A series of scans were simulated, at moderate stellar density (b=10), using PSFs with different FWHM's. The four FWHMs tested were 1.5, 2.0, 2.5 and 3.0 arcsec. Every frame in a single scan had the same FWHM, was properly dithered (0.15,42.23), with no subpixel response nor dead pixels. A PSF was determined for each scan, so the data was reduced with the appropriate PSF; future testing will be performed with intentionally mismatched PSFs.
The results of this test can been seen in Figures 8 -- 11. Photometric and positional accuracy suffers with very narrow (less than a pixel width) PSFs. The planned improvement of the offset determination (using three bands of information) will help reduce the errors in such cases. The photometry and positions are best when the PSF has a width betwen 2.0 and 2.5 arcsec.