CalDB Public Release Notes Version: 3.1.0 Effective Date: 2005-06-23T10:00:00 CalDB 3.1.0 is to be released to the public on 23 June 2005 with the public release of the CIAO 3.2.2 patch upgrade. The new CalDB upgrade includes the ACIS N0004 responses suite, including new T_GAIN, GAIN, CTI-corrector, and phase 2 response (mkacisrmf) files. The essential things being added are the cti-correction, t_gain correction, and phase 2 responses for chips 4, 8, and 9, which had previously been excluded from the ACIS CTI/T_GAIN/P2_RESP suite. CalDB 3.1.0 also includes two items that were released into CXC standard processing with CalDB 3.0.4, dated June 14, 2005. These are the ACIS event and split threshold files, and the new-format HRC "GAPLOOKUP" tables. Details are below. CalDB 3.1.0 is required for the proper performance of CIAO 3.2.2 acis_process_events, and for optimal performance of the tool mkacisrmf. It is data-neutral for HRC tools. I. Summary of Changes: A. ACIS -120C CTI correction file version N0004 Filename: acisD2000-01-29ctiN0004.fits Location: $CALDB/data/chandra/acis/bcf/cti/ The -120C ACIS CTI-corrector file used by the tool "acis_process_events" in either standard processing or CIAO is being enhanced to include the three remaining FI chips' trap density maps. Specifically, trap maps for chips 4, 8, and 9 (i.e. S0, S4, and S5) have been added to the file, which complete the enabling of the parallel CTI correction for all eight FI chips. Pipes/tools affected: ACIS L1 acis_process_events CIAO acis_process_events with parameter "apply_cti"=yes Threads affected: "Apply the ACIS CTI Correction" http://cxc.harvard.edu/ciao/threads/acisapplycti/ B. ACIS -120C T_GAIN correction files version N0004 Filenames: acisD2000-01-29t_gainN0004.fits acisD2000-05-01t_gainN0004.fits acisD2000-08-01t_gainN0004.fits acisD2000-11-01t_gainN0004.fits acisD2001-02-01t_gainN0004.fits acisD2001-05-01t_gainN0004.fits acisD2001-08-01t_gainN0004.fits acisD2001-11-01t_gainN0004.fits acisD2002-02-01t_gainN0004.fits acisD2002-05-01t_gainN0004.fits acisD2002-08-01t_gainN0004.fits acisD2002-11-01t_gainN0004.fits acisD2003-02-01t_gainN0004.fits acisD2003-05-01t_gainN0004.fits acisD2003-08-01t_gainN0004.fits acisD2003-10-19t_gainN0004.fits acisD2003-11-18t_gainN0004.fits acisD2004-01-01t_gainN0004.fits acisD2004-02-01t_gainN0004.fits acisD2004-05-01t_gainN0004.fits acisD2004-08-01t_gainN0004.fits acisD2004-11-01t_gainN0002.fits Location: $CALDB/data/chandra/acis/bcf/t_gain/ As part of the new ACIS N0004 suite, all existing T_GAIN files require an upgrade from version N0003 to N0004 in order to include new time-dependent gain corrections for ACIS CHIPS 4, 8, and 9. Previous calibrations had only included chips 0-3 and 5-7. Additionally, a new epoch of T_GAIN calibration data is being added (EPOCH 19, for November 1, 2004). There is also a small modification to EPOCH 18 (August 1 2004) because a more complete set of data from that epoch is now available. This requires that the 2004-08-01 N0002 file, the flat (i.e. non-interpolating) file that has been the current file for Standard Data Processing (SDP) since November 2004, must be upstaged with two new files, an interpolating file acisD2004-08-01t_gainN0004.fits, and a new flat file for current processing acisD2004-11-01t_gainN0002.fits. We maintain the convention here of using "N0002" for the flat file from SDP to alert users that an upgrade in the T_GAIN calibration is due at some point in the future. Pipes/Tools affected: ACIS L1 acis_process_events CIAO acis_process_events with parameter apply_tgain = yes Threads affected: "Apply the Time-Dependent ACIS Gain Correction" http://cxc.harvard.edu/ciao/threads/acistimegain/ C. ACIS -120C GAIN_CTI file version N0004 Filename: acisD2000-01-29gain_ctiN0004.fits Location: $CALDB/data/chandra/acis/bcf/gain/ With the introduction of a new -120C phase 2 response file (Item D below) there is an incumbent change in the ACIS -120C gain_cti file. (Currently only the OBSIDs with -120C operating temperature are calibrated for CTI correction in the system.) In this case, the gains for chips 4, 8, and 9 absolutely must be converted to the new response system, because they are being added to the phase 2 response file. The previous (N0003) gain_cti file had phase 2 response gains for all chips EXCEPT 4, 8, and 9, while for these chips, the gains were left to be identical to those derived from the old FEF_PHA file response format. In addition, the phase-2 response file has actually been recompiled for ALL of the FI chips, because the SC_MATRIX for the FI chips has been recalculated to a different, expanded grid, namely the same