Minutes of CALCOM Analysis meeting - August 22, 1997. ===================================================== Agenda: N. Pivnyuk - Straight track analysis for zero B-field runs V. Burkert - Luminosity and trigger studies S. Taylor - Update on TOF timing analysis for adjacent scintillators B. Niczyporuk - Drift chamber efficiency analysis ====================================================================== Individual reports: =================== Nikolai Pivnyuk: --------------- Hit Based Reconstruction for all sectors was performed for Run 3786. Goal : search for sources of background outside of the target area. ---- Input data : Run 3786 (Big Torus OFF, Mini Torus 1,000 amps, Empty target, ---------- Electron beam, Eo = 4 Gev ). Procedure : track reconstruction in the middle plane. Superlayers 3 and 5 were --------- taken into account only. Summary : Target position was reconstructed with accuracy ~ sigma = 2-3 cm. ------- The background was mostly due to the random, not related to one track segments in the region 2 and 3. No peculiar background sources were found. -------------------- Comment by V.Burkert: Although no isolated source of background was found, there appears to be an enhanced production of tracks coming from an extended area downstream of the target. This could be due to secondary interaction of electrons scattered in the target at small angles and their subsequent interaction with the mini-torus support/shielding structure. A possible remedy for this could be additional lead shielding around the beam pipe through the mini-torus. -------------------- V. Burkert: Luminosity and trigger studies ---------- 1. Runs 3736 - 3750 (4.045 GeV) were analyzed using CED to study the hit occupancies for isolated hits and for track related hits versus the beam current. The runs were chosen because they were taken with a variety of beam currents, and a minimum bias trigger was used. Currents of 1.5, 5.0, 7.5 and 10 nA were used corresponding to luminosities from (1.5 - 10.6) x 10^33 cm-2sec-1. By averaging the total number of hits for about 70 events, and subtracting the hits for events were no track segments were observed in region 2 or region 3, track related and non-track related occupancies were determined. While the track related occupancy was consistent with a constant behavior, independent of luminosity for all but the highest current, the non-track related occupancy showed a linear rise for all but the highest current. The first result (except for the 10nA run) is expected if there are no (or few) multiple tracks in the data, while the second result is consistent with accidental hits. The strong deviation of the 10nA data can indicate that the beam conditions changed and a higher than assumed luminosity was effective during data taking. The fact that up to about 8x10^33cm-2sec-1 (7.5nA) the number of track related hits is approximately constant indicates that tracks from multiple events should in general not be a problem for the data analysis at these luminosities. The study should be repeated with higher statistics. 2. The effectiveness of the calorimeter-based total energy trigger was studied using CED. A calorimeter energy threshold of about 1.5 GeV was used. Events were sorted into four categories: 1 - negative track pointing at the energy cluster in EC (electron candidate) 2 - neutral energy cluster at very forward angles (background hit mostly due to small angle electrons interacting in magnet support structures) 3 - neutral cluster(s) at large angles (pi-zero,..) 4 - positive track pointing to EC hit The breakdown was: 1:2:3:4 = 32% : 35% : 22$ % :11% with a few % uncertainty. The results indicate that the calorimeter trigger at this 4 GeV run selects electrons in about 30% of all events (#1). The situation may be improved significantly once the additional shielding is installed that should eliminate most of the small angle "neutral energy" triggers (#2). This shielding is planned to be installed before the upcoming commissioning run in September. See also Cole Smith' presentation at the CALCOM meeting on August 8 for the GSIM simulations that were used to design the additional shielding. Simon Taylor: Counter-to-counter TOF time match ------------ Six separate cosmic ray runs were taken triggering on each of the six sectors of TOF counters. These runs were filtered for those events that have adjacent counters firing. We looked at the the time difference between these adjacent counters for crossing tracks with the following cuts: (1) the ratio of the energies of the hits in the pairs of counters was between 0.2 and 5; (2) the magnitude of the position difference between the hits was less than 20 cm (3) no tdc zero's or overflows (4) positions identically equal to 0. were thrown out to try to eliminate a prominent spike at zero. Time walk corrections and left-right time matches for individual counters were applied. The measured time differences for each pair of counters in the forward carriage were all within a range of rougly +/-2.5 ns with some exceptions, most notably the difference between counters 22 and 23 in each sector, which can be explained by differences in construction between paddles 22 and 23. The resolutions as measured by the sigmas of the timing distributions varied between 200 ps and 1 ns depending on sector, with sectors 2 and 3 having the best timing resolution and sectors 1 and 4 having the worst, often due to multiple peaks in the time distributions. The timing differences have been entered into the mapmanager database for future use. Bogdan Niczyporuk "On Efficiencies and Resolutions of Drift Chambers" ----------------- Some preliminary results on DC performance are shown in Table I. The results were obtained exclusively from the analysis of Run#3166 (Sector 6). (1) First six rows in Table I show the average layer efficiency due to the following reasons: - dead channel: no TDS's or disabled (the biggest loss), or/and - drift time out of limits (see Tmax in Raw#7 of Table I), or/and - layer has been skipped in the LR-ambiguity procedure (the layer which contributes mostly to the unaccepted chi-square value) (2) The average superlayer resolutions are shown in Raw#9. Note that the resolution is approximately proportional to the cell size of superlayer. (3) The percentage of unresolved LR-ambiguities in the superlayers are shown in Raw#10. The big losses in SL1 are due to: (i) 4 layers only, and (ii) likely bigger background in that superlayer. Additionally, for tracks, with successfully resolved the LR-ambiguity in all 6 superlayers, the loss due to the overall chi-square cut is about 2 percent. Table I. Compilation of Results -------------------------------- |---|-----------||---------------------------------------------------------| |Col| Item || SL1 SL2 SL3 SL4 SL5 SL6 | |===|===========||=========================================================| | 1 | LA1 || 93.6 74.3 90.0 96.8 93.5 79.8 | | 2 | LA2 || 95.9 96.5 94.4 97.8 93.7 77.3 | | 3 | LA3 || 90.7 97.0 88.5 97.3 88.7 88.5 | | 4 | LA4 || 84.0 62.1 89.3 97.0 90.4 86.7 | | 5 | LA5 || NA 84.1 94.1 96.4 91.6 88.4 | | 6 | LA6 || NA 97.2 83.4 96.2 89.3 93.4 | |---|-----------||---------------------------------------------------------| | 7 |Tmax[ns] || 310.0 310.0 700.0 700.0 1900.0 2100.0 | |---|-----------||---------------------------------------------------------| | 8 |Dmax[cm] || 0.696 0.693 1.156 1.253 1.790 1.929 | |---|-----------||---------------------------------------------------------| | 9 |SIG/SL[cm] || 0.040 0.050 0.060 0.075 0.120 0.145 | |---|-----------||---------------------------------------------------------| |10 |LRfailed[%]|| 8.9 1.2 0.5 0.4 2.5 3.5 | |===|===========||=========================================================| The results compiled in Columns: 1-9 may be used in the simulation to estimate the track reconstruction efficiency.