Skip to main content

Hello,

I'm trying to take an elliptical beam (example from a laser - 26deg in X and 37deg in Y) as defined by the POP settings, using Gaussian Angle, and use a polynomial lens to circularize it.

My problem is that I get a better fiber coupling efficiency using a standard circular aspherical lens than what the optimization wizard gives me for the polynomial lens.

Are my assumptions wrong from the start or am I simply doing something wrong with the software?

In steps:

I define in the POP settings a Gaussian beam with an elliptical profile:

In one simulation, I select an Even Asphere and using the POPD merit function, I get 93% mode coupling efficiency from a lens initially designed for a circular beam. The initial optimization wasn't even for the elliptical beam. It was for a circular one.

But when I change the lens surface to a Polynomial one and set the different X and Y coefficients to variables, the simulation seems to only give me below 90% mode coupling efficiency. It never seems to really circularize the beam.

Should I have picked a different freeform surface? Is my understanding of what I want to do not matching with what the software is doing?

I sometimes see enormous Merit Function Editor lists of parameters and mine has one…

Can anyone identify a problem with either my assumptions or my simulations?

Invariably, thank you for the assistance.

Hey Tiago,

When you say “polynomial lens”, are you referring to an Extended Polynomial surface?  When you try to optimize this, what coefficients are you using?

To truly decouple an Extended Polynomial surface into a biconic/anamorphic, you should:

  • Not use any cross terms.  You should only make variable the terms with a power of 0 in either the X or Y, so the coefficients that look like X2Y0 or Y4X0
  • Since you have mirror symmetry, you should make sure you don’t use any odd terms in your coefficients, otherwise you won’t be able to circularize your beam
  • Like all aspheric optimizations, you should “walk” you optimization from the lowest order coefficient needed and only add the next coefficient if the figure of merit is not good enough (i.e., do not set 20 coefficients as variables and simply hit Optimize)

A slightly more simple surface might be the Biconic Zernike surface.  This elements the cross terms in the first bullet point but you should still only set the even coefficients as variables.

 


Hi and thank you for your clarification.

I was using a standard polynomial lens, and setting the first terms to variables, both X and Y. Was that wrong? Why do I need an extended polynomial instead of a polynomial surface?

I added this picture to my previous post but may have been too small, not clear and not seen:

Also, could you clarify this section?

  • Since you have mirror symmetry, you should make sure you don’t use any odd terms in your coefficients, otherwise you won’t be able to circularize your beam

Why is that so? I must say the field of optics and these simulation software are not my background (I'm in electronics and micro fabrication).

Nonetheless, I'm already going to try your suggestions.

Thank you for the explanation and clarification.


Hey Tiago,

The Polynomial surface should work.  

Another thing to keep in mind is that POP itself is not directly tied to the System Explorer’s Aperture settings so the 2D/3D Layout that you see might not reflect the actual starting beam shape of the POP analysis.  The default Merit Function Wizard (which gives you the enormously long list) will optimize on the System Explorer’s Aperture settings while the POPD operand will use the POP settings (after you hit Save in the bottom right of the POP Analysis Settings).  The DMFS will be significantly faster (and more likely to converge) than the POPD but the POPD might be able to guide the optimizer to make the beam circular when you use Xtr1 of 23 & 24.  The more operands you have, the easier it is for the optimizer to converge to a solution (you can think of an operand is a suggestion/guide to the optimizer for the “next step” in the optimization process)


Hello and again, thanks for the information. I'll try and simulate what you mentioned.

I must say that as a new user to this software, many of the things are quite daunting and I have a hard time figuring out which path to take, and why. I read the manual and I see where those values you mention are but have a hard time connecting all of them.

When I go online, I see long lists of operands and Merit Functions and that just seems dauting to grasp.

What would you say is a good path towards learning some of these things and how the software works. For me, after having some questions answered by David Nguyen (How to design a prism with a lens (one continuous solid)? | Zemax Community), I see my biggest use for this software from freeform lens design. Elliptical lenses to collimate elliptical beams or focus them in a circular shape. Or even a combination of two lenses, one cylindrical and one aspherical to shape the beam.

I see there is a Freeform Optics course in the Ansys Learning Hub subscription course. Would you say this would help for my needs?

Could you point me in the right direction to better learn some of these things or would you say simply coming here to the forums with my questions is better so that I share my problems, for others to learn from them as well?

Thanks for the help.


Reply