I recently read that Ansys is moving forward with AI tools, implementing Copilot in other software, but nothing about Zemax. Is there any research done about it? Any estimated date?

Any thoughts to the above?Imagine I’m creating some linear geometry, like a trough concentrator. I’d like to be able to optimize some of the geometry as defined in the x-z plane, but this geometry is just extruded some length along y.For a simple parabolic concentrator, the biconic surface works well, since one can define the y-radius to be 0 (infinite) and the x-conic to be -1. Set the rectangular flag and the extrusion length is defined by the max and min y-apertures. CPC rectangular is obviously intended for this purpose, but it requires a non-zero aperture so, as far as I can tell, one cannot generate “closed bottom” troughs...But what if the trough profile is something more complex? The help file for the extruded object suggests the extrusion can only be along the z axis, but maybe that is in the local coordinate system, and this limitation can be overcome by a rotation of the object? Then how to define the profile? The help file for the extruded object points to the user defined

Hello again friends from the Zemax community, I’m trying to optimize an optical system in where I have an input spherical wavefront (a single point) and the output is another spherical wavefront with a changed radius of curvature.  Since my image is not real at the target output plane, how could I use zemax merit function definitions to optimize such a system? Would optimizing for the spot size/wavefront error at the “virtual” image point in image space be a suitable solution here? Or is there any other approach that I could follow? Thanks a lot for your feedback and comments

Dear Zemax Team, I shortly designed a high numerical aperture system (NA>0.9) and did a tolerancing. The system is a rotational symmetric objective with spherical surfaces only. Useably quite simple for a ray tracer. Subsequently, I performed a MonteCarlo tolerancing by slightly shifting lenses and surfaces and by modifying surface thicknesses by a bit. Almost every system that is created in the MC analysis has a wavefront deformation as follows. It shows a large discontinuity at the edge of the pupil:  Obviously, this WFRMS is not real and hampers an evaluation of the system. If the entrance pupil radius of this tolerated system is slightly reduced (less than 1%), the resulting wavefront looks as follows. The discontinuity disappears since it originated at the edge of the pupil: Ray aiming will not help since the stop is in front of the objective. I tried it nevertheless, it doesn’t change anything.I am estonished since the displacement of surfaces is in the order of 1 µm, I didnt

One year ago, ​@Xianglei wrote a topic (thanks a lot!) about:Since its been quite painful to generate such plots with ZPL (thanks to ​@Erin.Elliott, we might get a fix in 2025 🤞):I’ve been using the ZOS-API option more and more to create animations. It is more stable, but currently lacks a couple of options. Since we’re approaching the Christmas period 🎅, this is my wish list to improve the ZOS-API Layout Exports feature (in order of importance):As far as I understand, we cannot specify the file type of the export. I’ve tried to use multiple extensions in OutputFileName (.JPG, .PNG, ...etc.) but the file size remains the same. Based on the file size, which is rather small (for a given image size), I’m assuming the export is made in .JPG regardless of the extension specified. Can the file type be added to the ZOS-API? Could we get a vector graphics file type? Although there’s a property RaysLineThickness, for me, it doesn’t seem to have an effect. If someone is anyway going to look at

Hello everyone, I was working on tolerancing a new design and comparing “old method” TFRN+ TIRR, to the ‘new method’ ISOA/ISOB/ISOC, and it seemed to me that the new method was more ‘pessimistic’ that the old one. Looking in details to compare the methods, I made a very simple system (1 PCX lens, R=50mm, N-BK7, Aperture 10mm) and compared different combinations :ISOC alone TIRR + ISOC ISOB + ISOCLooking at the results I stumbled upon the fact that the contribution in sensitivty of ISOC was not the same depending of the associated operand :ISOC alone : ISOC       2                 -0,50000000         0,00399992         0,00215149         0,50000000         0,00399992         0,00215149TIRR + ISOC : ISOC       2                 -0,50000000         0,00399992         0,00215149         0,50000000         0,00399992         0,00215149ISOB+ISOC : ISOC       2                 -0,45000000         0,00790629         0,00605786         0,45000000         0,00790629         0,00605786(Here I red

In the Merit Function Editor there is the option to generate a Merit Function Listing. At the top of the Merit Function Listing three summarizing lines are listed: “Merit Function Value”, “Constraints Merit Function”, and “Sequential Merit Function”. The “Sequential Merit Function” value results from the used “Default Merit Function” and the “Contraints Merit Function” value from all the operands added to the Merit Function (e.g. contraints controlling geometry, MTF, distortion...).My question: How do the two component values combine to yield the “Merit Function Value”?

