HPGSPC background subtraction

• Offset background subtraction

  • Original procedure

    This procedure is implemented as a shell script hp_dir (documentation TBD c/o IFCAI) which does the following steps :
    • accumulate the HK time profile of the COLPOS collimator position (raw HK)
    • generate "raw" time windows when the collimator is aligned (on), offset in the negative direction (off-) or offset in the positive direction (off+). In practice only off- was used in the standard rocking law.
    • intersect these windows with a previously generated "no occultation" time window
    • accumulate an on and an off- spectrum
    • using a special command which corrects the off- spectrum for contamination by the calibration sources on the collimator wall (reading a measured "difference spectrum" and scaling the line intensity for calibration source decay)

      hpgsoprpha 1 on.spectrum none off-.spectrum on_bs.spectrum 1

  • New procedure

    In the new procedures one follows the same steps, but for the last one which is replaced by
    • the sum to the off- spectrum of an analytical (negative) "difference spectrum" (containing just the Cd and Fe lines) scaled for the respective calibration source decay.
      This is currently implemented only with a piece of private IDL code.
    • the subtraction (soperate) of the resulting corrected background from the on spectrum

  • Occultation background subtraction

    If the rocking collimator is not operated, one might accumulate the on spectrum as above, but must also accumulate a spectrum (with the collimator aligned) during the occultation of the source by the Earth.
    • generate an occultation time window as

      iwindow occulted earth below -2

    • intersect such time window with the collimator aligned one
    • accumulate a raw occultation spectrum (occ)
    • measure the count rate in a pure background region of the on and occ spectrum

      scount on  B 270 470
      scount occ B 270 470

    • compute a correcting factor (negative, small)
      dflux=(oncr-occcr)/(470-270)

    • subtract such constant from the occ spectrum (where non zero)

      soperate occ + CONSTANT  occ1 dflux

    • multiply a standard correction factor (consisting of a low energy exponential tail and the two Xe lines) by the constant

      soperate corrections x CONSTANT  temp dflux

    • add the result to the scaled occultation background

      soperate occ1 + temp occ2 

    • subtract this from the on spectrum

      soperate on - occbg2 net