SDAST Meeting 24 - ISDC, 28. + 29. 11. 2002 =========================================== Attendants: Carol Anne Oxborrow DSRI [CAO] Niels Lund DSRI [NL] Niels Jørgen Westergaard DSRI [NJW] Jérôme Chenevez DSRI [JC] Sami Maisala Helsinki [SM] Stefan Larsson Stockholm [SL] Silvia Martínez Núñez Valencia [SMN] 13:45 Welcome & Agenda ---------------------- Changes to agenda: - NL's hardware presentation ~14:15 - Include planning of next steps at end ISDC News [PK] -------------- Operations running quite smoothly. Main problem TM interruptions if Goldstone is used. All isnstruments produce results. ISDC plans component list freeze mid December. Science Results [NJW] --------------------- Imaging basically works, sources can be found if offset is taken into account. Two effects: - Origin of coordinate system wrong by 4mm in backplane direction - Misalignment between startracker and JEM-X (~8') LMC X-1, X-2 and X-4 found in LMC observation. Newly delivered JEM-X IMOD file (007) contains correct zero point in detector (within 1 mm uncertainty). Rest of misalignment will be put in misalignment matrix of ISDC => ISSW (j_ima_basic_recon, j_src_spectra, j_src_lc) must include this matrix. CBJ's finds peak widths of ~1.3mm, degrading at low energies. ISWT Results [NL] ----------------- Good news: all instruments on and working. SPI: Great energy resolution ISGRI: Working well, able to work down to 14 keV OMC: Good images & photometry Bad news: Background countrates much higher than anticipated => crunch in telemetry capacity. SPI needs 50% increase in TM to get essential data down. Have 39 packets, need 59 packets for solution with single events (lower energies) packed into spectra. As JEM-X probably has limited lifetime, after PV phase only one detector will be operated (except for Crab observations). JEM-X TM share will be reduced to 9 packets (from 18). Baseline for PV phase: one detector in REST (3 packets) and one in FULL (6 packets). Necessary discussion within JEM-X Consortium: how to run with only one instrument. Proposition: do GPS with both JEM-X. Uploaded new rejection criteria, threshold can be lowered. Analysis Pitfalls ----------------- Missing or mismatched auxiliary data for early revolutions => workaround through patching of SWG, skipping levels, ... Invalid gain corrections. Calibration source leakage. Noisy image borders if vignetting corrections are applied. Timing analysis [SL] -------------------- No pulsations found in Cen X-3 (upper limit 10%) with various data. Should use j_src_lc now on both JEM-X detectors. Comparison of detector lightcurves with source extraction lightcurves demonstrates that vignetting corrections essentially work. Power spectrum still rough but shows expected behaviour for Cyg X-1 in this state. Gain calibration & deadtime [CAO] --------------------------------- Lower peaks (e.g. Ni peak) not usable in single spectra. Programs will be rewritten to use major peaks and knowledge about channel zero. JEM-X 2 has two segments with very distorted calibration spectra, due to dead anodes under calibration sources. Gain calibration shows rapid changes on switch on and then settling gain but with glitches. Glitches can be taken out in gain correction by modeling overall evolution. Problem for early science windows in revolution. Baseline: - Use only good anodes for JEM-X 2. - Compare anode results; in case of large differences use those close to default values Source finding problems [JC] ---------------------------- Known sources not found - known misalignment problems. Some sources identified spuriously, ***but at completely wrong positions*** => PK to check q_identify_srcs. ################################ ACTION on PK, due Friday Dec 6 Check q_identify_srcs ################################ Instrument Characteristics [NJW] -------------------------------- NJW summarized the required changes (see viewgraphs). j_cor_gain will handle PHA->PI translation correctly once new IMOD group is available (and gain fitting is stabilized). Repsonse information will change as well. ################################### ACTION on NJW+SL, due Friday, Dec 6 Resolve ARF above 35 keV ################################### Update of response expected next week. Analysis Scripts [SMN] ---------------------- Listed main changes since last meeting (see viewgraph 1). Continued with recent - not yet delivered - changes (see vg 2). Finished with planned changes. Spectra ------- Different types of spectra: source (j_src_spectra) and binned over detector. Background handling completely different: automatically subtracted