A Tale of Three 20's

[dmillard]

I just tested the three objectives pictured above (Nikon M Plan 20/0.4 ELWD, WD=10.5mm; Nikon LU Plan EPI 20X/0.45, WD=4.5mm; Mitutoyo M Plan Apo 20/0.42, WD=20mm) to determine their center and corner performance on my D300 (23.6 x 15.8mm sensor, 28.6mm diagonal).

The Mitutoyo and the Nikon LU infinity objectives were mounted approximately 75mm and 100mm from the edge of the front barrel of a 210mm f/6.1 Schneider Xenar lens, which had been focused to infinity. Using a stage micrometer and ImageJ, a free image analysis program, I found the magnification resulting from both combinations to be 21.65X ( I believe that the focal length of the Xenar is slightly greater than designated, since the Mitutoyo provides 24X magnification when used with a 240mm Apo-Ronar, as predicted). I matched the 21.65X magnification when I used the ELWD.

The wing was illuminated with a Vivitar 283 flash through three fiber optic light guides, the light slightly softened through a cylinder of Lee 250 ½ White Diffusion gel.

I took approximately 140 JPEG images at 2μm increments with each objective, and then stacked them with ZS PMax. The output images were then processed minimally in Photoshop (Levels adjustments and minor USM sharpening) before posting. The crops below are actual pixel size.


Olive Hairstreak, Mitoura grynea, Mitutoyo 20

Center Crops:


Mitutoyo


Nikon LU


Nikon ELWD

Corner Crops:


Mitutoyo


Nikon LU


Nikon ELWD

All three lenses performed fairly consistently across the image field at this magnification, with little apparent deterioration in quality from center to corner. The Mitutoyo seems to provide marginally better definition than the Nikon LU objective, and the Nikon ELWD objective is a little softer. On the other hand, it's a lot easier to use Considering the similar numerical apertures of these three objectives, I'm surprised at the visible difference in image quality.

For the next round of tests, I would like to see if reducing the distance between the front of the tube lens and the Mitutoyo objective mount has
any effect (Mitutoyo specifies 76.5mm for a 30mm image field, but as Charles Krebs points out in this thread here , they also state that "a smaller distance does not affect an optical performance".)

I would also like to see if there are any negative effects (on my sensor format) of using the intermediate lens focused at distances other than infinity.

[Charles Krebs]

David,
Great comparison, thanks for posting it.

[Craig Gerard]

David,

Most helpful and concise. Thankyou.

Looking forward to the 'next round'

Craig

[jotafoto]

Thanks David.
I guess you've made more cuts than necessary to ensure focus. With 3μm could have solved the stack, no?.
The conclusion is that it is better Mitutoyo, incredibly has more contrast and resolution than the Nikon LU, difficult to distinguish, but there is some difference.

I'm waiting on 20 ELWD finite, and I will do a test with the Nikon BD Plan 20 / 0,4 - Nikon BD Plan 20 / 0.4 ELWD, and Mitutoyo 20 / 0.42. I guess I'll get similar results, but I'll try FF on my Canon 5D. It might be better with a camera with a higher density of pixels. If I had the 5D mark II ...

Your tests are impeccable. Thanks for sharing
greetings
Javier.

[ChrisLilley]

Thanks for making this interesting comparison. I was shocked by how bad the ELWD was, even in the centre, compared to the other two. That blue-purple glow around all the white areas ... plus the overall hazy contrast.

Are there any bright rings of metal on the moun surface that would tend to produce haze?

The Mitutoyo is clearly superior in colour clarity, crispness and contrast - and has 20mm working distance too?

[rjlittlefield]

The Mitutoyo is clearly superior in colour clarity, crispness and contrast - and has 20mm working distance too?

...and easily obtained new from Edmund Optics for only $1,830 list.

David, have you told us somewhere how you obtained this gem?

BTW, I'll echo others' compliments on the quality of the test. Very clear and concise, easy to compare.

The only concern I might have about the M PLan test is that to match magnification at 21.65X, you probably had to increase the extension beyond the nominal 210 mm. By rough calculation, you're getting near the edge of tolerable variation as shown in this classic graph. I haven't tested enough to know for sure what those lines in the graph mean, but it's possible that some of the contrast loss is due to aberrations introduced by tube length variation.

--Rik

[Chris S.]

David--a really solid test with interesting results. Thanks!

Like Chris L., I am struck by the ELWD lens' purple fringing around the areas of high contrast. Looks as if the LU has a bit of the fringing, though much less, and the Apo none at all. So is this chromatic aberration? Or as Chris L. wondered, flare from bouncing light?

If it's any clue, I have 20x, 40x, and 60x ELWDs. I sometimes see purple fringing at 20x, more often see it with at 40x, and very often see it at 60x. It does not appear to exhibit symmetry, and it moves around depending on where I put the key light, being particularly present in shadow areas. It is most visible in high contrast areas. It is visually reduced by the use of diffuse light or enough fill light to make shadow areas not too dark. My sense has been that is flare, although I don't know. I have flocking everywhere I can think of in my rig. I haven't yet tried a flare-cut diaphragm, and there is certainly light hitting the fronts of the objectives, as tight working distances limit the practicability of a lens hood.

Edit: I didn't see Rik's reply until after I hit "submit." Interesting point about the tube length, Rik--I've been pretty loose about bellows extensions with my ELWDs. I've made it a point not to go shorter than nominal, but haven't worried about going a bit longer. I'll take this more carefully into consideration, in the future--maybe it makes a difference.

