Index: /trunk/FACT++/src/feedback.cc =================================================================== --- /trunk/FACT++/src/feedback.cc (revision 17172) +++ /trunk/FACT++/src/feedback.cc (revision 17173) @@ -10,4 +10,5 @@ #include "Console.h" #include "externals/PixelMap.h" +#include "externals/Interpolator2D.h" #include "tools.h" @@ -32,5 +33,4 @@ bool fIsVerbose; - bool fEnableOldAlgorithm; DimVersion fDim; @@ -52,4 +52,5 @@ vector fVoltGapd; // Nominal breakdown voltage + 1.1V vector fBiasVolt; // Output voltage as reported by bias crate (voltage between R10 and R8) + vector fBiasR9; // vector fBiasDac; // Dac value corresponding to the voltage setting @@ -64,5 +65,8 @@ double fUserOffset; - double fTempOffset; + vector fTempOffset; + float fTempOffsetAvg; + float fTempOffsetRms; + double fTempCoefficient; double fTemp; @@ -109,4 +113,5 @@ { fVoltGapd.assign(evt.Ptr(), evt.Ptr()+416); + fBiasR9.assign(evt.Ptr()+2*416, evt.Ptr()+3*416); Info("Nominal bias voltages and calibration resistor received."); } @@ -131,5 +136,5 @@ int HandleCameraTemp(const EventImp &evt) { - if (!CheckEventSize(evt.GetSize(), "HandleCameraTemp", 60*sizeof(float))) + if (!CheckEventSize(evt.GetSize(), "HandleCameraTemp", 323*sizeof(float))) { fTimeTemp = Time(Time::none); @@ -137,26 +142,18 @@ } - const float *ptr = evt.Ptr(); - - double avgt = 0; - int numt = 0; - for (int i=1; i<32; i++) - if (ptr[i]!=0) - { - avgt += ptr[i]; - numt++; - } - - if (numt==0) - { - Warn("Received sensor temperatures all invalid."); - return GetCurrentState(); - } - - avgt /= numt; // [deg C] - - fTimeTemp = evt.GetTime(); - fTempOffset = (avgt-25)*0.0561765; // [V] From Hamamatsu datasheet - fTemp = avgt; + //fTempOffset = (avgt-25)*0.0561765; // [V] From Hamamatsu datasheet + //fTempOffset = (avgt-25)*0.05678; // [V] From Hamamatsu datasheet plus our own measurement (gein vs. temperature) + + const float *ptr = evt.Ptr(4); + + fTimeTemp = evt.GetTime(); + fTemp = evt.Get(321*4); + + fTempOffsetAvg = (fTemp-25)*fTempCoefficient; + fTempOffsetRms = evt.Get(322*4)*fTempCoefficient; + + fTempOffset.resize(320); + for (int i=0; i<320; i++) + fTempOffset[i] = (ptr[i]-25)*fTempCoefficient; return GetCurrentState(); @@ -183,8 +180,8 @@ for (int i=0; i vec(416); + /* if (fEnableOldAlgorithm) { + // ================================= old ======================= // Pixel 583: 5 31 == 191 (5) C2 B3 P3 // Pixel 830: 2 2 == 66 (4) C0 B8 P1 @@ -445,5 +444,5 @@ // Offset induced by the voltage above the calibration point - const double Ubd = fVoltGapd[i] + fTempOffset; + const double Ubd = fVoltGapd[i] + fTempOffsets[i]; const double U0 = Ubd + overvoltage - fCalibVoltage[5][i]; // appliedOffset-fCalibrationOffset; const double dI = U0/Ravg[i]; // [V/Ohm] @@ -455,7 +454,4 @@ // Make sure that the averaged resistor is valid const double dU = Ravg[i]>10000 ? r*(I*1e-6 - dI) : 0; - - if (i==2) - cout << setprecision(4)<< dU << endl;; vec[i] = Ubd + overvoltage + dU; @@ -488,141 +484,145 @@ } } - else - { - /* ================================= new ======================= */ - - for (int i=0; i<320/*BIAS::kNumChannels*/; i++) + */ + + for (int i=0; i<320/*BIAS::kNumChannels*/; i++) + { + const PixelMapEntry &hv = fMap.hv(i); + if (!hv) + continue; + + // Number of G-APDs in this patch + const int N = hv.count(); + + // Average measured ADC value for this channel + const double adc = Imes[i]/* * (5e-3/4096)*/; // [A] + + // Current through ~100 Ohm measurement resistor + const double I8 = (adc-fCalibDeltaI[i])*fCalibR8[i]/100; + + // Current through calibration