How to determine true focal length

[couldabin]

I am building a panoramic camera that will use 120 film, and need to determine the true focal length of the lens being used. I'm starting with one of my 90mm 4x5 lenses. I came up with a methodology that may well be more Rube Goldbergesque than is necessary: photograph a landscape with the lens focused on the far distance, then use a transit (I happen to own one) to measure the arc between landmarks in the scene. The trick is to determine how far, physically, those landmarks are on the negative, which then allows you determine the focal length. I have an Epson 4990 but don't know if the I in DPI is truly ironclad. I scanned my negative at 4800 dpi, and if that provided an accurate measure of the distance, my 90mm is actually 87.65mm. Does anyone know if the DPI value can be trusted? Are there more elegant ways of establishing the true focal length?

 

[BradS]

The flange focal length is relatively easy to determine empirically...

1. mount the lens on your 4x5 camera
2. focus on something far, far away
3. measure the distance between the flange (the front of the lensboard) and the film plane

That's it!

If the lens is modern and made by one of the respected manufacturers, then specfifications are usually published and available...one simply needs to ask or dig with google.

 

[bernard_L]

Your method is correct in principle; For an overview of other methods, see, e.g.:http://fp.optics.arizona.edu/optomech/student reports/tutorials/Tutorial_Morel,Sophie.pdf
BUT... could you tell us why the difference between 87.65 and 90 is so important to you? do you plan to use your camera for photogrammetry? Then you should also worry about high-order distortions. If, in the spirit of the response of BradS, you are concerned about mounting your lens at the proper distance, and since your lens is not a "thin lens", the mounting distance has little to do with the focal length. See, e.g. telephoto or retrofocus designs.

To verify the dpi of your scanner, (to a fraction of %) scan a ruler.

 

[removed account4]

hi duane

i use a white wall and an window ..
i project and image from something far away onto the wall.
i take a tape measure and i measure the distance from the aperture
to the wall.
its pretty much like bradS says, but i do it against the wall.
its a crude measurement of focal length because it doesn't take into account
the lens design or that in some lenses the focal nodes might not be in the same place ...

have fun and good luck with your project !
john

 

[RobC]

I suspect OP is wanting to use a helicoid lens mount such as following and it is important that when lens is focussed at infinity, it's the correct distance from film. This has to be designed into the camera body front flange to film distance taking into consideration helicoid depth and lens rear flange to film distance. If ain't built to perfect tolerances you'll never get sharp at infinity with lens wound all the way in. Maybe design in some adjustment so it can be set accurately. Replaceable shims may be a good idea so it can be adjusted for each individual lens of same or close focal length.

http://www.schneiderkreuznach.com/e...t-lenses/helical-mount-and-accessories/#c1691

 

[couldabin]

BUT... could you tell us why the difference between 87.65 and 90 is so important to you?

The panoramic camera will have a film transport that moves the film as the camera body rotates. It goes without saying that the rate of movement has to be correct. The rate is a function of the image distance. Or, to put it more pragmatically, if you have a 90mm lens focused at infinity, a 360-degree image should be exposed over a distance of 22.26 iinches. If the lens is 87.65mm, that distance should be 21.68 inches. Setting aside the question of whether one can keep the film flat enough or avoid jitter (I'll just have to find those things out), we know going in that if the film isn't advanced the correct distance, acuity will suffer. The width of the exposure slit is particularly relevant -- the narrower it is, the less effect we see from having the film advance too fast or too slow. But if the speed is correct, then the slit can be varied for exposure purposes without giving up image quality.

If the slit is 1/16th of an inch (and I wouldn't want to have to rely on something too much narrower), then the movement resulting from 90mm vs 87.65mm would be 0.0016; a good lens can resolve 80 lpmm, which are 0.0005 apart. Knowing the true focal length is one pretty painless way to improve image quality.

I thought about scanning a ruler to benchmark the scanner's calibration, and will do that. But I'll scan several -- I've been surprised to see how much variation there is in my household's rulers. Perhaps a lab somewhere has one that is certified or something.

