Shimming a Sensor

<p>So I've been reading that an advantage of putting a medium format back on a technical/view camera as opposed to MF DSLR system is that with the former, the back can be shimmed precisely into the correct position (at least on an Alpa camera) so that focus is sharp at infinity. This seems reasonable to me as it is intuitive that any image quality advantage of a MF back must depend on precise (and motionless) capture (because blur is a great equalizer).</p>

<p>My first question, though, is how in the world would a consumer, or even a professional photo shop (absent factory equipment) shim accurately enough to get this right inasmuch as I read that tolerances need to be within .01 mm?</p>

<p>My second question is why an option to focus past infinity isn't sufficient? This may well be a dumb question, but when I shoot film on my 4x5 press camera, the entire lens moves as I focus (and way more than .01mm, of course). Using a loop on a focusing screen I adjust focus to precisely where I want it and sometimes this is beyond the infinity mark, but who cares? Wouldn't the principle be the same for any camera--that as long as the lens is free to move wherever you want it, precise flange distance prior to focus adjustments are unnecessary? What am I missing (and I'm willing to believe it can be almost anything).</p>

<p>Thanks in advance.</p>

<p>Sorry, but this question doesn't really make sense. A technical or view camera has no fixed "infinity focus" position. The front and back standards are free to move between some hard near limit and the maximum bellows extension. Not only that, but they're free to be tilted and swivelled as well, and the camera's designed to take any number of different lenses with differing back-focii. That's the whole point of a technical camera!</p>

<p>Even with a digital sensor, focus is done by eye, or by eye aided by electronic means. Shimming would only be relevant to accurately aligning a ground-glass focusing screen with the film or sensor plane; and that ain't rocket science. It can be done with equipment as simple as a micrometer depth-gauge and a few cigarette papers. Why you'd need to do that with a digital sensor though is unclear. Surely you'd just view and focus electronically through the sensor image?</p>

<blockquote>

<p>"So I've been reading that......."</p>

</blockquote>

<p>Wherever you've been reading it, I wouldn't waste my time reading any more from that particular source.</p>

<p>That there shouldn't be a problem requiring shimming is exactly my point, but I'll concede that I didn't put the question well. On the field camera I use (an old press camera, actually), using screws, I make a gross adjustment of the position of the lens on the camera's rails, then adjust the focus using a knob that moves a mark against the bed, where various distances are inscribed. As I change lenses, I never worry much about getting the gross adjustment exactly right, because the fine adjustment of the focusing knob has plenty of play, including play past where the bed is marked "infinity." I wondered, therefore, why similar play in fine adjustment wouldn't solve any problem on a technical camera as well. (I had in mind a similar setup, where the fine adjustment would be marked on the camera but I realize now that a traditional view camera has no such feature.)</p>

<p>My explanation in mind, my source for the supposed need to shim is from someone who strikes me as reliable source, Mark Dubovoy, a contributor on Luminous Landscape. He makes the point as follows and in the quote he does seem to make the point that positioning of the sensor has independent consequence, both distance from, and angle to, the lens, and for the reasons given I remain mystified as to the former:</p>

 

<blockquote>

<p>I do want to re-emphasize that the ALPA is the only camera that allows you to shim the digital back adapter so that the sensor is positioned exactly where it needs to be. I do not quite understand why so many people think this is such a mystery, or why some think this is not important.<br>

<br>

The situation is very simple: Nothing is perfect. Therefore, when you buy a digital back, the sensor will not be positioned exactly where it is supposed to be, and it very well may not be perfectly perpendicular to the focal axis either. It will be "within manufacturing tolerances", but not perfect.</p>

<p>When you mount the back on a camera, even the thickness of the paint or the anodizing on the camera wil make a difference. The point I am trying to make is that your sensor is almost 100% guaranteed to be out of focus. How much out of focus depends purely on the luck of the draw, it can be very slight or it can be more severe. The crucial issue is that even if it is very slight, it is quite visible in the files, and as we all know, anything that is not captured is lost forever.</p>

<p>If you use an ALPA camera, through a clever design of the back adapter and a set of shims, you can correct for this and shim the mount in such a way that when you mount the digital back on the camera your sensor is in the perfect position and therefore in perfect focus (by "perfect" I obviously mean really close to perfect, close enough to make a significant visual difference). The difference is much greater than you can imagine. </p>

</blockquote>

Backs are shimmed so that the sensor is in the place it should be: the plane in which the focussing screen is, and in which film (or another sensor) is placed when used instead of this particular sensor.<br>If it isn't, if you depend on changes to the focussing setting to get a sharp image on film after you focused using the focussing screen, how would you know by how much to defocus to get the image sharp on a sensor that isn't put in the correct position?<br>And if you know by how much to defocus the camera, why would you not build that correction permanently into the adapter (i.e. shim the thing) so that you never ever need to think about it again, instead of having to defocus by a tiny amount after focussing for every single shot?