resistors (R9) + // This is uncalibrated, biut since the corresponding calibrated + // value I8 is subtracted, the difference should yield a correct value + const double I9 = fBiasDac[i] * (1e-3/4096);//U9/R9; [A] + + // Current in R4/R5 branch + const double Iout = I8>I9 ? I8 - I9 : 0; + + // Applied voltage at calibration resistors, according to biasctrl + const double U9 = fBiasVolt[i]; + + // Serial resistors (one 1kOhm at the output of the bias crate, one 1kOhm in the camera) + const double R4 = 2000; + + // Serial resistor of the individual G-APDs + double R5 = 3900./N; + + // This is assuming that the broken pixels have a 390 Ohm instead of 3900 Ohm serial resistor + if (i==66) // Pixel 830(66) + R5 = 300; // 2400 = 1/(3/3900 + 1/390) + if (i==191 || i==193) // Pixel 583(191) / Pixel 1401(193) + R5 = 390/1.4; // 379 = 1/(4/3900 + 1/390) + + // The measurement resistor + const double R8 = 100; + + // Total resistance of branch with diodes (R4+R5) + // Assuming that the voltage output of the OpAMP is linear + // with the DAC setting and not the voltage at R9, the + // additional voltage drop at R8 must be taken into account + const double R = R4 + R5 + R8; + + // For the patches with a broken resistor - ignoring the G-APD resistance - + // we get: + // + // I[R=3900] = Iout * 1/(10+(N-1)) = Iout /(N+9) + // I[R= 390] = Iout * (1 - 1/ (10+(N-1))) = Iout * (N+8)/(N+9) + // + // I[R=390] / I[R=3900] = N+8 + // + // Udrp = Iout*3900/(N+9) + Iout*1000 + Iout*1000 = Iout * R + + // Voltage drop in R4/R5 branch (for the G-APDs with correct resistor) + const double Udrp = R*Iout; + + // Nominal breakdown voltage with correction for temperature dependence + const double Ubd = fVoltGapd[i] + fTempOffset[i]; + + // Current overvoltage (at a G-APD with the correct 3900 Ohm resistor) + const double Uov = U9-Udrp-Ubd>0 ? U9-Udrp-Ubd : 0; + + // Iout linear with U9 above Ubd + // + // Rx = (U9-Ubd)/Iout + // I' = (U9'-Ubd) / Rx + // Udrp' = R*I' + // Uov = U9' - Udrp' - Ubd + // Uov = overvoltage + // + // overvoltage = U9' - Udrp' - Ubd + // overvoltage = U9' - R*I' - Ubd + // overvoltage = U9' - R*((U9'-Ubd)/Rx) - Ubd + // overvoltage = U9' - U9'*R/Rx + Ubd*R/Rx - Ubd + // overvoltage = U9'*(1 - R/Rx) + Ubd*R/Rx - Ubd + // overvoltage - Ubd*R/Rx +Ubd = U9'*(1 - R/Rx) + // U9' = [ overvoltage - Ubd*R/Rx +Ubd ] / (1 - R/Rx) + // + + // The current through one G-APD is the sum divided by the number of G-APDs + // (assuming identical serial resistors) + double Iapd = Iout/N; + + // In this and the previosu case we neglect the resistance of the G-APDs, but we can make an + // assumption: The differential resistance depends more on the NSB than on the PDE, + // thus it is at least comparable for all G-APDs in the patch. In addition, although the + // G-APD with the 390Ohm serial resistor has the wrong voltage applied, this does not + // significantly influences the ohmic resistor or the G-APD because the differential + // resistor is large enough that the increase of the overvoltage does not dramatically + // increase the current flow as compared to the total current flow. + if (i==66 || i==191 || i==193) + Iapd = Iout/(N+9); // Iapd = R5*Iout/3900; + + // The differential resistance of the G-APD, i.e. the dependence of the + // current above the breakdown voltage, is given by + //const double Rapd = Uov/Iapd; + // This allows us to estimate the current Iov at the overvoltage we want to apply + //const double Iov = overvoltage/Rapd; + + // Estimate set point for over-voltage (voltage drop at the target point) + //const