How to determine focal length of unmarked lens?

[Account8600]

For an old unmarked lens, how does one determine its focal length? Can it be done without putting it in a camera?

[vinny]

Yes. Hold it up near a wall and find the spot where the image from a window comes into focus. Measure that, you'll be close. I just made that up but I think it works.

[Alan Gales]

I could pretend that I'm smart here but being honest you got me curious so I looked it up on the internet. Vinny is correct. What I found is that for a quick estimate you can focus the lens onto the floor from an overhead light source and measure the distance between the lens and the floor. A window and a wall should work just as well unless of course it's dark outside.

Remember that it's an estimate.

[Leigh]

The distance from the back of the shutter* to the image when focused at infinity** gives you
the "flange focal distance" or "flange focal length" of the lens.

Depending on the design of the lens, that may or may not be close to the optical focal length.

The optical focal length is the distance from the rear node of the lens to the image at infinity focus.
Since you have no idea where the rear node is, you have no way of determining the true FL.***

- Leigh

*NB: If the lens has no shutter, you would measure the distance from the diaphragm to the image.

**Def: Infinity >= 100 times the (assumed) optical focal length of the lens. Farther is better.

***Note: If you really want to measure the optical focal length, and have the time and equipment...

1) Mount the lens on a rail so it can move. Mount a flat piece of white paper (the "film") behind it.
2) Focus an image at infinity on the paper. Mark the lens position as the "zero" reference.
3) Put a flat object slightly shorter than the height of the paper in front of the lens.
4) Repetitively move the object and the lens until you have a focused image exactly the size of the object.
5) The distance the lens has moved from its "zero" point equals its optical focal length.

The accuracy of your result depends on the accuracy of your measurements, which should be 1mm or less.
Use a good magnifier to evaluate image focus.

[Alan Gales]

I agree with you, Leigh. Without knowing the rear node point you can't get an accurate focal length. What I read on the internet it said it would give you an estimate. They suggested measuring from the front of the lens to the image on the floor. Of course the actual distance of the node point to the front of the lens varies with different lenses.

I'm wondering how accurate the estimate could be?

[Leigh]

Hi Alan,

Please review my previous post. I was editing it while you were composing yours.
I added an accurate method for determining the optical focal length.


Of course the actual distance of the node point to the front of the lens varies with different lenses.
I'm wondering how accurate the estimate could be?
The position of the rear node is all over the map, from behind the rear element on short FL lenses to
hundreds of millimeters in front of the front element on telephoto designs.

You really must know both the design of the lens and its intended area of coverage to determine the rear node.
For example, a 200mm lens for a 35mm camera is very different from the same FL for 4x5.

But if you knew that, you would already know its focal length. :)

- Leigh

[Alan Gales]

Thanks, Leigh. What you posted makes a lot of sense to me.

[Emmanuel BIGLER]

Hello from France !

The following diagram is a ray tracing corresponding to Leigh's method
(sorry, the text is in French ; I hope that the ray tracing is self explanatory)
http://www.galerie-photo.com/images/form_opt-2005-08_h013.png

I have, on purpose, chosen an "exotic" case where the principal points (or nodal points) H (=N) and H' (=N') are "crossed" i.e. a situation where H' is ahead of H. This sometimes occur, but has no influence on the proposed method.
The method is very interesting in the sense that you do not have to care for the position of the nodal (or principal) points.

Once an approximate value for the focal length is found, you can refine the measurement by comparison of the image size delivered by another lens of known focal length, for a far-distant pair of objects. But for this you need to mount both lenses on something solid .. like a camera on a tripod ;)

[Mark Sawyer]

Seriously, just point the damned lens at something damned far away, and hold a damned piece of cardboard or something behind it and see where it damned well focuses, damn it. Measure from the damned cardboard to the damned aperture control, or where the damned aperture control would be if there damned well wuz one.

Geez, we're not large format lawyers...

[Leigh]

Geez, we're not large format lawyers...
Well, Mark...

Using your technique on a Nikkor T 12/800mm yields a measurement of 527.4mm (if your results are accurate).

The error is 272.6mm.

That's a damned big error.

