Comparing macro lenses using MTF - invitation, take 3

Have questions about the equipment used for macro- or micro- photography? Post those questions in this forum.

Moderators: Pau, rjlittlefield, ChrisR, Chris S.

Miljenko
Posts: 170
Joined: Sat Jun 01, 2013 11:53 pm
Location: Zagreb, Croatia
Contact:

Comparing macro lenses using MTF - invitation, take 3

Post by Miljenko »

Last week our fellow photomacrographer Filip (jurkovicovic) has sent me new batch of lenses/objectives for testing. This time there are 3 finite microscope objectives (Reichert 10x 0.25, Zeiss Jena 6.3x 0.16 Semiplan, Nikon Plan 10x 0.30), 3 Meopta Evitar lenses (3.0/14 and 2 samples of 4.5/30.76) and Sigma Life-Size Attachment.

Image

We have already established magnifications and modes for all those lenses while I have some additional ideas for extra set of data.
Since I've never seen those Evitar lenses, I'll be very keen to hear from Filip their history and possible details. They are already mounted on my testing rig and MTF testing is on the way. Results will appear in this forum somewhere at the weekend.
Best regards,
Miljenko
All things are number - Pythagoras

Lou Jost
Posts: 4638
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

"Evitar" is a poor choice of name for a lens... in Spanish it means "avoid".

That reminds me of a store I saw in Ecuador that had the reverse problem. The Spanish "d" sounds close to an English "t". A clothing store called itself "Closed", their transliteration of the English "Closet" They had a big "Closed" sign on their storefront. They didn't get many English-speaking customers.

Miljenko
Posts: 170
Joined: Sat Jun 01, 2013 11:53 pm
Location: Zagreb, Croatia
Contact:

Post by Miljenko »

No surprise lens manufacturers do not check various languages when selecting model names since mass production stuff like cars are not checked for naming either. Ford is manufacturing Kuga SUV for 10 years now and it's almost not sold in Croatia, Bosnia & Herzegovina, Serbia and Slovenia. Guess why?
Because of the name, of course which in all 4 languages means plague. Maybe population of 15 million people is not big enough for Ford to have big concerns about it but still customers prefer to buy similar cars from other manufacturers instead!
Last edited by Miljenko on Tue Feb 04, 2020 1:39 am, edited 2 times in total.
All things are number - Pythagoras

Miljenko
Posts: 170
Joined: Sat Jun 01, 2013 11:53 pm
Location: Zagreb, Croatia
Contact:

Post by Miljenko »

Filip was kind enough to dig some data for Evitar 2 (30.76mm) lenses and sent me couple of minutes ago. So here it is without any editing. I believe everything is clear even without translation. Unfortunately, he didn't manage to find any data on smaller Evitar 4 (3.0/14).
I believe MTF testing for all 3 Evitar lenses will be finished today.
Image
Image
Image
All things are number - Pythagoras

Pau
Site Admin
Posts: 5214
Joined: Wed Jan 20, 2010 8:57 am
Location: Valencia, Spain

Post by Pau »

Lou Jost wrote:"Evitar" is a poor choice of name for a lens... in Spanish it means "avoid".
Very true! :lol:
I will avoid Evitar lenses

Mitsubishi renames its Pajero car as "Montero" in Spain because the joking sexual connotations of the first word.
Pau

Miljenko
Posts: 170
Joined: Sat Jun 01, 2013 11:53 pm
Location: Zagreb, Croatia
Contact:

Evitar lenses test results

Post by Miljenko »

Yesterday I have finished testing first batch of Filip's lenses, all three bearing Evitar name. I don't have info on how he ended up with two samples of Evitar 4.5/30.76 but on many ocassions it was shown this can be a huge advantage due to frequent sample variations. And this is exactly what happened here, one sample performing very differently to it's sibbling. And there is another Evitar lens covered here, tiny Evitar 4 14mm f3.0 which didn't perform as expected maybe due to sample variation as well.

