Adapt Scanner Lens for Camera-Scanning??

[shijan]

Here are my side by side examples with FUJIFILM X-E4 APS-C sensor. DiMAGE SCAN 5400II lens is really perfect in the center. 100% no CA, no distortion, no flares, very sharp. But at this magnification it became soft in the corners.

 

[ColSebastianMoran]

Here are my side by side examples with FUJIFILM X-E4 APS-C sensor. DiMAGE SCAN 5400II lens is really perfect in the center. 100% no CA, no distortion, no flares, very sharp. But at this magnification it became soft in the corners.

Shijan, what is the magnification for this comparison? 1:2 or 0.5x? And, can you possibly share a 400% screenshot of the USAF pattern at the corner?

Robert O'Toole says the Dimage 5400 lens is best he's seen at 2x, and that would make it terrific at 1:2, but he doesn't comment about coverage. Savazzi, another dependable source suggests there is curvature of field at 2x. To test this, you could make your same shot, but focus in the corner, not in the center. ( I have no u4/3 body, so not easy for me to duplcate your test).

You wrote, "I can perfectly focus to the corners or to the center. But not both." I belive that's a sign that there is field curvature (assuming your alignment is good).

 

[ColSebastianMoran]

FWIW, here are 400% screen grab comparisons for the Dimage 5400 lens at 0.67x to a 24MPx APS body:

Skip centers; centers are just fine.

Corner comparison, no problem for either lens. G0-E1 is just about the max realizable resolution for one-shot to a 24MPx sensor. Dimage 5400 result looks just a bit better than my benchmark lens, the 70 Sigma.



Now let's look at the VERY extreme corner, within a mm or so of the real corner, still just fine. Any problem in the corner will make a mess of this tiny image area.

 

[shijan]

I use 5400 lens at 0.67x magnification. So just project full film frame to APS-C sensor.
Corner at your example looks perfect. So it could be some problem from my side:
- Lenses are always slightly different. So this could some unlucky bug in my lens.
- It also could be due film curvature tolerance combined with my film holder.
- Seems the main problem source is my bellows. I just noticed that bellows are wobble a little bit even when lock screw is fixed. My setup is vertical, so due weight lens became under some angle compare to sensor. In optical systems this problem named Astigmatism. Guess for 24mm F3.9 reversed 5400 lens (based on closeuphotography measurements https://www.closeuphotography.com/mi...lite-5400-lens) perfect alignment it way more critical than to 50mm F5.6 enlarger lens.

I created test shoot in horizontal position and corners looks sharper, but still not so sharp as in your example. (you can open full sized images in new tab or download them)

I use Novoflex Minolta Bellows modified with RAFCAMERA plates with M39x1 thread. This adapter was manufactured based on my idea and measurements, so you can see actual pictures of my bellows there:
https://rafcamera.com/plate-novoflex...ws-rear-m39x1m
https://rafcamera.com/plate-novoflex...s-front-m39x1f
I also use 18mm to RMS thread adapter for 5400 lens
https://rafcamera.com/clamp-18mm-to-rms-m
And RMS female to M39x1 male thread adapter with M42x1 front female thread (this adapter also was manufactured based on my idea)
https://rafcamera.com/adapter-rms-f-to-m39x1m-m42x1f
On top of that i put lens hood made from M42 extension rings.

By the way, i find interesting lens named Osawa Tominon EL 1:2,8/50 mm. It looks like some modification (or newer version?) of Tominon with wider aperture and with normal M39 thread. It should arrive in month or so and i could test it...

 

[ColSebastianMoran]

I use 5400 lens at 0.67x magnification. So just project full film frame to APS-C sensor.
Corner at your example looks perfect. So it could be some problem from my side:
- Lenses are always slightly different. So this could some unlucky bug in my lens.
- It also could be due film curvature tolerance combined with my film holder.
- Seems the main problem source is my bellows. I just noticed that bellows are wobble a little bit even when lock screw is fixed. My setup is vertical, so due weight lens became under some angle compare to sensor. In optical systems this problem named Astigmatism. Guess for 24mm F3.9 reversed 5400 lens (based on closeuphotography measurements
I created test shoot in horizontal position and corners looks sharper, but still not so sharp as in your example. (you can open full sized images in new tab or download them)

Ah, yes, your body is APS, I was confused. This lens, at 0.67x, oriented so that the double ring is to the sensor, looks very good to me. If the corners aren't sharp: 1) alignment, 2) film flatness, or 3) bad lens. To test, I would try a test focusing carefully on the extreme corner.

And, overall, limited DOF and no aperture control are drawbacks of these scanner lenses.

 

[shijan]

I attempt to reassemble my stand system and see if it will work with Macro Focusing Helicoid instead of bellows. Unfortunately that China-made helicoid also wobble and produce random off-centered blurry corners. Not precise at all. Wider extended - more wobble. So it is not a solution for 5400 lens.

