Canon 20mm f3.5 Macro Bellow Lens

SigmaTog wrote:

ken_in_nh wrote:

How'd you move your assembly for the stack? Step size?

Nice to see your lens and rig in action!

Thanks
The rig is getting too big & is vibrating with the slightest touch.
The steps were 19 into 1mm, ie 0.02mm.
My rig at present.
Tube length is 430mm flange to flange.
Is making the tube length 1/2 as long & using a 2x converter a useful arrangement ?

Absolutely. Your Canon lens is optimized for a relatively short bellows. It's not well corrected at 450mm, so you will have problems with chromatic aberration, spherical aberration, and curvature of field that you won’t have at 200mm.

And you're likely to pick up a decent quality improvement from controlling vibration better.

Of course, you don’t really need the teleconverter, because you don’t have the resolution. Are you familiar with the term "empty magnification"? That's when you try to operate past the "diffraction limit" of an optical system, so all you're doing is magnifying blur. Your 20mm f/3.5 is really only built for magnification up to 10x, and even then it's a bit "iffy".

I only used super-long extensions back in the day when I needed greater working distance. You really don’t want to build a high magnification system this way, because you're creating a bunch of optical problems. Extending a 20mm f/3.5 lens some 450mm (I've included the camera's internal depth) means you’ve got 21.5x magnification, which sounds great until you realize that the lens is about f/80 effectively. (3.5x430mm/20mm).

That means is a very strong "empty magnification" situation. The diffraction limit calculator says we're looking at an Airy disc of 107um at f80. I believe your Sigma has about 9um pixels, so the finest detail that can be resolved is about 5 pixels wide.

You could get the same, or better, pictures by dropping the tube to 200mm, cropping, and enlarging, with a lot more stability and controllability, and you’re in the range your lens was corrected for. You don’t need the teleconverter, because that would just be piling on more empty magnification.

What you really need, at 10x or 20x, is a faster lens. A much faster lens.

Microscope objectives are insanely fast. My favorite 20x objective is a Nikon CF M Plan 20x ELWD. It's an 11.1mm focal length (which doesn’t really matter) with a 10.5mm working distance (which does matter). It has a numerical aperture of 0.4, which converted to an f stop is f/1.1. That's over 3 stops faster than the 20mm f/3.5. More importantly, at 20x it's effectively f/22 instead of f/80, so you have an Airy disc of 30um instead of 107um.

Do I place the 2x behind the lens or in front of the cameras mount please ?

In front of the camera mount. A teleconverter is a negative lens (multi element, and pretty well corrected, but for analysis purposes it's effective a simple negative lens) of about -60mm focal length. If you put it near the positive 20mm lens, you'll get a net -40mm, and the system won’t focus light.

Interesting rail. Is it borrowed from your 3D printer?

And what's holding up the rear end of it?

ken_in_nh wrote:

Wow. That's an impressive looking rig!

It reminds me of something I did many years ago, using a length of PVC pipe as a very long extension tube. But I had a faster lens.

Regarding use of and placement of a teleconverter, I think experimentation is the only way to go, especially since the experiments are easy and can be fun...

That won’t be an easy experiment, and as I pointed out, the answer is already known.

You don't see kit like yours very often, partially because, I suspect, many people use microscope objectives for higher magnification work.

I only used super-long extensions when I needed greater working distance. You really don’t want to build a high magnification system this way, because you're creating a bunch of optical problems. Extending a 20mm f/3.5 lens some 450mm (I've included the camera's internal depth) means he's got 21.5x magnification, and it means the lens is about f/80 effectively.

That means he's in a strong "empty magnification" situation. The diffraction limit calculator says we're looking at an Airy disc of 107um at f80. I believe his Sigma has about 9um pixels, so the finest detail that can be resolved is about 5 pixels wide.

He could get exactly the same pictures by dropping the tube to 200mm, cropping, and enlarging, with a lot more stability and controllability.

For example, I use a 10x objective and my existing lenses as a tube lens, either my 100mm macro or my 55-250 telezoom. I can also do my pics for focus stacking using the internal focusing motors of my "tube" lens and the in camera focus bracketing function of my camera. Whether this sacrifices resolution or creates other issues I can't say because I haven't compared. I'm sure others will comment, since they always do...

Well, since you insist…

Lenses are sharpest when they're designed to operate at only one specific magnification. When you optimize a lens to cover a range of magnifications, you decrease resolution and increase aberrations across the entire distance range. This was a necessary compromise for "general purpose" photographic lenses for most of the 200-year history of photography.

Old-style "finite" microscope objectives are optimized to reach focus at a designated "tube length" from the objective's mounting flange. Most of mine are Nikon 210mm tube length objectives. As you diverge from 210mm, chromatic aberration and spherical aberration get worse, and the field becomes less flat.

Infinity objectives are optimized when going from their designed focal distance to infinity. Move your subject farther from the lens and it now focuses to a finite distance, and you'd need a camera lens that focused "beyond infinity" to bring it back into focus. Move the subject closer and you can bring it into focus by focusing the camera lens closer, but you're moving outside the optimally corrected range of your objective. It's a complex lens, so aberrations may get really strange doing this: the field may literally be "wavy" with multiple peaks and valleys.

It will get "worse", but we can’t easily predict how bad.

You might also consider whether reversing your lens would improve results at higher magnification, if you care.

It won’t help. His lens is optimized for magnification in the 5-10x range. It's not a symmetrical lens, so reversing it would be forcing it to operate in the 0.1-0.2x range.

One reverses a "regular" photographic lens because those lenses are designed to work in the 0.0-0.2x range, so reversing them means they are better corrected for 5x to infinite x (although they will be diffraction limited long before you reach infinity, lol).

Same thing with enlarger lenses: they're optimized to go from maybe 0.25-0.1x, so reversed you can go from 4x to 10x.