300mm tube lens?

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

Moderators: rjlittlefield, ChrisR, Chris S., Pau

JKT
Posts: 424
Joined: Fri Oct 28, 2011 9:29 am
Location: Finland
Contact:

Post by JKT »

zed wrote:because we all use cameras with sensor resolutions that mostly exceed the objective’s resolving power
That is precisely the point - the objective resolution is too good compared to the camera sensor. 20x/.75 objective has 0.36 um x 20 = 7.2 um resolution at sensor with nominal tube length and the RP pixel spacing is 5.75 um.

Additionally I suspect the image circle is not sufficient, though I could be wrong about that. The objective arrived before the adapter.

In any case, 300mm tube will help with the latter issue, while the camera pixel spacing remains still just under 2 pixels for the objective resolution keeping the combination sensor limited despite the magnification.

Lou Jost
Posts: 5987
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

I just now grabbed a dust mite and sandwiched it between two coverslips and ran a stack with my 400mm Nikkor ED and 20x 0.75 VC on an S1R FF camera.. I'll post the results here shortly.

JH
Posts: 1307
Joined: Sat Mar 09, 2013 9:46 am
Location: Vallentuna, Stockholm, Sweden
Contact:

Post by JH »

Lou Jost wrote:I just now grabbed a dust mite and sandwiched it between two coverslips and ran a stack with my 400mm Nikkor ED and 20x 0.75 VC on an S1R FF camera.. I'll post the results here shortly.
Looking forward to that. 400mm tube lens cleared my corners 300mm did not. I could not resist trying it without a coverslip, and got decent results without a coverslip using the 300mm tube lens on a bellow.

Best regards
Jörgen Hellberg
Jörgen Hellberg, my webbsite www.hellberg.photo

Lou Jost
Posts: 5987
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

I had problems with my mite, the stack was ruined because it wasn't dead and moved its legs....

Lou Jost
Posts: 5987
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Looks like it was dead by the time I did the stack with the 300mm. Here is the whole scene (this is processed and sharpened):

Image

And here a 100% crop of its setae:
Image

zed
Posts: 72
Joined: Thu Feb 07, 2019 1:35 pm
Location: Houston, TX
Contact:

Post by zed »

Let’s assume for a second that you have a 47MP camera (like Lou’s) and a 20x/0.75 objective. For this lens - assuming a white light source:

0.61(500*) / 0.75 = 407nm is the limit of lateral resolution for this optic.

* - this value is 1/2 the usable visible spectrum (350nm-650nm) and should increase if you use coverslips that are not 170 microns thick AND/OR mount in solutions with an RI less than 1.51. If you do not use a coverslip it gets even worse - I would use 1000nm to approximate a no coverslip condition (but even this is guesswork based on what optics you are using) - just avoid doing this is the best option because everything is toast at this point; resolution, sharpness and contrast.

407(20) / 2000 = 4.07 micron pixel size assuming we use 2X sampling for Nyquist.

If you focus your tube lens to the focal length required by the objective (200mm for Nikon) then the field of view for a full frame camera (36mmx24mm) would be 1.8mm x 1.2mm - therefore you would satisfy Nyquist with a frame size of 8845x5897 or 52Mp.

If you then increase the focal length of the tube lens it is true that you increase the projected size of the intermediate image plane - however you don’t increase your resolution because you have already satisfied Nyquist at 200mm with this camera and doing so would enter into empty magnification. So you can increase the magnification if you want - but you don’t improve the resolution and ultimately diffraction can reduce the sharpness and contrast that can negatively influence stacking.

Now if you are using a camera in the 24Mp range that is full frame I can see why one would think you can just increase the focal length - and of course you can as many of you have demonstrated. But this too is not without cost. The introduction of additional optics significantly reduces the throughput which translates to much longer exposure times - in the seconds range for continuous light - but then again you can always try to throw more light at it using high powered flash.

The point in my original post was not to state that your approach doesn’t yield a usable image. It was that if you’re already sampling at Nyquist there is no benefit going higher just for the field of view coverage. Cropping in post or using APS mode achieves the same result without having to use such a massive tube lens. In addition you have to be careful about calculating tube length to cover the field of view without taking into account pixel density.

Lou Jost
Posts: 5987
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

Your initial claim was that it was not a good idea to use a tube lens longer than 200mm. That's just wrong; if an image is good on APS using a 200mm tube lens, the image is at least as good using a 300mm tube lens on a FF camera with the same or greater number of pixels. Can you agree to that?

