## Tube lens question

**Moderators:** rjlittlefield, ChrisR, Chris S., Pau

### Tube lens question

Hello! Long time lurker here and first post.

I got my hands on a couple of microscope objectives over a year ago and what started out with a simple question along the lines of "I wonder, if I can somehow put those on my camera?", has sent me to this great forum and down a deep rabbit hole, where I have now built a focusing rig and an Arduino-driven controller to go with it (that's maybe a story for another post).

Initially, I was using my camera's zoom lens as a tube lens, which worked reasonably well, but a few month ago, I found a set of cheap T-mount extension tubes that even included a lens (of unknown specifications). I was hoping the included lens would work out as tube lens, which it does...sort of.

The lens turned out to be a 25 mm diameter specimen and looked like an achromatic doublet. I eyeballed the focal length by projecting a sharp spot of a ceiling lamp (my best approximation of infinite parallel rays) onto the floor. FL was about 100 mm, so not ideal.

Next, I tested various combinations of tube length and infinity space, imaging a ruler with my Olympus 4x (infinty) objective (camera has an APS-C sensor):

Infinity Space [mm] Tube Length [mm] Magnification

10 180 4.0

30 180 **3.7

50 180 3.4

70 180 *3.2

110 180 2.7

30 60 2.7

30 110 3.2

30 160 **3.7

30 210 4.4

*Noticable vignetting at 70 mm infinity space

** Equal mag for 160 vs. 180 mm tube length, some measurement error here

Apparently, I'm getting 4x magnification, but usable infinity space is rather short. Also, Edmunds, Thorlabs etc. recommend an objective-tube lens distance of at least 70 mm, so overall, this doesn't seem ideal.

To make a long story short, I'd like to get a more suitable tube lens and was looking into achromats on Edmunds.com. Would this one (FL 175 mm) work well:

http://www.edmundoptics.com/optics/opti ... ses/32884/ ? Would you rather go for the 200 mm FL equivalent (I have three Olympus objectives and one Nikon (40x), so 180 mm vs. 200 mm optimal tube length)? Are there any better or cheaper alternatives around, that anyone could recommend?

Thanks!

I got my hands on a couple of microscope objectives over a year ago and what started out with a simple question along the lines of "I wonder, if I can somehow put those on my camera?", has sent me to this great forum and down a deep rabbit hole, where I have now built a focusing rig and an Arduino-driven controller to go with it (that's maybe a story for another post).

Initially, I was using my camera's zoom lens as a tube lens, which worked reasonably well, but a few month ago, I found a set of cheap T-mount extension tubes that even included a lens (of unknown specifications). I was hoping the included lens would work out as tube lens, which it does...sort of.

The lens turned out to be a 25 mm diameter specimen and looked like an achromatic doublet. I eyeballed the focal length by projecting a sharp spot of a ceiling lamp (my best approximation of infinite parallel rays) onto the floor. FL was about 100 mm, so not ideal.

Next, I tested various combinations of tube length and infinity space, imaging a ruler with my Olympus 4x (infinty) objective (camera has an APS-C sensor):

Infinity Space [mm] Tube Length [mm] Magnification

10 180 4.0

30 180 **3.7

50 180 3.4

70 180 *3.2

110 180 2.7

30 60 2.7

30 110 3.2

30 160 **3.7

30 210 4.4

*Noticable vignetting at 70 mm infinity space

** Equal mag for 160 vs. 180 mm tube length, some measurement error here

Apparently, I'm getting 4x magnification, but usable infinity space is rather short. Also, Edmunds, Thorlabs etc. recommend an objective-tube lens distance of at least 70 mm, so overall, this doesn't seem ideal.

To make a long story short, I'd like to get a more suitable tube lens and was looking into achromats on Edmunds.com. Would this one (FL 175 mm) work well:

http://www.edmundoptics.com/optics/opti ... ses/32884/ ? Would you rather go for the 200 mm FL equivalent (I have three Olympus objectives and one Nikon (40x), so 180 mm vs. 200 mm optimal tube length)? Are there any better or cheaper alternatives around, that anyone could recommend?

Thanks!

