3X test - Objectives vs Lenses - Final Results Online

[RobertOToole]

https://www.closeuphotography.com/3x-lens-test

Lenses:
Canon 35MM f/2.8 Macrophoto Lens
Canon-MP-E 65mm f/2.8 1-5x Macro Photo Lens
Minolta DiMAGE Scan Elite 5400 Scanner Lens
Schneider Kreuznach Componon S 50mm f/3.5 Fixed Aperture lens
Tominon 35mm f/4.5 Lens

Last minute addition to the test and not in the group picture:
Olympus Zuiko Auto-Macro 38mm f2.8

Objectives:
Mitutoyo M Plan Apo 5X 0.14 objective @ 3x
Nikon 3x Measuring Microscopes Objective
Qioptiq Mag.x LD Plan APO 5x / 0.20 Objective @ 3x



This was an interesting test and I do hope that a few people out there that aren't too tired of looking at silicon wafer crops to actually look over the test results

The standouts were the Tominon 35mm due to the low cost and chromatic correction, and the Minolta Elite 5400 with APO correction and edge to edge sharpness. These two lenses when compared side-by-side next to some of the popular older well regarded lenses like the Canon 35mm MP, the Canon MP-E 65 and the Olympus 38mm f/2.8 with a target like the wafer were disappointing and not exactly as I like to remember since I have owned probably 6 or more Canon MP 35s and maybe 3 Olympus 38 f2.8s over the years.

Here are a couple of comparisons. Corner on the left, center on the right.

Be sure to click on an image for the full-size 2500 version.

Minolta 5400 vs Canon 35mm MP.



The Canon MP35 was my favorite 3x lens for years but now the Minolta 5400 has replaced it as my new favorite lens from 2-3x.

Tominon vs Olympus 38mm



The 38mm Olympus vs the Tominon comparison is tight. The Tominon performance is amazing for the cost, I think I paid $35 maybe, and it has much better CA suppression.

The Minolta Elite 5400 lens image quality proved to be best out of all the lenses and was surprisingly close to the pushed down Mitutoyo 5x M Plan.

Amazing performance. The Mitutoyo is better but the Minolta is surprisingly good at 3x.

Minolta 5400 vs Mitutoyo 5x M Plan



Camera: Sony A6300, model # ILCE-6300
Sensor size: APS-C. 23.5 × 15.6 mm. 28.21 mm diagonal. 3.92 micron sensor pitch
Flash: Godox TT350s wireless flash x 2 with one Godox X1s 2.4G wireless flash transmitter
Vertical stand: Nikon MM-11 with a Nikon focus block

A series of images was made with each lens in 5 micron steps. This was repeated for each aperture. Then the sharpest frame was then chosen using Photoshop at 100% actual pixel view. Separate images were selected for center, edge, and corner if needed. Each image was processed in PS CC with identical settings with all noise reduction and lens correction turned off, all settings were zeroed out (true zero) and the same settings were used for all of the images. All of the images shown here are single files.

Questions comments welcome.

Thanks for looking.

[Lou Jost]

Thanks so much for making these extensive tests. These series of tests at fixed magnification will be an important reference for everyone in our community.

The DiMage continues to surprise me too! Something inside me doesn't want to believe that such a little thing can be so good.

[mawyatt]

Robert,

Thanks again for all the tests and showing the results.

BTW I love looking at Si wafers

An amazing effort and hats off to you sir

Best,

[RobertOToole]

Thanks Mike !

This is a close up of one of the 100% crops areas at 42x, what are the different types of patterns we are looking at here?

BTW I can guess using the Microlithography basics I know but you can help?

I assume these are etched patterns used for quality control?

What is the poly vs the metal, are these at different materials on top of the substrate? The poly patterns look like much higher resolution.

BTW this is single image, not a stack, its a wonderful thing when you can get a wafer leveled for a photo at 42x!

Thanks in advance.

[RobertOToole]

Thanks so much for making these extensive tests. These series of tests at fixed magnification will be an important reference for everyone in our community.

The DiMage continues to surprise me too! Something inside me doesn't want to believe that such a little thing can be so good.

Glad to hear someone else feels the same way Lou. I am sure lots of people don't really want to hear more else about the Minolta 5400 but I am still getting surprised by this lens more than 9 months after the first time I tried one.

Thanks for taking time to comment.