double Uset = Ubd + overvoltage + R*Iov*N; + const double Uset = Uov<0.3 ? Ubd + overvoltage + Udrp : Ubd + overvoltage + Udrp*pow(overvoltage/Uov, 1.66); + + // Voltage set point + vec[i] = Uset; + + /* + if (fDimBias.state()==BIAS::State::kVoltageOn && GetCurrentState()==Feedback::State::kInProgress && + fabs(Uov-overvoltage)>0.033) + cout << setprecision(4) << setw(3) << i << ": Uov=" << Uov << " Udrp=" << Udrp << " Iapd=" << Iapd*1e6 << endl; + */ + + // Calculate statistics only for channels with a valid calibration + //if (Uov>0) { - const PixelMapEntry &hv = fMap.hv(i); - if (!hv) - continue; - - // Number of G-APDs in this patch - const int N = hv.count(); - - // Average measured ADC value for this channel - const double adc = Imes[i]/* * (5e-3/4096)*/; // [A] - - // Current through ~100 Ohm measurement resistor - const double I8 = (adc-fCalibDeltaI[i])*fCalibR8[i]/100; - - // Applied voltage at calibration resistors, according to biasctrl - const double U9 = fBiasVolt[i]; - - // Current through calibration resistors (R9) - // This is uncalibrated, biut since the corresponding calibrated - // value I8 is subtracted, the difference should yield a correct value - const double I9 = fBiasDac[i] * (1e-3/4096);//U9/R9; [A] - - // Current in R4/R5 branch - const double Iout = I8>I9 ? I8 - I9 : 0; - - // Serial resistors (one 1kOhm at the output of the bias crate, one 1kOhm in the camera) - const double R4 = 2000; - - // Serial resistor of the individual G-APDs - double R5 = 3900./N; - - // This is assuming that the broken pixels have a 390 Ohm instead of 3900 Ohm serial resistor - if (i==66) // Pixel 830(66) - R5 = 300; // 2400 = 1/(3/3900 + 1/390) - if (i==191 || i==193) // Pixel 583(191) / Pixel 1401(193) - R5 = 390/1.4; // 379 = 1/(4/3900 + 1/390) - - // Total resistance of branch with diodes - const double R = R4+R5; - - // For the patches with a broken resistor - ignoring the G-APD resistance - - // we get: - // - // I[R=3900] = Iout * 1/(10+(N-1)) = Iout /(N+9) - // I[R= 390] = Iout * (1 - 1/ (10+(N-1))) = Iout * (N+8)/(N+9) - // - // I[R=390] / I[R=3900] = N+8 - // - // Udrp = Iout*3900/(N+9) + Iout*1000 + Iout*1000 = Iout * R - - // Voltage drop in R4/R5 branch (for the G-APDs with correct resistor) - const double Udrp = R*Iout; - - // Nominal breakdown voltage with correction for temperature dependence - const double Ubd = fVoltGapd[i] + fTempOffset; - - // Current overvoltage (at a G-APD with the correct 3900 Ohm resistor) - const double Uov = U9-Udrp-Ubd>0 ? U9-Udrp-Ubd : 0; - - // Iout linear with U9 above Ubd - // - // Rx = (U9-Ubd)/Iout - // I' = (U9'-Ubd) / Rx - // Udrp' = R*I' - // Uov = U9' - Udrp' - Ubd - // Uov = overvoltage - // - // overvoltage = U9' - Udrp' - Ubd - // overvoltage = U9' - R*I' - Ubd - // overvoltage = U9' - R*((U9'-Ubd)/Rx) - Ubd - // overvoltage = U9' - U9'*R/Rx + Ubd*R/Rx - Ubd - // overvoltage = U9'*(1 - R/Rx) + Ubd*R/Rx - Ubd - // overvoltage - Ubd*R/Rx +Ubd = U9'*(1 - R/Rx) - // U9' = [ overvoltage - Ubd*R/Rx +Ubd ] / (1 - R/Rx) - // - - // The current through one G-APD is the sum divided by the number of G-APDs - // (assuming identical serial resistors) - double Iapd = Iout/N; - - // In this and the previosu case we neglect the resistance of the G-APDs, but we can make an - // assumption: The differential resistance depends more on the NSB than on the PDE, - // thus it is at least comparable for all G-APDs in the patch. In addition, although the - // G-APD with the 390Ohm serial resistor has the wrong voltage applied, this does not - // significantly