The real criterion for maximum tube lens focal length for a given objective depends on the sensor and the objective's resolution. The limit of empty magnification might be 200mm, it might be 300mm, it might be 400mm. Depends where severe over-sampling begins, and depends on the field number of the objective. There is nothing very special about 200mm in this. Experience shows that 200mm does well with typical objectives on APS sensors. That immediately implies that 300mm tube lenses would make a good standard on FF sensors.

When do we start getting empty magnification? Putting two pixels under each resolvable feature is not enough. It does not take into account the reduction in effective sensor resolution due to the Bayer array. Also, three pixels under a linear feature is generally thought to improve captured resolution.

JKT
Posts: 424
Joined: Fri Oct 28, 2011 9:29 am
Location: Finland
Contact:

Post by JKT »

zed wrote:The point in my original post was not to state that your approach doesn’t yield a usable image. It was that if you’re already sampling at Nyquist there is no benefit going higher just for the field of view coverage.
Not quite. What you wrote is that it is never useful to do what I suggested.

In this particular case I'm not quite at Nyquist limit even with the 300 mm tube lens ... or maybe I am - depends on whether the resolution formula uses 0.61 or 0.5. As Lou wrote I should still see improved detail with 400 mm tube lens, though not in direct proportion to increased focal length.

As to your mention of additional optics on the path reducing image quality. I don't quite see what the additional optics is if I choose lower diopter tube lens.


Thanks for the reference picture Lou! If the results are considerably worse, I will know I have a problem somewhere.

JH
Posts: 1307
Joined: Sat Mar 09, 2013 9:46 am
Location: Vallentuna, Stockholm, Sweden
Contact:

Post by JH »

Back to the initial question.
My apo ronar 360 mm is nice and was not expansivt. But needed a lot of stepping rings and tubes to adapt. The nikon 300 needs a piece of tape to hold focus at infinity using it vertical. The nikon 400 might be to expensive. Another option is to use a teleconverter.

Hope this helps
Jörgen
Jörgen Hellberg, my webbsite www.hellberg.photo

JKT
Posts: 424
Joined: Fri Oct 28, 2011 9:29 am
Location: Finland
Contact:

Post by JKT »

Yeah - using just a lens will require quite a lot of tubes.

I thought about the teleconverter, but I'm sceptical about the result. As a matter of fact - extremely sceptical with the teleconverters I have. :wink: I might try them for fun, though.

Lou Jost
Posts: 5987
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

I use the Nikkor ED-IF lenses which has internal focusing, so the tape is not needed. Some people report the image quality is better (less purple fringing) on the non-IF lenses, so the investment in tape might be worth it.

I have indeed noticed some purple fringing on some images.

zed
Posts: 72
Joined: Thu Feb 07, 2019 1:35 pm
Location: Houston, TX
Contact:

Post by zed »

Lou Jost wrote:Your initial claim was that it was not a good idea to use a tube lens longer than 200mm. That's just wrong; if an image is good on APS using a 200mm tube lens, the image is at least as good using a 300mm tube lens on a FF camera with the same or greater number of pixels. Can you agree to that?
My initial comment was that it is not practical to slap a giant 400mm camera lens and camera onto an objective. But as long as your sensor resolution does not surpass Nyquist then this is can be attempted, however the example you gave for justification would over sample with a 400mm tube lens - which is what I don’t agree with. In addition your example of sandwiching a specimen between coverglass in air introduces a severe RI mismatch which will reduce these theoretical numbers significantly. Keep in mind that these numbers are at best valid at the interface between coverslip (facing specimen) and mounting media. The further away you focus from this interface the worse these values get - even with a close RI match.
Lou Jost wrote:When do we start getting empty magnification? Putting two pixels under each resolvable feature is not enough. It does not take into account the reduction in effective sensor resolution due to the Bayer array. Also, three pixels under a linear feature is generally thought to improve captured resolution.
For a color camera I would tend to agree that 3X is a better estimate for sampling frequency for scientific grade imaging. Even if you assume a 3x factor (which is quite frankly overkill for people not making diffraction limited resolution measurements from the imaging data), the maximum array size for a 400mm tube lens in these conditions is 6639x4426 - so you are still oversampling with an S1R.
JKT wrote:Not quite. What you wrote is that it is never useful to do what I suggested.
What I originally said was that I cannot think of a practical reason why this would be useful - but if you are under sampling with a 200mm tube lens and want to increase the magnification to improve the corners in your image then this is fine as long as you don’t over sample. If you over sample then you might as well just crop the image from the 200mm rather than completely change out your tube lens. These days I tend to assume people are imaging with a camera in the 40-50Mp range - and with these cameras it just not practical to change out the tube lens when using a variety of different objectives.
JKT wrote:In this particular case I'm not quite at Nyquist limit even with the 300 mm tube lens ... or maybe I am - depends on whether the resolution formula uses 0.61 or 0.5.
Depends on your camera - which you don’t state. The Rayleigh limit (0.61) is the generally accepted estimate in microscopy - but the delta between Sparrow and Rayleigh is negligible for such low NA optics.
JKT wrote:As to your mention of additional optics on the path reducing image quality. I don't quite see what the additional optics is if I choose lower diopter tube lens.
Using a camera lens as a tube lens will always be less transmissive than simpler optics such as a Raynox or a Thor TTL-200. This translates into longer exposure times - and I don’t know about you guys but even with a continuous LED at 1100 lumens my exposures tend to hover in the 1/2 - 1 second range at ISO 100. Adding a considerable more hefty tube lens would introduce more potential for vibration artifacts.

