PDS_VERSION_ID = PDS3 RECORD_TYPE = STREAM OBJECT = TEXT INTERCHANGE_FORMAT = ASCII PUBLICATION_DATE = 2001-09-01 NOTE = "N/A" END_OBJECT = TEXT END XCALOVER.TXT Written as part of the NEAR XGRS submission to the Planetary Data System (PDS). GRS Energy Calibration As part of the data processing system a quick-look system for screening GRS data and an energy calibration system were developed and were used for routine GRS processing. The quick-look system includes the inspection of daily spectral sets. Most bad spectra are flagged automatically by the data quality indicator (DQI) flag in each spectral file delivered by the science data center (SDC) at APL. For other problems identified during visual inspection of daily summed spectra that were not flagged by the DQI, a problem code is selected and a file is automatically generated that contains the MET and problem code for each bad spectrum. Gamma-ray fluxes are very low and spectra (collected typically for 1200-2400s) must be summed together before any analysis can be done. Spectra must be on the same pulse-height (or energy) scale before summing in order not to degrade the energy resolution of the detector. Large gain changes seen early in mission were due to PMT aging and not due to temperature or voltage changes. Aging effect on the calibration is expected to be much smaller during orbital operations. An energy calibration system is used after the data has been reviewed with the quick-look system. The calibration system is used to deter- mine the gain and zero offset, corresponding to the linear relation- ship: Energy = Gain * Channel + Zero. Two calibrations are required; one for NaI and one for BGO. The raw and anticoincidence NaI spectra are taken simultaneously and will have the same energy calibration. The actual energycalibration of the coincidence spectra depend both on the energy calibration of the central detector and on the window settings. These will be considered separately. For this energy calibration two background lines that can be identified consistently in the spectra are needed. For calibration purposes during cruise, one-day summed spectra are used for analysis. The assumption is that the gain/zero of the spectra do not change appreciably in 24 hours. Where this is not true, such as commanded changes in the instrument settings, then the spectra are summed in smaller time intervals. Whenever possible, commanded changes that affect the energy scale were set to occur at or near day (UT) boundaries. During orbital operations, the higher fluxes expected at Eros could make it possible to use shorter accumulation times for calibration. Energy calibration for the NaI detector was done using a low-energy peak around channel 18 in the anticoincidence spectrum and a peak around channel 50 in the raw NaI spectrum. Analysis early in the cruise tentatively identified the low energy peak in the spectrum as 175 keV from the decay of 71mGe from the shield material and the other peak as the annihilation line at 511 keV. Later analysis of cruise spectra and comparisons with calculations indicated that the peak in question in the anticoincidence spectrum may actually come from the decay of 123I with an energy of 189 keV. Since the purpose of the energy calibration is to allow the summing of spectra without loss of energy resolution the calibration system continued to assume the peak was at 175 keV. Tests have shown that spectra all corrected using the same peak identification can be summed with good energy resolution preserved even if the identification is incorrect. A final transformation to the correct energy scale can be made after summing. The BGO energy scale was established using the prominent gamma-ray peaks identified at 511 keV and 2223 keV. It is expected that prominent gamma-ray lines detected during the orbital phase that will be used for energy calibration will be different than those used during cruise. The spectral lines are fitted with a hard-coded function in the calibration system consisting of a third-order polynomial to fit the background and a Gaussian peak. An interactive system is used to determine when the peak fits are satisfactory by visual inspection and then the resulting gain/zero parameters are stored for further use. The gain/zero results for each daily sum for the two detectors are used as input into the next step in energy calibration. The complete time histories of gain and zero for both detectors are used to interpolate the exact values to apply for each MET. Time intervals, based on the one day accumulation fits, are selected interactively where no abrupt change in the energy calibration is noted. A n-order polynomial is selected to fit the time series of gain or zero values where the polynomial order is selected based on an observed fit to the data. Whenever possible a lower order poly- nomial fit was preferred to a higher order polynomial to prevent wild oscillations in the fitted values. Highly erratic time series were broken up into smaller time intervals for fitting. The time series breakdown of the mission into intervals facilitates the use of a simple function to accurately approximate the real gain/zero for each detector. The fitting function is then used to calibrate any given spectrum by interpolation using the appropriate function for the time intervalunder consideration. The new coefficients sets are made available to the query calibration processing system on the XGRS database.