The ACIS flight software telemeters enough information to create spectra that have been corrected for both pixel by pixel bias variations and frame by frame bias variations.

Pixel by Pixel and Frame by Frame Bias Corrected Spectra

Figure 1. Corrected Center Pixel Pulse Heights

Peak at 404 ADU	Peak at 396 ADU
Peak at 400 ADU	Peak at 414 ADU

Figure 1 shows histograms of the corrected center pixel pulse heights for the data set. Each pulse height has been corrected for both the pixel by pixel variations in bias and dark current as well as the frame by frame variations in the average bias level for each quadrant.

This data set was taken in the Timed Exposure, faint telemetry mode. This means that:

• For each science run, the flight software telemeters an average bias value for each pixel of each chip being used. This is one of 1024 x 1024 pixel bias values in the average bias file.
• For each data frame, the flight software telemeters an average bias change for each quadrant of each chip being used. This is the delta overclock value.
• For each event, the flight software telemeters the position and pulse heights for the 3 x 3 pixel region of the event. This is a single row in the event list.

The corrected pulse height values in Figure 1 were created by subtracting the pixel bias value and the delta overclock value from each of the pulse heights in the event list.

Pixel by Pixel Bias Correction

At the beginning of the science run several frames were captured, x-ray events were discarded, and the frames were averaged, then the bias information was telemetered by the flight software. The resulting bias map has a single value used to correct each pixel.

Figure 2. Histograms of Pixel by Pixel Bias Levels

Average Value: 930 ADU	Average Value: 865 ADU
Average Value: 551 ADU	Average Value: 833 ADU

Figure 2 shows the histograms of the pixel bias levels for each of the four quadrants.

Figure 3 shows the histogram of the center pixel pulse heights without the pixel bias correction applied. While the general shape of each of the curves is the same, the peak for each quadrant is shifted positive by the average of the bias level.

Figure 3. Center Pixel Pulse Heights without Pixel Bias Correction

Peak at 1335 ADU	Peak at 1080 ADU
Peak at 950 ADU	Peak at 1248 ADU

Frame by Frame Bias Correction

The frame bias values are determined by taking average overclock values. At the end of each row in every frame the serial output shift registers are clocked additional times. The flight software computes an average overclock for each quadrant. The change of the overclock values from the beginning of the Science Run, delta overclock, is telemetered by the flight software for each data frame.

Figure 4. Frame Bias Values vs. Time for Each Quadrant

Average: -19 ADU	Average: -18 ADU
Average: -33 ADU	Average: -23 ADU

Figure 4 shows the frame bias values telemetered by the flight software for each data frame.

Figure 5. Center Pixel Pulse Heights without Frame Bias Correction

Peak at 384 ADU	Peak at 378 ADU
Peak at 367 ADU	Peak at 395 ADU

Figure 5 shows the histogram of the center pixel pulse heights without the frame bias (delta overclock) correction applied. While the general shape of each of the curves is the same, the peak for each quadrant is shifted by the average of the frame bias.

The Data Set

The data set chosen for this comparison is Science Run 72. The chip is s4 (aka. w457c4). The anode is silicon. The data set was captured on 09 May 1997 at 17:48 GMT.

The parameter block used was xfaint_win_s.te. xfaint_win_s.te is full frame, faint telemetry, trickle bias, with a window.

The window block used was xsik56b50.2d. For the s4 chip, this has two windows. The first window begins at row 9 (counting from 0) and is 28 rows high. The second window begins at row 972 and is 28 rows high. All 1024 columns were telemetered.

The first event for the science run was on 09 May 1997 at 18:03 GMT. The last event for the science run was on 09 May 1997 at 23:49 GMT.

An ERV event list and fits format average bias frame for this science run were created from the ACIS telemetry using PSCI. For a description of PSCI and the ERV format see here.

The ERV format event list was converted to ACIS RV format using an IDL program which combines the fits format average bias file and the ERV format event list. For a description of the ACIS RV format see here.