one used previously and currently for the BI chips. Although the actual calibrations for the gains of chips 0-3 and 6 are not changed, there a slight numerical updates in the new dataset, because it is derived directly from the p2_resp input data. That is, the gain-tweaks table has been recalculated, so the derived ACIS gain table has had to be recalculated. Hence to complete the N0004 ACIS response suite, the new gain file below must be added to the CalDB, and must become the new default gain file in processing, replacing the N0003 file that is now the SDP default ACIS gain file. Pipes/Tools affected: ACIS L1 processing (acis_process_events) CIAO acis_process_events Threads variously affected: Create a New Level=2 Event File http://cxc.harvard.edu/ciao/threads/createL2/ "Apply the ACIS CTI correction" http://cxc.harvard.edu/ciao/threads/acisapplycti/ "Apply the ACIS time-dependent gain correction" http://cxc.harvard.edu/ciao/threads/acistimegain/ "Apply an ACIS gain map" http://cxc.harvard.edu/ciao/threads/acisgainmap/ The new gain file should be applied whenever the ACIS CTI-correction and the T_GAIN corrections are done with the above new files. D. ACIS P2_RESP (phase 2 response) input file for "mkacisrmf" version N0004 Filename: acisD2000-01-29p2_respN0004.fits Location: $CALDB/data/chandra/acis/cpf/p2_resp/ Several issues have been raised concerning the existing -120C (N0003) P2_RESP file in particular concerning certain FI chip tiles' having restrictive PHA/energy grid limits in the scattering matrix. Additionally, there are new calibrations available for chips 4, 8, and 9. Hence the ACIS calibration team has built, tested thoroughly, and released a new calibration file for "mkacisrmf." The FI chips' scattering matrices are now calculated on the same expanded grid as the BI chips are. The data of the BI chips is totally unchanged; however the input calibrations for FI chips I0-3, and S2 are also unchanged, although the data in the file are now slightly modified due to the new grid. The data for S0, S4, and S5 (4, 8, and 9) are entirely new to this release. With these additions, a user can now employ mkacisrmf to calculate the most current response matrices ANYWHERE on the ACIS array. Pipes/Tools affected: CIAO mkacisrmf (version 3.2.2 and later) This file is NOT used in standard processing through Level 2. Threads affected: "Creating ACIS RMFs with mkacisrmf" http://cxc.harvard.edu/ciao/threads/mkacisrmf/ II. Technical Details A. ACIS -120C CTI file version N0004 The science reference for the CTI correction algorithm and the trap density maps may be found on the Chandra Bibliography Page http://cxc.harvard.edu/cdo/biblio/chandra_bib.html, specifically under the reference "A charge transfer inefficiency correction model for the Chandra Advanced Imaging Spectrometer," C.E. Grant, M.W. Bautz, S.M. Kissel, B. LaMarr, 2004, SPIE, 5501, 177. The contents of the CTI ARD file acisD2000-01-29ctiN0004.fits include HDU 1: A table with the following columns Column Name Format Dims Units TLMIN TLMAX 1 CCD_ID I 2 CHIPX_LO I pixel 3 CHIPX_HI I pixel 4 CHIPY_LO I pixel 5 CHIPY_HI I pixel 6 NPOINTS I 7 PHA 47E adu 8 VOLUME_X 47E 9 VOLUME_Y 47E HDU 2: A trap-density map for CCD I0 HDU 3: " I1 HDU 4: " I2 HDU 5: " I3 HDU 6: " S0 HDU 7: " S2 HDU 8: " S4 HDU 9: " S5 The CTI adjustment to the data is performed by adding charge to each pixel of a 3x3 event island to improve the energy of the CCDs. The amount of charge added charge_added = trap_density * volume Here, the "trap_density" is the sum of the trap density in every pixel between the read out and the pixel where the event occurred. The values of trap_density are stored in the eight images contained in HDUs 2-9. The value of "volume" is obtained from the table in HDU 1. The volume is computed by finding the row in the table that has the right CCD_ID, CHIPX_LO, CHIPX_HI, CHIPY_LO, and CHIPY_HI values. In that row, the sets of PHA and VOLUME_Y points are used to compute the volume based on the PHA value for the CTI adjusted event. (The volume occupied by a charge packet is a function of the pulse height of the event.) A linear interpolation scheme is used between the tabulated pairs of PHA and VOLUME_Y points. Note that the VOLUME_X column is designed for use with a serial CTI adjustment. Since the effects of serial CTI have not been calibrated, this column is full of zeros at the moment. However, it remains in the file as a place holder in case the serial CTI is calibrated. The derivation of the trap maps for chips 8 and 9 follows the same procedure as was used for the previously included chips.For the S0 chip there is an exception in the