mooge
Well-known
Hi all,
I've been measuring the lens flange to focal plane distance (FFP, let's call it) of a few of my rangefinder cameras using the apparatus shown below.
And, I guess, to clear things up right off the bat, I'm really concerned with the physical distance between the film rails to the lens flange, which I'm calling the FFP here. The flange-focal distance (FFD) is apparently not the same (see below).
The apparatus consists of:
1. A granite surface plate
2. a dial indicator on a cheapo indicator stand
3. a block, surface ground flat and parallel
4. a reference block, surface ground flat and parallel to about 27,80mm
The block serves to allow the camera to sit on the (inner) film rails, and the reference is to zero the indicator. In theory, if the FFP is correct the indicator will read -0,05mm (see below and the Erwin Puts link).
I've measured three rangefinders so far; a Leica M4 and M2, and a Canon 7.
The Leica M4 FFD varies from -0.15mm to -0.08mm from the reference height.
The Leica M2 FFD varies from -0.15mm to -0.04mm from the reference, with the distribution shown below. This has the most extreme variation of the three cameras tested, but not by much. This camera has been dropped numerous times in the 8 years or so that I've owned it.
The Canon 7 FFD varies from -0.16mm to -0.10mm +1mm (since it's L39 and not M mount) from the reference.
According to Erwin Puts, and the KS-15(4) repair manual (Leica M2), the film plane to lens flange distance should actually be 27,75mm, as opposed to the 27,80mm flange focal distance quoted elsewhere. Puts puts the tolerance of the film plane to lens flange distance as +/- 0,02mm whereas the KS-15(4) manual states +0,005/-0,01mm.
The reference block, I'm going to guess, is 27,80mm -0,01mm. My good calipers say 27,79mm and my 1-2" mic says 1,0944" or ~27,797mm; the mic is name brand but used and I don't have anything to check it with. Both the reference and offset blocks have been checked and are flat and parallel to well under 0,01mm; better than my indicator can measure.
The indicator is new and good, but the indicator stand is junk. The whole setup has a repeatability of +/- 0,01mm (at least), which seems really bad.
It seems like my measuring setup is good to, let's say, +/-0,03mm (at worst). The film plane to lens flange distance should be 27,75mm; so this will show up as -0,05mm on my indicator. With Puts's tolerances I should be aiming for -0,03mm...-0,07mm. And then there's the error in my measuring setup.
Questions:
1. I have some difference (>-0,05mm) on all of the cameras from the ideal 27,75mm. Is there something wrong with my setup?
2. Is +/- 0,02mm a reasonable tolerance for the FFP?
3. How much of a variation in FFP can be hidden with rangefinder calibration? (all three cameras have had their RF calibration done by me)
4. Does incorrect FFP affect wide-angle lenses less than longer focal length lenses? (I pretty much exclusively use a 12mm and 35mm).
Your thoughts?
Thanks,
Eugene.
I've been measuring the lens flange to focal plane distance (FFP, let's call it) of a few of my rangefinder cameras using the apparatus shown below.
And, I guess, to clear things up right off the bat, I'm really concerned with the physical distance between the film rails to the lens flange, which I'm calling the FFP here. The flange-focal distance (FFD) is apparently not the same (see below).
The apparatus consists of:
1. A granite surface plate
2. a dial indicator on a cheapo indicator stand
3. a block, surface ground flat and parallel
4. a reference block, surface ground flat and parallel to about 27,80mm
The block serves to allow the camera to sit on the (inner) film rails, and the reference is to zero the indicator. In theory, if the FFP is correct the indicator will read -0,05mm (see below and the Erwin Puts link).
I've measured three rangefinders so far; a Leica M4 and M2, and a Canon 7.
The Leica M4 FFD varies from -0.15mm to -0.08mm from the reference height.
The Leica M2 FFD varies from -0.15mm to -0.04mm from the reference, with the distribution shown below. This has the most extreme variation of the three cameras tested, but not by much. This camera has been dropped numerous times in the 8 years or so that I've owned it.
The Canon 7 FFD varies from -0.16mm to -0.10mm +1mm (since it's L39 and not M mount) from the reference.
According to Erwin Puts, and the KS-15(4) repair manual (Leica M2), the film plane to lens flange distance should actually be 27,75mm, as opposed to the 27,80mm flange focal distance quoted elsewhere. Puts puts the tolerance of the film plane to lens flange distance as +/- 0,02mm whereas the KS-15(4) manual states +0,005/-0,01mm.