Welcome to the forum! Long story short, search the forum for "Raynox". Their 125mm (DCR-250) and 210mm (DCR-150) Focal Length close up lenses work pretty well. You'll need a few cheap adapter rings but the total outlay is modest.

Many primes between 100 and 200 mm, and some zooms for part of their range, have been found to be good, too.

[edited for clarity]

Many primes between 100 and 200 mm, and some zooms for part of their range, have been found to be good, too.

[edited for clarity]

Last edited by ChrisR on Tue Apr 28, 2015 12:30 pm, edited 1 time in total.

- rjlittlefield
- Site Admin
**Posts:**22312**Joined:**Tue Aug 01, 2006 8:34 am**Location:**Richland, Washington State, USA-
**Contact:**

jnh, welcome aboard.

That achromat from Edmund would probably work OK, after you finally got it mounted, and decided which side should face forward, and with what spacing between it and the objective.

But a much simpler method that is known to work well is to purchase a Raynox DCR-150 "closeup" lens and a small set of adapters so as to mount it as shown at http://www.photomacrography.net/forum/v ... 195#143195. (If you're using a full-frame camera then in order to avoid vignetting you'll need to replace the cone adapter with a flat one, so as to put the objective closer to the Raynox.)

See http://www.photomacrography.net/forum/v ... hp?t=23898 for a comparison of Raynox DCR-150 versus other purpose-made tube lenses from Mitutoyo, Nikon, and Thorlabs.

Regarding your table of numbers, I think that by "infinity space" you mean the air gap between objective and tube lens. If I've read that correctly, then the fact that magnification changes with separation means that the gap is not infinity space (with parallel rays). If it were infinity space, then changing the separation would have no effect except definitely the major issue of vignetting and perhaps some minor issue with aberrations away from center.

To adjust the tube lens properly, the trick is to remove the objective, point the camera plus tube lens at some distant object, and adjust extension behind the tube lens until the image seen by the camera is focused.

You asked about

I hope this helps. Again, welcome aboard!

--Rik

That achromat from Edmund would probably work OK, after you finally got it mounted, and decided which side should face forward, and with what spacing between it and the objective.

But a much simpler method that is known to work well is to purchase a Raynox DCR-150 "closeup" lens and a small set of adapters so as to mount it as shown at http://www.photomacrography.net/forum/v ... 195#143195. (If you're using a full-frame camera then in order to avoid vignetting you'll need to replace the cone adapter with a flat one, so as to put the objective closer to the Raynox.)

See http://www.photomacrography.net/forum/v ... hp?t=23898 for a comparison of Raynox DCR-150 versus other purpose-made tube lenses from Mitutoyo, Nikon, and Thorlabs.

Regarding your table of numbers, I think that by "infinity space" you mean the air gap between objective and tube lens. If I've read that correctly, then the fact that magnification changes with separation means that the gap is not infinity space (with parallel rays). If it were infinity space, then changing the separation would have no effect except definitely the major issue of vignetting and perhaps some minor issue with aberrations away from center.

To adjust the tube lens properly, the trick is to remove the objective, point the camera plus tube lens at some distant object, and adjust extension behind the tube lens until the image seen by the camera is focused.

You asked about

*180 mm vs. 200 mm optimal tube length*. With infinity objectives, the main effect of changing tube length is to change magnification. Each objective delivers its rated magnification at some particular tube lens focal length. Using a different tube lens focal length merely makes the projected image larger or smaller; it does not drag the objective away from its optimal focus arrangement as would happen in a finite system. As a result, what's "optimal" depends very much on what you're doing. See discussion at http://www.photomacrography.net/forum/v ... 394#168394.I hope this helps. Again, welcome aboard!

--Rik

I actually have a Raynox DCR-250 and briefly considered using it (I read the related threads a while ago), but decided against it, since I didn't want to tie it up in my rig and dealing with a plethora of adapter rings didn't appeal to me either after getting this neat set of T-mount tubes.

Mounting the achromat shouldn't be a problem (I naively assume), as the lens that came with the T-mount tubes was mounted in one of those tubes. So, I'd just have to take out the old lens, pop in the new one and secure it with the retaining ring.

Which side should face forward is a bit more of a conundrum. Do I figure that out by trial and error or is there a better way to go about it?

