Solved

Non sequential trace with TIR in annular object

  • 3 September 2022
  • 7 replies
  • 237 views

I would like to trace rays into this object and let them TIR around the inside.

The attached gives an idea but is not correct.  I have used the swept object to generate the annulus.

It has in outside radius of 4mm and an inside radius radius of 2mm, concentric surfaces, Germanium.

I don’t know how to get rays in or out.

I would like the source rays to be collimated, or converging, or diverging, or astigmatic.

Can anybody help!

I know that I don’t know how to do the entrance or exit surfaces properly.

Harvey Spencer

glasair470@gmail.com

icon

Best answer by Ray 6 September 2022, 09:28

View original

7 replies

Userlevel 7
Badge +2

Hi Harvey,

 

When you say that your model is not correct, can you tell us what is wrong about it? Also, could you share your file with us?

Another object to consider is the Torus Volume. I’ve tried to replicate your model with the Torus Volume and a Source Ellipse, it gives something like so:

I’m also attaching this file to my answer (TorusVolume.zip).

You are also mentioning Entrance and Exit Ports, which are features of hybrid systems (sequential + non-sequential), and this might be an issue. For a ray to be propagated through a hybrid system, it has to enter through the Entrance Port, which should be fairly ok since you have a sort of well defined beam that enters the torus, and it has to exit through the Exit Port, and this is an issue. The Exit Port is a plane, and you have rays exiting the torus at different locations and angles (as shown in my screenshot). Only a small fraction of those rays would make it to the Exit port. I’ve tried to show this using the same Torus (I just deleted the Source Ellipse), but I had to move it to satisfy the Entrance Port requirements (more about that below):

As you can see, only the rays making it to the IMAGE surface are effectively traced through the system. Therefore, I suspect a lot of sequential analysis features will fail. For example, if the chief ray doesn’t make it to the IMAGE surface, the calculations that are based on the chief ray optical path length are impossible to make.

For your information, here is how I setup the object in sequential mode (this is how to setup your Entrance/Exit Ports). Have a look at this article also, which describes how to set the ports more generally.

First, to view the Entrance Port, use the Draw Ports? flag of the Non-Sequential Component Surface in the Lens Data Editor with a value of 1. If you toggle between the value 0 and 1, you can view where the Entrance Port is. Here is a screenshot with the value 0:

Here is a screenshot with the value 1:

Do you see the circle that appears at the centre of the Torus? This is your entrance port. Next, I’ve moved the Torus in the Non-Sequential Component Editor to Z Position = 5, and Y Position = -4:

The rays are not tracing through the system yet. This is because we need to setup the Exit Port. Then, if you toggle the Draw Ports? to a value of 2, this will show you the Exit Port. By default, it is overlaid with the Entrance Port (it would be the second circle to the right in the image above). We can change the Exit Port location with the columns Exit Loc X/Y/Z, and I’ve applied the value Exit Loc Z = 15 and I’ve change the next surface Semi-Diameter (this is the Exit Port surface in fact) to make it larger (10 mm radius):

I’ve attached this file to my answser also (TorusVolumeSEQ.zip).

Let me know if this helps.

Take care,

 

David

 

Thanks for the help!  I have since tried some stuff.  Actually the shape is not a torus by a slice of a cylinder with a hole in the middle, so the annular volume is perfect when rotated 90 degrees to let the light enter the “side” and not the front.  

I need to put a pair of flat facets on the outer surfaces to couple the light in from the source and out to a detector.

The locations and angles of the facets are not defined at this time.  Ideally, the light would make about 9 or 10 bounces around the ring and then exit.  

I need to model the source as either collimated, converging, or diverging light.  Don’t  know how to make a converging source with just a source object without using a focusing lens.  I suspect that to preserve beam quality you would want to start convergent light (maybe toward the center of rotation of the shape?) to overcome some of the net negative power of the neg/positive reflection pair.

I have attached the lens so far.  If you could suggest some additions to make the source work and how to include the facets better than this method of nested objects  I would be thankful.

Harvey

Userlevel 4
Badge

Hi Harvey,

 

To make a simple convergent beam in Non Sequential mode, you can use the “Source Distance” and make it negative. -10 means the rays converge to a point at 10mm from the source. You have to keep the cosine and Gauss factors at 0.

With a positive value, the source is divergent, with 0, it’s collimated. It however will act as a “point source”, not a realistic extended source. If you want a more extended source, as long as it is on axis, the LambertianOverfill DLL can be really useful (make the target diameter smaller than the source and your beam converges on the target.

Userlevel 7
Badge +2

Hi Harvey,

 

I see, sorry I didn’t understand the problem from your Power Point file earlier. I guess @Ray answered your Source question.

Regarding your other question, I’m not sure I also understand what you mean with a pair of flat facets are you talking about the Rectangular Volume (highlighted in orange in your last screenshot)? Nesting objects in this way is actually quite efficient. What do you mean by including the facets in a better way?

Take care,

 

David

Yes.  I want a flat  that surface that can be tilted at an arbitrary angle through which the source light enters the annular volume.  So that the entering beam does not pass through the cylindrically curved annular surface first, but is coupled into the ring through a flat portion.  Likewise, I want to be able to include another one somewhere TBD around the ring to couple light out.

 

Thanks for all the help.  I have managed to learn a bit about NSC and hove gotten it to work.

But there is always another question.  There are sneak paths that do not TIR of the inner surface of the donut and go into the center of the donut, which is modeled as wtaer.  I need to keep it as water so as to get the correct TIR at the surface between the Ge donut and water but would like the water to completely absorb the rays that don’t TIR and enter the water so they don’t come back out and into the donut again.  Any way to do that?

 

Userlevel 3
Badge +2

Hi @glasair,

I’m glad to hear that you’ve gotten your NSC model to work! Sorry about the delay in our response. In order to absorb the rays that do not undergo TIR, you could create an additional surface in the center of the structure, like the one pictured below, that has a coating set to ABSORBING.

 

 

Best,
Ethan

Reply