© Erwin Puts
Back to the table of content

7 Appendices

7.1 Appendix A: Serial numbers and production dates.

The study of the Leitz archives is a fascinating experience, but also a most humbling one. You should not expect, that you will find a neat and clear-cut listing of all products and their serial numbers. In fact the records are not as systematic as you would hope for. The general idea is simple. At first a new lens is designed and gets an internal identification. Then a few handmade prototypes are machined. Sometimes these prototypes get a real serial number, but often they get their own numbers, like 000123. When the prototype has passed all tests, there are two possibilities: the lens goes straight into the production cycle or a small batch is produced to simulate the manufacturing stage in order to study the feasibility of the production. Sometimes the lens needs some more fine tuning optically and this small batch is used for taking photographs by the testing department or selected outside photographers. Whatever the case, somewhere during this stage a range of serial numbers is reserved. The first batch might already be given these numbers, but that is not necessarily the case. The date, assigned to the serial number range is the reservation date, not the real production date. The lists of Leica lens numbers (and bodies) you see in several Leica books and in this book too are based on a big tome in folio format, that is used since 1933 and updated till today, which does register a date, a batch of serial numbers and a lens name. As example: September, 10, 1957, from 1535001 to 1537000, Summarit ( + a code). The date is the date that the serial numbers have been reserved, that is put in the big book. It does not tell you when actual production started and it does not even give information if this batch of 2000 Summarit lenses has been manufactured in one or more runs. It might be the case that the last numbers of this series have been assembled somewhere in early 1958. You cannot be sure if all the allocated numbers have been produced. It is likely, but it cannot be found in this list.

This document is updated by the manufacturing department, specifically that part where the serial numbers are engraved as they are responsible that numbers are not engraved twice or not at all. A second series of books have been kept by Leitz, at least till about serial number 2.100.000, which are the so-called 'Verkaufsbücher' (sales records). In these books you will find a listing of all serial numbers, consecutively numbered one per line and every line shows the serial number, the designation of the lens, the date it has been sold and the name of the person, or company where it went to. There are a number of gaps however in these listings where the serial numbers are empty. The question then is, what happened here. Did Leitz not produce these specific lenses, or have they been produced but not sold, that is used internally. Sometimes such empty numbers have been used to designate a special prototype as is the case with the Summaron 3.5/35mm. Most historians set the production date of this lens in 1946, but the engraving department notes that the first allocation is recorded in late 1948. What happened is this: Leitz stopped making lenses in 1944 with serial number 594852, incidentally an exotic lens, the Summar 1:0.85/150mm. When the war was over and Leitz resumed production in the summer of 1945, they started with a clean number, at # 595000. The first batch of serial numbers for the Summaron is allocated in 1949 In the sales records you see three numbers: 594853, 594859 and 594860, with designation Summaron and dates between 11-07-1946 and 11-05-1948. If you look carefully at the names of the recipients, the puzzle is solved. The three names are heads of the optical and mechanical departments and Mr Leitz himself. These three lenses are prototypes and have been duly registered as being handed out to Leitz personnel. The true production date of the Summaron then is 1949 and not 1946. The early batch of Summarex lenses can illustrate the problems with identifying lenses and fixing dates.

The first batch of these lenses starts with #593001, early 1943 and the first 100 (with designation 8,5cm)have been sold to the German Army. The next 400 lenses (the 'B'' type with designation 9cm), starting with 595101, have been sold to the public, but after the war. These 400 are not exclusively reserved for the 'B'type however and sometimes you find a non-B type. The sales records have dates till 1950 for this first batch of 400 lenses. This does indicate that the person in Havana, who has ordered a Summarex in 1949, gets a lens produced in 1943. Or it might also be the case, that Leitz did not manufacture all these Summarex lenses in 1943, but had spare numbers, which were filled with a production run in 1948. Both example are in themselves trivialities, but it indicates that the true history of the Leica products has yet to be written. And it specifically cautions you to be careful when studying published figures and dates too closely. The serial number list you find here give the numbers as reserved and dated by the engraving department. But it is clear that numbers allocated in the last week of a year, will have been manufactured early in the next year. The figures for 1943 and 1944 must be seen as indications as a cross check of documents do reveal differences, being even more problematic as they are handwritten and overwritten and stricken out and new numbers written in.

1934 195001 236000 41000

1935 236001 284600 48600

1936 284601 345000 60400

1937 345001 416500 71500

1938 416501 490000 73500

1939 490001 540000 50000

1940 540001 566000 26000

1941 566001 582250 16250

1942 582283 593000 10718

1943 593001 594750 1749

1944 594751 594852 101

1945 595000 601000 6001

1946 601001 633000 32000

1947 633001 647000 14000

1948 647001 679000 32000

1949 679001 756000 77000

1950 756001 840000 84000

1951 840001 950000 110000

1952 950001 1051000 101000

1953 1051001 1124000 73000

1954 1124001 1236000 112000

1955 1236001 1333000 97000

1956 1333001 1459000 126000

1957 1459001 1549000 90000

1958 1549001 1645300 96300

1959 1645301 1717000 71700

1960 1717001 1827000 110000

1961 1827001 1913000 86000

1962 1913001 1967100 54100

1963 1967001 2015700 48700

1964 2015701 2077500 61800

1965 2077501 2156300 78800

1966 2156301 2217200 60900

1967 2217201 2254400 37200

1968 2254401 2312750 58350

1969 2312751 2385700 72950

1970 2385701 2468500 82800

1971 2468501 2503100 34600

1972 2503101 2556550 53450

1973 2556551 2663450 106900

1974 2663451 2731921 68471

1975 2731922 2761150 29229

1976 2761151 2809400 48250

1977 2809401 2880600 71200

1978 2880601 2967550 86950

1979 2967251 3007150 39900

1980 3007151 3087000 79850

1981 3087001 3160500 73500

1982 3160501 3249100 88600

1983 3249101 3294900 45800

1984 3294901 3346200 51300

1985 3346201 3383200 37000

1986 3383201 3422890 39690

1987 3422891 3455870 32980

1988 3455871 3481900 26030

1989 3481901 3503150 21250

1990 3503151 3540467 37317

1991 3540468 3583830 43363

1992 3583831 3610679 26849

1993 3610680 3644475 33796

1994 3644476 3677030 32555

1995 3677031 3730290 53260

1996 3730291 3770929 40639

1997 3770930 3818624 47695

1998

1999

2000

2001

 

