The STAR experiment

Mainly a rundown of issues described in http://www.star.bnl.gov/HyperNews-star/pr

 Trigger      #events    CPU/evt (3.0 GHz)    Timescale for production

.

BSMD

The SMDs can be simulated in one of two modes.  In a rewrite of the slow simulator I call them kTestMode and kSimpleMode.  kTestMode doesn't do anything interesting, so let's focus on kSimpleMode.  Here are the steps to generate an ADC:

1. ADC = energy deposit * sampling fraction(eta) / calib
2. ADC += mRandom.Gaus(pedMean, pedRMS)
3. ADC *= mRandom.Gaus(1.0 + calibOffset, calibSpread)
4. force ADC between 0 and maxADC

The energy is then calculated from this ADC is the usual way:  energy = (ADC - pedMean) * calib.  Remember the BEMC calibration coefficients are stored as GeV/ADC.

BTOW - BPRS

The simulators for these two detectors have some additional modes (kPrimaryOnlyMode, kPrimarySecondaryFastMode, kPrimarySecondaryFullMode).  Here's the setup for the kPrimaryMode:

1. photoElectrons = energy deposit * (MIP photo electrons / MIP energy deposit)
2. smear using Poisson distribution
3. ADC = photoElectrons * sampling fraction(eta) * (MIP energy deposit / MIP photo electrons) / calib
4. follow steps 2-4 of simple simulator to add pedestals, smear calibration, and check for valid range

The other two modes also take the secondary photostatistics from the PMT amplification into account.  I'll focus on the full mode, as it's slightly easier to understand:

1. build a set of secondary electron conversion coefficients (1 / dynode) using the relative voltage distribution among the dynodes and an approximate value for the total PMT amplification.
2. Loop through dynodes and at each step
   1. nElectrons = mRandom.PoissonD( mSecondaryCoeff[i] * nElectrons + mDynodeNoise );
3. ADC =  nElectrons * mTotalDynodeGain *sampling fraction(eta) * (MIP energy deposit / MIP photo electrons) / calib
4. follow steps 2-4 of simple simulator above

The mTotalDynodeGain in step 3 is the inverse of the product of the secondary electron conversion coefficients.

sampling fraction -- each sampling fraction is a 2nd order polynomial in eta.  Here are the values currently being used:
BTOW = 14.69 - 0.1022*eta + 0.7484 * eta^2
BPRS = same as BTOW (pams/emc/inc/samplefrac_def.h includes different number)
BSMDE = 118500.0 - 32920.0*eta + 31130.0*eta^2
BSMDP = 126000.0 - 13950.0*eta + 19710.0*eta^2

Here are the numbers for BPRS from pams/emc/inc/samplefrac_def.h:
#define P0BPRS  559.7
#define P1BPRS -109.9
#define P2BPRS -97.81

max ADC value -- BTOW(3500), BPRS(220), SMDs(900)

Calibration offset (0.0) -- shifts all calibration coefficients during generation of ADCs.

Calibration spread (0.0) -- smears ADCs with a Gaussian.  A spread of 0.15 for the BTOW has been used in recent analyses.

The following params are only used by the slow simulator for the BTOW and BPRS:

MIP photo electrons -- BTOW(63.0), BPRS(6.0)
MIP energy deposit -- BTOW(19.8 MeV), BPRS(2 MeV)

The ratio of these two is used to translate the energy deposit from GEANT into a number of photoElectrons which is then fed into the StPmtSignal simulator.

approximate PMT amplification (1.5e+6).  Does NOT affect overall gain, which is a separate input parameter.  Basically only determines some widths in the PMT simulator.

relative dynode voltage distribution {2.,2.,1.,1.,1.,1.,1.,1., 2., 3., 4.}

cathode noise (0.0) A probability for a thermal electron (or the average number of thermal electrons per pulse) to spontaneously emerge from the photocathode within the ADC gate.

dynode noise (0.0) A probability for a thermal electron (or the average number of thermal electrons per pulse) to spontaneously emerge from a dynode within the ADC gate. This probability is assumed to be the same for all dynodes.

For these 2 detectors one can choose between simple, fast, and full simulators.  The only difference between full and fast is that at high numbers of photo electrons the PMT simulator switches to Gaussians instead of Poisson distributions.  We've been inadvertently running the fast simulator for some time now due to a flip in indices in the simulator code, but according to V.Rykov's documentation the two are in good agreement.  The simple simulator is the one used by the SMDs; no accounting for secondary photostatistics is done, so the ADC value is basically (energy deposited * sampling fraction * gain), plus pedestal noise and calibration smearing.

I continue using HIJING and trying to understand the energy-momentum-conservation
and acceptance effects in d+Au.

I continue using HIJING and trying to understand the energy-momentum-conservation
and acceptance effects in d+Au.

The PDSF FAQ can be found under http://wiki.nersc.gov ->STAR->FAQ. Here is an echo of the FAQ as of 8/22/2007.