NIKKOR - The Thousand and One Nights No.71

—A normal zoom lens redesigned to be more compact—

AI Zoom-Nikkor 35-105mm f/3.5-4.5S

I'll take a step back in time to the days of the F3 to introduce a child of the new NIKKOR lenses, the AI Zoom-Nikkor 35-105mm f/3.5-4.5S. A tremendous number of NIKKOR lenses, redesigned with the switch to auto indexing (AI), were released at this time, but just what sort of lens was this child? And what secrets lie behind the development of the lens? This tale delves into the secrets of the AI Zoom-Nikkor 35-105mm f/3.5-4.5S, which was born during a rough period of reorganization at Nikon.

by Haruo Sato

I. History of the AI Zoom-Nikkor 35-105mm f/3.5-4.5S

First, let's take a look at the history of the AI Zoom-Nikkor 35-105mm f/3.5-4.5S. The AI Zoom-Nikkor 35-105mm f/3.5-4.5S was released in 1983. It was developed as an interchangeable lens for renewed cameras like the F3HP and FM2. Early AI NIKKOR lenses used the same optics of existing new NIKKOR lens with a new barrel for updated exterior. However, when cameras like the F3 and EM were released in the 1980s, a large number of redesigned AI NIKKOR lenses offering improved specifications were developed. The AI Zoom-Nikkor 35-105mm f/3.5-4.5S was one of those lenses. This period was followed by the AF era. The well-received optics of the AI Zoom-Nikkor 35-105mm f/3.5-4.5S were adopted for the new AI AF Zoom-Nikkor 35-105mm f/3.5-4.5S released in 1986. During the transition period from founding of the AF era through its maturation, this lens also underwent improvements using the same optics with a new barrel. This ultimately resulted in the release of the AI AF Zoom-Nikkor 35-105mm f/3.5-4.5S (NEW) with a straight (push-pull) zoom mechanism and focusing ring in 1991. Optics were renewed and the lens achieved its final incarnation as the AI AF Zoom-Nikkor 35-105mm f/3.5-4.5D IF, released in 1994. The lens had a long run among NIKKOR lenses, selling for approximately 11 years.

II. Development history and the designer

Optics for the AI Zoom-Nikkor 35-105mm f/3.5-4.5S were designed by Tomowaki Takahashi, who worked in the First Research Department, Research Institute as the department was known at the time. Originally, Takahashi developed optical design software. He specialized in optimizing (automated design) software. A book titled "Lens design: From aberration coefficient to automated design" was written by Takahashi. At the time, Takahashi was focused on optical theory and software development. The Research Institute was in possession of powerful optical design software, developed and updated in-house, that had been used over the years, and Takahashi was responsible for developing some of this software. There was a very good reason for which Takahashi was chosen to design this photographic lens.

Once upon a time at Nippon Kogaku K. K. (now Nikon Corporation), optical designers and developers were scattered throughout various business units and departments. Naturally, each designer specialized in a particular type of product and was an expert in their chosen field. However, that does not really give them a broad or global point of view. That is because the types of optics in existence for products or fields other than their own are as numerous as are the stars in the sky. There are cases in which flexible and broad knowledge is required to look at obstacles we have never faced or overcome in a whole new light. This led to all of the company's designers being assigned to a single department. Thus, the organizational framework that the noted Nikon designer Zenji Wakimoto had described and long encouraged was finally achieved. This was especially beneficial for designers of photographic lenses that had long been separated into two departments, one involved in research and the other in business. These two departments were finally combined. However, this reorganization was not without some drama. Experts from the research and business fields had to compete to come up with the best optical design for a zoom lens for which they had yet to discover the optimal design. The winner of that competition would be appointed the leader of the design project. Design competitions are very common today, but at the time working practice was to have one designer for each product. The AI Zoom-Nikkor 35-105mm f/3.5-4.5S was the first project for which experts from the two fields were required to compete with one another to come up with the best design for a 35-105mm f/3.5-4.5 lens. They faced a number of constraints, including requirements regarding optical performance and size restrictions. A particularly difficult requirement was the 52-mm filter (attachment) size. As most are aware, the majority of interchangeable NIKKOR lenses from that time had a 52-mm filter (attachment) size. This was, in fact, a significant obstacle for this project. According to product planning, 35-105mm zoom lenses were positioned as normal zoom lenses for everyday use. It seemed only natural that the filter (attachment) size be 52 mm. Ultimately, the design proposed by Takahashi won the competition. While he went on to develop 35-135mm and 35-200mm lenses in addition to the 35-105mm lens, the AI Zoom-Nikkor 35-105mm f/3.5-4.5S was the first product that was released. Takahashi distinguished himself with development of these zoom lenses and earned the position of Manager in the new 1st Optical Section, Optical Designing Department. Progression from trial production to mass production went smoothly with the support of those around him. Optics originally designed and developed for the AI Zoom-Nikkor 35-105mm f/3.5-4.5S were used, modified, and manufactured right through the AF era, and the lens was very popular with a great number of users for quite a long time.

