Hi mcwikla,

If I try this on a dummy system, it seems to make some changes for me. Those changes are subtle though (but it sort of make sense since this dummy system doesn’t present that much of chromatic aberrations). The left Extended Source Encircled Energy is with all weights to 1 (and then I locked it), the right one only has the weight of 1 at the primary wavelength (other wavelengths are 0):

Do you have an example to share for us to troubleshoot? What level of chromatic aberration is present in your system?

Take care,

David

@David.Nguyen Thank you for responding. Please try applying different value from 0 or 1, for example 0.5. I was expecting that weight will define maximal normalized energy, namely if weight of 1 are set energy is normalized to 1 but if 0.5 is set, energy is still normalized to 1.

Hi mcwikla,

Left is again all weights at 1.0. Right is with the last wavelength weight at 0.5:

The values are close, but they are different. Have you tried increasing the sampling? I’m using 100 in the Rays x 1000 setting.

Take care,

David

@David.Nguyen 

Yes I can observe minor differences for increased sampling but how OpticStudio calculates confuses me and I didn’t find it explained neither in help pdf nor in blogs.

My goal is to evaluate energy reaching the detector from broadband light source. For example, let’s consider broadband source characterized by maximum intensity at 600 nm (weight = 1.0) and half intensity at 500 nm (weight = 0.5). Assuming 100% coupling efficiency for both wavelengths at the detector (image plane), calculating encircled energy for both wavelength will be normalized to 1.0 where I would expect to see it being normalized to 0.5 for 500 nm.

For now, only partial solution I see is to evaluate encircled energy for all wavelengths with weights of 1.0 and subsequently multiply it by normalized broadband light source spectral characteristic, but I was hoping to avoid it.

Hi mcwikla,

Reading the Help File about the Extended Source Encircled Energy under the settings Rays x 1000 it seems that:

The distribution of rays at each wavelength is in proportion to the wavelength weights.

My understanding is that you can choose a certain quantity of rays to be traced with the Rays x 1000 settings and then the rays are split in wavelengths according to their respective weight. For example, if you choose 100 000 rays and you have two wavelengths, one of them has a weight of 1.0 and the other has a weight of 0.5. Then, 66 667 rays will be traced at the first wavelength and 33 333 rays will be traced at the second wavelength.

This analysis gives the Fraction of Enclosed Energy, so, I think it makes sense that it is normalized to one somehow. But I also understand your problem: you want to evaluate the absolute encircled energy between the two wavelengths. Your partial solution makes sense to me, assuming you have perfect coupling as you said. I guess what you want is to trace the same number of rays for each wavelength, but the rays would carry a different energy depending on the weight of their corresponding wavelength. Have you tried using the non-sequential mode for doing this analysis? The non-sequential mode doesn’t have an analysis feature that would extract the encircled energy directly, but it can be computed from a Detector image, and it could be easier to define your source spectrum in non-sequential mode.

Take care,

David

@David.Nguyen 

It is possible that minor differences in Encricled Energy that we observe are just result of error - as you said: reducing weight, reduces number of rays, thus introducting bigger error - it makes sense to me.

Using NSC it’s possible to evaluate absolute energy for each wavelength using “Flux vs wavelength” analysis and yes it’s great, but for systems with low efficiency (in my case around 5%) you need a lot of rays to get reliable results becouse most of them are clipped, thus calculating same thing in SC is much faster.

I am wondering if SC mode allows to get absolute value of optical power reaching the image plane (or any plane). I know there is IMAE operand but it’s only gives efficiency with relation to starting value, thus it’s not influenced by wavelength weight.

Hi mcwikla,

I see, it makes sense.

I’m not sure what the best way would be, hopefully someone else can help you, but I know you can read a ray intensity from the ZOS-API batch raytrace tool.

Take care,

David