The slow simulator includes a detailed simulation of the detector response.
It creates electron cluster as hits in the drift volume and transforms them
into digits. A schema of the code can be found
here.
For the slow simulation version 1 of the TRD
(AliTRDv1) has to be selected in Config.C
The following picture shows a display of the hits in the TRD:
Generation of Hits
The StepManager()
function in
AliTRDv1
creates along the path of a traversing charged particle in the drift volume
of the chambers electron clusters. The distance between the points where the
creation of the primary delta-electrons takes place is set according to a
parametrization of the Bethe-Bloch formula (see below).
The data points are averaged values taken from GEANT. The plateau value is
1.55. For the number of produced delta-electrons at the minimum ionizing
point a value of dN1/dx|min = 19.3/cm is used.
The energy distribution of the delta-electrons is extracted from GEANT3
(red curve in plot below, compared to a 1/E distribution, blue curve).
The energy of one delta-electron is given by a random sample from this
distribution subtracted by the first ionization potential (12.1 eV).
Each produced delta-electron in turn can create a number of secondary electrons
by ionization. Their number is calculated by simply dividing the energy of the
delta-electrons by the ionization energy of 22.04 eV. The average number of
electrons produced by 3 GeV/c pions traversing 1 cm of gas is 282.
The AliTRDdigitzer
class transforms the hits from
AliTRDv1
into digits. In this process the following effects can be included:
Effect
Default Status
Diffusion
ON
E x B
OFF
Absorbtion
OFF
Pad response (1-dim)
ON
Gas gain + gain fluctuations
ON
Electronis gain + noise
ON
Conversion to ADC values
ON
All corresponding parameter can be customized (see the example macro
slowDigitsCreate.C).
The default parameter are defined in
AliTRDdigitizer::Init().
These one-dimensional pad response functions (which correspond to each layer)
decribes charge sharing in column direction:
The fluctuations in the gas gain are described by an exponential distribution:
The output of the digitization procedure are objects of the type
AliTRDdigit:
Since the data volume gets extremly large if a full event with realistic
multiplicity is processed, the digits are internally stored in a compressed
data structure, following a scheme invented by Marian for the
TPC.
The class AliTRDdigitsManager provides the
interface through which the digit objects are stored and reconstructed. It also manages
the dictionary between Monte Carlo tracks and digits.
Clustering
The class
AliTRDclusterizerV1
provides a clustering algorithm that looks in each timebin for local maxima
in the charge distribution in row- column-direction.
Merged cluster are separated by an unfolding algorithm. There is no
clusterfinding done in time-direction. The example macro
slowClusterCreate.C
demonstrates how to do the clustering. The resulting space points are stored as
AliTRDrecPoint objects:
Page maintaned by
Christoph Blume.
Last modification: 02/06/04 (A.Andronic)