 

[shijan]

So i search for possible bellows replacement and came to Asahi Pentax Bellows Unit II. It is simplified and reduced in size model compare to older Asahi Pentax Auto Bellows. http://www.pentaxforums.com/content/...ax_Bellows.pdf The only thing i don't like - those simplified bellows don't have secondary fine adjustment mechanism. It can only slide and lock. Also there are reports that nylon inserts may crack due age. But guess those could be replaced by 3D printed ones. Bellows have M42 threads on both sides, so with seamless M39 to M42 adapter ring i can connect any existing M39 lens. Similar M40 to M42 adapter ring also exists for original Tominon lens.
It was a luck to find low cost compact sized adapter M42 to FX designed specially for macro bellows https://www.ebay.com/itm/184887427125 https://www.ebay.com/itm/165689465658 (similar adapters also available to all other mirrorless cameras). So i can connect camera without extending minimum distance to the lens.

Another interesting M42 threaded candidate was KONICA AUTO BELLOWS AR (also known as Mamiya Sekor, CONTAX, YASHICA bellows). It is large sized, but rail consists of two parts and have few threaded mounting holes at the bottom, so it is possible to mount it directly to 15mm rod blocks without additional plates. This will reduce size and weight of overall setup.

Also rails in these bellows are compatible with Arca Swiss clamps! See explanation video:

 

[shijan]

So Asahi Pentax Bellows Unit II and some other parts for stand improvement now arrived. After assembly things became look much better. Minolta 5400II lens really needs insane amount of flatness and optical axis centering. It literally sees 0.1-0.2mm film curvature, so it is rather problematic to scan film in any basic film holder. This lens should work great for perfectly flat wet mount glass film holders.
Asahi Pentax Bellows Unit II is very precise and stable tool. It have only one wheel to smoothly adjust distance from lens to camera and have no secondary adjustment wheel to move all system. This makes focusing and adjustment rather problematic and uncomfortable because changing distance from lens to camera changes field of view. But i find a way to modify stand system to fix limitation in these bellows.
I also find technically precise way to adjust bellows left/right angle - expand bellows and put lens mount to the bottom plate or to light box. Then lock bellows and then tight screw that connects bellows to cheese plate. if bellows are too short and can't reach bottom plate, you may add some M42 extension rings.

100% corner/center crops:

Download full size image here:
https://drive.google.com/file/d/1ANR...usp=share_link

Improved Stand components:

SmallRig Multi-purpose Cheese Plate 1092

SmallRig 15mm Rods 15cm

SmallRig Super lightweight 15mm RailBlock v3 942

1/4"-20 Button Head Screws with 1 1/2" length (38.1mm). Those extended length screws used to attach two rail blocks to base plate. Dual rail block base creates very stable and rigid structure.

M5x18mm Hex Socket Head Cap Screws for bottom rail blocks.

CAMVATE Camera Baseplate Extension Long Cheese Plate. This low profile plate fits inside rail block, so bellows now mounted closer to rails center axis.

M42 to FX Modify Lens low profile adapter

M39x1 to M42x1 lens adapter

RafCamera RMS female to M39x1 male thread adapter with M42x1 front female thread

or RMS female to M39x1 male thread adapter

or RMS female to M42x1 male thread adapter

RafCamera 18mm clamp to RMS male thread adapter for Minolta 5400 DPI scanner lens

1/4"-20 Thumb Screw and 7mm (or larger) Steel Ball. First you slide bellows by hand and lock at some position. Then this thumb screw moves all system up/down and provides precision focusing adjustment. Don't forget to lock rail blocks clamps when focusing is done.

 

[shijan]

Another possible precise option is M42 WeMacro Tubes System (Formally a clone of larger sized and higher cost SM2 Thorlabs Tubes System described here https://www.closeuphotography.com/52mm-setup). Guess clamps could be attached to 15mm Rod Blocks. And you can use similar DIY thumb screw focusing mechanism. But compare to regular bellows this system is less comfortable. It also require some DIY bellows hood between lens and camera because unfortunately WeMacro don't manufacture Variable M42x1 Extension Tube. So fine tuning and focusing could be problem.

System consists of 50mm M42x1, 10mm M42x1 Extension Tubes (actually you can use any other tubes with 46.3mm external diameter) and Tube Clamps

There is also Variable M42x0.75 Extension Tube and Extension tubes made by SVBONY, but they use M42x0.75 thread (T2 mount) and external size about 44.5mm which makes them incompatible with WeMacro Tube clamps

 

[ColSebastianMoran]

I attempt to reassemble my stand system and see if it will work with Macro Focusing Helicoid instead of bellows. Unfortunately that China-made helicoid also wobble and produce random off-centered blurry corners. Not precise at all. Wider extended - more wobble. So it is not a solution for 5400 lens.

