PDA

View Full Version : Achromatic doublets, coverage and resolution



genotypewriter
22-Nov-2012, 08:25
It's very difficult to come across information on using achromatic doublets as a taking lens.

The exact thing I'm wondering about is, if the focal length of the achromatic doublet is 1000mm (i.e. a +1 diopter) and it's stopped down to f/125 (using a ~8mm aperture stop place in front of the lens), what is the diffraction-limited angular coverage like?

I guess this also depends the diameter of the lens as well, although the opening stays the same. Any idea on how big a difference we would see if the diameter went from say 50mm to 100mm (while the stop stayed the same size at ~8mm)?

Any input would be appreciated. Thanks in advance...

G

Dan Fromm
22-Nov-2012, 09:25
A number of long focus lenses for 35 mm still cine cameras are achromatic doublets. Century Tele Athenar and TA II lenses, for example. Mine weren't outstanding. Astro Berlin's long lenses (see http://camerapedia.wikia.com/wiki/Astro_Berlin) are also achromatic doublets.

I apologize for not responding directly to your question, but if you're looking for a decent and not too expensive lens that covers at least 4x5 -- if not, why ask about doublets? -- your best bet is probably a long process lens. They're around and dialyte types, e.g., Apo-Ronars, are quite competitive with modern tele lenses for LF. Like doublets they're not teles so require extension >= focal length.

genotypewriter
22-Nov-2012, 14:36
Thanks for your response, Dan.

I was being a lot more ambitious than shooting 4x5 with one. In my defense, I wasn't planning on enlarging and hence the f/125.

The issue with enlarging lenses, as you know, is things get expensive at those focal lengths and honestly I was hoping to go beyond the ~45 degree coverage limit. Haven't come across any WA enlarging lenses around the 1000mm/40" mark.

I have a single element diopter and it seems to cover 8x10 with enough sharpness when I stop it down to ~f/125. Didn't measure lp/mm or anything. I'm expecting an achromatic doublet to be even better.

Would be good to hear from someone who has experimented with one.

Hermes07
22-Nov-2012, 14:55
Afraid I can't answer your question directly either, but a few things to think about..

Firstly, if you're thinking along the lines of getting a 1000mm lens to use for 4x5, as Dan says, there are other options - 1000mm isn't THAT extreme so there's no reason to be resorting to a doublet. Process lenses are superb and aren't that uncommon in this focal length, but they'll be fairly expensive. Igor Camera has a 1000mm Dallmeyer Dallon Telephoto which (as a genuine anastigmat) will top what an Achromat can do and is a cheaper option.

If you're looking at using doublets in general then consider that a rough rule for the circle of usable definition from a classic landscape lens (single achromat with stop positioned a good distance in front of it) is about half the focal length. Although it can be reduced by narrowing the aperture, spherical aberration in a typical achromat bought today (they're likely to be biconvex if not specifically designed for photography) will be very pronounced if covering a negative.

Ultimately, it depends on exactly what you're trying to achieve. Narrowing the field of view will reduce the effect of many of the off-axis aberrations so it all depends on exactly how much coverage is critical to you. You mention apertures of 50mm/100mm which equate to f/20 & f/10. When you think that the traditional landscape lens was used at f/30+ on a blue-sensitive medium which wasn't going to be enlarged, and still found lacking, I think you will be disappointed at the performance of a doublet at anything wider.

genotypewriter
22-Nov-2012, 20:04
Firstly, if you're thinking along the lines of getting a 1000mm lens to use for 4x5, as Dan says, there are other options - 1000mm isn't THAT extreme so there's no reason to be resorting to a doublet. Process lenses are superb and aren't that uncommon in this focal length, but they'll be fairly expensive. Igor Camera has a 1000mm Dallmeyer Dallon Telephoto which (as a genuine anastigmat) will top what an Achromat can do and is a cheaper option.

