Minutes of the HEC/FCAL Testbeam Analysis Meeting from 9.12.99. ----------------------------------------------------------------------- 1. Testbeam Running 2000 (D. Sauvage) Dominique presented the run schedule for the year 2000 and the related constraints from the cryogenics. A larger LAr dewar should be foreseen if all three cryostats have to be operated frequently in parallel. Presently the HEC runs are scheduled for 30.5.00 - 14.6.00 and for 31.7.00 - 16.8.00. The second run may be extended by a few days. 2. The hec_adc code: Status and Planning (M. Lefebvre) Michel presented the status of the hec_adc version 3.8. The monitoring has been improved. The understanding of the TDC timing is substantially better (see note HEC-083). Digital filtering weights from the Mainz group and from Pavol have been implemented. Further improvements with respect to timing/calibration/weights are in progress. Michel also presented some results from Dominique Fortin's analysis. It is found that the pedestal value for each cell depends on the trigger type used (random in physics run, physics in physics run or events in noise run). The difference can be more than 4 adc counts. To calculate pedestals, one should therefore use events of the trigger type under study. Dominique also studied the Aug99 electron resolution for cubic fit and digital filtering methods. The linearity is improved if the response vs energy is allowed to have an offset. There is evidence that the electronics noise may depend on time. 3. Performance of Testbeam Electronics in 99 (L. Kurchaninov) Leonid discussed in detail the operation of the full electronic chain. The calibration signal deviates from a pure exponential shape due to a inductance in the chain and - at small signal level - due to the clock feed thru. The stripline connector adds a parasitic capacitance (1-2 pF), yielding a change of the signal shape at the few % level. The integration time of the preshapers is somewhat larger than designed, the ratio of rear to front gain is 1.9 rather than 2. Some bad hybrids have been fixed. Bad shapers have been replaced, two different types of line drivers have been used, yielding a difference in rise time of typically 2 ns. Few noisy channels have been identified, the noise could not be reproduced in the lab. It might be related to the specific cabling (patchpanels) at CERN. 4. Status of the Calibration 99 (P. Strizenec/L. Kurchaninov) Leonid presented the latest results. Two approaches have been studied: a) Modeling in detail the full electronic chain, 3 parameters (gain, preshaper time constant, shaper time constant) are fitted from the calibration pulse. b) the 98 method: a multipole function is used to describe the pulse. Eventhough the second method gives on average a better description of the calibration signal (0.5%-1% rather than 0.5%-1.5%) the unfolding to the real pulse yields an error on the amplitude which is comparable (0.5%-1%). To understand all details of the pulse shape, a special technical run is planned for next february/march. 5. Electron and Pion Data: Analysis of August 99 data (A.Minaenko) Andrej Kiryunin presented the results from Andrej. Based on the analysis of the electron and pion data, the reconstructed amplitude (using OF weights) is underestimated by typically 0.7%. The corresponding time shifts are rather small, typically (0.6,-0.6,3.1,1.2) ns for layer (1,2,3,4). Correcting each layer by this offset, all channels stay within +- 1 ns. The noise within a cluster is not constant for all runs, indicating the problems with few channels. The energy resolution for electrons is similar to the 98 results, 22% for the sampling term, but with a slightly increased constant term (0.2-0.7%). The spatial uniformity is affected by the few bad channels. The calibration constant nA -> GeV differs from last year, but the reason (calibration signal measurement) is understood. For the pion data the energy resolution is comparable to the 98 data: 75% for the sampling term and 5.5% for the constant term. As for the electrons, the spatial homogeneity is affected most by the few bad channels. 6. Muon Data: Results from 99 and Update of 98 Results (M. Levitsky) Andrej Kiryunin presented the preliminary results from Mikhail. The horizontal and vertical scans show the varying response due to pad structure, cracks and tierods. The deviation of the absolute signal from last year is related to the calibration (see above). Using the new set of OF weights for the 98 data, the analysis of last year data has been repeated. The signal change is 4% (8%) for the april (august) data. 7. Status of Analysis at Dubna (V. Kukhtin) Victor presented the status of the analysis of the 99 data in Dubna. Electron data have been studied either at the ADC level or using the cubic fit for the amplitude reconstruction. For the electron resolution a sampling term of typically 20.5% - 21 % is obtained. 8. Simulations for the HEC Testbeam Data 99 (A. Kiryunin) Andrej presented his results concerning the simulation of the testbeam data. Muons, electrons and pions have been generated at various impact points and covering the whole energy range of the testbeam data. The total energy leakage for pions is at the level of 3%-4%, depending on the impact point selected. The sampling term of the pion energy resolution is typically 60%, the constant term in the range 4.6%-5.3%, depending slightly on the impact point. The sampling term for the energy resolution of electrons ist typically 21%. 9. First Results from the Simulation for the Combined Test. (A. Kiryunin, D. Salihagic) Andrej presented the first results of the MC for the combined run. The total leakage for pions is at the level of about 4.5% in comparison to about 3.5% for the standard HEC testbeam setup. The energy resolution obtained for pions is very close to the resolution obtained with the ATLAS simulation. The second option of the HEC2 setup - as discussed in the INTAS proposal - will be studied soon. 10. Status and Plans for the GEANT 4 Testbeam data Simulation (D. Salihagic) Denis presented the first results of the testbeam data simulation within the framework of GEANT 4. Muons and electrons in the whole energy range of the testbeam data have been studied. For GEANT 4 the range parameter has been varied from 0.5 mm to 4 mm. The muon response does not show a strong dependence on this parameter and shows good agreement with the GEANT 3 result and the testbeam data. This is in contrast to the electron data, where the response varies by up to 3% within this range. Except for a range of 4 mm, the GEANT 4 simulation yields a larger response than the GEANT 3 simulation. The energy resolution for electrons is in GEANT 4 bettter than in GEANT 3, for all range parameters. On the other hand, the testbeam data agree rather well with the GEANT 3 simulation. 11. HEC Simulation with GEANT 4 (R. Mazini) Rachid presented the first results of the GEANT 4 simulation for the HEC. The geometry is very close to the real HEC, the tie rods are not yet implemented. The range cuts used correspond to 0.1 MeV for Cu (used in GEANT 3). In this case the GEANT 4 response for electrons is about 5% below the corresponding GEANT 3 (60 GeV) response. The sigma of the energy resolution shows at 60 GeV good agreement between GEANT 4 and GEANT 3. 12. NIM paper Discussion (All) In the discussion it became evident, that there is an agreement to continue to finish the NIM paper based on the testbeam data 98. In vie of the better understanding of the timing/calibration new calibration constants have been provided and the hec_adc software has been modified to be compatible with 98 data. The idea is, to concentrate on few results (response muons, response electrons, response hadrons) to understand how sensitive the previous numbers are with respect to the modifications. Presumable energy resolution or spatial homogeneity are less affected. For a few selected runs this should be proven first.