Minutes of the CALCOM meeting, March 6, 1998. ============================================= Agenda: ------- B. Mecking - Cold CLAS Torus Coil Geometry (write-up in preparation) G. Gavalian - Level 2 trigger data analysis J. Price - Cherenkov counter calibration status J. Mueller - Luminosity dependence of tracking efficiency Short contributions by : V. Burkert, J. McNabb. B. Niczyporuk, (1-2 transparencies) L. Qin, R. Schumacher, S. Stepanyan ==> Also included: February 20 contribution by M. Mestayer ======================================================== The next meeting will be Friday March 13. ---- Agenda: ------- G. Mutchler - TOF geometry calibration C. Loukachine - Correction for dE/dx TOF calibration + short presentations (1-2 transparencies) ===> For longer presentations, please let me know by Thursday noon <=== Volker Burkert ================================================================================== Individual presentations: ------------------------- G. Gavalian reported on his study of the level 2 segment finder using the raw data information. He finds that the segment finder works in principle however, there are several sectors and regions which do not report TDC informmation. The efficiency for finding 2 track segments in level 2 for a RECSIS/SEB reconstructed tracks is currently about 20%. This number seems to be independent of the polar angle. A large portion of the "inefficiency" appears to be due to false timing. J. Mueller: Luminosity Dependence of Hit Based Tracking ----------- I ran on the runs for the 4 GeV luminosity scan (9138-9143) and produced plots for the HBT efficiency as a function of Beam current. (see February minutes for details of analysis). This was done for release-1-12 and also for current tracking code. The calculated efficiecies ranged from 94-98 percent for the old code and 96-98 for the new code, with error bars in the 2% range due to the limitted statistics in these runs. As always, I then hand scanned the failures to look for possible futher improvements to the code. The failures fell into two types: 1) Well reconstructed electrons were found by tracking, but the polar angle predicted from the proton was outside my cuts (usually theta_meas-theta_pred = -3 to -4 degrees, where I was cutting at -3). These I moved from the fail to pass catagory. 2) Events where a good positive track consistent with an elastic proton is found opposite a CC hit and an EC cluster with energy consistent with the missing energy recoiling off the proton, but there is no track visable by eye in that sector. Since there were no hits to reconstruct into a track, this shouldn't be considered a tracking failure, so it was removed from the demoninator. It is possible that these may represent a hardware failure, and that needs to be looked at further. After these changes from the hand scan the efficiency was recalulated and found to be >99% at all beam currents taken during the luminosity scan. The luminosity scans have been very useful for tuning the code to be able to deal with higher luminosity data. It would be very useful for further tests to aquire a sample of events with production triggers at design luminosity. The current luminosity scan only went to 4E33. We need some data at higher luminosities to uncover the problems we will run into when we run there and rectify them. Three 500K runs at 6, 8, and 10E33 are requested. Finding events with tracking failures are useful for improving the code. I stated that people who have such failures make a file containing only those events and contact me. I then spent many hours in meetings with people on saturday. I will revise my statement here. If you have events where you think tracking failed, please make a file containing only those events and run them through the latest release of tracking. If those events are still problematic send me e-mail with your estimate of the problem and the location of the file. J. Price: Status of Single Photoelectron calibration of Cherenkov counters --------- A drift in the phototube gains for the Cerenkov detector have been observed. Most of the drift is seen in sector 2; this is not yet understood. More frequent calibration of the Cerenkov PMT gains is necessary to ensure that we are using the proper values for the PMT gains. This currently requires a dedicated set of runs totalling approximately three hours. John Price and Mike Klusman are studying ways to reduce the amount of time needed for this purpose. Details of the Cerenkov PMT calibration may be found at , under "Calibration" and "Results". M. Mestayer: Drift Chamber Calibration Status (February 20, 1998) ----------- 1) Pulser Data Analysis: Liming Qin analysed pulser calibration data taken in December, 97 to obtain the FIXED DELAYS for each drift chamber signal. These calibration values are much improved over previous data, primarily because the pulser system was fixed for Region 1. There had been a mismatch between the pulser polarity and that expected by the Region 1 SIP groups. Setting a time window to eliminate pulser-generated noise also helped improve the results. 