The reference block, I'm going to guess, is 27,80mm -0,01mm. My good calipers say 27,79mm and my 1-2" mic says 1,0944" or ~27,797mm; the mic is name brand but used and I don't have anything to check it with. Both the reference and offset blocks have been checked and are flat and parallel to well under 0,01mm; better than my indicator can measure.
The indicator is new and good, but the indicator stand is junk. The whole setup has a repeatability of +/- 0,01mm (at least), which seems really bad.
It seems like my measuring setup is good to, let's say, +/-0,03mm (at worst). The film plane to lens flange distance should be 27,75mm; so this will show up as -0,05mm on my indicator. With Puts's tolerances I should be aiming for -0,03mm...-0,07mm. And then there's the error in my measuring setup.
Questions:
1. I have some difference (>-0,05mm) on all of the cameras from the ideal 27,75mm. Is there something wrong with my setup?
2. Is +/- 0,02mm a reasonable tolerance for the FFP?
3. How much of a variation in FFP can be hidden with rangefinder calibration? (all three cameras have had their RF calibration done by me)
4. Does incorrect FFP affect wide-angle lenses less than longer focal length lenses? (I pretty much exclusively use a 12mm and 35mm).
Your thoughts?
Thanks,
Eugene.
craygc
Well-known
When you say "calipers" do you mean veriner calipers? If so you should have a depth measurement out the backend that you could use to measure between the flange face and the film rails more accurately than with the dial indicator; especially as you move between cameras.
Livesteamer
Well-known
I once explored this with a few Nikon F's, long ago when I worked in a gear shop with a big inspection department and a coordinate measuring machine. I owned the worlds ugliest Nikon F but when inspected it was square and precise to the limits of the CMM which was one ten thousandths of an inch. All for corners of the lens mount gave the same reading. My other, newer F's showed a little variance but less than a thousandths of an inch. I think film is forgiving of such small errors but digital requires greater precision. At least that's my impression. Joe
Hi Eugene,
Interesting post. I can offer a couple of observations perhaps, having done a similar process a few times to set up the parallelism of Rollei twin lenses.
Firstly I use a granite surface plate myself. Surface plates tend to be very expensive to procure in Australia. There's no trouble getting one of course, but larger engineering workshop sized ones tend to be more common from local sellers and are priced accordingly. Smaller ones can be found very affordably in Eg the USA but you get totally murdered on the postage of such heavy items.
Luckily I eventually found a seller who stocked a smaller plate for well under AUD $100 but it was, like your own, made of granite. Now, granite has certain benefits, Eg it tends not to raise burrs after an impact but might chip slightly, hence, it may retain its surface accuracy better with rough treatment in a workshop, whereas a metal one may be affected. But of course, magnetic stands will not affix to a granite plate, will they!
My solution when using a gauge stand (and it is one you may want to consider adopting) was to fasten the base of the stand to the edge of the plate with a small G clamp so that the stand was absolutely stable, relative to the plate. I was using the surface plate at my repair desk, so I accomplished this simply by sitting the plate on the desk with a slight overhang over the edge of the desk so that I might fit a clamp.
What this will mean is that as opposed to keeping the camera in the same position and moving the stand, you will most likely find yourself having to rotate the camera body to align with the tip of the gauge. A little inconvenient but of no import as far as accuracy is concerned. The reason you're using a surface plate is because it is flat and stable. If you are taking your various measurements around the lens mount at different places on the plate, this should not be of any consequence providing your reference block is also true. And it's likely to be much superior in consistency than relying on a stand resting on the plate by virtue of gravity, resting on a loose block underneath it. Even if you are also holding it in place firmly with your hand. I hope this makes sense?
Your predicament is also a little different to my own. When setting up a Rollei for instance, you're less concerned with the absolute measurement from the rails to (in the case of a Rollei) the lens board, than you are with getting all four corners in agreement. Not having interchangeable lenses, as long as the board is parallel to the film rails the focus is then dialled in optically and fixed via the stop on the focus knob. Yes, there's a bit more to it than that, but providing you haven't had to use an excess of shimming material or haven't cocked up the cam follower positions badly, it all comes out OK.
Of course the situation is not as straighforward with an interchangeable lens camera. Naturally the mount should be parallel to the film rails, but the absolute distance between the two points is also critical if any given lenses for the system are to mount to the body and then focus correctly. So I can see why you are trying to set a precise initial height for the dial gauge: this is not a point I've yet had to tackle.