Yes, wobble is an issue, but I have used these successfully in a vertical arrangement being careful to avoid any vibration.

A better current item is the UniFoc, very solid and precise. One seller offering right now in eBay, item # 183398997115



There is also the older item from Rodenstock/Linos called "Modular Focus"

 

[ColSebastianMoran]

in #12 above I had a link that's now broken. Here's a corrected link to a page on use of various lenses for camera scanning.

 

[ColSebastianMoran]

Another possible precise option is M42 WeMacro Tubes System (Formally a clone of larger sized and higher cost SM2 Thorlabs Tubes System described here https://www.closeuphotography.com/52mm-setup). Guess clamps could be attached to 15mm Rod Blocks.

Thank you, @shijan, for this. I now have the M42 WeMacro tubes and they are excellent.

 

[shijan]

Some additional thoughts about Dimage Scan lens and APS-C sensor according to some updates from here Minolta DiMAGE Scan Elite 5400 Lens — Close-up Photography

FUJIFILM X-E4 sensor:
15.6 x 23.5 mm

5400 II Scanner:
Scan area 5400: 35mm film 24.61 x 36.69 mm / 5,232 x 7,800 pixels
Scan area 5400 II: 35mm film 25.0 x 37.25 mm / 5328x7920 pixels
CCD Scanner sensor width: 44mm wide mask (scanner element is 47mm wide)

Lens Magnification: 1.83x normal mount and 0.545x in reverse (44mm/24mm)
Working Distance: Good. 58mm @ 1.2x.
Designed lens coverage: 24 x 36mm, full frame (44mm circle near perfectly fits to 24x36mm frame)

In reversed position for APS-C sensor to match perfect 0.545x ratio we need to shoot film circle 51mm. This is larger than normal 43mm circle of 24x36mm frame. So if we attempt to project film frame to full size of the sensor - we got 0.67 ratio instead of perfect 0.545. It sort of works, but corners are slightly blurry and hard to focus.
So in theory this problem could be fixed if move lens and camera slightly further from film and sacrifice sensor edges. So we got less pixel resolution, but sharp corners of the film.

Here is chart that shows visually how many sensor edges should be lost to match perfect 0.545x ratio for APS-C sensor and 24x36 film. I'll check this in real tests soon...

 

[CameraQuest]

Has anyone used the 105mm f/5.6 APO El-Nikkor for scanning?

 

[Freakscene]

Has anyone used the 105mm f/5.6 APO El-Nikkor for scanning?

i tried it. With a good system for holding it perpendicular to the sensor and the film, it’s basically unstoppable. It’s definitely the best enlarging lens ever made. But I replaced it with a Zeiss S-Planar 50mm f/1,6, filter, and a 4360 A° light source, because I only had monochrome film, and it was easier to manage the film-to lens distance for that lens. In the end when I really needed resolution, I used a Zeiss stereomicroscope with a monochrome Axiocam and z-stacks to take a series of photos through the emulsion (but not the base) and let z-stacks compile the images in software. I also subsequently moved to monochrome digital for capture, so I haven’t needed to scan film for a while.

A lot of choosing this depends on the negative and what you want out of it. I have a lot of historical technical scientific work where resolution matters. For my own work, however, on Neopan 400 or Tri-X and shot hand held, it made the grain crisper, but for that work, concentrating on optimising tonality generally works better.

 

[shijan]

So i finally made some tests with FUJIFILM X-E4 and Dimage Scan 5400II lens.
0.67 vs 0.545x ratio makes no real difference at all.
De-focused corners are 100% due film curvature and too shallow aperture, f/3.7.
So lens is fully suitable for scanning 35mm film with APS-C sensor. It also can be used to scan medium and large format film with APS-C sensor. But DOF is tiny, so you need really flat film and well aligned stand setup. Perfectly it should be wet mount or glass holder.

 

[x-ray]

A few lenses not mentioned here. My Imacon 848 uses a Rodenstock Rodagon enlarging lens. I’m not sure if it’s the Apo or not. I’m guessing it’s not as the Apo to my understanding is optimized for mural printing.

A few others not mentioned with the first being fairly reasonable in price and a true Apo lens. The Goerz Apo Red Dot Artar was the standard of Apo process lenses for color separations and the prepress industry for decades. It came in a 4”(~100mm) FL in a barrel or a #0 shutter. It’s super flat field and is a true Apo lens. It came in long FL’s up to around 70”. Rodenstock also made their equivalent but not sure if it came in that short of FL but was called the Apo Ronar. Schneider also made the Repro Claron process lenses and G Claron as well.

Also at higher prices, the Zeiss Luminar and the Leitz Repro Summar are excellent choices.