- PS CDR Dry Run --
A dry run for PS CDR
presentations will replace the regular working group meeting next
Tuesday, March 19, but it will begin at 2:00 PM instead of 3:00 PM. J.
Fowler requested that all subsystem cognizant engineers prepare the
following information about the status of their development work as of
next Tuesday: percentage completion of subsystem design,
documentation, coding, and testing; it is preferable to break down all four
categories by functional task, i.e., if a subsystem does functional
tasks A and B, and coding of A is basically finished while B is hardly
begun, the status should be reported as (e.g.) A 95% coded and B 5%
coded, rather than the subsystem being 50% coded. For example,
BANDMERGE initialization and input processing is 100% coded, the
cross band connection processing is 98% coded, the inconsistent linkage
correction processing is 0% coded, and the refinement and output
processing is 0% coded. Status reports for the BANDMERGE design,
documentation, and testing will also be broken down into these
processing tasks. The breakdown should be fairly high level and
meaningful to science team members. In some cases, such a breakdown may
not be applicable; whether it applies to a given subsystem is left to
that subsystem's cognizant engineer.
- Flattening Problem --
A flattening problem has been found in a scan being used for galaxy
development by T. Jarrett; this problem has been found before, and a
search for its solution hasn't yet had a high enough priority to be
allocated resources. It appears that persistence due to a bright source
contaminates the trimmed averaging in DFLAT, causing the persistence
signature to get into the sky offset correction frame itself, from
where it spreads the contamination to the frames in the domain of that
correction frame. Assuming this is the problem, a solution was
suggested by G. Kopan: use the nonlinearity detection code in DFLAT
to identify the bad correction frame and mask the corresponding pixels
in that frame, forcing DFLAT to use a correction based on averaging
over the entire scan. J. Fowler will check the DFLAT code to see if
this is a capability that can be turned on or requires code
modification.
- Seven-Apparition Photometry --
B. Light presented some intermediate results in a study of whether
aperture photometry is significantly improved for seven apparition
point sources when all seven apparitions are used, as opposed to using
only the first six. The results graphed showed no significant
improvement when all seven apparitions were used, but the
seven apparition statistics were diluted by being averaged with cases
in which seven apparitions were in principle possible but one was
excluded because of dead pixels or proximity to the edge of the frame.
Whereas one would expect the S/N of seven apparition results to be
better than six apparition results by the square root of 7/6, it is not
clear that the nonuniformity introduced is worth the slight gain in a
few cases.