Image

Meopta Evitar 2 4.5/30.76 as called officially is pretty sofisticated lens design consisting of 6 lens elements in 5 groups packed in nicelly machined and finished alluminum casing measuring 24mm in diameter, 22mm long with M22x0.5 thread. It was obviously designed as close focusing lens since it worked as it should only in normal orientation.
I started testing at 2x magnification using just plain extension tubes. Performance was barely average or slightly below with big differences between samples. Check the results chart which shows not only mediocre MTF resolution but the excessive CA as well, especially with sample #1. At 3x extended performance was even worse compared to many other lenses tested previously while interestingly sample #2 took the lead. CA went down for the lens #1 but doubled for #2.

Image

Then I pulled out my reference set of short tube lenses: Rodenstock Rodagon 60mm f/8 (opened to f/4 as suggested by Robert), Rodenstock Ysaron 4.5/90 and Gretag 6/120 which was discarded right away since 4x magnification was simply to much for Evitar lenses.

Image

But at 2x and 3x performance improved significantly for both 4.5/30 samples (except for sample #2 resolution at 3x). Resolution went up while CA dropped by a huge margin. This is more true for sample #1 which now performed well above average. It didn't change my hall of fame list at top 3 places but surely qualifies at top ten charts at 2x and 3x. When more aggressive presharpening is applied during raw to tiff conversion, MTF figures jump close to 4000 LW/PH at 2x and over 3000 LW/PH at 3x. So Filip, you have one great lens there; just make sure you find appropriate 60 to 90 mm tube lenses to pair that little gem.
Unfortunately, the same can't be said about 3/14 Evitar where the tube lens couldn't improve resolution and CA by much. Resolution is so poor that otherwise great Fujifilm X-T2 focus assist was struggling to show exact focusing plane. Performance did improve slightly with the tube lens (Rodagon 43mm f5.6 was used) but not to a level worth further testing.
So the conclusion is: Meopta Evitar 4.5/30.76 is one great lens if you are lucky enough to get the proper sample. It's a shame it doesn't appear very often on Ebay (as the matter of fact, I've never spotted one) so Filip might give us a hint how and where can we find some!
All things are number - Pythagoras

kutilka
Posts: 51
Joined: Thu Feb 04, 2010 9:33 pm
Location: Czech republic
Contact:

Post by kutilka »

I lent the Evitars to Filip for testing.
They are prototype pieces from Meopta Research and Development. Therefore the different quality.
They were used in the 1970s in microfiche readers and there were more versions.
I have about 10 Evitars still in the drawer ...
Excuse me for my English. (Google translator).
Image

To simulate the image test through the lens in ZEMAX:
http://www.manualniskla.cz/viewtopic.php?f=21&t=950

Miljenko
Posts: 170
Joined: Sat Jun 01, 2013 11:53 pm
Location: Zagreb, Croatia
Contact:

Post by Miljenko »

Thank you for clarifying Evitar history, Kutilka. And thank you for linking this thread on Manualni Skla forum. Yes, I have checked fully your thread on Evitar lenses. As the matter of fact, the whole site seems interesting so I might check it from time to time. And yes, Google translate works perfectly ok for that 50% of words that are different from Czech to Croatian!
So far I have tested more than 20 microfiche reader lenses with very different results. In most cases they are not good enough for serious macro work. However, some exceptional pieces appear here and there like this Evitar 2.
I believe you might find interested buyers if you offer some of that Evitar stuff on Equipment Exchange forum here.
All things are number - Pythagoras

Miljenko
Posts: 170
Joined: Sat Jun 01, 2013 11:53 pm
Location: Zagreb, Croatia
Contact:

Zeiss 6.3x 0.16 Semiplan tested

Post by Miljenko »

Next piece of optics from Filip's collection sent for testing on this occasion is Zeiss (from Jena, East Germany factory) 6.3x N.A.0.16 Semiplan. This vintage objective was presented here some time ago by Antonio (Yawns) with very positive impressions. Since all the photos I've seen on the web were of high quality, I've acquired one as well so this was an opportunity to test two samples side by side.