for source spectra, binned from model for binned spectra. Both results can be used with XSPEC. Open tasks: - include radiusLimit in j_bin_evts_spectra - pulse phase resolved source spectra - generic tools to combine spectra (spe_pick, FTOOLS) Lightcurves ----------- Fixed so far: - Parameters rowSelect & radiusLimit included - 180 deg rotation leading to missing fluxes in JEM-X 2 - Core dump, due to two different ways of counting bins - Removed ARF parameters Future: - Include correction of open fraction - Phase folding - Allow for REST data - Use background model - Deal with dead anodes (cut out, or use?) Explained verification of S/W performance through use of diagnosticMode parameter Scientific Validation --------------------- A lot of validation has been effectively done by all. Next steps - Misalignment matrix - Spectral fitting / comparison with RXTE/ISGRI/SPI - Recheck time stamps - Reanalyze Cen X-3, Cyg X-1 ======= Friday ============================================== Discussion around usage of detector maps in IC. These maps will be in raw cooordinates. j_cor_ will read this information and include it in the FLAG column. The current 8 bits are probably not enough. ################################################ ACTION on PK, due Monday, Dec 2 SCREW to have 16 bit event flags (STATUS column) ################################################ Hardware News [NL] ------------------ Summary of presentation made for ISWT. Strong Xe fluorescence K-line can be used for gain calibration. Data has be used by CBJ to refine selection criteria. Since JEM-X sensitivity above 35 keV is very low a cut at 50 keV has been introduced. New on-board rebinning uses full resolution at lower channels. The uneven widths of ADC channels will be visible. CAO pointed out that this information could be put into the JMXi-FULB/RESB/...-MOD tables which are _floats_ and not integers. Reported on big solar flare event. Will be good test case for deadtime corrections. 15% double triggers in HARDWARE_TRIGGER. Anode monitoring in OSM. Idea: integrate over several hours collapse in Y direction, compare with semi-fixed reference. ################################ ACTION on NL, due Friday, Dec 13 Define monitoring strategy ################################ Discussed anode deaths further. Egde effect of background mostly visible at Xe energies. Imaging ------- Summarized status (see vg). Some discussion about image formats. To-do list: - correct source position [2w] - systematics - image formats - IROS - Include NL's code in ISSW - Include SL's code as check on source finding Open fraction calculation does _not_ take radius selection into account. Background discussion --------------------- Preliminary results from Rev 0005 presented. Include in IC map of pixel widths. Use (linearized) PHA channels instead of unreliable PI values. Discussion of bkg model maps: raw or corrected positions? Conclusion: bkg model maps are in idealized, corrected observations. Dead anodes should be treated separately from the background model. Three intervals - Before HV change - After HV change, before linearization table change - After HV change, after linearization table change Calibration source regions to be excluded. SMN: Use Katja's scripts for WWW page with information? Prioritization -------------- Tasks ordered by analysis level. Symbols: ! = urgent, before Christmas + = required for first s/w release, by end of Jan ~ = can come later PRP ! Avoid spurious alerts in j_prp_verify ! Update gain fitting COR ! Allow for gain evolution model, skip glitches ! Flag events also based on detector model DEAD + Deadtime estimate with few % accuracy BIN_I ! Use event flags ~ use background model IMA ! Correct source positions for zero point and misalignment ! Correct source identification [PK] ! Implement IROS + Add significance map or similar image + Include NL's image deconvolution algorithm as alternative ~ Clean up systematic noise ~ Use pixel size information from IC ~ Include SL's code as check on results SPE & LCR ! Pulse phase resolved spectroscopy ! Use event flags + Support Restricted Imaging ~ use background model BIN_S & BIN_T ! Include radiusLimit ! Use event flags Scripts ! Help files + Advanced GUI IC files ! Include dead anodes and calibration sources in detector model ! Add pixelsize tables ! New spectral response + Update FULB|RESB|.... tables Documentation + ADD update + Analysis User Manual AOB --- Shortage of IRs for December. SL, SMN agree to check their availability. PK will be at ISDC, Dec 27-31 Next Meeting ------------ 1st week of April