- Leigh

[Leigh]

Once an approximate value for the focal length is found, you can refine the measurement by comparison of the image size delivered by another lens of known focal length, for a far-distant pair of objects.
Thank you very much, Dr. Bigler. The ray trace is quite interesting.

A word of caution regarding comparisons...
The measured optical focal length of a lens may differ from the value engraved on the lens.

This appears to have been a fairly recent marketing decision by the lens manufacturers. I expect a lens with nice even numbers (like 200mm) sells better than a lens with odd numbers (203mm), even if it's the same lens.

- Leigh

[Dan Fromm]

The measured optical focal length of a lens may differ from the value engraved on the lens.

This appears to have been a fairly recent marketing decision by the lens manufacturers.
- Leigh

How long ago is "fairly recent?" I ask because the design focal length of the 38/4.5 Biogon has always been 38.5 mm. I've had twenty (20) 38/4.5 Biogons that Zeiss supplied to AGI. AGI measured the lenses' focal lengths and engraved them on the barrels. The range was 38.3 to 38.8 mm. The camera for which these lenses were purchased is fixed focus and the lenses were shimmed to collimated them to it. Each of my little Biogons had its own shim, with thickness and lens' serial number marked on it.

Further on this point, my CEDIS-Boyer fiches techniques give nominal (as engraved in the trim rings or barrels and stated in marketing materials) and actual (+/- 1%) focal lengths for Boyer's product slate. I have a few process lens QC slips that give actual focal lengths. None matches nominal.

Even without more and older data, I find it easy to believe that nominal and actual focal lengths have rarely been identical.

Cheers,

Dan

[Tim Meisburger]

I'm with Mark. If its an old lens and unmarked the wall method works fine (often these are projection lenses). This is a practical method if you need to know if a lens will work on your camera. If its a new exotic lens, its easier to just read the engraving on the barrel.

[domaz]

How long ago is "fairly recent?" I ask because the design focal length of the 38/4.5 Biogon has always been 38.5 mm. I've had twenty (20) 38/4.5 Biogons that Zeiss supplied to AGI. AGI measured the lenses' focal lengths and engraved them on the barrels. The range was 38.3 to 38.8 mm. The camera for which these lenses were purchased is fixed focus and the lenses were shimmed to collimated them to it. Each of my little Biogons had its own shim, with thickness and lens' serial number marked on it.

Further on this point, my CEDIS-Boyer fiches techniques give nominal (as engraved in the trim rings or barrels and stated in marketing materials) and actual (+/- 1%) focal lengths for Boyer's product slate. I have a few process lens QC slips that give actual focal lengths. None matches nominal.

Even without more and older data, I find it easy to believe that nominal and actual focal lengths have rarely been identical.

Cheers,

Dan

To give another not recent example all the Wollensak APO Raptar's I have seen have the exact focal length handwritten in white-out like ink on the rear element. The focal length's seem to vary in about a 10mm range from the stated focal length's. Here's a page with a photo example (http://lensn2shutter.com/15inaporaptar.html) of such writing.

[Mark Sawyer]

Well, Mark...

Using your technique on a Nikkor T 12/800mm yields a measurement of 527.4mm (if your results are accurate).

The error is 272.6mm.

That's a damned big error.

- Leigh

Well, if you're going to go all telephoto on me... :rolleyes:

Sorry about my last post, but ray tracings and nodal points are a little overboard for someone who just wants to know the focal length of an old lens.

[Leigh]

How long ago is "fairly recent?"
I'm sure that varied by manufacturer. I was thinking generally after WWII.

l was thinking in particular of some lenses for the Graphic press cameras, which had odd FLs (152 or 157).

For example, in the Zeiss Hasselblad lenses, the lens marked 150mm is actually 151.2mm (151,2 if you prefer).

- Leigh

[Leigh]

Sorry about my last post ...
It was a bit much.

... ray tracings and nodal points are a little overboard for someone who just wants to know the focal length of an old lens.
The OP's question, which I attempted to answer, was "How to determine focal length of unmarked lens?"
It was not...
how to guess
how to approximate
how to fudge.

Given that the whole process I described only has five steps, I don't think it's excessively convoluted.

Forgive me for trying to provide an accurate answer.

