Judging high-NA infinity-corrected objectives

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

Moderators: rjlittlefield, ChrisR, Chris S., Pau

rjlittlefield
Site Admin
Posts: 23561
Joined: Tue Aug 01, 2006 8:34 am
Location: Richland, Washington State, USA
Contact:

Post by rjlittlefield »

Lou Jost wrote:As for the theory, I am also not sure, I hope Rik can illuminate the issue.
I'm not very good on theory of condensers. I'm not sure there is any good and simple theory, considering the number of different formulas for "effective" NA when the condenser's NA is less than the objective's. Then of course highest resolution can be different from best seeing. If that were not the case, there would be no use for oblique, let along the Gradient Universal Filter. But anyway...

As Pau says, the key thing is that the light from each point of the subject fully fills the rear aperture of the objective.

The entrance cone half-angle is just arcsin(NA), so at NA 0.75, it only takes 49 degrees off-axis to fill the aperture.

I expect the milk-immersed hemisphere will have no trouble doing that. In any case it's easy enough to check by just looking down the tube or at the exit pupil of the eyepiece.

--Rik

Lou Jost
Posts: 5944
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

The thing that's special about my milk condenser is that any point on the subject receives almost exactly the same amount of light from every direction (below the horizontal) if I ignore reflection off the cover slip that the subject rests on. I think that's a good thing...

JohnyM
Posts: 463
Joined: Tue Dec 24, 2013 7:02 am

Post by JohnyM »

With diffused only light, you can fill apertures up to 0.95
That means, it's not hard to dia light dry objectives with no optics. I myself use flat led panel with extra diffusion sheet for slide scanning.

Macro_Cosmos
Posts: 1511
Joined: Mon Jan 15, 2018 9:23 pm
Contact:

Post by Macro_Cosmos »

Hey Lou,

I'm not sure how I am going to approach the statements here. A 20x objective pushed to 40x cannot adequately replace a 40x. This is like comparing a 100mm f/1.4 lens with a teleconverter made into 200mm f/2.8 to a native 200mm f/2.8. The former is generally bound to be worse. If we control everything, that is say we have $2000 to make a 100mm f/1.4 and the teleconverter, compared to $2000 to just make a 200mm f/2.8, the 200mm f/2.8 is going to be better at 200mm. Likewise, if we push the 200mm down to 100mm, the 100mm is better.

I doubt the 20x pushed to 40x can outperform a native 40x/0.65. If it's 0.45 -- maybe. Since we're not talking about the image circle, I won't bring it up.

Also, the resolving limit of a native 40x will outdo that of a 20x pushed to 40x. This basically means one could squeeze more resolution out of native objective than one that's pushed out of its comfort zone.

I guess the advantage is kind of a "jack of all traits", one has a 20x objective that can be used at 40x if required. A native 40x/0.75 pushed down to 20x/0.75 won't outperform a native one. The 20x can also be pushed down to 10x; this is more useful than pushing the 40x to 80x, where weird stuff such as thermal expansion suddenly could become an issue.

Hope this makes sense.

JohnyM
Posts: 463
Joined: Tue Dec 24, 2013 7:02 am

Post by JohnyM »

No, inscribed magnification means only how it's performing on dedicated scope.
In fact most manufacturers provide various focal lenghts of tube lenses. Ie: nikon have 100mm and mitutoyo have 400mm.

It's very likely that 20x will outperform 40x, just because we'll use much smaller part of its image circle, than we do with 40x.

