Talk:Telecentric lens

Intro
I think the new intro may be too technical, and doesn't flow all that well with the rest of the article. I didn't want to just revert it back to the previous version. Perhaps it can be improved.--Srleffler 02:46, 1 May 2006 (UTC)


 * Your edit is an improvement. Do you think it needs more? Dicklyon 02:51, 1 May 2006 (UTC)
 * Yes. My concern is that an article should not start out assuming that the reader knows the meaning of terms like chief ray, aperture stop, image/object space, entrance pupil, etc. What is needed are a couple of sentences that give a reader with no knowledge of optics a sense of what a telecentric lens is, and what it might be used for. Then, the article can launch into the technical details. The previous intro at least tried to do that: it said that these lenses are used for machine vision, and explained some of the reasons why (differences between a telecentric and a regular lens). While it wasn't perfect, it was better than the current version for that. We have to remember that the reader might be, for example, a photographer who knows camera lenses but not optical design. Or an engineer who wants to learn something about machine vision but doesn't know much optical theory. Even a reader with a background in optics will appreciate a general overview before the details.


 * The present version also introduced some redundancy that wasn't there before. The distinction between image space and object space telecentric lenses is now explained twice, for example.


 * I'm not sure if the best solution here is to improve the current version or to revert back to the previous version and improve that.--Srleffler 03:03, 1 May 2006 (UTC)


 * OK, I'll try. My main point was to make the intro not just work for one of the two types. Dicklyon 03:07, 1 May 2006 (UTC)


 * Seriously... This intro is one of the more confusing and technical I've ever read. Does this have anything to do with "zoom"?Xvani (talk) 11:48, 30 November 2010 (UTC)


 * No, it doesn't.--Srleffler (talk) 00:53, 1 December 2010 (UTC)

Diagrams
I just moved a couple of diagrams to Talk:Cardinal point (optics) - you may want them here ?

Image:BFP.png

Image:BFP aperture.png

--195.137.93.171 (talk) 04:15, 7 March 2008 (UTC)
 * The lens depicted is not telecentric. --Srleffler (talk) 04:39, 7 March 2008 (UTC)
 * Correction: the version without the aperture is not telecentric, the one with the focal plane aperture is.--Srleffler (talk) 03:10, 8 March 2008 (UTC)

Escher image - 'isometric' effect
has a nice illustration, using an M C Escher-inspired Lego creation to show the 'isometric' effect. --195.137.93.171 (talk) 03:49, 8 March 2008 (UTC)

Come to think of it, isn't it worth explicitly mentioning the absence of perspective (infinity viewpoint or isometric effect) in the article ? I guess that's what 'geometric invariance' means, and it is implied by 'same magnification at all distances', but different people will understand different expressions more easily (eg artists, draughtsmen, athletes and 2D game players). On the other hand, 'orthographic' seems to mean the same thing in optics, but orthographic and isometric are subtly different for draughtsmen ! --195.137.93.171 (talk) 05:25, 8 March 2008 (UTC)

See also Pericentric lens ? Hypercentric lens ?
I deleted the above broken link - I would guess the original link-target has been removed as advertising a specific product ? (Mind you, there are many others in Category:Photographic_lenses!) It was unlikely to have been telecentric, producing extremely distorted images of cylindrical surfaces. ('See also' can sometimes usefully contain the opposite of the article's topic, but probably not usefully here.) There may have been a better place to mention it - any ideas? (Not Anamorphic lens unless it gets de-merged from cine formats !) There doesn't seem to be an existing article for 'silver ball mirror lenses'. I've added it to Anamorphosis#External_links, but that's Art, not Science ! --195.137.93.171 (talk) 04:11, 8 March 2008 (UTC)
 * Pericentric lens

It may be the same as the 'hypercentric lens' shown in the Schneider Kreutznach document, Fig 2-8. There the aperture is actually behind the focal plane, so perspective is inverted - distant objects seem larger than near ones! 'Looking around corners' is probably a misleading description.

Both sources agree that 'Entocentric' is the term for normal lenses, with the aperture in front of the focal plane. That might be worth adding to this article, too.

Can we find the original deleted article, if it ever existed ? It looks interesting, if extremely extreme.--195.137.93.171 (talk) 10:53, 9 March 2008 (UTC)

A telecentric lens is a compound lens ?
Can't a simple lens be telecentric ?

If you put a narrow aperture in the focal plane of a simple lens, it fits the definition.

OK - it may not be much good !

I find the terminology 'Telecentric lens' confusing, since it is not really a property of the glass, but rather depends on the location of the aperture. Question of definition, really, but it could de-mistify a little. Too late.

--195.137.93.171 (talk) 04:46, 8 March 2008 (UTC)

The Schneider link says double-telecentric requires two simple lenses. --195.137.93.171 (talk) 08:02, 8 March 2008 (UTC)

geometric invariance ?
I presume that the magnification still changes if you re-focus a system with a telecentric lens ? (ie if you change both object and image distances) People reading the article may get the impression that the magnification is a fixed value, intrinsic to the lens design. —Preceding unsigned comment added by 195.137.93.171 (talk) 05:14, 8 March 2008 (UTC)

Double-telecentric systems are afocal - "the object- and image side focal points (F'G, FG) for the complete system are at infinity" "A ray parallel to the optical axis leaves the system also parallel to the optical axis" - see the Schneider article in Ext Links, Fig. 4-5. This means that the magnification is indeed intrinsic to the lens design for the double telecentric case. Not for the single one, though.

