Soviet Lens Designers Pose for a Group Portrait (30x enlargement
from Neopan 1600 and an 85/2 Jupiter-9 Lens at f/2 and 100 feet)
Well, I've heard enough about Hexar/Leica compatibility wars. How about the compatibility problems between Nikons and Contaxes and Leicas and Soviet stuff?
1. How does a rangefinder couple to a lens?
We have to start with a discussion of how rangefinders couple to lenses.
Long ago, in a galaxy far away, there were two competing standards for cameras.
One was the Leica III, which involved a camera with a short-base magnifier that connected to lenses with a threaded mount. Inside each lens were two moving parts. One was the optical unit, which is all of the individual optical elements and the aperture ring. The other is the rangefinder cam, which transmits the focused distance of the lens into the camera body.
For a "50mm" lens, the lens and cam move forward and back at the same rate. In Leica thread mount, the cam is about 6mm from the lens flange at infinity and about 3mm from the lens flange. So the travel of the optical unit from near to far is 3mm. The entire rangefinder system assumes this rate of movement, which corresponds to the distance a 51.6mm lens (actual focal length) moves forward when focused from infinity to one meter.
Note: the 51.6mm assumption is controlled by the linearity adjustment of a Leica M body's rangefinder. I am not sure that how much it can be adjusted away from 51.6mm, or whether you would ever want to (unless it was off).
For lenses other than 50, it is necessary to translate optical unit movement, because the camera body does not know the focal length of the lens mounted. You need this translation because the optical unit 21mm lens moves about a 0.5mm between close and far, and a 90mm lens might move, say, 9mm. So if the contact surface for the rangefinder mechanism moved at the same rate of speed as the optical units for these lenses, the rangefinder would only be consistent at infinity between lenses.
On typical Leica-style rangefinder camera, the distance is translated mechanically. There are several systems in use today.
Wideangle lenses tend to use two helicoids. The optical unit is mounted in one. The threads of the helix are designed to move the optical unit a certain millimetric distance in a certain degree rotation. Let's say here it's 0.5mm forward movement in 90 degrees (a quarter-turn for focusing is fairly common).
Single helicoid. This term is something of a misnomer, but since John Van Stelten introduced it to me, I'll respect it. I'd prefer "single-adjustment helix." On the back of the optical unit helix ("focusing helix") is a tab, which engages a second helix in the rear of the lens. The rear helix ("rangefinder helix") has a tab-shaped cutout in it. The rangefinder helix is set up to move forward 3mm in that same 90 degrees. The back of this helicoid is what contacts the rangefinder mechanism roller in the camera. This system is called a single-helix system.
With this system, there is no adjustment of the absolute position of the rangefinder helix. So the infinity stop is set for where the optical helix is when the optics are focused at infinity, and the back of the helix is ground until the test rangefinder reads at infinity. This is the type of mechanism used on Canon rangefinder lenses.
Double helicoid. No, it's not Crick and Watson, but a system in which the two helixes (focusing helix and rangefinder helix) described above can rotate completely independently, but are tied together through the focusing ring.
This type of mechanism (which I think should be called the "double adjustment helix") allows for precise adjustment of the front-to-back position of the rangefinder helix. So in other words, you can adjust the infinity point of the rangefinder helix separately from the infinity point of the optical unit (the focusing helix). This arrangement is most commonly found on Leica-brand rangefinder lenses. It requires a lot of hand adjustment during assembly or CLA, and we know how much Leica loves to adjust things. You can tell that a lens probably has a double helix if the rangefinder coupling ring on the back does not rotate when you change the distance on the lens.
I would not go so far as to say that this is a more accurate method for assuring proper rangefinder focus; its real purpose seems to be to allow parts to be made to wider tolerances and then adjusted individually so the whole system comes into spec. Contrast this to the "statistical quality control" methods used to calibrate Canon lenses, which makes the proper lens adjustment a physical function of hardware (it can only come apart and go back together one way).
Curved cam. This is variation of the "no-helix" design below. Basically, the back of the optical unit focusing helix is ground so that as the lens is focused, its depth from front to back (well, where it contacts the RF roller in the camera) changes. This is what I understand to be in use with Minolta CL lenses. It allows for very, very compact lenses, much more compact than the two types above. On a 40mm lens, you might not even notice that the the back of the helix is cut at any angle other than dead parallel to the film plane.