Now let's take a look at AI Zoom-Nikkor 35-105mm f/3.5-4.5S development history. Work on the design began around 1981 and was ready for trial production just as the cherry blossoms were blooming in the spring of 1982. Some corrections and improvements were needed, and a second round of trial production was initiated. At that time, multiple rounds of trial production was not uncommon, so this was not at all unusual. Mass production began in the summer of 1982, and the lens was finally released in April of 1983 just as spring was being heralded.

III. Lens construction and characteristics

Figure 1. Lens cross-section

First, take a look at the cross-section of the AI Zoom-Nikkor 35-105mm f/3.5-4.5S (Figure 1). Please forgive me if the following is quite technical. This lens uses the standard for high-power normal zoom lenses that is a four-group zoom structure consisting of positive-negative-positive-positive (convex-concave-convex-convex) structure with which all groups move. This lens was one of the first high-power normal zoom lenses that incorporated a number of advanced technologies. Take a look at lens construction. With the exception of the first (focusing) group, each group acted as both variator and compensator, and it was difficult to specify their individual roles. However, if roles must be specified, I suppose the second group, which contributed most to variable power, could be considered the main variator. We can assume that groups three and four act as both variator and compensator, and see that they combine to form the master lens group. We can also understand that the path along which groups three and four move produces a floating effect that compensates for fluctuations in curvature of field caused by changes in the interval that occur with zooming (as variable power changes). Therefore, with a quick glimpse this structure looks like a three-group positive-negative-positive (convex-concave-convex) zoom structure with a floating group. I wonder if you've noticed something. If we look back at predecessors of this normal zoom lens, we will certainly find Takashi Higuchi. It seems that the blood of the 43-86mm zoom lens flows within it. The DNA of Higuchi's zoom lenses can be found the world over.

So, what characteristics does this four-group positive-negative-positive-positive (convex-concave-convex-convex) zoom lens offer that other zoom lenses do not? It seems that the reins of power have now been handed over to the five-group positive-negative-positive-negative-positive (convex-concave-convex-concave-convex) structure that makes positioning of an image-stabilization group easy. Let's compare four-group zoom to five-group zoom.

In fact, the difference between four-group zoom and five-group zoom is whether the third positive-lead group, which serves as the master lens in a three-group positive-negative-positive (convex-concave-convex) structure, is divided into two positive (convex) elements or three positive-negative-positive (convex-concave-convex) element. As I noted before, the master lens group in a four-group zoom structure adjusts the space between two positive (convex) elements with a floating motion to correct aberration rather than to control variable power. An exaggerated description would be that it is actually a three-group zoom structure in terms of variable power. Therefore, the zoom group must move significantly to achieve a large zoom ratio, thus requiring significant changes in the total length of the lens. On the other hand, since the master group in a five-group zoom structure is comprised of a positive-negative-positive (convex-concave-convex) triplet, moving the middle negative (concave) element changes the variable power. Therefore, with a five-group structure each group moves less, the total length of the lens changes less, and a higher zoom ratio can be achieved than with a four-group structure.

To this point, it may sound as if a five-group structure is superior to a four-group structure. However, that is not always true. The problem is size. A four-group zoom is smallest at the maximum wide-angle position, and the interval between the third and fourth groups, as well as total thickness, can be relatively small and thin. As the triplet comprised of the third, fourth, and fifth groups are quite thick, five-group zoom lenses cannot be compact. Compared to a four-group zoom lens, five-group zoom lenses are generally longer at the wide-angle position and shorter at the telephoto position. For all of these reasons, the decision to adopt a four-group zoom structure for the 35-105mm zoom, which was then regarded as the next-generation 43-86mm zoom, resulted in a great success.

