Scanner-Nikkor

[ray_parkhurst]

I read on one of Enrico Savazzi's pages (in Italian...translated) regarding the Printing-Nikkors that there was a Scanner-Nikkor lens that may have a similar construction. It seems Nikon made a new lens design using ED glass for their CoolScan IV and CoolScan V film scanners, and gave it the name Scanner-Nikkor. A little research on the Nikon site led me to believe that the lens is 7 elements, not quite as complex as the new versions of Printing-Nikkor which are 14 elements. Given the compact size of then CoolScan film scanners, I would presume that the focal length of the Scanner-Nikkor lens is quite short. The sensor may be small, such that the magnification is much less than 1:1. Does anyone have any experience with one of these lenses? If I am right, then there does not seem to be any hope for them in photomacrography, but it would be good to hear any real input on this, since all I have is speculation based on rumor.

[dickb]

I read on one of Enrico Savazzi's pages (in Italian...translated) regarding the Printing-Nikkors that there was a Scanner-Nikkor lens that may have a similar construction. It seems Nikon made a new lens design using ED glass for their CoolScan IV and CoolScan V film scanners, and gave it the name Scanner-Nikkor.

Are you perhaps referring to Marco Cavina instead?

A little research on the Nikon site led me to believe that the lens is 7 elements, not quite as complex as the new versions of Printing-Nikkor which are 14 elements.

The Marco Cavina article refers to prototype lenses for two scanners, the Nikon LS4000 (for 35mm size) and the Nikon LS8000 (for 120 film).

The LS4000 lens is described as a 7 elements design and the LS8000 lens as 14 elements.

Given the compact size of then CoolScan film scanners, I would presume that the focal length of the Scanner-Nikkor lens is quite short. The sensor may be small, such that the magnification is much less than 1:1. Does anyone have any experience with one of these lenses? If I am right, then there does not seem to be any hope for them in photomacrography, but it would be good to hear any real input on this, since all I have is speculation based on rumor.

No I haven't, unfortunately. Annoyingly, I vaguely remember reading about separate scanning-nikkor lenses years ago and I can't retrieve the information..

[ray_parkhurst]

Are you perhaps referring to Marco Cavina instead?

Indeed it was. Apologies to both Marco and Enrico for the confusion.

This was the link I was looking at:

https://www.nital.it/nikkor-historical/el-nikkor2.php

[dickb]

Given the compact size of then CoolScan film scanners, I would presume that the focal length of the Scanner-Nikkor lens is quite short. The sensor may be small, such that the magnification is much less than 1:1.

I came across a few images of disassembled LS8000s. With a few quick and dirty size comparisons I estimate the exposed end of the optic to be between 40 and 50mm in diameter. That is the end towards the front surface mirror and the film to be scanned. I don't know the dimensions of the Nikon imaging sensor, do you? The distance between the exposed optic and the surface to be scanned is at least 8 cm, maybe up to 10 or 12. The maximum width of the image to be scanned is 58mm. If that gives you enough information about the lenses' focal length, I'd like to hear it.

[ray_parkhurst]

Yeah, I don't think the FL can be deduced without some more info, such as the sensor size.

I'll keep looking around for more info, and maybe a broken scanner will show up for a decent price.

[jcb]

Got one in front of me that is said to come from a Nikon Coolscan 8000 scanner :

• weight : 395 g
  length : 84 mm
  min diameter : 49 mm
  max diameter : 55 mm
  focal length (quick and dirty estimate) : +/- 80 mm
  back of lens to image distance (free space behind the lens at infinity) 30/40 mm (quick and dirty again)
  fixed aperture (very quick and very dirty) : probably f/2.8 (pupil diameter/80mm)
  pupils magnification +/- 1.0
  no filter threads
  no mounting threads

Edit :
See message below. Taking more time to make better measurements the lens seems to be a 100mm f/3.5

Nikon (http://imaging.nikon.com/lineup/scanner ... n_8000_ed/)says :
Exclusive SCANNER NIKKOR ED high-resolution/high-performance lens (14 elements in 6 groups including 6 ED glass elements) for reduced color aberration and minimized image distortion

[Lou Jost]

These film scanners have intrigued me for many years. I have a Coolscan 5000 that I use almost daily now and it is sharp. The 8000/9000 versions for 35mm and medium format have the same rated resolution, but twice as many glass elements, as noted above--- 14 elements including 6 ED lenses, just like the Printing Nikkor, but with far greater coverage. That's got to be a GREAT lens in there.

