Subject: Gain aging proposal Dear All In fulfillment of AI30_13 I am sending you two proposals for implementing the individual aging of each calibration spectrum collection area. As we have seen, each of the calibration spectra from the two instruments is evolving somewhat differently during the course of the mission. In some cases quite suddenly and violently (JEM-X2, anode 4), in others slowly, yet eccentrically (JMX1, anode 4). While we have cut out JEM-X2 anode 4 from the gain calibration process, we are still seeing a steady upward creep of the Xe line position due to nonuniformity of the plate aging. Somehow we have to add an aging factor into the gain determined from the calibration sources to compensate for this evolution. ********** Proposal 1: Almost already implemented ************************* Lurking in the header of instrument model table JMXi-SPAG-MOD are a number of useful keywords: - USECAL_n is a set of flags indicating whether a particular source should be used for gain calibration or not. Currently these all have the value 1 except for USECAL_4 in JMX2-SPAG-MOD, meaning that we don't use this source for calibration due to its early unstable nature and the fact that though stable it is now way out of line with its pre-flight position. It could however, be put back into use with this proposal. - XE_CHAN indicates the postion of the Xe line found in uncorrected data (onboard ADC channel) during a certain reference period during which gain was constant. - PKCHAN_n is a set of floating point numbers that indicate the ADC channel position of each of the calibration peaks during the same reference period Currently, the values of XE_CHAN and PKCHAN_n are the same for all the instances of SPAG-MOD delivered to ISDC and correspond to the values of these positions at the beginning of the mission when they conformed to Carls values used during the SPAG table determination. Previously, it was not expected that these values would change. However, the values can be updated for each IMOD period (we have had 7 such periods so far) to reflect the slow evolution of each source peak position. Where a source changes too quickly to be described by a value covering 40 or so revolutions, the USECAL_n flag can be set to zero to take this source out of the processing just for that one IMOD period. The actual values required for each period are quite easy to obtain because they are all based on uncorrected data with no reference to which source or sources are being used for correction. Actual use of the keyword values is already implemented in j_cor_gain, so the only work involved is in finding new values of XE_CHAN and PKCHAN_n for each IMOD period, and making sure they get included in the next suite of IMOD tables delivered to ISDC. *************************************************************************** However, for fine-tuning of the aging factor, we will need something more complex, if for example, we feel that 7 time periods is not sufficient to describe the aging process. Proposal 2: Super deluxe We could introduce a new IMOD table for each unit, JMXi-XENO-MOD, that will contain four columns (one for each calibration source) and a row for each revolution. Tabulated by revolution and source will be the best-known average position of the Xe line found in corrected data (ISDC level COR) using only the specified source for the gain correction. This will mean that someone (presumably me) will have to find empty fields in as many revolutions as possible, run the gain correction with a different values of USECAL_n in the header of JMXi-SPAG-MOD, so that only one calibration source at a time is used. The Xe line position in `ISDC-corrected' KeV can then be found from the detector binned spectra using Jerome's program. We take the average of as many empty fields as time or the data will allow, and tabulate this in JMXi-XENO-MOD. Gain calibration program j_cor_gain will then read in the four values for a given revolution and use these to correct the corrected data. Evidently, at certain times some sources have changed their peak position with respect to the other sources so radically that no single per-revolution value could describe this behaviour (JEM-X2, source 4, revolutions 7, 14, 18, 20, 21, 22, 23) and a negative tabulated Xe line position could be used to indicate that the source should not be used for gain correction in this revolution. Practical limitations indicate that we cannot produce a value of Xe position for every single revolution of the mission. Where data is lacking the last previously determined value will be used. The same value could be used for many revolutions since the gain drift is only significant over many revolutions. However, we (I) must be careful to cover sufficiently time periods where an individual source may be changing its behaviour relatively rapidly. Where a unit is switched off, the last known line position will be tabulated. JMXi-XENO-MOD, like all IMOD files will have to be updated at regular intervals. In revolutions where no tabulated values yet exist, j_cor_gain can use the latest values in the table. We can perhaps extrapolate our known values into forthcoming revolutions so as to improve on the `latest values' estimate. We cannot however, possibly predict sudden radical changes in behaviour, so NRT processing may be somewhat lacking in the precision of its gain calibration. The above proposal will require the creation of a new data template, and addition of this data structure to the IMOD group data structure. The table itself must be filled and then updated regularly, perhaps at short notice of an anode running through a particular calibration area dies. While it is possible to tune the gain correction revolution-by-revolution, at least in re-processing, this option involves alot of work. ******************************************************************************* I hereby declare AI30_13 well and truly closed. Best wishes, Carol Anne PS If I don't hear from you by the New Year, I'll implement proposal 1. .DANISH.SPACE.RESEARCH.INSTITUTE.DANISH.SPACE.RESEARCH.INSTITUTE.DANISH. Dr. Carol Anne Oxborrow Email: oxborrow@dsri.dk Homepage: http://www.dsri.dk/~oxborrow Telephone (direct): +45 35 32 57 33 Telephone (secretary): +45 35 32 57 01 Fax: +45 35 36 24 75