PDA

View Full Version : True focal lengths of APO Grandagon lenses



Jerry Fusselman
18-May-2004, 22:39
What are the true focal lengths of the 35mm, 45mm, and 55mm Rodenstock APO Grandagon lenses?

Bob Salomon
19-May-2004, 05:17
35mm, 45mm, 55mm.

Why not ask the "true" focal lengths of a 150, 135 and 210mm?

What do you really want to know about them?

Nick_3536
19-May-2004, 06:45
I'm guessing he wants to know the distance to focus the lenses. Maybe to see if the lens will focus at infinity on his camera?

Bob Salomon
19-May-2004, 07:28
Flange focal lengths at infinity in Copal 0 are:

35mm = 43.2mm 45mm = 55.5mm 55mm = 67.6mm

Leonard Evens
19-May-2004, 07:53
Jerry,

Perhaps I can clarify what the others have said. The true focal length is the distance from the film plane to the lens when focused on infinity. It could differ slightly from the stated focal length for individual lenses but not by enough to matter. The problem is that this would apply only if a lens were an idealized point. Actual lenses have some significant physical extent, so one must choose some reference point in the lens. That point is called the rear principal point, and its actual location depends on the design of the lens. For most lenses it is pretty close to the center of the lens which in turn is pretty close to the lens board. For lenses of telephoto design, it may be far in front of the lens. For most wide angle lenses, it is some distance in back of the lens board. The reason for this is so that the lens need not be brought quite so close to the film plane when focused at infinity. It also helps with light drop off over the field. Lens manufacturers give as part of the specifications the distance that the lens board will be from the film plane when the lens is focused at infinity. That is called the rear flange focal length. Thus Bob Salomon has told you that for the Grandagon 55 mm lens, the rear flange focal length is 67.6 mm. That means the rear principal point is 67.6-55 =12.6 mm back of the rear of the lens board. That means that the lens functions for geometric optics purposes as if it were a plane element centered on a point 12.6 mm in back of the lens board.

Jerry Fusselman
19-May-2004, 07:56
When a lensmaker names a len's focal length, it ain't necessarily so. For example, Schneider's Super Angulon 47mm XL actually has a focal length of 48.0mm. Their current 47ml nonXL is really 47.5mm. Their SSXL 80mm is really 81.0mm. This information is on their website. (They use f' and F' as symbols for the true or "effective" focal length.)

I really do want their focal lengths, but with a little more precision than just using their names.

Bob Salomon
19-May-2004, 09:48
Sorry Jerry, that information is not in their literature. If you are need this for an OEM purpose we can put you in touch with the proper people. If that is the case please call me at 800 735 4373 x15.

Leonard Evens
19-May-2004, 13:06
Jerry,

The actual true focal length is likely to be different from any stated focal length. If you really need to know it as accurately as possible, the best thing to do is to measure it yourself. There are a variety of ways to do that, and you can probably find one that you like somewhere on the web. What I would do is to measure as carefully as possible the distances from the front of the lens board to the image plane for some number of subject distances. There are basically three variables you need to determine: the focal length, the distance from the lens board to the front nodal point and the distance from the lens board to the rear nodal point. So in principle, if you use the lens equation and some algebra, making these measurements for three subject distances should suffice, but I would do it for several more and take an average to improve accuracy.

David A. Goldfarb
19-May-2004, 13:56
In the case of the data on the Schneider website, isn't that just a sample test of one representative lens, which may not be accurate for a lens that you actually purchase? My impression was that lenses vary in focal length from sample to sample due mainly to variations in the refractive index of the glass, and this is why Leica and Linhof rangefinder cams are calibrated individually to each particular lens.

Bob Salomon
19-May-2004, 14:45
Actually there is a very easy way to find the effective focal length. Simply point a lens out the window at a distant object and measure the distance to the projected image that is formed on a card held behind the lens.

However if Jerry wants to have a lens of exact focal length Rodenstock does offer that service at an extra charge. They will deliver lenses (at an extra cost) that are certified as to the exact focal length and flange focal length distance. Or they will certify the focal length, flange focal length and back focal length (Schnittweite) of a lens or lenses. Price depends on the number of lenses calibrated. Price breaks are 1-4 lenses, 5 - 24 lenses, 25 - 99 lenses and 100 or more lenses.

The first calibration was done automatically on all Rodagon G enlarging lenses longer then 50mm as well as Apo Ronar lenses delivered for process cameras. As Apo Ronar lenses and the G series (except for the 50mm) are out of production this is not a commonly requested feature.

And many modern enlargers for mural printing with electronic AF can calibrate the lens in the set-up process eliminating the need for factory calibration.

And, as someone else pointed out, there is no "true" focal length. The actual focal length will be the marked focal length a small percentage. It is this small percentage that the factory will measure per the above.

in fact it has never been asked of us before for taking lenses for large format work.

Jerry Fusselman
19-May-2004, 21:33
I guess this is not common knowledge: Lens designers often put labels on lenses that not as accurate as they could be---even if rounded to the nearest convenient number. Focal lengths and apertures can both be quite a bit different from the manufacturer's label. The labeled-actual difference in focal length has been reported to be over 5% in many cases with 35mm lenses.

I want to know the focal lengths for these three lenses to 2 or 3 digits of accuracy. I don't need flange focal distances. I don't believe that Schneider would publish 48.0mm if it was 46.0mm just as often. I doubt that any 47XL on earth has a focal length that is 47.0mm or less. The reason Schneider publishes f' = 48.0mm, I am convinced, is because that's the way it was designed, not the way that a single lens turned out to be.

