In non-sequential mode, using two coaxial cylinders to represent the core and cladding should work okay for simulation of a multimode fiber (MMF), but there are a few details to take into account.  First, in the layout spreadsheet, the cladding cylinder should be listed before the core cylinder.  That way the nesting rules associated with object placement will ensure that the core replaces the appropriate portion of the cladding cylinder.  Here’s an example using the NA = 0.22 core and cladding refractive indices with a simple 0.20 NA source.

The coupling efficiency is 92.5%.  The 7.5% loss is simply Fresnel reflection from the input/output surfaces.  This raytrace was done with ray splitting (and hence polarization) turned on, as well as using a linearly polarized source beam.  If the fiber NA is increased, the coupling will remain unchanged as expected since all of the non-reflected rays are being guided.    

However, if the NA of the source is increased to 0.30, you can see cladding rays show up.  So, next, it’s a good idea to use an annular surface at the fiber output as an aperture that blocks cladding mode rays but passes rays that are guiding in the core (i.e., the absorbing annulus has an inner diameter equal to the core diameter). 

Also, the outer surface of the cladding cylinder should be made absorbing so that rays don’t TIR off of this surface and make it to the detector.  Now the coupling efficiency is reduced to 54.5%.  The example file is attached.

This approach for modeling MMFs is fine, but the fiber length is limited by the maximum number of allowed ray TIR bounces inside the core (which is 4,000).

Based on the maximum NA of the guided rays, this typically corresponds to a fiber length in the range of a few meters.

However, there is a simpler and more computationally efficient way to find the coupling efficiency for MMFs.  That is to recognize that a MMF is basically a spatial and angular filter for rays, and to use a detector that accounts only for those rays that make it through the filtering process.  This can be readily done in sequential mode as described here: How to model multi-mode fiber coupling.  A similar approach can be implemented in non-sequential mode using only a detector (no need to physically model the fiber).  This is done by placing a detector where the input plane of the fiber resides, inserting an annular aperture immediately in front of the detector (to form the spatial filter), and then using a filter string to look at only those rays that are contained within the acceptance NA of the fiber (i.e., the angular filter).  Here’s the layout with a input beam having an NA = 0.30:

To implement the angular filter, we note in the help documentation:

For a fiber NA=0.22, the maximum angle of incidence for guided rays is 12.7 degrees (with cos(12.7) = 0.9755). 

To get the coupling efficiency we run a raytrace and save the ray file. 

Then, in the detector viewer, we simply use this ray file and apply the appropriate filter string:

We then see that the coupling efficiency is 56.5%, which is slightly higher than the value we found previously using the coaxial cylinders, but of course this latter approach does not include Fresnel loss so we would expect a somewhat higher value.

Lastly, to model coupling into single-mode fibers, which involves a mode overlap calculation, the best approach is to use the various tools available in sequential mode. (see Single-mode fiber coupling in OpticStudio)

In non-sequential mode, using two coaxial cylinders to represent the core and cladding should work okay for simulation of a multimode fiber (MMF), but there are a few details to take into account.  First, in the layout spreadsheet, the cladding cylinder should be listed before the core cylinder.  That way the nesting rules associated with object placement will ensure that the core replaces the appropriate portion of the cladding cylinder.  Here’s an example using the NA = 0.22 core and cladding refractive indices with a simple 0.20 NA source.

The coupling efficiency is 92.5%.  The 7.5% loss is simply Fresnel reflection from the input/output surfaces.  This raytrace was done with ray splitting (and hence polarization) turned on, as well as using a linearly polarized source beam.  If the fiber NA is increased, the coupling will remain unchanged as expected since all of the non-reflected rays are being guided.    

However, if the NA of the source is increased to 0.30, you can see cladding rays show up.  So, next, it’s a good idea to use an annular surface at the fiber output as an aperture that blocks cladding mode rays but passes rays that are guiding in the core (i.e., the absorbing annulus has an inner diameter equal to the core diameter). 

Also, the outer surface of the cladding cylinder should be made absorbing so that rays don’t TIR off of this surface and make it to the detector.  Now the coupling efficiency is reduced to 54.5%.  The example file is attached.

This approach for modeling MMFs is fine, but the fiber length is limited by the maximum number of allowed ray TIR bounces inside the core (which is 4,000).

Based on the maximum NA of the guided rays, this typically corresponds to a fiber length in the range of a few meters.

 

However, there is a simpler and more computationally efficient way to find the coupling efficiency for MMFs.  That is to recognize that a MMF is basically a spatial and angular filter for rays, and to use a detector that accounts only for those rays that make it through the filtering process.  This can be readily done in sequential mode as described here: How to model multi-mode fiber coupling.  A similar approach can be implemented in non-sequential mode using only a detector (no need to physically model the fiber).  This is done by placing a detector where the input plane of the fiber resides, inserting an annular aperture immediately in front of the detector (to form the spatial filter), and then using a filter string to look at only those rays that are contained within the acceptance NA of the fiber (i.e., the angular filter).  Here’s the layout with a input beam having an NA = 0.30:

To implement the angular filter, we note in the help documentation:

For a fiber NA=0.22, the maximum angle of incidence for guided rays is 12.7 degrees (with cos(12.7) = 0.9755). 

To get the coupling efficiency we run a raytrace and save the ray file. 

Then, in the detector viewer, we simply use this ray file and apply the appropriate filter string:

We then see that the coupling efficiency is 56.5%, which is slightly higher than the value we found previously using the coaxial cylinders, but of course this latter approach does not include Fresnel loss so we would expect a somewhat higher value.

Lastly, to model coupling into single-mode fibers, which involves a mode overlap calculation, the best approach is to use the various tools available in sequential mode. (see Single-mode fiber coupling in OpticStudio)

 

Dear Sir, 

Thanks for explaning the important things in the above discussion. 

One more doubt I am having How you are changing the NA of fiber, by changing the core and cladding R.I. only?

( Because  Fiber NA = sqrt(core_R.I.^2 - cladding_R.I.^2), so NA is depending on the core and cladding R.I. only!)

Like attached screenshot, how I am changing the R.I. in zemax spreadsheet.