Image
Image

According to serial numbers it seems those were manufactured years apart. Dimensions are equal but the engraved lettering and signs are different. Internal casing looks the same except for the tip area which is shiny chrome plated on Filip's while dull nickel plated in case of mine.
Both were tested in 4x to 8x magnification range both extended on bellows and by the tube lenses.
Performance at 8x was very inferior (both resolution and CA) so I didn't bother to spent much time on that side of nominal magnification. Using 10x or 12.5x objectives short focused to 8x would give much better performance.
More than 500 shots were done and processed using Fuji X-T2, Capture One and Imatest in exactly the same way I've done it for the last 4 years.

So here are the results for both objectives:

Image

One parameter not included in this chart is light falloff at the corners.
It becomes visible below 5x but easily ironed out with Capture One or Photoshop. More so at 4.5x but not that successful at 4x. I'll deal with that in separate post.
You can easily see very different performance from those two samples when going through various magnifications. Sample #1 (Filip's) has in general higher resolution at expense of higher CA. Different too is response to the tube lens instead of plain extension tube.
In general, both objectives perform much better when shortened to 5x or even 4.5x . Central resolution at 4x when using the tube lens is the best I've ever measured, edge resolution not so much and the vignetting is hard to remove completely.
Filip's sample gains quite a bit of resolution with the tube lens (Rodenstock Rodagon 5.6/150) at nominal magnification but the CA gets unacceptably high.
Otherwise, both objectives didn't gain much of performance with tube lenses (except slightly at the magnification extremes) so you might decide not to fiddle with this additional component. Besides, working distance halves to 5mm when using tube lens; another drawback of this concept.

My conclusion after this exhaustive testing is very positive. Although not advertised as such, this objective controls CA very well but the „semiplan“ badge is not deserved since it's very far from „plan“ variety. Focusing plane is very curved which fortunately is not a big deal 'cause we're all stacking, aren't we?
All things are number - Pythagoras

Lou Jost
Posts: 4638
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Very interssting. Are these Zeiss objectives not designed for CA correction in the eyepiece? I had heard that some low-m Zeiss objectives do not need this correction.
Lou

Miljenko
Posts: 170
Joined: Sat Jun 01, 2013 11:53 pm
Location: Zagreb, Croatia
Contact:

Post by Miljenko »

Lou Jost wrote:Are these Zeiss objectives not designed for CA correction in the eyepiece? I had heard that some low-m Zeiss objectives do not need this correction.
Lou
Zeiss aus Jena objectives from 1968. to 1982. were of achromatic design, requiring mild eyepiece correction for CA but only for magnifications over 12.5x. Such objectives had "-c" sign at the end of it's designation.
Obviously, this 6.3x Semiplan objective is almost free of CA, especially at some magnifications which are incidentally outside it's default values. Stacking it to some tube lenses as done here gives even better performance.
Best,
Miljenko
All things are number - Pythagoras

Miljenko
Posts: 170
Joined: Sat Jun 01, 2013 11:53 pm
Location: Zagreb, Croatia
Contact:

Single lens element tube lenses tested

Post by Miljenko »

Next lens from Filip's collection sent for MTF testing is well known Sigma Life-Size Attachment. A single lens apochromatic „macro filter“ is designed as an attachment you put in front of ordinary (tele) lens in order to do handy close-ups. There are many both contemporary and vintage models on offer at Ebay with various focal lenghts which determine total magnification. Probably best known is Raynox DCR series of „Macroscopic Lenses“ covering wide range of focal lengths. It was established long time ago that infinite microscope objectives are quite happy with simple design tube just like those macro attachments.

