Minutes of August 14, 2001, Telecon, 1100 CDT

NEXT TELECON: 28 AUGUST, 1100 CDT.

Present: Babler, Churchwell, Clemens, Dickey, Remy Indebetouw,
Kulbinicki, Marston, Mathis, Meade, Stolovy, Watson, Whitney

I. Observing strategy
	
A. Good news: the actual starting date of the observations will be
within two days of that assumed in our AORs. This is important because
of changing roll angles. We will need relatively little overlap beyond
the 15 pixels already planned. Near longitudes > 50 - 60 deg, some
extra overlap will be required for about a quarter of the exposures,
but the net effect will be a decrease in efficiency of only 1%.

B. There will be some geometrical distortion of the images because the
camera is not in the center of the focal plane.The amount and location
depends on the roll angle. It should be considered in the SSC pipeline
reduction, but we will have many reference objects for accounting for
it anyway.

II. Instrumental issues

A. Saturation and latency (see email from Marilyn Meade on 5 August,
forwarded from Benjamin): We estimate about 1 - 3 point sources are
saturated per frame at 8 um. The saturation of extended sources is still being
determined.  For the shortest wavelengths, we may use 2MASS as estimate.
Latency still seems to be a relatively minor problem.
Wrap-around for saturated images will be checked by Tony.

B. Scattering in focal plane: very little is known! Will any scattered
light have recognizable characteristics? We might have to wait for
in-orbit checkout. Our GLIMPSE data might be the best way to identify
scattered light, since we will have a  lot of bright sources. However,
we estimate that visually inspecting every frame will require 
about 5 person-hours per day, so we may come up with alternative plans, such
as inspecting the BQD, which is a mosaic of an AOR.

C. Muxbleed/banding (spurious stripes across the image. With
muxbleed, every fourth pixel in a row is bright.)  Muxbleed 
is well-characterized and is expected to be corrected for reasonably
well.  Banding is more problematic.  We will rely on SSC expertise 
to remove almost all of these. We do not expect them to be as serious 
as saturation or latency. Our two passes will help reduce the effects 
of muxbleed but probably not banding. However, any corrections we can
do are not easily automated--though we should think more about this. 
There will be a document produced by the pipeline review in late August 
(attended by Mike Wolff) that will address these issues.

III. Contingencies

A. Our response to a reduction in IRAC sensitivity? IRAC is so much
more sensitive than any previous detector that we should just go ahead
as planned. If we need to increase our exposures from 2 sec to 12, the
mission will have almost failed anyway.

B. Tony announced that the shutter will be open continuously. This
implies no dark information and much longer, more complicated
calibration procedures. The extra calibration times will not be
charged to teams. Calibrations will be done less often than previously
planned (at intervals > 6 hr, but lasting ~hours).

IV. Miscellaneous

A. Definition of one-sigma fluxes (how to determine them, especially
in confused regions.) Consensus: use either background noise, for
uncluttered regions, or the confusion limit. DAOPHOT determines a
background level for the whole frame and also one for the vicinity of
each point source identified. It seems to work well for finding
real sources (see below).

B. Source extraction: Babler and SSC have tested extraction programs 
on the same simulated data set. The agreement looks good. There is
need to improve DAOPHOT speed if possible. Perhaps simply using the
optimized compiler and removing some IRAF parts of the code will
help.  The other solution is to add more processors to the pipeline.
We will need to decide soon between DAOPHOT or using the
SSC code, though having both to compare can be advantageous.

C. A new patch for SPOT will be released today.

to Minutes index page