Flange diameter etc

The new Nikon Z makes me wonder: How come Leica is able to retain their legacy mount and lenses and make them work with their digital cameras? I am aware of sensors with microlenses etc, and I'm sure there is some image engine manipulation but is there any other secrets to it?

Looking at the new "full frame mirrorless" out there, it looks like what Leica is achieving is unique: a digital M body has the same thickness as a film M except for a small additional thickness around the flange.

Meanwhile Nikon is introducing this considerably larger mount. I doubt that Leica has some secret technology that Sony or Nikon don't have access to.

This is rather puzzling to me. So what is it?

Thanks

p.s. maybe sensor needs more light for AF? and also I assume the ability to introduce very fast lenses?

rondom wrote:

The new Nikon Z makes me wonder: How come Leica is able to retain their legacy mount and lenses and make them work with their digital cameras? I am aware of sensors with microlenses etc, and I'm sure there is some image engine manipulation but is there any other secrets to it?

Looking at the new "full frame mirrorless" out there, it looks like what Leica is achieving is unique: a digital M body has the same thickness as a film M except for a small additional thickness around the flange.

Meanwhile Nikon is introducing this considerably larger mount. I doubt that Leica has some secret technology that Sony or Nikon don't have access to.

This is rather puzzling to me. So what is it?

Thanks

p.s. maybe sensor needs more light for AF? and also I assume the ability to introduce very fast lenses?

Hi,

Digital systems don't need a large flange diameter. But it could be that a larger flange diameter could be advantageous in some cases.

Leica does have some issues with older lens designs, but that is not related to the flange diameter.

The problems are:

• Cover glass
• Beam angle
• Crosstalk

Digital sensor normally have pretty thick cover glass. Two mm is fairly typical, but 4/3 has around 4 mm.

Optical glass in front of the sensor causes astigmatism, because light beams with different angles will defract differently on trough the optical glass.

So, Leica's initial design used extremely thin cover glass, 0.3 mm on the M8 if I recall it right. But, that left a problem with IR-contention.

Later models increased cover glass thickness to 0.8 mm.

The cover glass issue is mostly affecting older lenses. In newer designs the cover glass would be taken into account in the optical design.

The large beam angle on many Leica lenses causes two other problems. Crosstalk, where part of the light passing trough the microlens in front of the sensor enters the neighboring pixels. That causes contaminated color.

Also, large beam angles cause many of the photons entering the sensor to absorbed outside the "well" corresponding to the photodiode of the pixel.

To reduce that, Leica designed their own sensor, in cooperation with CMOSIS. They used a design with shallower pixels combined with a different microlens design.

Not least, bar coding was added to the lens and the cameras also estimate aperture. That information is used to enable some software correction of the image.

So, Leica needed to address quite a few issues with their film era lens designs.

Just to mention, it is quite possible replace cover glass on say a Sony A7rII with cover glass to M10 specification and that will improve lens performance with Leica lenses but it has been proven that the modification has a negative effect with some Sony lenses.

It seems that 2 mm cover glass is quite standard, and it is no problem as long as the lenses are calculated for it.

With DSLRs it was less of a problem, the DSLRs have a mirror box, moving the lens away from the sensor and that yields lower beam angles, in general.

Having a short flange distance allows to use more symmetric lens designs, which are optically simpler as symmetri limits many aberrations. But, modern designs tend to be more assymetric, needing more lenses but also offering more air to glass surfaces to control aberrations. Modern era standard lenses like Otus 50/1.4 and Sigma 50/1.4 are inverted telephoto designs with a great amount of complexity, while older 50/1.4 lenses usually were pretty symmetrical double gauss designs, with a field flattener added.

Best regards

Erik

rondom wrote:

The new Nikon Z makes me wonder: How come Leica is able to retain their legacy mount and lenses and make them work with their digital cameras? I am aware of sensors with microlenses etc, and I'm sure there is some image engine manipulation but is there any other secrets to it?

