Solved

Tolerancing/Optimizing w/ STAR

  • 15 August 2022
  • 2 replies
  • 138 views

Userlevel 1

Hi all,

I’m currently working on fitting FEA data to our system using the STAR module.  I’m able to fit the FEA data just fine but I’m looking to see if there are ways to compensate with element translation.  Is there a way to have the FEA data follow it’s parent surface while optimizing?  Would the same be possible in tolerancing.

icon

Best answer by Christophe.Weisse 1 September 2022, 11:29

View original

2 replies

Userlevel 3
Badge +1

Hello Alex,

Very interesting question - it gives the opportunity to share the answer with the community.

There are no OpticStudio features that are unavailable or grayed out when you import FEA data with the STAR module. You are then able to run an optimization or a tolerance analysis, but there are a couple of things you’d need to consider to have meaningful results.

First of all, strictly speaking if you change the system you’d need to re-compute the FEA data (including modifying the mechanical parts if the change is significant). The STAR module API is then key to automate the workflow.

If the change is small and/or the parameter you vary relates to an actual possibility (like moving some elements to re-focus for instance), then a unique set of FEA data could be a close enough approximation.

Then, different scenarios should be considered as well :

  • for surface tolerances, the STAR data must be refit any time the shape of the surface is changed. Including radius of curvature, and Zernikes for an irregularity surface. Currently this can be done through the STAR API.
  • for element tolerances, like the translation that you mention, the deformations will “follow” the part (if the part is moved, but the shape isn't changed) if the FEA data is in local coordinates. Data in global coordinates stays at that location relative to the global coordinate reference surface defined in the system explorer under Aperture. So in most cases it will not follow the movement of the surface and therefore you end up with the FEA data in the wrong location (not sure if that has a significant effect if the surfaces move just a bit).

I hope this helps,

Christophe

Userlevel 1

Hi Christophe,

Ah that makes sense.  I have my FEA coordinates in Global coordinates so that would make sense that Local coordinates would follow the surface after a transformation.

I am currently working on a Python script using the API to automate this.  So that does seem to be the only way to do it with Global coordinates.

 

I appreciate the response!

 

Cheers,

Alex

Reply