I feel as if my understanding of the optics of photomicrography is a bit lacking. Hopefully someone here can answer a few of my questions.
I have been reading Charles Krebs' articles on photomicrography, which stresses the need (for most sensor sizes) for placing a photo eyepiece to pick up the real image about 10mm into the photo tube. The photo eyepiece enlarges the image so that it fills the diagonal of the camera sensor.
What I've been reading also suggests that while it is possibly to directly project the real image from the finite objective onto a camera sensor, vignetting should occur as the image is about 17mm in diameter. Yet when I use my objective-on-a-bellows setup, I never get vignetting, even when I am not using a photo eyepiece. And yet, if the real image at the 160mm tube length mark is only 17mm in diameter, I should see vignetting. Why don't I? It is due to the lack of a narrow c-mount tube?
On this forum a popular technique is to attach objective lenses to bellows, not necessarily kept at the proper tube length for the lens. I have tried this technique, and it certainly works. This almost care-free approach seems contrary to the recommendations for photomicrography using traditional microscopes. Aren't we sacrificing some image quality by not using the objectives at their intended tube lengths?
A finite microscope is a hollow tube with an objective lens at one end. This is especially obvious if you remove the head and use a straight, optics-free monocular. A bellows with an objective lens at one end is exactly the same thing, just without a nice integrated specimen stage, stand, and lighting system. So why use photo eye pieces for traditional microscopes, but not for objectives-on-a-bellows setups, which seem no different?
I own a photo eyepiece which I can attach to my bellows via an M42 thread. I can set it up such that the distances between the objective lens and the photo eyepiece projection lens are exactly as they would be if placed parfocal into a conventional microscope. Given the advice from Charles Krebs, this would seem to be the best approach as it mirrors the technique used with traditional microscopes. Yet this doesn't seem to be a common approach when dealing with objectives-on-a-bellows setups. Why not?
Photo eyepieces and objectives on a bellows
Moderators: rjlittlefield, ChrisR, Chris S., Pau
-
Rylee Isitt
- Posts: 476
- Joined: Fri Apr 13, 2012 3:54 pm
- Location: Canada
- Contact:
-
rjlittlefield
- Site Admin
- Posts: 23622
- Joined: Tue Aug 01, 2006 8:34 am
- Location: Richland, Washington State, USA
- Contact:
Let's see if this helps...
Every microscope objective is designed to deliver rated magnification and best image quality when used with a certain tube length and cover glass thickness. Any deviation in those parameters will add optical aberrations.
If aberrations were the sole concern, then objectives would always be used exactly at their design point.
However, aberrations are not the sole concern, and in many cases the added aberrations are small enough that they are more than compensated by other advantages such as expanding the area of interest to cover the whole sensor. This is particularly true for objectives with small NA, such as 0.30 and below.
To put things in perspective, consider cover glass thickness. Most biological objectives are designed to be used with cover glass that is 0.17 mm thick. But it turns out that with NA of 0.25 and below, omitting the cover glass entirely has no visible effect on the image. The effect is detectable at NA 0.40, and from there on out it becomes progressively more important at a rate of NA to the fourth power. At NA 0.65, the effect of omitting cover glass is blatantly obvious -- see HERE for illustration.
Variations in tube length have the same sensitivity to NA. One standard reference is a graph of "Tolerance to tube length change versus objective NA", HERE. This graph is for objectives designed for 160 mm tube length, and you can see that even at NA 0.30, the tolerance shown is about 90 mm. The exact number is fuzzy and depends on the criteria chosen, but the point is that low NA objectives are quite tolerant of changes in tube length.
As a matter of practice, this means that with low NA objectives, you have the option of adjusting tube length and thus magnification over quite a wide range, without introducing aberrations that degrade the image too much.
So, the real question is not how big an image the objective creates, but rather how much of that image is good enough to use? This varies widely depending on the particular objective.
In very old scopes, the eyepiece field stop might have been only 16 mm diameter, with an expectation that only the center of even that small area would be good. Then 18 mm became sort of standard, and now wide-field eyepieces are 25 mm or more, with an expectation of high quality across the whole field.