After some personal correspondence with jcb (thank you for that info, jcb) I went ahead and bought a not-too-expensive Coolscan 8000 that came up on eBay as "Not working- For parts" . It is waiting for me at my family's address in the US and I will pick it up in a few weeks. Is a giant medium format Printing Nikkor inside? Wouldn't that be "Cool"? Broken ones do show up now and then, and might be a great source of exotic glass. I'll definitely report on the results when I test it. It is unthreaded so mounting it might be tricky.

[jcb]

It probably is neither a 95 nor a 105 mm printing nikkor even though the optical formula and design aim seem promising.
I have one and tried to measure the focal length by focusing as near as I could to 1:1 and then at 2:1 (infinity focus is not possible with a mirror DSLR). I get 98.3 mm and a suspicion that I am a little below reality due to the width of the lines on my target. Let's say 100mm (there are tolerances and a 100mm rarely is 100.00mm!)
Aperture (100 divided by the diameter of the front pupil) in this case would be 3.5.
From the pictures I have seen of the internals of the scanner, I would guess it is optimized for something near 1:1. The ultimate test would be to measure the length of the array on the sensor but I do not have one. The exact magnification ratio would be this length / 63.5mm (the maximum width that the scanner accepts). From memory, there seems to be a small difference in pupils size so that it may not be exactly symmetrical.
I am still finishing to mount it to test as a slide duplication lens so there are no results yet.

Edit : corrected the maximum width that can be scanned from 62 to 63.5mm after checking the specs on Nikon's web site.

[Lou Jost]

I have the whole device sitting in the US waiting for me so I can make those measurements and discover what m it is designed for. I will know in about three weeks and will report here.

I also hope to have a Printing Nikkor by then so can compare them directly.

[ray_parkhurst]

I'm looking forward to the results of these tests, as these seem very intriguing. I've looked on and off for a 5000 or 8000 cheap on eBay but never found one that still had its lens for a decent price. My presumption was that the focal length would be much shorter, given the compact size of the systems. I reasoned that there would be no need in such a scanner for a large sensor, and thus a short focal length could reduce the required magnification. Will be interesting to know the real situation in these.

[Lou Jost]

The medium-format scanners should have largish sensors, I would think. The lens must have a huge field of view; I imagine it has to cover the whole width of medium-format film. To do that without perspective distortion should require a longish lens.

[Chris S.]

The medium-format scanners should have largish sensors, I would think. The lens must have a huge field of view; I imagine it has to cover the whole width of medium-format film. To do that without perspective distortion should require a longish lens.

I don't understand this expectation, since these machines are scanners. They probably do not capture the image all at once, but sweep the lens along two axes of translation and assemble something like a panorama. If this is the case, a scanner for medium format media could use exactly the same optics as one for 35mm. It would simply require longer linear actuators, and probably a bit more buffer and computational horsepower to handle the increased data.

If this is the case, I would not be optimistic about the optics having a large high-quality image circle. One plausible design approach would involve a lens whose high quality image circle is both very good and rather small, paired with a smallish sensor capturing just this region. This would reduce the cost of lens and sensor, which were likely the most expensive components of these machines.

By comparison, Printing Nikkors were designed to image a full frame in a single shot--a very different job.

This said, the tests will be interesting. I would be delighted to be wrong about this.

--Chris S.

edited typos

[Lou Jost]

I assumed the scan only goes in one dimension. The lens looks like it is fixed in place, based on the photos of the innards on the internet.This would explain why the lens on the 8000 and the 9000 is so much more complex (twice the number of elements) as the LS 5000 scanner for 35mm film, which was released at the same time as the LS 8000. But I tend to be an eternal optimist and might be completely wrong. I'll find out in two weeks. But probably jcb already knows the image size.

Edit--My guess is that the lens is designed for about 2:1 reduction, so can be used in reverse at 2x. Something like the 95mm PN.

[jcb]

@Chris S.
The pictures of the Coolscan LS8000 disassembled show that the slide/negative moves only along the front-back axis of the scanner.
The lens moves but only as some kind of auto focusing device.
Nikon gives the same resolution figure of 4000dpi for every format scanned. This should confirm that there is no slide/negative movement along the left-right direction and that all scans are optically 120 film format scans with 35mm film scans reading only the central part of the sensor.
Nikon specs say that the maximum width that can be scanned is 63.5mm. We can expect that the good image circle of the lens, slide side is at least 63.5mm.
The sensor clearly is a line sensor. So the length of the sensor divided by the maximum film width that can be scanned (63.5mm) should give a good estimation of the reproduction ratio for which the lens was optimized.

@Lou Jost
The information regarding the sensor size will be most welcome.

[Lou Jost]

I should note that in these scanners, the film does not move at all. There is a mirror in the system. Maybe that moves?

jcb, I'll let you know the sensor size as soon as I dissect the device, around May 3.