CNC & CAD in Buildling a Camera - Tri-Color Specifically

[michaelbsc]

US Patent - Des. 115,165 - Curtis Color Camera

Not very exciting though. Haven't found the goods yet.

edit: Adding some as I come across them; haven't read any yet though.
USP 1,895,555 - Emil Reckmeier 1933 - (Dead Link Removed)
USP 1,951,896 - Bermpohl 1934

The 1951896 is the one I just found in my archives. There are pics of the mirror arrangements, but not specific angles and dangles. I had planned to get a protractor to get close, then calculate the needed angles necessary. But time never presented itself.

The advantage of these over a straight 90 degree splitter is the focal path is shorter, so you can get wider lenses, and the body is smaller.

As far as my earlier statement that one of the three would have the image reversed, that's of no real consequence, and was an attribute of many commercial systems of the day. Printers simply dealt with it.

I had the idea that the first mirror needs to have slight registration marks in opposing corners so you can easily line things up later. Frame generously.

My original thinking was to learn to process the three images in the K-14 chemicals, without the requirements of doing it in the real K-14 process, then aligning the three sheets to make a transparency. Hence my suggestion for K-15 name.

Of course, one could always simply develop for color separations to print.

 

[polyglot]

Yes, that Bermpohl is pretty much what I was thinking in terms of a crooked/folded implementation. I was expecting it to be more-crooked, but it still looks pretty compact.

Anyway. I don't reckon it would be too hard at all to draw up the casing for one of these and have it laser-cut from birch or bamboo plywood. Total cost of the body should be well under $200; the hard part IMHO will be obtaining large, high quality separation filters.

Edit: not to mention an all-analogue tricolour printing process!

 

[michaelbsc]

Edit: not to mention an all-analogue tricolour printing process!

My thought was that the beginning experiments for me would be tri color bromoil matrices. It seems far simpler to start than dye transfer or carbon.

Of course, not having done it I could be completely underestimating how hard tri color bromoil really is. But if perfection isn't the goal, then it should be manageable as a way to get the camera hardware figured out.

My enemy is time management.

 

[RedGreenBlue]

There's a precedent for plastic bodied tri-color cameras. The 1945 version 4x5 Curtis Scout camera had a welded acrylic body. The philosophy behind this was that if the camera was dropped it would be sent in for repair. It has been said that a frequent complaint of the Devin camera was if it was dropped the metal would bend and registration would be lost. The earlier version 2x3 Scout (1940-41 model) body was an aluminum casting. The final model, the 1950s Press Special had a body of cast magnesium.

 

[polyglot]

Worth noting that burgers has crossposted this to LFPF and got a very cool reply suggesting the use of a 3-way compact prism for beamsplitting. Such an arrangement should be on a similar order of depth as a traditional SLR mirror, which means you could maybe use retrofocus SLR lenses (RB67) to get normal/wide views.

Anyway, I suspect it'd be a good idea to implement decent tricolour printing (from consecutive shots on a rollfilm back) before building a camera like this. To me, the camera looks like a minor mechanical challenge of a similar order of complexity as a traditional camera build, just with a few more parts. The camera is of no use unless you can print (scan?) the separations, and the printing I think will be a real challenge.

 

[michaelbsc]

Worth noting that burgers has crossposted this to LFPF and got a very cool reply suggesting the use of a 3-way compact prism for beamsplitting. Such an arrangement should be on a similar order of depth as a traditional SLR mirror, which means you could maybe use retrofocus SLR lenses (RB67) to get normal/wide views.

Anyway, I suspect it'd be a good idea to implement decent tricolour printing (from consecutive shots on a rollfilm back) before building a camera like this. To me, the camera looks like a minor mechanical challenge of a similar order of complexity as a traditional camera build, just with a few more parts. The camera is of no use unless you can print (scan?) the separations, and the printing I think will be a real challenge.

The advantage of a Philips prism is no loss of image brightness in the separation. All the red light goes out the red path, and likewise for the rest.

So it is a far faster camera. No beam splitter loss and no filter factor loss. Ok, there is probably some loss, but much less. But the prism is very expensive.

 

[holmburgers]

RGB, thanks for mentioning the plastic Curtis camera. I had no idea about that one.

No doubt that the whole printing step is an ordeal in its own right. However, remember, I'm set on this carbon dye-imbibition scheme. Color bromoil would be fantasti michael. You might be able to simply use lithographic process inks for this. Also, making slides might be easier if using a mordant process or even dye-toning. Scanning is also a very real possibility. Who knows how long we might linger in a world where LF color film doesn't exist and LF digital sensors are too expensive. Also, imagine if you could attach light sources to the film backs of a 3-color camera; you'd have a color projector!

But that's not the point of this discussion in my opinion, at present at least. Who knows when & if this will get built, but already we've advanced the effort and contributed something useful to the gods of analog photography. Even if it was only a teaching tool, it'd be worth it.

I think a prism design should be v.2.0. And what about dichroic reflectors? There likely exist reflectors with very high efficiency at transmitting 2 primaries while reflecting 1.

A few more on lenses from Scott B. over at vintagephoto.tv:

I have two 6.5x9cm Devins and they both have 5 1/4 inch Hugo Meyer Aristostigmat lenses (133mm).

Two of my 2x3 inch Curtis cameras have 7 1/2 inch (190mm) Goerz American Optical Dogmar lenses, and one has a 135mm Wollensak Raptar. The camera with the 135mm lens is later than the two with the longer lenses.

I have a 5x7 Curtis with a 14 inch Eastman Ektar.

