Flange distance questions

Started Apr 10, 2017 | Discussions thread
AiryDiscus Senior Member • Posts: 1,708
Re: Flange distance questions

lattesweden wrote:


Some questions I hope someone here knows the answer too:

Mirrorless cameras have shorter flange distance due to the missing mirror box. But is there any advantage optically by doing this or could the DSLR systems do mirrorless versions of their bodys without changing their flange distance or would there be any optical downsides doing this?

Mirrorless lenses seems at times to be longer than a DSLR lens with the same focal lenght, is there an optimal flange distance for a given sensor size?

Thanks and best regards from Sweden!

For lenses with focal lengths longer the the flange distance, a symmetrical design can be used.

The rapid rectilinear is the first symmetrical design form, and has 2-4 elements.

In this rendition, which is the most common, the singlets on each side of the aperture stop are cemented into doublets to correct axial color and (to some extent) spherical aberration and coma.

Because of symmetry, this design has no field curvature, very low astigmatism, no distortion, and no lateral color.

It is useful for extreme fields of view - even e.g. 100 degrees, but only low speed (~f/16 or f/20) because it cannot correct spherical aberration.

The cooke triplet adds a negative lens at the aperture stop (in this, the original one, it is quite displaced, but it is usually at the stop) to add more variables to correct spherical aberration.

You can speed these up to about f/6 for big fields of view. They have higher order astigmatism because symmetry is broken which prevents them from being used for huge fields. About 65 degrees full field is the max. If you use a small field (e.g. 10 degrees FFOV) you can use them at up to e.g. f/3.

The double gauss raises the element count to 6 and restores symmetry, which reduces higher order aberrations

This design is extremely versatile and can be used for most applications outside of designs with extremely large fields of view. This includes very fast (e.g. f/1) designs, or wide fields of view (e.g. 80 degrees). A "wide angle adapter" can be added to the front that reduces the field of view seen by the double gauss lens, whcih is the design ethos used by most fast aperture wide angle lenses (Canon 24L, nikon 24/1.4, sigma 24/1.4, etc.)

Ludwig bertele designed a symmetrical wide angle lens form,

Here it is schneider's super angulon, but Zeiss refers to it as a biotar or biogon. Can you see the cooke triplet embedded in this? The cemented lenses in the middle are essentially squeezing a very large field of view through the aperture.

This is the crux of the symmetric design forms. All of them have no field curvature, no lateral color, no astigmatism, and no distortion because of symmetry. Symmetry fixes your problems naturally. Symmetry is beautiful.

In a symmetric lens, because the lens focuses the light but does very little else to bend it, the chief ray angle is equal to the half field of view, and the image clearance is equal to the focal length. This means if you want to make a 35mm double gauss, it's going to be 35mm from the image. If your camera has a 44mm flange distance, this is a problem. Digital sensors also have an angle of acceptance ( http://blog.teledynedalsa.com/2012/05/the-angle-on-optical-acceptance/ ) which means there is vignetting from the sensor itself at large chief ray angles, and potential other issues like color shifts in the corners of the picture.

The double gauss can be made telecentric, meaning the chief ray angle is nearly or exactly 0.

This form has a long working distance (image clearance) and results in a lens that is physically quite long. The optical quality is also not all that great over a big field of view, and large apertures are impossible without a very large number of lens elements.  It's still pretty symmetric, although there is a fair amount more  bending being done in the back than in the front.

If the goal is a compact lens for mirrorless, this is not a very good solution.  You could make it more compact or faster, but you're going to need a lot of aspherics and the ability to manufacture it consistently at a low cost is going to go away.

Here's a retrofocus lens' take on the chief ray angle problem,

Can you see how the chief ray (center of yellow bundle) angle is less than half what it was on the left on the right side of the lens?

You can find the focal length by extending the blue rays going into the lens and finding where they intersect themselves in the focusing area behind the lens.  For this lens, the focal length is shorter than the back focal length / image clearance.

This solves both the chief ray angle problem, and the image clearance problem since we can make a 35mm lens with more than 44mm of image clearance.

Unfortunately this design is very asymmetric, so you need a lot of elements to correct it.  Lateral color, distortion, and astigmatism are all difficult.  That's why most wide angle lenses, which are retrofocus, have e.g. 12-16 elements, and most double gauss 50mm f/1.8 or 50mm f/1.4 lenses have 6-8 elements.  Symmetry isn't doing the heavy lifting.

Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow