A DIY 123mm f/8 Cooke split triplet for 4x5

[OTS Matt]

Hi Everyone,

I was inspired by Nodda Duma's popular thread A DIY 177mm f/8 Cooke triplet for 4x5 from off-the shelf lenses that YOU can build (https://www.largeformatphotography.info/forum/showthread.php?120024-A-DIY-177mm-f-8-Cooke-triplet-for-4x5-from-off-the-shelf-lenses-that-YOU-can-build) and thought I'd take a crack at improving the performance by splitting the front and rear positive lenses of his Cooke triplet design.

Here's what I got:

215801

Focal length: 123 mm.
Pretty sharp on- and off-axis at f/8. The RMS spot size in the far corner of a 4x5 format is 108 um.
Biggest drawback is vignetting, which causes ~50% light loss at the far corner.

Here's the lens data, all lenses can be purchased from Thorlabs for a total cost of $258. Air gaps are measured from the center of the lens.

Lens 1: LA1979-A (https://www.thorlabs.com/thorproduct.cfm?partnumber=LA1979-A)
0.10 mm air gap
Lens 2: LA1399-A (https://www.thorlabs.com/thorproduct.cfm?partnumber=LA1399-A)
8.56 mm air gap
Lens 3: LD4293 (https://www.thorlabs.com/thorproduct.cfm?partnumber=LD4293)
2.40 mm air gap
STOP (aperture diameter 14.66mm at f/8)
1.27 mm air gap
Lens 4: LA1399-A (https://www.thorlabs.com/thorproduct.cfm?partnumber=LA1399-A)
2.71 mm air gap
Lens 5: LA1050-A (https://www.thorlabs.com/thorproduct.cfm?partnumber=LA1050-A)
Back focal length: 109.55 mm

I ran an image simulation to give an idea of its performance. If you click on the photo to enlarge and then hit next/prev you can pick out some of the differences.

215804
This is the stock image (640x480 pixel resolution) that's used as the input, i.e. this is the object that we'd be taking a photo of.

215805
This is the resulting image. As you can see, the field darkens as you go out toward the edges, but stays pretty sharp. There's also a small amount of distortion.

215806
This is the resulting image from Nodda Duma's OTS Cooke design at f/8 for comparison.

In the next few days I'll put together a mechanical design and post it for the DIY'ers who'd like to build it. My mounting preference is usually Thorlabs lens tubes with spacers cut to length using a lathe.

In the meantime, I'm not a LF photographer and am curious what the community thinks of trading off field uniformity for resolution at the edges of the field. If anyone has suggestions or comments for further improvements, I'd love to hear them. Unfortunately, it's difficult to have it all with stock lenses.

Matt

[Dan Fromm]

The vignetting is pretty much as expected for a 123 mm lens on 4x5.

[Oslolens]

My 121.3mm lens in a Copal #1 has coverage of 82 degree at f22 or 211mm according to Schneider, and it at can be shifted 34-39mm on a 4x5" sheet of film. How will this lens behave when it comes to shift?
Could you also do a similar thing using Nikon 5t, 6t, Canon 250d or 500d achromatic diopter lenses in combination with a minus-element?

Sent fra min SM-G975F via Tapatalk

[Nodda Duma]

Nice work. The 0.1mm airgap between the two front lenses is tight. Can you reverse the order of those two front positives and pull them apart slightly?

[OTS Matt]

Nice work. The 0.1mm airgap between the two front lenses is tight. Can you reverse the order of those two front positives and pull them apart slightly?

Thanks! I did a quick check and increasing the air gap between the two front lenses to 1 mm results in no appreciable difference in performance. In this case, the reoptimized air gaps are:

LA1979-A to LA1399-A: 1.00mm
LA1399-A to LD4293: 8.38mm
LD4293 to STOP: 2.54mm
STOP to LA1399-A: 1.32mm
LA1399-A to LA1050-A: 2.35mm
Back focal length: 109.43mm

I've generally found that I can cut spacers to length with a tolerance of 0.1mm, so this separation will be a bit safer in terms of ensuring that positive lenses 1 and 2 don't touch.