7.2 Appendix B: all Leica lens designs

 

Design	Specs	Name	Year of announce	Serial number
Zeiss	8.0/15	Hologon	1972	5.474.xxx
Zeiss	3.5/15	Super-Elmar-R	1980	3.004.101
Minolta	2.8/16	Fish-Eye-Elmarit-R	1975	2.682.801
ELC	2.8/19	Elmarit-R(1)	1975	2.735.951
Solms	2.8/19	Elmarit-R(2)	1990	3.503.151
Schneider	4.0/21	Super Angulon(1)	1958	1.583.001
Schneider	3.4/21	Super Angulon(2)	1963	1.967.101
Schneider	3.4/21	Super-Angulon-R(1)	1964	2.056.001
Schneider	4.0/21	Super-Angulon-R(2)	1968	2.283.351
ELW	2.8/21	Elmarit	1980	2.993.701
Solms	2.8/21	Elmarit ASPH	1997	3.796.510
Minolta/ELW	2.8/24	Elmarit-R	1974	2.718.151
Solms	2.8/24	Elmarit ASPH	1998	3.737.201
ELW	6.3/28	Hektor	1935	250.001
ELW	5.6/28	Summaron	1955	1.231.001
ELW	2.8/28	Elmarit(1)	1965	2.061.501
ELC	2.8/28	Elmarit(2)	1972	2.314.801
ELC	2.8/28	Elmarit(3)	1979	2.977.551
Solms	2.8/28	Elmarit(4)	1992	3.585.865
Solms	2.0/28	Summicron-M ASPH	2000	n.a.
ELW	2.8/28	Elmarit-R(1)	1970	2.440.001
Solms	2.8/28	Elmarit-R(2)	1994	3.664.831
Schneider	2.8/28	PC Super-Angulon-R	1988	3.470.571
Schneider	4.0/35	PA-Curtagon-R	1969	2.426.201
ELW	4.5/35	Elmar	1935	n.a.
ELW	3.5/35	Elmar	1930	n.a
ELW	3.5/35	Summaron	1948	706.001
ELW	2.8/35	Summaron	1958	1.615.001
ELW	2.8/35	Elmarit-R(1)	1964	1.972.001
ELW	2.8/35	Elmarit-R(2)	1979	2.928.901
ELC	2.0/35	Summicron-M(1)	1958	1.630.501
ELC	2.0/35	Summicron-M(2)	1969	2.307.451
ELC	2.0/35	Summicron-M(3)	1969	2.312.751
ELC	2.0/35	Summicron(4)	1980	2.974.251
Solms	2.0/35	Summicron ASPH	1996	3.767.100
ELW	2.0/35	Summicron-R(1)	1972	2.402.001
ELC	2.0/35	Summicron-R(2)	1976	2.819.351
ELC	1.4/35	Summilux	1961	1.730.001
Solms	1.4/35	Summilux aspherical	1988	3.459.071
Solms	1.4/35	Summilux ASPH	1994	3.636.101
ELW	1.4/35	Summilux-R	1984	3.271.401
ELW	2.8/40	Elmarit-C	1973	2.512.601
ELW	2.0/40	Summicron-C	1973	2.507.601
ELW	3.5/50	Anastigmat/Elmax	1924	n.a
ELW	3.5/50	Elmar -1	1925	104.xxx
ELW	3.5/50	Elmar -2	1930	125.xxx
ELW	3.5/50	Elmar -3	1954	1.140.016
Solms	3.5/50	Anastigmat	200??	n.a.
ELW	2.8/50	Elmar	1957	1.402.001
Solms	2.8/50	Elmar-M	1994	3.668.031
ELW	2,5/50	Hektor	1931	92..xxx
ELW	2.0/50	Summar	1933	167.001
ELW	2.0/50	Summitar	1939	487.001
ELW	2.0/50	Summicron collaps.	1953	920
ELW	2.0/50	Summicron -2	1957	1.400.001
ELC	2.0/50	Summicron -3	1969	2.269.251
ELC	2.0/50	Summicron -4	1979	2.909.101
ELC	1:2/50	Summicron-R -1	1964	1.940.501
ELC	1:2/50	Summicron-R -2	1976	2.777.651
Schneider	1.5/50	Xenon	1936	270.001
ELW	1.5/50	Summarit	1949	820.001
ELW	1.4/50	Summilux -1	1959	1.640.601
ELC	1.4/50	Summilux -2	1961	1.844.001
ELW	1.4/50	Summilux-R -1	1969	2.411.021
Solms	1.4/50	Summilux-R -2	1998	3.797.910
Solms	1.4/50	Summilux-M ASPH	2004	n.a.
ELW	1.2/50	Noctilux	1966	2.176.701
ELC	1.0/50	Noctilux	1976	2.749.631
ELW	2.8/60	Macro-Elmarit-R	1972	2.497.101
ELW	3.5/65	Elmar	1960	1.697.001
ELW	1.9/73	Hektor	1931	96.xxx
Solms	2.0/75	Apo-Summicron ASPH	2005	n.a.
ELC	1.4/75	Summilux	1980	3.063.301
ELC	1.4/80	Summilux-R	1980	3.054.601
ELW	1.5/85	Summarex	1942	541.053
ELW	4.0/90	Elmar	1931	n.a
ELW	4.0/90	Elmar collaps. (nullseries)	1954	633.001
ELW	4.0/90	Elmar collaps.	1954	1.010.001
ELC	4.0/90	Elmar, 3element	1964	1.913.001
ELW	4.0/90	Elmar-C	1973	2.505.101
Solms	4.0/90	Macro-Elmar-M collaps.	200?	n.a.
ELW	2.8/90	Elmarit	1959	1.585.001
ELC	2.8/90	Tele-Elmarit	1964	2.001.001
ELC	2.8/90	Tele-Elmarit-M	1974	2.585.501
ELC	2.8/90	Elmarit-R -1	1964	1.965.001
ELW	2.8/90	Elmarit-R -1	1984	3260001
ELW	2.8/90	Elmarit-M	1990	3.462.071
ELW	2.2/90	Thambar	1935	226.001
ELW	2.0/90	Summicron -1	1957	1.119.001
ELC	2.0/90	Summicron -2	1959	1.651.001
ELC	2.0/90	Summicron-M -3	1980	3.163.007
Solms	2.0/90	Apo-Summicron-ASPH	1998	3.815.625
Solms	2.0/90	Apo-Summicron-R ASPH	200?	n.a.
ELC	2.0/90	Summicron-R	1969	2.400.001
ELW	4.0/100	Macro-Elmar-R	1968	2.279.851
ELW	2.8/100	Apo-Macro-Elmarit-R	1988	3.412.891
ELW	6.3/105	Elmar	1932	n.a
ELW	2.5/125	Hektor	1954	1.051.001
ELW	4.5/135	Elmar	1931	n.a
ELW	4.5/135	Hektor	1933	172.001
ELC	4.0/135	Elmar	1960	1.733.001
ELW	4.0/135	Tele-Elmar	1965	2.046.001
Solms	3.4/135	Apo-Telyt-M	1998	3.838.125
ELW	2.8/135	Elmarit-M -1	1963	1.957.001
ELW	2.8/135	Elmarit-M -2	1964	2.151.551
ELW	2.8/135	Elmarit R(1)	1964	1.967.001
ELC	2.8/135	Elmarit-M -3	1968	2.404.001
ELC	2.8/135	Elmarit R(2)	1973	2.655.901
ELW	4.0/180	Elmar-R	1976	2.785.651
ELC	3.4/180	Apo-Telyt-R	1975	2.748.631
Schneider	2.8/180	Tele-Elmarit for M	1965	2.082.501
ELW	2.8/180	Elmarit-R	1968	2.161.001
ELW	2.8/180	Elmarit-R	1980	2.939.701
Solms	2.8/180	Apo-Elmarit-R	1998	3.798.410
Solms	2.8/180	Apo-Elmarit-R -2	200?	n.a.
Solms	2.0/180	Apo-Summicron-R	1994	3.652.221
ELW	4.0/200	Telyt-V	1959	1.710.001
ELW	4.5/200	Telyt	1935	230.001
ELC	4.0/250	Telyt-R	1971	2.406.001
ELC	4.0/250	Telyt-R	1980	3.050.601
Solms	4.0/280	Apo-Telyt-R	1993	3.621.833
ELW	2.8/280	Apo-Telyt-R	1984	3.280.401
ELC	4.8/280	Telyt-V	1961	1.850.001
ELC	4.8/350	Telyt-R	1980	2.991.151
ELW	5.0/400	Telyt	1955	1.366.001
ELW	5.0/400	Telyt	1936	332.001
ELW	5.6/400	Telyt	1966	2.212.101
ELW	6.8/400	Telyt-R	1971	2.370.001
Solms	2.8/400	Apo-Telyt-R	1992	3.569.973
Minolta	8.0/500	MR-Telyt-R	1980	3.067.301
ELW	5.6/560	Telyt-(R)	1966	2.212.301
ELW	6.8/560	Telyt-(R)	1972	2.411.041
ELW	6.3/800	Telyt-S	1972	2.500.651
Solms	2.8/280	Apo-Telyt-R module	1996	3.754.626
Solms	2.8/400	Apo-Telyt-R module	1996	3.754.626
Solms	4.0/400	Apo-Telyt-R module	1996	3.754.626
Solms	4.0/560	Apo-Telyt-R module	1996	3.754.626
Solms	5.6/560	Apo-Telyt-R module	1996	3.754.626
Solms	5.6/800	Apo-Telyt-R module	1996	3.754.626
Solms	4.0/35-70	Vario-Elmar-R	1997	3.773.930
Solms	4.0/28-35-50	Tri-Elmar	1998	3.753.126
Solms	4.0/28-35-50	Tri-Elmar -2	200?	n.a.
Minolta	3.5/35-70	Vario-Elmar-R	1983/8	3.171.001
Solms	2.8/35-70	Vario-Elmarit-R ASPH	1998	3.812.110
Sigma	3.5-4.5/28-70	Vario-Elmar-R -1	1990	3.525.796
Sigma	3.5-4.5/28-70	Vario-Elmar-R -2	1997	3.787.860
Minolta	4.0/70-210	Vario-Elmar-R	1984	3.273.401
Minolta	4.5/75-200	Vario-Elmar-R	1978	2.895.401
Minolta	4.5/80-200	Vario-Elmar-R	1974	2.703.601
Solms	2.8/70-180	Vario-Apo-Elmarit-R	1995	3.697.501
Solms	4.0/80-200	Vario-Elmar-R	1996	3.698.001
Solms	4.2/105-280	Vario-Elmar-R	1997	3.790.510