If you're looking at using doublets in general then consider that a rough rule for the circle of usable definition from a classic landscape lens (single achromat with stop positioned a good distance in front of it) is about half the focal length. Although it can be reduced by narrowing the aperture, spherical aberration in a typical achromat bought today (they're likely to be biconvex if not specifically designed for photography) will be very pronounced if covering a negative.

Ultimately, it depends on exactly what you're trying to achieve. Narrowing the field of view will reduce the effect of many of the off-axis aberrations so it all depends on exactly how much coverage is critical to you. You mention apertures of 50mm/100mm which equate to f/20 & f/10. When you think that the traditional landscape lens was used at f/30+ on a blue-sensitive medium which wasn't going to be enlarged, and still found lacking, I think you will be disappointed at the performance of a doublet at anything wider.

Appreciate your interest, Hermes, but it would've been a better use of your time if you had read what I wrote before replying.

Maris Rusis
22-Nov-2012, 20:05
It's very difficult to come across information on using achromatic doublets as a taking lens.

The exact thing I'm wondering about is, if the focal length of the achromatic doublet is 1000mm (i.e. a +1 diopter) and it's stopped down to f/125 (using a ~8mm aperture stop place in front of the lens), what is the diffraction-limited angular coverage like?

I guess this also depends the diameter of the lens as well, although the opening stays the same. Any idea on how big a difference we would see if the diameter went from say 50mm to 100mm (while the stop stayed the same size at ~8mm)?

Any input would be appreciated. Thanks in advance...

G

I have used achromatic doublet lenses as taking lenses with focal lengths from 900mm down to 100mm.

The largest aberration that severely limits sharp coverage is field curvature. An approximate rule of thumb suggests that the radius of the curved field is about equal to the focal length. In general achromats work better as long focal lengths on small formats and abominably as short lenses on big formats. Some of the other first order aberrations (spherical, astigmatism, coma, distortion, vignetting, lateral colour) can be tamed somewhat by stopping down. In practice I find the limit to stopping down is a 3mm aperture for a 8"x10" photograph. Smaller than this results in an image softened by visible diffraction. A 1000mm lens will work fine at f300 and a 100mm lens is ok at f30, and so on. Non-diffraction aberrations will continue to affect the image....more or less.

One thing is certain: no matter how you compromise on aberrations an achromat is not a cheap and versatile substitute for a fully corrected camera lens. I wasted a lot of time finding this out.

genotypewriter
22-Nov-2012, 20:56
...

Thank you, Maris. The field curvature radius ~= focal length is a bit shocking. As mentioned, I won't be enlarging the shots so I was hoping that stopping down to the point of diffraction limit nearing the naked eye's resolution might help a bit with the curvature and other aberrations. Any idea on how big a difference the ach. doublet's diameter makes if we keep the aperture fixed (at a small size)? Through a simple lens design tool I figured that a larger diameter lens is better (at the same aperture) but wasn't sure how it affected the field curvature and other aberrations.

One more question for you, if you don't mind... I can't remember the lens design name off the top of my head but I've seen a symmetrical one that looks very similar to two opposing doublets (convex sides out). Is there any room for improvement if two ach. doublets are used this way instead of just one?

Thanks, again.

Struan Gray
23-Nov-2012, 02:07
How much of a difference is made by having a physically larger lens (of the same glass and surface curvatures) depends on where you have the aperture stop, and how large it is. Bigger stops (larger holes) need bigger lenses, as do stops located further away from the glass.

To a certain extent you can trade field curvature against blurring aberrations like coma by moving the stop - this is why traditional landscape menisci have the stop some way out in front of the glass. It's no magic cure though.

Using two lenses in a symmetrical arrangement does help to squelch aberrations. It also reduces the total focal length, roughly by half. Your choice :-)

If you want to tinker with simple lenses, it's easier to play than predict what will happen. If you want to experiment with long lenses without spending a lot of money there are ways to avoid the fuzz and fiddle of singlets and achromats. One is to get a cheap off-brand tele for 35 mm. Many of them will cover LF film once you remove the tubes and baffling that otherwise vignettes them. Better, and easier, is to get a reasonably long lens and remove the front group. 420, 465 and 480 mm process lenses (apo-ronars and the likd) in barrel are still relatively cheap, and removing the front element will roughly double the focal length *and* give you an easy way to mount it on your camera. Some sort of coating comes free too.

