Pushing down objectives... how low can you go?

[heartprairie]

"Koppace" 20x/0.34 on Lensbaby Velvet 56mm f/1.6 (focus @ infinity, non-reversed) on APS-C camera yields 5.6x magnification. (post initially stated 9x, which was incorrect)
Planarity is lost, though the Velvet 56 isn't planar in the first place. Center resolution remains good.

Unlike most photography lenses, the Velvet 56 has a similar apparent pupil size when held reversed. Perhaps this is is beneficial.
The Konica Macro-Hexanon 105mm f/4, a bellows-type lens which I use on full-frame, is like this too.

Samples with Velvet 56

Crop of above image. Fun fact: those 'gray' subpixels are white when fully on.

[rjlittlefield]

I think I'm missing the point.

Usually when people ask how far they can push down an objective, they mean "while retaining decent image quality at least across the entire height of the frame".

There's no value in pushing farther down than that, except for the unusual case where you want to show some wider context and simply don't care how blurred it is.

But in the case that you've shown here, it looks like only the extreme center of the image retains decent quality. At the resolution posted, image #2 looks like it might be OK for about 1/2 the frame height. Then in the tighter crop of image #3, we see that the sharp area is even much smaller than that, maybe only 15% of the total frame height.

To be brutally honest, these images look to me like junk.

So I am confused. What are you trying to accomplish? What message do you take away from these images?

--Rik

[Scarodactyl]

I guess the big question is whether those edges are just non-planar or if they are also distorted when in focus. I'd guess the latter but it can go both ways.

[heartprairie]

I think I'm missing the point.

I have several other lenses of similar focal length and only this one was able to fill the frame.
Unfortunately it has very poor performance when it comes to planarity.
I am certain some members here would have lenses without this deficiency and am eager to see what others can accomplish.

Usually when people ask how far they can push down an objective, they mean "while retaining decent image quality at least across the entire height of the frame".
There's no value in pushing farther down than that, except for the unusual case where you want to show some wider context and simply don't care how blurred it is.

Besides showing context, merely seeing a wider view can be useful in identifying a region one would like to image at higher magnification.
It is novel to swap the tube lens rather than objective to accomplish this.
I hope someone will present a combination better suited to imaging.

But in the case that you've shown here, it looks like only the extreme center of the image retains decent quality. At the resolution posted, image #2 looks like it might be OK for about 1/2 the frame height. Then in the tighter crop of image #3, we see that the sharp area is even much smaller than that, maybe only 15% of the total frame height.

It is expected that a single frame will have a limited area in focus due to non-planarity.

To be brutally honest, these images look to me like junk.

To me, they're a lot better than I expected of a soft focus lens.

So I am confused. What are you trying to accomplish? What message do you take away from these images?

I would like to further discussion around using lenses of relatively low focal length as tube lenses.
From the images I shared, the specific combination I utilized would be best suited for looking across a subject, rather than imaging it.

[rjlittlefield]

I hope someone will present a combination better suited to imaging.

Ah, OK!

I hope that other people will add more examples, but to start the list with some of my own postings:

• viewtopic.php?t=15876 "Nikon CFI 10X Plan Achromat on Canon 100 mm macro lens, giving 5X. This combo is much sharper than the MP-E 65, not surprising since the MP-E is running around f/16 effective while the combo is about f/10. What is surprising is that the combo holds up well clear to the corners, despite that this far exceeds the field size that the objective was designed to handle."
• viewtopic.php?t=16348 "Shot with a Nikon objective, CFI BE 10X NA 0.25 WD 6.7 (part number MRN70100) used at 5X in combination with a Canon 100mm f/2.8L IS USM as tube lens, on a Canon T1i camera, sensor size 15.1 megapixels (4752 x 3168) on 22.3 x 14.9 mm." (Sharp clear into the corners.)
• viewtopic.php?p=103825#p103825 "using a 25X NA 0.55 lens configuration consisting of a Nikon CF Plan 50X NA 0.55 inf/0 EPI ELWD (8.7 mm) with tube lens Canon 100mm f/2.8L IS USM macro."
• viewtopic.php?f=27&t=47326 "Shot at roughly 5X using Mitutoyo M Plan Apo 10X NA 0.28, on a Canon EF Macro 100 mm f/2.8 IS USM lens on Canon R7 body, using in-camera focus bracketing"

