Possible to manufacture a Horseman 970 Rangefinder Cam?

[abruzzi]

I don't know how hard they are to find or their cost, but Horseman made a ground glass that simply fits in any graflok/internation 6x9 back. I have a couple of them as I've found them useful. I think one of them came from something I bought from B.S. Kumar so you might check with him.

 

[Chuck1]

Be careful there's one for polaroid backs(spaced further than for film)

 

[loccdor]

I did receive the 270mm Tele-Arton today. It's the (I think) newer version that looks like this:

I measured the infinity distance roughly in the 152mm range figure that I found online. Maximum extension on the Horseman 970 of 300mm puts it approximately at 0.915 meters focus. That should make it usable as a portrait lens at a fixed 1.5 meter head and shoulders distance if I calibrate its infinity stop in conjunction with the 1.5m rangefinder setting. All other distances will probably be off.

Don't yet have the lensboard or 180mm cam for it, coming soon.

 

[abruzzi]

The last version has outside printing rather than the inner trim ring.

The oldest one (I think its oldest, its a Compur shutter unlike mine or yours which are in a copal.) has more of the lens behind the the lens board than in front of it.

Schneider 270mm f/5.5 Tele-Arton Linhof-Select GOOD+

Schneider 270mm f/5.5 Tele-Arton Linhof-Select GOOD+ Very pretty 4x5 lens by Schneider. It is marked "Technika" and "Linhof". Speeds sound accurate. Glass is cl

www.seawood.shop

 

[loccdor]

@abruzzi Must be the middle version then. The flange focal distances for the 3 versions I saw floating around were 152mm, 158mm, and 178mm. Could be mine is 158mm, I used a ruler and the film plane was my sweater.

 

[loccdor]

The lens board came today and I was able to take some measurements. The 180mm cam comes tomorrow.

Schneider Tele-Arton 270mm f/5.5 Lens (middle version, black but with text on inside of ring)
On 6x9, this is the 35mm equivalent of a 115mm f/2.4
Flange Focal Distance (infinity): 158mm
Max extension with Horseman 970 when infinity stop is set correctly: 228mm, which focuses 1.65 meters from film plane, 0.14x magnification
Max extension with Horseman 970 when infinity stop is at absolute end of rail: 242mm, which focuses 1.37 meters from film plane, 0.18x magnification, infinity is lost (8.3 meters farthest focus)
Note that this is a with a flat Copal #1 board and you could get it a little farther with a special board
Also note that with the flat board the lens has to be rotated somewhat oddly so that the user has access to all the controls on the small 80mm horseman board

Infinity focus

Infinity focus stop

1.65 meters focus

Odd orientation

This is what AI calculates will happen when I use it with the 180mm cam after giving it all my data. It's interesting that it would apparently give me two usable rangefinder distances (if it's real). I'll test it.

Tele-Arton 270 mm on 180 mm Horseman 970 cam​

Infinity stop set to 158 mm​

Bellows (mm)	Δ ext (mm)	RF indicates	Actual focus	Linear error	% error
158	0	∞	∞	0	0%
165	7	14.3 m	18.6 m	–4.3 m	–23%
172	14	7.7 m	8.3 m	–0.6 m	–7%
180	22	4.8 m	5.0 m	–0.2 m	–4%
190	32	3.6 m	3.6 m	≈ 0 m	≈ 0%
200	42	2.7 m	2.6 m	+0.1 m	+4%
228	70	1.8 m	1.65 m	+0.15 m	+9%

 

[loccdor]

Previous calculation was not quite there. Too much extrapolation. Here is the real world measurement.

Using the 180mm cam, with infinity stop mis-set to maximum of 172mm at end of rail, the rangefinder is accurate at the distance of 11 ft 1 inch (3.38 m), corresponding to 10mm of extension (182mm total). That's a portrait frame 1.02 meters tall and 0.68 meters wide, a "bust" half-length type of portrait. Useful. Focus can be fixed there and used for all shots.

 

[ic-racer]

You should still be able to use the camera's built-in rangefinder. Though it won't be coupled.

With the 180mm cam in place, the rangefinder's readout on top of the camera should indicate the correct subject distance, irrespective of the lens in place.

After focusing with the rangefinder, you could read off that number and then (with a focus-rail mounted distance guide) move the lens to the matching focusing distance.

