Mamiya C330 focusing

[clayne]

Keep at it Peter. I know you can lick this. Surprised those screws still wouldn't come out.

 

[PeterB]

The Film plane isn't where it should be!

Keep at it Peter. I know you can lick this.

I will prevail. Man this is harder than I first thought.

I put a roll of film through and shot a focus test sheet at 45°, and much to my amazement it was still out consistently by an amount between 1/2 to 1 depth of field distance for all distances tried.

I think I know what the problems are.

1. The first and biggest problem is that the emulsion plane is at least 0.2mm behind the plane of the inside film rails where I have been placing the Perspex with Scotch tape on it to focus. There is too much of a gap in the film transit area ! Using a CoC of 0.0474 my Depth of Focus at f2.8 varies between 0.266mm and 0.352mm when the object is either at infinity or 30cm from the lens plane. So close ups are therefore better. with such a small DoFocus, all it takes is for the film to move back from the rails to the plate by 0.133mm and the focus point on the object will be out by 1/2 a Depth of Field ! That is huge with such critical focussing requirements. Why didn't Mamiya or othe camera mfg fornthat matter put this disclaimer in their user manual?!?

2. The technique I had been using to compare the viewing focus to the film plane focus was to temporarily attach a sharp pin on the body atop the focusing knob and draw a fine mark on the focusing knob below the pin. I would refocus and see how far the mark had moved relative to the pin. I discovered this approach is flawed, and I should have realised this earlier. There is a small amount of mechanical hysteresis and other non repeatabilities which mean that even when focussing using the same view option (either viewer or film plane) that the mark didn't always return to the same spot!!! I have now changed methods and must use my ultra short term memory to try and see if the alternate view is in focus or not by turning the knob ever so slightly one way and then the other to see if it was in or out of focus. Harder but much more repeatable ! (focussing would be made easier if I had a split screen focus, but I only have a clear round circle in the middle of the viewing screen.)


Any ideas on how I should solve the problem of the really wide film transit area (it is at least 0.3mm I reckon)?

One idea is to place some contact sheet (adhesive on one side) on the pressure plate to force the film to press against the inside film rails. Another idea is to use some blank + developed film as a focus plane with a clear glass or Perspex block replacing the pressure plate, then I would need to readjust the height of my viewing screen up by about 0.2mm again.

 

[MattKing]

Peter:

The C330f backs are changeable. Maybe yours has been damaged or installed wrong.

 

[PeterB]

Peter:

The C330f backs are changeable. Maybe yours has been damaged or installed wrong.

Thanks for the suggestion Matt, but it would appear that Mamiya were indeed smarter than I gave then credit for above. I forgot to include the thickness of the paper backing in the 120 film I use! I did some measurements this afternoon.

The 120 HP5+ film plus paper is 0.14+0.1=0.24mm thick
The film channel is 0.35mm deep when the pressure plate is set to 120
The film channel is 0.25mm deep when the pressure plate is set to 220

I thought, beauty, I will just set my pressure plate to 220 when I run my film through and the emulsion will be within 0.05mm of the focal plane I tested to. nearly but not quite. I loaded a fresh roll of film, took off the lenses and poked a wooden skewer onto the film. low and behold the centre area (majority) of the film was bulging ever so slightly towards the front of the camera... I couldn't measure this accurately but my guess was about 0.1mm bulge. The edges of the film didn't bulge and were probably pressing against the pressure plate. I tried tensioning the film by advancing it as I poked, but that slight bulge persisted. This is an interesting observation because the natural curl in the film should be pressing it very flat onto the pressure plate. This bulge seemed to be the same amount irrespective of whether I set the pressure plate to 120 or 220.

Anyway these discoveries now give me reassurance that the main problem was most likely due to point 2. I made in my last post above about the non repeatability of the focusing knob's position and likely unrelated to film and perspex focal plane differences.

I will run a roll through again tomorrow with the plate set to the 120 position as it always has been.

 

[PeterB]

So I have repeated the focus test using this focus test chart at distances between 24cm and 2m. In summary I think my results are much better than before.
The chart is really only good to about 1.2m as it is too hard to accurately see and focus on the thin vertical lines after that.