Looking at the new "full frame mirrorless" out there, it looks like what Leica is achieving is unique: a digital M body has the same thickness as a film M except for a small additional thickness around the flange.

Meanwhile Nikon is introducing this considerably larger mount. I doubt that Leica has some secret technology that Sony or Nikon don't have access to.

This is rather puzzling to me. So what is it?

Thanks

p.s. maybe sensor needs more light for AF? and also I assume the ability to introduce very fast lenses?

Below is a diagram that might help explain this

Looking at the top diagram, the black line to the right represents the frame diagonal. The two L shaped lines are the lens flange, for the Nikon Z at a diameter of 55mm, 16mm from the image plane. The two dotted black lines from the centre of the frame are the light cone from an f/0.95 lens to the centre of the frame. The lens' exit pupil could be anywhere along the lens' axis, and wherever it is, must be large enough to project this light cone (that's what makes it an f/0.95 lens). I've placed exit pupils at 28mm (green), 50mm (red) and 100mm (blue). You can see that the 28mm one can easily project to the corner of the frame, as can the 50mm one. The 100mm one just makes it (in fact, so close that I'm wondering whether an f/0.95 lens with a 100mm distant exit pupil was a design goal).

The lower diagram represents the Leica mount, with a 39mm clear mount at 27.8mm. Here you see that the 28mm exit pupil can easily project to the corner of the frame but that the 50mm and 100mm ones can't. We can fit an f/0.95 lens to the Leica, so long as it has an exit pupil close to the image plane.

Digital image sensors with microlenses work better with distant exit pupils, and most 'designed for digital' lenses have an exit pupil in the range of 100mm or so. Thus the new Nikon mount can take a 'designed for digital' f/0.95 lens, the Leica mount can't. The Leica lens will work with a digital camera, but there is liable to be a lot of vignetting in the corners, because the microlenses can't work well with the light as oblique at the corners as the close exit pupil makes it.

rondom wrote:

The new Nikon Z makes me wonder: How come Leica is able to retain their legacy mount and lenses and make them work with their digital cameras? I am aware of sensors with microlenses etc, and I'm sure there is some image engine manipulation but is there any other secrets to it?

Looking at the new "full frame mirrorless" out there, it looks like what Leica is achieving is unique: a digital M body has the same thickness as a film M except for a small additional thickness around the flange.

Hi Leica had to design their sensor specifically for their m lenses. That is why many Leica lenses only perform well on m cameras. Nikon has to ensure that all existing f mount lenses perform well ... it is different problem to solve.

it is also probably the case that the initial lenses have high optical performance as the goal rather than size. Perhaps there will be smaller lenses with a different priority in the future.

See Leica SL lenses for example. They are huge, right? There is no magic, only trade-offs

bobn2 wrote:

Thanks for this, Bob. I knew that exit pupil size and position was key to understanding what Nikon was aiming for with the geometry they chose, but I hadn't seen it diagrammed. Your drawing was a nice help to me.

I noticed in the 50mm S-Line block diagram that Nikon has placed a fairly large (negative? @ 52mm in diameter) element almost as close to the sensor plane as physically possible. (Maybe 14mm or so?) And a fairly large group just in front of it.

Any comments on what this affords them, if anything? I think I know that entrance pupil size and position are heavily influenced by the frontmost element, so I'm guessing the rearmost element heavily influences the size and position of the exit pupil.

Nikon has made a special point about the aberration control they can achieve by getting elements closer to the sensor. (They talked about working very hard to shave millimeters off the flange distance by redesigning and minimizing interposing structures like the shutter mechanism.) Any thoughts on whether or how some of these factors might be reflected in this block diagram? The rear structure of this lens must be unique in 35mm photography right? The EF mount has the width but not the short flange distance, and none of the short flange distance mounts have the width. Maybe something similar has been done for M4/3rds?

peripheralfocus wrote:

bobn2 wrote:

Thanks for this, Bob. I knew that exit pupil size and position was key to understanding what Nikon was aiming for with the geometry they chose, but I hadn't seen it diagrammed. Your drawing was a nice help to me.