 

 

7.3 Appendix C: best image quality

Leica lenses are capable of very high image quality, but only if all the elements of the picture taking process are controlled and tuned to maximise the image clarity. In an ideal world a photograph would be a faithful and accurate reproduction of the subject. The topic of image degradation could be expanded easily to a full book, and it may seem rash to treat the subject in a few pages. A few very important guidelines may be helpful. It is an almost perverse fact of life that a better lens will demonstrate our technical shortcomings most clearly. A slight defocusing error, some camera vibration, a shade of overexposure will be magnified more when the lens performance is higher, This is quite logical as defects are recorded with greater precision too. The clean and crisp drawing of the main subject outlines and the precise recording of the very fine subject details (textures and minute details) are trademarks of the current Leica lenses. The reduction of flare and the containment of halo rims around specular highlights add to the impression of sparkling clarity.

Leica lenses are capable of recording fine image structures up to 100 lp/mm, that is details that on the film area occupy 5 micron of space. Any defocus blurs, scattering of light (halation) due to over exposure and movement during exposure will degrade these fine structures significantly. But also the high edge sharpness of the subject outlines will be degraded by these effects.

 

7.3.1 Film selection.

This is one of the most intensely discussed topics in Leica circles. The choice of film is closely related to the size of the final image carrier. The performance of the Leica lenses is most evident when big enlargements are made from a negative or projected from a slide. The small details of 5 micron are invisible to the eye when a negative is enlarged only 5 times (that is 13x 18cm). Under ideal circumstances the eye can discern details that are equivalent to 6 lp/mm. The finest details that the Leica lens can record, are visible to the unaided eye only when we enlarge at least 15 times. Or use slides. I am convinced that the perennial discussion among Leica users whether lenses of the older generation are as good as or better than the current ones, is as yet not over, because the evidence used to support the statement is not able to exploit and show the best features of the current lenses.