I think that, if I was starting out where you are, I would probably see if any of my outside micrometers were able to align across the film rails and your lens mount. Obviously you'd not be able to cover all positions around the clock, but if you could get four or five measurements you'd have a fairly accurate picture of what the precise distance between the two points is, a micrometer being more accurate than either a dial gauge or verniers. A dial gauge set up I think, would be the easiest way to correct the parallelism but it's not necessarily the easiest way to measure the distance you need to know.
Getting both spot on might need some back and forth between the two (if you have a micrometer, that is, otherwise make do with verniers I suppose). I don't want to get political, here, but the fact is I'm not into Leica stuff at all, so don't even know how the factory fine tune the alignment in the first place. Shims under the lens mount and/or film rails (if they come off, that is, some cameras do, some don't) is probably the most common way this is achieved. (Unless it's an ALPA hybrid, that is, where everything is just machined to be perfect, with a ridiculously thin piece of shim tape between the front and rear body castings, at the worst). But I'll tell you about that project some other time...
As far as the parallelism is concerned--I had a quick glance at your post earlier this evening before processing a few films, and I was pondering what sort of tolerance might apply. Franke & Heidecke specified 0.05mm across the lens board corners for their TLRs. But a 35mm camera would want even tighter accuracy in order to deliver its best at the smaller format, I would have thought. So, a figure of half that or less seemed a reasonable sort of starting point to me, and, on reading your post in full, I was gratified to note that your references, indeed, point towards something like that, depending on whether or not you work with the good Mr Puts suggestion, or the manual specification. In your shoes, I'd be interested to know why he suggests a more generous tolerance. Presumably he has good reasons for this.
I can't speak to the varying degrees of accuracy needed for different focal length lenses. Except to say that as a wide angle lens tends to have greater depth of field than the medium telephotos used at the other end of a RF lens range, I should have thought that if there was a discrepancy in the adjustment of a particular body from factory specs that, even at minimum focusing distances where DOF will be at its minimum, the shorter lenses would tend to be more forgiving, yes? Edit: this might however, be complicated by any focus shift characteristics of whichever lenses you have a preference for, perhaps?
In closing, bear in mind you are trying to achieve two crucial, and related, but separate measurements. You probably can't reasonably expect to get both the parallelism and the required lens register bang on in the first pass. If the flange to focal plane distance is at least in the ballpark, getting the flange parallel to the rails might be a good first step. Having done that, re-measuring the distance, and the dimension of any shims in place beneath the flange (assuming, once again, Leica use them, Contax do, umm, have you experienced the joy that is the precision of a pre-war Contax RF yet?) should, then, in conjunction with some basic arithmetic, enable you to substitute shims of adjusted thickness that will achieve the precise distance you need, without impacting the correct parallelism in the process.
That's the theory anyway. The reality is probably as above but two or three repetitions until it's all bang on, but I would think that would get you there. The last two times I did a Rollei (measured at both the infinity and minimum focusing distances) I got the lens boards to, from memory, 0.01mm and 0.03mm across the corners, well within the factory spec, and results from the cameras in question reflected that accuracy. So it is do-able. But I think with your set up (or mine, for that matter) getting down to the bottom end of the manuals 0.005mm spec would be nigh impossible, at least, with much repeatablity.
Hope this helps.
Cheers,
Brett
Interesting post. I can offer a couple of observations perhaps, having done a similar process a few times to set up the parallelism of Rollei twin lenses.
Firstly I use a granite surface plate myself. Surface plates tend to be very expensive to procure in Australia. There's no trouble getting one of course, but larger engineering workshop sized ones tend to be more common from local sellers and are priced accordingly. Smaller ones can be found very affordably in Eg the USA but you get totally murdered on the postage of such heavy items.
Luckily I eventually found a seller who stocked a smaller plate for well under AUD $100 but it was, like your own, made of granite. Now, granite has certain benefits, Eg it tends not to raise burrs after an impact but might chip slightly, hence, it may retain its surface accuracy better with rough treatment in a workshop, whereas a metal one may be affected. But of course, magnetic stands will not affix to a granite plate, will they!
My solution when using a gauge stand (and it is one you may want to consider adopting) was to fasten the base of the stand to the edge of the plate with a small G clamp so that the stand was absolutely stable, relative to the plate. I was using the surface plate at my repair desk, so I accomplished this simply by sitting the plate on the desk with a slight overhang over the edge of the desk so that I might fit a clamp.