Hermes07
23-Nov-2012, 02:08
Appreciate your interest, Hermes, but it would've been a better use of your time if you had read what I wrote before replying.

I wrote this reply before reading your second post.

Now that I have, let me be clearer.

Without knowing exactly what doublet you're talking about, no-one will be able to answer your question accurately. The covering power will depend on the shape of the lens and the stop position, and will vary significantly based on these. A bi-convex doublet with a close stop will show far more curvature of field than a meniscus doublet with a distant stop. You can improve the image by increasing the stop distance, thereby choking the coverage to eliminate the inbound off-axis rays that produce the field curvature. Adding a second doublet to make a mock rapid rectilinear/aplanat has the same effect - improving the central image but reducing the angle of illumination.

Questioning a lens' performance only becomes meaningful once you define exactly what it is you want it to do. I've used a meniscus with satisfactory performance at f/90-f/128 before but this was with my AOV and covering requirements - what are yours?

William Whitaker
23-Nov-2012, 09:10
When I hear "achromatic doublet", I keep thinking refractor telescope...

E. von Hoegh
23-Nov-2012, 09:10
Thank you, Maris. The field curvature radius ~= focal length is a bit shocking. As mentioned, I won't be enlarging the shots so I was hoping that stopping down to the point of diffraction limit nearing the naked eye's resolution might help a bit with the curvature and other aberrations. Any idea on how big a difference the ach. doublet's diameter makes if we keep the aperture fixed (at a small size)? Through a simple lens design tool I figured that a larger diameter lens is better (at the same aperture) but wasn't sure how it affected the field curvature and other aberrations.

One more question for you, if you don't mind... I can't remember the lens design name off the top of my head but I've seen a symmetrical one that looks very similar to two opposing doublets (convex sides out). Is there any room for improvement if two ach. doublets are used this way instead of just one?Thanks, again.

It's called an Aplanat, or Rapid Rectilinear. They still have astigmatism and the associated spherical abberations, but are actually extremely sharp in the central zone. If you could find one (or find a pair of 2000mm doublets to assemble) a 1000mm R-R would give very sharp 8x10 negatives.

E. von Hoegh
23-Nov-2012, 09:11
When I hear "achromatic doublet", I keep thinking refractor telescope...

Yes, an achromatic doublet + an eyepiece = telescope.

Struan Gray
23-Nov-2012, 09:23
I've mounted a couple of antique telescope objectives on my cameras, a 500 mm and 1000 mm, both f10-ish; but I never bothered to take a picture with them. They are extremely sharp on axis (one is genuinely diffraction-limited wide open), but get blurry really quickly once you're off axis. They really only cover APS-C at best, even if you like funky swirl.

Ironically, less good telescope objectives may perform better as imaging lenses. However, even in an air-spaced doublet you have enough degrees of freedom that optimising for on-axis high-magnification imaging is going to give you a different doublet than an ideal wide-field image-forming version.

William Whitaker
23-Nov-2012, 09:55
Yes, an achromatic doublet + an eyepiece = telescope.

Well, I may be simplistic, but refractor telescopes I've had were pretty well taxed covering a 35mm negative, at least at prime focus. Perhaps that was just because of internal stops, but I would suspect the performance would be pretty awful very far off axis.

Mark Sawyer
23-Nov-2012, 10:08
I think the answers would apply better if we knew the OP's questions better. Is this for 8x10? (Just guessing from post #3.) Are you looking for conventional results, just on a budget, or is the f/125 also for greater depth of field?

I think you might consider putting a negative diopter behind a conventional lens, which would give you the angle of view of a longer lens, but might give you a shorter bellows extension, (the telephoto effect). Putting the negative diopter in front would increase the focal length and give you the conventional bellows extension if you want a long extension.