<blockquote>

<p>And if you know by how much to defocus the camera, why would you not build that correction permanently into the adapter (i.e. shim the thing) so that you never ever need to think about it again, instead of having to defocus by a tiny amount after focussing for every single shot?<br>

</p>

</blockquote>

<p>This makes sense. No question. It's just that the article to which I referred seemed to suggest that the problem absent the ability to shim, say with a MF DSLR, was not merely a nuisance but an uncorrectable defect. And if one is manually focusing (or using a camera with a good autofocus) and sharp focus is <em>available </em>despite the lack of precise calibration, then this strikes me as an awfully small problem. Thus, I'm still not sure I understand <em>unless</em> the idea is to get the focus precisely right without visually confirming (or having auto-focus do the equivalent).</p>

The idea is that using digital sensors, the exact position of the sensor, i.e. the correct lens to sensor distance, is more important than it is when using film.<br>Be that as it may, the underlying truth (and not a small problem per se) is that, of course, the sensor needs to be where the sharp image is formed.<br>The focussing screen is at the exact same distance. And if an AF system is used, it is callibrated to that same distance as well.<br>If a sensor in a DB should not be in the correct position, it is important that it can be adjusted so that it is. Else, despite that you will see a sharp image on the focussing screen when focussing manually, and/or despite that the AF system will focus accurately, the sensor will not capture a sharp image.<br><br>The sharp image will be available. Somewhere in front of or behind the sensor. But unless the sensor position conforms with the focussing screen position and AF system calibration, you have no way of telling when it can be captured.<br>And that's the idea: being able to indeed capture a sharp image when the focussing screen or AF systems suggests that you should.<br><br>Now you can argue about what the defect is exactly: the fact that the sensor position would need adjusting to begin with, or the fact that such an adjustment would not be possible. But if the sensor is not in the correct position, it would be double trouble if you then couldn't adjust its position to correct that defect.

<blockquote>

<p>The focusing screen is at the exact same distance. And if an AF system is used, it is calibrated to that same distance as well. If a sensor in a DB should not be in the correct position, it is important that it can be adjusted so that it is. Else, despite that you will see a sharp image on the focusing screen when focusing manually, and/or despite that the AF system will focus accurately, the sensor will not capture a sharp image.<br>

</p>

</blockquote>

<p>Now I get it. So the shimming is of the back and how it fits into its frame, not of the frame and how it fits into the camera. I assumed the latter, and thus couldn't understand why a fine focusing wouldn't just substitute for slight misalignment of the frame. Thanks for clarifying.</p>

<i>"the shimming is of the back and how it fits into its frame, not of the frame and how it fits into the camera"</i><br>It could be either, as long as the result is that the sensor is moved into the right position.

<blockquote>

<p><em>"the shimming is of the back and how it fits into its frame, not of the frame and how it fits into the camera" </em>It could be either, as long as the result is that the sensor is moved into the right position.<br>

</p>

</blockquote>

<p>Well I thought I understood this, but now I'm uncertain. Forget digital for a second and imagine an old fashioned camera with a film holder that gets swapped with a focusing screen. Now consider:<br>

<br>

Scenario One: The focusing screen and the film in its holder sit in precisely the same place in the camera, but that place is slightly off from where it is intended and so when the camera's focusing knob is set to what is marked as "infinity," distant objects are not in sharp focus.<br>

<br>

Scenario Two: The focusing screen sits in a slightly different place in the camera from where the film in the holder sits.<br>

<br>

Because my experience is with a 1940s press camera, I had assumed that the shimming problem referred to Scenario One, which doesn't strike me really as a problem. (I'm reminded of the famous "11" scene from <em>This is Spinal Tap</em>.) I perfectly well understand why Scenario Two is a serious problem that needs to be fixed. From your prior response, I thought you were simply telling me that shimming a digital back in a technical camera essentially addresses Scenario Two, which is why I thought I understood, but now I'm not so sure.</p>

Scenario Two is indeed the one.<br><br>The sensor, for some reason (it is said that the position is far more critical using a sensor than using film, so tolerances accepted in camera manufacture that were no problem using film are a problem when using a sensor, so the sensor position needs to be adjusted to the individual camera the thing is used behind. That would imply you can't use it on another camera of the same make and type without readjusting the sensor position) is not where it is supposed to be.<br>So the sensor position is adjusted.<br><br>That adjustment can be done by changing anything that has an effect on the sensor position, so shimming anything that is between camera and sensor would work. Moving the back itself a bit, or tweak the adapter that is used to attach it to the camera: either would work.