What this will mean is that as opposed to keeping the camera in the same position and moving the stand, you will most likely find yourself having to rotate the camera body to align with the tip of the gauge. A little inconvenient but of no import as far as accuracy is concerned. The reason you're using a surface plate is because it is flat and stable. If you are taking your various measurements around the lens mount at different places on the plate, this should not be of any consequence providing your reference block is also true. And it's likely to be much superior in consistency than relying on a stand resting on the plate by virtue of gravity, resting on a loose block underneath it. Even if you are also holding it in place firmly with your hand. I hope this makes sense?
Your predicament is also a little different to my own. When setting up a Rollei for instance, you're less concerned with the absolute measurement from the rails to (in the case of a Rollei) the lens board, than you are with getting all four corners in agreement. Not having interchangeable lenses, as long as the board is parallel to the film rails the focus is then dialled in optically and fixed via the stop on the focus knob. Yes, there's a bit more to it than that, but providing you haven't had to use an excess of shimming material or haven't cocked up the cam follower positions badly, it all comes out OK.
Of course the situation is not as straighforward with an interchangeable lens camera. Naturally the mount should be parallel to the film rails, but the absolute distance between the two points is also critical if any given lenses for the system are to mount to the body and then focus correctly. So I can see why you are trying to set a precise initial height for the dial gauge: this is not a point I've yet had to tackle.
I think that, if I was starting out where you are, I would probably see if any of my outside micrometers were able to align across the film rails and your lens mount. Obviously you'd not be able to cover all positions around the clock, but if you could get four or five measurements you'd have a fairly accurate picture of what the precise distance between the two points is, a micrometer being more accurate than either a dial gauge or verniers. A dial gauge set up I think, would be the easiest way to correct the parallelism but it's not necessarily the easiest way to measure the distance you need to know.
Getting both spot on might need some back and forth between the two (if you have a micrometer, that is, otherwise make do with verniers I suppose). I don't want to get political, here, but the fact is I'm not into Leica stuff at all, so don't even know how the factory fine tune the alignment in the first place. Shims under the lens mount and/or film rails (if they come off, that is, some cameras do, some don't) is probably the most common way this is achieved. (Unless it's an ALPA hybrid, that is, where everything is just machined to be perfect, with a ridiculously thin piece of shim tape between the front and rear body castings, at the worst). But I'll tell you about that project some other time...
As far as the parallelism is concerned--I had a quick glance at your post earlier this evening before processing a few films, and I was pondering what sort of tolerance might apply. Franke & Heidecke specified 0.05mm across the lens board corners for their TLRs. But a 35mm camera would want even tighter accuracy in order to deliver its best at the smaller format, I would have thought. So, a figure of half that or less seemed a reasonable sort of starting point to me, and, on reading your post in full, I was gratified to note that your references, indeed, point towards something like that, depending on whether or not you work with the good Mr Puts suggestion, or the manual specification. In your shoes, I'd be interested to know why he suggests a more generous tolerance. Presumably he has good reasons for this.
I can't speak to the varying degrees of accuracy needed for different focal length lenses. Except to say that as a wide angle lens tends to have greater depth of field than the medium telephotos used at the other end of a RF lens range, I should have thought that if there was a discrepancy in the adjustment of a particular body from factory specs that, even at minimum focusing distances where DOF will be at its minimum, the shorter lenses would tend to be more forgiving, yes? Edit: this might however, be complicated by any focus shift characteristics of whichever lenses you have a preference for, perhaps?
In closing, bear in mind you are trying to achieve two crucial, and related, but separate measurements. You probably can't reasonably expect to get both the parallelism and the required lens register bang on in the first pass. If the flange to focal plane distance is at least in the ballpark, getting the flange parallel to the rails might be a good first step. Having done that, re-measuring the distance, and the dimension of any shims in place beneath the flange (assuming, once again, Leica use them, Contax do, umm, have you experienced the joy that is the precision of a pre-war Contax RF yet?) should, then, in conjunction with some basic arithmetic, enable you to substitute shims of adjusted thickness that will achieve the precise distance you need, without impacting the correct parallelism in the process.
That's the theory anyway. The reality is probably as above but two or three repetitions until it's all bang on, but I would think that would get you there. The last two times I did a Rollei (measured at both the infinity and minimum focusing distances) I got the lens boards to, from memory, 0.01mm and 0.03mm across the corners, well within the factory spec, and results from the cameras in question reflected that accuracy. So it is do-able. But I think with your set up (or mine, for that matter) getting down to the bottom end of the manuals 0.005mm spec would be nigh impossible, at least, with much repeatablity.
Hope this helps.