Of course this increase in field diameter has gone along with improvements in objective design to cover those wider fields with high quality image. For example the modern Nikon CFI Plan Achromat 10X NA 0.25 has been tested to cover over 50mm diameter with high quality image. In contrast, the older Nikon CF N Plan Achromat 10X NA 0.30 just barely covers an APS sensor, roughly 27 mm. As a result, you can push the CFI down to even 5X and it will still cover an APS sensor, whereas the CF N cannot be pushed down at all or the corners start to go seriously astigmatic. Other objectives commonly found on eBay aren't even that wide; for example the Nikon CF BD Plan 5X NA 0.13 is fine at image center, but falls far short of covering an APS sensor.
--Rik
Every microscope objective is designed to deliver rated magnification and best image quality when used with a certain tube length and cover glass thickness. Any deviation in those parameters will add optical aberrations.
If aberrations were the sole concern, then objectives would always be used exactly at their design point.
However, aberrations are not the sole concern, and in many cases the added aberrations are small enough that they are more than compensated by other advantages such as expanding the area of interest to cover the whole sensor. This is particularly true for objectives with small NA, such as 0.30 and below.
To put things in perspective, consider cover glass thickness. Most biological objectives are designed to be used with cover glass that is 0.17 mm thick. But it turns out that with NA of 0.25 and below, omitting the cover glass entirely has no visible effect on the image. The effect is detectable at NA 0.40, and from there on out it becomes progressively more important at a rate of NA to the fourth power. At NA 0.65, the effect of omitting cover glass is blatantly obvious -- see HERE for illustration.
Variations in tube length have the same sensitivity to NA. One standard reference is a graph of "Tolerance to tube length change versus objective NA", HERE. This graph is for objectives designed for 160 mm tube length, and you can see that even at NA 0.30, the tolerance shown is about 90 mm. The exact number is fuzzy and depends on the criteria chosen, but the point is that low NA objectives are quite tolerant of changes in tube length.
As a matter of practice, this means that with low NA objectives, you have the option of adjusting tube length and thus magnification over quite a wide range, without introducing aberrations that degrade the image too much.
It's a misconception that the objective creates an image that small. At rated magnification, almost all objectives will create an image that is much larger than 17 mm diameter. It is a field stop in the eyepiece that establishes a hard border on what you can see.vignetting should occur as the image is about 17mm in diameter
So, the real question is not how big an image the objective creates, but rather how much of that image is good enough to use? This varies widely depending on the particular objective.
In very old scopes, the eyepiece field stop might have been only 16 mm diameter, with an expectation that only the center of even that small area would be good. Then 18 mm became sort of standard, and now wide-field eyepieces are 25 mm or more, with an expectation of high quality across the whole field.
Of course this increase in field diameter has gone along with improvements in objective design to cover those wider fields with high quality image. For example the modern Nikon CFI Plan Achromat 10X NA 0.25 has been tested to cover over 50mm diameter with high quality image. In contrast, the older Nikon CF N Plan Achromat 10X NA 0.30 just barely covers an APS sensor, roughly 27 mm. As a result, you can push the CFI down to even 5X and it will still cover an APS sensor, whereas the CF N cannot be pushed down at all or the corners start to go seriously astigmatic. Other objectives commonly found on eBay aren't even that wide; for example the Nikon CF BD Plan 5X NA 0.13 is fine at image center, but falls far short of covering an APS sensor.
Probably the most important reason is parfocality -- having the camera and visual eyepieces be focused at the same depth so that "what you see is what you get". Matching the field size to the sensor size is also important, but at low NA and without the need for parfocality, that could also be accomplished just by extending the camera away from the scope so as to increase the tube length.So why use photo eye pieces for traditional microscopes, but not for objectives-on-a-bellows setups, which seem no different?
Cost, convenience, and flexibility, enabled by no need for parfocality.[using a projection eyepiece] doesn't seem to be a common approach when dealing with objectives-on-a-bellows setups. Why not?
