Note that for observations which are members of a multi-observation stack, this step is the same as for the Pre-Calibrate pipeline, but run with the updated aspect solutions.
A new bad pixel file is created to ensure that all observation-specific hot pixels and afterglow events are included.
The OBI is then reprocessed to apply the latest calibration—gain, ACIS time-dependent gain, ACIS charge transfer inefficiency (CTI), HRC degap, etc. A unique CALDB is used for catalog processing; it is frozen for the duration of the processing run so that the same calibration is applied to all the datasets in that version of the catalog. The same processing tools are used as in the Chandra Standard Data Processing pipeline: acis_process_events for ACIS data and hrc_process_events for HRC data.
Standard filtering is applied to the data to apply the good time intervals (GTIs) and remove bad events with bad grades or status flags. ACIS datasets are destreaked. Average dead time corrections are calculated for HRC datasets.
Identify and remove background flares
Under certain conditions that are often met in CSC2, removing background flares improves the signal-to-noise ratio for source detection. In order to remove background flares in each chip, candidate source pixels are identified and then excluded from the analysis. Source pixels are identified by creating a histogram of the data for each chip. The mean and standard deviation of the histogram values are determined. All pixels which have a value greater than (mean+3σ) are considered source pixels.
A light curve is then created with the Gregory-Loredo algorithm using the full area of the remaining background pixels in the chip. The light curve minimum rate is calculated, and a threshold value is set at 15 times this minimum. For each chip, light curve times for which the ratio of the light curve rate to its minimum is greater than the threshold are excluded from further catalog analysis. These are the times when the lightcurve flare takes place.
The GTIs are revised to exclude those periods of background flares.
Create background maps
The mkvtbkg algorithm uses a Voronoi Tesselation method to create a full field background map in each energy band. The maps are created at 1 pixel resolution for ACIS and 2 pixels for HRC. The algorithm also detects candidate source regions including highly-extended ones.
Create full field data products
Standard Chandra/CIAO tools are used to create image products and other full field data products, listed below.
Output data products
The Calibrate pipeline produces these data products:
- full-field event file (evt3.fits): similar to the evt2.fits file from the Chandra Data Archive, but has different GTIs due to the background flare filtering.
- aspect solution file (asol3.fits).
- bad pixel file (bpix3.fits): for ACIS, this is the same as the bpix1.fits file from the Chandra Data Archive. The HRC bpix3.fits may have a newer degap correction applied than the bpix1.fits version.
- field of view (fov3.fits): this file may be slightly different than the fov1.fits from the Chandra Data Archive, due to the different GTIs from background flare filtering.
- aspect histogram (ahst3.fits): generated by the asphist tool, as users would in CIAO data analysis.
- full-field exposure maps (exp3.fits): for each source detection energy band, at 1 pixel resolution for ACIS and 2 pixels for HRC. The exposure maps are created as shown in the CIAO thread "Multiple Chip ACIS Exposure Map."
- full-field background maps (bkgimg3.fits): for each source detection energy band
- full-field images (background subtracted, exposure corrected) (img3.fits): for each source detection energy band
- full-field image mask, (pixmask3.fits). Mask pixel values are 1 if the pixel has non-zero valid exposure; otherwise the pixel value is 0 or NaN.
- Background contour file, (poly3.fits). This file contains polygon regions for extended sources detected by the background map creation algorithm.