Muon Offline Meeting -- Minutes ================================= 4 - Sept - 2002 ---------------------------------------------------------------------- Michael Gold -- CMU/CMP Data and Simulation Comparisons ======================================= Michael used J/psi and W samples for his studies, which will be reported in detail in an upcoming cdf note. In his plots, the red points represent data, while the blue solid line represents the simulation. Generally they are normalized to the same number of entries. Michael applied a pT requirement on the muons, and required both CMU and CMP. For the W sample it may be necessary to reject cosmic rays more thoroughly. 1) The stub chi-squared distribution does not look good in the data. This probably reflects the arbitrary and certainly incorrect values for the sigma's used in the reconstruction. 2) Matching cuts: Michael has achieved a good simulation of dZ. For dX there still are tails and the data are wider than the simulation, which may be a reflection of the less than perfect alignment available at the time these files were reprocessed. Michael showed a nice check of the pT dependence. 3) Michael showed the wire hit efficiency as a function of the drift distance. Data is, of course, lower than the simulation. It is not clear why the simulation is less than 100% efficient, since no inefficiencies are simulated at this point. There also is a mysterious dip in CMU hit efficiency at low drift distance. This needs to be investigated further. 4) The distribution of residuals is not flat as a function of drift distance. These are absolute, not normalized, residuals. The data are about 20% higher than the simulation on average. There is some very puzzling oscillatory behavior at the extremes of a CMU wedge. To make progress, these data clearly need to be reprocessed to pick up the final alignment constants for CMU and CMP. This is a group priority. ---------------------------------------------------------------------- Victoria Martin -- CMX Alignment ============= Victoria used the Feb-June high-pT lepton sample to make a first pass at the CMX alignment. For the "final" alignment, she has added in the new data taken since June (about 50% more luminosity). She allows for global x and y displacements, and for a local x displacement, which is equivalent to a phi rotation. Each of the four arches was treated as a unit. She reprocessed the CMX data in order to make use of David's fix to the east channel maps. In all cases she obtained reasonable fits, though there are features which hint at imperfections in the mechanical assembly of the wedges. Phil Schlabach reminded us of the struggles years ago to put the arches in the correct configuration on the blue frames. Victoria's results are very encouraging. Using dX (fitted to a sum of two Gaussians) as a measure of success, she finds for the core resolution (sigma from the narrower Gaussian): before after West arches 2.4cm 1.22cm East arches 1.5cm 1.22cm Victoria's constants are ready. She and David will work out exactly how to modify the code, and reposit said code by the end of the week if possible. [It was actually done on Saturday morning.] Victoria is working on a CDF note which describes all of the alignment work she has done. ---------------------------------------------------------------------- Dave Dagenhart -- CMX Drift Velocities ==================== Dave used the CMU-CMX J/psi sample. The mean of the dZ distribution is sensitive to the drift velocity, and be used to tune it. He took data from just two days (17-19 August), from a time after the alcohol temperature had been changed and the HV had been stabilized. He was using the jbmu03 data set, and reran with 4.7.1 and the Mextrapolator. Dave showed a very clean J/psi peak with about 2000 entries. He checked the T0's and found a very sharp edge for those hits assigned to stubs. He confirmed the 105ns value currently in use. Taking the nominal value of 44 um/ns, he found the west dZ offset by -8.9 +/- 0.5cm, and the east by 16 +/- 0.5cm. The sigma of the distribution is 10.3 +/- 0.4cm. He quickly and efficiently iterated the drift velocity to put the mean dZ at zero for east and west separately. As explained by Phil and Abraham Gallas, there are separate alcohol baths for the east and west side, which could differ in the actual temperature by up to a centigrade degree. Consequently, one should expect distinct drift velocities for the east and west sides. Dave checked the final result and found dZ distributions which are well centered at zero. These values will be used for all data taken in August and later. He will return to the spring data once he has a list of valid wedges for each run. (Michael Schmitt is supposed to provide that.) ---------------------------------------------------------------------- Ken Bloom -- Z->mumu with Stubless Muons =========================== Ken reminded us that the ICHEP results were based on pairs of CMUP muons. The CMUP requirement greatly reduces the acceptance, so he looked into the possibility of a Z sample with one CMUP muon and one stubless muon (which may sometimes have a stub in fact). He presented a comparison of the eta and phi distributions for the second muon, including the several cases of muons with stubs in all possible subsystems including CMX and BMU. This provides interesting tests of the geometry, and no great problems were spotted. With the nominal CMUP-CMUP analysis, there were 57 candidates in the Z mass window. In this new version, there are 340! Ken made a prediction of 330 events based on the increased acceptance. He also estimated the QCD background and found 20 +/- 24 events, which is not very different from the (6+/-5)% background estimated for the ICHEP analysis. He found zero like-sign candidates. He concludes that the geometry description looks good at this level. Also, the matching distributions are convincing. It would appear that this sample is quite valid for physics, though of course systematics have yet to be fully evaluated. ---------------------------------------------------------------------- Hyunsoo Kim -- Comparison of Muon Isolation Run II vs Run I. ============================================ Hyunsoo showed a direct comparison of the isolation variable for Run II and Run I data. This variable is the sum of calorimeter energies in a cone of 0.4 around the muon, with the energy assigned to the muon removed. There is a clear threshold at 100 MeV which comes from the fact that we impose a 100 MeV threshold on the towers entering the sum. This may need review. Also, the Run II data have a much larger spike at zero, but this may come from cosmic rays. ---------------------------------------------------------------------- M.Schmitt / Northwestern