Normal lenses can use one helix, two helices, curved cams, or no cams. If a normal lens is equal in actual focal length to the assumption for the Leica rangefinder system (i.e., 51.6mm), then you can use the same helicoid for pushing the optical unit forward and for contacting the rangefinder mechanism in the camera. This is the solution used on Summars, older Elmars, and some older Industars. Given that the "single helicoid" design above actually has two helicoids, we'll call this the "no-helix design."
Alternately, depending on the price point, a manufacturer can also make a single helix mount or double helix mount as described above. Most LTM normals use the single helix mount because it is now inexpensive to make, and most of of the rangefinder helix fits within the threads of the the Leica thread mount.
Telephoto lenses can use one helix, two helices, or a true cam. This is where things get fun.
The true cam is an interesting solution to two problems. The first problem is that with the "single helicoid" and "double helicoid" systems, you still need to make two very finely threaded helixes. This is very expensive and time-consuming. The second problem is that the optical units of telephoto lenses move a lot farther fore and aft when they focus. This makes it the "single helicoid" system difficult to implement, because it then requires the linkage tab to be really long. This leads to focusing slop.
The true cam solves this problem by only making it necessary to have one helicoid (total), just like a "no-helicoid" normal lens.
Telephoto lenses that have a true cam have a small tab visible at the back of the mount. This tab runs about half the length of the lens in a track (so it can only move front-to-pack). The lens elements sit in a helix, the tail of which is a tube with a spiral cut into it (from front to back). The tab has a pin that engages this spiral cut. So when you focus, the tab is pulled fore and aft along the spiral (think of it like a worm gear). The pitch of the spiral translates the tab's movement to be equivalent to our magic 51.6mm lens. You can see this at work in Nikon rangefinder teles, for example.
Depending on the make, true cams can be adjusted either by moving the tab relative to the mating pin (where it connects to the spiral) or by simply grinding down the end of the tab to the right position. Canon 135s are notorious for not not having properly-ground tabs.
The competing system was the Contax, which had the focusing helix for a "50mm" lens in attached to the body but was in all other ways (and with all other lenses) similar to the Leica system. Contax was standardized around a standard lens of 52.3mm. Contax bodies had two different bayonet mounts. An inner bayonet (which mated to 50mm lenses) and an outer (which mated to all others).
Within a given system, for lenses to work properly with bodies, three conditions must be present. First, the actual focal length of the optical unit must be within a certain spec so that the conversion from focus helix to rangefinder helix will be accurate This problem becomes all the more acute when you are using a fast lens close up. That's why when Leica made the dual-range Summicron, it made focusing mounts for several different actual focal lengths. Then optical units could be matched with the mounts that would focus them the most accurately. Second, the rangefinder helix focal-length assumption (51.6 or 52.3mm) must be the same for the body and the lens. Third, the back focus of the lens must be the same as the body focus of the body. And obviously, everything must be ground, machined and computed properly.
2. Enter the Nikon
It's well known that Nikon rangefinder equipment and Contax rangefinder equipment are not compatible. Many people attribute it to Nikon's imperfect copying of the Zeiss design (this was, in fact what explains why early Canon lenses don't fit Leica bodies; someone screwed up the pitch of the threads on the mount). This is simply not the case.
Nikon made most of its reputation making recomputed Zeiss designs for Leica lenses (more on this, click here). When it came time for Nikon to make a camera body, it initially decided to make a body which had a Leica-style M39 lens mount and a Leica-style rangefinder. The prototype body did borrow some styling from the Contax II, although it was missing a number of Contax features: thumbwheel focusing, built-in helicoid for 50mm lenses, etc. Notably, the Nikon prototypes also had simpler Leica-style (rather than Contax-style) cloth curtain shutters and rangefinder mechanisms based on a 51.6mm standard lens. By all accounts, it would be a copy of Contax styling that would in every other way be a Leica. It would naturally be a platform for Nikkor lenses already made in Leica mount.
Scrambling for a ready-made lens mount, Nikon naturally chose the Contax, which had a convenient 3-prong bayonet. Nikon then designed in (or rather reincorporated) the thumbwheel focusing and built-in helicoid (going to the Contax-style internal helicoid allowed Nikon to make its 50mm lenses much simpler and cheaper - basically they are just optical units. Nikon adopted the "right to infinity" focusing direction of the Contax (which continues today in Nikon SLR lenses), which is the opposite of Leica-style "left to infinity," which is standard on almost every other camera on earth.
But... despite adopting the back-focus distance of the Contax, Nikon retained the 51.6mm normal lens of the Leica system -- meaning that Contax lenses would focus properly at infinity (and rangefinder-align properly at infinity), but as you move closer, the difference between the 52.3mm assumption of the Contax and the 51.6mm assumption of the Leica makes for progressively bigger focusing errors as you get closer and closer. This was the problem with using non-normal lenses; when you use 50mm Contax lenses in the Nikon, you made it even worse.
