Compensating MIcro Nikkor 55mm f/3.5 revisited

[Bernard Delley]

This old short focal lens is occasionally useful for 35mm slide  digitizing.
I have the AI version and recently got a compensating sample too.

 I use the Mac Book Retina screen as my high precision micro-engraved luminescent target for a 1:1  repro size test. The tricolor pixels of the target are at a 115 micro-meter pitch. Each color patch is less than 38 microns wide.

Short summary of my test: I see NO significant difference between the compensating and the non-comp version .

Here are some details: Normal alignment to the screen was done using the mirror image of the macro rail.
For the compensating lens I chose nominal f/8. Because of the compensation the aperture is opened about one stop. Yet since its 1:1 repro ratio the image distance is 2*f makes the actual aperture ratio = 1/11.

This is the result: OOC jpg large medium, picture control neutral.
attachment 1: lower left quadrant of image at 50%  with compensating 55mm lens #271483

The non-compensating lens is set at nominal f/5.6, resulting in an actual aperture ratio of 1/11 at the 1:1 repro ratio. the result is: (please ignore the test structure to the right, it helps to make absolutely sure that I don not accidentally confuse the test shots)
attachment 2: lower left quadrant of image at 50% with non-compensating 55mm lens #1010507

Both lenses appear to have the same longitudinal chromatic aberrations requiring judicious manual LV max mag focus adjustment for about equal sharpness of the three color patches.  There is lateral chromatic aberration clearly skewing the tricolor in the lower right corner.  To the left this works out that the patches partially overlap and the image desaturates, to the right (not shown) the arrangement of the color patches  works out to the worse. This is not a lens decentering!

I really see no significant difference in image quality between the two lens versions.  If the compensating version were optimized for a closer image distance than the later version (Bjørn Rørslett wrote so), this should show up in less aberrations at 1:1 repro ratio.   Or did Nikon change the optimization before abandoning the compensating mechanics - my lens is from the latest compensating series (1968? according to photosynthesis) .

[Bjørn Rørslett]

Neither type goes to 1:1 on its own, unless you incidentally used the early '60s 5.5 cm f/3.5 Micro-Nikkor. So how did you bring the magnification to 1:1 ? Were exposure times identical for your two comparison shots?

Also pay attention to the fact that Nikon optimised these Micro-Nikkors for 1:10 not 1:1.

[Bernard Delley]

sorry, I forgot to mention that I used the 27.5mm M extension tube. As the Micro Nikkor was originally designed for microfiche, optimisation for 1:10 repro ratio is certainly reasonable. 1:1 is far away from optimum, and with the tube we are stretching it.  I actually got the compensating version because your old site seems to indicate that the compensation version was stronger at close range than the later 55mm f/3.5 P-C and AI versions. I do not see this in my comparison. Can you explain how you arrived at that statement? 

[Bjørn Rørslett]

Actual observation based upon images from the various Micro-Nikkor 55 lenses at my disposal (4 or 5 over the years)

[Intrepid]

They were (5.5cm, 55mm) all optimized for 1:10 AFAIK (and also think that is what Bjørn says/said).  The compensating part is misleading.  It just changes the aperture readout to the camera to compensate for the magnification (this relates to the metering part- quite a neat accomplishment for that era).

The Olympus OM 50/3.5 Macro, for example, OTOH, involved a floating element in the design (in old Nikon speak "CRC") that compensated for aberrations at close ranges much like the latest  Micronikkor 60/2.8.

[Roland Vink]

I have manuals for compensating and AI 55/3.5 micro lenses. Both state the lens is optimized for 1:10 magnification with performance remaining high through the entire focus range. Optical drawings from old Dealer Sales Manuals do show a very minor change in dimensions between the earlier compensating and the later non-compensating versions. There are many reports that the older version is better at close range but less good at distance, while the later versions sacrifice some macro optimization for better all-round performance. If so, I don't see how both can be optimized at 1:10 unless it's an error and nobody told the copywriter of the manuals...

Since you shot at f/11, diffraction might be masking some differences between the lenses, or it might just happen that you have a "bad" compensating or a "good" non-compensating lens, making them appear equal. It would be interesting to try the same test at medium and far distances.

