Determining the focal lenght of a brassie

[argus]

I bought this nice looking brass lens with perfect glass yesterday, but there is no trace of the focal length.
- manufacturer: E. Peters (Bruxelles) so it must be Belgian
- f-stops range 6 - 8 - 11 - 16 - 22 - 32
I can see 4 reflections in the glass and the lens covers 8x10 with ease.

I am not great at math, so tried average out the focal lenght with this simple indoors test:
subject is a painting on the wall, 7 meters away.
size on the groundglass with a 300mm lens is 43 mm (scale appr. 1:18)
size on the groundglass with 360mm lens is 55mm (scale appr. 1:14)
size on the groundglass with ? mm is 72mm (scale appr. 1:10)

Is there any way to compute the focal length with given data?

I tried the simple rule of the thirds, but the outcome is weird:
(300 / 43)*72 = 502
(360 / 55)*72 = 471

Could the bellows extension influence this result?

Thanks,
G

 

[Ole]

The bellows extension does influence the results.

The perfect way to do this is for focus on infinity - which means you need a very distant subject.

The imperfect way is the way I do it: I have more lenses, and have made myself a "calibrated focal length curve". To keep it in calibration I don't use a camera, but project the image of a window onto a wall. Neither is likely to move out of position.

 

[jimgalli]

I cheat for the quick and dirty answer. Set the aperture at f10. Measure the opening. Multiply by 10. ie. 38mm aperture at f10 = 380mm lens.

 

[Struan Gray]

The size on the ground glass should be equal to f*tan(theta) where theta is the angle subtended by the object as seen from the lens' position. If you only move the rear standard to re-focus I would be surprised if theta changed enough to make a difference. If you moved the front standard, or if the brassie is so cantilevered out in front of the lensboard that it 'sees' from a significantly different position, you will get an error when you assume simple proportionality to f.

You can either find the entrance pupils and do the maths, or, easier, find a more distant object for your tests. Ole's method is very accurate once you have enough 'known' lenses to plot the curve, but with only two knowns you get a straight line, which just your stated problem in graphical form.

PS: a goodish measure of how far away is far enough is to use the hyperfocal distance for the f-stop you measure at.

 

[Struan Gray]

A quickie calculation.

f = MSo/(1+M)

where, f=focal length, So=subject distance, M=magnification.

So is measured from the front principle point, which for practical LF purposes is also the centre of the entrance pupil, which is where the center of the aperture appears to be when looked at through the front element.

I get So to be 5.7 meters for the 300 mm lens and 5.4 meters for the 360 mm lens, so you're too close to ignore small shifts in the position of the lens, even allowing for slop in the magnification measurement.

5.7 meters gives 520 mm for the focal length of your brassie, 5.4 gives 490 mm.

480 was a common enough focal length for longer lenses, but with an unknown maker all bets are off. I have a Belgian portrait lens made by OPI in Gand/Ghent, which is 440 mm, a length I have never seen elsewhere.

Go long son, go long.

 

[argus]

Thanks a lot for the answers, I'll take Jims suggestion and measure the aperture tonight.

G

 

[Early Riser]

I always thought that the distance from the focal plane to the nodal point, with the lens focused at infinity was the focal length. Therefore if you focus at infinity and measure the distance from the film plane to the nodal point on the lens, you'll have your focal length. BTW if you focus on something close up and render it at 1:1 in the ground glass, that distance between film plane and nodal point is twice the focal length.

This does not work though with true telephoto lenses or retro focus lenses.

 

[Struan Gray]

Finding the nodal point is left as an exercise for the reader.....

 

[noseoil]

Just measure from the film plane to the the lens board which is close enough. Why worry about the focal length? Just use the lens and see how the prints come out! tim

 

[Nicholas Lindan]

I bought this nice looking brass lens with perfect glass yesterday, but there is no trace of the focal length.

In addition to other methods:

Get a 1:1 image on the ground glass. The focal length is 1/4 the
film to subject distance.

It has the advantage of not needing to know anything about the lens's esoteric points, nodal or otherwise. Works with telephoto and retrofocus lenses.

The principle point is in the middle - With all the points being bandied about I am not sure if 'principle' is the right word, but it is the place where a pinhole with the same focal length as the lens would be.

 

[Ole]

The principal point doesn't have to be in the middle. With more-or-less symmetrical designs it's usually fairly close to the middle, often it's not. Besides, there's two of the little buggers.

The same goes for the lens board. Most modern lenses have the principal points somewhere close to the lens board, but most "brassies" do not. The aperture can be assumed to be fairly close to the center, but in some cases it's very far from it (e.g. the Steinheil Gruppen-Antiplanet I got myself for chrismas).