Leonard,

Are you suggesting a simultaneous estimation of the location of the two principal points along with focal length, or do you have some way to estimate the location of the principal points? I asked two news groups how to find the front principal point a few years ago, but none of the ways that were given were accurate enough for this.

Bob,

Your method of estimating focal length also requires knowledge of where the rear principal point is. If you can provide that information, that would be great! I am not interested in having a new lens made just for me, my existing APO Grandagons are working fine, but I want to know the designed focal length for t hese lenses. Though nobody asked you before, I am still interested.

I disagree with "The actual focal length will be the marked focal length a small percentage." This way of looking at it is not consistent with Schneider's own website. The true statement is "The actual focal length will be the designed focal length a small percentage." I am not worried about the small percent that comes from manufacturing variance. For these wide-angle lenses, the manufacturing variance in focal length is probably less than 0.1mm---don't you agree? I want to know the designed focal length for the three Rodenstock APO Grandagon lenses.

Bob Salomon
20-May-2004, 05:31
Then send me the serial numbers of your lenses and I will see if the factory kept a record. Otherwise they will have to be measured. Exactly why do you need this?

Leonard Evens
20-May-2004, 08:44
Jerry,

Photographic Optics by Arthur Cox suggests several methods to determine these numbers by optical measurements. Here is a variation of one of them.

Hang a tape measure so it is plumb. Set up the camera so it is centered on the tape and the gg is also plumb. Focus carefully on the tape using a high powered loupe, and measure carefully the size on the gg of some known length on the tape. Calculate the scale of reproduction by dividing the image size by the subject size and call that M1. Measure the distance from the gg to some specified reference point such as the back of the lens board or whatever is convenient. Call that V1. Now move the camera and do the same thing over again for a different subject distance getting numbers M2 and V2. Then

f = (V1 - V2)/(M1 - M2)

A similar method works using subject distances rather than image distances and the reciprocals of the scale of reproduction in the formula.

Note that since you are only interested in the difference in the image distances, you can manage to measure that to a few tenths of a mm if you go about it correctly. Namely instead of measuring the distance to a reference point as suggested above, measure instead the distance you have to move along the rail from the focus position in one case to that in the other position. I've put a scale on my focusing knob which allows me to make measurements of this kind down to a few tenths of a mm. See my essay www.math.northwestern/~len/photos/pages/dof_essay.html for a description. However, you would be lucky to be able to measure the image sizes to better than half a mm. So just using two pairs of values would not give you a very accurate answer. However, it you repeated the procedure for many pairs and took an average, in principle at least, you could improve your accuracy without limit. also, making several image size measurements for different subject lengths and averaging the results for the scale of reproduction in each case would probably help improve accuracy.

But I haven't actually tried this so I don't know if practice lives up to theory. The problem is that there could be some systematic errors such as the tape not being quite parallel to the gg These factors should again average out if you repeat the method independently from scratch.

Of course, once you know the focal length, you can locate the principal points by measuring subject and image distances from a fixed reference point and solving some equations, but the algebra and associated numerical analysis could be involved.

When I get a chance, I will actually try this with one of my lenses.

Emmanuel BIGLER
21-May-2004, 03:35
To Jerry F. : Jerry. If you make a picture of a distant object for which you know the size and distance, when the distance is large you can assign an angular size to, say, a grid or
an angular distance between two distant objects (trees, buildings etc...)

The conversion factor between the angular size for the distant object (expressed in radians) and the image size in the focal plane (measured in whatever unit is best for you, e.g. inches ;-) is named : focal length and the experiment will give you the right value whatever the lens design can be... when the object is far enough.
A precise method which needs an optical bench is to adjust the position of the lens around a certain pivot point and look where the image of a distant object become stationary w/respect to the rotation of the lens. The method is named 'nodal slide method' since you use the property of the rear principal plane to be the rear nodal plane as well, and images of a distant object formed on a fixed plane w/respect to the object are stationary when rotating the lens around its rear nodal point. Note that the object should be placed at infinity, in the optics lab a collimator is used to create an "artificial object" at infinity. In practice monorail view camera users can attempt to do this experiment if they can slide the tripod head independantly from the camera standards. The image should be fixed w/respect to the landscape, rotate the tripod head to perform the required nodal point rotation. The focal length is the distance between the rotation axis and the image plane in infinity->focus position.
Note than you cannot use the front swing movements to do the experiement since you cannot in general adjust the position of the lens on the front standard ! A less precise method that does not need any knowledge of the principal plane separation HH' or position as well is two measure how far you have to extend the bellows
of a view camera to switch from infinity->focus to the 1:1 magnification ratio in 2f:2f position. Measuring the required bellows extension to reach 1:1 is equal to one focal length (checked with a gridded object and image). The method is not very precise but interesting by the trick : no need to know where the principal planes are, valid for a thick retrofocus or a telephoto.

In the particular case of 35, 45 and 55 mm view camera lenses, I'm afraid that giving a figure within 0.1 mm of focal length requires a professional optical bench and an precise optical pointing system. There are other tricks & methods well-known to the professionals since the XIX-st century (Bessell's or Cornu's methods) that allow you to determine the gaussian elements of any thick compound optical system. To me this is an academic exercise for my students but is irrelevant for photographic use ;-);-)