--Rik
I'll tackle a few parts:
Yep! The only objectives I have which actually giive a "circular image field" (probably not the right phrase but I expect it makes sense!) are, surprisingly perhaps, 2x objectives. The rest all fill a 35mm frame with light, but l;largely with unusably poor quality away from the centre circle. The circle varies with objective family, as you've read.
If you alter the distance from sensor to objective, yes you alter the magnification, but the quality goes off too. If you go longer, you'll be hitting diffraction as well as using the objective outside its designed area of use.
The "better" the lens, the more critical it's likely to be. There's a pair of examples, quality versus magnification:
a 4x Plan apo here http://coinimaging.com/nikon_4apo.html
an Economy 10x here http://coinimaging.com/nik10lwd.html
See how the 4x's perrformace peaks, but the 10x doesn't.
There are Apo 10x objectives (NA = 0.45) which probably behave more like that 4x apo.
(Apos typically have a smaller or tighter image cicle and less working distance, too).
Why don't "we" use projection eyepieces? Cos a sensor doesn't need one like a human eye does. You CAN use one, eg to enlarge a small image-circle lens field onto your camera sensor, so you get more magnification and use all your pixels, but you need a microscope to arrange the tubular bits, really. ( A teleconverter has a siimilar effect).
Also IF your sensor has small enough pixels then you may as well just use that smaller image field - you have all the information you're going to get.
I'll leave you to read up on what resolution you get from a particular objective and how many pixels the "airy disc" would occupy (see Nikon's Microscopyu pages http://www.microscopyu.com/ ). If you use an APS/DX sensor, then you'll be "matching" an objective's output farily well. As a rough example, simplified for illustration: say your objective is capable of 1 micron resolution, it's a 10x objective, so you get dots down to 10 micron on your sensor. Well a current sensor might have pixels about 5 microns across, which is somehwere about right.
That'll be far too loose an explanation/line of reasoning for some folk hereabouts, but you get the idea.
Only some bjectives will covber an APS sensor, but often the corners aren't as critical.
When you get to Infinite objectives, things shift a bit - that's for after you've done some more reading!
Aren't we sacrificing some image quality by not using the objectives at their intended tube lengths?
Yep! The only objectives I have which actually giive a "circular image field" (probably not the right phrase but I expect it makes sense!) are, surprisingly perhaps, 2x objectives. The rest all fill a 35mm frame with light, but l;largely with unusably poor quality away from the centre circle. The circle varies with objective family, as you've read.
If you alter the distance from sensor to objective, yes you alter the magnification, but the quality goes off too. If you go longer, you'll be hitting diffraction as well as using the objective outside its designed area of use.
The "better" the lens, the more critical it's likely to be. There's a pair of examples, quality versus magnification:
a 4x Plan apo here http://coinimaging.com/nikon_4apo.html
an Economy 10x here http://coinimaging.com/nik10lwd.html
See how the 4x's perrformace peaks, but the 10x doesn't.
There are Apo 10x objectives (NA = 0.45) which probably behave more like that 4x apo.
(Apos typically have a smaller or tighter image cicle and less working distance, too).
Why don't "we" use projection eyepieces? Cos a sensor doesn't need one like a human eye does. You CAN use one, eg to enlarge a small image-circle lens field onto your camera sensor, so you get more magnification and use all your pixels, but you need a microscope to arrange the tubular bits, really. ( A teleconverter has a siimilar effect).
Also IF your sensor has small enough pixels then you may as well just use that smaller image field - you have all the information you're going to get.
I'll leave you to read up on what resolution you get from a particular objective and how many pixels the "airy disc" would occupy (see Nikon's Microscopyu pages http://www.microscopyu.com/ ). If you use an APS/DX sensor, then you'll be "matching" an objective's output farily well. As a rough example, simplified for illustration: say your objective is capable of 1 micron resolution, it's a 10x objective, so you get dots down to 10 micron on your sensor. Well a current sensor might have pixels about 5 microns across, which is somehwere about right.
That'll be far too loose an explanation/line of reasoning for some folk hereabouts, but you get the idea.
