Hi Andrey,





I.95 means 95% transmission and 5% reflection. If you want a 95% efficient reflector then you would use I.05 for 5% transmission 95% reflection. You could construct a 95% reflector with no transmission named "someName" as "IDEAL someName 0 .95 1" for 0% transmission, 95% reflection and 100% reflection for TIR.





A common cause of too many intersection is the many internal Fresnel reflections inside a glass volume. This can be eliminated by setting a minimum ray intensity in the non-sequential section of the system explorer. For example, setting the "minimum relative ray intensity" to 0.01 means that any segment that falls below 1% of the initial ray intensity will not be traced. You should set this value to be high enough to still trace all the important energy, but low enough to prevent tracing rays that do not significantly contribute to the result. When you trace rays, the ray trace dialog will report the total energy lost as 'lost energy (thresholds)'.











 

---

I changed min rel ray intensity to 1e-1 and set # of analysis rays to 100, and there's 12 LEDs so it's 1200 rays, yet I still get the same error.





I removed the coating from the object and the object material is MIRROR.

---

Hi Andrey and David,





I just had a look at your file, and I believe I know what is causing the issue.





David is right, I.95 means 95% transmission and 5% reflection.





<<A common cause of too many intersection is the many internal Fresnel reflections inside a glass volume. This can be eliminated by setting a minimum ray intensity in the non-sequential section of the system explorer. For example, setting the "minimum relative ray intensity" to 0.01 means that any segment that falls below 1% of the initial ray intensity will not be traced. You should set this value to be high enough to still trace all the important energy, but low enough to prevent tracing rays that do not significantly contribute to the result. When you trace rays, the ray trace dialog will report the total energy lost as 'lost energy (thresholds)'.>>





However, from the archive file, I can see that the checkbox ''polarization'' is unchecked.


This means that the polarization is not taken into account, therefore, a ray keeps its intensity without any losses.


When a ray enters the CAD, it is traped in it and reaches the maximum allocated number of intersection:








All I did to solve the problem is to turn polarization on by checking the box in the settings of the raytrace:











Now, the losses are taken into account, and the ray is terminated when it reaches the maximum number of intersection or the minimum intensity.





All the best,


-Thomas





PS: the rayfile is here: https://www.osram.us/ledengin/products/luxigen/lz4.jsp

---

Thank you Thomas, now I'm able to progress but still running into this error when I try to increase certain limits.





Question#1: Why are the rays penetrating into the CAD object especially when object material is MIRROR? It's a solid body when exported, why doesn't glue distance affect the ray penetrating into the CAD object even if there are holes between the sheets of the object? Is there a way to prevent this penetration? I don't believe the rays are originating inside my object since they should be confined to the LED, which has a 5.6mm diameter lens and my object has a minimum clearance of 5.7mm dia around the LED.





Question#2: How does setting the object material MIRROR/ABSORB affect the rays that interact with the object when Coatings are used?


For example, if I use "IDEAL name T R TIR" -> "IDEAL aluminum 0 0.9 1" means 10% would be absorbed by the coating, so none of the rays would reach the object material correct?


What if I use "IDEAL aluminum 0.05 0.9 1", then 5% of rays would transmit through the coating to the object material, would they then 100% absorb and 100% reflect based on the object material?

---

I'm also curious why OpticsStudio is not saturating my CPU, it's at like 12.5% avg, no single CPU is saturated, quad-core running 8 threads. It's using 28GB of RAM out of 32GB, but it doesn't appear to be paging the disk. It's taking a long time to ray trace 12M rays, hoping there's a way to speed it up other than reducing the # of rays traced.

---

Hello Andrey,





-1 Raytracing in CAD objects is always error prone. Due to how CAD objects are defined, with faces and nodes, determining the surface normal can be tricky, and leads to geometrical error (for example, a ray hitting a node won't have a physically correct behaviour). However it doesn't happen a lot, so you can usually ignore the errors (they only represent a small amount of energy)





-2 If you trace with polarization on, it takes coatings into account


    If you uncheck polarization, it is binary, reflect or transmit (based on the property)





Best,

---

Thomas, can you please clarify on Question#2 what would happen to the rays that are transmitted through the coating, what do they hit next? The object material? So if it's a mirror material object, then does it hit the coating (95% reflection) and thus get trapped until the energy decays?

If I set the object to to ABSORB, and the coating to reflect 100%, I think that should solve my issues with forever bouncing rays inside my CAD object while keeping the simulation effectively an ideal mirror, right?

---

Answering my own question from help file: "Note that a material of ABSORB  absorbs all incident rays even if a reflective coating is applied to the face of the NSC object."

---