I noticed in the 50mm S-Line block diagram that Nikon has placed a fairly large (negative? @ 52mm in diameter) element almost as close to the sensor plane as physically possible. (Maybe 14mm or so?) And a fairly large group just in front of it.

Any comments on what this affords them, if anything? I think I know that entrance pupil size and position are heavily influenced by the frontmost element, so I'm guessing the rearmost element heavily influences the size and position of the exit pupil.

Nikon has made a special point about the aberration control they can achieve by getting elements closer to the sensor. (They talked about working very hard to shave millimeters off the flange distance by redesigning and minimizing interposing structures like the shutter mechanism.) Any thoughts on whether or how some of these factors might be reflected in this block diagram?

Thanks for that, I hadn't seen it before. The core of the lens is a symmetrical double Gauss (though with aspherical and ED elements. At the front, you have what's effectively an afocal converter, which will give the lens a shorter focal length. The back end I think has two functions, one it's a negative lens, so it acts as a teleconverter, second, a negative element close to the sensor is often used as a 'field flattener'.

So in summary, there's a normal prime with a TC behind it, which lengthens the FL and moves the exit pupil away from the image plane (as well as increasing the f-number) and then a wide converter at the front, which brings it back to 50mm FL and also helps in the optical correction of the lens.

bobn2 wrote:

peripheralfocus wrote:

bobn2 wrote:

Thanks for this, Bob. I knew that exit pupil size and position was key to understanding what Nikon was aiming for with the geometry they chose, but I hadn't seen it diagrammed. Your drawing was a nice help to me.

I noticed in the 50mm S-Line block diagram that Nikon has placed a fairly large (negative? @ 52mm in diameter) element almost as close to the sensor plane as physically possible. (Maybe 14mm or so?) And a fairly large group just in front of it.

Any comments on what this affords them, if anything? I think I know that entrance pupil size and position are heavily influenced by the frontmost element, so I'm guessing the rearmost element heavily influences the size and position of the exit pupil.

Nikon has made a special point about the aberration control they can achieve by getting elements closer to the sensor. (They talked about working very hard to shave millimeters off the flange distance by redesigning and minimizing interposing structures like the shutter mechanism.) Any thoughts on whether or how some of these factors might be reflected in this block diagram?

Thanks for that, I hadn't seen it before. The core of the lens is a symmetrical double Gauss (though with aspherical and ED elements. At the front, you have what's effectively an afocal converter, which will give the lens a shorter focal length. The back end I think has two functions, one it's a negative lens, so it acts as a teleconverter, second, a negative element close to the sensor is often used as a 'field flattener'.

So in summary, there's a normal prime with a TC behind it, which lengthens the FL and moves the exit pupil away from the image plane (as well as increasing the f-number) and then a wide converter at the front, which brings it back to 50mm FL and also helps in the optical correction of the lens.

Hi Bob,

Very nice analysis! Thank you!

Best regards

Erik

peripheralfocus wrote:

bobn2 wrote:

Thanks for this, Bob. I knew that exit pupil size and position was key to understanding what Nikon was aiming for with the geometry they chose, but I hadn't seen it diagrammed. Your drawing was a nice help to me.

I noticed in the 50mm S-Line block diagram that Nikon has placed a fairly large (negative? @ 52mm in diameter) element almost as close to the sensor plane as physically possible. (Maybe 14mm or so?) And a fairly large group just in front of it.

Any comments on what this affords them, if anything? I think I know that entrance pupil size and position are heavily influenced by the frontmost element, so I'm guessing the rearmost element heavily influences the size and position of the exit pupil.

Nikon has made a special point about the aberration control they can achieve by getting elements closer to the sensor. (They talked about working very hard to shave millimeters off the flange distance by redesigning and minimizing interposing structures like the shutter mechanism.) Any thoughts on whether or how some of these factors might be reflected in this block diagram? The rear structure of this lens must be unique in 35mm photography right? The EF mount has the width but not the short flange distance, and none of the short flange distance mounts have the width. Maybe something similar has been done for M4/3rds?