My Devin and Bermpohl 5x7 cameras also have 30cm/14 inch lenses.

My two 9x12cm Bermpohls have 210mm and 215mm lenses.

The 133mm Curtis Stellar lens is truly unusual in the world of large format lenses as it is a retrofocus design. There is also a 90mm retrofocus Color Stellar lens that was fitted to the 2x3 Curtis. These lenses date to ca. 1956 and were designed and manufactured by Curtis Laboratories. I'm certain that the Stellar lenses were the widest to appear on a 2 or 1 reflector tri-color camera.

Also, I've got a fresh batch of patent #'s to search. I'll try get to this soon.

 

[RedGreenBlue]

I should clarify what I wrote about the 5x7 inch cameras. Above I am quoted as having written 30cm/14 inch. In my email to Chris I meant to write that there are 30cm AND 14 inch lenses.

...Scott

 

[RedGreenBlue]

This is an interesting thread and I think it would be a fun and challenging project.

There is good discussion about what printing method to use. Certainly, the quickest and simplest way to make prints from the negatives would be to scan, assemble in Photoshop and make ink jet prints. It may not be the funnest and wettest method but guaranteed, you would have some beautiful prints.

I recently acquired a homemade bipack tri-color camera. As it's a bipack there is only one reflector. I have not had the time to examine it closely and probably won't in the near future. The photos are from the former owner, the son of the maker. OK, it's not in the same league as a 2 reflector camera, but it's an example of a color project that someone took on and completed. Whether it produced useable images is unknown.

...Scott

 

[ric_kb]

The beam splitter prisms were used in CCD cameras and an alternate "x" version for projectors, combining the sep beams into the single white out...

 

[mr rusty]

just in case you haven't seen it, butkus has the devin manual online! Interesting read. You learn something every day here!

http://www.butkus.org/chinon/devin/devin.htm

 

[holmburgers]

Check out this post over at LFF. http://www.largeformatphotography.i...r-Three-Color)&p=858137&viewfull=1#post858137

A home-made tri-color! (though he's not certain that it's functional) -> http://www.flickr.com/photos/43175600@N00/3403386677/sizes/l/in/photostream/

There are many leads in that post, and I haven't had the time to go through them. I think I need to take a week off work...

Scott, the bipack camera is really something special! And since there are no bi-packs today, you could just use it for 2-color photography and make some Capstaff Kodachromes. It's a busy day in analog photography -> http://new55project.blogspot.com/2012/03/early-color-transparencies-at-george.html#comment-form

 

[greybeard]

Back to your choice of materials for a moment: thermal expansion is not really an issue, but warpage might well be. The odd piece of Delrin (acetal) is not particularly smooth or parallel-sided, and unless you are quite careful about removing the same amount from both sides to flatten it, internal stresses are likely to cause it to warps. Metals which are rolled flat tend to behave the same way, but there is a grade of aluminum alloy called "tooling plate" that is carefully processed to prevent exactly that. It is quite easy to machine, and available in various thicknesses.

The screw-alignment idea is workable but fiddly and probably bulky; you might consider assembling the body, making a final finish pass on each image-plane surface, and then making any needed further corrections using shims. Unless you have access to a coordinate measuring machine, or an optical table with an autocollimator and the ancillary stuff, you probably won't be able to get closer than a half-thousandth shim will take you anyway.

Sounds like a fun project; maybe make the film interface to mate with Hasselblad magazines (you just know what people are going to ask when they see it...)

 

[holmburgers]

If you haven't followed the LFF links to this French website, check out these two specifically, for starters. They're in English, from 1937.

A Practical Bi-Pack Color Camera by Warren T. Mithoff: http://trichromie.free.fr/trichromie/index.php?post/2010/08/17/01

An Experimental One-Shot Color Camera by Warren T. Mithoff: http://trichromie.free.fr/trichromie/index.php?post/2010/08/17/02

 

[holmburgers]

Just wanted to thank everyone for their input so far. Like I said in the beginning, this thread was created as a place for ideation and it's done well so far in that capacity. I'm sure more ideas will come with time, and maybe someday some practical steps.

Until then, we gotta keep shooting E6 and C-41...

 

[Mark Fisher]

If you go with plastic, I'd recommend black ABS. It is easy to machine, tough and bondable. Delrin/acetal machines well, but it is a bit brittle (relatively) and not very bondable. UHMW is can't really be glued at all. Coefficient of expansion really isn't all that important for what you are doing as long as the material is homogeneous (wood is not). For construction, CNC is the way to roll. All this said, wouldn't it be easier to just use three filters and three sheets of film in a 4x5? I realize it isn't as cool, but it would work as long as the scene isn't moving much.

 

[holmburgers]

Thanks Mark; being able to use adhesives would undoubtedly be helpful, although I was imagining that UHMW-P could be bolted together if there was a good way to "sink" thread-inserts into it. But ABS sounds good.

Taking sequential exposures is fine for still lifes, but one-shots were invented to fulfill the need for people & action photography.

Though more importantly I guess, I see the act of building something, discussing some kind of optical system, or investigating any number of seemingly archaic things as an end in it's own right. I'm not necessarily striving to find practical solutions that serve photography, rather, the study is enough. If in the end these investigations can be served by photography, well that's awesome.

 

[Bill Burk]

Make sure the glue is black to light and infrared. Hate to see a light leak from a seam.

How about making a three-lens camera that partitions one sheet of film in thirds.

Angle the three lenses so they aim at one focus point.

Then just sit back and enjoy the tricolor rainbow where parallax throws some of the image out of register.