JKT
Posts: 424
Joined: Fri Oct 28, 2011 9:29 am
Location: Finland
Contact:

Post by JKT »

Actually I did mention the camera (RP), but only in the second post. The word "average" in the first could be misleading as I meant it as average over all DSLRs, not for FF only. For the latter 26 MP is pretty lousy.

As far as extra optics go, I was asking for tube lenses, wasn't I. :) Century Precision optics seems to have +4, +3.5 & +2.6, though the latter two are quite large and expensive. The first and +7 are known to be good, so there is a chance that the other two might be too.

With my three flash system I'm not too worried about the exposure time.

Oh well - we seem to have reached an agreement. :)

zed
Posts: 72
Joined: Thu Feb 07, 2019 1:35 pm
Location: Houston, TX
Contact:

Post by zed »

OK - You have to give me more than 'RP' :) - what camera is that?

Good luck with everything!

Lou Jost
Posts: 5987
Joined: Fri Sep 04, 2015 7:03 am
Location: Ecuador
Contact:

Post by Lou Jost »

My initial comment was that it is not practical to slap a giant 400mm camera lens and camera onto an objective.
As I said, it makes no difference to my set-up whether I use a 135mm or 400mm f/5.6 lens. I've even used a giant white Canon 300 L f/2.8 fluorite lens as a tube lens on this rig (though that is probably near the weight limit). Also, sometimes being heavier actually helps against vibrations. Depends on how the lens is mounted and the frequency of vibrations.
For a color camera I would tend to agree that 3X is a better estimate for sampling frequency for scientific grade imaging. Even if you assume a 3x factor (which is quite frankly overkill for people not making diffraction limited resolution measurements from the imaging data), the maximum array size for a 400mm tube lens in these conditions is 6639x4426 - so you are still oversampling with an S1R.
The Nyquist and 3x criterion apply to monochrome images. Factor in the loss of resolution due to the Bayer array and the number of needed pixels would approximately double. The S1R with 400mm tube lens is not oversampling in this example.

You didn't comment on my main observation. If 200mm is a normal and useful tube lens size for APS, then approx 300mm gives you that same aerial image, with same FOV, on the FF sensor. There is nothing speical about the 200mm focal length except that it gives the nominal magnification on the sensor. Over- or under-sampling can happen with any size tube lens.
with these cameras it just not practical to change out the tube lens when using a variety of different objectives
Maybe this is why we think so differently about the role of tube lenses. In my set-up there is not much difference between changing tube lenses vs objectives. But if it was hard to change tube lenses, again the standard for FF should be 300mm if 200mm is your standard for APS. Why would someone invest in a FF camera if the intention was to always crop the image to APS size?
Using a camera lens as a tube lens will always be less transmissive than simpler optics such as a Raynox or a Thor TTL-200.


This seems nit-picky. The difference in T vs f number for modern lenses is very small, and close-up diopters often have less sophisticated coatings than camera lenses, so the difference in transmission between camera lenses and close-up lenses is usually going to be negligible (and might even favor camera lenses sometimes).
Last edited by Lou Jost on Tue Feb 25, 2020 7:06 pm, edited 2 times in total.

Post Reply Previous topicNext topic