An optical mystery from adapted Mamiya N 43mm f/4.5 L lens

[rdeloe]

I have an optical mystery and would appreciate the thoughts of members who have more knowledge of optical design than I do.

I adapt Mamiya 6 and 7 lenses (among others) for use on my digital view cameras with a Fuji GFX 50R. To use a Mamiya 6 or 7 lens this way, they have to be modified (shutter blades, range finder and aperture controls and electronics all have to be removed). I mount them on custom mount boards for my F-Universalis and other setups.

I received a new lens this week: the Mamiya N 43mm f/4.5 L for the Mamiya 7 system. When I put it on my F-Universalis and made some test pictures, I was in shock. Image quality was horrendous. The centre was sharp from wide open, but the outer 20% of the whole GFX image was mush. I don't mean "a bit soft". I mean unusable mush. Conversion of these lenses is a one-way street, so I feared I had an attractive but expensive paper weight on my hands.

I've adapted and modified lots of lenses, so I sat down with it for a careful study. On a whim, I removed the factory-installed shim (spacer) between the rear lens group and the mount and made some test pictures. Image quality was now excellent. I experimented further and discovered that optimum image quality required a very thin spacer to replace the 0.5mm spacer installed by Mamiya; it's the thickness one gets by turning the lens group back 1/4 turn from fully tight without a spacer.

Based on everything I've learned over the years, this should be impossible. The Mamiya technician I use, Bill Rogers from Mamiya Repair, confirmed that it makes no sense. He puts every lens that makes it to his work bench on his autocollimator to check for issues; he had checked mine before starting the conversion, and the spacing was correct according to his measurements. This means it wasn't a case of someone installing the wrong spacer. The spacer the factory installed was corrected for this lens.

Interestingly, I also have adapted the Mamiya N 65mm f/4 L and the Mamiya N 210mm f/8 L, for the Mamiya 7, and the Mamiya G 50mm f/4, for the Mamiya 6. These lenses all work perfectly with whatever spacers were originally installed.

The only thing I've lost since replacing the thick spacer is the ability to focus to infinity on an adapter that is the correct length for Mamiya 7. That is a non-issue on my F-Universalis, so I now have an outstandingly sharp lens (albeit with very constrained movement potential because of the large rear lens group). It's a contender for the sharpest lens I've ever mounted on my GFX 50R, and that includes the native GF lenses I've used. The Mamiya N 43mm f/4.5 L aliases so much that I'd have to shoot it at f/16 all the time if I wanted to avoid aliasing. It's astonishing.

Anyway, the lens works so I should just be happy and accept that sometimes life gives us mysteries we can't solve. And yet, here I am on a camera forum that attracts some deeply knowledgeable people. I'm hoping someone can solve this mystery for me!

 

[buildbot]

I know nothing, but I am quite interested to know why this would be too!

 

[wallpaperviking]

I also have nothing to add but am curious to hear what other have to chime in...

How much movement can you get out of it with your 50R? Thanks.

 

[richardman]

I'm interested in such conversion, not for the same camera, but ;-)

 

[rdeloe]

I also have nothing to add but am curious to hear what other have to chime in...

How much movement can you get out of it with your 50R? Thanks.

I bought the lens knowing I wouldn't get much movement with it on an F-Universalis and GFX 50R outfit, and I was not wrong!

• I get some tilt and swing (although the rear lens is so close to the centre that an LCC is required with those movements).
• I was expecting around 5mm of shift, rise and fall. I get that much shift, and I can do a 5mm front rise, but I get 0mm front fall (or rear rise) because the "cowl" that protects the rear element catches on the part of the GFX camera that provides electrical contact with GF lenses. That I did not expect.

Really the main reason I bought this one was to pair up with my Mamiya N 65mm f/4 L in a two-lens outfit for use on an adapter that goes straight to the camera. It's a redundant focal length on my F-Universalis outfit because I have an excellent 35mm and 50mm lens.

Now down the road, my Mamiya N 43mm f/4.5 L might get another crack at being a lens for movements. I'd need two things: a BSI sensor (so I wouldn't have to do LCC frames all the time, or perhaps at all, when tilt/swing), and something with a shorter flange distance (some kind of a medium format back). In the meantime, it's going to work hard on my simple adapter.

 

[wallpaperviking]

I bought the lens knowing I wouldn't get much movement with it on an F-Universalis and GFX 50R outfit, and I was not wrong!