Switching the order of the two front positives yields similar results in terms of spot sizes. With the LA1399-A leading, the field curvature is nearly completely corrected, but at the expense of larger uncorrected astigmatism.

[OTS Matt]

My 121.3mm lens in a Copal #1 has coverage of 82 degree at f22 or 211mm according to Schneider, and it at can be shifted 34-39mm on a 4x5" sheet of film. How will this lens behave when it comes to shift?
Could you also do a similar thing using Nikon 5t, 6t, Canon 250d or 500d achromatic diopter lenses in combination with a minus-element?

Sent fra min SM-G975F via Tapatalk

To make this sort of lens work well, the surface curvatures, glass choice, and spacing are all critical. I haven't seen this sort of detail published for Nikon or Canon lenses, which means that adding a negative element would unfortunately be pretty unlikely to improve image quality. If anyone knows the lens data for these lenses, I'd be happy to give it a try.

Regarding the field of view, this split triplet has a 62.6 degree field of view at the diagonal. I'm not sure what you mean by shift. Lateral or axial shift, and for what purpose?

[OTS Matt]

The vignetting is pretty much as expected for a 123 mm lens on 4x5.

Good to know, thanks. I might play around a little with a stronger negative element to increase the focal length and reduce the vignetting then before moving over to the mechanical design.

[Dan Fromm]

Good to know, thanks. I might play around a little with a stronger negative element to increase the focal length and reduce the vignetting then before moving over to the mechanical design.

Vignetting is given by cos^4(theta) where theta is the angle off axis. Increasing the focal length will reduce theta and vignetting, otherwise there's little you can do.

The shift referred to in post #3 above is movement of the lens' axis off the gate's axis. Direction doesn't matter with decentering movements. Right, left, up, down are all the same.

To answer the shift question in post #3, a 123 mm lens that covers 62.6 degrees will just cover 4x5 with no room for decentering movements. If you scale the prescription up to make a 150 mm lens, it will cover 182 mm and will allow some movements on 4x5. You haven't designed a wide angle lens.

[OTS Matt]

Vignetting is given by cos^4(theta) where theta is the angle off axis. Increasing the focal length will reduce theta and vignetting, otherwise there's little you can do.

The shift referred to in post #3 above is movement of the lens' axis off the gate's axis. Direction doesn't matter with decentering movements. Right, left, up, down are all the same.

To answer the shift question in post #3, a 123 mm lens that covers 62.6 degrees will just cover 4x5 with no room for decentering movements. If you scale the prescription up to make a 150 mm lens, it will cover 182 mm and will allow some movements on 4x5. You haven't designed a wide angle lens.

That's right. This lens system was designed for imaging on a 4x5 format and I didn't consider a larger field of view. While light will get through the system at a wider angle, the performance drops off pretty rapidly. I'm not an LF photographer and wasn't aware that decentering the lens was a common technique. What minimum diagonal field size would you recommend for designing for a 4x5 format?

Unfortunately, scaling the prescription is tricky with stock optics. I was able to achieve an improved design with a 165mm focal length, which I'll be posting soon, but it's not what I'd consider to be wide angle and is unfortunately quite a bit more expensive due to the lenses needed.

[Dan Fromm]

What minimum diagonal field size would you recommend for designing for a 4x5 format?

The minimum is 150 mm. Every 4x5 photographer has ideas about how much more is absolutely necessary. They got along for quite a while with 150/4.5 Tessars and similar, which cover around 170 mm. Late 1950s - early 1960s plasmat types cover around 210 mm. More modern plasmat types cover around 220 mm.

If you're thinking about wide angle lenses, a 4/4 double Gauss type might be possible with stock meniscii. Might. This one of four classic (= pre-WW II) w/a forms. The others are wide angle Protars, "dagor" types (symmetrical 6/2) and the Hypergon (two deep thin meniscii). All very slow, I think the fastest really wide ones are f/9 or so and most are f/14 or slower. I doubt that any of them can be cobbled up from stock optics, could be mistaken.

[Nodda Duma]

Dan I looked at COTS lens designs which would need meniscus lenses. Meniscus COTS lenses are rare and the right set that would be needed don’t exist unless something has changed in the past 2-3 years.