 

7.3.1.2 Slide film.

After many years of testing lenses and films, I can testify that the current slide films of ISO25 to ISO100/200 are the best medium to enjoy Leica pictures and their optical performance. A large scale projection at a screen of 3 meters wide however will show every defect and technical shortcoming of the photographer and may be a humbling experience at first. I personally see it as a challenge to improve my expertise in this area and a strong visual exposure to the effects of image degradation are very helpful if improvements are to strived after. Generally speaking, films from all reputable manufacturers can be used, as long as the ISO value is up to or below 100. For purposes of optimalization of the recording capacity and disregarding the colour characteristics, I find emulsions with a tight but small grain pattern and high MTF values from 1 to 20 lp/mm to be the best to use. I am using always Kodak films, Kodachrome and Ektachrome, depending on the circumstances, because these films suite me well. I have also extensively used all Fuji and Agfa films for comparison purposes, and can also recommend the Agfa RSX II with ISO50, and the Fuji Velvia ISO50 and Provia ISO100 series. In my practical shooting I use the Kodak E200, Kodachrome 200, Kodachrome 64 en 25, Ektachrome 100 SW and Ektachrome 100VS. This is not a film test, but the choice of films gives me the opportunity to look at the parameters with which film type the qualities if current Leica lenses can be exploited. The E200 ISO film has remarkably fine grain for its speed and certainly when compared to the K200. Still the K200 pictures bring out more of the image potential of the VAE than the E200 films. Why? The finer grain (or more precise: the very small and closely packed clouds of dye) and the lower character curve of the E200 make it difficult or impossible to see the very fine details. The K200 on the other hand while preserving the edge contrast and the crispness of the Leica imagery, has large grain, which reduces the richness of the texture and colour hues of fine detail. The 100SW and 100VS have a higher inherent contrast and are better suited to preserve the VAE performance. The dye clouds are indeed very fine and closely packed, but as with the E200 series this characteristic still reduces the ability for a crisp rendering of fine details somewhat. The K25 and K64 establish again their reputation as the sharpest slide film.

These films, coupled with Leica lenses produce imagery of the highest order, that is as yet still unbeatable. The colour rendition of the Kodachrome series is completely different from that of the 100 SW and 100VS. The excellent preservation of high light gradation and microcontrast in the white (that is overexposed) areas, effectively enhance the subject brightness range that can be recorded. The inherently higher contrast of the 100SW/VS can be exploited without washed out white or whitish areas. The Kodachrome films would be my first choice when I need to explore image quality 'a bout de souffle'. The long process cycle in many countries might put off some users. The E100 family then would be alternative choice to sample the quality of the Leica lenses. The E100 images at 30 times enlargement show crisply rendered extremely fine detail with that famous Leica clarity. Apo correction of course helps the clarity and textural gradation of very small colour patches.

 

7.3.1.3 Black and white films

For best image quality films in the category of ISO 100 and below are the first choice. My approach here is to go for the film first and then choose a developer. The quest of finding the best match between any film and developer will generate hundreds of suitable combinations. I tested many of these combinations and I must say that the influence of the developer on the final result is less important than the other variables, like accurate focus, good exposure and so on. Only when all variables are tightly controlled, the role of the developer might become crucial. Films I use are the Kodak Technical Pan, Agfa APX 25 and Ilford 100Delta.

For these films I use Paterson FX-39, but you will be very happy with Technidol for the Techpan and Agfa Rodinal or Ilford DD-X for the other films. My strategy is to use only a few well chosen-developers as I then know what their character is when using different films. Leica photography should be performed under all circumstances and often an ISO 400 or even ISO3200 (to be exposed as 1600 at most for good shadow detail) is needed. In the ISO400 range I use the Ilford 400Delta if I can carefully expose. Otherwise the Ilford HP5 and Kodak Tri-X are excellent choices. The personality of these films is different and here a wide choice is a must. Never stick to one film, use every film and learn about its character. Stick to one developer to keep things manageable. A secret favourite of mine is the classical Kodak Plus-X (ISO125), with a tight grain pattern and a beautiful tonal scale. If you would have a start I recommend as developer for all films (exception Techpan) the classical Kodak D76 or newer Xtol or the Ilford DD-X. These developers can handle almost every film on the market, with a very good balance of speed, granularity and acutance. Always use the nominal ISO value and never push your films more than half a stop. If film speed is an issue, use the 3200 films or go to colour negative films.

 

7.3.1.4 Colour negative films

I am not happy with colour negative films in general. The extremely fine dye clouds and the bad processing kill all inherent image quality. If you can find a good lab that does manual enlargements for a reasonable price, you might try any one of the current ISO160 films (Fuji, Konica and Kodak). The only area where colour neg is a viable and interesting alternative is the high speed film. The new Kodak Supra and Portra 800 films can be exposed without bad side effects at EI 1600 and deliver excellent image quality. These films have more exposure latitude than their monochrome comrades and will let you use your Leica in challenging situations where photography is exciting and Leica lenses are at an advantage.

 

7.3.2 Shutter speed and tripod use.

It is a truism that any movement of camera and subject will degrade image quality.

For really outstanding picture quality a tripod is a must, even when using a 50mm lens. The classical rule that the lowest possible shutter speed for handheld picture taking is the reciprocal of the focal length. is nonsense. I shot thousands of pictures at a range of shutterspeeds from 1/4 to 1/8000 with all types of lenses. Statistically it is not possible to get fine imagery below 1/250 (big chance factor is involved when shooting that slow). At 1/250 to 1/500 the chances of a good quality picture are higher but it is not fully secure. Above 1/1000 (M-system) and certainly at speeds of 1/2000 and 1/4000 (R-system) the true image potential can be enjoyed. There is a tendency among many Leica users that a tripod is anathema for true Leica photography. This is a bad proposition. A tripod has to be used when it is needed: slow film, slow shutter speeds, small apertures and/or maximum resolution and sharpness. If you take pictures with the camera only supported by your own body, the highest shutter speed that is useable should be selected. There is no need to set the aperture to 5.6 or higher and often it is better to use a wider aperture as image quality is higher at these apertures. The only rule of the game is this: when you select shutter speeds below 1/60 and often you have to) make a rapid series of pictures.

The chance factor will work to your advantage: the chance that one picture is degraded because of camera motion is reduced by a large factor if you have more pictures to choose from.

