Data Processing

The ACIS FEP processing mode was timed exposure faint 3x3 with an event threshold of 38 ADU for the front illuminated devices and 20 ADU for the back illuminated devices. The split threshold was 13 ADU for all measurements.

The ACIS BEP processing mode was timed exposure Faint mode with a window filter. The size of the windows varied depending on the flux rate, from as small as 20 rows to as many as the full 1024. In all cases, the windowing was split into two strips, one toward the top of the CCD and one toward the bottom of the CCD. Because of the differences in the quantum efficiency of the front illuminated devices from the back illuminated devices, different window sizes were used for the front illuminated and back illuminated devices. The window sizes used in each data set are listed in Table  4.4.

The flight software telemetry packets were converted to FITS format average bias files and ERV format event lists using PSCI (for a description of the formats and the PSCI program see http://acis.mit.edu/ttools/psci.html).

The average bias and the ERV format event lists were combined and converted to ARV format event lists using an IDL script. The IDL script subtracted the delta overclock value and the average bias values for each pixel of each event.

Each ARV format event list was then summarized using an IDL script. This IDL script fit a gaussian to the main peak of the spectrum. Only events in ACIS grade bit map 0x1d05 0x0047 0x0004 0x0047 0x1133 0x0003 0x1100 0x0000 0x1d04 0x0000 0x0004 0x0000 0x0000 0x0003 0x0000 0x0000 (ASCA grades 02346) were considered. Using the fit parameters, the position of the peak, full width half max of the peak (defined to be 2.35 times the fitted sigma), and total number of counts under the peak are determined. The number of counts under the peak is taken to be $\sqrt{2 \pi} \sigma N_{max}$, where N_max is the number of counts in the peak channel. The IDL script computes the flux in counts per centimeter squared per second assuming a frame time of 3.34 seconds and an input window size. All of these values are written to a text format file.

Finally, the text format summary files are read by another IDL script which uses the peak location and peak width values at each energy to tabulate and plot the linear energy scale and spectral resolution values. This script also applies a pileup correction and the filter efficiency to the counts under the peak. See Section  4.4 for a description of pileup.