Skip to main content

Hello,

I have been tolerancing some achromatic doublets using vendor presets in OpticStudio (specifically the Edmund Optics presets) and had a question regarding the expected changes in effective focal length.

Most of the lenses are manufactured to a tolerance of +/- 1-2% on the focal length, but the observed changes that I am seeing are smaller even when the tolerances are relaxed. For example, a 50 mm diameter 170 mm focal length doublet only has a focal length variation by 0.5 mm for 1000 monte carlo runs (wavelength is 632.8 nm)

Is there something that Zemax does not consider when tolerancing that is leading to this discrepancy? I also imagine that I could be missing something obvious!

Thank you,

Musab

Hi Mudsab 

 In my opinion, the tolerance items manufactures use to calculate the f and the ones you use may be different. Since from the site, only dimension tolerance can be got. Other information ,say n tolerance, is not released to public.

 

best regards 

Yang

 


@Musab.Blain

I think, in general, a focal length variation spec of +/- 1% is very conservative.  A lens supplier wants high yield, so they use a loose focal length tolerance that is easy to hit.

You can try the following:

  • Apply the “commercial” default tolerances for Edmund Optics (be sure all lens surfaces are included)
  • Use uniform statistics (not Gaussian) with a STAT operand (Type = 1) in the TDE
  • Add wavelength variation over the entire visible region (0.425 to 0.675 um) since this is how Edmund specifies lens performance.  Doing so will impact the range of effective focal length variation.

For wavelength variation, use a WAVE operand in the multi-configuration editor, along with a TMCO operand in the TDE as outlined in the answer to this post: 

 

When I do this for Edmund 32886 (D = 50 mm, f = 150 mm), I see a focal length variation of  -0.44 to + 0.60 mm (or approximately +/- 0.33%).

 

So, I don’t think you are doing anything wrong.  It’s just that the 1-2% focal length spec is looser than would be expected based on lens fabrication tolerances plus glass dispersion.

Regards,

Jeff


Wavelength variation?  Not sure what you mean…. 


If you buy an achromat designed for use in the visible, then you might be using it at a wavelength anywhere from 425 nm to 675 nm.  The measured focal length will change with wavelength, and my assumption is that this focal length variation is part of the tolerance spec.  Otherwise, it’s hard to see how the focal length can change by +/- 1% with only fabrication tolerances (when measured solely at the design wavelength).  However, I can’t find an exact definition of Edmund’s focal length tolerance.

Jeff 


Thank you for the replies!

That’s a good point that you raise Jeff. I didn’t consider that the tolerancing is really only considering focal length variations as a direct consequence of manufacturing and not the variation you would see across the visible wavelength band.

I’ve checked the best and worst case monte carlo files and used some EFFL operands to check the focal length at the F,d,c spectral lines. Taking that into consideration the +/- 1% specification makes more sense now!

Musab


Reply