So, three different 10X objectives pushed down to 5x, and a 50X objective pushed down to 25X. I don't have an example handy, but the Mitutoyo 20X also pushes down nicely to 10X on APS-C. It actually makes the sharpest 10X in my kit.

Using shorter than normal tube lens to give reduced magnification is well known practice. In fact one of the first postings here at PMN which evaluated the Raynox DCR's for tube lenses, back in 2012, tested the Raynox DCR-250 and concluded "very good corners at 6.2X on APS-C with the Nikon CFI 10/0.25". That is at viewtopic.php?t=18145 .

You mentioned "Unlike most photography lenses, the Velvet 56 has a similar apparent pupil size when held reversed. Perhaps this is is beneficial." Pupil ratio is less important than pupil location. Lenses that have their entrance pupils far back tend to vignette when used as tube lenses. This is a common problem that typically limits the use of zoom lenses as variable magnification tube lenses. The Canon 100mm f/2.8L IS USM that I used for the above postings works particularly well because at infinity focus its pupil is far forward. Most enlarger lenses also work well in this regard.

It is expected that a single frame will have a limited area in focus due to non-planarity.

Fair enough, but then usability of the outer regions can be tested by focus stacking. A good example is at viewtopic.php?p=57815#p57815 , where two finite 10X objectives are short-focused down to 6.9X and one of them still works well while the other does not.

Shooting single frames that are focused in the corners can also be used, but the test/demo is not so convincing because there are common aberrations that will prevent good focus stacking even though each individual frame looks OK.

Is this helpful?

--Rik

[heartprairie]

For ~100mm, I already have a 105mm lens that looks quite promising, just awaiting parts for achieving proper extension.

Is there a formula for calculating the resultant magnification with a shorter tube lens?
With my LU Plan Fluor 10x on the Velvet 56, I'm getting around 3x magnification, which does not seem proportional to the outcome with my 20x objective.

Unsurprisingly, resultant image quality is very poor.
The 'D' in 'DIV' is clearly being stretched out of shape, similar to the aberration in your example of a bad pushing down combination.
Have attached comparison image from Laowa 100mm with 1.4x TC.

[rjlittlefield]

Is there a formula for calculating the resultant magnification with a shorter tube lens?

When the rear lens is a well corrected rectilinear and is focused on infinity, final magnification scales simply in proportion to the focal length of the rear lens.

So, if the objective gives 10X with a 200 mm tube lens, then it will give 5X with 100 mm, 6.25X with 125 mm, and so on.

If the rear lens is not focused on infinity or is not rectilinear, then magnification is difficult to predict.

You wrote:

With my LU Plan Fluor 10x on the Velvet 56, I'm getting around 3x magnification, which does not seem proportional to the outcome with my 20x objective.

I don't know your setup or results well enough to interpret that.

One thing I don't see mentioned is that different lines of objectives can be designed to use different length tube lenses. Mitutoyo and Nikon are mostly rated for 200 mm, but I recall 180 mm for some other brands that I don't have.

--Rik

[rjlittlefield]

Switching subjects for a moment...

Crop of above image. Fun fact: those 'gray' subpixels are white when fully on.

This is the first time I have seen a pattern like that.

What sort of device has this screen?

--Rik

[Scarodactyl]

The most important thing to test when pushing down an objective is how the edges/corners perform at their best focus. Nonplanarity is no big deal if you're stacking and some high performance objectives sacrifice planarity for better IQ. But if the edges/corners are mush at any focus you should almost always just use a lower mag lens instead.

