Mitutoyo 100x/0.55 vs 50x/0.55

[nielsgeode]

Hi all,

What is the advantage of the Mitutoyo 100x NA 0.55 over the 50x NA 0.55? The NA is the same, so the amount of resolution should be the same? Especially with such high-magnifications, any modern sensor will far outperform the resolution of these lenses? This makes me think that you can just as well take a 50x and crop it in post, rather then spend a fortune on a 100x? Am I correct?

Cheers,
Niels

[enricosavazzi]

Hi all,

What is the advantage of the Mitutoyo 100x NA 0.55 over the 50x NA 0.55? The NA is the same, so the amount of resolution should be the same? Especially with such high-magnifications, any modern sensor will far outperform the resolution of these lenses? This makes me think that you can just as well take a 50x and crop it in post, rather then spend a fortune on a 100x? Am I correct?

Cheers,
Niels

Assuming that both objectives are diffraction-limited and have the same NA, there is no difference in the actual resolution on the subject side. However, the 50x objective should provide an absolute size of the field of view about twice as wide (i.e. 4 times the area) of that provided by the 100x.

In my opinion, the 50x in practice is more versatile than the 100x when either lens is coupled to a relatively high pixel-count sensor, because it gives you a (slightly) longer working distance as well as a wider native circle of view on the subject side, and leaves you the option to crop the image if you need a detail of the central portion of the image.

The 100x remains a useful choice when mounted on a videomicroscope with limited pixel count, or for observation by eye through a bi/trinocular head. In these cases, the jump from 20x to 100x would be excessive, and it makes sense to have 20x, 50x and 100x objectives.

[nielsgeode]

Thanks a lot! I own the 50x for some time now and I was offered a 100x. That's why I asked. However, I have the feeling my 50x does not perform according to specs. This is because it is só much worse than my Mitty 20x 0.42. How could I (approximately) test it?

[enricosavazzi]

[...]I have the feeling my 50x does not perform according to specs. This is because it is só much worse than my Mitty 20x 0.42. How could I (approximately) test it?

The M Plan Apo 20x has NA 0.42, which means a computed resolution on the subject side (0.5 lambda / NA) of 0.63 micron.

The 50x gives 0.48 micron, which is only slightly better. Coupled with 2.5 times higher magnification, this means a worse resolution on the image side than the 20x, so it is expected that the 50x will not be as good as the 20x, assuming both objectives are faultless. In practice you get an image resolution of 31.5 micron with the 50x, vs 12.6 micron with the 20x. On a typical sensor with 5 micron sensels, the difference is quite visible.

The 50x will also be somewhat more sensitive to loss of resolution if the light source is not well diffused (causing the utilized aperture to be lower than the nominal NA).

Comparing with a known-good objective is the easiest way to test, but if you get an image resolution substantially worse than the expected 31.5 micron (let's say you only get a 100 micron image resolution), you can suspect that something is wrong with the objective and/or illumination.

EDIT: The 100x objective gives you a (calculated) image resolution of 63 micron (about 12 pixels across), which is visibly blurry even without much pixel-peeping.

[Macro_Cosmos]

The 100x is the SL version, forget about it. You'll get the same if you push that 50x to 100x with a 400mm tube lens.

The 50x has an NA of 0.55, 20x being more than half of that has an NA of 0.42. Whilst offering a higher NA, it's not higher in resolution if you account for the 2.5x magnification.

Your bad 50x results could be:
1. Lens is damaged, ie decentered or something
2. Step distance. The DoF of the 50x is 0.9um, people commonly use 1um which isn't enough. Find a way to do 0.5um and you should see an improvement

[rjlittlefield]

2. Step distance. The DoF of the 50x is 0.9um,

Not really true. That number is what Mitutoyo quotes, but they specify their "Depth of focus" as maximum distance from perfect focus. It's only half what we usually mean by DOF, and what Nikon quotes as DOF for their objectives at same NA. The maximum step size for a remarkably well behaved subject (image quality limited only by diffraction) is twice as large as Mitutoyo's number, about 1.8 microns under the usual assumptions (total DOF = lambda/NA^2, lambda = 0.55 microns).

Find a way to do 0.5um and you should see an improvement

Often true, but the reason is not because of simple diffraction-limited DOF. It's because the "utilized aperture" effect makes 3D features appear to move laterally as focus changes. The direction of movement depends on both illumination and local surface angles of the subject, with the result that in many cases different features move in different directions and the subject appears to sort of "squirm around" as focus is changed. Using a smaller step size gives smaller lateral movement between steps, which can improve the quality of the composite image even when there's no significant difference in sharpness.

--Rik

[nielsgeode]

I have a new rail that can accurately do steps much smaller than 1 micron. I think 0.2 micron is not a problem at all. Today I shot stacks with my 20x at various step sizes and tomorrow I will do the same with my 50x. Given they are PixelShiftMultiShot Images at 42 MP (yes overkill), processing takes time. For the 50x I will not do the PSMS feature, but just shoot 42 MP single shots.

When they are finished I post them here so people can judge