This is not a good lens design...

Element 2 is a paper thin meniscus with R1 ~= R2.  This will be extremely difficult to meet good tolerances on wedge of the two surfaces on.  It is also cemented to a strong biconcave, a process that will be made more difficult by its thinness.

Element 3 looks inside out.

Element 8 has a nasty thin airspace that strongly varies in thickness with radial position.  This will be extremely sensitive to despace.

The airspace between elements 8 and 9 is almost constant, this is already sensitive.  If there is an aspheric surface on the back of element 8, even worse.  Elements 10,11 are both thin and cemented.  This is not very good.

Finally, the rearmost element could be merged into the doublet before it with likely no consequences.

I do not have high hopes for either optical performance or consistency.  Though elements 3,4 and 11 are the way they are to mess with astigmatism, so I'm sure it will bear Nikon's more recent fingerprint of very uniform performance over the field.

Is there anything they could have done to maintain uniform performance over the field and  address your other concerns? Or are the two goals mutually exclusive?

AiryDiscus wrote:

I'm sure it will bear Nikon's more recent fingerprint of very uniform performance over the field.

AiryDiscus wrote:

peripheralfocus wrote:

bobn2 wrote:

Thanks for this, Bob. I knew that exit pupil size and position was key to understanding what Nikon was aiming for with the geometry they chose, but I hadn't seen it diagrammed. Your drawing was a nice help to me.

I noticed in the 50mm S-Line block diagram that Nikon has placed a fairly large (negative? @ 52mm in diameter) element almost as close to the sensor plane as physically possible. (Maybe 14mm or so?) And a fairly large group just in front of it.

Any comments on what this affords them, if anything? I think I know that entrance pupil size and position are heavily influenced by the frontmost element, so I'm guessing the rearmost element heavily influences the size and position of the exit pupil.

Nikon has made a special point about the aberration control they can achieve by getting elements closer to the sensor. (They talked about working very hard to shave millimeters off the flange distance by redesigning and minimizing interposing structures like the shutter mechanism.) Any thoughts on whether or how some of these factors might be reflected in this block diagram? The rear structure of this lens must be unique in 35mm photography right? The EF mount has the width but not the short flange distance, and none of the short flange distance mounts have the width. Maybe something similar has been done for M4/3rds?

This is not a good lens design...

Element 2 is a paper thin meniscus with R1 ~= R2. This will be extremely difficult to meet good tolerances on wedge of the two surfaces on. It is also cemented to a strong biconcave, a process that will be made more difficult by its thinness.

Element 3 looks inside out.

Element 8 has a nasty thin airspace that strongly varies in thickness with radial position. This will be extremely sensitive to despace.

The airspace between elements 8 and 9 is almost constant, this is already sensitive. If there is an aspheric surface on the back of element 8, even worse. Elements 10,11 are both thin and cemented. This is not very good.

Finally, the rearmost element could be merged into the doublet before it with likely no consequences.

I do not have high hopes for either optical performance or consistency. Though elements 3,4 and 11 are the way they are to mess with astigmatism, so I'm sure it will bear Nikon's more recent fingerprint of very uniform performance over the field.

One could speculate how it happens. It looks like the output of computer optimisation, with the manufacturing constraints set too loose. Then the production people are asked 'can you do it'? and with a newfound corporate ethos of accepting challenge, they say 'yes'. It could also that they think they can achieve high production tolerances.

I'm guessing that the rearmost element is separate from the doublet because the doublet is the group that gives internal focus.

I am following this discussion with great admiration for this forum. Science and tech forum is really very special.

I hope a camera manufacturer decides to introduce a full frame camera that aims compactness as a priority.

