Back focus using manual bellows

[ray_parkhurst]

When front-stepping, the magnification decreases somewhat for closer frames, while when rear-stepping the closer frames have a more significant increase in magnification. No mag change when moving both, but perspective does shift. Has anyone done a study of all three for a given subject at appropriate mag to compare the end results and process?

[rjlittlefield]

No mag change when moving both, but perspective does shift.

That depends on the objective. Many objectives are close enough to telecentric that they have no discernable shift of perspective and can be treated by software as if they were perfectly telecentric.

For 10X and above, I normally recommend turning off Scale adjustment, which essentially treats the objective as telecentric.

At 5X and below there can be relevant perspective shifts, but again it depends on the objective. The Mitutoyo 5X is so close to telecentric that it works fine with Scale adjustment turned off.

Has anyone done a study of all three for a given subject at appropriate mag to compare the end results and process?

I am not aware of one.

--Rik

[heartprairie]

You could try placing an iris diaphragm immediately behind the objective and narrowing the aperture slightly. I have tried this with a 10x objective and saw noticeably lesser fringing in a single frame test. They are readily available with M42 threading, search "M42 iris".

[bbobby]

@bbobby - yes, I think you may be correct about that section being not quite in focus.

...

Thanks for reminding me about my limit with the D90- I have been looking around for a new camera but due to my stock of Nikon F mount lenses I would prefer to stick with that mount which limits my choices. I have been considering a D810.

Keith

When I mentioned 4x Plan Apo in my previous post and later that the full frame camera not always is better when using it with objectives I wanted to give u example and check the drawers... Today I found the long forgotten Nikon 4/0.20 Plan Apo and decided to give it a shot... Camera used - Sony A7R4 (JPGs straight from the camera)... as other mentioned - amazing in the center, though it covers full frame the image is soft in the corners with lots of red/blue fringing and a bit of vignetting... holds better on crop size sensor of course BUT away of the center is not so perfect...
microscopyu.com shows resolution value for this objective 1.375 µm and it is capable of that in the center, Group 8 Element 4 is clearly resolved on the USAF target (1.38 µm) and with 4 shots pixel shift used to compensate for the Bayer filter I can see Element 5, which is 1.23 µm... why I am able to resolve more - no idea...

Last year found dead butterfly on a parking lot and today I tried picturing the wing. Stack of about 60 images, ~10 µm step.

[keithostertag]

Thanks for the examples, very good for me to see since I am pretty inexperienced and have not yet learned what to expect.

Yea, I've decided to stay with APS-C, but going with the Nikon D7200 - it seems like a sweet spot between price and performance. And yes, it is important to me to avoid fringing but I still want sharpness corner-to-corner. Not really concerned about going to mirror-less at this time.

Keith

[rjlittlefield]

microscopyu.com shows resolution value for this objective 1.375 µm and it is capable of that in the center, Group 8 Element 4 is clearly resolved on the USAF target (1.38 µm) and with 4 shots pixel shift used to compensate for the Bayer filter I can see Element 5, which is 1.23 µm... why I am able to resolve more - no idea...

I think you've fallen prey to a common misinterpretation of the numbers, conflating µm per cycle with µm per line width.

Here's the way I work the numbers, using the formulas discussed at viewtopic.php?p=124831#124831 .

Nikon's 1.375 µm corresponds to the bandwidth limit of NA 0.20 for 550nm light. That number means µm per cycle. It's the closest distance that two bright spots can be, and still have the lens see any darker space between them.

But according to the specs HERE (at the Technical Information tab), USAF 1951 Group 8 Element 4 is 362 line pairs per mm. That calculates to be 1.38 µm per line width, but it's 2.76 µm per line pair, which also means 2.76 µm per cycle. 2.76 µm is a lot bigger than 1.375 µm, so what you're seeing is well within the range of the objective.

Working the numbers, NA 0.20 should have no trouble resolving anything in Group 8. Even Group 9 element 3 is within reach, though barely.