--195.137.93.171 (talk) 08:16, 8 March 2008 (UTC)

double telecentric - semiconductor printing
I think this type should not be regarded as secondary - it should be numbered 3 in the heading and have its own section.

Why ? I am sure that semiconductor printing applications must use the double telecentric type, although this is currently listed in the 'object space' section.

I have worked in a 'wafer fab', but that doesn't mean I've ever seen the lenses ! I would prove this by explaining another property of the lenses that is not very clearly explained in the article. The lenses do not suffer from 'wide-angle lens distortion' - most obvious when spheres (or heads) get elongated into ellipses at the edges of the image. (Yes, I know it is not 'really' distortion if you consider that the photographic print should be viewed from a distance of lens-focal-length x enlargement ... Viewing-distance does not apply to semiconductor chips !)

I am sure that a lens must be double-telecentric to eliminate this perspective effect, so the semiconductor wafer industry must use double-telecentric lenses. Given the volume of the industry, that strongly implies they are probably about as common as the other two single-telecentric lenses. (Of course, there is x-ray and e-beam lithography ... but electron optics is similar - I don't know about x-ray lenses).

The Edmund Optics page in the first link gives an illustration of "Doubly telecentric 0.25X reduction camera lens for microlithographic use operating at the ArF laser line (0.193 microns)"

--195.137.93.171 (talk) 05:47, 8 March 2008 (UTC)

Um ... I'm having some second thoughts. 'Wide-angle lens distortion' does not stop a simple lens imaging a planar array of squares to a planar array of squares. (Consider the pinhole-camera case.) It will elongate a sphere into an ellipse, but a circle object in the plane of focus will be foreshortened by perspective to exactly counteract the elongation at the image ! I think this means that single-telecentric lenses will actually distort a plane of squares, double-telecentric lenses won't distort, but simple non-telecentric lenses won't distort either. Doesn't that mean that (wide-angle) single-telecentric lenses have problems for metrology, too ? Double-telecentric lenses may be even commoner ... Or do they work so differently from a single lens that you can't think of them as simple lenses at all? (See 'compound lenses' post above.)

There must be other good reasons for using double-telecentric lenses in the semiconductor industry. The light-sensitive layer being exposed will have a finite thickness, which could cause trouble with cones of light arriving at an angle. Underlying surfaces will have reflectivity that varies with angle. Standing-waves within the light-sensitive layer will cause interference patterns if the light is coming from a point, rather than everywhere vertical. That is probably enough to justify using a double-telecentric lens.

My post above is entirely true, but pretty much irrelevant to the semiconductor industry. Worth explaining 'Telecentric Lenses and Wide-angle lens distortion' in the article, though.

My conclusion is unchanged, just the reason for thinking it !

--195.137.93.171 (talk) 06:48, 8 March 2008 (UTC)

Microscope objectives & condensers ?
Thinking about 'apertures at the focal plane' reminds me of some (pretty rare) microscope objectives that have a variable aperture somewhere near the focal plane. Plus condensers, too. Does anyone know if they are telecentric ? It would make sense. suggests so, but may not be an authoritative forum. It seems almost all objectives and condensers may be telecentric. 'Infinity corrected' is different, but an objective can be both that and telecentric. Maybe since 'telecentric' is ubiquitous it hasn't been used as a buzzword in microscopy ?

That would make microscopy the oldest and biggest application of telecentric lenses.

--195.137.93.171 (talk) 06:00, 8 March 2008 (UTC)

Looking for info on Nikon, Olympus, Leitz and Zeiss websites turns up suspiciously little outside of optical profile projector lenses. One exception is the Nikon AZ100 Multizoom, but that is not a Koehler-type microscope. A review of it says "Nikon's AZ100 Multizoom microscope features unique to the industry telecentric optics"

Inconclusive !

--195.137.93.171 (talk) 09:11, 8 March 2008 (UTC)

Rollback by Materialscientist?
@Materialscientist, what it the reasoning for your rollback? — Preceding unsigned comment added by 2601:281:8300:2740:AC26:CDC1:17D8:9127 (talk) 05:42, 13 November 2021 (UTC)
 * I don't know what Materialscientist's reasoning was, but I notice that you removed a bunch of content in your edits. Generally, if you're going to remove content it's a good idea to say why either in the edit comment field or on the talk page. Users will respond better to your edits if they understand what you are doing. --Srleffler (talk) 02:15, 15 November 2021 (UTC)