This board seems to have a higher percentage of imprecision and WAGs* than others I frequent.
That's probably because photography is both art and science, and artists are usually imprecise.

- Leigh

*Def: WAG = Wild Ass Guess

[jnantz]

Well, if you're going to go all telephoto on me... :rolleyes:

Sorry about my last post, but ray tracings and nodal points are a little overboard for someone who just wants to know the focal length of an old lens.

+1

and what you said before has worked on every lens i have owned ...
(except a symmar when the front cell is removed )

[Mark Sawyer]

I guess one person's Wild Ass Guess is another person's Close Enough For All Practical Purposes...

[hoffner]

This board seems to have a higher percentage of imprecision and WAGs* than others I frequent.

- Leigh



There is a good reason to it. There was a time when knowledgeable intelligent people participated more often. For reasons not difficult to know, that time is largely over for many of them.

[jnantz]

this thread has made me curious ...
are there very many lenses out there that are OLD and UNMARKED
and have their focus node farther back than "normal"
i know symmars when converted have a far-back focal node
far enough back that it the 370 needs more than 445mm of bellows to focus at infinity ...

but that is a converted lens .. i have heard it referred to as a plasmat design ...

its too bad there isn't a thread somewhere that points out a variety of modern and historical designs
and suggests if the focus node is way back or up front ...

i've never had trouble using the window and white wall method ... but i hoped the nodes were normal

[Leigh]

I guess one person's Wild Ass Guess is another person's Close Enough For All Practical Purposes...
I am all in favor of practical porpoises.

They're fun to watch, jumping out of the water and such.

- Leigh

[Emmanuel BIGLER]

this thread has made me curious ...
are there very many lenses out there that are OLD and UNMARKED
and have their focus node farther back than "normal"

Hi John.

Telephotos are as old as the 1890's.

In my 360 [actual F.L. = 353 mm] Schneider-Kreuznach Tele Arton, the image nodal point N'=H' is floating in air 63 mm in front of the first lens vertex and the flange focal distance, very close to the distance between the iris and the focal point, is only 210 mm. And the entrance pupil is of course not located where the iris is: it is located in air 6 mm behind the last lens vertex, inside the camera ;)

But, too bad, the 360 Tele-Arton is both modern, and well-documented!

Old umarked lenses are really facetious.
Old lenses usually have signed a secret agreement with Dr. Murphy, stating:
"In order to supply ammunition to endless discussions on the US LF forum, any old & unmarked lens will have its nodes very far away from the iris".

And since we are on the US forum, here is an updated version of my ray tracing diagram with texts in English.
https://farm4.staticflickr.com/3922/14210410819_b0608e4366_b_d.jpg

The excellent technique explained by Leigh is as old as the 1870's, and appears in numerous French textbooks & Internet web sites under the name "Méthode de Davanne & Martin".

Remember during the Cold War and Soviet times: there was always a Popov scientist or engineer from Russia who had invented everything before Western inventors did.

In Photography we have in France a lot of, not Popov, but Martin inventors: Martin is the most common family name in France (even Dupont is not as frequently encountered as Martin ;) )

Hence my diagram cannot compete with the excellent description of Davanne & Martin's method by M. Charles Fabre in his huge "Traité encyclopédique de photographie, complément A, 1892".
I love those old engravings in scientific texbooks of the XIX-st century. This reminds me illustrations in Jules Verne's novels, at a time when Science & Progress were supposed to bring all possible Good to Mankind, without any evils.

http://www.suaudeau.eu/memo/Fabre/fabre_objo/Fabre_obj_a/019.html

M. Fabre uses the word "glace" which does not mean "ice" but a flat piece of glass of good quality, like in "galerie des glaces à Versailles". glace-glass is a typical faux ami in modern times, but was not in the good old days.

The whole book as well as other works by Charles Fabre is accessible here, thanks to M. Suaudeau.
A real photographic gold mine, for those who can read French, but the images themselves are worth looking at!
http://www.suaudeau.eu/memo/Fabre/Fabre.html

The section dedicated to shutters is fascinating!
http://www.suaudeau.eu/memo/Fabre/fabre_obtu/Fabre_1889/150.html

Have fun !