Now let's take a look at the aberration characteristics of the AI Zoom-Nikkor 35-105mm f/3.5-4.5S. We'll start at the 35-mm maximum wide-angle position. At first glance, aberrations generally appear beautiful. What is noteworthy here is that spherical aberration that can be observed to a certain extent in terms of amount has an exemplary form. There is little asymmetrical coma, and flare is restricted to off-axis portions of spherical aberration. There is little sagittal coma, and there is very little astigmatism for up to 90% of image height with meridional and sagittal deviations near the maximum angle of view. A good balance between correction of curvature field and the amount of spherical aberration generated has been achieved. In short, while some aberration remains, the remaining amount and balance can be regarded as exquisite. In addition, distortion is -4.6% at infinity, but drops to -1.4% at short distances. Next let's look at aberration at a mid-range focal length. Spherical aberration changes gradually resulting in over-correction at around the 70-mm focal length. Curvature of field follows suit for a significant decrease. There is also less coma, and excellent resolving power can be expected. Distortion changes to a positive value. One unfortunate characteristic is the change in spherical aberration to a positive value. This correction results in unattractive background blur or bokeh which can be resolved by stopping down the aperture by a half stop to a full stop. Finally, we'll look at aberration at the 105-mm maximum telephoto position. At the telephoto position, the most notable characteristic is the correction of spherical aberration. There is basically no spherical aberration up to about 70% of the frame from the center, but it increases gradually in the negative direction at peripheral portions. Moving out from the center of the frame, consistently negative curvature of field is exhibited to the extreme edges. This overall aberration compensation shows the lens' tendency to achieve pleasing, three-dimensional rendering characteristics. Distortion of roughly +2% can be considered normal for a zoom lens.

Now let's look at point-image formation. At the wide-angle position, point images are expressed to reflect aberration compensation tendencies. While the core of the point image is a bit large, it is well formed. Practically no sagittal coma flare is noticeable to an image height of around 70%, and the point has a nice shape that is close to round. As we might expect, sagittal coma flare does appear near the maximum angle of view, and point images look a little like birds with their wings spread. At mid-range focal lengths, point image cores are small and some flare remains. Off axis, the tendency toward outer coma gradually increases. At the telephoto position, flare is reduced significantly, but the tendency toward outer coma remains.

Now we'll look at MTF values. If we look at characteristics from infinity to shooting distances of several meters, we see that contrast at all focal lengths is excellent at both 10 and 30 lines/mm. 30 lines/mm values at the wide-angle position are relatively consistent to the edges of the frame, but plunge significantly at the extreme edges. I think this was probably the designing policy. I'm quite certain that Takahashi's objective here would have been consistent reproduction of contrast to as close to the edge of the frame as possible. The cost of this can be seen in the sagittal focal plane at the extreme edges of the frame. Contrast at 10 lines/mm drops a little at mid-range focal lengths. This is primarily the result of flare caused by elements of spherical aberration generated with over correction. On the other hand, we see high values with measurement at 30 lines/mm near the center of the frame. At the telephoto position, MTF values are extremely good throughout 70% of the frame from the center, with contrast that surpasses that of fixed focal length lenses at both 10 and 30 lines/mm. However, this does not mean that performance at the extreme edges of the frame is extremely bad. Contrast seems to drop gradually. I would say that rendering characteristics seem quite natural.

IV. Actual performance and sample images

Next let's look at results achieved with actual images of distant landscapes. Details regarding performance are noted for each aperture setting. Evaluations are subjective, and based on individual preferences. Please keep in mind that my opinions on sample images and evaluations below are for reference purposes only.

Maximum wide-angle 35-mm position

f/3.5 maximum aperture
Despite slight flare at the center of the frame, three-dimensional resolution is achieved. Some flare and color bleed is generated the closer to the edges of the frame we get. Good resolution is preserved nearly to the corners of the frame but resolution at the extreme edges drops drastically at the maximum angle of view. Image formation remains good to the corners of the frame, though, for pleasing results even with this sudden drop in resolution.

f/4 to f/5.6
Stopping down the aperture to f/4 reduces flare and increases sharpness and contrast, especially at the center of the frame. There is very little change in the corners of the frame. At f/5.6, sharpness is increased to the edges of the frame.

f/8 to f/11
Detail and sharpness increases even more at f/8. At f/11, excellent image quality is achieved with sharpness maintained throughout the entire frame except the most extreme edges of the frame.

f/16 to f/22
Even more consistent rendering is achieved throughout the entire frame, but resolution decreases. The effects of diffraction are visible and resolution suffers slightly.