• I get some tilt and swing (although the rear lens is so close to the centre that an LCC is required with those movements).
• I was expecting around 5mm of shift, rise and fall. I get that much shift, and I can do a 5mm front rise, but I get 0mm front fall (or rear rise) because the "cowl" that protects the rear element catches on the part of the GFX camera that provides electrical contact with GF lenses. That I did not expect.

Really the main reason I bought this one was to pair up with my Mamiya N 65mm f/4 L in a two-lens outfit for use on an adapter that goes straight to the camera. It's a redundant focal length on my F-Universalis outfit because I have an excellent 35mm and 50mm lens.

Now down the road, my Mamiya N 43mm f/4.5 L might get another crack at being a lens for movements. I'd need two things: a BSI sensor (so I wouldn't have to do LCC frames all the time, or perhaps at all, when tilt/swing), and something with a shorter flange distance (some kind of a medium format back). In the meantime, it's going to work hard on my simple adapter.

Good One! That is what I was thinking, a 43mm on a BSI digital back...

 

[itsdoable]

Barring spacing issues with the reassembled Seiko shutter shell, the 43 has a short back focus, so the perifery rays are oblique, and have a longer effective path length through the cover glass of the sensor, which primarily shows up as field curvature. Can you refocus out at the edge? Respacing and adding a mild negative lens is one of the remedies.

 

[4x5Australian]

Rob, from your description of the events, I understand that:

(a) Before the modification of the Mamiya 7 43mm lens, Bill Rogers measured the spacing between the front and rear lens cells and found it was correct. (Your terms in italics.)

(b) After the modification, you found the image quality had become horrendous (gone to mush),

(c) Subsequently, you found that the lens quality was made excellent (again) when the spacing shim between the front and rear lens cells was changed for a thinner one.

I presume that the spacing between the two lens groups of the Mamiya 7 43mm lens is simply the cross-sectional dimension of the Seiko shutter plus a shim. Is that correct?

Just as a matter of interest:
Did Bill Rogers use a collimator to determine the distance between the cells, and then compare that measurement to a specified figure?
Or was he judging image quality by reference to a star test?

In any case, it appears that the reason the image quality deteriorated badly was that the cross-sectional dimension of the shutter housing increased slightly during the procedure.

The suggestion above by 'itsdoable' sounds plausible, and I like it. However, from your description it seems that no amount of re-focusing was able to make any of the peripheral field sharp. Was that the case?

Rod

 

[vjbelle]

• I was expecting around 5mm of shift, rise and fall. I get that much shift, and I can do a 5mm front rise, but I get 0mm front fall (or rear rise) because the "cowl" that protects the rear element catches on the part of the GFX camera that provides electrical contact with GF lenses. That I did not expect.

I don't have enough knowledge in lens design to add any help but was intrigued by your noticing the issues with the GFX camera and rear lens elements that could prevent movement once inside the cavity of the camera. I was confronted by the same issue as I had wanted to try to use one of the lenses listed by Arca as 'No Movements' when used with a GFX camera and it was that part of the GFX camera that provides electrical contact that would prevent some movement for fall. A little movement for rise is available. Once anything is inside the cavity of the camera movement becomes very dicy.

Victor B.

 

[Paul Spinnler]

The reason why you convert Mamiya 7 lenses is that they are relatively compact and still have a large IC? Are you happy with the IQ of all of them besides the 43?

 

[rdeloe]

Barring spacing issues with the reassembled Seiko shutter shell, the 43 has a short back focus, so the perifery rays are oblique, and have a longer effective path length through the cover glass of the sensor, which primarily shows up as field curvature. Can you refocus out at the edge? Respacing and adding a mild negative lens is one of the remedies.

Thank you for chiming in. This sounds very plausible.

A few things point to extreme field curvature:

1. The centre of the frame was fine with the original spacer. At f/4 focused on the centre of a flat target, there's a zone of good image quality from the centre out towards the edges. The transition from sharp to unsharp is quite rapid.
2. Depth of field helps. By f/22 the edges and corners show considerable improvement.
3. Focusing at the edge leads to an improvement in the formerly blurred image, but it does not get close to acceptable, and when I focus there, the formerly good central part goes bad. This is exactly what you would expect from extreme field curvature.
4. To calculate the size of the shim needed for best performance, I set up a large target on a flat wall with Zeiss stars in the centre and corners, focused on the centre at f/4, starting with no shim, and then gradually increased the distance between the lens groups by unthreading the rear one 1/4 turn at a time. Optimal was 1/4 turn, and after that it was very obvious with each 1/4 turn that the corners were going bad.

I haven't added any new glass to correct, so all the correction has happened due to changing the spacing between the front and rear groups. As I mentioned in the starting post, the difference is profound -- from unusable to one of the best lenses I've ever used.