 

7.3.3 Accurate focusing.

This is a most important topic. (For accuracy calculations, see appendix 7.5). There is only one sharpness plane and that should be located exactly in the solid space of the subject where you want it. Zone focusing, or the hyperfocal distance or the depth of field scales are all approximations, based on the circle of least confusion. With bigger enlargements, the depth of field is automatically reduced. Always focus with the rangefinder or ground glass on the subject plane you want to be critically sharp. If you need the horizon to be sharp, set the lens at infinity and do not use the depth of field scales to artificially extend your range. At closer distances (1 to 2 meters and sometimes even more) another phenomenon will be visible sometimes. That is the effect of the focus shift. When a lens is stopped down the rays from the outer zones are cut off and we get a narrower bundle of rays. As the bundle is narrow, the blur circle is also of smaller diameter and we get more depth of field. But at the same time, the plane of best focus also shifts a little. This can be seen with all high speed lenses of longer focal length, that is from 50mm. The Noctilux is an example: when you take a picture at 1.5 meter at full aperture and then another one, stopped down to 1:5.6 the original focus plane might be closer to the camera. It looks as if you have focused on the wrong plane. The rangefinder cam coupling is calibrated for the wider apertures. When you focus on the eyes of a model and stop down to 5.6 to get enough DoF, and look at the picture, you might be under the impression that you missed the correct plane of sharpness or that your rangefinder is inaccurate. If fact neither happened, but you see the effect of the focus shift. This phenomenon is hardly visible, but under critical circumstances it might.

 

7.4 Appendix D: groups of focal lengths

The computed focal length is the theoretical focal length that will have a certain variation in the production process as some tolerances must be accepted. Leica acknowledges this fact and use a coding system to give the actual focal length. In the past this procedure was valid for M and R lenses, but after some period, the adjustment of focal length for the R-lenses could be abandoned. On many M-lenses you will see a number inscribed after the infinity mark. The meaning of this number is as follows.

 

Nominal focal Length	Code Number	Actual focal length
Elmar 50mm	2	50.1
	3	50.4
	4	50.7
	5	51
	6	51.3
	7	51.6
	8	51.9
Other 50mm	0	50
	10	51
	11	51.1
	13	51.3
	15	51.5
	17	51.7
	19	51.9
	22	52.2
75mm	47	74.7
	50	75
	53	75.3
	56	75.6
90mm	95	89.5
	0	90
	5	90.5
	10	91
135mm	45	134.5
	50	135
	55	135.5
	60	136

 

 

7.5 Appendix E: the rangefinder accuracy

The accuracy of the distance measurement is a very important parameter in the quality of the image. The M and R systems are very different here. The M-range finding is based on a separate range finding mechanism, that is mechanically coupled to the lens movement. This coupling must be of high accuracy and precision as the two main movements (focusing ring on the lens and movement of the rangefinder patch) are invisibly and mechanically linked. The focusing method of the SLR is completely different. Here we focus through the lens on a ground glass screen and there is a direct visual check if we have focused correctly

7.5.1 Rangefinder focus accuracy.

Obviously any measuring instrument has some tolerances, mechanical and optical/visual. The rangefinder of the Leica measures the distance of an object by superimposing two images of that object and noting the degree of coincidence of both images. If both images fully align, the distance measured is correct. As our eye is the critical factor here, the limit of accuracy is dictated by the eye' s visual resolution. Every equation that tries to compute the rangefinder accuracy has this limit of visual resolution incorporated. The blur circle that relates to the depth also defines necessary accuracy of field. The eye has a maximum limit of resolution of 0.06 mm at a viewing distance of 25cm, translating to 8 line pairs/mm. Often a more practical limit of 0.1mm is used, which translates to 5lp/mm. Even this value is too high for most uses and so the industry settled to a more convenient 2 lp/mm as the norm for optical formulae. These 2 lp/mm refer to a distance between two adjacent objects (points or lines) of 0.25mm (1 mm divided by 4). As we are talking here about the print or transparency , we need to translate this figure to another one on the negative. Assuming an 8 times enlargement factor we divide the 0.25 mm by 8 and we get 0.03mm: the famous diameter of the blur circle. We know that in reality we only have an infinitely small sharpness plane that is 'artificially' extended into three dimensional space by this DoF mechanism, combined with the resolution limit of the eye. The rangefinder in theory measures a point in space at one exact distance.

There is always a certain latitude in measuring inaccuracy: the focusing error. Slightly before and slightly behind the real distance the instrument will give identical readings. As a bottom line for rangefinder accuracy we must state that the distance of the focusing error is at least equal or less than the DOF distance. That is the most minimum demand. As the rangefinder is based on triangulation, we do not use in our equations lp/mm but the equivalent angular resolution. For the limit of 0.06mm the angular resolution is 1 minute of arc. For the often used 2 lp/mm the angular resolution is 3.4 minutes of arc. The former figure relates to optimum viewing conditions and the latter one to normal conditions. We are almost there! The triangulation method obviously is more accurate when the base length is larger. The Leica M rangefinder has an effective base length of 49.86mm for the M2/4/5/6 and 58.863 mm for the HM series. Contrary to the opinion of many authors I must state that the physical base of ALL Leica bodies from M1 over the M3 to the latest M6 is identical (69.25mm). The only difference is the magnification (0.58, 0.72, 0.85 or 0.92). The CL has a physical base of 31.5mm.

 

Model	Frames for Focal lengths	Magnification	Effective base length
M1	35, 50	0.72	No rangefinder
M2	35, 50, 90	0.72	49.86
M3, MP	50, 90, 135	0.92	63.71
M4,M4-2, M4P, M5	35, 50, 90, 135	0.72	49.86
M6J	35, 50, 90, 135	0.85	58.56
M6, M6TTL, M7, MP.72	28, 35, 50, 75, 90, 135	0.72	49.86
M6TTL, M7, MP .85	35, 50, 75, 90, 135	0.85	58.56
M6TTL, M7, MP .58	28, 35, 50, 75, 90	0.58	40.17
CL	40,50,90	0.6	18.9

 

Any equation that computes the RF accuracy will use at least three variables: effective base-length, visual resolution in angles and blur circle diameter. These are intimately related. There are several different equations to be found in the literature and unfortunately the results differ greatly. The factors that enter into the equation are the diameter of the blur circle which is normally taken as 0.03mm, which in my view is too large, the resolving power of the eye, which has different values, based on point discrimination and vernier acuity and the measuring circumstances like contrast level and fatigue of the eye. If we assume for all these factors reasonable values, we can calculate the following table, on the basis of a blur circle of 0.03mm and 0.01mm and a realistic power of discrimination of the eye. The table gives you the limiting values for normal accuracy based on point discrimination and critical accuracy based on vernier acuity. Column three lists the values when you are relying on superposition and contrast and column four lists the values when you need the highest accuracy based on vernier acuity. If you are planning to make really big enlargements or project your slides on very wide screens, these numbers are indicative of the care required when focusing at the wider apertures.