Only some bjectives will covber an APS sensor, but often the corners aren't as critical.
When you get to Infinite objectives, things shift a bit - that's for after you've done some more reading!
-
Rylee Isitt
- Posts: 476
- Joined: Fri Apr 13, 2012 3:54 pm
- Location: Canada
- Contact:
Great replies. It makes perfect sense to me now.
I have a nice macro lens to use as a tube lens should I ever want to fiddle with infinite objectives. And should I ever upgrade my microscope, I'll probably go with a modern infinite scope. But that's years from now, probably.
Thanks all for the clarification. All makes sense!
I guess that's where I lucked out in a sense. My photo eyepiece has a removable c-mount adapter with a T2 thread. Removing this, you can use a T2 to M42 adapter (Baader part # 1508015 "Russian Camera Adapter") to directly attach the photo eyepiece to a M42 bellows. It allows you to set the bellows to the tube length the lens is designed for, and to magnify the center of the image just as you would using a photo eyepiece in a normal situation. But now I understand why this is seldom done.you need a microscope to arrange the tubular bits, really
I have a nice macro lens to use as a tube lens should I ever want to fiddle with infinite objectives. And should I ever upgrade my microscope, I'll probably go with a modern infinite scope. But that's years from now, probably.
Thanks all for the clarification. All makes sense!
Errrm, the tube lens has to focus at infinity, so macro lenses aren't the obvious choice, though some are good at all distances.macro lens to use as a tube lens
Canon 100mmL and Nikon 105 f/2.8 MF look fine though, and probably many others at that length, if you have adequate field coverage and want the lower magnification.
-
Rylee Isitt
- Posts: 476
- Joined: Fri Apr 13, 2012 3:54 pm
- Location: Canada
- Contact:
Most macro lenses can still focus to infinity. The MP-E65 and a few similar lenses (with either built-in bellows or extension tube-type designs) cannot. But they are not the majority of macro lenses, at least not among modern off-the-shelf offerings. Quite a few macro lenses come in 100mm or longer to increase working distance. Canon offers one at 180mm, although it's the 100mm 2.8 that I use.
-
Peter De Smidt
- Posts: 233
- Joined: Thu Jan 05, 2012 9:10 am
- Contact:
-
Rylee Isitt
- Posts: 476
- Joined: Fri Apr 13, 2012 3:54 pm
- Location: Canada
- Contact:
Peter,
That's understandable. The 100L did get a good mention here. I have its cousin, the non-L version - similar in spec, but without the weather sealing or IS. The lack of IS does mean the two lenses are going to be optically different, though - and I'm sure the addition of the L standard means higher quality glass as well. I am not sure how the non-L version would fare.
That's understandable. The 100L did get a good mention here. I have its cousin, the non-L version - similar in spec, but without the weather sealing or IS. The lack of IS does mean the two lenses are going to be optically different, though - and I'm sure the addition of the L standard means higher quality glass as well. I am not sure how the non-L version would fare.
-
Charles Krebs
- Posts: 5865
- Joined: Tue Aug 01, 2006 8:02 pm
- Location: Issaquah, WA USA
- Contact:
Rylee,
On objectives, photo-eyepieces and such...
Most older compound microscopes microscopes used 23.2mm diameter eyepiece tubes and trinocular tubes. While it varied somewhat, most also had the "real" image (known as the intermediate image) located 10mm down from the the top edge of these tubes. Many trinocular tubes were also fixed in position and size. As a result there was no way to place a cameras film or sensor at the location of the intermediate image. This then necessitated additional optics to "place" the the intermediate image onto the film. The intermediate image was also limited by the 23.2mm inside tube diameter (but generally was considerably smaller than that). In the film days the 35mm format (24x36mm) was the small format and commonly the microscope cameras of the day were designed for larger sheet film or Polaroid print sizes. As a result, the small 16-20mm diameter intermediate image in the trinocular tube needed to be enlarged in order to cover the larger film formats. Even the "small" format of the day, 35mm, generally used a 2.5X magnification to "fit" the image onto the film with its 43mm diagonal. Also very important... many of the microscope optical systems of the 60's, 70's, and 80's were designed to provide final chromatic correction with the eyepieces (After all, eyepieces of some sort were needed both for viewing and for photography using the microscope. So the designers incorporated them as an integral part of the optical system).