Another possibility -- create a simple line drawing in, say, Pagemaker (which I have), then print it. Printers have a vested interest in representing digitally described objects to a uniform size. Then scan the printed image at 4800 dpi and see how big that shows it to be.

Any other suggestions are welcome.

 

[couldabin]

Yes, I plan to use shims and a high-powered loupe to establish the mounted distance.

I suspect OP is wanting to use a helicoid lens mount such as following and it is important that when lens is focussed at infinity, it's the correct distance from film. This has to be designed into the camera body front flange to film distance taking into consideration helicoid depth and lens rear flange to film distance. If ain't built to perfect tolerances you'll never get sharp at infinity with lens wound all the way in. Maybe design in some adjustment so it can be set accurately. Replaceable shims may be a good idea so it can be adjusted for each individual lens of same or close focal length.

http://www.schneiderkreuznach.com/e...t-lenses/helical-mount-and-accessories/#c1691

 

[couldabin]

I promise to post at least one image (more if this actually works), but it will be several weeks before I get to that stage.

 

[RobC]

you will require to find the rear nodal point if you are rotating anything. There are various descriptions of how to find it if you goole it.

 

[RobC]

how are you going to print it ? :confused:

 

[couldabin]

Not in my darkroom.


And I wouldn't have space on my wall to hang it.

how are you going to print it ? :confused:

 

[couldabin]

I'm assuming the rear nodal point is the image distance. My thought was that I would determine the true focal length as described, and use that number to calibrate movement. The actual focusing, though, would be done with a loupe and temporary groundglass. Changing the image distance won't be attempted very often, since it necessitates machining a drive disk with a correspondingly different diameter.

Am I correct in thinking that rear nodal point = image distance?

you will require to find the rear nodal point if you are rotating anything. There are various descriptions of how to find it if you goole it.

 

[couldabin]

I see now that rear nodal point most certainly isn't image distance. Thanks -- I'll be taking that into account!

 

[RobC]

I see now that rear nodal point most certainly isn't image distance. Thanks -- I'll be taking that into account!

It may actually be the entrance pupil you need to be concerned with but I have no idea how things change at different focussing distances, i.e. varying lens to film distance and the corresponding rotation point or even whether its particularly relevant to your design.

 

[MartinCrabtree]

Please do post up images of the final camera. I love stuff like this.

 

[Nodda Duma]

Something to keep in mind for the method you are thinking of using: Your focus will be off by a factor related to the distortion in the lens. I *think* the distortion data typically provided by the manufacturer is enough to reduce the error to (perhaps) a manageable amount, but it's one of the reasons I've never relied on the method.

In the past I've used an autocollimator viewing into the front of the lens, with a mirror on a linear slide behind. When the return reticle is focused by sliding the mirror, I then directly find the flange focal length by measuring the distance from mounting flange to mirror surface. For a real lens, the flange focal length is what's really important because that's how you know where to mount it. Typical autocollimators have a green filter on their light, so this method gives best focus for the range of wavelengths that get through the filter. You also have to know how to properly focus when using your Mk. I eyeball when looking through a reticled eyepiece. A fancier version of this method makes use of an interferometer (which is more accurate than focusing by eye), but you need access to an interferometer and the measurement is good only for the wavelength the interferometer operates at (typically 632.8 nm).

Nowadays I just put the lens on my MTF bench which spits out the focus ~+/- 0.1% as part of its measurement suite for whatever spectral filter I put in front of the source.

But I'm a spoiled lens designer


This paper summarizes some of the more common methods for measuring focal length as used in the industry:

http://fp.optics.arizona.edu/optomech/student reports/tutorials/Tutorial_Morel,Sophie.pdf

perhaps you can make use of one of these approaches.

-Jason

 

[couldabin]

This paper summarizes the more common methods for measuring focal length as used in the industry:

http://fp.optics.arizona.edu/optomech/student reports/tutorials/Tutorial_Morel,Sophie.pdf

perhaps you can make use of one of these approaches.

Thanks. I'm not sure I have access to that kind of equipment. I'm going to apply the simple magnification formula and see how that compares.