One thing I didn't consider when taking the shots (or even when adjusting the height of the viewing screen when calibrating the focus) was to keep the test object perfectly vertical, and the camera perfectly vertical for close up shots (lets say between 24cm and 50cm). When the lens plane is not parallel to the object's plane then the closer you get, the greater the focus error is. I think this is fixed using Mamiya's paramender. Up until now I only thought the paramender provided a vertical offset, but if it provides an offset while keeping the lenses in the same plane (assuming the lens plane isn't vertical) then that would solve this problem. Surely other TLR users have this problem with close up photos if they don't use something like a paramender.

I now want to test focus for distances between 2m and infinity. Picket fences and home made collimator here I come ! Actually did you know that you can make a collimated light source using a 2nd SLR camera ?

 

[PeterB]

Finally comprehensive success with fixing and calibrating my Mamiya C330f focus !

Finally I have achieved comprehensive success with fixing and calibrating my Mamiya C330's focus !

I took the long and hard path when a few different decisions earlier on would have seen this solved months ago. Here is my advice for anybody attempting to fix/calibrate their TLR's focus.

1. Use a focus test chart for close up checking. i.e. distances that have the chart fill at least 2/3 of the viewing frame, preferably much more.
2. Ensure the test chart is 45degrees, but most importantly ensure the lines and paper chart are as close to vertical as possible using a small spirit level.
3. When using the test chart ensure the camera's lenses are vertical using a spirit level.
4. You will quickly see if the camera's lenses do not focus exactly on the test chart's centre line. Use this fact to align the viewing and taking lenses much more quickly than by trying to continually toggle the focus. (perhaps using a split screen focus screen would have helped here, but I didn't have one).
5. When testing the focus at other distances, you have two reliable options
   1. Use a set of vernier callipers to compare the distance your lens plane extends from the camera's body when focusing through the view screen and then at the film plane. If you don't have a split screen focus then vernier callipers are SOOO much more repeatable compared to trying to focus on a flat image and remembering how in/out of focus the prior view was, OR
   2. Find something like a picket fence or a brick wall that runs away from you at an angle and place a marker next to one picket or vertical mortar line and do comparative focusing on that.
6. When focus testing at infinity, use one of two methods
   1. A set of vernier callipers as above, OR
   2. The back sighting method (or a collimator if you have one!) to check/calibrate your camera. Fortunately in my case when it was in cal. close up, it was also in cal at infinity.
7. When you have finished using a ground glass/perspex sheet+magic tape at the film plane, run through a roll of film. Take shots of the focus screen at close distances, then for any further objects up to infinity choose any object with good amount of detail/local contrast and take 3 shots. One in focus, one with the focus slightly too close, one with the focus slightly too far. Then using a loupe and a light box, you can compare the three negatives for sharpness.
8. Know what tolerances you are aiming for. In my case I wanted at least 0.1mm repeatability since my Depth of Focus was only about 0.25mm. My callipers had a resolution of 0.01mm. When you focus on a line and measure the bellows extension with the callipers then re-focus your manual adjustment of the focus knob, you will not give better than a +/- 0.1mm repeatability.

In another post to this thread I will write about the problems I encountered and what I would have done earlier on to get a faster solution.

Here is a sample focus chart showing my success:

 

[PeterB]

Lessons Learned

Lessons Learned:

