Skip to main content

Hi everyone,

 

I find that one of the most useful tools in OpticStudio is the Physical Optics Propagation tool. However, I think that it is also one of the most challenging tools to use, especially when the model is composed of a relatively large number of elements, since one could easily find sampling issues or inadequate grid size which could cause some artifacts to appear downstream, and one should check the intensity and phase plots for each surface, and there is a lot of adjusting parameters like sampling, pilot fit, re-compute pilot beam parameters, etc, that can have a large impact on the results.

 

 

The first thing that I usually do even when the resulting irradiance and phase plots at the image look reasonable, is to check if there are any warning messages at the Prop Report tab. The most common warning messages that I get are:

 

 


  • Warning: Low sampling of pilot beam detected
  • Warning: Inadequate guard band detected
  • Warning: Transfer function may have too many waves of phase to accurately model beam

     

 

For these, I found good resources to address them here: 

 

 

However, I recently got a warning message that I never got before:

 


  • Warning: Pilot beam waist smaller than wavelength detected, scalar diffraction propagation algorithms may be inaccurate.
     

I'm having a hard time trying to solve it. My system consists of a collimated laser beam of Ø7 mm refracting and reflecting through several components of an optical bench, and finally expanding through an afocal telescope to free space with a diameter of Ø130 mm. The beam shape after the telescope is no longer Gaussian since there is a central obscuration at the telescope exit, but in the far-field, the irradiance pattern recovers the Airy pattern and expands with some divergence which I'd like to quantify.

 

 

The warning message that I get is shown here:

 

 

 

 

 

201230-025012-pop20pilot20beam20waist.png

 

 

The wavelength that I'm using is 1.55 um, and the pilot beam waist at this surface is 1.19 um. I've tried recomputing the pilot beam parameters at that surface, at the surface before and/or the surface after, changing the Output Pilot Radius fit criterion to plane, resampling after refraction, and cannot see that there is a value with which I can get rid of that warning message.

 

 

Probably related to that warning, I start to see some artifacts in the far-field like these:

 

 

  201230-025758-pop20pilot20beam20waist20-20intensity.png

 

 

201230-025814-pop20pilot20beam20waist20-20phase.png

 

 

 

 

 

Moreover, when I try to calculate the divergence as very well explained in the Divergence of a POP beam post, the results that I get are really much larger than what I'd have expected:

 

 

201230-030306-image.png

 

 

For comparison, the analytical results for a Gaussian beam with a beam waist of 65 mm at 1.55 um is about 7.6 urad, and I'm getting several times that reference value, which can also be noticed by the M-square computation.

 

 

I'm attaching the ZAR file as a reference.

 

 

So, I have the following questions:

 

 


  1. Is there a recommended approach to recompute or override that pilot beam waist with confidence?
  2. Would you recommend to recompute the pilot beam parameters in the far-field, after the Rayleigh length? 
  3. And finally, is there any list of POP warning messages and how to solve them? I couldn't find this listed in the help files, and I think that would be a great reference to have. 
     

Looking forward to receiving some hints.

 

 

José

 

 

 

 

Dear José,



The issues you are looking at are indeed connected. Generally, the algorithms used in POP are only good for slow beams. Fast beams, i.e. beams with large convergence or divergence, cannot be accurately modeled with scalar diffraction. When the pilot beam waist becomes so small that it appears to be smaller than the actual wavelength of light used, this is a good warning that the beam is too fast. There is not generally a POP setting or tool to get around this because the problem lies with the fundamental algorithm.



The discussion of this can be found in your Help system, on the page at The Analyze Tab (sequential ui mode) > Laser and Fibers Group > About Physical Optics Propagation > Algorithm Assumptions. This page also includes the likely solution. Fast beams tend to be accurate enough with Geometric optics so that diffraction effects are minimized. A reasonable rule of thumb with POP is that if POP can't handle the physics, you probably don't need it anyway, and should use Geometric. And if Geometric effects are overwhelmed by diffractive effects, that is a good time to use POP.



You can always use the surface properties to tell POP to use rays for a given propagation between surfaces. In the Physical Optics section of the Properties, the first check box is 'Use Rays To Propagate To Next Surface:'. If you check this, you can just use rays and ignore the problems in that portion of the system.



Kevin


Dear Kevin,

Is there a way to model  a physical optics propagation  of “fast beams” in ZEMAX?


Hi @Vadim.Chuyanov - I see you moved your question to a new post (here). I’ve asked Kevin to take a look to see if he has anything to add to Jeff’s response. Thanks!


Hi @Vadim.Chuyanov, (and thanks Allie for directing my attention here),

Jeff’s reply in the other thread seems pretty sound to me, though I can’t vouch for his DIFFRACT software link. He is correct that handling a fast beam does take a vectorial approach that the POP option in OpticStudio does not have. In many cases, fast beams are dominated by their geometric aberrations, in which case the additional information gleaned from diffraction is minor by comparison. If you cannot find a combination of sampling sizes and pilot-beam recomputations and field resamplings and other options, it may be necessary to switch the fast sections of beam to a ray propagation by clicking on that box in the surface properties.


Hi Kevin,

This is a slow system.   I don’t see why the pilot beam size would be so small at surface 52…  Something weird is going on.  

Jose, I can't reproduce the irradiance plots in your post. They say surface 71 but your ZAR file only has 65 surfaces.  When I open the file the irradiance at the last surface is zero...


Reply