Thank you all again

rondom wrote:

The new Nikon Z makes me wonder: How come Leica is able to retain their legacy mount and lenses and make them work with their digital cameras? I am aware of sensors with microlenses etc, and I'm sure there is some image engine manipulation but is there any other secrets to it?

Looking at the new "full frame mirrorless" out there, it looks like what Leica is achieving is unique: a digital M body has the same thickness as a film M except for a small additional thickness around the flange.

They tried to design th digital bodies to work within the parameters of the M system. It did take them some generations.

For example: sensors with micro lenses designed for the angle of incidence of their M lenses output.

Meanwhile Nikon is introducing this considerably larger mount. I doubt that Leica has some secret technology that Sony or Nikon don't have access to.

This is rather puzzling to me. So what is it?

Thanks

p.s. maybe sensor needs more light for AF? and also I assume the ability to introduce very fast lenses?

I thought that mirrorless cameras were supposed to make Lens design easier and simpler😀

KWNJr wrote:

I thought that mirrorless cameras were supposed to make Lens design easier and simpler😀

Open up the possibilities. You could either have them 'easier or simpler' or use the additional freedoms to make the lenses much better, which seems to be what Nikon is doing. I suspect this reflects that at the moment, they see the Z line as a premium product. I should think that come the Z4 or Z5, we'll see some 'easier', 'simpler' and hopefully 'smaller' and 'cheaper' lenses.

SmilerGrogan wrote:

Is there anything they could have done to maintain uniform performance over the field and address your other concerns? Or are the two goals mutually exclusive?

AiryDiscus wrote:

I'm sure it will bear Nikon's more recent fingerprint of very uniform performance over the field.

It depends on the level of performance of the lens.  It is difficult to add aspherical elements to double gauss designs to reduce spherical aberration, and in a normal double gauss they are ineffective at fixing the astigmatism as well.  In a double gauss with a fairly radical change to the front and back like this, they can be very effective for controlling astigmatism and distortion.

I would remove both aspheres that are there now.  Delete element 3, break elements 1,2 and aspherize element 1 surface 2.  Make the new element 3 the zoom (focus) group.  Delete the last element.  Aspherize the final surface, or second to final surface.

That would probably make a simpler, more manufacturable design with similar optical performance, but I couldn't say without raytracing the design.

bobn2 wrote:

KWNJr wrote:

I thought that mirrorless cameras were supposed to make Lens design easier and simpler😀

Open up the possibilities. You could either have them 'easier or simpler' or use the additional freedoms to make the lenses much better, which seems to be what Nikon is doing. I suspect this reflects that at the moment, they see the Z line as a premium product. I should think that come the Z4 or Z5, we'll see some 'easier', 'simpler' and hopefully 'smaller' and 'cheaper' lenses.

I should of said: “easier and/or simpler”🙂

This of course, like you said, does not prevent anyone from making them more complex, and hopefully better.😀

AiryDiscus wrote:

This is not a good lens design...

I debated whether to make this post because it's going to inevitably sound preachy and holier-than-thou and elderly, and it will therefore probably fall on deaf ears. But so be it; the baleful habits that the Internet encourages are getting me down lately.

One way to approach the world intellectually is to believe you are smarter than everybody else. Another way to approach the world intellectually is with humility and respect for the work and thoughts of others, with the sense that you can learn from them, that they can enrich your knowledge and skill. In my extensive experience, the second way makes for a happier, richer, more successful life.

The man who is overseeing the development of the new Nikon Z lenses is named Yasuhiro Ohmura. You can look him up on SPIE, where you are both members. (He's even on LinkedIn). He has a number of patents and SPIE publications, all in the field of lithography lenses. A random sampling of the optical design department and section leaders who work under him includes Takeshi Suzuki, Motohisa Mouri, Hiroki Harada, and Akihiko Obama. I stopped counting their lens patents at around 40. Some of them earned their first optical patents in the 1990s, so just by sheer dead reckoning, I'm guessing they have as much as 100 years of lens design experience between them. They supervise another 8 or 10 younger optical engineers working on the Z system lenses.

