Description of mMvbdNdEta

The calling sequence for this module (with links to the source code) is:

mMvbdNdEta dMvdGeo dMvbGeo dMvdPar dMvbPar dMvbDbase.idl dMvbRaw dMvdVertexOut dMvddNdEtaPar dMvddNdEtaOut dMvdMultOut dMvdIo

Calculate dN/d(eta) distributions from the MVD using the digitized information from the barrel. In the MVD strips (barrel part of the detector), the current algorithm uses only the inner shell. In central Au+Au Hijing events, the occupancy in the inner barrel is high (around 50%) and most particles hit more than one strip -- therefore the algorithm makes no attempt to find individual hits. Instead, the barrel is divided into groups of adjacent strips (currently the algorithm uses groups of 64 adjacent strips) which are in the same row (azimuthal segment). For each of these groups of channels, a dN/deta value is calculated as follows:

  • the total ADC value for all channels in this group is added up,
  • the range of pseudo-rapidity (delta-eta) is calculated,
  • the pseudo-rapidity (eta) at the center of the group of strips is calculated,
  • a geometric correction (for non-normal incidence) is applied,
  • the number of particles is estimated by dividing the corrected ADC signal by the ADC signal expected for a particle at the measured pseudo-rapidity.
  • dN/deta is the estimated number of particles divided by delta-eta. The average dN/deta for a set of events is calculated by taking the average (using a profile histogram) in each bin of eta.

    Results are returned in

  • dMvddNdEtaOut[next].dndeta = dN/deta/dphi value
  • dMvddNdEtaOut[next].dndetaerr = estimated uncertainty on dN/deta
  • dMvddNdEtaOut[next].eta = eta at center of bin
  • dMvddNdEtaOut[next].deta = size of eta bin
  • dMvddNdEtaOut[next].phi = phi value
  • dMvddNdEtaOut[next].dphi = delta phi
  • dMvddNdEtaOut[next].softdndeta = software ID for algorithm
    softdndeta is set to 1.xx (currently 1.01) for the algorithm which combined all rows (azimuthal segments) at a given eta value into a single entry. It is set to 2.xx (currently 2.01) for the algorithm which produces separate dN/deta/dphi values for each row. Every time the module is called, it puts both sets of entries into the output.

    There is a complicated calibration parameter involved in this algorithm. The original algorithm converted ADC signals to mips by asssuming all particles were "mips". This is clearly wrong. It is also clear that the average dE/dx should change (at the 10% level) with pseudo-rapidity. In an effort to include this correction, a calibration/fit was made to the average ADC signal vs. eta. This fit is of the form: Average Signal = a + b*(Eta^2) + c*(EtaAvg^4) + exp( d + f*abs(Eta) )
    where "Average Signal" is in ADC channels. The five parameters in this algorithm are from a "database" file filled by mMvdDbase. They are stored as:
    a=dMvbDbase[index].calib_const[0];
    b=dMvbDbase[index].calib_const[1];
    c=dMvbDbase[index].calib_const[2];
    d=dMvbDbase[index].calib_const[3];
    f=dMvbDbase[index].calib_const[4];
    where index refers to channel numbers. Despite the effort put into this effort, we still get better results by assuming all particles are mips. We do not understand this, put think that background particles passing through the MVD at an unexpected angle of incidence are causing calibration problems. The Yonsei group is working on this problem an a solution is expected soon (in a month or so).

    The module also fills

  • dMvdMultOut->totalmult = estimated total multiplicity in inner barrel,
  • dMvdMultOut->totalmulterr=estimated uncertainty
  • dMvdMultOut->softmult =1.xx (currently 1.01)

    John Sullivan
    comments to: sullivan@lanl.gov
    updated 23-Dec-1999