Resolution at (rather) high magnification

[enricosavazzi]

These are a few tests I run on fossil foraminifers. The setup is:
Leica Galen III stand, LED illumination for focusing and framing, electronic flash (Nikon SB-800) for exposure in iTTL mode, Nikon D200 camera. The distance between sensor plane and mounting flange of the objectives is about 210 mm. This gives a magnification of 12x (verified with scale, at least within 10%) with the JML 21 mm lens and 54x (derived by comparison with other test images) with Nikon M Plan 40x 0.5 ELWD lens (210 mm finite tube length, no cover glass). All stacked with Zerene Stacker PMax, retouched only to remove defect trails. Slice depth is 20 µm with JML, 5 µm with Nikon lens. No optics are present between lens and sensor plane.

The first two pictures are reduced full frame and 1:1 cropped detail with JML.



The remaining two pictures are reduced full frame and 1:1 cropped detail with Nikon lens. Note that the whole picture records the same subject area as the preceding detail one.




It is obvious that using the Nikon 40x lens gives more detail, compared to using the JML lens, cropping and enlarging. However, the image from the Nikon 40x lens does not really show that much detail when enlarged. The subject may also be "difficult", being translucent and showing internal as well as surface detail. Used as a microscope objective, the Nikon lens does give excellent visual results. Is this amount of detail satisfactory given that the magnification exceeds the specifications of this lens, or does the same lens model perform better in someone else's setup?

[Cyclops]

Comparing image # 2 and # 3,with the white line, really shows a difference! The line is really quite fuzzy with the JML lens, almost like an image from a lo-res digital camera!

[Charles Krebs]

Enrico,

Is this amount of detail satisfactory given that the magnification exceeds the specifications of this lens, or does the same lens model perform better in someone else's setup?

Hard to say. For one thing I do think this is a difficult subject to use for assessing resolution.

Another thing is that both lenses are being used far into the area where diffraction issues are substantial.

The 21mm JML is about an effective f45 and the 40X is about an effective f54. So instead of trying to see which is "sharper", in this case it becomes more trying to see which is less blurry.

If the 40X were used at it's designed magnification (40X) the effective aperture would be about f40. And 0.50 is a very modest NA for a 40X... a price that is paid to get the longer working distance. A common NA for a 40X is 0.65, and when used at 40X that provides an effective aperture of about f30. I think in both cases the differences (compared with the 40/0.50 at 54X) would be noticeable.

I realize that my constant reference to "effective apertures" involves some over simplification of this issue. But as a quick and simple benchmark I have yet to see a case where it did not track directly with the resolution results I see. (This would be with good quality optics, used close to their "design" geometry. and effective apertures from about f22 and greater)

The distance between sensor plane and mounting flange of the objectives is about 210 mm. This gives a magnification of 12x (verified with scale, at least within 10%) with the JML 21 mm lens and 54x (derived by comparison with other test images) with Nikon M Plan 40x 0.5 ELWD lens (210 mm finite tube length, no cover glass).

This puzzles me. I don't understand why you would be getting 54X with the 40/0.50 M Plan.

[enricosavazzi]

Hi Charles,

I re-did the measurements after finding out that the distance was actually well over 210 mm and shortening the tube length. With the new setup, the distance between sensor plane and mounting flange of the objective is quite close to 210 mm, but I still cannot measure it exactly because of all the fat microscope stuff in-between. The modified trinocular head I use for focusing contains no optics when switched to photo mode. I took pictures of an eyepiece reticle with 100 divisions in 10 mm.

The Leitz Photar 25 mm on an 8 mm RMS extension ring gives 7.5x.
The JML 21 mm gives 9.8x.
The Nikon M Plan 20x ELWD gives 20.3x.
The Nikon M Plan 40x ELWD gives 40.7x.

These results are more reassuring. The discrepancy between JML and Nikon lenses may depend on the fact that we don't know where the exit pupil is located. I mounted it with its rear (open) barrel end flat against the objective revolver.

[elf]

but I still cannot measure it exactly because of all the fat microscope stuff in-between. The modified trinocular head I use for focusing contains no optics when switched to photo mode.

Can you measure it without the camera? I'd try inserting a 250mm dowel down the tube (taking precautions not to damage the objective), then measure the length of the dowel that's inside the tube.

[rjlittlefield]

I'd try inserting a 250mm dowel down the tube (taking precautions not to damage the objective), then measure the length of the dowel that's inside the tube.

Remove the objective too, so there's just an empty tube. That makes it easy to tell where the shoulder of the mount is, instead of having to factor in the length of the objective's threads.

Be sure to use a clean dowel, though. The last one I grabbed from my shop turned out to have a frightening amount of fine sawdust on it.

--Rik

[enricosavazzi]

Can you measure it without the camera? I'd try inserting a 250mm dowel down the tube (taking precautions not to damage the objective), then measure the length of the dowel that's inside the tube.

Hi Elf,

yes, that is pretty much what I did (actually with a length of plastic-clad copper wire, not to scratch anything). It is not a precise measurement because the wire bends slightly and is also difficult to hold exactly vertical, but the measurement is close enough.

Another problem is that the length of the photo tube must be visually calibrated to be parfocal with the eyepieces, so I cannot set it exactly to a given length. But at least it's in the ballpark, and I cannot make the optical path within the modified trinocular head any shorter without another major metal surgery and without bringing moving parts of the head dangerously closer to the camera (the head is from a Nikon Labophot and of the type that switches to camera mode by pushing the binoviewer to the left).

I am attaching a (quite cluttered) shot of the current setup below.