Focal length	Aperture	Effective base needed 0.03	Effective base needed 0.02	Leica 0.58	Leica 0.72	Leica 0.85	Leica 0.92
21	2.8	1.6	2.1	40.17	49.86	58.86	63.71
24	2.8	2.1	2.7	40.17	49.86	58.86	63.71
28	2.8	2.8	3.6	40.17	49.86	58.86	63.71
28	2	3.9	5.1	40.17	49.86	58.86	63.71
35	2	6.13	7.96	40.17	49.86	58.86	63.71
35	1.4	8.8	11.4	40.17	49.86	58.86	63.71
50	2.8	8.9	11.6	40.17	49.86	58.86	63.71
50	2	12.5	16.3	40.17	49.86	58.86	63.71
50	1.4	17.9	23.3	40.17	49.86	58.86	63.71
50	1	25	32.5	40.17	49.86	58.86	63.71
75	1.4	40.2	52.3	40.17	49.86	58.86	63.71
75	2.8	20.1	26.1	40.17	49.86	58.86	63.71
90	2.8	28.9	37.6	40.17	49.86	58.86	63.71
90	2	48.5	63.1	40.17	49.86	58.86	63.71
135	3.4	53.6	69.7	40.17	49.86	58.86	63.71
135	2.8	65.1	84.6	40.17	49.86	58.86	63.71

 

 

 

7.5.2 SLR focusing accuracy.

The most often used focusing aid is the split optical wedge, which is also based on the property of our vision, known as vernier acuity. The eye can judge very fast and accurate if two lines are broken or aligned. The physical base length of the 'rangefinder' depends on the slope angle of the wedges used and the focal length of its ocular. But the magnification of the image on the ground glass by the lens has to be added into the equation. The equation for the base length of the Leica SLR is Focal length / aperture times focal length /61.53 (focal length of ocular). The 50mm lens has an effective base of 9.82mm. Compare these values to the ones from the RF-system and you will see that the RF has higher accuracy. On the other hand we should note that we can focus quite well with an SLR and a 50mm lens. So there is some margin in accuracy needed. With a 135mm lens, the SLR base length becomes 42mm and from there the RF method is less accurate. In fact the real advantages in accuracy are lost around the 90mm focal length as small mechanical inaccuracies are enlarged disproportional in the RF-system. The SLR-system does not need such elaborate mechanical linkages. As long as the film plane and the ground glass are accurately aligned, sharp focus is ensured when the user sees a high contrast image on the screen or uses the split wedge to align vertical lines.

 

7.6 Appendix F: Lens manufacture in detail.

From an evolutionary viewpoint, we can observe that the inherent image quality of Leica lenses shows improvements that increase steadily, if not exponentially at least in the period from 1988. Creating and computing a new lens design to a higher degree of accuracy and to a higher level of aberration correction is only one part of the equation. The precision of the manufacture of the lens elements and the mechanical components, the care of assembly and the small tolerances in the testing equipment all have to be synchronized to the same level of quality. I would ask the reader's attention to what I would refer to as the quiet revolution in lens manufacture in Solms. The result of any lens design looks disappointingly simple. A lens element is fully described when we know the radius of curvature of both surfaces, the diameter, the thickness, the glass type and the distance to the next element. For a triplet lens the following list would completely define the design.

Lens element	Radius A	Radius B	Thickness	Glass type	Distance to next surface
1	25.5	1100	4.8	SK7	3.55
2	-81.4	26.4	1.24	F15	8.81
3	206	-55.7	3.38	SK7	86.17

 

The distance of the last element is of course the back focal length, that is the distance from the last surface to the film plane. For a lens as the new Apo-Summicron-M ASPH 1:2/90mm, that has 5 lens elements, the list is longer, but has the same type of information. These data are the result of a lengthy creative design process that may take many months, if not years to finalize. If you look at the numbers, you see there are three figures behind the decimal point, so the theoretical accuracy is in the realm of a thousands of a millimetre. This unit is often referred to as a micron or 1/1000 of a mm or 1/1.000.000 of a meter. To put the smallness of these numbers into context, remember that the average wavelength of the visible light is 0.5 micron. A triplet lens cannot be corrected to a very high degree and small tolerances will hardly impair the performance. The image degrading by the residual aberrations as larger than by small deviances from the manufacturing accuracy. In practice you see a production tolerance that amounts to tenths of an mm, or a hundred micron. One tenth of an mm (100 micron) may seem large in the perspective of optical calculation and tolerance, but in the shop and even in large scale manufacture, this distance is not easy to hold consistently. If we now look at the design specifications of the original Summicron 1:2/50mm, we see figures like a thickness of 1.42 and a radius of 101.78 (as the actual data may not be published of course, the numbers are just for illustration purposes). It is clear, that a lens that has a higher potential optical quality, should be manufactured with a level of accuracy and precision, that matches the increased performance. Theoretically the precision would be in the range of a hundredth of a millimetre (10 micron), but with the equipment available in that period, one would be very happy if the accuracy would be in the region of a few hundredths of a millimetre or 20 to 30 micron. Again, we should reflect for a moment on the fact that a reduction of the tolerance margin from 3/100 of a millimetre to 1/100 of a millimetre (or a reduction from 30 micron to 10 micron) is a factor three in higher accuracy. This is not so easy to accomplish.

We should also remember that no machine does function within the theoretical zerotolerance of specifications. Equipment has its own tolerance and any piece that is machined will be some % off the specified figure. It would be very nice, if the statistical variation of the production would be within 5% of the specified value. But even this requirement is not feasible, as the mechanical nature of every machine will generate a systematic drift in the specified values. So over a longer period of time adjustments have to be made to bring the machine into line again. Here the human operator is the key player, as he has to sense and check for drifts and out-oftolerance variations. Any lens element then will depart for a small value from the zero-tolerance value and this has to be accounted for in the stage of mounting the elements and the assembly. One option is to pair components that show the same deviation in value but with opposite signs. In would be statistically unlikely to assume that such pairs exist in all cases, so additional measures are needed. Another option is to use a technique, called adjustment of compensators. One can study the design of a lens and the sensitivity of an element to tolerancing and its effect on image quality.