In optics, these are people that anyone here could learn from. If you could sit down with them over sake in Tokyo one night, you could start the conversation by detailing for them exactly what's wrong with their lens designs, even though the only data you have is a drawing scanned from a brochure. Or you could start with the assumption that such accomplished engineers must have had reasons for what they did, and maybe it would be interesting to know what those are. It's the difference between "What are the reasons for doing it this way?" and "Here's how you screwed up." Again, it's about respect for others.

I don't know these men, but if they're anything like most of the other Japanese camera engineers I've met (lots), they would be curious about what you're doing, the things you're researching or working on—in short, they'd be interested to know what they could learn from you. And maybe you could show them something they hadn't thought of before. But they would never hear you if they don't approach the world with some humility and respect for the capabilities and talents of others. It's a habit of mind that serves a person well.

peripheralfocus wrote:

AiryDiscus wrote:

This is not a good lens design...

I debated whether to make this post because it's going to inevitably sound preachy and holier-than-thou and elderly, and it will therefore probably fall on deaf ears. But so be it; the baleful habits that the Internet encourages are getting me down lately.

One way to approach the world intellectually is to believe you are smarter than everybody else. Another way to approach the world intellectually is with humility and respect for the work and thoughts of others, with the sense that you can learn from them, that they can enrich your knowledge and skill. In my extensive experience, the second way makes for a happier, richer, more successful life.

The man who is overseeing the development of the new Nikon Z lenses is named Yasuhiro Ohmura. You can look him up on SPIE, where you are both members. (He's even on LinkedIn). He has a number of patents and SPIE publications, all in the field of lithography lenses. A random sampling of the optical design department and section leaders who work under him includes Takeshi Suzuki, Motohisa Mouri, Hiroki Harada, and Akihiko Obama. I stopped counting their lens patents at around 40. Some of them earned their first optical patents in the 1990s, so just by sheer dead reckoning, I'm guessing they have as much as 100 years of lens design experience between them. They supervise another 8 or 10 younger optical engineers working on the Z system lenses.

In optics, these are people that anyone here could learn from. If you could sit down with them over sake in Tokyo one night, you could start the conversation by detailing for them exactly what's wrong with their lens designs, even though the only data you have is a drawing scanned from a brochure. Or you could start with the assumption that such accomplished engineers must have had reasons for what they did, and maybe it would be interesting to know what those are. It's the difference between "What are the reasons for doing it this way?" and "Here's how you screwed up." Again, it's about respect for others.

I don't know these men, but if they're anything like most of the other Japanese camera engineers I've met (lots), they would be curious about what you're doing, the things you're researching or working on—in short, they'd be interested to know what they could learn from you. And maybe you could show them something they hadn't thought of before. But they would never hear you if they don't approach the world with some humility and respect for the capabilities and talents of others. It's a habit of mind that serves a person well.

Here's what my "deaf" ears got out of your comment.

1. By pointing at specific flaws in a single lens design that is not even attributed directly to Ohmura-san, I have put forward that I am smarter than him.
2. His experience developing 40+ element photolithography systems, microscopes on steroids, makes him an excellent designer of camera lens objectives manager of the design of consumer camera lenses.
3. That some of his subordinates hold large numbers of patents makes them good optical designers.  One can only patent a good optical design.
4. Experience (=age) is what it takes to be good at something, not skill or "talent."  Indeed, we all must age into being masters of our field.
5. It takes 13-15 optical designers to produce 3 lens designs over 10 years.
6. By pointing at specific flaws in a single lens design that is not attributed to Ohmura-san, I am being disrespectful towards him.
7. Someone (i.e. you) who can name a significant portion of Nikon's camera lens designers has never met them, even though their identities are often extremely difficult to expose.