Cheers, and thanks again, David.

--Chris

[Chris S.]

I guess you've made more cuts than necessary to ensure focus. With 3μm could have solved the stack, no?

Javier, you may be as surprised as I was to see that Edmunds specifies the depth of focus for the Mitutoyo 20x Apo here as 1.6μm. I'm sure there are judgments implicit in that number, which may or may not coincide with what we find to work in stacking. But when I test this lens in my own rig--which I definitely will--I'll try some even smaller increments to see if they offer any benefit.

Cheers,

--Chris

[morfa]

Excellent test David – many thanks!

The purple fringing (or whatever it is) in the M Plan ELWD seems perfectly consistent with my experiences with a Nikon BD Plan 20/0.40 ELWD (optically identical to the M Plan, as far as I know).

I'm impressed by how well you've managed to reproduce the conditions between the different objectives in spite of their differing properties. If you have the time I'd love to see a photo of the setup (preferably with the Xenar in place)!

[rjlittlefield]

I'll be very interested to see the results of tests with various step sizes.

I too am puzzled by the value 1.6μm. There is a simple limiting formula that DOF = 0.5/(NA^2) microns. This is the first term from a formula at microscopyu, rounded to use 500 nm instead of 550, and ignoring sensor resolution and CoC issues that can only increase DOF. But this formula produces the number 0.5/(0.42^2) = 2.83μm. As noted, there must be judgments implicit in the number, but the discrepancy seems large and in the wrong direction.

Perhaps Edmund is quoting the maximum distance from perfect focus, which is only half of what we photographers would normally consider DOF.

--Rik

[Oskar O]

Thanks for an excellent comparison, very interesting.

May I ask what is the Xenar designed for? A cine lens? Old telephoto?

As noted, there must be judgments implicit in the number, but the discrepancy seems large and in the wrong direction.

Stupid question: is it typical to list depth of focus rather than depth of field for objectives? And is there a standard measurement for acceptable CoC?

[RogelioMoreno]

David,

Thank for this excellent test and for sharing.

That Mitutoyo is amazing! Hope to get mine in the near future.

Rogelio

[rjlittlefield]

is it typical to list depth of focus rather than depth of field for objectives?

I don't know. I raise this possibility because new member brianc1959, who lists his occupation as lens designer, recently posted a formula for "DOF" that gave half the value of mine, then followed up with another post "Note that the focus step size can be double the DOF figure I mentioned above since the maximum error will occur halfway between the steps." If Brian is listening, perhaps he can clarify about the different conventions.

And is there a standard measurement for acceptable CoC?

Most people use about 1/1500 of the frame diagonal. That's appropriate for normal size prints viewed at normal distances, but when you're pixel-peeping into a 15+ megapixel image, it's too big.

--Rik

[Charles Krebs]

I think Edmund O is using numbers that Mitutoyo provides.

Take a look at page 32, #9 in this document, it may shed some light on this:
www.krebsmicro.com/MitutoyoE4191-378.pdf

(And if you look at the objective specification tables in this PDF, you will see the same numbers that Edmund uses).

I'm not sure if there is a "hard" rule, but it is most common to refer to "depth-of-field" as the acceptable distance range that yields sufficient resolution in the subject space, and "depth-of-focus" as the acceptable distance range of the image at the film/sensor space. But it is not unusual at all to see "depth-of-focus" used where (I think) most expect to hear "depth-of-field". (And as we all know there is a significant number of variable factors that go into these calculations.)

[dmillard]

Thank you for all your comments.

Thanks David.
I guess you've made more cuts than necessary to ensure focus. With 3μm could have solved the stack, no?. . . . .

I'm waiting on 20 ELWD finite, and I will do a test with the Nikon BD Plan 20 / 0,4 - Nikon BD Plan 20 / 0.4 ELWD, and Mitutoyo 20 / 0.42. I guess I'll get similar results, but I'll try FF on my Canon 5D. . . . .
Javier.

Hello Javier,

I used the formula on the Nikon Microscopy U website for the Nikon LU lens to obtain a depth of field of 3.85μm, then used about a 50% overlap to minimize any effects of stacking on the lens tests. I'm not usually that zealous!

I'm looking forward to seeing the results of your comparisons.

...and easily obtained new from Edmund Optics for only $1,830 list.

David, have you told us somewhere how you obtained this gem?

Hello Rik,

I was very fortunate to pick mine up on eBay for a little over $400 from a seller who sold a large group over a short period of time. I was a little concerned that the described "excellent" condition was an eBay exaggeration, but the objective exceeded my expectations.

If you have the time I'd love to see a photo of the setup (preferably with the Xenar in place)!

Hello John,

I'll post a picture of my setup this weekend.

May I ask what is the Xenar designed for? A cine lens? Old telephoto?

Hello Oskar,

The Schneider Xenar is a single-coated Tessar formula lens (4 elements in 3 groups). The 210mm Xenar (discontinued about ten years ago) was intended for use with view cameras. On Schneider Optics' website, under Vintage Lens Data, they describe the image circle as being 580mm, which I'm sure is an error, but they also describe the effective focal length as being 215mm, which explains my greater than expected magnification.

David

Edited 1/12/11 to correct my faulty memory concerning objective cost.