As example for a Summicron-type lens (a double-Gauss design), we can establish that a certain surface radius can depart from the specification by ~0.16mm and another by a mere ~0.01mm before image quality will be degraded. Now a tolerance of 0.01mm is very tight. Additional study showed that the tilt and decenter values had a higher impact on the image quality. Greater care at the mounting stage could compensate for a slightly higher level of tolerancing at the radius and irregularity tolerances. Another way of approaching the problem of the unavoidable manufacturing errors is the study of the statistical distribution of errors than can normally be expected to occur. We first have to specify the required limit of the finished product. Let us say we specify that a lens is considered as accepted at the final check if the measured MTF value at image height of 9mm is no less than 90% of the calculated value. The value may be higher of course, but the minimum is 90%.

Assume that 85% of all manufactured lenses will meet this requirement. Then we have 16% of all lenses that will show errors greater than specified. If we study the production quality results, we see that most elements are already at their minimum tolerance values. But a further study shows that a tightening of the tolerances for decentering will ensure that 98% of the lenses will meet the goal of defined image quality. The lesson to draw from all of this that one should look at a lens as a tightly coupled set of components, where the original design, the selection of materials, the manufacturing of the components, the machining of surfaces, the care of mounting and assembly and the care of testing all contribute to the final result. It does not make sense to single out one aspect (glass type or accuracy of radius or the availability of multi layer coating) as the important characteristic of a lens. As noted, decentering and accurate spacing of lens elements are very important variables with a significant influence on the final image quality. The quiet revolution. If we look at lens manufacture, we see at the surface the same procedures as were established during 's times. The main stages: lens grinding and polishing, centering and edging, mounting of lens elements, and assembly of lens elements into the lens barrel and addition of aperture mechanism and focusing mechanism. Between 1988 and 1993 the designers at Solms ran into the limits of the production technology when designing lenses like the Apo-Telyt-R 1:2.8/400mm and the Apo-Summicron-R 1:2/180mm. This last lens has a resolution that is almost diffraction limited, which means that far more than 250 lp/mm can be resolved with high contrast. This lens can resolve details with a diameter of 0.002mm or 2 micron! Compare this figure to the Summicron lens I referred to above, which can resolve details as small as 0.02mm or 20 micron and we see a 10 fold increase in detail rendition, which has to be accompanied by a corresponding increase in manufacturing accuracy. Recall that the average wavelength is 0.5 micron and the smallest detail that can be recorded is 2 micron, pretty close to wavelength dimensions. The rays that are reflected from a tiny detail in the object, will converge to a point on the film plane, and this convergence we can represent as the tip of a very sharply pointed pencil. If we lightly touch the pencil on the paper (read film plane) we will draw a very tiny spot. If we press the pencil-tip with force through the paper surface, we will create a much larger spot. The same happens with the lens. If the distance from the last lens surface to the film plane is not accurately held, the rays will not focus with the sharpest point on the film plane, but before or after this plane. Instead of a sharp spot of dimension of 2 micron we get a spot of maybe 10 micron and most imaging quality will be lost, at least as calculated by the designer. A difference of a hundredth of a millimetre is critical already. Especially the contrast of a lens will suffer over proportionally if the spot diameter is enlarged and thus diffused. Lenses of previous generations had a much higher margin here as they are sensitive to tolerances in the region of 0.1 millimetre or somewhat smaller. The effective image quality depends as much on the manufacturing precision as on the design. If fact, they are mutually dependent. The high quality that the Apo or aspherical lenses deliver cannot be delivered if glasses with very specific properties are not available. This glass is in the standard catalogues of current glass manufacturers. The story that Leica uses glass that is specifically formulated by or for them and is only available to them is not true. What is true, is that the special glass types need to be treated in a very special way, and here lies the secret of Leica. Thy have the technology and expertise to employ these glass types.

As example on may mention the sensitivity of some glass to changes in temperature: heating and cooling will imbalance the molecular structure of the glass and if these changes have to be avoided, special coating techniques have to be employed, as example. Surface finish is another example. Sometimes the glass is so sensitive to humidity or oxidation that a glass surface needs to be coated within hours after treatment. Aspherical surfaces have a shape that departs from the pure sphere and in the past these surfaces could only be produced by a specially constructed device and by extensive manual adjustments. This procedure was not only expensive and error prone, but restricted the designer to a few aspherical shapes, which in turn limited the design possibilities. One needed machinery with a higher level of flexibility and that was found in the new generation of CNC (computer numerical controlled) machines. So the causal chain to ensure high imaging quality is as strong as the weakest link. Better imagery asks for specific glass types, specific shapes of the glass surfaces, accurately held tolerances and high quality testing equipment. Current manufacturing technology in Solms is a mix of high tech computer controlled equipment, that has been bought off the shelf or has been designed specifically for Leica in cooperation with the manufacturer. Special training and a long period of experience is needed before one can operate these machines. As the specifications now are increased a tenfold in accuracy (as compared with the previous periods) and dimensions are measured in one thousand of a millimetre, the operation of the machines has to be controlled on that level too and now the anticipation of the behavior of the machine is part of the game: if the tolerance now drifts for a few microns you can throw away the lens. Many machines are computer controlled and almost every workplace has its own testing equipment. Everyone is responsible for its own part of the process. It is fascinating to observe that in one room a onemillion Dmark machine is quietly and slowly polishing a aspherical surface into the required shape, in another a laser driven machine is grinding the rim of a lens to the accurate mechanical axis and that in a third one a woman is busy with blacking the rim of the lens with black paint by hand as the lens is rotated on a small electromotor. This time honored process, the same as for forty years, cannot be improved and there is still no mechanical substitute for it. A finished lens then is the cooperation of the creativity of the designer, the most accurate production process in the mechanical industry and the careful attention of a female member of the construction team. It is remarkable that this part of the process resists all attempts to mechanization. The causal chain then shows the interdependency of all stages, but also of the impact of improvements in one stage on all others. If you can polish a lens surface to a higher level of accuracy, the coating technique has to follow as does the precision of centering and mounting. Otherwise the gain in one department is lost in the rest of the process. Current tolerances are 2 microns for radius of curvature of lens surfaces, 2 microns for thickness and 5 microns for distance between lens elements. Grinding and polishing. Now that the designer has specified the surface curvatures in thousands of millimetres, that is three digits after the decimal point, with tolerances of less than 2 microns, it is imperative that every glass surface is individually machined into shape. In the past one would fasten 20 to 70 lens elements (blanks or pressings as they are called) to a common support and grind/polish them in one batch. This was economical of course, but it restricted the designer to those curvatures that could be processed in this way. And tolerances had to be larger. In theory one can polish the surface to a very high degree of accuracy, but in practice this cannot be accomplished. As this process requires the glass to be heated, it is unsuitable for some of the glass Leica wishes to use. Therefore individual grinding and polishing is now the rule. Not all glass elements are processed in the Solms factory. Some glass is outsourced as the cost of production would be too high to bear on the glass. The new CNC-machines have specially designed digital motors to position the grinding tool to an accuracy of 4 million positions in a 360o movement. As no machine works with a zero tolerance, the true values wander slightly around the ideal position and these deviations are constantly monitored and adjusted. Very important is the fact that the operator has to understand what the numbers mean, how the machine will react to every adjustment he makes. Grinding is a painful process for the glass. In fact glass parts are chipped away and the surface looks like a moon crater. The polishing has to be done to a depth below the damaged surface and now a precision of 1 micron is required. One person controls 4 of these polishing machines, that will smooth and close the glass surface. An interferometer is used to check the accuracy. In the past one used so called test plates to check the surface. A test plate has a shape that is identical to the one that has to be checked but in the 'negative' form. A positive surface will be checked with a negative plate. The principle behind this test is the same as that used with the interferometer. We all know the phenomenon of Newton rings, that is the irregular rings we see when a negative dos not fit exactly to the glass carrier. These rings are irregular and differ in thickness and shape. In the shop one wavelength is specified (often sodium lamp or HeNe laser) and by an interferometric comparison one can study the regularity of the rings. Are all rings concentric and regular, a perfect fit is found. With this device one interference ring refers to one half of the wavelength. If an accuracy of 2 micron is required, one may accept at most irregularities in the first 4 rings from the centre. (4 rings is 4 times a half wavelength equals two wavelengths equals 2 micron). Some newer machines can grind and polish in one step, which fosters the accuracy.