This single lens (actually, it consists of two lens elements made of different kind of glass, glued together to control CA) is extended with tubes or bellows in front of the camera with the objective mounted right onto it. Extension length is adjusted so the focus plane is at infinity. So far, it seems most macro attachments when used as the tube lenses perform better when reversed. Some authors (Robert OToole comes to mind) have discovered that focusing „beyond infinity“ using somewhat shorter extension provides better total performance. This shortening produces proportionally less magnification which in turn might put taking lens (objective) into favorable condition but I'll leave this speculation to more qualified folks here.

Different manufacturers unfortunately decided on various TL focal lengths but probably most frequently used Nikon and Mitutoyo objectives require 200mm or diopter value five.

This is exactly how „long“ is Sigma Life-Size Attachment lens and since I have other two single lenses with roughly the same focal length, this is one nice triple tube lens shootuot using Mitutoyo 5x 0.14 objective.

Image

Other two macro attachments are Kenko #5 (55mm thread version) and Raynox DCR-150. All three were reverse mounted, focused at infinity and shortened to 10% lower total magnification. Since Raynox has longer focal length (208mm declared), I have tested it at infinity setting and shortened to 5x and 4.5x magnifications. Based on declared focal length, infinity setting is supposed to produce 5.2x magnification but my sample provides close to 5.35x with Mitutoyo 5x 0.14, Nikon 10x 0.30 objectives and few other lenses with known focal lengths. This means it's actual focal length is 214mm. I would appreciate very much if someone could confirm or decline this finding.

The results of all MTF tests are presented in numerical chart below. I believe Filip is not going to be happy with outcome since Sigma attachment tested less favorable of all three. Much cheaper Kenko No.5 performed better although Sigma came close when shortened for 4.5x magnification. However, Sigma's CA was worse of all three obviously to inferior design and/or selection of glass for both elements.

Image

The winner of this tube lens comparison is Raynox DCR-150. Incidentally, it's peak performance happens when short focused to extent where it behaves as if it was 200mm lens. Mitutoyo produces exactly 5x magnification at that setting and the best resolution I measured with this objective with any tube lens. Kenko #5 does have much lower CA but with DCR-150 CA is not that bad at all. Of course, this is probably not the best performance you can get from Mitty 5x since I have a feeling my sample is not the best around and I don't possess any of those more expensive 200 mm tube lenses to test with. Nevertheless, I'm happy to have side benefit out of this MTF comparison of finding the best tube lens for my Mitutoyo 5x.

And yes, Filip will probably have to consider buying Raynox DCR-150 for his infinity objectives.
All things are number - Pythagoras

Scarodactyl
Posts: 472
Joined: Sat Apr 14, 2018 10:26 am

Post by Scarodactyl »

I like the kenko lenses, 94% of the resolution and 50% of the CA vs the storied raynox, all in a package that reliably costs less than 20bux. For those occasions where I've needed an extra tube lens for this or that (most recently on stereo microscopes) it's nice to have such a good, cheap option at hand.
I wonder if there is much variation copy to copy on these.

Miljenko
Posts: 170
Joined: Sat Jun 01, 2013 11:53 pm
Location: Zagreb, Croatia
Contact:

Post by Miljenko »

Scarodactyl wrote: I wonder if there is much variation copy to copy on these.
Not likely. Large variations are expected with complex lens designs consisting of 10 or 15 elements. Here we have only one compound lens so I expect to see no more than 2 or3% resolution and/or CA sample to sample variations.
All things are number - Pythagoras

Lou Jost
Posts: 4638
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Because you have actual numbers, we can test the hypothesis that short-focusing gives better results. If a tube lens was neutral with respect to focusing, then the resolution should rise as the magnification drops; a 50% drop in magnification should double the resolution. In this example, if the actual resolution were more than double, then we could say that the tube lens performs better when short-focused.

So for the short-focused Sigma LSA, the predicted neutral resolution (obtained by simply shrinking the 5x image to 4.5x) would be about 1890. You observed 1987, so yes, this tube lens does improve when short-focused.

Post Reply Previous topicNext topic