I’m strictly in sequential mode, but I’m not getting rays in the generated STEP file.All surfaces selected, Number of rays = 3Ray pattern = XY, Ray layer = 1, Lens layer = 0Wavelength = mid value,  all fields selected, configuration 1 is selected, tolerance = 1e-6File type = STEP, Parasolid set to V31 (current)System has zero vignetting, but no rays whether the Delete Vignetted box is checked or not. Surfaces as solids box is checked.Is 2026 R1 broken?  Help, anyone? 

Hi All, Is there a simple way to model a cylindrical fresnel lens in NSC mode?Didn’t find anything in the knowledge base. Thanks,Nadav

Hello,After following the next article: The easiest way to tolerance an off-axis mirror I have been struggling to manage the tolerances in my system since it has 2 different OAPs and as in the example, I have set one of them as the global coordinate reference Apart from that issue, I really don't think my coordinate breaks are working properly to tolerance the OAPs (I only want to check the decentered and tilt tolerances).Lastly, I don't really know why final focus is not hitting the center of the image.Thanks

Is Erin's guide (linked below) still the easiest method?  I've just learned about the new nesting feature, but that's not an option for Pro users.https://community.zemax.com/people-pointers-9/the-easiest-way-to-tolerance-an-off-axis-mirror-4756

I have implemented a simple system composed of:Diode Source with 2deg divergence in x and y; Diffuser implemnted with cylinder volume with a single scaterring surface following a gaussian model with sigma = 0.27 Polar detector to monitor the angular distribution of energy after the diffuser in the range -90, +90 deg; After exporting the detector data using the ”Detector Viewer", I have fitted the data points with the gaussian model described in the Help documentation: BSDF = A*exp(-x**2 / sigma ** 2):The value of the fitted sigma (sigma_i = 23.194) has nothing to do with the specified sigma of the diffuser (sigma = 0.27), which tells me that my interpretation of the scattering model is wrong. Where does the discrepancy come from? 

Hi all, Don’t ask me how I got there… but I found a possibly undocumented (or not well documented let’s say) behavior of RemoveWavelength in the ZOSAPI (SystemData.Wavelengths).In the user interface, if you keep removing wavelengths from the system, eventually you get down to a single wavelength, and this one cannot be suppressed. The enable checkbox grays out.If you use RemoveWavelength enough times with a piece of code like so (I am using ZOSPy):while True: if not oss.SystemData.Wavelengths.RemoveWavelength(1): breakprint(oss.SystemData.Wavelengths.GetWavelength(1).IsActive)print(oss.SystemData.Wavelengths.GetWavelength(1).Wavelength)The prints will be as follows:False0.0And if you use AddWavelength afterwards, this new wavelength will be at position 1 in the Wavelength Data Editor.I am mentioning it because normally the ZOSAPI follows closely what the user interface does and this seems to be an exception. I like the exception because it means we can really clear the editor

I am in non-sequential mode trying to design coupling optics to fiber couple a Lambertian source into a 980um/0.57NA fiber.  After completing my design which consists of two even aspheric lenses, I want to perform a tolerance analysis.  In sequential mode there are many operands available to tolerance lens surfaces. How do I go about tolerancing the lens surface of each aspheric surface in non sequential mode?  What options do I have for lens tolerancing in non sequential mode?