Lest you think I am making this up, you can find this story on Nikon's corporate web site if you look. Here's the URL. You can also see the 3 Nikon Leica screw-mount prototypes in Rotoloni's article here — look under "Contax Style" on the right side.
3. Ivan the Incompatible
Without belaboring the story of what happened after World War II, the Soviets carted an entire Zeiss Contax factory back to Ukraine (this is why the Contax IIa and IIIa are of a slightly different design from the prewar II and III). The Soviets made tens of millions of cameras and lenses that were copies of Contaxes and Leicas (the Contax clones were sometimes made from actual Zeiss tooling).
Among these products are Leica thread mount (M39) lenses with rangefinder couplings: 28/6 Orion; 35/2.8 Jupiter-12, 50/2 Jupiter-8, 50/1.5 Jupiter-3, various 50/3.5 Industars, the 85/2 Jupiter and 135/4 Jupiter. All of these are basic Zeiss designs that the Soviets put in mounts for Leica. Some of these mounts look like the Jena LTM Sonnars; some look like wartime Zeiss lenses. These lenses are generally inexpensive and plentiful (there are a few thousand real wartime Sonnars; and probably a hundred or a thousand times as many Soviet copies).
50/1.5 Jupiter stopped down to f/5.6
These lenses have mediocre reputations, probably not helped by the fact that Ebay is encouraging Russian and Ukrainian sellers to empty every drawer of dog lenses into the United States. Two lenses that seem to get a really bad rap are the 50/1.5 Jupiter and the 85/2 Jupiter. Having used both, I can say that the Nikkor 50/1.4, a copy of the same Zeiss Sonnar absolutely murders the Jupiter-3 when used on the same Leica camera. Same with the 85/2 Nikkor and the Jupiter-9.
But why? The traditional answer is quality control. I don't think that's all of it. One gentleman, Mr. Michael Darnton, has repeatedly suggested that the LTM lenses are simply incompatible. There may be something to that. People always say the LTM lenses are "soft," and no one ever seems to complain about the same lenses in Kiev (Contax) mount. Nor do they seem to complain when these lenses are plugged into Soviet LTM bodies. People say the "quality control" is better on the Kiev/Contax mount lenses. I think this is not realistic, since they are made in the same plant.
Robert Ludwig wrote the following after this article first appeared:
I have empirically confirmed Dante Stella's assertion of Leica LTM - Soviet 39mm rangefinder lens incompatibility by testing 3 screwmount Leicas, 12 non-Soviet Leica compatible lenses (Leitz, Canon, Angenieux, and Nikkor having focal lengths ranging from 19mm to 135mm), 23 Soviet 39mm rangefinder lenses (Industar and Jupiter in focal lengths of 35mm, 50mm, and 85mm) and 15+ Soviet cameras (FED, Zorki, Leningrad and Drug2) manufactured after 1950. The median resultant focusing error is approximately 2.5 inches at a subject distance of 5 feet for Leitz cameras and Soviet lens combinations tested except for one Jupiter 8 customized for the Leica - like cylindrical cam Drug2. The source of the error (incompatibility) is not known to me. Since the error is systematic and uni-directional, a knowledgible photographer using a Soviet lens on a non-compensated Leica or other non-Soviet camera LTM compatible camera can partially compensate for the focusing error by using a focusing target approximately 2 inches from the subject.
When I inquired as to the mechanics of the test, he wrote:
Yes, you may quote me.