You could have the AI Bot calculate the distance scale for the lens and affix it to the focus rail. The examples I posted above were for the 4x5 FA Horseman, but something similar could be created for your lens.

BTW, that lens looks great on your camera!

 

[loccdor]

Thanks @ic-racer , that sounds like a good idea, though I'd have to be really precise with it for a 270mm lens. The rangefinder readout's difference between 10 and 11 ft is pretty much just a hair.

 

[loccdor]

I decided I would try to make the 150mm cam, prices for it are crazy. This is the reference image I settled on, it needs to be exactly 40mm wide:

And here I am massacring a filter case in preparation. We'll see how it goes.

 

[loccdor]

Well, that was 3 hours of painful knife work. Soon I'll figure out if it can do the job...

 

[loccdor]

Results look encouraging! It focuses past infinity which means I didn't take too much material off. It fits in place and operates the rangefinder smoothly. Theoretically, all I need to do now is to sand it carefully to the exact curve...

 

[ags2mikon]

If it doesn't work you can always say you hate those filter cases anyway. When you sand that use wet and dry 1000 grit and use it wet. It will take longer but it will be much smother.

 

[loccdor]

If it doesn't work you can always say you hate those filter cases anyway. When you sand that use wet and dry 1000 grit and use it wet. It will take longer but it will be much smother.

Haha! Thank you.

 

[ic-racer]

You might try aluminum. This latch for the ground glass back on my Horseman FA 4x5 camera broke. The original is steel but I made a duplicate out of aluminum with a Dremel and hand tools. Still working fine over ten years.

 

[loccdor]

you are also more than welcome to send me the scan, and i’ll trace it in fusion 360, so that you have a 2D vector (.dxf)

Hey @ivan35mm , I gave up on making a 150mm cam through sanding plastic. I don't think I can get it accurate enough and the material doesn't seem like it will hold its shape for long. I was trying to make a DXF file for sendcutsend.com in Inkscape for this image, to make with steel, but I don't know the program and I'm struggling. Would you be able to make the DXF for me if you have time? The dimensions of the image are 40mm x 18.29mm. The cam is 1mm thick if it matters. Thank you again for the help.

 

[reddesert]

This is what AI calculates will happen when I use it with the 180mm cam after giving it all my data. It's interesting that it would apparently give me two usable rangefinder distances (if it's real). I'll test it.

Tele-Arton 270 mm on 180 mm Horseman 970 cam​

Infinity stop set to 158 mm​

Bellows (mm)	Δ ext (mm)	RF indicates	Actual focus	Linear error	% error
158	0	∞	∞	0	0%
165	7	14.3 m	18.6 m	–4.3 m	–23%
172	14	7.7 m	8.3 m	–0.6 m	–7%
180	22	4.8 m	5.0 m	–0.2 m	–4%
190	32	3.6 m	3.6 m	≈ 0 m	≈ 0%
200	42	2.7 m	2.6 m	+0.1 m	+4%
228	70	1.8 m	1.65 m	+0.15 m	+9%

I wasn't looking at this earlier, because I don't know much about rf cams or the Horseman. However, I am pretty sure that some columns of numbers for focus distance that the AI has given you are incorrect. I don't know if that's due to, e.g., how you set the calculation up, or the problem is too unusual for the AI to parse it correctly, or it interpolates in some incorrect way.

If you set the infinity stop correctly for the lens infinity focus, and extend past infinity by some amount "e" (Delta in your table), and f=focal length, then d_image = f+e, and the subject distance from lens is given by:
1/(d_subject) = 1/f - 1/(f + e)
This is true whether the lens is a telephoto design or not. So for example, looking at the boldface line in your table: for an extension past infinity of 32mm, and f=270mm, it should be focused at 1/d_subject = 1/270 - 1/302, so d_subject = 2.6 meters. This is lens to subject distance, not film to subject, which is 2.9 meters. Either way, not the AI's 3.6 meters. This formula at bellows draw 228mm gives film-subject 1.55m, fairly close to your measurement of focus 1.65m.