Hi all,First things first, I am still using Zemax 2012, so this may be a limitation. I have a relatively simple setup: input fibre, half ball lens, PCX lens to collimate a beam, the beam passes through a test glass, and is collected onto an output fibre using a PCX lens and half ball lens combination. I optimise the energy throughput from fibre to fibre (using IMAE). The lens and test glass diameters are fixed. The optical beam is allowed to overfill the first PCX lens, but any rays passing through the first lens MUST NOT overfill the test glass or the second PCX lens. For short throw distances (distance from the first PCX lens to the test glass, and from the test glass to the second PCX lens) the Hammer Optimiser works just fine. I have the IMAE merit function to optimise the energy throughput, and I have MXSD merit functions on the test glass and second PCX lens to ensure that the beam is completely within them (no overfill).For short throw distances (like 500mm) the optimisation wor

The manual states that ISFN returns a paraxial-only f/# and WFNO returns a real-ray f/#.  Nothing is said about ISFN accounting for vignetting.  WFNO allegedly accounts for vignetting following a SETVIG or SVIG operand.  However, I’m not seeing this happen with WFNO in a vignetted system, as the focal ratios are identical to 6 digits for both unvignetted and vignetted analysis.  I’m also not seeing either SFNO or TFNO accounting for vignetting, either.  The manual does not state whether they are based on paraxial or real-ray analysis.  Is this another example of users having to write a macro to make Zemax do what the manual says it is supposed to already be doing?  

Does Zemax have the internal to export the envelope of beam footprints on surfaces to a mechanical engineering file?

HiI am trying to analyze scatter contribution in my imaging system for off axis mirrors using different surface roughness. For example I have calculated the A, B & g values for 10nm surface roughness and used those values to create an ABg file in the ABg scatter catalog,the TIS calculator showed 0.0058030 (The value is less than 1%).While trying to assign this created ABg file to the mirror the following error on Total integrated scatter pops up.Can anyone help me in understanding this error pops up  when my calculated TIS is less than 1% ? 

I am using us_array.dll to represent the lenslet array in sequential mode. Is there any way to rotate the lenslet array by an angle?

And also, how do I know which one to use when comparing against real life lenses and optical systems?

Hi guys!I am working on a tolerancing analysis of an optical system that includes a beam splitter, and I have come across a coupling issue that I would like to understand better.In the nominal configuration, all elements are centered and the system is fully aligned.When I introduce a Y decenter of +20 mm on the beam splitter on the lens data editor, this physically displaces the element along the vertical arm of the system. As a result, the optical path length in that arm should change accordingly. A +20 mm decenter on Y should reduce the effective thickness by 20 mm, from 239.617 mm to 219.617 mm. However, Zemax keeps the thickness at 239.617 mm regardless of the decenter, which is not physically correct.My question is: is there a way to link the Y decenter of the beam splitter to the thickness of the preceding surface, so that when the beam splitter is decentered by a given amount the thickness is automatically compensated by the same value? Ideally I would like to keep using TPAR to

Hi All,I am currently designing a low‑cost imaging lens for a sports application. In my lens model I use the REAY operand to control image height, the RAID operand to control chief‑ray angle (CRA), DIMX to control f‑theta distortion, LACL for lateral‑color correction, RSCH for spot size and MTFT/MTFS for resolution across the field. During optimization I can obtain either good MTF or good lateral color with CRA and distortion depending on the operand weightings, but I cannot find a stable balance even when I adjust the weights gradually. In both cases, my geometric spot‑size results were poor. Please suggest optimization techniques to achieve a better trade‑off, or alternative approaches to indirectly control these parameters.Note: candidate glass materials must have hardness greater than 450. Thank you in advance for your support.

Dear all, In the manual, it is written : “Samp: The grid size. A value of 10 would yield a 10 x 10 grid of rays”. I need to use CENX(Y) with the highest accuracy. What is the maximum acceptable value of samp ? Is there another way to increase the accuracy of CENX(Y) ? Many thanks. Best regards, Thierry. 

Dear all,I’m working on a project of ultra-precise micro-arcsecond relative astrometry space observatory (Theia). Especially, I need to map the image plane (0.5°x0.5°) with a very high accuracy. For instance, I have to study the variations of positions of the PSF (with CENX/CENY) due to some perturbations (distortion, tilts of the mirrors...).By default, the number of fields is limited to 2025 for my premium version, but I would need 100x100 fields.It’s rather easy to generate such a new grid, but the parameter “i” (field number) of CENX/Y refers to the (default) field table.Is there a way to use CENX/Y with this new grid ?Many thanks,Thierry.

Hello,I would like to model the propagation of a polarized light beam through a biaxial crystal.While methods for uniaxial crystals seem to be well established, I am unsure how to approach this problem for biaxial crystals.How do I handle this case?Thank you very much for your help.Best regards,Dirk