1. If your C330 photos are out of focus, here are the things to check in approximate order.
   • Push against the filter thread of each lens and ensure the lens plate under it is flush against the camera. If it moves, identify where the gap is. In my case the lens plate was slightly bent and I needed to shim between the silver retaining clip and the plate.
   • Replace the foam under the viewing screen. Keep in mind though that even if the foam is deteriorated it won’t necessarily be the cause of the problem. It needing replacing in my case but wasn’t the cause of the out of focus.
   • Ensure you didn’t lose a lens shim. Mine are keyed so I couldn’t accidentally swap them between the viewing/taking lenses.
   • Adjust the height of the viewing screen by screwing up or down the three posts under it (on the camera body). To ensure the top of these posts lie in a plane perpendicular to the lens plane, use a set of Vernier callipers or a dial gauge to perform measurements. You should be able to get better than +/-0.05mm accuracy here.
2. If the two lens offsets/heights from the camera body are different, that is OK. (Presuming your other lenses are also adjusted accordingly, but in my case I have no other lenses).
3. Use proper test equipment suitable for the task at hand. For example I purchased a dial gauge to measure the parallelism of the lens plane on the camera body/bellows. I made the mistake of not having a sufficiently flat surface under the film rails. I used what I thought was a square/flat block of wood, on a laminated benchtop, but the combination wasn’t level to <0.05mm. This mistake lead me down a very long and wrong path (see below). I solved this problem by having a thick block of Perspex cut to size to fit snug on the film rails and than sat that and the dial gauge on another flat sheet of larger Perspex. (see prior images in this thread for setup photos)
4. The reason I thought the lens plane’s supposed non-parallelism was worth pursuing was because the service manual instructs one to test for parallelism and had a mechanism for adjusting it if it was out. I then wasted lots of time pulling the side of the camera off only to be unable to unscrew a final plate (thankfully). After that didn’t work, I then tried shimming the lens plate away from the camera and removing one lens shim. All then looked good with the comparisons using a loupe at the lens plane, but that method only adjusted the centre of the viewing field for equal focus !! The way I found out it was wrong is when I had run a few rolls of film through and kept getting a non vertical/sloping depth of field in my shots !! (even after meticulously levelling the camera and target sheet of paper).

I hope this experience helps at least one other person in the future if their C330 or TLR is not focusing correctly. In fact, I’m willing to bet there would be quite a few cameras of all flavours (both analog and digital) out there with their focus out of calibration, but their users are oblivious to it !

 

[flavio81]

Bumped this thread because it is very important.

Some observations:

Do NOT do #1.4. In other words, never fiddle with the three lens posts that calibrate the focus screen height. They are very difficult to move and thus it is 99.99% likely they are perfectly set just like the day they are at the factory,

Now, in my experience the most likely causes of focusing errors on the Mamiya C330 (and similar are):

- #1.2 - the foam under the viewing screen. Really frequent problem!!
- #1.3 lens shims -- not necessarily lost but sometimes the taking or viewing lens is not screwed in tightly
- #1.1 lens plate not flush against the camera - with age the metal clip that retains the lens on the camera loses a bit of springiness and thus lets the lens plate wiggle a little bit. The solution is extremely easy and simple: Use pliers and bend back the metal clip a litte bit again, so, when the lens is installed, it pushes back the lens plate fully, with no wiggle.

For testing what i use is very simple -- a medium format focusing screen pressed firmly against the guide rails. I calibrate focusing to infinity, because that's where it is more sensitive to misalignment.

Another important note -- the testing and attempts performed by PeterB above assumes to calibrate the camera so at wide open aperture (i.e. F2.8) the objects are sharply focused. However, TLR cameras are not necessary calibrated in such a way. Due to curvature of field and focus shift (which makes the focus plane shift as you change the aperture), lenses are usually calibrated in a way such as center and corner resolution are sufficiently sharp at two or three stops down. Because for many of such lenses, calibrating for perfect center focus wide open will mean the corner performance will be weak at most apertures until you reach f16 or f22.

So, in short, calibrate infinity focus for a sharp center at f5.6 or f8, not wideopen.

The same will happen on folder cameras, for example. If you calibrate for sharp center at wideopen and then test the camera at smaller apertures like f8, you'll see the center remains very sharp but the middle and edge zones are unsharp.

 

[ic-racer]

I always laugh to myself when I learn of people using medium format film backs on 4x5 cameras without perfectly aligned center detents on the standards or even on cameras with no center detents. Like what are the chances you are going to get this lined-up correctly in the field.

 

[PeterB]

Another important note -- the testing and attempts performed by PeterB above assumes to calibrate the camera so at wide open aperture (i.e. F2.8) the objects are sharply focused. However, TLR cameras are not necessary calibrated in such a way. Due to curvature of field and focus shift (which makes the focus plane shift as you change the aperture), lenses are usually calibrated in a way such as center and corner resolution are sufficiently sharp at two or three stops down. Because for many of such lenses, calibrating for perfect center focus wide open will mean the corner performance will be weak at most apertures until you reach f16 or f22.