For the tests and calibrations at 1.5meter, I used a steel tape for measurement, a tripod, and one 1.5 million candlepower spotlight placed a distance of 6 feet from the vertical focusing target. The reason for the high illumination was to facilitate detection of coincidence (or alignment,in the case of my Leningrads) of the rangefinder images. I used a Leica IIIg, a Leica IIIf and a Leica IIIa that were calibrated (with minor variations) to focus accurately at infinity and 1.5 meters (or 5 feet in the case of the Canon 35mm f1.8 and the 50mm Summicron) with my collection of German, Japanese and French LTM lenses. On the Soviet side, I calibrated three FED 2s (early, medium and late), three FED 3s (early, late and "brand new"), two Zorki 6s, two Zorki 4ks, two or three Zorki 4s (both early and late), one FED 4b, one FED 5c ("brand new") and two Leningrads to focus accurately with a "representative" Leitz lens - a 50mm f2. Summitar. With three exceptions, I verified these calibrations using my collection of non-Soviet LTM lenses. (My two f2.0 Canon 100mm lenses and my Nikkor 85mm lens will not mount on the Leningrads due to interfere between the lens barrel rings at the rear of these lenses and a shelf below the Leningrads' top plate) (I also used a Drug2, but did not a satisfactory calibration for Leica lenses. The Drug2s have a Leica-like roller cam and were shipped with Jupiter 8s that seem to have customized to the camera). The example of the Drug2 Jupiter 8 that I own only deviates from the Leitz compatible LTM lenses by approximately 1.5 inches at 1.5 meters. To calibrate the camera to the Leitz lenses it would have been necessary to bend the support for the vertical roller.) The rangefinder coupling of Drug2 Jupiter 8 at a specified lens distance of 1.5 meters is the closest to that of my Leitz compatible LTM lenses. The next closest among the Soviet lenses is a "brand new" black Jupiter 12 (The Soviet version of the 35mm f2.8 Zeiss Biogon.) The only model of the post 40s Soviet 39mm LTM lenses known to me that I did not test is the 135mm f3.5 Jupiter 11.
In calibrating the Soviet cameras, holding the shutter open and inspecting the action of the camera rangefinder cams from the rear, I first aligned the pivoting rangefinder cam so that it properly engaged the lens focusing rings (or pawls in the case of the Canon, Nikkor and Leica longer focal length lenses). In some instances it was necessary to bend the attachment levels and tilt the rangefinder cam to insure that the cam properly maintained contact with the lens coupling ring (or pawls) throughout as the nominal focus was moved between infinity and the closest distance accommodated by the lens. The thin focusing ring on my 35mm Summaron lens also necessitated this type of alignment. When mounting and dismounting lenses having pawls the lens focus was moved to its closest positions, the backs were removed, shutter opened and cam pulled to its most rearward position. The step was necessary to prevent the tips of the cams from "hooking", damaging, and potentially altering the cam alignment. This procedure is also recommended when using lenses having pawls with Soviet cameras having pivoting rangefinder cams.
The contention concerning Soviet 39mm rangefinder lenses supported by my experiment is the following: "At an indicated lens distance setting of 1.5 meters, the rearward extension of the cam coupling ring with respect to the lens mounting flange of each of my collection of more than 23 39mm Soviet rangefinder lenses that will mount on my Leicas is slightly less than that of a typical Leica or Leica compatible Canon, Angenieux, or Nikkor lens. As a consequence, when used on a camera calibrated for a typical Leitz or Leitz compatible lens (whether German, Japanese, or Soviet), to achieve an indicated lens distance setting of 1.5 meters for a subject at a distance of 1.5 meters it is necessary to focus on an object roughly 2.5 inches behind the desired point of focus (i.e., subject)."
Obviously, a conceptually identical experiment could have been performed using precise measuring tools to compare the extensions of the lens focusing ring (or pawl) of the Soviet and non-Soviet LTM lenses when fully collapsed and at a nominal distance setting of 1.5 meters. It was not necessary to use cameras to substantiate the first part of the above contention!! However, a camera or computation scheme would then necessary to translate the extension discrepancy to inches or to attach quantitative significance to the discrepancies.
There obviously could be several explanations for this result: (1) Soviet
lenses tested have mismarked distance scales; or (2) the pitch of the helicoid mechanisms controlling the rearward extension of the lens focusing ring within each of the ensemble of Soviet lenses tested is designed to specifications different from that of the Leitz compatible lenses; (3) the non-Soviet lenses tested have mismarked distance scales at 1.5 meters (or 5 feet as applicable) or focus incorrectly. Given my 50 years of experience with Leicas, the later alternate is not tenable. It is also contradicted by Leitz' and Canon's production of the various supplemental non-reflex housing close-up attachments that rely on the calibration of the camera rangefinder and precision of the lens registry. (I have owned and used the Leica IIIf with the 50mm Summitar lens since I was 12 years old. This IIIf was CLA'ed with complete replacement of the rangefinder mechanism within the last 4 years; the rarely used IIIg is pristine in original condition and the IIIa is my user camera. It was made the same year that I was born. My use of this 64-year-old camera for stage photography is something of a put-on to owners of auto-focus, auto-exposure and digital cameras).
To me the only reasonable explanation is (2). This is also supported by the dismal experiences I have had with low light stage photography using the Nikkor P C 85mm and two Canon 100mm f2.0 lenses on Soviet LTM rangefinders calibrated for Soviet normal lenses.