I was, however, able to sneak one token COTS lens into a $50k 3-lens imaging system at work. :D

[Dan Fromm]

Jason, you stinker, I had to look up COTS. Commercial Off The Shelf. I'm not surprised that ones that could be used to make an imitation Wide Field Ektar don't exist. Old economist test of feasibility: if it is worth doing and can be done, someone somewhere is doing it.

I took a quick, possibly incompetent. look at Edmund's on-line catalog (just a coupla' miles away, until it closed I visited their surplus room fairly often) and found no meniscii. Surplus Shack has + and - meniscii, with practically no specs given in the catalog. Didn't look at Thorlabs.

[Nodda Duma]

Dan that made me chuckle. For future reference, GOTS = government off the shelf GFE = government funded equipment.

Thorlabs has a very few…all the wrong shape for minimizing spherical aberration.

Last time I met their regional rep, I considered suggesting they stock singlets that are useful for imaging designs. But, had other things to discuss so it never came up.

[OTS Matt]

I agree - I tried a few designs with the Thorlabs' meniscus lenses and unfortunately wasn't able to get anywhere.

Here's an updated optical design that achieves similar performance at f/6.7. I swapped out the two inner plano convex N-BK7 lenses for CaF2 versions. The engineer in me is sad that this new iteration was achieved by trial and error rather than any particular design insight.

Effective focal length: 164.6 mm. Sharp across the field (150mm image diagonal) at f/6.7. Vignetting is less than before, with about 65% relative illumination in the corners. Unfortunately, the CaF2 lenses cost quite a bit more; the total lens cost is $1k. The CaF2 lenses also don't come anti-reflection coated, same as the fused silica bi-concave lens. Also not a wide-field lens - the 150 mm diagonal corresponds to an angular field of view of 48.4 degrees.

215964

LA1353-A
5.76 mm gap
LA5230
16.52 mm gap
LD4735
3.25 mm gap
STOP
16.00 mm gap
LA5220
1.00 mm gap
LA1353-A
Back focal length: 131.08 mm

Again, here's the source bitmap and resulting simulated image.

215965
215966

Any comments about suitability for LF photography are appreciated. My suspicion is that $1k is getting pretty steep for just the lenses in a DIY project, so I think I'll go back to figuring out a mounting arrangement to share for the 123mm f/8 design.

[Dan Fromm]

Any comments about suitability for LF photography are appreciated. My suspicion is that $1k is getting pretty steep for just the lenses in a DIY project, so I think I'll go back to figuring out a mounting arrangement to share for the 123mm f/8 design.

The big question is what y'r new confection (no sarcasm intended) does better than currently available normal +/- lenses for 4x5, especially at the shooting apertures normally used. Most of us rarely shoot wide open because of depth of field and coverage considerations.

[OTS Matt]

The big question is what y'r new confection (no sarcasm intended) does better than currently available normal +/- lenses for 4x5, especially at the shooting apertures normally used. Most of us rarely shoot wide open because of depth of field and coverage considerations.

Good question! I doubt these stock lens systems will ever be better than what you could achieve with custom optics, but they can be pretty good if you really want to DIY. Perhaps the better question (and maybe the subject of a different thread) is: what would be worth building yourself because it's not otherwise available?

[Dan Fromm]

Good question! I doubt these stock lens systems will ever be better than what you could achieve with custom optics, but they can be pretty good if you really want to DIY. Perhaps the better question (and maybe the subject of a different thread) is: what would be worth building yourself because it's not otherwise available?

From time to time people here have looked into custom optics, usually for wide angle lenses for formats larger than 8x10. So far the costs have been prohibitive.

As for making something because it isn't available, see above. More seriously, used lenses for formats no larger than 8x10 are around, often at reasonable prices, and with more than enough coverage at focal lengths from a bit less than 0.4 x normal to longer than easily usable. To give you an idea, look at the lens catalogs to which the link in the first post in this https://www.largeformatphotography.info/forum/showthread.php?138978-Where-to-look-for-information-on-LF-(mainly)-lenses discussion points.

