Cooling the ACA CCD below -10 degC: Initial Data Analysis

In preparation for cooling the ACA CCD below -10 degC, I have analysed the available flight data from the mission, to examine the performance of the TEC (thermo-electric cooler) which cools the ACA CCD, and to estimate further cooling capability. Cooling the ACA CCD reduces the dark current in the CCD, which allows more accurate centroiding of star and fid light images. The dark current decreases by about a factor of 2 for each reduction of 5 degC in the CCD temperature, according to data from the manufacturers (Scientific Imaging Technologies Inc.).

The analysis uses header-3 telemetry MSIDS from ACA image slots 6 and 7, available when those ACA image slots are using 8x8 readouts. The data used are the CCD temperature, the ACA temperature (I take the overall ACA temperature used in the field angle polynominal calibration as the ACA temperature), and the DAC set point for the TEC controller.

ACA HDR3 telemetry:

Image Slot 6: HD3TLM76,HD3TLM77 =  CCD temp in units of 0.01 degC.

Image Slot 7: HD3TLM66,HD3TLM67,HD3TLM72,HD3TLM73 = ACA temp in units of 1/256 degC.
              HD3TLM76,HD3TLM77 = TEC DAC input count.

Since the start of the Chandra mission in 1999, the ACA temperature has increased by about 1 degC, so that the temperature differential across the TEC has increased by about 1 degC as the CCD temperature has remained approximately constant at -10 degC. However, only for about the past year has the ACA been operated to use 8x8 pixel readouts in image slots 6 and 7, during non-science obsids.

All analysis code and data files are in /proj/sot/ska/reports/TEC/.

Processing steps:

racdot_search.pl
output: sdot67.8x8
The perl script racdot_search.pl searches through DOT files to find obsids where both image slot 6 and image slot 7 have 8x8 readouts, for getting HDR3 telemetry.
drive_arc5gl.pl
input: sdot.8x8
output: bar2/fd.*.* (image slot 7 HDR3 data)
input: sdot67.8x8
output: bar3/fd.*.* (image slot 6 HDR3 data)
The perl script drive_arc5gl.pl retrieves ACA level 0 telemetry files for image slots 6 and 7 from the archive using arc5gl, for obsids found in step 1 above, and then uses FTOOLS fdump to extract time and HDR3 telemetry msids from these telemetry files into ASCII files named fd.$OBSID.$N
tec_merge.pl
input: fd.*.*
output: foo
The perl script tec_merge.pl converts the HDR3 telemetry MSIDs into real temperature and TEC DAC values.
plots2.pro
input: foo
The IDL procedure plots2.pro plots the ACA temperature and TEC DAC data against time, and plots the TEC temperature differential against the TEC DAC setpoint, to establish the relationship between the two parameters.
Result: about 12 steps of TEC DAC control produces 1 degC change in differential temp.

The results of the analysis are shown above. The plot in the lower right shows the relationship between TEC DAC control point and temperature differential across the TEC. The red line on the plot shows a curve with slope = 12 DAC_counts/degC, and offset = -27.5 DAC_counts at 0 degC.

The maximum TEC DAC count is 511. If the above linear relationship between TEC DAC count and TEC temperature differential is extrapolated to this limit, the TEC can produce a maximum temperature differential of about 45 degC. At the current operating temperature of the ACA, this would produce a CCD temperature of about -30 degC.

The above analysis indicates that the ACA CCD can be cooled to at least -15C, resulting in a decrease in dark current by about a factor of 2. The current plan is to gradually decrease the CCD temperature by 0.5 degC increments, spending a week at each intermediate temperature to allow analysis of ACA and TEC performance.

Rob Cameron

Last modified: 2003-Feb-19