Currently, two algorithms are being used to calculate dN/deta from the MVD pads. The first algorithm is similar to the algorithm used for the strip detectors. The total ADC value for all pads in the same row (rows are the azimuthal segmentation) in a pad detector are added up. This is divided by the expected dE/dx for a "mip" to estimate the number of particles. The range of eta is calculated from the known geometry of the pads and the vertex location to calculate dN/deta. As with the strips, a profile histogram is used to get the average value of dN/deta in each eta bin. A sample histogram is available as a postscript file. The histogram is from an older version of the code, but the algorithm is unchanged.

Calculating the number of hits by taking the ADC value and dividing by the average signal per mip gives us an algorithm which is sensitive to La=ndau fluctuations, showers in the detector, etc. The following algorithm is an attempt to solve this problem.

The second algorithm used to get dN/deta from the pad detectors assumes that a single particle hits only one pad. The occupancy of the pad detectors is around 15-20% for central hijing events. The algorithm estimates the number of particles associated with the observed ADC value in an individual pad using an ADC distribution from a "calibration" procedure and the occupancy of the pad detector (which is measured for each event). The "calibration" ADC distribution is input to the program via a data-file (ver_calib.dat). If the input file is not found, a default distribution is assumed. This file contains the ADC distribution expected for a single particle at normal incident angle. Currently, this is taken from pisa/staf results in which the hit is known to be due to one particle. Eventually, this would come from low multiplicity events. The algorithm does not associate an integer number of particles with each hit. Instead a mean number of particles which would be associated with a given ADC channel is calculated -- this is not an integer. For example, if we see an ADC signal at 2*(1 mip signal), there is a certain probability that this was caused by a single particle and Landau fluctuations and a probability that it was caused by two particles, and maybe even a chance that there were three hits. The relative probablilities of associating 1, 2, ... hits with the ADC value depends on the occupancy (Poisson distribution assumed), but the result would be between 1 and 2 hits for this example. Some examples of the mean number of hits associated with a given ADC value for different assumed occupancies are shown in this postscript file. This sample histogram comes from an older version of the code, but the algorithm is unchanged. The same (now root) histograms are available in the current code as NcalcVsADCocc1, NcalcVsADCocc10, NcalcVsADCocc25, NcalcVsADCocc50, and NcalcVsADCocc100. This number of hits is then converted into a dN/deta value using the vertex location and the pad geometry. The average dN/deta for a set of events is calculated by taking the average (using a profile histogram) in each bin of eta. A sample (again from an old version of the code) is availabe as a postscript file. Again, the algorithm is unchanged from this older run.