After aligning the Soviet rangefinder prisms and pivoting cams at infinity, 3 meters (or 10 feet as applicable) and 1.5 meters (or 5 feet as applicable) using the lens distance scales and metal tapes, the quality of my low light stage photographs using calibrated Soviet rangefinders is comparable to that obtained with my Leicas - even though, the effective base lengths of the Leicas is greater than those of the Soviet rangefinders. At a recent concert given by the Canadian guitarist Keith Glass and vocalist Lynn Miles, I took some excellent photos of Glass at f2.8 using 50mm Summicron and 85mm Nikkor lenses mounted on FED 3a's in a setting that so dark that I had to set my shutter according to the shutters' sound - I could not see the markings on the shutter speed dials.
With this evidence supporting the contention that the lens distance scales / rangefinder coupling ring travel on the Soviet LTM lenses differs systematically from that on the Leitz compatible LTM lenses obvious questions are "How did this discrepancy arise? and "Why wasn't is corrected?" Could this be a consequence of your conjecture concerning the maintenance of a Contax vs.. Leica focal length by the Soviet engineers? Incidentally, I am a mathematician and applied financial economist. I did not bother to dress-up or summarize the empirical results in the jargon of mathematical statistics since the data was persuasive without the window dressing.
The comment I left on your website is incorrect in one respect. When I wrote it, I had forgotten that I did not include my 19mm Canon lens in the experiment.
I would guess that to promote efficiencies of scale, the Soviets standardized all of their camera systems to a 52.3mm standard lens. Why do I think so? Ever see any of the Soviet lens charts that show the actual focal lengths of 50mm LTM rangefinder lenses as 52mm? The commiecameras.com site has a verbatim copy of a Soviet spec sheet for all of the lenses.
First, this allowed the immediate use of Contax tooling to make optical units for both Kiev and Fed/Zorki lenses. Second, standardizing on 52.3 would allow optical units to be used for both Kiev and Fed lenses interchangeably, depending on what needed to be assembled that day. Third, it also cuts down on the number of rangefinder helix types to design. And who cares about whether or not the resulting M39 bodies were truly compatible with Leicas? What good worker would have such a capitalist-lackey Leica camera? And what self-respecting Westerner would be caught with a Soviet lens? For a very long time, Soviet cameras were just not an export product.
You can hook a Soviet lens up to a Leica (or Cosina or Konica or whatever). Turn the lens to infinity - the rangefinder will line up at infinity, they lens will say infinity, and the image will be focused at infinity. But get closer than infinity, and your problems begin. While you might not notice a gross focusing error using a 28/6, a 35/2.8 a 50/3.5 or 50/2 with a real Leica, the 51.6/52.3mm difference will cause an error with a 50/1.5 or an 85/2 close-up (like about 1m). Oddly (or not), these are the lenses everyone thinks are a little soft (or a lot).
But quality control can play a role too (although probably not as big as people like to think). Even if a Soviet lens is designed around the 51.6mm standard, the following things can cause a Soviet lens to be soft optically:
– Lens elements that are out of spec;
– Optical units which are not to dimensional specs;
– Optical units whose actual focal length significantly deviates from the norm;
– Soft aluminum alloy mounts which are not dimensionally stable or hard enough to keep their shape;
– Single-helicoid designs with rangefinder helixes cut at the wrong angle;
– Single-helicoid designs with rangefinder helixes cut to the wrong length;
– Wear on the rangefinder bearing surface of the lens (the Feds and Zorkis have a sled-shaped contact, not a little wheel, and the sled rubs against the bearing surface; and
– For 50mm Industar collapsibles, optical units that are not close enough to spec to match the focus/rangefinder helixes they are mounted to.
This is not to say that Soviet lenses are bad; it's just that a lot can go wrong in a factory where there are few incentives to perform quality control (gee, if I don't like "FED" brand, I can always buy "Zorki" brand equipment, right?). And to be fair, these things afflict even Leica lenses (although to nowhere near the extent). Having had a bunch of 1970s and 1980s black Soviet LTM lenses, I can tell you the rangefinder helicoids are often cut too short or too long, and the matte black paint in the barrel (to kill internal reflections) is often missing. The best Soviet products are 1958-1962, and all chrome (well, aluminum).
Maybe the best thing to do with Soviet lenses is put them on Soviet bodies. Or spring for the extra cash and buy a Leica lens. As a spot of hope, you can always get the SLR lenses (especially the Jupiter-9), which will eliminate the need to worry about rangefinder couplings and actual focal length (they are also a lot cheaper - like 1/3 the price for a Jupiter-9).