Robert

[Lou Jost]

Your question to Mike is interesting. It never occurred to me that those line triplets might actually be there for a reason. I imagined they were some kind of branding or authorship marks. But it makes perfect sense that they are testing the quality (resolution) of the photolithography/mask/resist system...by looking at those marks, chips that had mask alignment ossues could be easily and automatically detected and discarded.

[ray_parkhurst]

Your question to Mike is interesting. It never occurred to me that those line triplets might actually be there for a reason. I imagined they were some kind of branding or authorship marks. But it makes perfect sense that they are testing the quality (resolution) of the photolithography/mask/resist system...by looking at those marks, chips that had mask alignment ossues could be easily and automatically detected and discarded.

Usually there is no visible issue with those marks. Any systemic yield issues tend to show up in PCMs that check electrical characteristics of the different layers and interconnects. I do think AVI (Automatic Visual Inspection) will check such marks and look for defects though.

[Chris S.]

Robert, another capital test! (You've been posting these at a rapidity that challenges considered response. Wow. Given that conducting such broad, thorough tests and clearly presenting the results takes tons of time, one wonders if your day has many more hours in it than 24?)

This 3x test is especially interesting, in that 3x has been a difficult magnification for which to choose an optic--not quite high enough to demand a microscope objective, but stretching the bounds of almost everything else. Must confess to being happy that three lenses I own--the Mitutoyo 5x, Canon MP 35mm, and Minolta 5400--performed very well in your tests. While I admire the small but real advantage of the Qioptiq, I don't see one of those as being in the future for many of us. Given its price, and a sense I have that its better optical output would not be visible in most real world work, it seems to lie well into the domain of diminishing returns. This said, should I suddenly find myself transformed into a billionaire, I may order a full set.

--Chris S.

[mawyatt]

Thanks Mike !

This is a close up of one of the 100% crops areas at 42x, what are the different types of patterns we are looking at here?

BTW I can guess using the Microlithography basics I know but you can help?

I assume these are etched patterns used for quality control?

What is the poly vs the metal, are these at different materials on top of the substrate? The poly patterns look like much higher resolution.

BTW this is single image, not a stack, its a wonderful thing when you can get a wafer leveled for a photo at 42x!


Thanks in advance.

Robert,

These wafers are from the 80~90 era and thus well behind what's used today. So these answers are best guesses.

The various etched layers might be for wafer alignment (15 * for example), test and QC (assist in failure analysis). Probably for fully automated fab systems, even back then.

The METAL is top metal (back then may have had ~6 metal levels) and the numbers represent the metal line width from 1 to 1.7 microns.

The CONTACT is likely and underlying metal and the "dots" are metal that would be utilized for thru vias (contact from one metal layer to another).

The POLY is a poly silicon layer (not crystal like the wafer, but silicon molecule random orientation). These poly layers are usually doped P or N type and the lines can be used for resistors and/or high resistivity interconnects. Poly based resistors are generally not very precise and have high non-linear temperature coefficients and complex behavior.

Note how the top metal (aluminum) and poly differ in edge definition, with the poly being much better defined, this is likely related to the etching process.

If my best guesses are correct then it appears you have a superb optical system easily resolving micron level features. Please let us know what your setup was for these 42X superb images.

Anyway, hope this helps,

Best,

[typestar]

Dear Robert,

thankyou very much for this test, again.

I only can imagine, how much time and efforts you invested in all your testing and - like many in this community - I love and appreciate your GREAT work!

Considering the age of optical construction (designed for the large Polaroid copy camera) -- it is great to see how well the "cheap" Tominon 35 mm fights here (and glad to have it)
and how it beats the ("younger" and much more expensive ) Olympus 38mm
(It would be of interest how the 50 mm Tominon would work here...)

All the best!

Christian

[RobertOToole]

Robert, another capital test! (You've been posting these at a rapidity that challenges considered response. Wow. Given that conducting such broad, thorough tests and clearly presenting the results takes tons of time, one wonders if your day has many more hours in it than 24?)

All this testing has taught me a much more careful, and methodical photographer. I have a much better 'flow' to my work now so I can relax and have fun when I am shooting without getting hung up on technical issues.

So shooting is the easy part. Processing the info is what eats up my time.