 

[gzinsel]

I realize i am coming up short for "the correct" answer, But did you consider the literature from the lens Manufacturer. I have read the fujinon literature fromfuji has/ . . . literature that usually gives fairly precise statistics. For example the 420mm L, I own is actually ( by there accounts) 419.5mm I suppose we should not trust the literature from the Manufacturers ? just asking?

 

[John Koehrer]

Won't the lens have to pivot on the optical center?
There were cameras made like this early in the last century so there may be some patent drawings available.

 

[couldabin]

The available data suggests that the effective focal length (which I believe is the one used for calculating magnification -- anyone out there, correct me if I'm wrong on this) is 90.4mm. I did the test shot to see if that matched what I found, and it doesn't. My method may not be correct, of course, so I'm trying other approaches.

 

[couldabin]

It would need to pivot around the rear nodal point in order to keep the image stationary relative to the film (again, if anyone out there recognizes this to be wrong, please let me know). On the other hand, getting the film to advance at precisely the right speed is entirely a function of the image distance, which is a function of the object distance and focal length. Thus, any uncertainty about either leaves us with a dilemma as to how fast to advance the film. This is why I'm taking some pains to find a method that conclusively establishes the focal length. It's also why I thought that simply photographing a scene, using the lens in my 4x5, then comparing the arc between landmarks in the scene to the physical distance on the negative would give an absolute measure of how far the film would have to be advanced as the camera rotates.

So while placing the rotational axis ad the rear nodal point eliminates image creep due to rotation, it It seems to me that allowing for adjustment of the rotational axis could allow for compensation of incorrect (too fast or too slow) film movement.

 

[RobC]

I always thought it was the rear nodal point but I recently read it should be the entrance pupil.

Noblex cameras used a curved film plane so that lens was always a constant distance from film and rotated the lens but around which point I don't know for sure. As I said, I thought it was the rear nodal point but it may be the entrance pupil which could be anywhere depending on your lens design and type.

You need someone with more lens knowledge than I have for this.

 

[Ian C]

The swing-lens panoramic camera that “wipes” the image through a narrow slit onto a circular band of film requires that the lens pivot on its second nodal point. This can be found from the flange distance and true focal length (which can vary somewhat from the marked focal length). Since we usually use the hyperfocal distance for the chosen aperture we’d actually use the image distance from the second nodal point to the image to determine the location of the second node relative to the mounting flange.

The center of the entrance pupil is indeed the correct pivot point when we use a conventional camera and take multiple shots with the intention of blending the images (darkroom prints or “stitched” digital images). The correct pivot point for this geometry is the no-parallax point, which is located at the center of the entrance pupil. It can only be found by experimentation using a simple nodal slide.

These are two distinctly different geometries. Each requires a unique rotation point.

 

[Nodda Duma]

Rotate about the entrance pupil to rotate the camera around through the scene. Rotate about the exit pupil if you swing the film around and the lens stays stationary.

). The correct pivot point for this geometry is the no-parallax point, which is located at the center of the entrance pupil. It can only be found by experimentation using a simple nodal slide.

There is another way to find the entrance pupil position if you have an optic that can serve as an eyepiece (an eyepiece with a reticle like out of a grain focuser minimizes errors due to accommodation):

Line up the eyepiece in front of the lens, and focus on the vertex of the front element surface (easy reference). Move the eyepiece in towards the lens until the aperture stop comes into focus. The distance you move the eyepiece gives you the physical distance to the entrance pupil plane referenced from the vertex of the front surface.

You can repeat this measurement from the rear of the imaging lens using the back surface vertex as the reference to find the exit pupil position.

Actually just used this yesterday at work for positioning an imaging objective in a jig that needs to pivot on its entrance pupil.

 

[RobC]

I had understood that the entrance pupil could be inside the lens, to the front of the lens or behind the lens depending on lens design. I can't see how the eyepiece method will always work for inside or outside, especially if the entrance pupil is behind the eyepiece or behind the lens. Am I wrong ?