LPC Muon Meeting - Minutes ========================== 24 - October - 2006 http://indico.cern.ch/conferenceDisplay.py?confId=7692 ----------------------------------------------------------------- Claudio Campagnari - Update on Si Track - mu Segment Matching (UCSB) ======================================== Claudio presented an update on an algorithm which starts with tracks found in the silicon tracker and propagates them to the muon chambers to pick up muon segments. They have made significant progress on the algorithm, which he discussed in some detail. The main point of interest is the arbitration between two silicon tracks both of which are consistent with a given muon segment. The UCSB team has implemented quite simple ideas for this arbitration based on the separation in X and Y in the muon systems. This works remarkably well for a first try, reducing wrong associations by a large factor (depending on pT) with very small losses in efficiency. This performance was based on simulated ttbar events, and the muons coming from on-shell W decays and those inside b-jets were investigated separately. They foresee further refinements, for example, penalizing a candidate muon if a segment does not exist where one is expected. They plan to implement a Producer, and are thinking about how to expand its functionality. In the subsequent discussion, it was suggested to try to use other information to perfect the arbitration procedure, such as the angle between the extrapolated track and the muon segment(s), and the calorimeter energy associated with the extrapolation. People wanted to know how low in pT could one go with this approach. ---------------------------------------------------------------------- Andy Kubik -- Update on CSC Timing Studies (NWU) ============================ Andy presented progress in his studies of the timing as measured with the signals on the strips. On the one hand, he observes some dramatic non-uniformities throughout the MTCC detector. On the other, he is pushing toward a precise relative time measurement so that he can distinguish the left-going from the right-going muons. It appears this will be possible. Discussion: He should check with Stan Durkin about the best function to use to describe the signal shape (Andy presently is using a Gamma function). Also, the question was raised whether it is really a problem if the second time bin does contain some signal. (Andy does observe a large number of channels in which this is the case.) ---------------------------------------------------------------------- Ingo Bloch - Segment Finding Update and ROC MTCC Monitoring (FNAL) ============================================== Ingo has investigated the improvements to the performance of the segment reconstruction in CMSSW_1_1_0_pre3. He finds little impact for the geometry fix and for the new wire digis, which is not a bad thing. He has begun to study overlaps and extrapolations with the DT's, and ultimately, the HCAL. He has discovered a puzzling variation of the error on phi as a function of the local Y coordinate. There is no good hypothesis for this effect. Concerning the ROC monitoring work, this is up and running since a while. Ingo reminded us of the excellent web-based tools for accessing information. He highlighted one particular problem with ME+2/2/32 in which a set of strips appears not to be working correctly. He has also tried to use the RecHitB package from Dominique Fortin, which some minor difficulties which he expects to resolve quickly. ---------------------------------------------------------------------- Stoyan Stoynev - Studies of CSC's with Muon Passing through Two Stations (NWU) ======================================================= Stoyan reminded us of how he selects MTCC events in which one muon passes through both stations 2 and 3. He demands a good muon segment in station 2 and extrapolates it to station 3, and then studies what happens in station 3. He finds that the number of hits on a segment is quite high, essentially six. He showed measurements of the hit efficiencies in each layer of each chamber in station 3, and noted a humped shape to the efficiency curve. He measured offsets in the coordinate measured by the strips. This offset varies in a peculiar way across the chamber in what he calls an "S-curve." He showed that the overall offset is present when he restricts his sample to muons with near-normal incidence, but the curviness of the S-shape goes away. He then proceeded to study the shape of the curve as a function of the angle of the muon, but needs a larger event sample to obtain clear results. Discussion: The humped shape of the efficiency curve is likely due to a trigger bias in reading out the chamber. There is also a question of geometric effects but opinions on this differ. Tim Cox explained that the fix to the strip geometry implemented by Jordan Tucker is *not* a change in the width of the strips, but rather a change in the way different layers are staggered. This opens up the question of what might be producing the S-shapes that Stoyan sees. ---------------------------------------------------------------------- Alex Tumanov -- h -> 4mu Reco Studies (Rice) ===================== Alex has pushed forward his studies of the Higgs 4-muon signal. He selects four reconstructed muons, and compared "loose" and "tight" mass cuts. The events gained by loosening the mass cuts do not improve the signal, and seem to produce tails in the mass distributions which need to be understood. He found the efficiency for muon reconstruction is too low compared to results other people have obtained with the muon gun. Discussion: Alex should separate the impact of geometric acceptance from reconstruction efficiency by considering only those generated muons with |eta| < 2.4. Also, his distribution of N muons per event show events with many muons, and this is surprising for such a clean signal. This needs to be studied. ---------------------------------------------------------------------- Jean-Roch Vlimant - Muon Seeded Tracking (UCSB) ==================== Jean-Roch presented an update of their studies on muon-seeded track reconstruction. The idea is to start with muon segments, swim them toward the interaction point, apply a beam constraint, and then look for hits inside a road of optimal width. There are several elements to the study, including the way that errors are estimated while extrapolating inward and then outward, and the distribution of hits found. They had observed terrible errors from the muons reconstructed with 0_9_2, but using 1_1_0pre3, their results are much more sensible. They observed pull distributions that are a little wide, by 20 - 30%. They note that adding the beam constraint to the procedure for setting the road substantially reduces the width of the road, and thus leads to better results. Discussion: Jay Hauser noted that the hit distributions for the "reconstructed" sets have far fewer entries than for the "perfect" sets. He asked whether there is a difference in the pull for hits found with the perfect muon when one requires that the muon was also reconstructed. ---------------------------------------------------------------------- submitted by M.S.