Problems with the STIS CCD ERR Array

During the course of getting the Eta Carinae TP database together we have discovered a problem in the way that CALSTIS processes the ERR array (which contains the estimated noise value for each pixel). One aspect of this is the irksome coherent noise illustrated in the image below.

The ERR array displayed is a typical spectrum randomly selected from our observations of Eta Carinae with the HST STIS. The data has been fully processed by the version of CALSTIS implemented in the STScI MAST archive pipeline. (Note that this is not our version of CALSTIS normally used to reduce the Treasury Project data. Our version acts in a similar way since it is the genetic brother to the STScI pipeline.) The gray scale has been set using a linear scale with limits set using the DS9 zscale algorithm.

We wish to emphasize that this problem applies to _ALL_ STIS CCD that is reduced using the STScI automatic reduction pipeline. Our investigation of this problem is ongoing and details can be found in the Treasury Project Techmemos:

• Our Adopted Scheme for Subpixel Modeling - Kris Davidson, University of Minnesota

• The Meaning of "ERR" HDU's or Noise Files - Kris Davidson, University of Minnesota

Last updated 13 October 2004