[Asle F]

The different optical drawings are to be find at http://www.mir.com.my/rb/photography/companies/nikon/nikkoresources/6070nikkor/micro/index.htm

I have both a compensation type and a later P.C. I haven't done any comparing of them, I use the compensating lens when using manual exposure with or without flash, and the P.C. when using TTL-metering and aperture priority automatic exposure. For infinity I use the later 55mm/2.8

[Bernard Delley]

Roland,
The f/11 does not mask the lateral color aberration. The 50% reduced sample just lets you see this when viewed full size. Of course 1:1 is a challenge for this lens. However, I find remarkably little (none! ) sample variation between these two lenses as shown with this lateral color. For me that hints at the same optical construction.  -- The link shows two optical drawings for the 55mm f/3.5, its the same 5lenses in 4 groups construction with the same measures shown, the same that I also found in other places.
When I have the time I will post further such screen test. For sure, at 13.8 reduction, very near the mentionend 1:10, the AI 55mm f/3.5 shows superb performance.  infinity is another matter, I would have to use stars for testing. I noticed that with these manual lenses that the infinity stop may not be that precise. It is a spec leading to sample variation with little connection to the glass.

BTW I would be so curious to see such a computer screen test for the Voigtlander APO Lantaher 125mm f/2.5 at 1:1 repro ratio.
Can somebody here do this? ( I do not have one)
When I get around, I will finalize a similar test at 1:1  1:1.5   1:2  1:3 of Micro Nikkor 55mm f/3.5 versus AI-S 60mm f/2.8. I will try to post 100% samples (jpg basic).

[Bjørn Rørslett]

I can already tell you, without any testing, that the APO 125 is not near its peak performance at 1:1. In fact, its otherwise superb chromatic correction starts slowly to break down and one can observe occasional blue fringing.

Testing a lens against a computer monitor is problematic because you cannot conclusively decide whether colour artefacts are caused by the lens or the target. You should find something better suited.

[Erik Lund]

...Optical drawings from old Dealer Sales Manuals do show a very minor change in dimensions between the earlier compensating and the later non-compensating versions....

I have measured that indeed the dimensions of the optics are different.

I would use another target than a monitor, to my eyes the two images look very different...

[Bernard Delley]

The outer dimensions of the lenses are obviously different  between  compensation and non-compensating version. I would be impressed if you did succeed measuring some of the glass dimensions, like the curvature of front and back lens or the overall "thickness" of  the glass.

Fortunately I am somewhere out in nature looking for images, no testing for some time.

While doing the computer screen tests, the behavior of the artifacts looks plausible as caused by the optics . Also 1:1 looks very good on the AF-S 60mm f/2.8, except for indications of longitudinal CA -- not lateral.

[Erik Lund]

The outer dimensions of the lenses are obviously different  between  compensation and non-compensating version. I would be impressed if you did succeed measuring some of the glass dimensions, like the curvature of front and back lens or the overall "thickness" of  the glass.

Fortunately I am somewhere out in nature looking for images, no testing for some time.

While doing the computer screen tests, the behavior of the artifacts looks plausible as caused by the optics . Also 1:1 looks very good on the AF-S 60mm f/2.8, except for indications of longitudinal CA -- not lateral.

I measured the diameter of the lens element.

[Asle F]

The link shows two optical drawings for the 55mm f/3.5, its the same 5lenses in 4 groups construction with the same measures shown,

I see different measures in the two constructions. The distance between the rear lens and F' is 37.9mm vs 38.3mm. The total lengths from the rear lens to the front lens is 27.8mm vs 27.9mm. And the distance between the front lens and F is 37.9mm vs 37.4mm

[Erik Lund]

BTW it's a simple procedure to convert a Compensating Micro-Nikkor to a non compensating aperture. It's just a matter of separating/splitting the compensating aperture guide, reverse the parts and put them back in again.

This was actually the first lens that I chipped for my F4e with a CPU  from an 35-70mm AF lens, shorted at 55mm and f /3.5 so the lens data was correct, on Nikon F5 it also enabled full matrix metering long time ago

[Bjørn Rørslett]

Nikon were continuously tweaking their Micro-Nikkor designs. The earliest 5.5 cm f/3.5 had very steeply curved rear element, and if memory serves a 5/3 compared to the later 5/4 design, to name one example. This forerunner to the more modern Micro-Nikkors performed poorly reverse-mounted unlike the two later incarnations.

although these lenses do well on digital cameras, one should not forget the optics were developed and optimised in the film era, thus some aspects of the optical performance had low priority and assumed not important with film.