Hi Soroush,

Thanks for your question here on the forums!

For the Huygens PSF calculation, if “Use Centroid” is selected, Zemax calculates the geometric centroid of the beam.  These values are then printed out as the center coordinates on both the graph and in the text listing.  If “Use Centroid” is not selected, then the center coordinates correspond to the chief ray of the field point for which the PSF has been calculated.  The graph is always centered on the center coordinates.

Regardless of whether “Use Centroid” is selected or not, Zemax will calculate the shift between either the geometric centroid of the beam (“Use Centroid” selected) or the chief ray (“Use Centroid” not selected) and the centroid of the PSF itself (determined from the second moment analysis of the PSF).  This shift is reported as the centroid offset in the text listing of the PSF.  The sum of the center coordinates and the centroid offset is reported as the centroid coordinates in the text listing of the PSF. 

The geometric centroid is only being used to center the plot, the actual centroid of the PSF is returned in the text listing.

I hope this helps, but if you have any further questions, please let us know and we will be happy to help!

Best,

Csilla

Thanks Csilla.

Best,

Soroush

Hi @Csilla Timar-Fulep 
Can you help clarify how the centroid offset is calculated? 
“the centroid of the PSF itself (determined from the second moment analysis of the PSF)”

What is “second moment analysis”

how can I manually calculate the same centroid offset? I tried using weighted average, but I get >200 nm error only in the Y axis.  
(however, X axis is within 1 nm from the reported Centroid_offset so it good enough). 

I also double checked that the values saved in my saved PSF are identical to those reported in the txt output.
 

say I have the matrix in matlab of the PSF, P. I tried:

        X = (X -65)*ImageDelta*1e-3; % to GET XY in mill, since image delta is in mu

        Y = (Y -65)*ImageDelta*1e-3;

        W = P/sum(P(:));

        M10 = sum(X(:).*W(:));

        M01 = sum(Y(:).*W(:));

        Centroid_offset_calc = ;M10,M01];

        % this is the woring answer by <1nm inX, but 200 nm in Y…

THANKS, Noam

Hello,

I’m running into the identical problem as Noam, and would be glad reading any insights into that issue since in my system the shift reaches over 1 µm which becomes significant wrt the specifications I’m trying to match.

Thanks in advance

Benoit