Triplets form the mainstay of present day apochromatic designs.  This is because by adding a third element it becomes much easier to manufacture high-performance lenses cost-effectively.  And more combinations of glasses are available to do the job.  Yet even here there are limits as one pushes for faster and faster focal ratios.  If diffraction limited performance and exquisite color correction are demanded at f/7 for a 150mm lens, then it may be necessary to add a fourth element.  And a radical departure from the closely spaced lens arrangements so far seen may be contemplated.  The Petzval lens in which there are two widely separated doublets may be necessary (cf. Chapter 5).  Of the rich variety possible in three-element lens combinations, we will survey a sampling in the present chapter.

H. Dennis Taylor developed and patented the first practical triplet in 1892 [British Patent no. 17994; cf. S. Czapski & O. Eppenstein, Grundzüge der Theorie der optischen Instrumente nach Abbe, 3rd ed. (Verlag J.A. Barth, 1924), p. 567; H. Chrétien, Calcul des combinaisons optiques, 4th ed. (Paris, 1958), pp. 261-267; and G.R. Nankivell, "The Cooke Photovisual Objective and the 22.9cm Refractor at the Carter Observatory, New Zealand," Journal of the Antique Telescope Society 24 (2002), pp. 4-8].  He seems to have been alerted to the possibility of making a triplet through the published writings of Abbe and Hastings [cf. H.D. Taylor, Adjustment and Testing of Telescope Objectives, 5th ed. (Adam Hilger, 1983), pp. 78-87, especially p. 79; and C.S Hastings, "On Triple Objectives with complete Color Correction," The American Journal of Science and Arts, 3rd series, vol. 18 (1879), pp. 429-435].  In 1894, Taylor published a detailed description of his triplet, specifying the glass types, their order in the construction, and the relationships of their surface curvatures.  He also illustrated his lens with a layout [cf. H.D. Taylor, "Description of a Perfectly Achromatic Refractor," Monthly Notices of the Royal Astronomical Society 54.5  (1894), pp. 328-337; reprinted in Adjustment and Testing, pp. 78-87, especially pp. 82-83].  

The glass types in question are all "ordentlich" or "regular" types produced at the time by Schott.  Facing the sky came a light barium flint (O.543); in the middle a borosilicate "short flint" (O.164); and on bottom a light silicate crown (O.374).  O.164 was a replacement for O.658, an earlier borosilicate short-flint which Taylor had specified in his patent.  O.164 allowed somewhat better correction than O.658.  Henri Chrétien, the famous French optical designer, studied Taylor's objective and has left valuable information regarding the refactive indices for all four types of glass [cf. Chrétien, p. 262].  Using Chrétien's indices and Taylor's illustration and description, we can form an approximation of his objective in ZEMAX.  What follows is an approximation following Taylor's second form of objective employing O.164.  This allows a direct comparison with the layout specified in Taylor's article. 

Next comes the layout:



The thinness of the convex central element is intentional, since Taylor's illustrations as well as surviving Photo-visuals show that he purposely employed thin glass [cf. Taylor, Monthly Notices, p. 332; Adjustment and Testing, p. 82; and Nankivell, pp. 4-8].  The ray fan plot and spots come next.  Since this and all subsequent designs are aplanatic, I do not give ran fan plots for the off-axis images:




Optically, the design is very good for a lens not employing fluorite or fluor-crown, and indeed it may be too good.  The color curve supplied by Chrétien [p. 265] indicates that rays shorter than F- should begin to focus longer and longer as in an achromat, so that the g-rays should form a large diffuse (but faint) halo around the Airy disk.  In the above lens, on the contrary g- still focuses well and errs in focusing too short, if anything.  Thus, one needs to take the above design with a grain of salt.  The lens geometry looks right for a Cooke Photo-visual; the radii of curvature look plausible; the fact that at the 80% zone where the entering rays focus most tightly, red and blue simultaneously focus short of green--all these features are correct for a Photo-visual.  But, apparently violet should focus long and it does not.

Cooke and Sons began selling Taylor's Photo-visuals by the mid-1890s, and Taylor himself made a presentation regarding the design in April 1894 to the Royal Astronomical Society [cf. Observatory 213 (April, 1894), pp.132-134].  Two of the largest lenses were constructed almost immediately:  a 9-inch in 1896 for Edward Crossley, a wealthy English amateur, and a 12.5-inch at about the same time for Cambridge University [cf. Nankivell, p. 5; and D.W. Dewhirst, "A Cooke Photovisual Lens in a Compensated Cell," Sky and Telescope 49.1 (1975), pp. 24-25].  They also made a number of other Photo-visuals by 1905 [cf. W.J.S. Lockyer, "Note on the Permanency of some Photo-visual Lenses," Monthly Notices of the Royal Astronomical Society 68 (1908), pp. 19-29].