Now this raises some more questions of course. By changing the spacing, what did I do to field curvature?

• Did I "flatten" the field, so that it's now properly flat across the whole image circle?
• Or did I just shift the curved field in a way that put more of the flat part where I need it on my 33mm x 44mm sensor?

If it's the first one, then the lens would be usable for shift across its entire circle of good definition.

But if it's the second one, then I've just kicked the can down the road, so to speak. In other words, excellent on a 33mm x 44mm sensor with no shift (or 5mm -- which I can use) may not mean excellent with a larger shift than 5mm because -- as mentioned in #1 above -- the transition from good to bad image quality is very rapid.

 

[rdeloe]

Rob, from your description of the events, I understand that:

(a) Before the modification of the Mamiya 7 43mm lens, Bill Rogers measured the spacing between the front and rear lens cells and found it was correct. (Your terms in italics.)

(b) After the modification, you found the image quality had become horrendous (gone to mush),

(c) Subsequently, you found that the lens quality was made excellent (again) when the spacing shim between the front and rear lens cells was changed for a thinner one.

I presume that the spacing between the two lens groups of the Mamiya 7 43mm lens is simply the cross-sectional dimension of the Seiko shutter plus a shim. Is that correct?

Just as a matter of interest:
Did Bill Rogers use a collimator to determine the distance between the cells, and then compare that measurement to a specified figure?
Or was he judging image quality by reference to a star test?

In any case, it appears that the reason the image quality deteriorated badly was that the cross-sectional dimension of the shutter housing increased slightly during the procedure.

The suggestion above by 'itsdoable' sounds plausible, and I like it. However, from your description it seems that no amount of re-focusing was able to make any of the peripheral field sharp. Was that the case?

Rod

Hi Rod. Thanks for jumping on this question. You'll see in my response to itsoable that I added some more info and questions. Let me answer yours here first.

(a) Yes, Bill measured using the procedure and tools described in the Mamiya 7 service manual. He has a short Instagram video showing how this is done. Here's the link: http://instagr.am/p/CghWa0DrTNa/ The service manual is excellent. It fills in the blanks on what Bill is doing in the video. In a nutshell, you use the special tools to measure the front and rear lens groups and then refer to the table in the manual for the spacer that's needed. He confirmed that the correct spacer was installed relative to the measurements he took.

(b) Bill sent me the lens back with the factory spacer installed, and the conversion completed (shutter blades removed, aperture control and range finder lever removed, electronics removed). I put it on my F-Universalis using the same board I use for my Mamiya N 65mm f/4 L (which worked flawlessly when I got it back from Bill). As explained to itsdoable, above, image quality outside of the central role was mush across the full aperture range. Ye Olde Brick Wall at f/4 is attached so you can see how we go from excellent in the centre to mush in the corners and edges, and how fast the transition is.

(c) Correct -- when I removed the shim and wound the rear lens group down tight, I concluded that I might not have in fact be the owner of a lens-shaped paper weight. As explained in the earlier post, I removed the shim on a hunch. It's a used lens and you never know what a previous owner or repair person did. Sometimes the wrong shim is installed. So I pulled it out to see. I was delighted to see the improvement, but also mystified when Bill reported it was the correct shim. I tested to see whether no shim or a thinner shim was required, and it does make a good difference at f/4. Attached is a diagram showing what the upper-right Zeiss star looked like for the first four shots in the test sequence. The 1/4 turn version was clearly the best.

The spacing is simply a function of the distance between the two lens groups that are threaded into the shutter. It's a very complex optical design, but still unit focusing (so no floating elements). To increase the spacing between the two groups, Mamiya simply sticks various thickness spacers between the shutter and the rear group (always the rear group per the Mamiya service manual).

And yes, as I noted for itsdoable, no amount of re-focusing at the edges fixed the problem -- which to me (in hindsight!) says very strong field curvature makes sense. I've used some lenses where the combination of small aperture and careful focusing at the edges could balance it out across the frame and get decent image quality. The field curvature of this lens with the "correct" spacer made this impossible.

 

[rdeloe]

The reason why you convert Mamiya 7 lenses is that they are relatively compact and still have a large IC? Are you happy with the IQ of all of them besides the 43?

It all started with 50mm.

