A number of detectors have already been defined. To define your own, it is best to start with one of them and modify as necessary.
Full Detector Description
The detector geometry is defined using the compact format. Instructions for creating this detector descritpion file can be found at Creating a New Detector Description.
The resulting lcdd file needs to be checked for overlaps.
The easiest way to do this is to create a macro file containing the following command
To check for overlaps:
> slic -g detector.lcdd -z -m geom_chk.mac
There may be a few spurious overlaps within calorimeters due to roundoff errors.
All other overlaps should be resolved.
Fast Monte Carlo Files
The current fast Monte Carlo smears charged tracks covariantly, and creates clusters for neutral particles, treating photons and neutral hadrons separately.
The fast Monte Carlo requires the following properties files to provide input:
The minimum transverse momentum as well as angular fiducial regions are defined here. The names of the track smearing files are also listed here. These files are generated using the lcdtrk program. These files should reside in the directory TrackParameters.
The file SimpleTrack provides the ability to define momentum smearing parameters independent of the lcdtrk parameters.
This file contains the fiducial cuts on momentum and angular coverage as well as expected energy resolution terms for the electromagnetic and hadronic calorimeters. Electromagnetic and hadronic calorimeter resolution terms will be scaled to achieve the jet energy resolution specified by parameter JETResolution if the boolean parameter JETParameterization is set to true. To perform this scaling properly you must also specify the fraction of the jet energy that is electromagnetic (JETEMEnergyFraction) and the fraction due to neutral hadrons (JETHadEnergyFraction). You can choose to only scale the hadronic energy resolution (JETHadDegradeFraction=1), to only scale the electromagnetic resolution (JETHadDegradeFraction=0), or to scale both using an intermediate value for JETHadDegradeFraction. More information on the jet energy resolution algorithm can be found on pp. 162-163 of the SiD Detector Outline Document.
This provides a simple method for simulating the confusion terms in identifying individual particles. In addition, the parameter wt_charged_track_calorimeter_energy determines how often a charged hadron's reconstructed energy will be set equal to the calorimeter energy. For example, if wt_charged_track_calorimeter_energy=0 then the tracker momentum will always be used for the charged hadron's reconstructed energy; if wt_charged_track_calorimeter_energy=1 then the calorimeter energy will always be used for the charged hadron's reconstructed energy.
This file contains the parameters for smearing the transverse momentum p of charged tracks according to the simple formula ?p/p 2 = a + b/psin? . This method of smearing track parameters is
enabled by setting the MCFast class constructor argument simple=true. The file parameter ConstantTerm = a and the file parameter ThetaTerm = b . By default, the errors for the track angles and impact pararameters are taken from the files defined in TrackParameters.properties. However, the angle and impact parameter errors can be scaled relative to these default errors using the SimpleTrack.properties parameters TanLambdaErrorScale, PhiErrorScale, D0ErrorSCale and Z0ErrorScale.
A digisim.steer file is needed in the digisim directory. This file contains timing and threshold
cuts for the hits in the calorimeters. Using the same file as a previous detector is fine, with
the follwing caveats: Scintillator and gas calorimeters have different hit thresholds, so use the
appropriate file. Also, files for detectors prior to sid01 do not contain muon calorimeter
Calorimeter Sampling Fractions
(see also the lcsim-cal-calib package)
Sampling fractions need to be provided for each calorimeter defined in the detector. Currently there
needs to be one file for each calorimeter in the directory SamplingFractions. The files may contain
different values for different layers or ranges of layers (needed for Ecal), and a digital flag for
the hits to be treated digitally. These numbers are used to give an energy estimate of hits and
clusters without angle or particle type corrections.
Generating the numbers
Once the digisim.steer file is in place, a driver QSFCalibrationFromData (soon to be in CVS) can be
run to generate the numbers for each calorimeter (or sub-calorimeter). The procedure minimizes
dE/sqrtE per event, using only the calorimeter information. Three sets of numbers are generated.
(Ecal+Hcal, Ecal+Hcal+Muon, Ecal+Hcal+MuonEndcap) The input files must be the qq events at fixed
energies. The numbers will vary significantly depending on the cm energy, so typically the driver
is run over a combination of 10k events each at 100,200,500 GeV cm energy. The numbers can then be
placed in the Sampling fraction files.
The calibration files reside in the photonCalibration and hadronCalibration directories. The
constants are derived by a driver (QCalibrationFromData, soon to be in CVS) which uses the cheat
reconstruction. Therefore, in addition to the sampling fraction files, all the FastMC files must
also exist. The procedure is to attach calorimeter hits to the neutral particles, calculate sampling
fractions independently for photons and neutral hadrons minimizing dE/sqrtE assuming an angle
correction, and calculate response corrections (for nonlinearity). The calibration depend on the
distributions in energy, angle, and particle type, and is typically run on the ZZ events at 500 GeV.
The files photon-(slic version).properties and nh-(slic version).properties can then be created, and
are used by the classes QPhotonClusterEnergyCalculator and QNeutralHadronClusterEnergyCalculator.
To run the pflow reconstructions, the file LongitudinalHMatrix.hmx is needed. Ask Norman Graf.