By making comments about how I have only seen a drawing scanned from a brochure you have denied me the same assumption of competence you give these men.  I routinely poo-poo the US-centric optics industry for its attitude towards many things (aspheres, robotics, use of anomalous partial dispersion glasses or cost associated with these things to name a few) when our Japanese friends, i.e. Nikon and Canon figured these things out at tremendous scale and low cost quite a long time ago.

It is not that I do not respect the capabilities and talents of others, but rather that in the areas of optics I am very skilled or knowledgeable, it takes more than a long work history to impress me.  There are many bad lens designers in the world, old and young.  I do not care who designed this lens.  It is not a good lens design.  Let it stand on its own merits, rather than attaching the duration of someone tangentially related's career to it as if that makes it a better solution.

quadrox wrote:

AiryDiscus wrote:

Here's what my "deaf" ears got out of your comment.

1. By pointing at specific flaws in a single lens design that is not even attributed directly to Ohmura-san, I have put forward that I am smarter than him.
2. His experience developing 40+ element photolithography systems, microscopes on steroids, makes him an excellent designer of camera lens objectives manager of the design of consumer camera lenses.
3. That some of his subordinates hold large numbers of patents makes them good optical designers. One can only patent a good optical design.
4. Experience (=age) is what it takes to be good at something, not skill or "talent." Indeed, we all must age into being masters of our field.
5. It takes 13-15 optical designers to produce 3 lens designs over 10 years.
6. By pointing at specific flaws in a single lens design that is not attributed to Ohmura-san, I am being disrespectful towards him.
7. Someone (i.e. you) who can name a significant portion of Nikon's camera lens designers has never met them, even though their identities are often extremely difficult to expose.

I am not the person you responded to, but your response irked me a bit, so here goes:

...

Maybe we should cut Brandon a bit of slack here. He is very young and very bumptious, but he's also clever and knowledgable and a useful person too have around here.

I think perhaps what he's doing here is interpreting the specific from his learned knowledge, which is good. I think all his criticisms of the design carry some weight in terms of the general wisdom of the topic, and it's not unusual for people in the know to have valid criticisms in areas of technology. My other technological passion is motorbikes, and it's absolutely usual for engineers with experience to be critical of design features of different manufacturers' engines. For instance, Honda engines generally run the camshafts directly in a 'bearing' cut straight into the aluminium of the cylinder head, they don't use a bearing shell at all. This causes all kinds of problems and engineers dealing with it are regularly critical. Do they know better than the Honda designers who designed the engine? Yes and no. The main problem with the Honda design is, if you have a minor lubrication problem, and the oil film in the cam bearings fails, you trash the whole cylinder head. If they had bearing shells, you'd just trash the shells. However, to the design engineers, this doesn't matter. Their solution provides an engine that is cheaper and simpler to manufacture, and has fewer parts to potentially fail, or cause problems due to mis-assembly. They expect their products to be serviced regularly anyway, so lubrication failures should be vanishingly rare, and when they do happen, it's not Honda's problem that the owner needs a complete new head rather than a few bearing shells. To the engineers using Honda engines (or maybe mis-using them, in the case of those racing them) they see the failures and see this as a design weakness. It's a matter of perspective.

So far as Brandon's criticism of this design, many of them were to do with the design requiring very high manufacturing tolerances, both in terms of profiling the glass and assembly. Presumably, Nikon would never have signed off the design had they not believed that they could meet the tolerances. They will have done a full tolerance analysis (a part of the design of an mass production lens) and where they need to develop production engineering to the required precision, they will have done so. So, the fact that it might be hard to manufacture is not our problem, it's Nikon's, and clearly they believe that it isn't a problem.

The other aspect of Brandon's criticism is whether the required tolerances can be kept in extended use. That to some extent is an unknown, and is like the Honda camshaft issue above, in the end, it's not a Nikon problem, but it is one that Brandon will have seen a lot of in the work he does, so in a way, he's in a better position to comment than are the Nikon designers, who won't see lenses coming back for service for a year or several. It's similar in a way to the American consumer's preference for a large capacity rumbling V8, it's not high tech but they know it's liable to run forever.