<p>Lobalodo... the article you refer too is one in Lula ...no? I personally wouldn't trust any source that is promoting any specific maker and "makes up" articles that are trying to show this maker as more careful than the others... <br /> The image area, in all cameras (SLRs, or technical, or view, ...or whatever) is calibrated <em>with respect to the focusing screen </em>of the camera... All MFDBs with constant fit <em>makers </em><em>assume that the focusing screen of the camera that the MFDB has been originally designed to fit </em>are made accurately and they make their backs for that ideal position... if the image area is off, there is a good chance that the flaw is with the mirror/focusing screen assembly (since all makers primarily design their backs for a MF dslr)... When Imacon (later Hasselblad) started their CF backs and Sinar their older backs, they took into account that there maybe small variations within samples of the same camera maker and they where providing a shimming plate for the back, so that the user could achieve "near perfect" calibration with respect to the focusing screen for the exact camera that he would use the back on... (my imacon 528c or the Sinarback e-motion I used to own was like that) ...that, was the best method! The user would have both interchangeable fit on his back ...and accurate calibration for his own camera... Now, for cameras in which a sliding device replaces the screen (after focusing) with the sensor, the shimming can be either on the back or on the screen itself! ...if the screen has to be taken off the camera and replaced by the image area and focusing is not accurate, again you will have to calibrate the screen or the back itself (I suggest the screen)! If a rangefinder design is used... then its best to first insure that the rangefinder focusing distance to subject is accurate and well calibrated with respect to what the lens is focusing at... but inaccurate shimming of the sensor of a back <em>with respect to the fit that it was originally designed for... </em>this I have never heart off... Regards, Theodoros. http://www.fotometria.gr</p>

<p>The situation with cut film and technical/monorail cameras is a bit different from using a digital sensor. For a start no two camera makers seem to have been able to agree where the register of the focal plane should actually be. It varies around 5mm by up to half a millimetre, or more in some cases. Notably, early MPP cameras and Graphmatic holders are well out of step with the rest of the world. To make things worse, most cut film holders have no real way of clamping the film flat and in exact register - some Linhof holders were an exception by incorporating a spring loaded pressure plate. All in all, the positioning of film in an LF camera is a pretty hit-and-miss affair.</p>

<p>Anyway; all this shouldn't really matter if a digital sensor is used to electronically view, focus and capture the image. It's only an issue if the camera is focused the old-fashioned way by eyeballing a ground glass screen before sliding or fitting the sensor into place. Even so the sensor itself has no need to be shimmed. All that needs to be done is to ensure that the ground glass is in the right plane, and this is far easier to achieve than messing about with the sensor.</p>

<p>After viewing the design of the ALPA cameras, I can see why their sensor position is such an issue. These are NOT technical cameras, for the simple reason that they allow no shift, swing or tilt of the front or rear standards, and some models don't even allow lens rise. Neither do they allow free interchange of lenses. They're simple rigid-body cameras that happen to take an MF digital back for untethered use. So, yes in this case accurate alignment of sensor with GG image is essential, but it still ain't rocket science to do so. And it's entirely wrong to class these as technical cameras.</p>

<blockquote>

<p>"Therefore, when you buy a digital back, the sensor will not be positioned exactly where it is supposed to be,...."</p>

</blockquote>

<p>So this guy at Luminous Landscape thinks he can do better, and with no specialist equipment, than Hasselblad and Leaf can? What a dreamer!</p>

<p>In scenario one you simply reset the infinity stops on what I assume is your Speed/Crown. The problem (scenario 2) is when the sensor position (film holder or digital back) deviates from the image you have focused on the groundglass.<br>

With digital backs a difference of a few to several microns can adversely impact the picture. Hence the need for precise shimming. </p>

A pressure plate behind a large sheet of film? Sinar's sticky back solution should have worked better. Yet very expensive, and not easy to load.<br>Anyway: changing the focussing screen position on a camera to cater for one, single back (amongst many) you may want to put behind it makes very little sense. It would mean that camera can be used with that particular back exclusively, or it will need undoing whenever you use any other back.<br>The best thing to do (and hence what is done) is to build the correction into the thing that is used to attach a (digital) back to a camera. Shim the adapter. Not the sensor (i agree), but not the focussing screen either. The adapter.<br>Shimming the adapter has the effect of positioning the sensor <i>"exactly where it is supposed to be"</i>. Not by changing it's position inside the DB it is built in, but by moving the back behind the camera.<br><br>You could do that yourself without specialist equipment. All you need is a place where a shim could be put. And do it. Specialist equipment to determine how much shimming to apply is not needed. Just put some in, take a test shot, and see what the effect was.<br>PhaseOne did have capture software that would help you adjust focus using a sort of live view and a graph display (available as a feature in their scanback software. Don't know whether it's available too in their 'regular' back's software).

<p>For large format film it means adjusting the gg to ensure it is as close as possible to the film plane, which varies by holder. It also means using tape/vacuum backs and a lot of testing of each holder. Most large format photographers do not do this rigorously, relying instead on the size of the film to mitigate some loss of sharpness, and by keeping track of their film holders to identify and remove ones that are problematic. Most medium format photographers rely on the precision of the manufacturing process for their equipment, and they fuss and fret over the ability of the film holder to maintain flatness. They identify holders/lenses/cameras that are probably within manufacturing tolerances but not ideal and replace them. With MFDBs, there are so many pixels crammed into a small sensor area that shimming the adapter often ends up on the to do list, alongside fussing and fretting over the other items mentioned. Add movements to the equation and you need a very precise camera that zeros perfectly and can achieve fractions of a degree of tilt/swing. Several microns of deviation can have a large impact.</p>