The 1:1 (FL=1/4 film-to-subject) method is fine in theory, but impossible in practice. You forget that glass has thickness, and that 1cm of glass corresponds to about 1.5cm air. So you will need to correct your measurement for the optical thickness of the lens. This makes finding the nodal points easy by comparison. And I haven't even mentioned the practical difficulties involved in getting exactly 1:1...

My suggetion is to get another longer lens to compare it with, so you can make a "calibrated curve". See if you can find a +1 and a +2 close-up "filter" in a rummage bin. Or an old "Proxar" - that's the same thing. A +1 is a 1000mm lens, a +2 is 500mm. Very nice to have for those FL measurements...

 

[argus]

Just measure from the film plane to the the lens board which is close enough. Why worry about the focal length? Just use the lens and see how the prints come out! tim

That's what I intended to do but I want to prevent myself from buying a next brass lens in the same focal length.

According to Jims' method, it would be a 450mm lens.

Thanks for helping me out in this!

G

 

[Nicholas Lindan]

> The principal point doesn't have to be in the middle.

Not the middle of the lens - I worded that a bit confusingly, but equidistant between the subject and the film plane.

> The 1:1 (FL=1/4 film-to-subject) method is fine in theory, but impossible in > practice.

All I can say is I use it and it works well. I just went and tried again as a reality check: Lens marked 105mm => film-to-subject is 42.9cm / 4 = 107.2mm.

Maybe it is one of those things that works well in practice but is impossible in theory?

> And I haven't even mentioned the practical difficulties
> involved in getting exactly 1:1...

???? I must be missing something. I find it pretty easy, takes 10 seconds or so. Easier if you have a geared rail-to-tripod mount, which I don't.

 

[Nicholas Lindan]

Well, it does have one practical limitation: If it's a 400mm lens then the method requires 800mm [~ 2 1/2 feet] of bellows draw...

 

[Ole]

> The principal point doesn't have to be in the middle.

Not the middle of the lens - I worded that a bit confusingly, but equidistant between the subject and the film plane.

Which one of them?

Lenses have two principal points, and the distanec between them can be considerable: http://www.schneideroptics.com/info...e_format_lenses/tele-xenar/data/5,5-360mm.htm

Even something as symmetrical as an old Symmar has a measurable separtion: http://www.schneideroptics.com/info/vintage_lens_data/large_format_lenses/symmar/data/5,6-360mm.html

If the rear cell is stronger than the front cell (true for most LF lenses), the lens will be slightly retrofocus. This means that the front nodal point is behind where it would be in a theoretical flat lens.So the extension at 1:1 will be slightly more than 2x focal length, just as the extension at infinity focus is slightly more than the focal length. But the difference between those two extensions should be one focal length - regardless of optical construction!

When using what is just about "the ultimate monorail" I find it very difficult and time-consuming to achieve an exact 1:1 ratio. I find it far easier with that camera to identify the front principal point through the adjustable axial front tilt axis, and measure that distance...

 

[phfitz]

Hi there,

A quick and dirty, fast and nasty way that works every time.

attach a mirror to the front of the lens
hold a point-light sorce (Mag-Light with the head removed) to the G.G. and focus the returned image
measure from GG to aperture blades or waterhouse slot

more than close enough for government work.

 

[Nicholas Lindan]

Ole - 419882 wrote
> Lenses have two principal points, and the distanec between
> them can be considerable.

Lagrange's two principle points/planes model of a lens applies to paraxial rays and small angles. The rear - P2 - point is where one would swing the lens for taking panoramas and is the 'effective focal length' of the lens when used at infinity.

At 1:1 there is only one principle point/plane: not hard to see as the distance from the lens to the subject is equal to the distance to the object so both principle points must be at the same point.

At points between 1:1 and infinity the P2 point shifts toward the film plane.

So you are correct, the principle point at 1:1 isn't, for practical lenses, at the same location as the P2 point.

> When using what is just about "the ultimate monorail" I find it very difficult
> and time-consuming to achieve an exact 1:1 ratio.

Ah, I think I see your difficulty. If you use a Sinar style camera and focus with the back standard then finding 1:1 will drive you balmy.

Move the back standard to alter magnification and focus with the front standard and you will find you can zoom in to 1:1 with no problem. It is the way magnification is set on an enlarger - something one does every day.

 

[Ole]

> When using what is just about "the ultimate monorail" I find it very difficult
> and time-consuming to achieve an exact 1:1 ratio.

Ah, I think I see your difficulty. If you use a Sinar style camera and focus with the back standard then finding 1:1 will drive you balmy.

Move the back standard to alter magnification and focus with the front standard and you will find you can zoom in to 1:1 with no problem. It is the way magnification is set on an enlarger - something one does every day.

No, I use a Carbon Infinity. I can focus with either standard.