ChrisR
Site Admin
Posts: 8668
Joined: Sat Mar 14, 2009 3:58 am
Location: Near London, UK

Post by ChrisR »

With a flat panel, the light amplitude will decrease with increasing angle of incidence from the axis so it won't be evenly lit. Cos^-2 or whatever. If the diffuser isn't perfect it'll be worse.
Lou''s globe should fix that.
Chris R

Pau
Site Admin
Posts: 6052
Joined: Wed Jan 20, 2010 8:57 am
Location: Valencia, Spain

Post by Pau »

Macro_Cosmos wrote:I doubt the 20x pushed to 40x can outperform a native 40x/0.65. If it's 0.45 -- maybe. Since we're not talking about the image circle, I won't bring it up.
A 40/0.65 is an achromat or plan achromat, while the 20/0.75 is a modern plan apo and we are getting only the best part of the image circle. It will outperform the 40X for sure
Macro_Cosmos wrote:Also, the resolving limit of a native 40x will outdo that of a 20x pushed to 40x. This basically means one could squeeze more resolution out of native objective than one that's pushed out of its comfort zone.
I don't understand it. Resolution (measured at the subject side as it should be) is dependent of NA not magnification. With infinite corrected objectives magnification can be changed with the tube lens FL as others have pointed. No problem here except about coverage when you go downwards.
Macro_Cosmos wrote:The 20x can also be pushed down to 10x; this is more useful than pushing the 40x to 80x, where weird stuff such as thermal expansion suddenly could become an issue.
I don't understand the role of thermal expansion here

At another thread*, the test I proposed was about the effect of spherical aberration due to cover mismatch, not about magnification or coverage.
Would the resolution drop due to SA induced by the absence of cover glass make a high NA objective performing like a lower NA one not affected by this issue?
* http://www.photomacrography.net/forum/v ... 586#243586
Pau

Lou Jost
Posts: 5944
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Macro_cosmos, most of the things you've said are the very common misconceptions I tried to dispel in the first entry of this post. Most of your comments are mostly correct for finite objectives, but not for infinity-corrected objectives on tube lenses. As Pau has noted, and as Rik has noted elsewhere, the rated magnification is not the "native magnification", it's just the rated magnification when you use a 200mm tube lens (or 180mm for Olympus). You can use shorter or longer lenses without penalty, as long as the FOV and NA allow it.

You are right about this:
A native 40x/0.75 pushed down to 20x/0.75 won't outperform a native one.


That's because the size of the good image circle is limited on most objectives rated at 40x compared to objectives rated at 20x. But it is that image circle size that matters, not the magnification. And that's why I was talking about pushing up rather than down. Image circle considerations always improve on pushing up, but can cause problems when pushing down.

I will reiterate what I have noted in other posts: the relevant parameters for an infinity-corrected objective are its NA and its FOV, not is m. There's literally nothing special about the rated m, except as an arbitrary standard that gives clues about its designed FOV. A 20x 0.75 on a good 400mm tube lens theoretically will indeed outresolve a 40x 0.65.

In practice of course some lenses will be dogs.

Lou Jost
Posts: 5944
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Let me put it another way:

Infinity-corrected objectives (and the reversed front lenses of stacked combos)
don't have intrinsic magnifications. They have intrinsic FOV and NA.

Like you, I used to think 200mm tube lenses were special. Further study, tests, and nudges from Rik and others, have shown that they are not.

Pau
Site Admin
Posts: 6052
Joined: Wed Jan 20, 2010 8:57 am
Location: Valencia, Spain

Post by Pau »

Lou Jost wrote:Let me put it another way:

Infinity-corrected objectives (and the reversed front lenses of stacked combos)
don't have intrinsic magnifications. They have intrinsic FOV and NA.
True, just to point a bit more:
Sensor coverage (image circle) also depends of the tube lens FL and sensor size.
FOV on the subject can also be limited by the secondary magnification (tube lens and/or eyepieces and other lenses in the optical train) and sensor size but the maximum attainable quality FOV is intrinsic of the objective design.

Seeing it from another perspective a 20/0.75 is like a 40/0.75 but with much wider FOV. Resolution will be the same but sharpness at the sensor will be much improved when used a 20X

With finite objectives secondary magnification also changes magnification and sensor coverage.
Pau

Lou Jost
Posts: 5944
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Yes, and to hammer this home, the images of a given subject formed by a 10x infinity-corrected objective on a FF camera with a 200mm tube lens, or on an MFT camera with a 100mm tube lens, or on a 6cm x 7cm camera with a 400mm tube lens, are all identical (assuming you don't move the objective, and assuming you use good tube lenses, and assuming the pixel dimensions of the sensors are identical).

