From Table 1, it may be seen that most of the pixels masked off are not completely dead, but are excessively noisey or have erratic or low responsivities. This implies that a sequence (> 100) of data frames illuminated with photons at the approximate level of a typical dark sky background will be required to reliably evaluate the number and pattern of usable pixels in a given chip.
Provided with such data, pixel masks can be generated, and used to mask off simulated data frames containing a moderate density of stars. The completeness, reliability, photometry, and astrometry of the point sources extracted from this data can be evaluated and compared to similar runs with no pixels masked off. This would provide a objective measure of the effect of the quality of a given chip on the resulting catalog quality. A histogram of the the level of coverage at the expected frame overlap during a survey scan will also be generated. This histogram will show likely areas of zero or low coverage.
This chip evaluation pipeline will be assembled and tested prior in advance, so that once data becomes available for a given chip, an evaluation of that chip may be provided within a few days.
The masks corresponding to Table 1 are shown in Figure 1.
Figure 1: Bad Pixel Masks for Protocamera Chips vs Time.