plates on universal joints. With 4 stages, 81 supports were finally used at the back of the mirror [5.3]. All modern mirror support systems are based either on this principle due to Grubb or the astatic principle due to Lassell discussed below (see also RTO II, Chap. 3).

Rosse's largest 1.82 m primary had a focal length of 16.5 m giving f/9.0, only slightly shorter than Herschel's normal f/ratios. Now from Eq. 3.11 we see that the asphericity for a given conic section defined by the Schwarzschild constant bs varies linearly with the size of a mirror and as the inverse cube of its f/no. For a parabola, the difference from the circle with the same vertex curvature c is simply

Fig. 5.2. Whiffle-tree support system in 4 stages designed by Thomas Grubb for the Rosse 6-foot reflector completed in 1845 (courtesy Rolf Riekher)

Inserting the f/no as N = f '/2y gives with f' = r/2

The above values for the Rosse 6-foot telescope give from (5.1)

Rosse parabolised from the sphere of equal curvature c according to formula (3.11) by flattening the outer parts of the mirror. The modern method abandons the equal curvature reference sphere, so that the radial aspherising function can be freely chosen according to

where the suffixes s and p refer to the reference sphere and parabola respectively. For example, if the two terms in (5.2) are made equal for the full aperture ym, then material must only be removed in a zonal operation. This is further discussed in RTO II, Chap. 1. The amount of material to be removed is reduced and the zonal position of the zero of the function can be chosen at will.

Rosse apparently was the first to introduce zonal testing using masks. He knew that the paraboloid no longer produced theoretically perfect geometrical optical images on axis if a test object were placed fairly near. This effect was first recognised by Herschel and its avoidance requires fulfilling the so-called Herschel condition (see ref. [3.3] and [3.6] of Chap. 3) which is rarely possible in practical optical systems. Rosse therefore calculated the theoretical focus shift for different zones and measured this with his zonal masks.

Rosse recognised the disadvantage of field coma introduced by the Herschel tilted single mirror form (Fig. 1.1) and reverted to the Newton form

gelb Sc'i: f KiiirJlrl'ütr

Fig. 5.3. Lord Rosse's 6-foot (1.82 m) telescope completed in 1845 (courtesy Deutsches Museum, Munich)

gelb Sc'i: f KiiirJlrl'ütr

Fig. 5.3. Lord Rosse's 6-foot (1.82 m) telescope completed in 1845 (courtesy Deutsches Museum, Munich)

with a second reflection, made possible essentially by the better composition and reflectivity of his speculum. Figure 5.3 shows the mounted telescope. Recognising the immense mechanical problems experienced by Herschel with his largest telescope, Rosse prudently renounced the possibility of a mounting permitting general sky coverage. Instead, the huge iron tube swung between walls on trunnion bearings through an angle of 160° along the meridian. By raising or lowering the bearings on each side with wedges, the tube could be inclined ± 12° giving an observation period of about 11 hours for an object near the equator.

This telescope represented a remarkable advance but was not yet a modern telescope. Using visual magnifications up to 1300, limited by the "seeing" in Ireland, Rosse was able to discover the spiral structure of external galaxies (Riekher gives a comparison drawing and photograph in [5.3]) some 75 years before their physical nature was proven. He also found that parts of some spiral nebulae could be resolved into stars [5.6].

About the same time as Rosse, William Lassell also made fundamental advances. His casting techniques and alloys were similar to those of Rosse and equally successful up to his largest size of 1.22 m (4 feet) with f/9.2. Lassell's largest telescope was set up in Malta in 1861. This telescope possessed three notable new features

- A primary mirror support system based on astatic levers (see RTO II, Chap. 3)

- An equatorial mount based on a fork-type equatorial axis

- An open slat tube to permit natural ventilation of the air in the tube

Figure 5.4 shows a reproduction of the original single astatic lever support invented for a 9-inch telescope in 1842 [5.7] and Fig. 5.5 the complete 1.22 m

i

Telescopes Mastery

Telescopes Mastery

Through this ebook, you are going to learn what you will need to know all about the telescopes that can provide a fun and rewarding hobby for you and your family!

Get My Free Ebook


Post a comment