On my setup (GFX 50R with various digital view cameras -- currently an F-Universalis), I had a good 35mm and 90mm lenses, but in between was a gap. The 60mm lens I used (APO-Componon HM 60/4) has a too tiny image circle, and was a bit long. I wanted something closer to 50mm. Everything else I tried didn't cut it for various reasons. The Rodenstock APO-Grandagon 55mm had lots of lens cast and banged into my camera because of the short flange distance and long rear lens group. The Pentax 645 55mm wasn't good at the edges until f/11 unshifted. The Pentax 67 55mm f/4, 3rd generation, is very good but huge and hard to mount. Some others I tried were just dodgy (e.g., an old Nikkor-O for Bronica).

I'd just about given up when I found a post by Christoph Kugler, who adapted Mamiya N lenses (among others) to his Phase One back. I chose the Mamiya G 50mm f/4 for Mamiya 6 because it has a much shorter rear lens group than the Mamiya N 50mm for Mamiya 7. It proved to be compact, relatively light, excellent from wide open, plenty of image circle, and no lens cast. On my F-Universalis, the rear lens group clears the Rotafoot, so there are no mechanical constraints on movements.

The place I bought the 50mm from also had a Mamiya G 150mm f/4.5 for a great price, so I took a chance on that one. It's also an outstanding lens: small, compact, sharp wide open, excellent colour and contrast -- just fabulous. I think the optics in the Mamiya N 150mm f/4.5 L for Mamiya 7 are the same as the G version (but I haven't tried it).

I added the Mamiya N 65mm f/4 L a month or so agao because I wanted a "normal" focal length lens. Not only is it superb as a technical camera lens, but also it's almost exactly how I would make a lens for use directly on the camera. The aperture ring is up front where I like it, the focus is smooth and in the right place for my hand. It's just a pleasure to use as a "walk around" lens when I don't want to carry my pack with the F-Universalis and all the trimmings.

If you have kids or grand kids you might know the book, If You Give a Mouse a Cookie... “If you give a mouse a cookie, he's going to ask for a glass of milk. When you give him the milk, he'll probably ask you for a straw. When he's finished, he'll ask you for a napkin.”
That's why I bought the 43mm. On my F-Universalis I don't really need something between 35mm and 50mm, but that 65mm was lonely and needed a 43mm to make a nice two lens set for going out just with the 50R (and no tech camera).

And the 210mm is just because I've always wanted something longer than 150mm! I've gone through multiple unsatisfactory options on that focal length too. The 210mm is very good, but my initial evaluation is that it's not quite in the same league as the 150mm.

 

[rdeloe]

I don't have enough knowledge in lens design to add any help but was intrigued by your noticing the issues with the GFX camera and rear lens elements that could prevent movement once inside the cavity of the camera. I was confronted by the same issue as I had wanted to try to use one of the lenses listed by Arca as 'No Movements' when used with a GFX camera and it was that part of the GFX camera that provides electrical contact that would prevent some movement for fall. A little movement for rise is available. Once anything is inside the cavity of the camera movement becomes very dicy.

Victor B.

Very dicey!

If I really wanted to use the N 43mm on my F-Universalis + GFX 50R outfit, I could replace the factory cowl with one that cleared the electrical contacts. I built one and tried it so I know it works. However, because it's a rail camera it's very easy to accidentally come too close, and I'm not keen on banging into things inside the sensor cavity on a regular basis. I'm also worried about hitting the glass; it's such a big, bulbous piece of glass and the factory cowl barely offers enough protection as it is.

My bottom line for this lens was that it worked well with my "straight" adapter for GFX. Thankfully it's superb as "just" a straight lens.

 

[vjbelle]

When I get my M-Two the first thing I'm going to test is to see if I really can focus to infinity with the Schneider 35XL. The Arca lens compatibility sheet lists a Schneider 35L and as far as I know there never was such a lens so I'm thinking it's just a typo. It would be great to have that lens available. No distortion and sharp if not shifted much. The rear lens portion will be inside the cavity but it's fairly small so I think I could actually have 7mm worth of rise. It's not a lens I would use much but good to have.

Will see tomorrow......

Victor

 

[rdeloe]

When I get my M-Two the first thing I'm going to test is to see if I really can focus to infinity with the Schneider 35XL. The Arca lens compatibility sheet lists a Schneider 35L and as far as I know there never was such a lens so I'm thinking it's just a typo. It would be great to have that lens available. No distortion and sharp if not shifted much. The rear lens portion will be inside the cavity but it's fairly small so I think I could actually have 7mm worth of rise. It's not a lens I would use much but good to have.

Will see tomorrow......

Victor

Very exciting! It's nice to see more Arca Swiss F-Universalis and M-Two users. Fingers crossed the lens actually works.

 

[Paul Spinnler]

It all started with 50mm.