- rjlittlefield
- Site Admin
**Posts:**22312**Joined:**Tue Aug 01, 2006 8:34 am**Location:**Richland, Washington State, USA-
**Contact:**

If the diameters match and the thicknesses are compatible, that should work fine.jnh wrote:So, I'd just have to take out the old lens, pop in the new one and secure it with the retaining ring.

There are only two possibilities, so try 'em both and see which one works better.Which side should face forward is a bit more of a conundrum. Do I figure that out by trial and error or is there a better way to go about it?

There are guidelines and computations that can predict this, of course. But those are difficult and error prone, so to be sure that you got the right result you would end up testing anyway. In this case it's more effective to just skip all the thinking and run the tests.

--Rik

At Edmund Optics, perhaps you've been looking at the arrangements shown for Mitutoyo tube lenses. Some of these do contain a poorly written phrase that makes it seem as if the tube lens and objective should be spaced apart from one another. However, deeper reading of Mitutoyo's literature indicates that no such space is actually needed. My experience with the MT-1 tube lens confirms this.jnh wrote:Also, Edmunds, Thorlabs etc. recommend an objective-tube lens distance of at least 70 mm, so overall, this doesn't seem ideal.

Rik's tests with the Thorlabs ITL200 tube lens showed something quite different--that some separation does improve optical quality. Ditto with the Raynox lens. But I would suggest not generalizing these findings, and suspect that most lenses pressed into service as converging lenses will perform best when placed as close to the objective as mechanically convenient.

You've likely seen it, but my Bratcam tube lens is a T-tube assembly. And somewhere among my files, there is a partial, hypothetical parts list for making something similar with Thorlabs SM2 tubes and an ITL200 tube lens.

Cheers,

--Chris

I've seen the completed assembly in another thread and - given how meticulous you are at documenting things - was hoping to find more details, but a search didn't turn up anything. So, thanks for the link - impressive setup.You've likely seen it, but my Bratcam tube lens is a T-tube assembly.

The MT-1 is a bit expensive, though. Hence, I was looking into that achromat...

You need but ask, Jnh, as you have done here. Like most at our forum, I'm delighted to share.jnh wrote:I've seen the completed assembly in another thread and - given how meticulous you are at documenting things - was hoping to find more details, but a search didn't turn up anything. So, thanks for the link - impressive setup.

I have a few spare MT-1 tube lenses that I purchased at discount. I'd be happy to provide you one at my cost, which, if memory serves, is about half the current list price.jnh wrote:The MT-1 is a bit expensive, though. Hence, I was looking into that achromat...

This said, my sense is that you should not accept my offer, given your list of infinite objectives--unless you intend to upgrade them. With your objectives, the achromatic converging lens you are looking at seems appropriate. If you ever trade up to apochromatic objectives, you might then expect that a better converging lens, such as the Mitutoyo MT-1, might be a solid investment that would bring out the best in these particular objectives.

I use the Mitutoyo apo MPlan objectives in 2x, 5x, 7.5x, 10x, 20x, 50x, and 100x magnifications. So for me, integrating a Mitutoyo tube lens--designed to work with them--was a no-brainer.

Cheers,

--Chris

I actually received the achromat from Edmunds last night and after mounting it, just had a chance to play around with it for a short time. Now, magnification with this lens is not affected by the length of infinity space, which is nice to see (and the expected behavior). However, tube length seems to have only a small effect on magnification: with a 10x objective (an Olympus UPlanFL UIS2) I get roughly 9x for 100 mm tube length and 11x for 250 mm tube length (just eyeballed, not measured exactly).

Nonetheless, image quality looks promising (just using the focus peeking function of my Fuji X-T1).

- Charles Krebs
**Posts:**5865**Joined:**Tue Aug 01, 2006 8:02 pm**Location:**Issaquah, WA USA-
**Contact:**

What is the focal length of this new Edmund lens? Something seems odd with your "numbers".However, tube length seems to have only a small effect on magnification: with a 10x objective (an Olympus UPlanFL UIS2) I get roughly 9x for 100 mm tube length and 11x for 250 mm tube length (just eyeballed, not measured exactly).

