Designing a mirror system in NX7.5

[RUGmechanic]

Hi All,

For a space project I am designing a (concept) optical system. One of the optical sub systems is a so-called offner relay imaging system.
It consists of 2 concave mirrors, see attached picture. The surface of the secondary mirror is located at the focal point F of the primary, at a focal distance f=R/2
(For any given concave mirror, theory tells me that the focal distance is 1/2 the radius and that rays of light coming in parallel at the mirror axis are focussed at F).
So far, so good...

Designing the system, I draw a line from light inpact to the mirror centre. This is the normal vector and the angle of impact is the angle of reflection and the reflected ray should go through F.
But here is the problem: In NX, the reflected ray is not going through F.

Anybody any idea of what might be the problem?

using NX7.5....

Older budweiser

 

[jerry1423]

There is a small chance that is is going thru where you want it but it's just not showing it properly - in other words it could be a graphics issue with your computer.

 

[cowski]

The rule that 'rays parallel to the principle axis pass through the focal point' is a good approximation for rays close to the principle axis. The ray you have drawn is far enough away from the principle axis that it significantly deviates from that rule.

 

[JohnRBaker]

What you're seeing IS the correct behavior for 'spherical' mirrors. However, it's incorrect to assumption that a mirror like this has a single focal point. You're thinking about the behavior of 'parabolic' surfaces. To see what I mean, draw a series of 'beam trajectories' parallel to the first one and see where they cross the centerline of your mirror arrangement (see attached image).

Now don't get me wrong, your model is correct, it's just your assumptions which are not. I think if you look into exactly what offner relay imaging systems are used for you will discover that they do not depend on the mirrors having single focal points. In fact, I suspect that if you replaced the spherical mirrors with parabolic mirrors that this scheme wouldn't even work.

John R. Baker, P.E.
Product 'Evangelist'
Product Design Solutions
Siemens PLM Software Inc.
Industry Sector
Cypress, CA

http://www.siemens.com/plm

http://www.plmworld.org/museum/

To an Engineer, the glass is twice as big as it needs to be.

 

[FrankSwinks]

Is it possible to rotate the angle mirror so that the system is in focus? Attached is a jpeg image showing what I mean.

Frank Swinkels

 

[JohnRBaker]

Unfortunately the source of the light is not a single point but rather a wide parallel beam. To see what I mean open the attached example part file (NX 7.5) and double click on the sketch. Now go to...

Edit -> Sketch Parameters...

...and select the Green 120mm dimension and then move the 'slider' at the bottom of the dialog to dynamically update the sketch showing where individual rays from a wide horizontal beam would strike the secondary mirror. You can also play the same game with the Green angular dimension which will show what effect changing the angle of the planar mirror would have.

Anyway, it's an interesting exercise that if for no other reason, shows how you can use sketches to simulate engineering problems and their possible solutions.

John R. Baker, P.E.
Product 'Evangelist'
Product Design Solutions
Siemens PLM Software Inc.
Industry Sector
Cypress, CA

http://www.siemens.com/plm

http://www.plmworld.org/museum/

To an Engineer, the glass is twice as big as it needs to be.

 

[JohnRBaker]

I found a slight problem with the last sketch, but its been fixed in the file attached below.

You can still follow the same procedure outlined in last post to see the light beam behavior, just use this sketch instead.

Sorry for the misstep, but it's a bit late here in SoCal ;-)

John R. Baker, P.E.
Product 'Evangelist'
Product Design Solutions
Siemens PLM Software Inc.
Industry Sector
Cypress, CA

http://www.siemens.com/plm

http://www.plmworld.org/museum/

To an Engineer, the glass is twice as big as it needs to be.

 

[RUGmechanic]

I am new into optics (had some basic education on this 20 years ago), so my first search for data was wikipedia . There you find some misleading information, like the single focal point.
Using NX, creating different parallel lines indeed shows there is no single focal point. Thanks for that John.
This can also be demonstrated by using simple mathematics using sin, cos and tan. There are some more sophisticated methods.

To make things more difficult (er, challenging, I am an engineer after all ), different wavelengths have different "focal points". Then there are aberations (already mentioned: chromatic, different foci for different wavelengths and things like astigmatism, coma, defocus, etc...) and last but not least: Field curvature.

For the project I am working on the optical design (3D) is done with Zemax software normally, data transfer by STEP, but for now NX is used to set up a concept model. Like John did, I use the sketcher to set it up parametrically. Works fine now, but it takes a lot of time rebuilding in case a mirror is added and/or removed. This is the case now, because the offner turns out to be to large for the launch vehicle, so I must "fold" the beams. After this first concept, I am planning to build a 3D model, based on the sketches.

Thanks all

Older budweiser

 

[RUGmechanic]

Ok then, so far so good.
It is a lot of work to design a system by using only sketches and lines and constrain them in the right way.
Positive thing is it gives me insight in construction techniques and shows how abberation works.
I am wondering though if it is possible to design more efficiently by (in this example) reflecting a solid in a concave mirror (see attachment). That would save a lot of work.
The way I would do it now is create an intersection curve, draw a few lines normal to the mirror surface and starting at the intersection line, and do a sweep.

Any suggestions?



Older budweiser

 

[kdiomedes]

Sorry for the late reply but to back up what John said about the single focal point explained in simple math. If in the attached photo point A was the same as F point then the triangle CBA would be the only triangle on earth with its hypotenuse being the sum of the other two edges. ? angle is 18,004 degrees because the lines CE and DB are parallel.

As for your previous reply what I would do for a very small diameter beam is:
1)find the intersection point of the beam (centerline) on the concave surface,
2)draw a datum axis that would be normal to the concave surface but would pass through the point created earlier,
3)create a datum plane that would pass from the previous datum axis and the centerline of the beam,
4)create another datum plane, 90degrees turned from the previous plane that would pass through the datum axis
5)mirror the beam with the new datum plane.

All the above are OK for a centerline but for a beam with a countable diameter, things are more complex.