Interchanging lenses in Copal Shutters. Why does it work?

[sanking]

Question for the optical gurus.

Many modern lenses intended for large format work are threaded to screw directly into Copal #0, #1, #3 or #3S shutters and quite a number of people are using the same shutter for two or more lenses, without any consideration to spacing between the elements. And this seems to work reasonably well in practice. My question is, why does it work so well? Is it because most lenses made for a given shutter size also have identical or similar spacing between the elements, or is it because most lens designs work well enough in spite of varying spacing distances between the elements?

[Steve Hamley]

Sandy,

But I think the key phrase is "... reasonably well in practice." The human eye is not a good way to measure lens quality, or at least eyes are not a consistent one across humans. Most of my modern Schneider and Rodenstock lenses have very thin shims on the front cells, some must be only a thousandth or so, while none of my Fuji A- or C- lenses do. I doubt you'd see the difference with or without, but I assume you could with sophisticated optical measuring equipment, otherwise the shims wouldn't be there.

But I'll guess improved manufacturing quality control with tighter tolerances for both shutters and lens cells has helped. In an ideal world, if I'm manufacturing lenses, I want to avoid hand-shimming the cells, an expensive proposition. So I'd design cells that have proper spacing for a standard shutter without shimming, and have a lens cell manufacturing process with appropriate tolerances. I'd also use a lens design that was usable with the shutter and cell manufacturing tolerances if it was compatible with the other lens performance criteria. I suspect this is what Fuji did.

Thanks!

Steve

[Bob Salomon]

When you buy a new shutter it comes with blank aperture scales. That is because the shutter manufacturer has no way of knowing what focal length or speed lens will be mounted in it.

So at a given spot on the aperture scale (a specific diameter opening) different lenses will have different apertures.

So a 35mm Apo Grandagon 4.5 and a 90mm 6.8 Grandagon both mount in a 0 but their aperture scales are very different.

Will you meter through the lens? Or guess? Also the threads on the lens cells were not designed for constant changing so if you were to cross thread a cell it can really get expensive.

[sanking]

"When you buy a new shutter it comes with blank aperture scales. That is because the shutter manufacturer has no way of knowing what focal length or speed lens will be mounted in it.

So at a given spot on the aperture scale (a specific diameter opening) different lenses will have different apertures."

The need for a different aperture scale is fairly obvious. But, are there not sources that sell aperture scales for different focal length lenses that fit Copal shutters?

As for the danger of cross threading you make a good point. However, I am personally thinking more on the lines of permanently replacing the existing cells in a shutter with those of another lens, not constantly threading different elements in and out so my major concern is optimum spacing of the elements.

[Michael S. Briggs]

One of the ideas behind mass production is interchangable components. It is much cheaper for the lens manufacturer if they can take any front lens cell, any rear lens cell, and any shutter and screw them together and obtain a complete lens that meets their performance specs. If they have to make optical or mechanical measurements to select components that will go together to make a lens with good performance, or to make adjustments, then their costs will go up, probably significantly because skilled labor will be required. The goal is interchangable components because this will normally reduce costs. In some cases holding the tolerances on each part to achieve interchangability might be more expensive than having to test components and select or make adjustments.

Copal delivers shutters that meet certain specs for the allowed variations in distance between the front and back, concentricity of the front and back threads, etc. These specs are listed in the booklet that sometimes comes with a Copal shutter. The allowed variations are quite small.

An optical design consideration that few photographers think about is designing a lens so that it will have the desired performance regardless of any mixture of expected variations in the components, e.g., curvatures of elements, thickness of elements, spacing of elements, thickness of the shutter, etc. The manufacturer would prefer to have lens designs and manufacturing procedures that allow interchangable components. With modern technology this is probably more possible than in the early days of photography.

The only LF lenses on which I have seen evidence of adjustments is wide-angle / wide coverage type such as the Nikkor-SW. The various Grandagons, Super-Angulons and Fuji-SWs are of the same design type. Probably this design type is more sensitive to either the optical powers of each cell, or to the spacing between the front and rear cells, or to the distance from the aperture, and so the manufacturer finds it cheaper to adjust the distance of the front cell from the aperture and rear cell rather than changing the manufacturing method to achieve smaller variations. My guess is that the variations that require adjustment are in the optical elements (probably the powers of each cell) and not in the shutter. Probably small variations in the powers of the cells can be compenstated for by varying the spacing of the cells. If my guess is correct, than the cells can be freely moved to any shutter. If not, then the spacing shim should be changed if the cells are moved to another shutter.

[David A. Goldfarb]

The frame spacing issue seems like a fairly simple problem to solve. Don't the lens manufacturers just build a certain amount of spacing into the mount for the cells, so that they screw into the shutter with the right spacing without need for shims or adapters? I suppose that if there were different shutter options, they could design the cells for one shutter and use an adapter for the other shutter.

Frame spacing issues aside, it would also be fairly easy to have a custom aperture scale made so that one shutter could be used for several lens sets.

I have an Ilex 5 shutter that contains a 10" WF Ektar, and I use a couple of barrel lenses front mounted on that shutter, so I don't have the frame spacing or aperture scale issues, since the lens barrels have their own irises. The advantages are less bulk and weight in my kit, and consistent shutter speeds for all three lenses. The disadvantage is that while this setup works well for 8x10", I get enough vignetting from front-mounting to notice on 11x14", so I'll probably eventually have these barrel lenses traditionally mounted in Copal 3 shutters at some point down the line.

[Arne Croell]

The main answer has been given by Steve and Michael: Optical systems are not (only) designed for best performance, but also to make that performance as insensitive to manufacturing tolerances as possible. This is a step in optical design that is called "tolerancing", and was possible to a much higher degree after the advent of computers. I would assume this step will include the shutter spacing. Together with the advances in reducing mechanical tolerances in machining in the last 50 years allows the exchange of lens cells without problems. In case the lenses come with shims (usually between front cell and shutter)these should be transferred to a new shutter when switching since in all likelyhood the need for the shims did not come from the machining tolerances of the shutter, but to allow for changes of the optical numbers in the cells. In addition to the super wide angles, I have seen shims also in modern Plasmat constructions, e.g. the Apo-Sironar S and -W, and the Apo-Symmar L.

[Ernest Purdum]

I think that it is much more likely that any shims found with mounted cells were put in by the lens maker to compensate for for lens variations than for variations in shutter thickness.



As long as if there are shims they remain when the cells are installed in the new shutter, I think the chances for proper performance are good.



(Sometimes in dealing with process lenses such shims are attributed to optimization for differing subject/image ratios. I doubt that this is correct.)

[Arne Croell]

Ernest, the adjustment of process lenses (mostly dialytes) for different enlargement ratios has been done occasionally. I know it for sure for the Docter Apo-Germinars. However, these shims are much thicker than the normal ones - the one for my 300mm Apo-Germinar is 0.7mm thick. The normal shims are much thinner, typically 1/10 of a mm or less.

[Michael S. Briggs]

Here are the tolerance specs from a booklet for a Copal No. 1 Press shutter:

distance between end surfaces (variation in thickness): within 0.025 mm (= 0.001 inch),

parallelism of end surfaces: within 3 arc minutes (=0.05 degrees),

eccentricity of front and rear lens mounting threads: within 0.03 mm.

In the tolerancing procedure that Aren describes, an optical designer can perturb a design by these amounts and check that the performace remains acceptable.