- rjlittlefield
- Site Admin
**Posts:**22312**Joined:**Tue Aug 01, 2006 8:34 am**Location:**Richland, Washington State, USA-
**Contact:**

I share Charlie's concern. The numbers are surprising. As background, consider that an Olympus 10X infinity objective is just an 18 mm focal length lens that has been optimized for high resolution over a small field. When you pair it with a 175 mm FL achromat, you get a combo that acts more or less like a 16.3 mm lens -- more like if the separation is small, less like as the separation gets larger. To get the numbers you've gotten, I calculate that you would need a lot of separation between the objective and the achromat, something on the order of 100 mm.jnh wrote:However, tube length seems to have only a small effect on magnification: with a 10x objective (an Olympus UPlanFL UIS2) I get roughly 9x for 100 mm tube length and 11x for 250 mm tube length (just eyeballed, not measured exactly).

Can you clarify exactly what your setup is, and perhaps confirm those numbers by measurement?

--Rik

That is pretty spot on (how do you derive that?). Based on the recommendations on the Edmunds website, I opted for a separation of >70 mm between objective and tube lens. And 100 mm happened to work out well with the layout of the motorized linear stage (and the mounting hardware for the T-mount tubes) in relation to the object holder.something on the order of 100 mm.

As has been pointed out above, this big separation between tube lens and objective is not necessarily required (it just happens to work well with my setup - and I was toying with the idea to add a beam splitter for axial illumination sometime in the distant future, so a bit of space would be required for that as well). I'll make some actual measurements with different tube length and lens/objective distances in the next few days.

- rjlittlefield
- Site Admin
**Posts:**22312**Joined:**Tue Aug 01, 2006 8:34 am**Location:**Richland, Washington State, USA-
**Contact:**

I know three methods, for some definition of "three".jnh wrote:That is pretty spot on (how do you derive that?).something on the order of 100 mm.

(Warning: the following explanation uses a rear lens focal length of 180, not your 175. But if you're willing to read it at all, I trust that you can make the necessary substitutions.)

Method 1. Play "what if?" using optics design software, such as the free program WinLens3D. This is the first method I used, setting up a simple system of two "thin lenses" with FL's of 18 and 180 mm, and seeing how magnification varied with extension for a specified separation. I got a value of "around 150 mm". But that's between principal planes, which are probably buried inside your lenses, so I subtracted off some distance, rounded to a nice tidy number, and made the posting.

Methods 2a and 2b. Since you asked, I thought harder about more direct methods. As background, note that effective focal length of any focusing lens system can be determined by measuring magnification at two different extensions and then calculating as efl = (ext2-ext1)/(m2-m1). So, given your numbers of 9X at 100 mm and 11x at 250 mm, we get that efl of your combo must be (250-100)/(11-9) = 150/2 = 75 mm. Then note that the effective focal length of a combo with two elements can be calculated (per Wikipedia) as 1/efl = 1/f1 + 1/f2 - d/(f1*f2), for separation d. Finally, use either of the following two methods:

2a. Fire up something like Goal Seek in Excel to figure out what numeric value of d is necessary to make 1/75 = 1/18 + 1/180 - d/(18*180). The answer is 154.8, give or take a little due to iterative approximation.

2b. Slog through the symbol manipulation of algebra to conclude that d = (1/efl - (1/f1 + 1/f2)) * (f1*f2), or "simplified", (f1*f2)/efl - f2 - f1, then plug in efl = 75, f1=18, f2=180, and crunch the numbers. The answer is 154.8, exactly.

If we think hard enough about these relationships, one arrangement of lenses eventually pops out that I find fascinating. (Please excuse the digression, but since you kinda sorta asked, I figure I can get away with finally writing this up.)

When the separation between principal planes of the two lenses reaches exactly f1+f2, then the effective focal length of the combination goes to infinity.

This makes perfect sense if we think about shining a collimated laser beam into one end: the collimated beam gets focused to a point (OK, an Airy disk) by one lens, then refocused by the other lens so as to form a collimated beam with a different diameter. In this role the combo is a beam expander/compressor.

However, we also know that when the rear lens is "focused at infinity", the combo becomes insensitive to separation between the lenses. This behavior still applies even when the separation is f1+f2, so we conclude that this special combo can still serve as an imaging magnifier with magnification f2/f1.

