Changes between Version 8 and Version 9 of IPP_Progress_Report_20110725

Timestamp:

Aug 2, 2011, 2:21:04 PM (15 years ago)

Author:

eugene

Comment:

--

IPP_Progress_Report_20110725

@@ -8 +8 @@
 We have been fighting with hardware problems this past week.  One machine, ipp030, had a motherboard failure.  Cindy replaced it with the motherboard from one of the compute nodes (ippc10).  We had many repeated crashes of ippdb03 (backup database server), and have arranged to move its services to another compute node (ippc11).  We had a problem with the raid on ippdb00 (nebulous server), compounded by inconsistencies in the reported drive IDs.  We had multiple crashes of ipp026, and have removed it from the nebulous list for now.  It is unclear to me if this is just a string of bad luck, or if there is an external reason for multiple hardware problems.  Cindy et al checked on the load on the PDUs, but there does not seem to be any thing unusual.  A/C seems to be fine.  
 
-I have been making progress on improving the non-linear fitting for extended sources.  I improved the model guess by using a large set of fake data to generate a better set of relationships between observed moments and model parameters.  I also improved the speed of the non-linear fitting by (a) using fewer pixels for the fit than for the model subtraction (these had previously been coupled) and (b) allowing for a 1D PSF convolution (saves about a factor of 2).
+I have been making progress on improving the non-linear fitting for extended sources.  I improved the model guess by using a large set of fake data to generate a better set of relationships between observed moments and model parameters.  I also improved the speed of the non-linear fitting by (a) using fewer pixels for the fit than for the model subtraction (these had previously been coupled) and (b) allowing for a 1D PSF convolution (saves about a factor of 2).  
+
+Chris & I investigated some odd detections reported by MOPS.  It seems that these are coming from reflections off vertical structures in the camera close to the focal plane.  One set are due to the calcite cells and the second set are (we think) due to the frame around the edge of the focal plane.  The resulting artifacts are extremely PSF-like and land within 60 pixels of the edge of the camera (in the second case) or around the edge of the calcite cell.  This is one of those annoying cases where we can add a suspect mask bit to warn about the danger, but we probably don't want to exclude all possible detections in this region.  Fortunately, these are fairly rare, and only were noticed by MOPS in observations near the Galactic plane.  Also fortunately, they should not appear static with respect to the stars because small shifts between the TTI images are doubled for the reflections.
 
 === Serge Chastel ===