Mark Sawyer
23-Nov-2012, 10:12
Well, I may be simplistic, but refractor telescopes I've had were pretty well taxed covering a 35mm negative, at least at prime focus. Perhaps that was just because of internal stops, but I would suspect the performance would be pretty awful very far off axis.

People who use achromatic doublets are usually looking for poor performance on and of axis. Think Imagon, Kodak Portrait Lens, Spencer Port-Land, Plasticca...

Stopping it down to F/125 (the OP's plan) should cure a lot of that.

E. von Hoegh
23-Nov-2012, 11:06
Well, I may be simplistic, but refractor telescopes I've had were pretty well taxed covering a 35mm negative, at least at prime focus. Perhaps that was just because of internal stops, but I would suspect the performance would be pretty awful very far off axis.

Yes, a telescope achromat is not the same as a photographic achromat - which (photographic achromat) is actually an achromatic doublet/meniscus.

The telescope achromat will give a sharp image over a degree or three, more or less.

Maris Rusis
23-Nov-2012, 17:59
[QUOTE=genotypewriter;957213]Thank you, Maris. The field curvature radius ~= focal length is a bit shocking. As mentioned, I won't be enlarging the shots so I was hoping that stopping down to the point of diffraction limit nearing the naked eye's resolution might help a bit with the curvature and other aberrations. Any idea on how big a difference the ach. doublet's diameter makes if we keep the aperture fixed (at a small size)? Through a simple lens design tool I figured that a larger diameter lens is better (at the same aperture) but wasn't sure how it affected the field curvature and other aberrations.E]

The size of the achromat outside of the aperture stop makes no difference. Only glass that sees light contributes to image structure.


One more question for you, if you don't mind... I can't remember the lens design name off the top of my head but I've seen a symmetrical one that looks very similar to two opposing doublets (convex sides out). Is there any room for improvement if two ach. doublets are used this way instead of just one?

Thanks, again.
Opposing achromats are the basis for the rapid rectilinear photographic lens. The underlying idea is that the achromats on either side of the central stop will have aberrations of equal magnitude but of opposite sign thus cancelling out. If the light "bending" at each achromat is equal, say in a symmetrical imaging system operating at 1:1 repro ratio, the image quality is virtually apochromatic! But only for a small on-axis field. That's why process lenses tend to be long for the fields covered.
For unequal image conjugates, say infinity on one side and film on the other, unequal acromats that minimise (and equalise) light "bending" at each surface give the best compromise. There are nice nuances of optical engineering and some effortful mathematics lurking here; interesting though!

genotypewriter
23-Nov-2012, 22:22
Thanks for the responses everyone!


To a certain extent you can trade field curvature against blurring aberrations like coma by moving the stop - this is why traditional landscape menisci have the stop some way out in front of the glass. It's no magic cure though.

Using two lenses in a symmetrical arrangement does help to squelch aberrations. It also reduces the total focal length, roughly by half. Your choice :-)

If you want to tinker with simple lenses, it's easier to play than predict what will happen. If you want to experiment with long lenses without spending a lot of money there are ways to avoid the fuzz and fiddle of singlets and achromats. One is to get a cheap off-brand tele for 35 mm. Many of them will cover LF film once you remove the tubes and baffling that otherwise vignettes them. Better, and easier, is to get a reasonably long lens and remove the front group. 420, 465 and 480 mm process lenses (apo-ronars and the likd) in barrel are still relatively cheap, and removing the front element will roughly double the focal length *and* give you an easy way to mount it on your camera. Some sort of coating comes free too.

Interesting to hear about the distance at which the stop should be placed. I might experiment with that. I tried removing the front group of a 480mm Tessar (conventional) but then the rear group had a much shorter focal length and I couldn't get an image to form from the front group (divergent?)



Without knowing exactly what doublet you're talking about, no-one will be able to answer your question accurately. The covering power will depend on the shape of the lens and the stop position, and will vary significantly based on these. A bi-convex doublet with a close stop will show far more curvature of field than a meniscus doublet with a distant stop. You can improve the image by increasing the stop distance, thereby choking the coverage to eliminate the inbound off-axis rays that produce the field curvature. Adding a second doublet to make a mock rapid rectilinear/aplanat has the same effect - improving the central image but reducing the angle of illumination.