influences the ohmic resistor or the G-APD because the differential - // resistor is large enough that the increase of the overvoltage does not dramatically - // increase the current flow as compared to the total current flow. - if (i==66 || i==191 || i==193) - Iapd = Iout/(N+9); // Iapd = R5*Iout/3900; - - // The differential resistance of the G-APD, i.e. the dependence of the - // current above the breakdown voltage, is given by - //const double Rapd = Uov/Iapd; - // This allows us to estimate the current Iov at the overvoltage we want to apply - //const double Iov = overvoltage/Rapd; - - // Estimate set point for over-voltage (voltage drop at the target point) - //const double Uset = Ubd + overvoltage + R*Iov*N; - const double Uset = Uov<0.3 ? Ubd + overvoltage + Udrp : Ubd + overvoltage + Udrp*pow(overvoltage/Uov, 1.66); - - // Voltage set point - vec[i] = Uset; - - if (fDimBias.state()==BIAS::State::kVoltageOn && GetCurrentState()==Feedback::State::kInProgress && - fabs(Uov-overvoltage)>0.033) - cout << setprecision(4) << setw(3) << i << ": Uov=" << Uov << " Udrp=" << Udrp << " Iapd=" << Iapd*1e6 << endl; - - - // Calculate statistics only for channels with a valid calibration - //if (Uov>0) - { - const int g = hv.group(); - - med[g][num[g]] = Uov; - avg[g] += Uov; - num[g]++; - - if (Uovmax[g]) - max[g] = Uov; - - const double iapd = Iapd*1e6; // A --> uA - - data.I[i] = iapd; - data.Iavg += iapd; - data.Irms += iapd*iapd; - - data.Uov[i] = Uov; - - med[2][num[2]++] = iapd; - } + const int g = hv.group(); + + med[g][num[g]] = Uov; + avg[g] += Uov; + num[g]++; + + if (Uovmax[g]) + max[g] = Uov; + + const double iapd = Iapd*1e6; // A --> uA + + data.I[i] = iapd; + data.Iavg += iapd; + data.Irms += iapd*iapd; + + data.Uov[i] = Uov; + + med[2][num[2]++] = iapd; } } @@ -638,5 +638,5 @@ ostringstream msg; - msg << setprecision(4) << "Sending new absolute offset: dU(" << fTemp << "degC)=" << fTempOffset << "V, Unom=" << overvoltage << "V, Uov=" << (num[0]+num[1]>0?(avg[0]+avg[1])/(num[0]+num[1]):0); + msg << setprecision(4) << "Sending new absolute offset: dU(" << fTemp << "degC)=" << fTempOffsetAvg << "V+-" << fTempOffsetRms << ", Unom=" << overvoltage << "V, Uov=" << (num[0]+num[1]>0?(avg[0]+avg[1])/(num[0]+num[1]):0); Info(msg); } @@ -647,5 +647,5 @@ { ostringstream msg; - msg << setprecision(4) << "Current status: dU(" << fTemp << "degC)=" << fTempOffset << "V, Unom=" << overvoltage << "V, Uov=" << (num[0]+num[1]>0?(avg[0]+avg[1])/(num[0]+num[1]):0); + msg << setprecision(4) << "Current status: dU(" << fTemp << "degC)=" << fTempOffsetAvg << "V+-" << fTempOffsetRms << ", Unom=" << overvoltage << "V, Uov=" << (num[0]+num[1]>0?(avg[0]+avg[1])/(num[0]+num[1]):0); Info(msg); } @@ -687,11 +687,12 @@ data.Iavg /= num[2]; data.Irms /= num[2]; + data.Irms -= data.Iavg*data.Iavg; data.N = num[2]; - data.Irms = sqrt(data.Irms-data.Iavg*data.Iavg); + data.Irms = data.Irms<0 ? 0: sqrt(data.Irms); sort(med[2].data(), med[2].data()+num[2]); - data.Imed = num[2]%2 ? (med[2][num[2]/2-1]+med[2][num[2]/2])/2 : med[2][num[2]/2]; + data.Imed = num[2]%2 ? med[2][num[2]/2] : (med[2][num[2]/2-1]+med[2][num[2]/2])/2; for (int i=0; i("num-calib-ignore"); fNumCalibRequests = conf.Get("num-calib-average"); + fTempCoefficient = conf.Get("temp-coefficient"); return -1; @@ -1053,4 +1041,5 @@ ("num-calib-ignore", var(30), "Number of current requests to be ignored before averaging") ("num-calib-average", var(300), "Number of current requests to be averaged") + ("temp-coefficient", var()->required(), "Temp. coefficient [V/K]") ;