[heartprairie]

Switching subjects for a moment...

Crop of above image. Fun fact: those 'gray' subpixels are white when fully on.

This is the first time I have seen a pattern like that.

What sort of device has this screen?

--Rik

It's the rear screen of a Sony a7R III.

[heartprairie]

With my LU Plan Fluor 10x on the Velvet 56, I'm getting around 3x magnification, which does not seem proportional to the outcome with my 20x objective.

I don't know your setup or results well enough to interpret that.

One thing I don't see mentioned is that different lines of objectives can be designed to use different length tube lenses. Mitutoyo and Nikon are mostly rated for 200 mm, but I recall 180 mm for some other brands that I don't have.

--Rik

It's a modern Nikon objective, model number is MUE10100.
The relevant markings on the objective merely state ∞/0 (i.e. slipless), without specifying a focal length.
I think it's safe to assume it's intended to be used with a 200mm tube lens.
The Velvet 56 was focused to infinity without any additional extension, mounted directly on a Samsung NX2000 camera with APS-C sensor.

I have subsequently tested the aforementioned Nikon 10x objective on a different APS-C camera, with a different lens - specifically, a Tamron SP 85mm f/1.8.
This appears to perform much better with regard to planarity.
Again, lens is focused at infinity.
Resultant magnification is ~4.3x.

I now realize the magnification I had stated for the 20x objective with Velvet 56 was incorrect.
If the focal length of the objective is known, it appears magnification is simply: tube lens focal length / objective focal length
Alternatively, if native tube lens focal length is known: tube lens focal length / native tube lens focal length * native magnification

1024px center crop

[rjlittlefield]

Those images are looking much better.

If the focal length of the objective is known, it appears magnification is simply: tube lens focal length / objective focal length
Alternatively, if native tube lens focal length is known: tube lens focal length / native tube lens focal length * native magnification

Correct.

Manufacturers usually do not quote focal length for an objective, but the algebra works to calculate that as simply
objective focal length = tube lens focal length / magnification

For much analysis you can imagine that an infinity objective is just an ordinary lens, reversed, and for that purpose you can also calculate its nominal f-number as
f-number = 1/(2*NA)

As example, your LU Plan Fluor 10x NA 0.30 can also be analyzed as a 20 mm f/1.67 lens, reversed.

The big difference is that an ordinary 20 mm lens would be compromised to work sort of uniformly OK over an area the size of the whole sensor, when stopped down quite a bit, while the microscope objective is optimized to work really well over a much smaller field, wide open.

One final formula, handy for thinking about resolution, is that the effective f-number, as seen by the sensor, is just
effective f-number = magnification / (2*NA).

So, when operated at 10X your NA 0.30 objective gives effective f/16.7, but when pushed down to 5X it gives effective f/8.33 .

This is why my sharpest 5X is a 10X pushed down --- my best nominal 5X objective is a Mitutoyo M Plan Apo 5X NA 0.14 objective that gives effective f/18, while my 10X NA 0.28 pushed down gives effective f/9.

--Rik

[rjlittlefield]

Adding to the list of combinations that work well for imaging:

• viewtopic.php?f=27&t=47432 , "Mitutoyo M Plan Apo 20X NA 0.42 on Componon 135 mm f/5.6 as a tube lens. Nominal magnification would be 13.5X, measured is 13.6X".

--Rik

[heartprairie]

I decided to try a lens on M43 that had similar equivalent focal length to using the Velvet 56 on APS-C

Planarity looks better but vignetting is stronger, and distortion is quite strong. What's causing the distortion, I wonder.

'Koppace' 20x/0.34 on 7Artisans 35mm f/2 WEN (M mount) adapted to Panasonic GX8 (M43)

[Lou Jost]

That image is making a maddening optical illusion, with a white dot in each crossing of the black lines with each other, especially in my peripheral vision. I can't make it stop! Is anybody else going nuts looking at this?