Mid-range 50-mm position

f/3.8 maximum aperture
Rendering is similar to that achieved at the wide-angle position, but resolution seems to increase even further. There is a little flare at the center of the frame, possibly due to positive spherical aberration. However, resolution is still good. Chromatic aberration is reduced, nearly eliminating color bleed. Resolution does still fall in the corners of the frame. Blur characteristics (bokeh) appear harsh. Precaution may be required so that double-line blur is not noticeable.

f/5.6 to f/8
Stopping down the aperture to f/5.6 reduces flare and increases contrast. Details become very clear everywhere except the most extreme edges of the frame. At f/8, sharpness increases even further. Image quality tendencies remain the same.

f/11 to f/16
Consistent rendering is achieved throughout the entire frame. Contrast increases even further. Of all aperture settings, the best image quality is achieved at f/11. An aperture setting of f/11 is best for landscape photos. Performance in the corners of the frame is also improved. At f/16, resolution drops somewhat.

f/22 to f/23
Even more consistent rendering is achieved throughout the entire frame, but resolution decreases. The effects of diffraction are visible and resolution suffers slightly.

Maximum telephoto 105-mm position

f/4.5 maximum aperture
Some flare occurs from the center of the frame to the edges. Resolution is also a little poor due to the effects of coma and chromatic aberration. However, this soft rendering is consistent to the edges of the frame. This might make for good portraits.

f/5.6 to f/8
Resolution and sharpness are increased by stopping down the aperture a little. Consistent and pleasing results are achieved to the corners of the frame. For some reason, though, background blur is harsh.

f/11 to f/16
At f/11, flare is nearly completely eliminated for an even sharper feel. Consistent rendering is achieved throughout the entire frame. Of all aperture settings, the best image quality is achieved at f/11. An aperture setting of f/11 is best for landscape photos. At f/16, resolution drops somewhat.

f/22 to f/32
Even more consistent rendering is achieved throughout the entire frame, but resolution decreases. The effects of diffraction are visible and resolution suffers slightly.

If sharpness is the goal, the best results would likely be achieved at an aperture setting of f/8 to f/11 at all focal lengths (positions). For portraits, using maximum or near maximum aperture at telephoto position would produce most desired results.
Now let's confirm these rendering characteristics with some sample photos. These images have not been edited or enhanced so that you may judge the characteristics of this lens for yourself.

Sample 1
RAW image captured with Nikon D800E w/ AI Zoom-Nikkor 35-105mm f/3.5-4.5S (at 35-mm equivalent) at f/3.5 maximum aperture, 1/1000 s (-0.3 EV), ISO 400, with white balance set to Shade, D-Lighting set to Auto, Portrait Picture Control specified (May, 2018)

Sample 1 was captured at the wide-angle 35-mm position at the maximum aperture of f/3.5. If we look at the pattern in the wood panels, and the model's face and hair, we see that the image exhibits a sufficient degree of sharpness. There is also no clear drop in sharpness in the intermediate and peripheral portions of the frame. Consistent rendering is maintained throughout the entire frame. However, we do see a sudden and sharp breakdown in the image in the extreme corners. This breakdown is especially noticeable in the unfortunately harsh background bokeh. However, with little color bleed, the overall results are quite pleasing.

Sample 2
RAW image captured with Nikon D800E and AI Zoom-Nikkor 35-105mm f/3.5-4.5S (at 50-mm equivalent) at approximately f/4.0 (maximum aperture), 1/4000 s (-1.0 EV), ISO 400, with white balance set to Shade, D-Lighting set to Auto, Portrait Picture Control specified (May, 2018)

Sample 2 was captured at around 50 mm at the maximum aperture of approximately f/4. We still see a good level of sharpness, and background bokeh is probably a little harsh, but not more so than would be expected of a typical zoom lens. With little color bleed, the overall results are quite pleasing.

Sample 3
RAW image captured with Nikon D800E w/ AI Zoom-Nikkor 35-105mm f/3.5-4.5S (at 70-mm equivalent) at f/4.5 maximum aperture, 1/500 s (-1.0 EV), ISO 400, with white balance set to Shade (w/ fine tuning of color temperature), D-Lighting set to Auto, Portrait Picture Control specified (May, 2018)

Sample 3 was captured at the near the maximum telephoto position of 105 mm at the maximum aperture of f/4.5. We see a good level of sharpness at this position, and the gentle bokeh is pleasing. Coloring is good, and there is little sense of chromatic aberration.

Sample 4
RAW image captured with Nikon Z 7 w/ AI Zoom-Nikkor 35-105mm f/3.5-4.5S (at 50-mm equivalent) at f/4 maximum aperture, 1/4000 s, ISO 100, with white balance and D-Lighting set to Auto, Landscape Picture Control specified (July, 2019)

Sample 4 was captured at around 50 mm at the maximum aperture of f/4. It exhibits the necessary sharpness, and there is very little peripheral illumination falloff (vignetting). Coloring is good, and chromatic aberration is well suppressed.

