Preliminary gains of PRES-1 from MIP's
(found with with 00xx00 in SMD and Tower=MIP)

  • Input: ~1.6M minB pp events from 2004 run. Only data from sector 5 are analyzed below.
  • 8 '0' before and after 'x' in the SMD pattern.
  • MIP response from towers: ADC-ped in [0.5,1.5]*predicted_MIP_response
  • Pres-2 and Post : ADC>ped+3*sigPed
  • no conditions on Pres-1

  • Conclusions:
    * MIP peak is well visible and matches the Landau shape.
    * with current statistics the gains of preshower 1 can be determined with accuracy of ~13%


    Fig 1.Example of one pres1 tile response (05PB08). Inclusive spectrum on the top, MIP gated one on the bottom. Gaussian shape was fitted to pedestal (blue), Landau to the spectrum above pedestal (red). Simultaneous fit is in green.

    60 plots for every tile of preshower-1 in etaBin=1, etaBin=2, etaBin=3, etaBin=4, etaBin=5, etaBin=6, etaBin=7, etaBin=8, etaBin=9, etaBin=10, etaBin=11, etaBin=12,


    Fig 2. Summary of results for the whole sector 5 (~60 tiles).
    Top left : MPV of Landau. Most of MPV is within 15 +/-7 ADC channels. ( no correction for eta dependence of scint thickness)
    Top right: relative error of MPV - precision of gains with 1.6 M eve would be of 13%
    Bottom left: ratio of width of Landau peak to its MPV. With current HV an average tile response to MIP is 2 'sigma' away from pedestal.
    Bottom right is for Scott. Lets assume MPV for an inclusive spectrum can be associated with 1 p.e. response of a tile. MIP should produce few p.e.. The correlation between MPV from the inclusive spec vs. from the one gated with MIP is visible. Also the later is always larger (note, the X/Y scales are different).
    If you believe in the above then in reality MIP is producing 1.5 -2 p.e. per tile.

    Small print:
  • used peds were ~1 month old
  • I did not care much if Landau fit failed for the inclusive spectra
  • Landau sometimes is too low for the MIP gated spectra
  • MIP gated spectra were rebinned over 2,3, or 4 bins if the integral of ADC-ped over range [5,40] was below certain thresholds.
  • some tiles do not show up
  • one could explore if ratio of (pres_1 ADC-ped) / (tower energy) is not better defined, but first tests (not shown here) indicate it is not.

    Macro used to generate plots: pl3.C

    Sum over all subsectors for each η Bin