Sironar Digital/Grandagon 55mm: Really Soft Corners

[MsLing]

I just got a Sironar Digital 55/4.5 lens, which was mounted in a Sinar DB lens panel. I measured its length before moving it to another Copal shutter that only cost less than $30. It's 70.94mm. Then I measured again after mounting it, 69.50mm, a horrible shutter, far away from standard, 20±0.025mm. That's why my late batch Symmar-S which mounted in this shutter isn't sharp I guess. Anyway, I only ordered 10 pieces of 0.02mm shims, they weren't enough to adjust this shutter and I had to buy some ready-made products to fix my shutter. Finally, yesterday, I used 4pcs 0.2mm, 3pcs 0.1mm and 2pcs 0.02mm shims and successfully adjusted it to 70.94mm.

Today, I took it out to test it, with Sinar P and Kodak Pan-X 3412 film. All sheets were shot at f/11～16. I do know Rodenstock did some tiny trims on Apo-Grandagon 55mm and renamed it, then sold it as Apo-Sironar-Digital.After processing, it's sad that contrast with really high resolution centers, all corners are soft and fuzzy. There are two examples below. Maybe only 90mm acceptable IC I can get from this lens. I plan to test it again with and without shims and shoot at f/22 tomorrow.

BTW, if there's anyone have experience with this lens? What result you got?

 

[MsLing]

I'll change shims' place to see if place of iris will influence it. According to datasheet, Grandagon and Sironar Digital are both recommended to work at f/8-11. It's not normal I think

 

[koraks]

Forgive me if this is a bit of a silly remark, but isn't this lens intended for smaller digital backs and thus cannot be expected to actually cover 4x5" film format? As far as a quick online search suggests, this lens should project a usable image circle of about 125mm. The diagonal of 4x5" nominal is around 160mm. If you've used any movements or tilt/swing, the usable image circle at the film plane will fall even more short.

 

[MsLing]

Forgive me if this is a bit of a silly remark, but isn't this lens intended for smaller digital backs and thus cannot be expected to actually cover 4x5" film format? As far as a quick online search suggests, this lens should project a usable image circle of about 125mm. The diagonal of 4x5" nominal is around 160mm. If you've used any movements or tilt/swing, the usable image circle at the film plane will fall even more short.

Hmmm, I didn't use any movement. The main reason why I use 4x5 is to check if I can get a 125mm high quality image circle because I plan to use it with digital and 6x9 back. But obviously 125mm is impossible on this copy.

 

[koraks]

Sharpness degradation does start closer to the center than I would have expected - although I really can't tell for sure. What usable image circle did you get on these two sheets of film?
Also, have you tried the lens without the additional shims?

 

[ic-racer]

Keep testing. But if 90mm of circle is all that meets your standards, then you may have higher standards than the manufacturer. A not-unheard-of scenario in photography.

 

[Light Capture]

It needs to be optimized at infinity for that particular shutter.
What's your method for shimming?

Proper procedure should be focusing center to infinity and shimming it until corners get as sharp as possible. It involves refocusing as needed and evaluating.
If you want to optimize it to some other magnification than really parallel to film target is needed and then it's optimized for that distance in the same way.
This is simplified way. If you wanted maximum performance all of this would be done with resolution targets and measurements. The best results would be extrapolated from measurements and optimal adjustment would be taken.

Other causes could be distortion of the shutter of some kind. Front and back rim distance needs to meet the specs and they also need to be parallel to the specifications. If it's off, front and back cells won't be parallel and will cause what you see.

From looking at the negative, if what's seen is what's really on the negative, it seems that negative gets visibly blurry before 125mm circle. The only way to say if it meets the specs or not is to shoot resolution target and measure.

 

[Lachlan Young]

There's some old Sinar data in their 2000-ish catalogues that claims that the DB mounted version of this lens has a 101mm image circle as opposed to other mounts that deliver 120mm image circles.

They likely had it adjusted for optimal performance on their (fairly small chip) backs of that time - and it's worth noting that they wanted to suggest it as the digital equivalent of a 210mm on 4x5 - which essentially means that it was probably shimmed to get the best centre/ edge performance trade-off across that smaller field. You may also be running up against Rodenstock and Sinar having modified the whole barrel to this end - the overall length of a 120mm image circle version is specified as 73.8mm

 

[MsLing]

What usable image circle did you get on these two sheets of film?

About 80-90mm.

Also, have you tried the lens without the additional shims

Doing this now.

 

[MsLing]

It needs to be optimized at infinity for that particular shutter.
What's your method for shimming?

Proper procedure should be focusing center to infinity and shimming it until corners get as sharp as possible. It involves refocusing as needed and evaluating.
If you want to optimize it to some other magnification than really parallel to film target is needed and then it's optimized for that distance in the same way.
This is simplified way. If you wanted maximum performance all of this would be done with resolution targets and measurements. The best results would be extrapolated from measurements and optimal adjustment would be taken.

Other causes could be distortion of the shutter of some kind. Front and back rim distance needs to meet the specs and they also need to be parallel to the specifications. If it's off, front and back cells won't be parallel and will cause what you see.