So, in short, calibrate infinity focus for a sharp center at f5.6 or f8, not wideopen.

Thanks for all that information. I'm trying to wrap my around this last statement you made. This is common practice when printing from an enlarger, and in a search I wasn't able to find advice against it. How many fractions of a mm will the focal plane (at the film) shift by when stopping down from wide open to f8?

Regards
Peter

 

[PeterB]

Thanks for all that information. I'm trying to wrap my around this last statement you made. This is common practice when printing from an enlarger, and in a search I wasn't able to find advice against it. How many fractions of a mm will the focal plane (at the film) shift by when stopping down from wide open to f8?
Regards
Peter

OK, partly answering my own question here. The effect is real and caused by lenses with spherical aberration. see here. http://photographylife.com/what-is-focus-shift

However the magnitude of the effect is still to be established. This effect being mentioned surely could not transfer from the focus calibration method to when using the lens+camera to take photos. The two are independent exercises. I agree that if when taking (non calibration) photos if one focuses wide open then stops down you might need to re-focus, but one could tell that by looking on the ground glass and noticing if the image is more or less in focus. Now to decide if this effect is really something worth considering, one could quickly do that experiment on a bright day. focusing wide open then stopping down to f8 and seeing if the scene needs re-focussing.

Alternatively a little maths may be in order. The lens I did the focus calibration with below, its smallest depth of focus (when focused at infinity) was about 0.25mm. So for this spherical aberration to have an effect worth worrying about it, changing from f2.8 to f8 must shift the focal plane by at least say 0.1mm. Does it ? There seems scant information in forums, I think I will need to reference a lens design text book or have somebody else chime in.

regards
Peter

 

[Xmas]

Focus shift depends on the lens and a system camera should allow for worst case lens(es).

The bowing of film away from the pressure plate is normal and also needs to be allowed for.

The Mamiya TLR (and Autcord) keep film flatter than most other TLRs cept the Rollis with glass plates, cause the film comes straight off the 120 (220) reel.

Lots of people wont go wider than f/8.

 

[Nodda Duma]

Spherical aberration will cause your focus shift. You could roughly characterize the amount of sph abb based on your focus shift but it's pretty much easier to just mark the offset for each aperture setting.

 

[PeterB]

Spherical aberration will cause your focus shift. You could roughly characterize the amount of sph abb based on your focus shift but it's pretty much easier to just mark the offset for each aperture setting.

Nodda, thanks for your reply. As an experienced LENS DESIGNER, I was wondering if you could quantify the shift in the focal plane when stopping down from f2.8 to f8 in the lens used which is described below. i.e., by how much will the focal plane shift (in mm). This lens' smallest depth of focus (When focussed at infinity) is about 0.25mm.

The lens is a Mamiya Sekor 80mm f2.8. It is a 5 element 3 group design based on the Cook Triplet. (taken from here and Dead Link Removed)

 

[flavio81]

Thanks for all that information. I'm trying to wrap my around this last statement you made. This is common practice when printing from an enlarger, and in a search I wasn't able to find advice against it. How many fractions of a mm will the focal plane (at the film) shift by when stopping down from wide open to f8?

Regards
Peter

No idea, but my suggestion is just to calibrate the lens using a ground glass on the film plane, but with the aperture stopped down one or two stops.

 

[flavio81]

Nodda, thanks for your reply. As an experienced LENS DESIGNER, I was wondering if you could quantify the shift in the focal plane when stopping down from f2.8 to f8 in the lens used which is described below. i.e., by how much will the focal plane shift (in mm). This lens' smallest depth of focus (When focussed at infinity) is about 0.25mm.

The lens is a Mamiya Sekor 80mm f2.8. It is a 5 element 3 group design based on the Cook Triplet. (taken from here and Dead Link Removed)
View attachment 97640

Peter,

The reason i pitch for not calibrating the center at wide open aperture is not only for compensating for focus shift at the center, but for taking into account the curvature of the focusing field (which if i recall correctly it will also change with aperture), so you get even (read: high performance) resolution at the center AND corners.

If Nodda is a lens designer i'm sure he will be able to explain this in depth.

PS: The 80/2.8 is not really based on the triplet but on the Leitz Elmar 90/2.8 configuration.