P.S. I have in my attic another amusing method to guess-estimate the focal length of an old facetious unmarked lens, where the lens is mounted on a camera & tripod with panoramic head; you measure the displacement of an image on the ground glass, then measure the angle of rotation of the whole camera, and find the focal length, again without any knowledge regarding the nodes or the iris.

[jnantz]

hi emmanuel

thanks as always !
those methods make perfect sense
but it seems like an awful lot of work ...
et je suis vachement paresseux .. :)

john

[Steven Tribe]

P.S. I have in my attic another amusing method to guess-estimate the focal length of an old facetious unmarked lens, where the lens is mounted on a camera & tripod with panoramic head; you measure the displacement of an image on the ground glass, then measure the angle of rotation of the whole camera, and find the focal length, again without any knowledge regarding the nodes or the iris.

I think you are (attempting?) to describe Sir Howard Grubbs' method of "A mechanical means of estimating conjugate foci". Best seen in Chapter XX in Trail Taylor's collection of articles "The Optics of Photography".

[Emmanuel BIGLER]

but it seems like an awful lot of work ...

Exactly like making pictures with a LF camera: such an awful lot of work, while you could do almost as good with a smartphone ;)

[Emmanuel BIGLER]

I think you are (attempting?) to describe Sir Howard Grubbs' method of "A mechanical means of estimating conjugate foci".

Thanks for the info, I have the deepest respect for Sir Howard Grubbs, but I'd never heard about him.
I found the text, dated 1904,
https://archive.org/details/opticsofphotogra00taylrich

The method "A mechanical means of estimating conjugate foci" is described on pages 116 and subsequent,
https://archive.org/stream/opticsofphotogra00taylrich#page/118/mode/1up
but this is not what I'm speaking bout.
However I admire the simpicity of this mechanical system which is an analog mechanical calculator for object / image relationship according to Newton's formulae.

And I cannot resist: what do you think of this sentence:

In a combination of lenses, whether symmetrical or
non-symmetrical, there is no fixed point which can be
termed the 'optical centre.' The mistake, however, is
frequently made of assigning it a position where the
stop is placed.

Isn't it marvelous, the answer to one of our FAQs, clearly stated in a 120-year old British textbook ;)
and the textbook starts as:
This little treatise is not theoretical but practical,
and it is not intended for the makers but the users of
photographic lenses.


---------------

I've tested the rotation method I'm speaking about (I'll post a drawing soon.) without any prior knowledge of this old but invaluable litterature. I tested with my tele-arton 360 and of course the method works fine.
At least if we only want something like an estimation of the focal length within an accuracy of 10%, which is presumably largely enough for what we are interested in LF photography.

[jnantz]

Exactly like making pictures with a LF camera: such an awful lot of work, while you could do almost as good with a smartphone ;)

thanks for the nudge emmanuel ;)

that's the easy part, wanting / needing to have exact /precise measures for FL/ Fstops &c is the hard part ( for me at least )
at a certain point, i gave up even caring what the FL/fstops on my lenses --- and i just am used to whatever
FL it might be, and whatever exposure the iris might give me ( i have a general idea but not an exact measure ) . since i shoot paper and glass + expired films a lot i have a little more "wiggle room" than someone exposing a material with very little room for error ...
not to mention the glass and paper i coat myself
doesn't really have a fixed iso, and it varies depending on the location atmospheric conditions, time of year +day. so in the end i don't mind being precisely imprecise ;)

like with everything LF or otherwise YMMV

[Account8600]

I thank you all for responses to my question on determining the focal length of an unmarked lens. I did not know I was opening a Pandora's box. The exhibited knowledge (at least most of it) was remarkable.

For the record the lens in question came on a Kodak View #2 view camera (improved model) I acquired at an estate sale, and I am now restoring. The lens/shutter is the Victor model from Bausch & Lomb, but has no other markings. Even though it was on the camera, I am not sure the image would fill the 8x10 size film, so I started my journey into focal length, image sizes, and nodal points.

Thanks again for your interest and help.

[Mark Sawyer]

I thank you all for responses to my question on determining the focal length of an unmarked lens. I did not know I was opening a Pandora's box...