TDM generation procedure. The "alpha" parameter in the qdp fitting engine was fixed at 0.53, rather than allow it to drift too high in the fitting process, which had previously resulted in an overcorrection of the PHA for the Ti and Mn ECS lines (i.e. the higher energies above 2 keV). By fixing that parameter, this irregularity in the trap map fitting algorithm could be corrected, and the downstream calibrations of T-GAIN, response, and detector gain could be produced with more reasonable results, and the testing thereof could be made successful. Calibration testing plots of these trap maps may be found at "Testing CTI Correction Calibration" visible at the following location: http://space.mit.edu/home/cgrant/cticorr/, and in particular for S0, S4, and S5 (the only new data added) at http://space.mit.edu/home/cgrant/cticorr/#S0 http://space.mit.edu/home/cgrant/cticorr/#S4 http://space.mit.edu/home/cgrant/cticorr/#S5 B. ACIS -120C T_GAIN version N0004 The basis of the ACIS time-dependent gain correction is given at http://cxc.harvard.edu/cal/Acis/Cal_prods/tgain/index.html The major change is inclusion of CTI-corrected S0,S4,S5 chips. There is also a subtle change affecting other FI chips (see below). The files can be found in /data/alexey3/chandra3/CAL/tgain they are called corrgain*.v3.fits. Calibration now covers all epochs through 2004/11/01-2005/01/31 NOTES: * This calibration includes measurements for CTI-corrected S0,S4,S5 chips. The procedure is almost equivalent to that for the "standard" FI chips (I+S2). An exception is the case of S4. This chip seems to have the strongest time evoltion. The energy dependence of the t-dependent CTI effect in S4 is better fit by E**0.75 than by E**0.5 which is used for all other chips. * Since there are detectable gain drifts in all FI chips (see http://space.mit.edu/~cgrant/gain/index.html), now one fits "A*E+B*E**0.5" to residuals in all FI chips. Previously this was done only for I2 which shows the strongest gain change. Inclusion of the linear term does improve residuals. The changes (relative to the previous measurements) are subtle at the ECS energies (1-2 PHA channels near Mn at most) but they can be more important when correction is extrapolated to high energies (i.e. for fitting background lines). * Tiny bug fix for S1 (now one uses the real gain table to find nominal energy for the given PHA; previously it was assumed that the gain was linear within the tile). The corresponding change in t-gain should be negligible. * No change in procedures for S3. The only news is the calibration for Epoch20. * Includes several more ECS data sets into analysis for Epoch19. The previous version was released on Oct 18, 2004 and thus missed several obsids from the end of that month. New file should be slightly more accurate. C. ACIS -120C cti-corrected gains version N0004 The phase-2 response file has actually been recompiled for ALL of the FI chips, because the SC_MATRIX for the FI chips has been recalculated to a different, expanded grid, namely the same one used previously and currently for the BI chips. Although the actual calibrations for the gains of chips 0-3 and 6 are not changed, there a slight numerical updates in the new dataset, because it is derived directly from the p2_resp input data. That is, the gain-tweaks table has been recalculated, so the derived ACIS gain table has had to be recalculated. A description of the gain table derivation in principle is given in the memorandum http://cxc.harvard.edu/caldb/calibration/CIP_files/gain.pdf. D. ACIS -120C Phase 2 response (P2_RESP) N0004 The FI chips' scattering matrices are now calculated on the same expanded grid as the BI chips are. The data of the BI chips is totally unchanged; however the input calibrations for FI chips I0-3, and S2 are also unchanged, although the data in the file are now slightly modified due to the new grid. The data for S0, S4, and S5 (4, 8, and 9) are entirely new to this release. With these additions, a user can now employ mkacisrmf to calculate the most current response matrices ANYWHERE on the ACIS array. Currently there is only a single -120C p2_resp file to be applied to all observations taken at -120C, with beginning effective date of UTC=2000-01-29T20:00:00. Recent evaluations of later-t-gain-epoch ECS OBSIDs have revealed an additional degradation effect in the pulse-height response. Adjustments to the ACIS response calibrations suite for later epochs are currently under development, but are not yet ready for release. The description of the calibration procedure is included in the memorandum http://cxc.harvard.edu/caldb/calibration/CIP_files/rmf.ps