2) After installation of these "T0" constants, Liming looked at the drift time spectra for particle data. The leading edge of the time distributions of most wires are aligned to about 5 - 10 ns. 3) We are still using DOCA vs. Time tables developed by Franz Klein for September data. KyungSeong Joo, Rustum Nyadov and Alex Skabelin are working with Liming to improve the function and finally to come up with an automatic method for re-calibrating the function using data. Remarks from the audience: - Dennis Weygand reminded us that we need to weight drift chamber hits by a POSITION-DEPENDENT sigma in order to obtain optimum resolution. - Bernhard Mecking pointed out that an alternative (and better?) way to obtain the "T0"'s is to fit the leading edge of the time distributions for each wire (or each group of 8 wires). - Elton Smith asked how well we know the velocity of the signal propogation along the sense wire. He suggested that we determine this important TIMING parameter before we go further with distance versus time functional fitting. Short presentations: ==================== V. Burkert N* data for the study of detector geometry ---------- I showed preliminary data for inclusive single pi-0 and pi+ rates (not corrected for acceptances and rad. corrections). The comparision of the two channels shows very clearly the strong Delta excitation in the pi-0 channel with (almost) no non-resonant background. Higher resonaces in the N*(1520) and N*(1680) region are visible but small. The Delta in p-pi0 is so clean that it can be used as a check of how systematic uncertainties in the geometry (acceptance) affect the multipole analysis in the positive parity p-wave channel. The pi+n channel shows the expected reduction (from isospin) of the Delta, but also the strongly enhanced (also from isospin) N*(1520) and N*(1680) states. John McNabb ----------- Results from moving the magnetic field in Recsis: Run's 7889(target empty) & 7841(target full) BEAM Energy = 1.6 Gev Time Based Tracking (release 1-12) the following information was compiled from a study done by moving the magnetic field within RECSIS to see the effects on the tracking. All were done using the survey information for the drift chambers. Three positions of the magnetic field were 1) nominal, 2) -3cm in Z (upstream), -1 cm radialy outward. I have listed the results for the sum over all sectors, If your intrested in the information by sector it is available. The measurements are the mass(and width) of the elastic peak, the location of the Chi^2 peak for +/- charge. The total number of tracks found for +/- charge, the position of the vertex(Z) of tracks from the downstream wall for +/- charge , and the difference between the + and - results for the wall position. Note that both shifts resulted in a larger number of trakcs being reconstructed by about 13%. measurement charge Nominal 3cm upstream 1cm radially outward elastic peak 964 MeV 958 MeV 945 MeV sigma 22 MeV 23 MeV 26 MeV chi^2 peak + 3.0 2.2 2.0 - 1.99 2.1 2.1 # of tracks + 17360 19659 19710 - 19322 21701 21852 downstream wall + 2.5 cm 2.5 cm 2.4 cm - -.9 cm .2 cm -.4 cm Vertex Z-Difference 3.4 cm 2.3 cm 2.8 cm L. Qin: New DC time calibration ------- A new DC cable delay map has been put into CLAS_PARMS area. The improvements are from the better pulsing data and corrections on pulser propagation time on ADB and signal travel time on STB. With the current DC and SC time constants and x(t) function #3, the chi2 distributions for all six sectors peak at 1.0. The chi2 distributions and the sigma of residual distributions plots as a function of doca can be found at http://www.cebaf.gov/~kuhn/dcintro.html. They are for run 7867 and 8774. R. Schumacher: Mass Resolution as a Function of Momentum. ------------- We considered CLAS's hadron mass resolution for pions and protons computed from the measured times of flight and measured momenta. As a function of momentum one can see 1) the linear rise (worsening) of resolution with increasing momentum, due to the tracking resolution of the spectrometer, and 2) the roughly 1/p**2 drop (improvement) of resolution due to multiple scattering with increasing momentum. For a run using 60% of full field, using only survey data for chamber alignment, as well as a less-than-optimal time to distance relationship, we find: sigma at sigma at sigma at 250 MeV/c 600 MeV/c 1800 MeV/c A) For pions 20 MeV 40 MeV 180 MeV B) For protons 50 MeV 30 MeV 55 MeV These results are "averaged" over all angles, all possible paths through the B field, and all the differing amount of multiple scattering material. The significance of these results is that they help understand our ability to separate kaons from pions, and underline the need to improve the momentum resolution of the system. 2) the roughly 1/p**2 drop (improvement) of resolution due to multiple scattering with increasing momentum.