On my setup (GFX 50R with various digital view cameras -- currently an F-Universalis), I had a good 35mm and 90mm lenses, but in between was a gap. The 60mm lens I used (APO-Componon HM 60/4) has a too tiny image circle, and was a bit long. I wanted something closer to 50mm. Everything else I tried didn't cut it for various reasons. The Rodenstock APO-Grandagon 55mm had lots of lens cast and banged into my camera because of the short flange distance and long rear lens group. The Pentax 645 55mm wasn't good at the edges until f/11 unshifted. The Pentax 67 55mm f/4, 3rd generation, is very good but huge and hard to mount. Some others I tried were just dodgy (e.g., an old Nikkor-O for Bronica).

I'd just about given up when I found a post by Christoph Kugler, who adapted Mamiya N lenses (among others) to his Phase One back. I chose the Mamiya G 50mm f/4 for Mamiya 6 because it has a much shorter rear lens group than the Mamiya N 50mm for Mamiya 7. It proved to be compact, relatively light, excellent from wide open, plenty of image circle, and no lens cast. On my F-Universalis, the rear lens group clears the Rotafoot, so there are no mechanical constraints on movements.

The place I bought the 50mm from also had a Mamiya G 150mm f/4.5 for a great price, so I took a chance on that one. It's also an outstanding lens: small, compact, sharp wide open, excellent colour and contrast -- just fabulous. I think the optics in the Mamiya N 150mm f/4.5 L for Mamiya 7 are the same as the G version (but I haven't tried it).

I added the Mamiya N 65mm f/4 L a month or so agao because I wanted a "normal" focal length lens. Not only is it superb as a technical camera lens, but also it's almost exactly how I would make a lens for use directly on the camera. The aperture ring is up front where I like it, the focus is smooth and in the right place for my hand. It's just a pleasure to use as a "walk around" lens when I don't want to carry my pack with the F-Universalis and all the trimmings.

If you have kids or grand kids you might know the book, If You Give a Mouse a Cookie... “If you give a mouse a cookie, he's going to ask for a glass of milk. When you give him the milk, he'll probably ask you for a straw. When he's finished, he'll ask you for a napkin.”
That's why I bought the 43mm. On my F-Universalis I don't really need something between 35mm and 50mm, but that 65mm was lonely and needed a 43mm to make a nice two lens set for going out just with the 50R (and no tech camera).

And the 210mm is just because I've always wanted something longer than 150mm! I've gone through multiple unsatisfactory options on that focal length too. The 210mm is very good, but my initial evaluation is that it's not quite in the same league as the 150mm.

Do you know where Christoph adapted these lenses? Or did he do this himself?

How big are the ICs? 120mm?

 

[rdeloe]

Do you know where Christoph adapted these lenses? Or did he do this himself?

How big are the ICs? 120mm?

Christoph had his Mamiya lenses converted by Mamiya in Germany, but by the time I contacted him in 2020, they weren't doing it anymore. I called Bob Watkins at Precision Camera first but he was not interested in doing the job. Then I found Bill Rogers. There's a Mamiya repair person in Belgium too who, I'm told, will do it. Bill is the best though and he has lots of practice now.

In Mamiya's lens data reports, they put the X axis on their MTF charts out to 44.5mm, which means 90mm image circle. That makes sense for 6x7 film. Some of the lenses do have a larger circle of illumination and correspondingly larger circle of good definition, so shifting a bit past 20mm on 33mm x 44mm (which is a 90mm IC) may be possible.

On a setup like mine there are other factors in play. For example, I think there's more good image circle on the 65mm than I can use. I get mechanical vignetting from the upper edge of the sensor cavity in my 50R at around 20mm because of how close the rear lens is to the sensor. I have other lenses that don't hit this threshold until 30mm (not Mamiya 7 lenses mind you).

A camera with a shorter flange distance would be a whole other story. I expect someone using a Hasselblad CFV 50 ii on my setup would get access to more of the image circle.

Out here in the wilds of adaptation land where I dwell, you have to try to find out for sure.

 

[Paul Spinnler]

Thanks, very informative. The point about the M7 lenses though is that they are not thaaat cheap, one could as well use a good analogue era lens (eg Grandagon); do you find them especially sharp or comparable to modern tech cam lenses? They are compact, I guess.

Very curious to see whether the 43 can be fixed. Incidentally this could be a killer architectural lens

 

[itsdoable]

RE: lens cast - I'm surprised you didn't notice any lens cast on the 43mm if there is significant field curvature from the sensor glass, they usually go hand in hand.