Rik nicely described the progression in coverage of objectives. Initially they needed to provide a "quality" image of about 16mm or so. "Wide-field" came next and generally was thought of being about a 20mm image circle. At that time "Super-Widefield" was sometimes used to describe a 22mm image circle. (There are no "official" industry specifications for these varying "wide-field" designations, so it is best to talk in terms of the actual image diameter). 22mm was as large an image that could be practically used if all the microscopes tubing is of 23.2mm diameter. So if you wanted to place a camera on a microscope, and the sensor had a diagonal larger than that, an eyepiece of some sort (or an "afocal" arrangement) was used to provide a good "fit" for the smaller image in the tube onto the larger film.
As it stands today, 35mm (24x36mm, 43mm diagonal) is considered a "large" format for photomicrography. 1/2" sensors measuring about 4.8x6.4mm, with a diagonal of 8mm, are one of the more common sensor sizes used in dedicated microscope cameras (which is why you now commonly camera adapters with around a 0.5X magnification). DSLRs with an APS-C sized sensor have a format diagonal of about 27 or 28mm (depending on the manufacturer), and would be considered a pretty large format for use on a microscope.
While today, the 23.2mm tubes are still the most common by far, there are more compound microscopes with 30mm eyepiece tubes, and eyepieces that can "see" image sizes of 25mm or 26.5mm. Certain objectives have been designed to provide images large enough to accommodate this.
When you "divorce" yourself from the microscope and start putting objectives on bellows, many considerations change. You no longer have the restrictions of the narrow tube diameters, and you can place the sensor/film at any distance you choose (Although for optical reasons it is often best not to deviate greatly from the "design" point). You can easily place the sensor at the location of the intermediate image. You do need to remember that most objectives were not designed to provide very large image circles. You also need to consider if the objective used was designed expecting to receive final image correction via an eyepiece... these are generally not a good choice for the bellows approach. If the objectives are chosen and tested carefully, there is no need to incorporate an eyepiece into a bellows arrangement.
Outside of the chromatic correction aspect, probably the biggest consideration when using a bellows mounted objective is the size of the quality image produced. Naturally you can crop out the edges and corners if they are not satisfactory, but I think the goal for most of us is to use optics that give usable results over the entire frame. This is easier to do with an APS-C sized sensor (or smaller), but a few do use 24x36mm sensors with some pieces. Really the only way to know what works well is to give it a try or see what other have had success using. Some objectives are clearly "over-achievers" and do quite nicely for sensor sizes larger than they would need to have been designed to cover. Many are excellent for about a 15-16mm circle in the center but then deteriorate rapidly beyond that.
On objectives, photo-eyepieces and such...
Most older compound microscopes microscopes used 23.2mm diameter eyepiece tubes and trinocular tubes. While it varied somewhat, most also had the "real" image (known as the intermediate image) located 10mm down from the the top edge of these tubes. Many trinocular tubes were also fixed in position and size. As a result there was no way to place a cameras film or sensor at the location of the intermediate image. This then necessitated additional optics to "place" the the intermediate image onto the film. The intermediate image was also limited by the 23.2mm inside tube diameter (but generally was considerably smaller than that). In the film days the 35mm format (24x36mm) was the small format and commonly the microscope cameras of the day were designed for larger sheet film or Polaroid print sizes. As a result, the small 16-20mm diameter intermediate image in the trinocular tube needed to be enlarged in order to cover the larger film formats. Even the "small" format of the day, 35mm, generally used a 2.5X magnification to "fit" the image onto the film with its 43mm diagonal. Also very important... many of the microscope optical systems of the 60's, 70's, and 80's were designed to provide final chromatic correction with the eyepieces (After all, eyepieces of some sort were needed both for viewing and for photography using the microscope. So the designers incorporated them as an integral part of the optical system).
