Wow. This one's a keeper for the archives! Thanks again Richard. +++++ On 26-May-05, at 2:17 AM, Richard Knoppow wrote: > [...] Kingslake points out that in > these days of computer design some modern lenses are > difficult or impossible to classify as being derived from > one of the classics. Even some old lenses can be thought of > as either of two designs or maybe more. For instance, the > classic aerial survey lens the Zeiss Topogon, designed by > Robert Richter, four elements all deeply curved meniscus. Is > it a double Gauss lens or is it a compounded Goerz Hypergon? > Well, its really both. The Hypergon, for those not familiar > with it, is an extremely wide angle lens with coverage of > over 130 degrees, designed in 1900 by Emil von Hough, the > designer of the Dagor. This lens consists of two very > steeply curved meniscus elements on either side of a stop. > The elements are very thin and the outer surfaces nearly > form a sphere. The lens has a very flat field and very > little astigmatism, and low coma and distortion due to its > symmetry. However it is not corrected for spherical or > chromatic aberration so can be operated only at very small > stops, less than f/20. The fall of of illumination is even > more than the rule of thumb cos^4 theta so the lens was > equipped with a spinning obstructive stop to even it out. > The Topogon has four elements, the outer ones positive thin > meniscus as in the Hypergon but the inner ones are negative > meniscus lenses as in a double Gauss type. The additional > elements allow it to be corrected for spherical and > chromatic aberration. The Bausch & Lomb Metrogon has an > additional element which further corrects the spherical. > What kind of lenses are the Topogon and Metrogon? As above > they can be thought of as either double meniscus or as > double Gauss lenses. > There are more difficult cases in some modern lenses, for > instance, few zoom lenses can not be classified as being > derived from any of the older types, they are just their own > thing. > What is interesting is to learn how the various > aberrations are corrected in the different types and what > tricks the designers found to correct them. For instance, > one trick used by Bertele in the Ernostar and Sonnar was to > use thick, low index, sections instead of air spaces. The > advantage of this was the elimination of flare while > retaining some of the benifits of the air space. Paul Rudolf > found a way of using a cemented interface to vary the > dispersion of the cemented pair virtually at will without > having any effect on other optical characteristics. He used > this trick, called a "buried surface" in the original Planar > to get the effect of a glass type which was not obtainable. > Bertele uses the same trick in the f/1.5 Sonnar. Another > trick, already mentioned, is the splitting of a strong > element up into two or more weaker elements. Simply > splitting them reduces some aberrations which is helpful > when the angles of incidence in the lens become large as in > very fast lenses or wide angle lenses. Because most of these > tricks can be adapted to any design they are not really a > basis for classifying a lens even though the trick may have > originated with a particular type or be characteristic of > it. > Computer analysis of designs has made a huge difference in > design technique. The method of evaluating a design is the > trace rays of light through it. About three rays are > necessary to get any idea of what its doing. By hand methods > a single ray trace will take perhaps half an hour. If a hand > calculator is used this can be reduced to perhaps five > minutes. Any of the common computer optical design programs > (OSLO, Zemax, etc.,) operating on a fast PC, can make > millions of tray traces in a fraction of a second. Its > possible to get a very complete analysis of a prospective > design very quickly and to derive presentations of the > information which were not practically possible before > computers. The ray tracing is so fast that the computer > program can be set to vary certain parameters to optimise > the design, but as Kingslake and Warren Smith point out the > program can't always tell when it is getting into > impractical areas so it needs human guidance. > This is not to say that all old designs were less than > optimum. Brian Caldwell, a well known lens designer and the > author of the program LensVIEW, says than many of the old > Zeiss designs are so close to optimum that computer > optimisation, even with changes in glass to modern glass, > does not improve them significantly. This is partly due to > very careful calculation but also because the old method of > design was to evaluate the presciption mathematically until > it looked pretty close and then build a model of it. > Optimisation was then done by poking at the actual lens > until it performed as well as could be gotten. > Some advantages of modern design are less significant > than might be thought. For instance aspherical surfaces have > been around for a long time. Zeiss used them in some > experimental lenses included in the survey of lenses called > the Zeiss Index. An asphere can be duplicated by several > spherical surfaces. The advantage of the asphere is > simplification of the lens. Modern manufacturing processes > allow economical production of aspheres. In the past each > one had to be hand figured. > Another advantage of the last sixty years has been the > avialability of glass with very high indices of refraction > and relativly low dispersion, or low index-high dispersion > glass. All glass bends light. The amount it bends is related > to the Index of Refraction. The idex is the bending of light > compared to a vacuum, which has an idex of 1. Actually air > is so close to 1 than it is usually considered to have an > index of 1 except for the most precise work. The index of > refraction is also the ratio of the speed of light in the > medium to the speed in a vacuum. > Now, things would be fine if the value of the index of > refraction a constant. It isn't: it varies with wavelength, > generaly going up as the wavelength decreases. The effect is > known as dispersion. This is why a prism splits white light > into a spectrum. The same effect is produced by simple > lenses. In fact, two prisms, base to base, are an elementary > positive lens. This effect is known as chromatic aberration. > It is corrected by combining a positive element with a given > amount of dispersion with a weaker negative element with > abuot the same amount of dispersion. If the dispersions are > nearly the same they will cancel. However, in order for this > conbination to have any power the positive element (assuming > we want a achromatic positive lens) must have more power > than the negative lens. Practically, this means combining a > positive lens with a high index but relatively low > dispersion with negative lens with lower index but the same > dispersion. Before Abbey and Schott came up with the Barium > glasses known as Jena glass, in the late nineteenth century, > all glasses followed a line where increasing index was > accompanied by increasing dispersion. This meant that > positive and negative lenses had to be assembled in a > certain way to cancel the chromatic aberration. The > disadvantage of this is that the reverse combination of > positive and negative was needed to correct astigmatism. So > that it was impossible (they thought) to make a lens of old > glass that was both chromatically correct and free of > astimatism. As it turned out, this was not true but a > chromatically correct anastigmat of old glass was not > produced until the 1920's by K. Martin of Busch (the Omnar). > Even so the invention of the new type glasses furthered > optical design enormously. The range of glasses was > increased even more by the development of rare-earth glasses > at the United States National Burea of Standards beginning > in the early 1930's and developed by Eastman Kodak in the > late 1930's. > By increasing the index of a glass, especially if the > dispersion can be kept low, the curvature of the surfaces > for a given amount of power can be reduced. Since several > aberrations are proportional in some way to the angle of > incidence of light at the lens surfaces the glass > automatically reduces the aberrations. This allow either an > improvement in performance with a given amount of complexity > or a duplication of performance with a simpler system. > > I've written too much. > > --- > Richard Knoppow > Los Angeles, CA, USA > dickburk@xxxxxxxxxxxxx --- Rollei List - Post to rollei_list@xxxxxxxxxxxxx - Subscribe at rollei_list-request@xxxxxxxxxxxxx with 'subscribe' in the subject field OR by logging into www.freelists.org - Unsubscribe at rollei_list-request@xxxxxxxxxxxxx with 'unsubscribe' in the subject field OR by logging into www.freelists.org - Online, searchable archives are available at //www.freelists.org/archives/rollei_list