11.1-bis LECS event corrections

However the distinction between mandatory and optional corrections is somewhat different.

• Time adjustments

  LECS packets present regularly a number of anomalies with a much greater frequency than other NFI packets, which are affected only by a few "spacecraft anomalies 80" (two consecutive events out of order, which cause all events until end of packet to be assigned by Telespazio times 64 s in the future) in a few cases of strong source observations. These anomalies are listed here :
  • the packet header start time has an insufficient (1s) resolution, and must be replaced by the first event time, unless the first event is affected by the 64 s anomaly. In the latter case the first event time is unconditionally corrected before proceeding.

  • the packet header end time also has an insufficient (1/256 s) resolution, and must be replaced by the time of the last event, unless this is affected by an error, which may be one of these :
    • the last event time is incorrectly set to zero. In this case it is replaced by the lower precision header end time !
      This has been changed : now the zero time is replaced by the time of the previous event (this will mantain the sequence of times).
    • the last event time is affected by the 64 s anomaly. In such case the event time is unconditionally corrected before proceeding.
    • the packet header end time is totally spurious (too small). A separate warning is issued, and the header time is replaced by the time of the last event as usual.

  • any other sequence of event times in the packet may be affected by the 64 s anomaly (events are assigned a time 64 s in the future). These events are spotted and currently optionally rejected or corrected (should be corrected) as controlled by the badtimes variable.

• Fixed shifts

  The telemetry values of X,Y, PHA and VETO (but not the Burst Length) are always subtracted a fixed shift of 3 pixels/channels.

• out-of-FOV rejection

  While SAXDAS performs this explicitly as the rejection of events outside a given radius in raw coordinates, there is no such thing in XAS. One can emulate the function using the standard arbitrary spatial region selection described below.
  One usually does that implicitly selecting smaller regions around the source.

• Positional dependency of gain

  This correction is mandatory if other corrections are wished. The raw PHA (energy) of an event is corrected dividing it by a factor g(x,y) which is the relative gain w.r.t. the detector centre (mapped in raw detector positions), i.e. the gain is effectively brought back to the detector centre. The LECS correction map is defined only inside the FOV (an ad-hoc arrangement is used for calibration sources).

• Linearization

  This correction is mandatory if other corrections are wished. The raw detector positions x,y of each event are first converted to mm on the focal plane using a cubic relationship (with crossed terms in xy, but no energy dependency), and then to "new" pixels of user selected size. Similarly to the MECS, but contrary to SAXDAS no compensation for misalignments (rotation) is applied at this stage.
  The LECS linearization coefficientes are defined only inside the FOV ; outside it one does a plain geometric scaling. This can cause a spurious ring at the edge of the FOV.

• Time dependency of gain

  This optional correction multiplies the positionally-corrected PHA of each event by a correcting factor. This takes into account the gain history described elsewhere (relative gain vs time) and also normalizes to a common energy scale (so-called PI channels).

• Veto/BL filtering

  While in SAXDAS this is implemented as a combined operation, in XAS one has two separate functions.
  • One is the rejection of events outside a nominal VETO range. This range is 30-65. It is implemented in XAS as a normal range selection, which must be done explicitly.
    Important !This correction must be used for all normal accumulations (in order to be consistent with response matrices), but must not be applied when accumulating gain histories, because it will reject all calibration source events !

  • The second one is the application of energy-dependent BL thresholds for an user specified acceptance. This correction is mandatory if other corrections are wished (although is effectively disabled if one requires a 100% acceptance).

    Essentially one uses BL thresholds which are different according to the energy (PI channel) of the photon. Events outside threshold are flagged for discard. Thresholds are interpolated from a discrete set of calibration measurements of BL spectra (as shown in Fig. 26 of the A&AS 122,309 paper) according to an user specified acceptance. This in practice means that standard Burst Length ranges must be left wide open (0 255), since they are taken care of by this correction.

• BL-dependent energy correction

  This so called PI to PIC correction, or penetration correction, is a complex correction which takes an energy value, already corrected for gain disuniformities, i.e. a PI channel, and corrects is as a function of the photon energy and Burst Length value. Since the energy is not known, a randomization based on a precomputed probability matrix is necessary. The correction is required only above 2.9 keV.
  This correction is optional and must be turned on independently on the other mandatory corrections (which are a prerequisite).

  For details see the A&AS 122,309 paper, in particular section3.3.

• Selection using arbitrary spatial regions

  This is not properly a correction, and is done in exactly the same way as for the MECS

• Dead time and celestial coordinates

  The dead time correction and the conversion to sky coordinates are not implemented in XAS.

• Always done corrections

  The correction of time errors is always performed except for the events controlled by the badtimes variable which are rejected by default.
  It is recommended to always enable their correction by

  xasset badtimes Correct

  (IT COULD BE SET AS DEFAULT IN THE FUTURE)

  The fixed shifts of 3 PHA/X/Y/VETO pixel/channels is always applied unconditionally.

• Mandatory corrections

  These are enabled as for the MECS:

  xasset correction Enabled

  But for the LECS they must not be applied when accumulating gain histories !!!

  Additionally all primary accumulations programs will currently four additional arguments (three as used also for the MECS) following all others. THIS IS PROVISIONAL AND WILL BE CHANGED IN THE FUTURE TO USE ENVIRONMENT VARIABLES. These extra arguments may take the following forms :

  N npix seed acceptance

  S pixsize seed acceptance

  In the first form one requires npix pixels to cover the original field of view, in the second case one requires the pixel size to be pixsize mm. In either case seed is a seed for a random number generator.
  The current default uses 256 pixels covering the field of view, which gives a pixel size (of approximately 0.1281 mm or 14.2824 arcsec). The FOV radius of 20 mm or 37 arcmin corresponds to 78 such natural pixels. The user can choose pixels of any size.
  This is analogous to SAXDAS : although default LECS event files use 8 arcsec pixels (0.0717524 mm requiring a 457*457 image to cover the nominal range including the calibration sources) SAXDAS code per se allows pixels of any size.

  The fourth argument, the acceptance, is specific of VETO/BL filtering. The recommended (default) value is 0.99 (99%). A value of 1.00 (100%) effectively disables the BL filtering.

• BL-dependent energy correction

  The PI to PIC correction is disabled by default but can be enabled using

  xasset piccorr Enabled

  provided other mandatory corrections have been enabled too.

• Time dependency of gain

  The (optional but highly recommended) correction of the time dependency of gain is enabled as for the MECS:

  xasset gainhistory gainfile

  where gainfile is a gain history produced by the appropriate command

• Arbitrary spatial region selection

  The (optional) selection of data using spatial regions is enabled exactly as for the MECS, using a map or a region file.
  xasunsetting the selectfile variable is sufficient to disable the selection.

[Previous][Next] [Up]