PIPE2D-509 involved determining how much of an FRD change was required to be detectable at given SNR. Further FRD work has been focused on how FRD effects the PSF and in-focus image, particularly focused on quantifying the effect on PSF rather than quantifying FRD variation. However, given that the fine defocus motor permits 100 um defocus, it would be valuable to revisit the defocus question with a emphasis on determining PSF variations due to FRD variations. The goals of these tickets are related though not the same.

The maximum % variation in a full boxcar extraction was tested across 120 positions at three levels of defocus: In-focus (0 um defocus), intermediate defocus (32 um defocus), and 'maximum' fine defocus (100 um defocus). The results are shown in DefocusPositions.png, where in focus is shown in downward-pointing triangles, intermediate focus as circles, and maximum defocus as upward pointing triangles. The maximum percent variation does generally increase with higher defocus as to be expected, though specific positions across the detector can have different trends. The analysis may be impeded by the use of the maximum % variation metric rather than quantifying else wise, such as 1D boxcar extraction or SNR variations (see PIPE2D-601, PIPE2D-602). Currently the magnitude of the effect is being examined more closely, since determining how much defocus would help PSF characterization would help prioritize calibrations for sky subtraction.

Surprisingly little change in the 2D residuals was detected with defocuses up to the fine defocus motor limit, so future analysis has focused on in-focus data.