Don't worry about it, we argue about everything here! Part of the fun, and hopefully no one takes it too seriously. Good luck with your lens quest! Here's a handy table listing quite a few lenses and their coverage:

http://www.graflex.org/lenses/lens-spec.html

For 8x10, you'll need a minimum of 320mm, but more is nice for movements. BTW, some lenses have more coverage than their manufacturer's specs, (and of course, some don't...)

[Leigh]

BTW, some lenses have more coverage than their manufacturer's specs, (and of course, some don't...)
The image circle diameter that appears on manufacturer's datasheets may well be smaller than the illuminated circle.

Nikon in particular defines their IC based on the image quality, which will meet specs within the defined area.
This is smaller, sometimes significantly smaller, than the illuminated circle.

- Leigh

[Mark Sawyer]

Indeed, though their criteria may be different than a photographer's. The Nikkor 450-M, for example, has a listed image circle of 440mm, but some very good photographers with critical eyes use them on 16x20, which requires an image circle of over 600mm. And of course, some photographers like the qualities of the illumination outside the circle of sharp coverage, even when it gets extreme, as is currently popular with the Petzval lenses.

Oh boy, another Pandora's Box... :rolleyes:

[Leigh]

And of course, some photographers like the qualities of the illumination outside the circle of sharp coverage, even when it gets extreme, as is currently popular with the Petzval lenses.
That's certainly true, an excellent example of the "artistic" aspect of the photographic discipline.

Of course, a lens manufacturer cannot evaluate that aspect in any reasonable manner.
They can only provide scientifically-measurable parameters. They can't assess how a user will interpret that.

- Leigh

[Cor]

Let's throw in my 2 cents (well actually I am quoting Richard Knoppow here):

Here is how to determine the focal length and size of the
entrance pupil of any lens.
First, determine the focal length by autocollimating. This
requires a mirror capable of covering the front of the lens.
It can be done using a view camera. A first surface mirror
is ideal but a standard makeup or shaving mirror will do.
Just be sure to use the plane side, not the magnifying side.
Place the mirror over the lens. Distance does not matter
but the mirror should be parallel.
Place a small light, a pencil flashlight is ideal, against
the ground glass, near but quite at the center. The mirror
will reflect an image of the light back to the ground glass.
For easier focusing you can draw a small cross on the ground
side with a pencil (will come off again). Focus the image as
sharply as you can. This focuses the lens _exactly_ at
infinity. Mark this distance at some convenient point on the
camera.
Now, set the camera up for an exact 1:1 image of some
object. A small ruler is handy since you can tape a similar
one to the ground glass to match it. When adjusted for 1:1
the diference between the focus point and the infinity focus
point will be _exactly_ the focal length of the lens.
To measure the _effective_ size of the stop set up a
pinhole source at the exact infinity focal plane (which you
determined in the first measurement). Place a sheet of some
translucent material over the front of the lens, and record
the diameter of the circle of light. This is the effective
size of the aperture. Devide this into the focal length to
get the f number. The size of the effective aperture may not
be much different than the physical aperture in many lenses
but the difference will be significant for some. This method
takes into account the magnification of the stop by the
lens.
If you make a test stop of known size you can determine
the difference between the physical size and effective size
and use that to calculate the size needed by the actual
stops.

Since you now know the focal length you can determine the
locations of the principle points by focusing the lens
exactly on infinity and measuring back one focal length
toward the lens from the focal plane. That is the location
of the principle point for that end of the lens. When the
lens is in its "normal" position this is the second or rear
principle point. To get the first or front principle point
turn the lens around and refocus it. Knowing the principle
points is sometimes useful.
The internal structure of lenses which are not too
complicated can often be determined by shining a small light
into each end of the lens and counting the reflections.
Easier if you can examine the cells separately. Glass/air
surfaces, even when coated, are bright, cemented surfaces
are dim. Sometimes its possible to determine the power, or
rather the sign, of the surface this way, but it can be
tricky.
My guess is that this is probably a Petzval portrait lens.