Then finally (this is the bizarre part!) we note that as we change extension the magnification will change by an amount (ext2-ext1)/infinity = zero!

In other words, when used for imaging purposes, that special arrangement of lenses has constant magnification!

The behavior of this combo is sufficiently bizarre that it causes WinLens3D to simply fail. Given f1=18 and f2=180, specifying a separation of

*exactly*198 produces the cryptic diagnostic "Error: 6 Overflow // SysDatEventDirect // Suggest you reverse last action if possible".

However, WinLens3D is perfectly happy calculating for separations that are very close to 198, and in that case its results confirm the nearly constant magnification with that nearly special separation. For example, specifying d = 197.99 confirms that with total rear extension of 180 mm (the rear lens focal length), the magnification is still 10X exactly, while decreasing the rear extension to 100 mm drops the magnification only to 9.9998X and increasing it to 250 mm raises it to only 10.0002X . Further, the behavior is nicely continuous (even though the math isn't), so increasing the separation beyond f1+f2 causes the effective focal length to become negative (wrapping from plus infinity to minus infinity and then approaching zero again), and in that case increasing the extension reduces the magnification and vice versa.

Yeah, I know, this is far more than you ever wanted to hear. But at least it's written and not spoken, so you've been able to easily skip over it. Cheers!

The

**"big picture"**message is that increasing the separation makes magnification less sensitive to rear extension. If you want maximum ability to adjust magnification by changing extension, then mount the two lenses close together.

--Rik

Quite to the contrary. It's as much a hobby as it is a learning experience, so your treatise is very much appreciated.Yeah, I know, this is far more than you ever wanted to hear.

And if I may add a point (2c) to solving that equation, the lazy way would be Wolfram Alpha: https://www.wolframalpha.com/input/?i=1 ... 818*180%29

(sorry, the url qualifier doesn't seem to work properly and the parser omits the last few characters from the link)

- rjlittlefield
- Site Admin
**Posts:**22312**Joined:**Tue Aug 01, 2006 8:34 am**Location:**Richland, Washington State, USA-
**Contact:**

I'm glad to hear you found that interesting!

Thanks for the reminder about Wolfram Alpha. It's a great tool.

I tend not to think about using symbolic solvers ever since I made a point of showing my algebra students how badly they could screw up without warning by having the solver think that for example "i" means sqrt(-1) instead of a variable. I see that Wolfram Alpha has now resolved that problem, at least in the fine print where they offer "Use i as a variable instead".

Nonetheless, I am intrigued and a bit disappointed to see that Wolfram Alpha quotes the exact result as d = 774/5 and offers a link to "Approximate form", which when clicked displays the result "d <approximately equal> 154.80" and offers a link to "More digits", which when clicked adds a bunch more 0's, all the while managing to overlook the fact that 154.8 is actually exact. It would be great if the world's premier math wizards could include one final incantation to check for that not uncommon end case.

Here is a tweaked URL that should work OK:

https://www.wolframalpha.com/input/?i=1 ... 8%2A180%29

--Rik

Thanks for the reminder about Wolfram Alpha. It's a great tool.

I tend not to think about using symbolic solvers ever since I made a point of showing my algebra students how badly they could screw up without warning by having the solver think that for example "i" means sqrt(-1) instead of a variable. I see that Wolfram Alpha has now resolved that problem, at least in the fine print where they offer "Use i as a variable instead".

Nonetheless, I am intrigued and a bit disappointed to see that Wolfram Alpha quotes the exact result as d = 774/5 and offers a link to "Approximate form", which when clicked displays the result "d <approximately equal> 154.80" and offers a link to "More digits", which when clicked adds a bunch more 0's, all the while managing to overlook the fact that 154.8 is actually exact. It would be great if the world's premier math wizards could include one final incantation to check for that not uncommon end case.

Right, the forum's html parser has trouble with a number of characters that are acceptable by browsers. This can be worked around by using the % notation, in this case substituting %2A in place of the * (asterisk).(sorry, the url qualifier doesn't seem to work properly and the parser omits the last few characters from the link)

Here is a tweaked URL that should work OK:

https://www.wolframalpha.com/input/?i=1 ... 8%2A180%29

--Rik