Unfortunately, all these lenses began to degrade soon after their completion.  Indeed, Sir Howard Grubb and A.C. Ranyard had both expressed concern regarding the permanency of their glasses at the April 1894 meeting.  But Taylor and his employers defended them and expected no decay [cf. Observatory 213 (April, 1894), pp. 147-148].  Yet by 1908 there was ample evidence of glass degradation [cf. Lockyer, "Note on the Permanency..," pp. 19-29].  Irronically, whereas Taylor had most been at pains to allay concerns about the permanency of the borosilicate short-flint, it was the inner surface of the light silicate crown at the rear of the lens which was most affected, and to a lesser extent the inner surface of the front light barium flint [cf. Taylor, Monthly Notices, pp. 333-334; Adjustment and Testing, pp. 83-84; and Lockyer, "Note on the Permanency..," pp. 22ff.].  Yet the short-flint also decayed with the passage of time; and like many other short-flint glasses, if actual droplets of water ever sat on its surface for a length of time, deep corrosion ensued.  Recently, the 9-inch Crossley lens was destroyed in this way after 100 years of service [cf. Nankivell, pp. 6-7].  It is a danger facing many short-flint apochromats.  Even if nothing catastrophic occurred, owners of the early Cooke lenses could expect them to need repolishing every 25 years of so [cf. Dewhirst, p. 25].

The optical success of Taylor's lens spurred Cooke and Sons to search for more stable glasses.  They had already replaced O.658 with O.164.  Next they replaced the ordinary crown O.374 with O.599 [cf. Taylor's reply to Lockyer in Lockyer, p. 29].  Schott too recognized the problem and by 1902 had removed O.374 from the market [cf. H. Hovestadt, Jena Glass and its Scientific and Industrial Applications (McMillan, 1902), pp. 388-393].  For its part, Zeiss recognized the value of Taylor's design and quickly marketed their own similar triplet, called the "B" objective, designed by Albert König [cf. A. Sonnenfeld, "Der Köngische Apochromat B," Zeitschrift für Instrumentenkunde 61 (1941), pp. 261-264; R. Riekher, Fernrohre und ihre Meister, 2nd. ed. (Verlag Technik, 1990), p. 214].  Since Zeiss had such close ties to Schott, they were apparently able from the outset to choose better, more stable glasses than Taylor had used.  At any rate, nothing I can find in the literature suggests that the Zeiss B lenses degraded [cf. also J.G. Baker, "Planetary Telescopes," Applied Optics 2.2 (1963), pp. 111-129, especially  p. 120].

König's student, Horst Köhler, has indicated the modern Schott glass types and geometry of the "B" objective, as well as radii of curvature and spacings for a small version.  From this information it is possible to produce an approximation of the "B" at 150mm f/18, in order to show how the lens would compare to Taylor's [cf. A. König and H. Köhler, Fernrohre und Entfernungsmesser, 3rd ed. (Springer Verlag, 1959), p.134, nr. 9]:
The ray fan plots and spot diagrams come next:


The color correction is not quite as good as Taylor's objective, even apart from the performance at 0.436 micron.  On the other hand, the "Zeiss B"-type has almost no spherochromatism, and only a small amount of zonal spherical aberration, which itself could be removed with suitable aspheric figuring in the outer zones.  Therefore, the lens could be built at a faster focal ratio, f/15 being the target at which Zeiss aimed.

The main problem with these two objectives--in addition to possible glass deterioration--is the strongly curved interior surfaces and the large airgap between the middle short-flint element and the final crown.  These features made the lenses very sensitive to errors in the tilt and centration of the elements relative to one another, as well as to temperature changes.  Even slight alignment errors would produce strong coma in the image.  Taylor and König recognized this and made provision for it in the design of their lens cells, introducing carefully made spacers and retaining rings which expanded and contracted with temperature changes in such a way as to keep the lens elements properly oriented with respect to one another [cf. Taylor, "The Cooke Photo-Visual Objective," in  The Adjustment and Testing of Telescope Objectives, 5th ed. (Adam Hilger, 1983), pp. 54-57, especially p. 55; Sonnenfeld, pp. 262-263; J.G. Baker, "Planetary Telescopes," Applied Optics 2.2 (1963), pp. 111-129, especially  p. 118; and Dewhirst, pp. 24-25].  August Sonnenfeld, the designer of the Zeiss AS objective, in his discussion of the B-type said that "One must...expect from the user that he value the B-objective like a highly sensitive physical measuring instrument...and treat it accordingly."  Obviously, one couldn't throw a "B" into the back of the car and race out to the country for an evening's observing, as one can with a present day apochromat!