Also remember that photographers have been somewhat specialized for quite a long time. They take pictures -- the artsy types capture images, I don't see the difference between that and taking pictures -- using off-the-shelf equipment. Very few photographers design and build equipment. Most of the very few go no farther than making view cameras, which aren't that much more complex than boxes (ducks and runs).

[Hugo Zhang]

If you can build something like this, I think many 810 camera users will be interested. Something with 100 degree coverage.

https://www.ebay.com/itm/114788016899

[Dan Fromm]

If you can build something like this, I think many 810 camera users will be interested. Something with 100 degree coverage.

https://www.ebay.com/itm/114788016899

Hugo, some years ago Michael Kadillak tried to persuade an optical shop to make such a lens -- they told him that they didn't want to just make a Dagor type, they could make something better -- and to organize a group buy. The economics -- cost of producing as many as Michael thought he could shift at the cost of production -- didn't come close to working.

But why be a piker? Why not go for a high coverage lens for larger formats? Don't settle for a Dagor when there's the f/14 Perigraphe VIa. Thinking of which, the longest focal length cataloged was 800 mm, claimed coverage 2 meters at infinity. Just what's wanted for shooting 1.2 x 1.6.

Also remember that the OP has been talking about making useful photographic objectives composed of commercial off-the-shelf lenses. This is in some ways a harder problem than making custom optics, especially with w/a lenses.

[OTS Matt]

To give you an idea, look at the lens catalogs to which the link in the first post in this https://www.largeformatphotography.info/forum/showthread.php?138978-Where-to-look-for-information-on-LF-(mainly)-lenses discussion points.

Thanks Dan. This looks like a great resource that you've assembled and maintained!

[OTS Matt]

If you can build something like this, I think many 810 camera users will be interested. Something with 100 degree coverage.

https://www.ebay.com/itm/114788016899

Hugo, thanks for the suggestion. 36 bids, from $299 up to $2,550... that lens is hot! Looks like a challenge to me, but I don't know if COTS lenses will allow me to get anywhere. I'll have to take a look.

[Dan Fromm]

Hugo, thanks for the suggestion. 36 bids, from $299 up to $2,550... that lens is hot! Looks like a challenge to me, but I don't know if COTS lenses will allow me to get anywhere. I'll have to take a look.

Matt, if you're not acquainted with dioptrique.info, take a look at it. It is in French, has prescriptions for > 1,000 lenses, many taken from patents. Eric Beltrando, its owner, is one of my friends. He was a friend of CEDIS-Boyer's owner and reoptimized Boyer prescriptions for properties of newly-arrived batches of glass. He wrote the program that produced the performance curves etc. on his site. I mention Eric and dioptrique here because he insists that Dagors (f/6.8 and f/7.7) have much less coverage than the 85 degrees Goerz claimed. He says the same about f/9 w/a Dagors, see http://www.dioptrique.info/OBJECTIFS11/00533/00533.HTM. He calculated that the f/18 Protar V has much more coverage. See http://www.dioptrique.info/OBJECTIFS1/00034/00034.HTM

I hope that I'm mistaken, but I doubt you'll be able to assemble anything like these lenses using COTS lenses.

[Hugo Zhang]

Dan,

Yes a Protar V covers more, but not by that much than Zeiss Dagor. 16 inch image circle vs 15 1/2 for this 18cm. F/9 is much brighter than f/18. Believe me as I have used both. This 18cm lens covers more than a 24cm f/6.8 Dagor. If you are hiking with an 810 camera and need a small wide angle lens, it's a good choice. I know people can get coated 210mm G-Claron and Computar for less tha half prices with similar coverage.

https://www.cameraeccentric.com/static/img/pdfs/zeiss_3.pdf

[Dan Fromm]

Hugo, I referred the OP to Eric Beltrando's calculations. Eric insists that Dagors of all speeds don't meet their coverage claims. In fact, he calculates that for "exacting applications" f/6.8 Boyer Beryls (Dagor clones) cover only 55 degrees.

I think our disagreement is due to different coverage concepts, the limit of illumination and the limit of good image quality.

In any case, making a working 6/2 double anastigmat from COTS lenses will take considerable luck. The glasses have to have the right combinations of refractive index and dispersion and the lenses have to have the right curvatures, otherwise cementing them together won't be possible.