Use of Edmund Optics and Opto Engineering links
The following two links in the External links section are by commercial businesses that sell telecentric lenses. However, they provide detailed information on telecentric lenses. These pages do not have any direct sales links nor pricing info. If a non-commercial source(s) has an equivalent or better information, please replace them all that those source(s). This link was 404, but I replaced the link with an archival copy from Wayback Machine: Adakiko (talk) 13:33, 3 December 2022 (UTC)
 * Technical description of telecentric effect by Edmund Optics
 * Telecentric lenses tutorial from Opto Engineering
 * Another good explanatory page by Donald Simanek
 * These pages are viable external links. They are not overly commercial and mainly provide a general introduction to the technology. It's not unusual for Wikipedia articles to link to this kind of information on commercial websites.
 * I removed the link to Simanek's article. The original link was dead, and in the edit war it got replaced by a second link to the Edmund article.--Srleffler (talk) 20:23, 3 December 2022 (UTC)

"Sufficiently small"
In this edit user:Goodphy edited a paragraph in the article as shown below. "The simplest way to make a lens telecentric is to put the sufficiently small aperture stop at one of the lens's focal points. This allows only rays including the chief rays (light rays that pass through the center of the aperture stop), that will be about parallel to the optical axis on the other side of the lens, to pass the optical system for any object point in the field of view."

I dispute the claim that the aperture stop needs to be "sufficiently small" to make a lens telecentric. Putting the aperture stop of a lens at one of the lens's focal points makes the lens telecentric. Period. It does not matter how big or small the aperture stop is. Per WP:CHALLENGE, I request a source for this disputed claim.-- Srleffler (talk) 05:11, 13 November 2023 (UTC)


 * Hi Srleffler. I made this statement because it is logically valid for me. If the aperture stop at a focal point of an optical system is so large, then what is the point to have the aperture stop on it? If the aperture stop plays optically something, then it must be finite in its size to a certain level depending on applications. If it is too large, like being infinite, then it is same situation as there is no aperture stop. Goodphy (talk) 09:48, 13 November 2023 (UTC)


 * That statement is wp:original research. Per wp:V, please cite a wp:RS. Thank you Adakiko (talk) 11:10, 13 November 2023 (UTC)
 * Wikipedia policy explicitly forbids including content derived from your own reasoning. That's what we call "original research". Everything must be verifiable in reliable sources, and if material is challenged it must either be sourced or removed.
 * To answer your questions, though, the aperture stop can't be excessively large, by definition. The aperture stop is whatever aperture limits the size of the cone of rays that will pass through the system from an axial object point. If the aperture at the lens's focus were excessively large, then some other aperture in the system would become the aperture stop and the stop would no longer be at the focus. A real optical system always has other apertures—the lenses have finite diameter.
 * Why having the aperture stop at the focal plane in itself makes a lens telecentric is a bit unintuitive, but try thinking about it this way: Imagine the rays from a point on an object in front of the lens. The cone of rays that will pass through the lens are defined by the aperture stop, and the center of that cone of rays is (approximately) the chief ray from that object point. At the image plane corresponding to the object's distance, the rays will all converge on the chief ray, producing an image point in the image plane. Away from the image plane, the rays form a blur circle centered on the chief ray. If the aperture stop is at the front focal plane of the lens, though, the chief ray is parallel to the optic axis, so the blur circle from any given object point is centered at the same distance from the axis that it would have at the image plane. This means that as you move away from the image plane, the magnification of the image remains constant, even as it becomes blurrier due to defocus. This is the characteristic behaviour of an image-space telecentric lens. It doesn't require the aperture stop to be small; it only requires it to be the aperture stop. --Srleffler (talk) 05:34, 14 November 2023 (UTC)
 * Hi.
 * Thank you to provide the detailed explanation. I agree that if the aperture stop is stated located at a certain position, it means that the aperture stop opening size is constrained to have the stop to work on there. I'm still thinking that how much the stop opens impacts how much telecentricity will be earned. Unless clear examples or citations are found, I will remove my wording about stop size requirement for the telecentric. Except for this, my modified statement looks good. Goodphy (talk) 10:05, 14 November 2023 (UTC)
 * Thanks. That fixes part of the problem. The sentence "This allows only rays including the chief rays (light rays that pass through the center of the aperture stop), that will be about parallel to the optical axis on the other side of the lens, to pass the optical system for any object point in the field of view." still needs some work. It has really bad sentence structure, and there are too many ideas jammed into that one sentence to be workable. Secondly, it only applies to an object-space telecentric lens. Third, it loses the point of how a telecentric lens actually works. What makes a lens telecentric is not that all the rays are "about parallel" to the optical axis. The properties of a telecentric lens come specifically from the fact that the chief ray is parallel to the axis. It's the chief ray through an optical system that determines the size of the image, and the chief ray is the center of the blur circle when a lens is not quite in focus. For the first two issues, the original sentence is much better: "This makes the chief rays (light rays that pass through the center of the aperture stop) on the other side of the lens parallel to the optical axis for any point in the field of view." Note that that sentence works for either an object-space or an image-space telecentric lens. I would restore that sentence, and add a sentence clarifying how the chief ray being parallel to the axis relates to the lens's telecentric properties.--Srleffler (talk) 06:11, 18 November 2023 (UTC)