The reason for the steep curves in these triplets is because the dispersions of the glasses do not differ sufficiently from one another.  We saw the same problem in Chapter 4a especially in regard to the SSKN8/KzFSN4 doublet.  In part to overcome this problem, Zeiss later developed the "F" or "dense flint" objective.  This became possible during the middle of the 20th century with the advent of extra dense flint glasses which displayed abnormal dispersions.  One such abnormal dispersion flint is called SF11.  The "SF" abbreviation stands for "Schwerflint," that is, "heavy flint" in German.  

SF11 can be combined with another dense flint such as SF1, SF4, SF9, etc. and a dense phosphate crown such as PSK3 or one of the "BaK" or "SK" barium crowns.  Zeiss's "F" objective contained the glasses PSK3, SF4, and SF11, according to H. Köhler who developed it along with R. Conradi [cf. König and Köhler, p. 61; p. 135, nr. 12; and p. 139].  We can form an approximation of the "F" at 150mm and f/15 as follows:
Although the color correction seen here is not as good as in a Taylor or "Zeiss B"-type of apochromat, it is still respectable in comparison to an achromat.  And since the interior curves of the "F" are weaker than in a Taylor or Zeiss B, it is possible to build the system even faster, as Zeiss did, producing the "F" as a half-apochromat down to about f/11 [cf. König and Köhler, p. 61; p. 135, nr. 12; and p. 139].  Other combinations of dense flints and crowns can give better color correction, but the present set has the advantage of significantly weakening the curves, which eases fabrication, mounting, and sensitivity to tilts and decenters of the elements [cf. Baker, p. 119 on the possibility of a superachromatic dense flint triplet].

The next step forward in the design of practical apochromatic triplets for amateur astronomy came when it was realized that although cementing closely spaced large lens elements was not possible, filling the gaps with a very thin layer of oil (or special gel), and sealing the edges of the lens was feasible.  I do not know when this realization occurred.  Certainly, James G. Baker, the famous American optical designer, understood it by 1963, when he suggested building a 510mm f/30 oil-spaced half-apochromat [cf. Baker, p. 125].

At present, a common method of sealing the edges of oil-spaced lenses is by means of polyimide pressure-senstive tape, which goes by the tradename of "Kapton."  This tape is very effective at sticking to the glass despite the oil, and forms a durable long-term barrier against leakage.  And as for spacing oils, many different types can work (including plain cooking oil), but ideally one would like an oil or gel which matches the index of the glasses involved and evaporates very slowly.  A great number of oil-spaced triplets have been manufactured over the last 20 years and I personally know of lenses over 10 years old which show no leakage of oil or deterioration.  So oil-spacing can be considered a permanent or at least long term solution.  If the oil layer ever becomes dirty or damaged in some way, it is possible to separate the lens elements, renew the oil, and retape the lens without harm to the glass.

By means of oil-spacing and taping it was possible for Roland Christen in the early 1980s to revolutionize amateur interest in high-performance apochromats [cf. R. Christen, "An Apochromatic Triplet Objective," Sky and Telescope (Oct., 1981), pp. 376-380; "Revised Triplet Design," Sky and Telescope (April, 1982), pp. 411-412; and "Design and Construction of a Super Planetary Telescope Objective," Telescope Making 28 (Fall, 1986), pp. 20-23].  Christen's original lens designs were essentially oil-spaced versions of Taylor's, using higher quality recent glasses.  As Christen noted at the time, the oiled interior surfaces of such lenses need not be figured accurately, since they contribute essentially nothing to the wavefront errors [cf. Sky and Telescope (April, 1982), pp. 411-412].  Moreover, oiling makes the tilt and decenter problems of the old Cooke and Zeiss triplets completely disappear.  Thus it becomes comparatively easy to fabricate high-performance apochromats, especially in the small sizes popular today.  

I myself have built several triplets and found them among the easiest optics to make accurately.  In two instances it was not necessary to do any figuring at all:  the lenses were complete as soon as they had been carefully polished out.  The first of these lenses, a 90mm f/12 short-flint triplet won a merit award for optical excellence at the 2002 RTMC Astronomy Expo; and the second, a 140mm f/12 ED triplet is equally good in color correction and figure.  Both have smooth wavefronts of 1/8th wave pv (at 532nm) or better.  Nothing more than the simplest lens cells has been necessary for either of the lenses.

What follows is a short-flint triplet closely modeled on those published by Christen in the above-mentioned articles:

This is an elegant symmetrical design, much faster in focal ratio than any of the preceding lenses.  It proved possible to increase the speed because the design uses a short-flint glass (KzFS1) of more abnormal dispersion than those seen above.  Indeed, Christen stated at the time he was building his first lenses that his batch of KzFS1 was even better in its partial dispersions than the catalog values [cf. Sky and Telescope (Oct., 1981), p. 378].  So, the ray fan plots and spot diagrams given below are no doubt worse than what he actually achieved.  Yet they give an idea of the excellence of the corrections: