Solved

Introducing a known wavefront error in sequential mode

  • 2 July 2020
  • 2 replies
  • 573 views

I’m currently modelling an imaging system in sequential mode and want to check the effect of a beamsplitter quoted with a lambda/4 surface flatness and lambda wavefront error. As an approximation, is there a simple way to introduce the wavefront error at a dummy surface used to represent the beamsplitter?


If I wanted to more accurately account for the surface flatness in reflection, would I need to use non-sequential mode? And how would I incorporate this property into the surface in that case?


Thanks!

icon

Best answer by Mark.Nicholson 2 July 2020, 18:57

View original

2 replies

Userlevel 6
Badge +2

Hi Tiffany


I think that it is quite equivalent to tolerancing. The point to figure out is what does the lambda/4 mean? Is it a peak-to-valley error, is it a RMS? What would model best this error? This comes down to how it is measured and how it is manufactured.


To tolerance the irregularity in OpticStudio, there are basically two choices: model the irregularity as a sum of spherical and astigmatic terms or model the irregularity with a Zernike surface.



In both cases, you can apply that error on the beamsplitter surface and then run the tolerancing. I have attached an example.


If you add a SAVE operand in the tolerancing, it will save the system at - tol and + tol.


The files are saved in the same folder as your file and they are called TSAV_MAX_0001.ZMX, ...


 


For more information on the Zernike terms, have a look at those articles:


How to use TEZI to tolerance for manufacturing-related surface sag error


Constructing mid-spatial frequency tooling errors for evaluation and tolerancing


and this forum thread:


Tolerancing- understanding of when using TEZI, the max tolerance value is RMS error of surface


 


Sandrine


 

Userlevel 7
Badge +3

Hi Tiffany,


You can also use a Zernike surface to pre-aberrate the beam by a known aberration. Se  https://my.zemax.com/en-US/Knowledge-Base/kb-article/?ka=KA-01392 for an example.


The problem is that while people will tell you that the compoent is 'flat to lambda/4' or whatever, you need to know how the aberration is distributed. If it's a nice low-order focus error then it can be easily compensated by a focal shift, and the higher-order and less-rotationally-symmetric the lambda/4 is, the less you're able to correct it.


HTH,


- Mark

Reply