Some formats have denser sensors than others, and those can give sharper final images than sensors with fewer pixels. But that's not the lens' fault.

Macro_Cosmos
Posts: 1511
Joined: Mon Jan 15, 2018 9:23 pm
Contact:

Post by Macro_Cosmos »

Pau wrote: A 40/0.65 is an achromat or plan achromat, while the 20/0.75 is a modern plan apo and we are getting only the best part of the image circle. It will outperform the 40X for sure
Not sure about this. I didn't mention the structure of it, perhaps it was unclear. Let's control everything and make it an infinity apo objective. The mit 50x has an NA of 0.55 or something, not even 0.65. Welp, trade-off for the WD.
Pau wrote: I don't understand it. Resolution (measured at the subject side as it should be) is dependent of NA not magnification. With infinite corrected objectives magnification can be changed with the tube lens FL as others have pointed. No problem here except about coverage when you go downwards.
I was conflating sharpness and resolution, so I was wrong here.
Pau wrote: I don't understand the role of thermal expansion here.
Yeah that was brought out of nowhere. It was probably 4am when I posted the rather incoherent response. I was thinking about the thermal problems of specimens when one reaches very high magnifications. I've seen the problem demonstrated, hence the "could". I've never tried myself however.
Lou Jost wrote:Macro_cosmos, most of the things you've said are the very common misconceptions I tried to dispel in the first entry of this post. Most of your comments are mostly correct for finite objectives, but not for infinity-corrected objectives on tube lenses. As Pau has noted, and as Rik has noted elsewhere, the rated magnification is not the "native magnification", it's just the rated magnification when you use a 200mm tube lens (or 180mm for Olympus). You can use shorter or longer lenses without penalty, as long as the FOV and NA allow it.
Yeah I was conflating the concept of sharpness with resolution here. Of course how much "resolution (defined in my perspective)" one can get will depend on technique and lots of other issues, but that's just me confusing myself. I like to think of resolution in the sense of electronics (ADCs and etc), a rather silly false equivalence to get myself trapped in since optics is entirely different.

The unrelated problem here would be the availability of such tube lenses. Plenty 200mm tube lenses can be found, but I've never came across dedicated 400mm tube lenses. I'm sure they do exist somewhere. Meanwhile, one would have to use lenses designed for cameras. The 400mm f/5.6 nikkor you metioned is kind of a rarity. It's not expensive but kind of difficult to come by. Nikon also has an f/3.5 version which is rather expensive (and a great lens optically).

Lou Jost
Posts: 5944
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

"Yeah I was conflating the concept of sharpness with resolution here"
I don't think it makes a difference here. Maybe I'm not understanding the distinction you are making. A 20x 0.75 lens should give sharper images and more resolution at 40x than a 40x 0.65 (if both are infinity-corrected).
"The 400mm f/5.6 nikkor you metioned is kind of a rarity."
That may be true, it is not as cheap as some of the 200mm lenses we often use as tube lenses. But the difference in price needs to be weighed against the added cost of buying a 40x 0.65 apo lens if you already have the 20x 0.75, and the results with the 20x on the 400mm lens will be better than if you had bought the 40x 0.65 and stuck it in a 200mm lens.

Many people on this forum successfully use longer tube lenses than 200mm, including some older models that can be bought cheaply.

ChrisR
Site Admin
Posts: 8668
Joined: Sat Mar 14, 2009 3:58 am
Location: Near London, UK

Post by ChrisR »

Just being difficult, here ;) - It's possible that a 40x NA 0.75 would be better than a 20X 0.75 because it didn't have to be designed to cover such a wide field. Aberrations could therefore have been be easier to control.
Chris R

Lou Jost
Posts: 5944
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Chris, that's a good point.

Post Reply Previous topicNext topic