This section lists issues that are beyond the scope of the current LSF library files and tools.
Regions: The current implementation of the library assumes that
that the region is wide enough (greater than 
1
)
for the details in the core of the psf not to matter (cf. figure
17). For narrow extraction regions we will need
multiple 1-D grating libraries or a full 2-dimensional grating PSF
library.
Pileup: To compare the observed PSF with the theoretical PSF we assume that the observed data is corrected for all instrumental effects (eg. gain variations effective area, ...). It is not clear at the moment how to correct the observed data for pileup or to modify the LSF for the effects of pileup.
Frame Transfer Effects: For bright sources a significant number of photons can strike the detector during the time it takes to read the CCD creating a streak in the data. The number of counts in this streak is just the ratio of the frame transfer time to the frame time. For standard TE mode data this is about 1.2% of events in the feature. The PSF tools do not account for this.
CCD Gaps: The PSF's and derived LSF's are still generated using MARX and the MARX detectors which have the detetor gaps and edges. The library avoids these gaps to generate the PSF's but this also means that we cannot generate accurate PSF near the detector edges. This will be solved when the CXC datasystem project ray tool is available and works with the gratings.