Cheers,
Brett
mooge
Well-known
Hi Brett (I was hoping to get your thoughts on this, actually) -
Shame that machine tools are hard to come by in Australia. I read that certain types of metal stock is also hard to find, or expensive, for you guys. I've got at least two metal stores and two machine tool shops in my area that cater just to hobbyists, so we have it pretty good...
I just checked a couple points on my M4 and the mic is pretty much spot on with what the indicator says (within 0,01mm... ish)
That said, I'm gonna stick to the indicator being the best method. With a mic I can pick up only on 4 points (and I'd need a deep throat 25-50mm mic, which does exist, to my surprise) and it's really hard to get an accurate mic reading off of the thin film rails. The indicator tells me exactly what I want to know - the distance of points on the lens flange from the plane of the film rails. .
My indicator stand is magneted to a piece of 1x2" steel bar, flycut on both sides (so slightly concave... go figure). I'm more worried about the arms of the indicator stand moving, but I might give your clamp trick a go sometime. A non-crap indicator stand (and/or height gauge) is on the list of stuff to get.
I can understand you not being into Leica stuff. I love Leicas, but they're also stupid expensive (ask me, I have three...) and kind of a pain in the ass sometime. More like all the time. I haven't worked on a real contax but worked on a Kiev 4 years ago (my first rangefinder!).
Ok, so it seems like my setup is okay (mic verified). I still have a mystery 0.05mm that all three cameras are off by. so to the last two questions:
3. considering that the FFP is shorter than the manufacturer's spec (= lens focuses beyond infinity), the rangefinder can be adjusted to correct for this, yeah?
4. My 12mm heliar is not RF coupled, and from inf to 0,3m it moves less than 1mm. Does that mean that a difference in FFP has a bigger affect on the sharpness or plane of focus of this lens?
Thanks all.
Shame that machine tools are hard to come by in Australia. I read that certain types of metal stock is also hard to find, or expensive, for you guys. I've got at least two metal stores and two machine tool shops in my area that cater just to hobbyists, so we have it pretty good...
I just checked a couple points on my M4 and the mic is pretty much spot on with what the indicator says (within 0,01mm... ish)
That said, I'm gonna stick to the indicator being the best method. With a mic I can pick up only on 4 points (and I'd need a deep throat 25-50mm mic, which does exist, to my surprise) and it's really hard to get an accurate mic reading off of the thin film rails. The indicator tells me exactly what I want to know - the distance of points on the lens flange from the plane of the film rails. .
My indicator stand is magneted to a piece of 1x2" steel bar, flycut on both sides (so slightly concave... go figure). I'm more worried about the arms of the indicator stand moving, but I might give your clamp trick a go sometime. A non-crap indicator stand (and/or height gauge) is on the list of stuff to get.
I can understand you not being into Leica stuff. I love Leicas, but they're also stupid expensive (ask me, I have three...) and kind of a pain in the ass sometime. More like all the time. I haven't worked on a real contax but worked on a Kiev 4 years ago (my first rangefinder!).
Ok, so it seems like my setup is okay (mic verified). I still have a mystery 0.05mm that all three cameras are off by. so to the last two questions:
3. considering that the FFP is shorter than the manufacturer's spec (= lens focuses beyond infinity), the rangefinder can be adjusted to correct for this, yeah?
4. My 12mm heliar is not RF coupled, and from inf to 0,3m it moves less than 1mm. Does that mean that a difference in FFP has a bigger affect on the sharpness or plane of focus of this lens?
Thanks all.
olo530
Member
I’m not sure if I should start my own thread or ask here. In any case, why on small and medium format cameras the flange focal distance is set with such tight tolerances and on large format cameras it’s moved back and forth by hand without any regard for 0.01 mm differences in either distance or how parallel to the film it is?
Freakscene
Obscure member
The flange to lens plane distance is more important for depth of focus rather than depth of field. Wide-angle lenses, which at distances other than infinity have low magnification, have less depth of focus. Wide angles are therefore much more sensitive to variations in flange to lens plane distance than normal or long lenses.
Marty
Marty
retinax
Well-known
Is this not over-compensated by the larger depth of field that wide-angles have (at the same distances)? Otherwise, why all this "one needs a more precise rangefinder for longer lenses"? Is that all bogus?
Freakscene
Obscure member
No. They are different issues. The depth of field (in front of the lens) is deeper for wide angles, but the depth of focus (behind the lens) is shallower. So wide angle lenses are less sensitive to rangefinder accuracy but more sensitive to flange-to-film/sensor plane distance being correct. The reverse applies to long lenses.
Marty