Telecentric Lenses: Benefits & Applications

[fisico30]

Dear Forum,

does anyone know what a telecentric lens or optical system does?

I just know that it provides a projected instead of perspective view of small objects...
Why is that useful?
thanks
fisico30

 

[Bobbywhy]

fisico30, Did you Google it? I did and got lots of hits. You might start here:

http://en.wikipedia.org/wiki/Telecentric_lens

 

[Andy Resnick]

Telecentric lenses are useful for metrology, becasue the magnification does not vary with object distance- telecentric lenses are used a lot for parts inspection. Telecentric lenses can be telecentric in object and/or image space, have a limited depth of focus, and a limited range of telecentric working distances. Because the entrance pupil is located at infinity (for object-sided telecentric lenses), the size of the front element is equal to the field of view.

 

[fisico30]

Hello Andy,
thanks for the reply.

So the same object, positioned at different distances from a converging lens, will have images of different sizes: the further the object the smaller its image.
In telecentric optical system, no matter how object distance, the image will look the same: given two identical objects, the far object will look exactly like the object close by...
This happens only within the telecentric range...

I am still not clear on what it means that the entrance pupil is located at infinity (for object-sided telecentric lenses). I know what the entrance pupil is: the image of the aperture stop formed by the lenses to the right of the AS.
But what does it mean that it forms to infinity? That it only accepts rays that are parallel to the optical axis?

What is the difference between the object-sided telecentric lenses and image-sided telecentric lenses? What different effect do we get?


In a non-telecentric system, does a circular ring look like an ellipse? Why? In a telecentric system instead the image is a circle too?

Thanks
fisico30

 

[Andy Resnick]

I am still not clear on what it means that the entrance pupil is located at infinity (for object-sided telecentric lenses). I know what the entrance pupil is: the image of the aperture stop formed by the lenses to the right of the AS.
But what does it mean that it forms to infinity? That it only accepts rays that are parallel to the optical axis?

What is the difference between the object-sided telecentric lenses and image-sided telecentric lenses? What different effect do we get?


In a non-telecentric system, does a circular ring look like an ellipse? Why? In a telecentric system instead the image is a circle too?

A lot of questions- here we go:

The entrance pupil is the image of the aperture stop in object space, but also, all rays passing through an optical system must enter through the entrance pupil (and exit through the exit pupil). Putting the entrance pupil at infinity does indeed imply that the numerical aperture is very low, but it not zero: AFAIK f/6 is a fast telecentric lens. Another way of thinking about a pupil at infinity means that the 'principal ray' is parallel to the optical axis. Here's a few ray trace diagrams:

http://www.lhup.edu/~dsimanek/3d/telecent.htm

Object vs. image-sided telecentric lenses simply acknowledge that object space and image space are different. There are doubly-telecentric lenses available. The marketing claims of image-sided telecentricity tend to center on the possibility that microlens arrays are sometimes used with a CCD to increase the light gathering efficiency, and "for critical applications, image telecentricity eliminates artifacts from these microlenses"- but I do not know if these claims have ever been objectively evaluated.

Your last question is very interesting, and I may not have the right answer- I think if you had a disk oriented at some angle to the optic axis, the telecentric lens and 'conventional' lens images will be superficially similar- there will be an ellipse in both- but a grid drawn on the disc will still consist of parallel lines in the telecentric system, whereas a conventional lens will show converging lines in one direction.

There's some images here that may help:

http://www.tmworld.com/article/323776-Telecentric_lenses_simplify_noncontact_metrology.php

Even though image 1b was taken with a telephoto rather than a truly telecentric lens, you should get the idea.

 

[fisico30]

Hello Andy,
thanks for all the infos. I will read it carefully.

I think the magnification constancy offered by telecentric systems occurs only for objects having the same physical size but located at different distances from the optical system. That is why telecentric systems are important in quality control where identical objects are monitored...If the objects have different physical sizes their magnification will still be different even with a telecentric system;

The entrance pupil is the image of the controlling aperture stop formed by the imaging elements preceding it. As far as the entrance pupil being at infinity: I know that the principal ray is the one passing exactly through the center of the aperture stop (AS). That implies that the chief ray is parallel to the horizontal axis;

The entrance pupil is the limiting aperture that the light rays "see". If the aperture stop is at "infinity", why would it necessarily mean that the numerical aperture is very low? That means that few rays are able to enter the system;

Also, let's consider two identical objects located at different distances from an optical system and focus our attention on the same point on both objects.
The bundle of rays emerging from the point that is more distant is different from the bundle of rays emerging from the closer point: one bundles reaches the optical system with more divergence.
Does a telecentric system select, let pass only those rays that have the same angle, divergence, from both object points and disregards the others? Is that how a telecentric system works at a fundamental level?
Aside from formulas and equations, why would a normal lens make a smaller or larger image of the same object if it gets located at different distances from the lens?

thanks and Merry Christmas,
fisico30

 

[Andy Resnick]

I'm having trouble following you. That is, your use of the term 'magnification' in your first paragraph is confusing me: an object will appear the same size regardless of how far away it is (for a telecentric system), and two objects, differently sized, will appear to be differently-sized by the same amount regardless of how each object is from the telecentric lens.

The ray trace diagrams may help you better understand the geometry of principal rays.

 

[fisico30]

The chief ray is that ray that passes through the aperture stop. In the object is an on-axis point source, its chief ray will go though the AS.
What if the object is a planar object, located where the on-axis point source is? Will the chief rays from the off-axis points making up the planar, extended object all pass through the center of the AS too?

Will these off-axis points see the same aperture as the aperture stop for the system?

If an object is 3-dimensional, can we still select one aperture in the optical system to be the AS? A 3D object is composed of point sources located on different planes...

thanks
fisico30

 

[Andy Resnick]

Will the chief rays from the off-axis points making up the planar, extended object all pass through the center of the AS too?

Yes.

Will these off-axis points see the same aperture as the aperture stop for the system?

Not always. 'Yes' for well designed lenses with a low-to-moderate field of view, 'No' for wide-angle lenses and poorly designed lenses (e.g. vignetting).

If an object is 3-dimensional, can we still select one aperture in the optical system to be the AS?

Yes. The AS is a property of the lens, not the object.

 

[fisico30]

Mr. Resnick,

I keep being told that the principal ray is different from the chief ray. I know what the chief ray is (the ray going through the AS). What is the principal ray?
I thought it was the same thing...

thanks
fisico30

 

[Andy Resnick]

Maybe it's simply confusion about terminology: the two 'principal' rays most useful for optical design are the chief and marginal rays:

http://spie.org/x33110.xml

 

[pslebow]

Great discussion. It appears that since only ray parallel to optics axis pass through the pupil, much of the light from any given object point is thrown away. Are telecentrics inherently less efficient?

 

[Andy Resnick]

I wouldn't say 'inefficient', since the transmission is likely over 98% (assuming all the elements are coated). The numerical aperture/f-number of telecentric lenses are small/large, tho. The Edmund Optics lenses are f/6 (na 0.1) at best, the Zeiss ones are similar.

 

[pslebow]

Yes - it seem to be a tradeoff between "ideal" telecentric performance and f# - the bigger the pupil the more rays not parallel with the axis reach the sensor. In someways its like a pinhole camera relying on the aperture to be a spatial filter.