[OTS Matt]

Matt, if you're not acquainted with dioptrique.info, take a look at it.

I hope that I'm mistaken, but I doubt you'll be able to assemble anything like these lenses using COTS lenses.

Thanks Dan. This looks like another great resource. If you haven't seen it, there's a similar collection of lens designs from patents, etc. at http://www.lens-designs.com/. All of the designs are apparently in the public domain.

I definitely agree that it'll be rare to find COTS lenses to approximate the original prescriptions. The trick will be to see whether there are some lucky substitutions that permit acceptable quality.

As for the split triplet designs I've posted, I think I'll hold off continuing with the mechanical design until there's some DIY interest expressed in building it.

[reddesert]

My 121.3mm lens in a Copal #1 has coverage of 82 degree at f22 or 211mm according to Schneider, and it at can be shifted 34-39mm on a 4x5" sheet of film. How will this lens behave when it comes to shift?

You're describing a Super Angulon type wide angle, which is a significantly more complex design than a Cooke triplet, with more elements and much greater bending of the elements.



Could you also do a similar thing using Nikon 5t, 6t, Canon 250d or 500d achromatic diopter lenses in combination with a minus-element?

Sent fra min SM-G975F via Tapatalk

If you take two off the shelf achromats like these and put them in a symmetric configuration (convex lenses out) about an aperture stop, you'd be making a lens like a Rapid Rectilinear. One could likely make at least adequate images this way. Once the lenses are chosen the only degrees of freedom are the lens spacing and the position of the aperture stop, so you don't have the same ability to optimize as the designer of the RR, but if the lens system is symmetric that minimizes several of the aberrations.

[Oslolens]

You're describing a Super Angulon type wide angle, which is a significantly more complex design than a Cooke triplet, with more elements and much greater bending of the elements..
Butbut, no. The super angulon has 100 degrees and f8. Mine is 121.3 super symmar (120mm nominal) f5.6 plasmat -bastard or worse according some ;)

I'm sure someone will find of-the-shelf lens elements to build a "miracle" large format lens. Mine would be a small 170-200mm covering 8x10" with room to spare. Sold the 180mm Fujinon-W inner ring. Going to test the 210 G-Claron against an early Nikkor-W with 77mm front.

If one can add elements from slr kit-lenses, the choice is widened, but it would take some time to investigate what the early AF 35-70, 18-70 and 28-80 are made of.


Sent fra min SM-G975F via Tapatalk

[reddesert]

I was inspired by Nodda Duma's original thread to look around at what one could make with inexpensive off the shelf lenses. As he says in that thread, a problem for even simple designs is that it's hard to find COTS flint glass elements. Many basic lens designs start from convex crown glass and concave flint glass. But in standard optics catalogs, most of the single elements are made of the same crown glass for the convex, concave, meniscus, etc (typically N-BK7 or an equivalent, which is the least expensive type of optical glass per pound AFAIK). Jason wound up using a fused silica lens for the concave "flint."

Surplus Shed sells a fair number of achromatic doublets and they also sell some of the elements as individual lenses, which means you can get a concave flint, but without any tech details other than diameter and focal length. Many of them are cheap enough (<$10) to fool around with. I got a variety of crown and flint elements with the idea of experimenting with a few rough designs. I am not an optical designer nor do I have Zemax, but there are some tools to visualize simple designs, such as https://arachnoid.com/OpticalRayTracer/, also found some calculations at pencilofrays.com and some Python modules. But next step, I need to make a spherometer to measure the radii of curvature of the elements I have.

I'm curious about the lens mounting methods. Not having machine tools, I got some threaded barrels and rings that are used to assemble telescope eyepieces (also from Surplus Shed), but it is still unresolved how to center and retain the elements in these.

[Dan Fromm]

If we had ham we could have ham and eggs if we had eggs.