Polishing takes on average 20 minutes per surface or close to an hour per lens. As soon as one surface is finished it will be protected by a cover before the next process starts. A surface polished to a few microns of accuracy needs to be centred with the same precision. The rim of the lens has to be grinded so that the mechanical axis (defined by the edge of the lens) coincides with the optical axis that is the line between the centres of curvature of the two surfaces. The rim is machined to a precision of 0.01mm with a computer based laser device and it takes about 15 minutes to centre one lens element. If automatic mechanical centering cannot be done, Leica uses a specifically designed procedure to centre the lens optically. A simple calculation can tell us that a lens with let us say 9 elements already takes more than a day just to manufacture the glass elements. Aspherical surfaces are more complicated to manufacture. The basic shape is spherical but the asphericity has its own optical axis (or more with a complicated shape) and all techniques of grinding and polishing are based on the idea of random movements of the grinding and polishing tool. A sphere has everywhere the same shape so it does not matter where the tool is moving as long as the curvature shape is obeyed. Here a new computer based process has been devised. Now the mechanical axis is the base to work from.

The designer has computed the specific shape and this surface contour is fed into the CNC machine that will follow the shape contour as specified and will polish the required form. This is a very elaborate process and a simple check with an interferometer is not possible as there are more curvature shapes on the surface.

Leica uses special holograms that precisely represent the aspherical shape and with the help of these holograms an interferometer check can be done. The precision now is less than a _ wavelength. One aspherical surface takes more than one hour to finish. When I visited the factory the lenses that happened to be processed were the asphericals for the Vario-Elmar-R 4/35-70mm. The coating process has been discussed in previous chapters. Before the glass can be coated it has to be cleaned. In 30% of the cases the glass is cleaned by ultra sonic techniques, but not always as some glass is still to sensitive and has to be cleaned differently. I do mention this to indicate that Leica s very attentive that every component gets the treatment it deserves or needs. I can add here that Leica uses the new Advanced Plasma Source technique that does not require heating and cooling. The glass is less strained and its original properties stay intact. The APS technique deposits a smoother and more stable surface coating and so more sensitive glass can be used. The circle of the causal chain is closed again. Generally 4 to 6 layers are used per surface, but is it possible to deposit 36 layers per coating. Every batch is checked as a deviation per layer of less than 10 micron impairs the spectral transmission properties. Then the glass is put in stock in gas filled cabinets or dry cabinets, whatever is best. Ultrasonic cleaning is now used when the glass is prepared for mounting. Again there is additional manual cleaning and inspection by female workers several stages during the mounting and assembly. Lens elements are mounted in a close-fitting sleeve and there are several methods to retain the glass in the mount. One can use a threaded lock ring with or without spacers or the lens can be cemented in place by a plastic cement, that has also a slightly centering effect, care has to be taken that the cement does not overflow and here Leica uses computer-controlled machines that adjust the cement flow very accurately. The individual lens mounts are now aligned so that the mechanical and optical axes coincide. The diameter and distance are very important parameters for the performance of the lens and here tolerances of 1/100 to 5/1000 of a millimetre are specified. The aperture mechanism is installed and is checked with a torque meter to ensure the click stops are within ergonomical specifications. The aperture mechanism of the R-lenses has a specially designed braking mechanism to reduce the bouncing of the blades when the aperture closes to its preselected position. Would the blades be allowed to bounce, this will reduce the aperture opening momentarily and cause underexposure. The mounting of the several lenses or (in the case of cemented lenses) lens groups into the lens barrel is a labourious and exacting procedure, and the machining of the barrel itself has to be done to very small tolerances. Depending on the complexity of the lens , several tests are done to ensure that every stage will deliver a sub product within tolerance. After assembly of the lens, there is a check (with MTF equipment) to see if the lens performs as specified. As additive tolerances can combine in assembly, there is always a possibility that the performance is just outside the tolerance band. By using a compensator mechanism it is possible to fine tune the lens into the tolerance space.

Not every lens does get this check as Leica knows by experience which lens types have to be checked individually or by sample. After this step the lens is ready for its final inspection. Every lens that leaves the factory has been individually checked. (see chapter 2.4.2 foe additional information).

 

© Erwin Puts
Back to the table of content