I am sitting on the results from 18-20 lenses @ 2.1x. Not sure when I am going to find the time and energy to tackle those stacks of files

This 3x test is especially interesting, in that 3x has been a difficult magnification for which to choose an optic--not quite high enough to demand a microscope objective, but stretching the bounds of almost everything else. Must confess to being happy that three lenses I own--the Mitutoyo 5x, Canon MP 35mm, and Minolta 5400--performed very well in your tests.

Agreed. The Minolta 5400 surprised me getting close to the Mity at 3x. That lens is like a miniature Godzilla eating up all the competition as it goes a long.

While I admire the small but real advantage of the Qioptiq, I don't see one of those as being in the future for many of us. Given its price, and a sense I have that its better optical output would not be visible in most real world work,

Agree. Hope it doesn't appear as showing off, it's hard not to include it when its just sitting on the shelf within arms reach.

it seems to lie well into the domain of diminishing returns. This said, should I suddenly find myself transformed into a billionaire, I may order a full set.--Chris S.

I actually mention the same case in the comments for the Tominon 35mm. It has very little lateral CAs and covers a big image circle since it was designed to cover 4 x 5 film. You can beat it but its going to cost you 5-10x the price of the lens. I paid $35 for this lens (my 4th copy but thats another story). The 4-element, 3-group reverse Tessar-type Tominon beats just about every enlarging lens I have ever tried including all of the those labeled APO


Thanks for the nice comments Chris.

[Lou Jost]

I really like that you included the Mag.X. It is good to see what is possible, not just what is practical or easy. I lust after that lens!

[RobertOToole]

Robert,

These wafers are from the 80~90 era and thus well behind what's used today. So these answers are best guesses.

The various etched layers might be for wafer alignment (15 * for example), test and QC (assist in failure analysis). Probably for fully automated fab systems, even back then.

The METAL is top metal (back then may have had ~6 metal levels) and the numbers represent the metal line width from 1 to 1.7 microns.

The CONTACT is likely and underlying metal and the "dots" are metal that would be utilized for thru vias (contact from one metal layer to another).

The POLY is a poly silicon layer (not crystal like the wafer, but silicon molecule random orientation). These poly layers are usually doped P or N type and the lines can be used for resistors and/or high resistivity interconnects. Poly based resistors are generally not very precise and have high non-linear temperature coefficients and complex behavior.

Note how the top metal (aluminum) and poly differ in edge definition, with the poly being much better defined, this is likely related to the etching process.

If my best guesses are correct then it appears you have a superb optical system easily resolving micron level features. Please let us know what your setup was for these 42X superb images.

Anyway, hope this helps,

Best,

Thanks for the input, I appreciate it.

It surprises me is that I had enough DOF at 42x to get all of what I wanted to show in a single image. The Metal markings look like they have a lot more depth than the poly marks.

The lens is a Mitutoyo 50x 0.55 pushed down to 42x with a short focus ITL200.
Bought this lens for a project. I couldn't pass on it, although It has a little fleck or chip in the front element, the price was right at $100 or so. The barrel was clean so if the image quality was off, I could use it for the barrel since I have another 50x with clean class with a marked up barrel. Turns out the performance, to my eyes, is indistinguishable between the two. Sometimes I get lucky!

Thanks again Mike.

[RobertOToole]

Dear Robert,

thankyou very much for this test, again.

I only can imagine, how much time and efforts you invested in all your testing and - like many in this community - I love and appreciate your GREAT work!

Considering the age of optical construction (designed for the large Polaroid copy camera) -- it is great to see how well the "cheap" Tominon 35 mm fights here (and glad to have it)
and how it beats the ("younger" and much more expensive ) Olympus 38mm
(It would be of interest how the 50 mm Tominon would work here...)

All the best!

Christian

I agree, the Tominon 35 is one of those unknown bargains out there and a great performer.

At one point I did have a full set of Tominons, 17-135 but the 17 and 35 were the most interesting performance wise. I ended up selling the others.

Thanks for taking time to comment.

Robert

[ray_parkhurst]

...
The Metal markings look like they have a lot more depth than the poly marks.

Poly is below contact, which is below the metals. Poly and contact are both doped, so are quite thin, which is why the resolution is better. Metals are thick, so cannot be spaced as closely, and have worse resolution due to alignment, etching variation (under or over etching producing undercut or shorts), and grain boundary issues. Design rules for line spacings get larger as you go higher.