[reddesert]

Here is a link to a sample chapter from a book on lens design that discusses designing and optimizing a Cooke triplet. Lens designers know this stuff, but it is useful for non-lens designers to see the level of complexity and degrees of freedom that are used in designing what most photographers think of as a modest lens. A lens that is either much faster or has a much wider field, like a Super Symmar, is a much greater level of complexity.

https://www.willbell.com/tm/ChapterB.3.pdf This is Chapter B.3 from G.H. Smith, "Practical Computer Aided Lens Design," published by Willmann-Bell, and unfortunately Willmann-Bell just closed up shop so you should download it now. For educational purposes, read the text about the choices made and skip over the computer printouts.

[David Lindquist]

Matt, if you're not acquainted with dioptrique.info, take a look at it. It is in French, has prescriptions for > 1,000 lenses, many taken from patents. Eric Beltrando, its owner, is one of my friends. He was a friend of CEDIS-Boyer's owner and reoptimized Boyer prescriptions for properties of newly-arrived batches of glass. He wrote the program that produced the performance curves etc. on his site. I mention Eric and dioptrique here because he insists that Dagors (f/6.8 and f/7.7) have much less coverage than the 85 degrees Goerz claimed. He says the same about f/9 w/a Dagors, see http://www.dioptrique.info/OBJECTIFS11/00533/00533.HTM. He calculated that the f/18 Protar V has much more coverage. See http://www.dioptrique.info/OBJECTIFS1/00034/00034.HTM

I hope that I'm mistaken, but I doubt you'll be able to assemble anything like these lenses using COTS lenses.

My circa 1933 Carl Zeiss Jena catalogue doesn't give coverage angles for their various lenses but does give the "Diameter of circle covered at small stops"; 10 1/2 and 15 1/2 inches respectively for the 18 cm f/6.8 and f/9 Dagors (this catalogue is for the U.S. market). Doing the arithmetic and trigonometry this gives coverage angles of 70 and 95 degrees. So for the f/6.8 Dagor they are showing a bit less than what Goerz American claimed. As for the Wide Angle Dagor, Goerz American claimed 100 degrees at f/45 until at least the early 1950's; later they dialed this back to 90 degrees.

This Zeiss Jena catalogue shows the 18 cm Series V (f/18) Protar covered a 16 inch circle "at small stops".

David

[Nodda Duma]

If we had ham we could have ham and eggs if we had eggs.

I’m going to use that in a meeting in the near future

[Bernice Loui]

Marketing value of advertised and want to believe image circle of any given view camera lens. This ideology and desire persist to this day without the much needed balance of what lenses-optics the camera is capable of dealing with and actual image goal requirements.

During the mid 1980's first wide angle lens acquired for 4x5 was a 4-3/8" golden wide angle dagor. Worked mostly good for 4x5 if stopped down smaller than f22. This "Golden" wide angle dagor was advertised as good coverage or image circle large enough for 5x7 with camera movements. Few short years later, 5x7 happened, the same "Golden" wide angle dagor was applied to the 5x7 film format, it just barely covered 5x7 at f45 with slight darkening of the corners unless the GWDR was used direct on film center zero camera movements. Did not take long before the very real optical performance limitations of the GWDR found it a new home after being replaced by a 115mm f6.6 Grandagon. Absolute superior optical performance to the GWDR, never looked back and have continued with majority of modern wide angle lenses since.

There is a place for vintage wide angle lenses like this, such as hiker-back packers with light weigh field folders where space and weight is paramount. Except too many place these values as the overly generalized features that is most important to all view camera users.. which is never accurate or true.

That said, Dagors have their place and remain a top choice for specific image goals. Much about knowing what the image goals are then applying the lens and remainder of the image making system to meet these image goal needs.


Bernice







My circa 1933 Carl Zeiss Jena catalogue doesn't give coverage angles for their various lenses but does give the "Diameter of circle covered at small stops"; 10 1/2 and 15 1/2 inches respectively for the 18 cm f/6.8 and f/9 Dagors (this catalogue is for the U.S. market). Doing the arithmetic and trigonometry this gives coverage angles of 70 and 95 degrees. So for the f/6.8 Dagor they are showing a bit less than what Goerz American claimed. As for the Wide Angle Dagor, Goerz American claimed 100 degrees at f/45 until at least the early 1950's; later they dialed this back to 90 degrees.

This Zeiss Jena catalogue shows the 18 cm Series V (f/18) Protar covered a 16 inch circle "at small stops".

David