Sample 5
RAW image captured with Nikon Z 7 w/ AI Zoom-Nikkor 35-105mm f/3.5-4.5S (at 45-mm equivalent) at f/3.8 maximum aperture, 1/4000 s, ISO 100, with white balance and D-Lighting set to Auto, Landscape Picture Control specified (July, 2019)

Sample 5 was captured at around the wide-angle 45 mm at the maximum aperture of f/3.8. It exhibits the necessary sharpness, and there is very little peripheral illumination falloff (vignetting). Some ghost is visible, but for a lens of this time, performance in this regard is excellent. Depending upon the way it is expressed, ghost can be either the poison or the remedy. There is also very little flare, and colors are good.

Sample 6
RAW image captured with Nikon Z 7 w/ AI Zoom-Nikkor 35-105mm f/3.5-4.5S (at 50-mm equivalent) at f/4 maximum aperture, 1/4000 s, ISO 100, with white balance and D-Lighting set to Auto, Landscape Picture Control specified (July, 2019)

Sample 6 was captured at around 50 mm at the maximum aperture of f/4. We can see that the image is sufficiently sharp, and blur characteristics for the background bokeh are pleasing. There is also very little flare, and chromatic aberration has little effect on colors.

Sample 7
RAW image captured with Nikon Z 7 w/ AI Zoom-Nikkor 35-105mm f/3.5-4.5S (at 50-mm equivalent) at f/4 maximum aperture, 1/320 s, ISO 100, with white balance and D-Lighting set to Auto, Landscape Picture Control specified (July, 2019)

Sample 7 was captured at around 50 mm at the maximum aperture of f/4 so that we might check blur characteristics. I intentionally framed the photo so that the difficult lattice pattern would be in the background. However, there is no breakdown in background bokeh, and the results are tolerable. Very little axial chromatic aberration also makes for pleasing results.

V. Tomowaki Takahashi

Tomowaki Takahashi was my immediate supervisor when I first began working at Nippon Kogaku K. K. (now Nikon Corporation). He was a true gentleman with a kind and warm nature. He laughed a lot, and I don't remember ever seeing him angry. While I haven't received much guidance from him regarding optical design, he always listened to my outrageous ideas with an open mind. Despite the fact that I often wanted to take my own way, he used to watch out for me like a father does for his child. When I proposed the 24-120mm, he helped me to negotiate with management for approval. He was also my manager and signed off on drawings I presented for my first lens, a 24-50mm zoom. He was definitely a superior in whom I had great trust and faith. As I noted before, Takahashi's specialty was the development of optical software. Naturally, he was also an expert in geometric and wave optics. His knowledge of automated design was especially deep, and, at that time, he developed cutting edge automated design (self-correcting) software. He made full use of the automated design software that he had himself developed in designing the AI Zoom-Nikkor 35-105mm f/3.5-4.5S. However, at that time, automated design was not truly or completely automated. It did not replace human designers. These days, artificial intelligence (AI) has greatly expanded and improved on development possibilities. It is likely that in the not-too-distant future, designers will be able to design a new lens with the press of a button. That was definitely not the case 30 years ago, though. Human designers were completely responsible for specifying initial design values, targets, and weight, calculating changes in RDN, and manipulating factors to forcibly achieve those values and targets. Takahashi was a researcher during that period of development transition. I would like to share with you a memory that Takahashi once shared with me.

When he was young, Takahashi married and was living in corporate housing in Tokyo. However, he didn't like the floorplan and took it upon himself to forcibly remove one of the house's pillars and a wall. At that time, corporate housing consisted of wooden single-story houses with a small yard. They were typical duplex-style Japanese homes of the Showa era. It seems that Takahashi wanted to combine a 4.5-tatami room and a 6-tatami room into one large room. I heard this story when I was thinking of moving into corporate housing. I asked, "But wasn't that bad for the structure and integrity of the house?" Takahashi replied, "No. It's still standing," (meaning that it hadn't fallen down yet). It seems that was in fact not a problem. It was, however, an incredibly bold move. I'm sure he probably had permission to do what he did, but most people don't usually go so far. After hearing the story, I decided I would definitely NOT move into that house (building). I also believed that it was this way of thinking and going about things that empowered Takahashi to promote his works, and I admired him greatly for it.