So then the question becomes why your images don't show that, and the answer is "sensor resolution". Group 8 Element 6, which your images do resolve but just barely, is 456 line pairs per mm. With 4X magnification that becomes 114 cycles per mm = 8.77 µm per cycle, or 4.39 µm per sample at the Nyquist limit. Your camera has pixel size 3.76 µm, which is a little smaller than the Nyquist limit, but not small enough to give a high quality capture. So, your captured image is limited by sensor resolution.

Another way of reaching the same conclusion is to consider that 4X NA 0.20 is running at effective f/10. (That formula is Feff=magnification/(2*NA).) Under standard assumptions of 550 nm light, f/10 has a limiting resolution on sensor of 5.5 µm per cycle. Even at Nyquist limit, that would require 2.75 µm per sample, much smaller than your pixel size.

Edited to add: For one more confirmation, go to https://www.microscopyu.com/tutorials/m ... resolution . Set up Nikon's calculator for 4X, NA 0.20, video coupler magnification 1.0, and you'll see that it says "Required Pixel Size 2.8 µm ".

Because your image is sensor limited, It would be interesting to stick a teleconverter on your camera and see how much more of that currently unresolved resolution you could pick up.

--Rik

[keithostertag]

Rik-

Could you elaborate further... why would adding a teleconverter increase the resolution? Is that what you are suggesting?

Keith

[rjlittlefield]

Could you elaborate further... why would adding a teleconverter increase the resolution? Is that what you are suggesting?

Start with the discussion at viewtopic.php?t=21751 , "What does a teleconverter do and why might you use one?"

Then if still unclear, ask again.

--Rik

[bbobby]

...
Nikon's 1.375 µm corresponds to the bandwidth limit of NA 0.20 for 550nm light. That number means µm per cycle. It's the closest distance that two bright spots can be, and still have the lens see any darker space between them.
...
Because your image is sensor limited, It would be interesting to stick a teleconverter on your camera and see how much more of that currently unresolved resolution you could pick up.
--Rik

Oh, they mean per cycle/line pair with that number... I do know the difference, but I carelessly assumed it is the line width... by bad... Thank you for pointing that!
I see the other table where "Required Pixel Size" for "4x (0.20)" is 2.8... 2.8 pixel*2=5.6 for a pair/4 is 1.4
1.375/2=0.6875 line width corresponding to Group 9 Element 4 in a perfect world.
This time I used Canon FD bellows and I am not sure if the older FD teleconverters will be good enough... will give them a try... maybe better to mount it on Nikon bellows, I got the latest 7-element versions of the Sigma 2x teleconverters... and 16 shot pixel shift function may be helpful... and my Olympus got a bit smaller pixels so hopefully this weekend there will be some free time...

[keithostertag]

Rik-

Yes, the discussion at that link does explain it well, thanks much!

Keith

[bbobby]

...
Working the numbers, NA 0.20 should have no trouble resolving anything in Group 8. Even Group 9 element 3 is within reach, though barely.
...

Because your image is sensor limited, It would be interesting to stick a teleconverter on your camera and see how much more of that currently unresolved resolution you could pick up.

--Rik

Better late than never, some people say...
I saw today a 3-4 FD teleconverters collecting dust on a shelf and wanted to test my supposedly better vertical setup so... here it is few images of the same Nikon Plan Apo 4x NA 0.2 160/- with FD 2X-B Extender (to be exact) shot with Sony A7R4.
You are right, Group 9 Element 2 is clearly there.
This objective is great in the center (5-10 mm circle), but it got severe CA and astigmatism away from it (at least my copy).

Whole image with 2x teleconverter.

Crop with 2x teleconverter, JPG straight from the camera.

Crop with 2x teleconverter, from 4 images pixel shift, exposure +1 in PS, sharpness +15.

Crop with 2x teleconverter, from 16 images pixel shift, exposure +1 in PS, sharpness +15.

And this is crop from 16 images pixel shift without teleconverter.

[rjlittlefield]

Nice confirmation -- thanks!

--Rik