From looking at the negative, if what's seen is what's really on the negative, it seems that negative gets visibly blurry before 125mm circle. The only way to say if it meets the specs or not is to shoot resolution target and measure.

I didn't shim front block, obeyed what Sinar did on the DB panel. I'll try method you mentioned, thanks. And I mounted my Apo-Symmar in DB panel, then measured. It was 0.50mm longer than in original Copal shutter. So I think what @Lachlan Young said makes sense.

 

[MsLing]

OK, I compare image with and without 0.08mm shims. Image with shims looks better than without one. But I doubt that I didn't focus Well. I am going to try more and less shims

 

[MARTIE]

Hello,
1. Did you make any photos with this lens in the dB mount to find out if was sharp?
2. We're shims used in the dB mount?
3. Can you post what source documentation you are using for your measurement calculations?

I could be way off, but I've looked at 3 pdf's and all appear to give completely different measurements, including differences between the 'Sironar digital' and 'Rodenstock Grandagon' versions of the 55mm f4.5.

I'll include a screenshot from the 'Apo-Sironar digital' brochure.

 

[MsLing]

Hello,
1. Did you make any photos with this lens in the dB mount to find out if was sharp?
2. We're shims used in the dB mount?
3. Can you post what source documentation you are using for your measurement calculations?

I could be way off, but I've looked at 3 pdf's and all appear to give completely different measurements, including differences between the 'Sironar digital' and 'Rodenstock Grandagon' versions of the 55mm f4.5.

I'll include a screenshot from the 'Apo-Sironar digital' brochure.

Oh I don't have a DB shutter or a Expolux shutter so I can't do any shot with the original DB panel.

Original DB mount used a small metal ring like the one in Copal version Rodenstock lenses as shim.

I only use vernier caliper to measure my lens. And I read all Rodenstock and Sinar datasheet as you show, its length is quite different.

 

[MARTIE]

So perhaps in the absence of precise and accurate knowledge of the correct measurements, you'd just unscrew the front lens block until the image was sharpest on the ground glass, with a powerful magnifier, and then shim the lens to that overall measurement?

Or is that too simplistic?

 

[MsLing]

So perhaps in the absence of precise and accurate knowledge of the correct measurements, you'd just unscrew the front lens block until the image was sharpest on the ground glass, with a powerful magnifier, and then shim the lens to that overall measurement?

Or is that too simplistic?

Theoretically it is. However, unfortunately, neither ground glass nor LCD diffuse film can't be regarded as a optimal choice. When using small aperture, grains are terrible even under 10x lupe. And to see all high frequency details, we need to use at least 20x. Another worse fact is that my 20x lupe don't allow me focus after a piece of 2mm glass. Another reason why ground glass is unusable is that due to its short focal length and large incident angle, corners are not that visible.

Digital can be used for adjusting lens. Users of Getdpi always do this. However, also unfortunately, WA bellow restricts movement, therefore, I can't check corners. That's why I had to use film : (

Ps. I place shims on rear block because I'm worried that if iris' place would influence the image quality or not. I only have Grandagon-N's patent so I can't do a simulation.

 

[MsLing]

I compare image with 0.00, 0.04, 0.08, 0.12, 0.16mm and 1.5mm shims, 0.02-0.06mm maybe good enough at center and corner.

 

[MsLing]

After final check with digital, I decided to use 0.02mm shim. BTW, additional shims do increase resolution of center, but highly reduce corner's

 

[koraks]

I'm starting to wonder if something odd is going on with your lens - e.g. this being some kind of Frankenstein combination of the front group of one lens and the rear section of another. The kind of unsharpness variation you're getting across the frame reeks of an unusually poorly compensated optical system - the kind you get when you flip one section back to front, have vastly incorrect spacing (much beyond tenths of a millimeter), remove elements from groups or, indeed, combine groups from different lenses into a new one.

Can you post clear photos of both sections of this lens (back & front)?

 

[MsLing]

OK, here are pictures. After remove shims, corners look quite nice.

 

[Lachlan Young]

After final check with digital, I decided to use 0.02mm shim. BTW, additional shims do increase resolution of center, but highly reduce corner's

That alteration makes a lot of sense if you consider that high end sensors of that era were often working with a 24x36 or 36x36 area (delivering maybe 6-16mp).

 

[reddesert]

Oh I don't have a DB shutter or a Expolux shutter so I can't do any shot with the original DB panel.

Go out at twilight when shutter speeds are longer and use a hat as a shutter. I forget if you'll have control over the aperture, but if not you can shoot it wide open. Even wide open, the corners of most modern lenses should be much better than the example in your first pictures.

There's a common thread here which is that you have two remounted lenses that give poorer results than you expect. Measuring barrel lengths, shimming, and comparing to lens spec sheets can only get you so far. Test the lenses in their previous mounts.

 

[MsLing]

Go out at twilight when shutter speeds are longer and use a hat as a shutter. I forget if you'll have control over the aperture, but if not you can shoot it wide open. Even wide open, the corners of most modern lenses should be much better than the example in your first pictures

I can't control its aperture without shutter so it can only wide open. I'll try your suggestion next time. Processing these 0.06mm sheets is an annoying thing, at least I don't want to do one more time this week @_@