Rik nicely described the progression in coverage of objectives. Initially they needed to provide a "quality" image of about 16mm or so. "Wide-field" came next and generally was thought of being about a 20mm image circle. At that time "Super-Widefield" was sometimes used to describe a 22mm image circle. (There are no "official" industry specifications for these varying "wide-field" designations, so it is best to talk in terms of the actual image diameter). 22mm was as large an image that could be practically used if all the microscopes tubing is of 23.2mm diameter. So if you wanted to place a camera on a microscope, and the sensor had a diagonal larger than that, an eyepiece of some sort (or an "afocal" arrangement) was used to provide a good "fit" for the smaller image in the tube onto the larger film.
As it stands today, 35mm (24x36mm, 43mm diagonal) is considered a "large" format for photomicrography. 1/2" sensors measuring about 4.8x6.4mm, with a diagonal of 8mm, are one of the more common sensor sizes used in dedicated microscope cameras (which is why you now commonly camera adapters with around a 0.5X magnification). DSLRs with an APS-C sized sensor have a format diagonal of about 27 or 28mm (depending on the manufacturer), and would be considered a pretty large format for use on a microscope.
While today, the 23.2mm tubes are still the most common by far, there are more compound microscopes with 30mm eyepiece tubes, and eyepieces that can "see" image sizes of 25mm or 26.5mm. Certain objectives have been designed to provide images large enough to accommodate this.
When you "divorce" yourself from the microscope and start putting objectives on bellows, many considerations change. You no longer have the restrictions of the narrow tube diameters, and you can place the sensor/film at any distance you choose (Although for optical reasons it is often best not to deviate greatly from the "design" point). You can easily place the sensor at the location of the intermediate image. You do need to remember that most objectives were not designed to provide very large image circles. You also need to consider if the objective used was designed expecting to receive final image correction via an eyepiece... these are generally not a good choice for the bellows approach. If the objectives are chosen and tested carefully, there is no need to incorporate an eyepiece into a bellows arrangement.
Outside of the chromatic correction aspect, probably the biggest consideration when using a bellows mounted objective is the size of the quality image produced. Naturally you can crop out the edges and corners if they are not satisfactory, but I think the goal for most of us is to use optics that give usable results over the entire frame. This is easier to do with an APS-C sized sensor (or smaller), but a few do use 24x36mm sensors with some pieces. Really the only way to know what works well is to give it a try or see what other have had success using. Some objectives are clearly "over-achievers" and do quite nicely for sensor sizes larger than they would need to have been designed to cover. Many are excellent for about a 15-16mm circle in the center but then deteriorate rapidly beyond that.
-
Rylee Isitt
- Posts: 476
- Joined: Fri Apr 13, 2012 3:54 pm
- Location: Canada
- Contact:
Charles,
That's a great description of the issues. Thanks!
I can remove the head from my compound scope, revealing a universal dovetail mount. I have a universal dovetail to T2 adapter, and am looking around for a T2 coupler with external threads, since I need a male T2 thread to work off of. If I can find this part, I can mount my bellows to my microscope and my camera to my bellows. This may be useful when I'm just doing slide work.
Based on this, I don't need my photo eyepiece for any of my work, which is something of a shame in a way, since it's a very nice (and expensive) piece of kit. Well... I guess that's a lesson learned!
I'm still looking for "bargain" finite CF objectives so that I don't have to worry about corrective eyepieces. I expect that'll cost me a pretty penny, though.
That's a great description of the issues. Thanks!
I can remove the head from my compound scope, revealing a universal dovetail mount. I have a universal dovetail to T2 adapter, and am looking around for a T2 coupler with external threads, since I need a male T2 thread to work off of. If I can find this part, I can mount my bellows to my microscope and my camera to my bellows. This may be useful when I'm just doing slide work.
Based on this, I don't need my photo eyepiece for any of my work, which is something of a shame in a way, since it's a very nice (and expensive) piece of kit. Well... I guess that's a lesson learned!
I'm still looking for "bargain" finite CF objectives so that I don't have to worry about corrective eyepieces. I expect that'll cost me a pretty penny, though.