Questioning a lens' performance only becomes meaningful once you define exactly what it is you want it to do. I've used a meniscus with satisfactory performance at f/90-f/128 before but this was with my AOV and covering requirements - what are yours?

About the placement of the stop... wouldn't increasing the distance in front of the lens (while keeping the stop diameter the same) be similar to decreasing the diameter while keeping it in one place?

As mentioned in an earlier post, I'm trying to see if achieving an angular coverage larger than 45 degrees is possible from a 1000mm achromatic doublet. From the sounds of what everyone's saying getting diffraction limited performance doesn't seem possible at f/125 far off the axis. So at this stage it would be interesting to know how rapidly the resolution drops. This is not so much about the format size/ratio but the said resolution.



It's called an Aplanat, or Rapid Rectilinear. They still have astigmatism and the associated spherical abberations, but are actually extremely sharp in the central zone. If you could find one (or find a pair of 2000mm doublets to assemble) a 1000mm R-R would give very sharp 8x10 negatives.

Yes, that's the name. Then I've seen designs like the Dogmar which seem like two air-spaced doublets? Any idea on how big an image circle a 500mm rapid rectilinear could produce for contact-type prints at around f/125 (for 6 lpmm)?



Ironically, less good telescope objectives may perform better as imaging lenses. However, even in an air-spaced doublet you have enough degrees of freedom that optimising for on-axis high-magnification imaging is going to give you a different doublet than an ideal wide-field image-forming version.

Interesting... I'm guessing it's because the cheap ones don't have cemented doublets?



I think the answers would apply better if we knew the OP's questions better. Is this for 8x10? (Just guessing from post #3.) Are you looking for conventional results, just on a budget, or is the f/125 also for greater depth of field?

I think you might consider putting a negative diopter behind a conventional lens, which would give you the angle of view of a longer lens, but might give you a shorter bellows extension, (the telephoto effect). Putting the negative diopter in front would increase the focal length and give you the conventional bellows extension if you want a long extension.

Hi Mark... as mentioned in the original post, I'm interested in knowing the size of the diffraction limited image circle that a good 1000mm achromatic doublet can produce at f/125. Ideally it'd work out to be greater than 45 degrees (830mm diameter) as there are process lenses that can do that but I don't want to go that way. But now I see I'm being very ambitious. I already knew such simple lenses are only good for relatively small ICs and hence the reason why I thought f/125 might improve my chances.




The size of the achromat outside of the aperture stop makes no difference. Only glass that sees light contributes to image structure.

Forgot to mention that the thicknesses of the elements would also change proportionally. It can be easily verified that the bigger glass is better but what's not clear to me is by how much it would be better.


Opposing achromats are the basis for the rapid rectilinear photographic lens. The underlying idea is that the achromats on either side of the central stop will have aberrations of equal magnitude but of opposite sign thus cancelling out. If the light "bending" at each achromat is equal, say in a symmetrical imaging system operating at 1:1 repro ratio, the image quality is virtually apochromatic! But only for a small on-axis field. That's why process lenses tend to be long for the fields covered.
For unequal image conjugates, say infinity on one side and film on the other, unequal acromats that minimise (and equalise) light "bending" at each surface give the best compromise. There are nice nuances of optical engineering and some effortful mathematics lurking here; interesting though!

Interesting stuff, again. I've also seen the rear group being slight smaller than the front group while in general shape they're the same.

Mark Sawyer
24-Nov-2012, 00:48
Hi Mark... as mentioned in the original post, I'm interested in knowing the size of the diffraction limited image circle that a good 1000mm achromatic doublet can produce at f/125. Ideally it'd work out to be greater than 45 degrees (830mm diameter) as there are process lenses that can do that but I don't want to go that way. But now I see I'm being very ambitious. I already knew such simple lenses are only good for relatively small ICs and hence the reason why I thought f/125 might improve my chances.