Suppose the rangefinder works by sensing the total bellows length, and we set it up using the 180mm cam for a 180mm lens of normal design that has 180mm of bellows draw at infinity. Then on the boldface line, you have 190mm of bellows draw. So the RF "thinks" that it has f=180mm and e=10mm. This gives 1/d_subject = 1/180 - 1/190, so lens-to-subject, d_subject = 3.4 meters. The rf scale might be calibrated to film-to-subject distance, which would be 3.6 meters. That does agree with the AI's number. However, under these assumptions, when you have the bellows at 158mm, that would be too little bellows for a 180mm lens, and the RF should read beyond infinity, not infinity. So I'm also suspicious of the AI's numbers in that column.

With the caveat that I don't know if the RF is exactly keyed off total bellows length. But I suspect it does something like that, I doubt it mechanically senses where the infinity stop is. This is why, in principle, one needs different cams for different lenses of focal length f, because they have different infinity stop locations.

 

[loccdor]

I wasn't looking at this earlier, because I don't know much about rf cams or the Horseman. However, I am pretty sure that some columns of numbers for focus distance that the AI has given you are incorrect. I don't know if that's due to, e.g., how you set the calculation up, or the problem is too unusual for the AI to parse it correctly, or it interpolates in some incorrect way.

If you set the infinity stop correctly for the lens infinity focus, and extend past infinity by some amount "e" (Delta in your table), and f=focal length, then d_image = f+e, and the subject distance from lens is given by:
1/(d_subject) = 1/f - 1/(f + e)
This is true whether the lens is a telephoto design or not. So for example, looking at the boldface line in your table: for an extension past infinity of 32mm, and f=270mm, it should be focused at 1/d_subject = 1/270 - 1/302, so d_subject = 2.6 meters. This is lens to subject distance, not film to subject, which is 2.9 meters. Either way, not the AI's 3.6 meters. This formula at bellows draw 228mm gives film-subject 1.55m, fairly close to your measurement of focus 1.65m.

Suppose the rangefinder works by sensing the total bellows length, and we set it up using the 180mm cam for a 180mm lens of normal design that has 180mm of bellows draw at infinity. Then on the boldface line, you have 190mm of bellows draw. So the RF "thinks" that it has f=180mm and e=10mm. This gives 1/d_subject = 1/180 - 1/190, so lens-to-subject, d_subject = 3.4 meters. The rf scale might be calibrated to film-to-subject distance, which would be 3.6 meters. That does agree with the AI's number. However, under these assumptions, when you have the bellows at 158mm, that would be too little bellows for a 180mm lens, and the RF should read beyond infinity, not infinity. So I'm also suspicious of the AI's numbers in that column.

With the caveat that I don't know if the RF is exactly keyed off total bellows length. But I suspect it does something like that, I doubt it mechanically senses where the infinity stop is. This is why, in principle, one needs different cams for different lenses of focal length f, because they have different infinity stop locations.

Yep, that was off, I think because the AI was assuming some linearity in the cam when in reality it's a curve.

What I found in the real world is that I can mis-set the infinity stop of the Tele-Arton 270mm to 172mm (stop flush with absolute end of rail) instead of 158mm, and the rangefinder becomes accurate at a focus distance of 11 feet, 1 inch. I also did experimenting on the other side with a 150mm Tessar f/4.5 and found that I can set the infinity stop somewhere between 130-135mm and the rangefinder becomes accurate at 1.5 meters. These are both while using the 180mm cam and would be useful for portraits.

 

[Besk]

The infinity stop should be the same. It is the curve of the cam that changes as the standard is racked forward. The cam of a short focal length has a steep drop compared to that of a long focal length lens.

BTW, my friend, who used to make custom cams for various lenses never was able to create a mathematical formula to calculate the curve. Instead he did it the old fashioned way - by hand.

I have a cams my friend made for a Xenar 240mm f/5.5 and one for a Rotelar 270mm f/f5.5 which have a very slight slope to the end versus one for a Schnieder Super Angulon 90mm where the slope stops about 1/3 of the way from infinity.

 

[loccdor]

BTW, my friend, who used to make custom cams for various lenses never was able to create a mathematical formula to calculate the curve. Instead he did it the old fashioned way - by hand.

AI has told me that the curve on the top of the cam is either described by an Archimedean spiral or logarithmic spiral equation. More math than I currently know, but I think I could figure it out if I sat down and studied it for a few days.

I have a cams my friend made for a Xenar 240mm f/5.5 and one for a Rotelar 270mm f/f5.5 which have a very slight slope to the end versus one for a Schnieder Super Angulon 90mm where the slope stops about 1/3 of the way from infinity.

That's quite cool!