---
Richard Knoppow
Los Angeles, CA, USA

[Leigh]

Here is how to determine the focal length and size of the entrance pupil of any lens.
...
When adjusted for 1:1 the difference between the focus point and the infinity focus point will be _exactly_ the focal length of the lens.
That is the same method that I presented in post #4, although I omitted the entrance pupil calculation.

- Leigh

[Dan Fromm]

Leigh, photography forums are echo chambers.

Cheers,

Dan

[Leigh]

Leigh, photography forums are echo chambers.
Hi Dan,

I do believe all fora are echo chambers. :cool:

- Leigh

[Jerry Bodine]

I used a method similar to Cor's and Leigh's to check the focal length of a Componon enlarging lens (labeled 105mm) that, for some reason, was not conforming to the formulas for determining "magnification vs. neg-to-easel distance", which I use to find the exposure change when changing the enlarger head height, although all of my other enlarging lenses did conform. Since I use a monorail camera (Sinar Norma) it was easy to set it up with the Componon focused on infinity out the window and mark the position of the front standard on the camera's scale. Then I used an auxillary frame in front of the front standard to support a ruler and moved the front standard until the gg image of the ruler was at 1:1 and marked that position of the front standard. The distance between the two marks was the true focal length of the lens. The focal length was actually 110, not 105. Then it conformed to the formulas. I didn't bother to checkout the true aperture.

[Louis Pacilla]

Leigh, photography forums are echo chambers.

Cheers,

Dan :rolleyes::)

[c.d.ewen]

Louis: You are a subversive ;)

I might as well throw in another FL & f/stop method, although stepping into a discussion that quotes Mr. Knoppow is presumptuous of me. I just find this method easier.

Go find a distant landscape scene (focus at infinity), with a couple of widely separated prominent points, and take a picture of it. Take out your sextant (if you've really got GAS, you've got one) and shoot the angle between the points. Scan the negative and using PS (or a good ruler), measure the distance between the points. They don't have to be equidistant from the center, but it'll make you feel better if they are. Divide the angle and distance in half and, son-of-a-gun, you've got the angle and base measurements of a right triangle. Do the math.

A variation on Mr. Knoppow's method is to focus the lens at infinity, then take the camera into the darkroom. Tape a piece of opaque paper with a centered pinhole to the GG. Put a piece of photo paper across the front of the lens, and shine a flashlight through the pinhole. Develop the paper and measure the diameter of the exposed hole. Do the math.

Charley

[Emmanuel BIGLER]

Go find a distant landscape scene (focus at infinity), with a couple of widely separated prominent points, and take a picture of it. Take out your sextant ...

C.D. Ewen's method is equivalent to the rotation method I was referring to previously.
Instead of measuring the angle between two distant points, I measure the image point separation on the ground glass for the image of a single distant objet, when I rotate the whole camera on a tripod + panoramic head.
For this angular measurement, you simply need a wooden stick attached to the camera back and sliding on top of a table, you measure the displacement of the end of the stick on the table, and you can even do the maths without any calculator.

If "L" is the length of the wooden stick and "d" the displacement of the end of the stick on the table, the angle is radian is approx d / L if d is kept not bigger than about L/3.
And if AB is the displacemetn of the image on the ground glass, then the focal length is approx
AB . L / d

Imagine that the displacement of the image AB is 10 cm = 4", and that the wooden stick si 0.9 m = 3 feet long.
If the stick moves by 30 cm, 1 foot, the angle is approx 1/3 of a radian, then the unknown focal length is approx
10 cm x 90/30 = 30 cm = 1 foot = 12 inches.

The method is less and less precise for small focal lengths but often old unknown lenses were in a range of focal lengths above 150 mm ... never in the range of 15 to 50 mm.

And again you can rotate around any point, you do not need to know where the nodal points or the pupils are. Provided that the object is really far-distant. At home through the window looking to my old city, I have the choice of various towers, spires etc that serve as excellent test-objets at infinity.
I even have old tiled roofs serving as resolution test targets ;-) (this picture was not taken from home, though, but from this elevated viewpoint I have the best choice of old tiled roofs, including Burgundy-style, colored-varnished tiled roofs!) (https://www.flickr.com/photos/43175600@N00/399611771/sizes/o/)

[ailsa]

such a right calculator