Finding something very close to 1:1 is easy. Finding something close enough to 1:1 to get a reproduction ratio that is 1:1 for all practical photographic purposes is also easy. But finding 1:1 exactly enough to use it to measure the focal length is a whole different question.

With the CI it's actually easier to determine the "swing point" at infinity and use that for measuring the focal length!

 

[Helen B]

Lagrange's two principle points/planes model of a lens applies to paraxial rays and small angles. The rear - P2 - point is where one would swing the lens for taking panoramas...

...only for swing lens cameras, of course. (Just a point of clarification, not of disagreement)

At 1:1 there is only one principle point/plane: not hard to see as the distance from the lens to the subject is equal to the distance to the object so both principle points must be at the same point.

At points between 1:1 and infinity the P2 point shifts toward the film plane.

I think not. The lens doesn't change its principal points depending on where it is focussed. That's why this method won't work for telephoto or retrofocus lenses: significant (ie significant in relation to the distances that are being measured and computed) principal point separation.

Best,
Helen

 

[Nicholas Lindan]

> only for swing lens cameras, of course. (Just a point of clarification, not of disagreement)

??? You can make a stitched panorama with any camera... To get things to line up at the seams you swing the camera around the rear nodal point.

> I think not. The lens doesn't change its principal points depending on where it is focussed.

The location of the paraxial infinity-focus principal planes don't change. But, they only have validity at infinity focus. If Lagrange's thick lens model is used at other distances the location of the principal planes as they apply to tracing the path of rays to and from [but not inside] a lens are at a different location - because of this the model is only used at infinity. I've never applied it except with colimated light. I just did a thumb-through - every optics book I have only uses this model for infinity subjects.

Set the camera to 1:1 and find the principle point by rotating the camera - it's not in the same place it is when the camera is focused for infinity.

> That's why this method won't work for telephoto or retrofocus lenses

Ah, er, don't know what to say... The whole reason for using this method is because it does work with retrofocus/telephoto lenses. It's not a point of argument or discussion. The sun rises, bears poop in the woods ...

It's fast, it's _plenty_ accurate for determining focal length. I am flumoxed by the idea that it's not accurate or easy to do.

Grab a camera and try it. It's not hard to do. Try it with your enlarger. Try it with a 35mm and a bellows attachment/extension tube/ what have you...

... this thread is turning into a Kafka novel.

 

[Helen B]

Nicholas,

I thought that I would come to regret writing that post, and even as I was writing it I thought "something's wrong, don't write this". Furthermore, many telephotos have more-or-less coincident principal planes at infinity. What was I thinking? It's a strange sensation. Age? Migraine? I should just spend the day in bed.

"You can make a stitched panorama with any camera... To get things to line up at the seams you swing the camera around the rear nodal point."

The position of the rear nodal point is of no consequence if the lens is fixed with respect to the film - as is the case with 'normal' cameras. In that case the correct point of rotation is the centre of perspective of the lens: the entrance pupil. That's why I wrote my attempt at clarification.

The change in perspective for cameras that rotate the lens about the rear nodal point is not perceptible on film because only a narrow slit of film is exposed at once - for those swing lens cameras the image-side relationship is the most important.

Best,
Helen

 

[Helen B]

I'm having a bit of difficulty understanding why the principal planes should always become coincident at 1:1, and would be genuinely grateful for an explanation or reference - I have most of the standard optical reference books. I'm also a little puzzled about why you are calling them Lagrange's principal planes rather than Gauss's. Maybe we are referring to two different properties?

Thanks,
Helen

 

[Dan Fromm]

Helen, I think Nicholas is assuming that internodal distance is zero. Whatever it is, it doesn't change with focused distance.

Or perhaps he's confused because at 1:1 it is the case that subject-to-front node distance = rear node-to-film distance. Many people have referred to the rear node on bulletin boards as the lens' "optical center;" once they start using that terminology they're at risk of, um, equivocation.

I continue to admire your patience.

 

[Helen B]

I wonder if this is the source of confusion: If you look through lens data you will see that some lenses do indeed have their principal point data marked 'at infinity'. That's because those lenses have floating elements, usually for close range correction in the case of a lens intended for still photography. A 'fixed focal length' lens for cinematography can have up to three floating groups. When you pull focus during a shot the edges of the frame need to stay in the same position - ie the angle of view should not change.

Lenses that do not have any floating elements have fixed principal points, as already mentioned.

Best,
Helen

 

[juanito]

Hi Argus,

According to Michael J. Langford's Basic photography book(chapter 3 in my spanish version), the way to know the focal length of a lens knowing his f No. is to multiply ths f No. by the effective diameter of the lens wide open.
He wrote this formulae:
f No.=Focal length/effective aperture diameter.
So in your case just mesure the effective aperture diameter of the lens and multiply it by 6, not by 10.

Juanito