I suspect that, with a simple, inexpensive doublet of unknown charactaristics, you're going to run into other resolution limits long before you hit the diffraction limits. The diffraction limits will depend in the wavelength of light, the distance you're focusing at, etc. And rather than being a hard "limit", it will just be a judgement call of "is this an acceptable size for an airy disc, depending on the degree of enlargement (if any), or not?"

genotypewriter
25-Nov-2012, 23:14
I suspect that, with a simple, inexpensive doublet of unknown charactaristics, you're going to run into other resolution limits long before you hit the diffraction limits. The diffraction limits will depend in the wavelength of light, the distance you're focusing at, etc. And rather than being a hard "limit", it will just be a judgement call of "is this an acceptable size for an airy disc, depending on the degree of enlargement (if any), or not?"

Hi Mark,

As I mentioned in post #3, I'm not going to enlarge the images. So, approximately 11.8 lines/mm (which I consider as the naked-eye's resolution for my purposes) is limited by diffraction with ~550nm light at f/125. Yes, there's always a bit of give or take due to one parameter (e.g. wavelength) and but another parameter might take or give (human vision) too. At the end of the day it's an approximate value but it's still a usable value when planning things. Same could be said about the validity of wind tunnel simulations in the real world but people still do them :)

As for the doublets I had in mind, by no means they're inexpensive. One of them I have in mind is around 140mm in diameter, coated, used for close-up work and costs over 2 grand... another reason for me to deliberate over this to this extent.

Øyvind:D
26-Nov-2012, 07:16
As for the doublets I had in mind, by no means they're inexpensive. One of them I have in mind is around 140mm in diameter, coated, used for close-up work and costs over 2 grand... another reason for me to deliberate over this to this extent.

SURPLUSSHED.COM: Achromat & menisc: 2 of each will not break your wallet! To get them mounted on each side of aperture might ;-) SK Grimes can help you! Or buy a casket lens set, mine fit perfectly to a 150mm f4.5 Tessar-shutter.
L5630 ACH ø22.4mm 1531 Y $7.00: 765mm f/32
L5931 PMN ø30.0mm 1357 Y $4.00: 678mm f/22
L8093 PMN ø42.0mm 2000 Y $4.00: 1000mm f/25
L8112 PMN ø30.0mm 1000 N $4.00: 500mm f/17

genotypewriter
28-Nov-2012, 18:38
SURPLUSSHED.COM: Achromat & menisc: 2 of each will not break your wallet! To get them mounted on each side of aperture might ;-) SK Grimes can help you! Or buy a casket lens set, mine fit perfectly to a 150mm f4.5 Tessar-shutter.
L5630 ACH ø22.4mm 1531 Y $7.00: 765mm f/32
L5931 PMN ø30.0mm 1357 Y $4.00: 678mm f/22
L8093 PMN ø42.0mm 2000 Y $4.00: 1000mm f/25
L8112 PMN ø30.0mm 1000 N $4.00: 500mm f/17


Thanks. I've already experimented with some positive meniscus lenses... 72mm diameter 1000mm focal length ones. I tried using a stop with a single lens only and it improved the sharpness off-axis a lot. When I combined the two with the backs reversed (with some gap in between but without any stops), I didn't find them producing any better/larger image circles than single element. With achromats I expect the same experiments to give better results. Hope I'm on the right track.

Øyvind:D
29-Nov-2012, 05:13
I tried using a stop with a single lens only and it improved the sharpness off-axis a lot. When I combined the two with the backs reversed (with some gap in between but without any stops), I didn't find them producing any better/larger image circles than single element. I think you need to stop down to see difference. Try this site http://translate.google.no/translate?sl=auto&tl=en&js=n&prev=_t&hl=no&ie=UTF-8&layout=2&eotf=1&u=http%3A%2F%2Fwww.dioptrique.info%2Fbase%2Fobj_1914%2Favant-1900.HTM

Compare Wollaston (press "Images" in upper left corner) to other constructions.