java/trunk/hps-java/src/main/java/org/lcsim/hps/readout/ecal
--- java/trunk/hps-java/src/main/java/org/lcsim/hps/readout/ecal/FADCEcalReadoutDriver.java 2014-02-11 21:24:17 UTC (rev 215)
+++ java/trunk/hps-java/src/main/java/org/lcsim/hps/readout/ecal/FADCEcalReadoutDriver.java 2014-02-12 23:40:01 UTC (rev 216)
@@ -55,11 +55,11 @@
private double tp = 14.0;
//TODO: set riseTime, fallTime, pulseDelay
//pulse rise time in ns
- private double riseTime = 14.0;
+ private double riseTime = 10.0;
//pulse fall time in ns
- private double fallTime = 14.0;
+ private double fallTime = 17.0;
//pulse delay time in ns
- private double pulseDelay = 14.0;
+ private double pulseDelay = 3 * riseTime;
//delay (number of readout periods) between start of summing window and output of hit to clusterer
private int delay0 = 32;
//start of readout window relative to trigger time (in readout cycles)
@@ -91,6 +91,12 @@
private boolean constantTriggerWindow = false;
private boolean addNoise = false;
//TODO: change to 2014 value
+ private double lightYield = 120.; // number of photons per MeV
+ private double quantumEff = 0.7; // quantum efficiency of the APD
+ private double surfRatio = (10. * 10.) / (16 * 16); // surface ratio between APD and crystals
+ private double gainAPD = 150.; // Gain of the APD
+ private double elemCharge = 1.60217657e-19;
+ private double gainPreAmpl = 0.550e12; // Gain of the preamplifier in V/C, true value is higher but does not take into account losses
private double pePerMeV = 2.0; //photoelectrons per MeV, used to calculate noise
public FADCEcalReadoutDriver() {
@@ -203,20 +209,29 @@
for (Long cellID : eDepMap.keySet()) {
RingBuffer eDepBuffer = eDepMap.get(cellID);
+// TODO : gain is not necessary anymore for 2014 run, as I have introduced all the variables in the same function
+// If a variable gain should be introduced from a database, it should be done in pulseAmplitude in order to have all the information at the same place
+// Pedestal should stay here
FADCPipeline pipeline = pipelineMap.get(cellID);
pipeline.step();
double pedestal = EcalConditions.physicalToPedestal(cellID);
- //normalization constant from cal gain (MeV/integral bit) to amplitude gain (amplitude bit/GeV)
- double gain;
- if (fixedGain > 0) {
- gain = readoutPeriod / (fixedGain * ECalUtils.MeV);
+
+ double currentValue;
+ if (useCRRCShape) {
+ //normalization constant from cal gain (MeV/integral bit) to amplitude gain (amplitude bit/GeV)
+ double gain;
+ if (fixedGain > 0) {
+ gain = readoutPeriod / (fixedGain * ECalUtils.MeV);
+ } else {
+ gain = readoutPeriod / (EcalConditions.physicalToGain(cellID) * ECalUtils.MeV);
+ }
+ currentValue = gain * eDepBuffer.currentValue();
} else {
- gain = readoutPeriod / (EcalConditions.physicalToGain(cellID) * ECalUtils.MeV);
+ currentValue = eDepBuffer.currentValue() * 4095.0 / 2.0; //12-bit ADC with 2 V range
}
- double currentValue = gain * eDepBuffer.currentValue();
- pipeline.writeValue((int) Math.round(pedestal + currentValue));
+ pipeline.writeValue(Math.min((int) Math.round(pedestal + currentValue), 4095)); //ADC can't return a value larger than 4095
Double sum = sumMap.get(cellID);
if (sum == null && currentValue > triggerThreshold) {
@@ -417,7 +432,7 @@
energyAmplitude += RandomGaussian.getGaussian(0, noise);
}
for (int i = 0; i < bufferLength; i++) {
- eDepBuffer.addToCell(i, energyAmplitude * pulseAmplitude((i + 1) * readoutPeriod + readoutTime() - (ClockSingleton.getTime() + hit.getTime())));
+ eDepBuffer.addToCell(i, energyAmplitude * pulseAmplitude((i + 1) * readoutPeriod + readoutTime() - (ClockSingleton.getTime() + hit.getTime()), hit.getCellID()));
}
}
}
@@ -452,7 +467,7 @@
return true;
}
- private double pulseAmplitude(double time) {
+ private double pulseAmplitude(double time, long cellID) {
if (useCRRCShape) {
if (time <= 0.0) {
return 0.0;
@@ -466,11 +481,27 @@
return 0.0;
}
}
- } else {
- return 0.0; //TODO: new pulse shape definition for 2014
+ } else { // According to measurements the output signal can be fitted by two gaussians, one for the rise of the signal, one for the fall
+ // Time corresponds to t-(t_eve+pulseDelay) such that the maximum of the amplitude is reached pulseDelay ns after the event t_eve
+ // Without the coefficient, the integral is equal to 1 - 1 per 1000
+ if (time <= 0.0) {
+ return 0.0;
+ }
+ double norm = ((riseTime + fallTime) / 2) * 1e-9 * Math.sqrt(2 * Math.PI); //to ensure the total integral is equal to 1; gives 3.3839e-8 for default rise and fall times
+
+ if (time < pulseDelay) {
+ return lightYield * quantumEff * surfRatio * gainAPD * gainPreAmpl * elemCharge * funcGaus(time - pulseDelay, riseTime) / norm;
+ } else {
+ return lightYield * quantumEff * surfRatio * gainAPD * gainPreAmpl * elemCharge * funcGaus(time - pulseDelay, fallTime) / norm;
+ }
}
}
+ // Gaussian function needed for the calculation of the pulse shape amplitude
+ public static double funcGaus(double t, double sig) {
+ return Math.exp(-t * t / (2 * sig * sig));
+ }
+
private class FADCPipeline {
private int[] array;
@@ -518,4 +549,4 @@
return array[((ptr - pos) % size + size) % size];
}
}
-}
\ No newline at end of file
+}
