Muon Offline Meeting -- Minutes ================================= 3 - April - 2002 ---------------------------------------------------------------------- Ken Bloom: Announcement ============ Ken showed a plot of pT vs dX and challenged us to guess its origin. The answer is: the BMU! Camille Ginsburg has used an earlier version of the BMU code to make this plot: dX is very narrow for pT above a few GeV. She will try to remake the plot using the latest versions of the code. Note these muons are volunteers: we have no BMU trigger in operation... ---------------------------------------------------------------------- Anyes Taffard: W/Z Samples =========== Anyes has rewritten Petra Merkel's filter, adding talk-to's with values which allow for one `standard' muon and a loose one or MET. The official CR Tagger is employed, of course. Stubless muons are allowed, even for W's, so that we can study the acceptances and efficiencies for the standard cuts. As per Henry's suggestion, she calculates the MET three ways, and a W candidate is selected on the basis of any of these results passing cuts. Anyes has checked her code with simulated samples and real data samples created by Alexei and Petra. She rejects a few events in each case, for legitimate reasons. The W and Z Stripper module will tag CR's but not remove them from the output. People felt the dX cuts need discussion and justification. This point will be settled over the next few days: anyone with an opinion, advice, or a study should contact Michael. Anyes showed a larger Z peak based on 14 pb-1 of data. Alexei needs to modify his scripts to use the new Stripper modules. Christoph should be ready to release the new code for Anyes. But first she needs to add in BMU. ---------------------------------------------------------------------- Anyes Taffard -- CR Tagger Update ================ Anyes has been working on improving the cosmic ray tagger. She has implemented a better calibration for the hadron TDC's, and has started to use the ToF information. The evidence from the ToF already shows an excellent performance of this module. Note she has recently released some examples and a little bit of documentation. Anyes has been working on the 1-legged case. She found that the OppMuon method (in which the 3-vector for the single leg is inverted and extrapolated) often fails to find the correct calorimeter tower, so she is looking in a wider cone around that extrapolation. She is also planning to use stubless muons and the ToF. There was some question from Alexei about a 2.6% loss of Z->mumu from her tagger. [Soon after the meeting, Anyes and Alexei worked on this and found the result to be fallacious due to a problem with the puffer module. The loss of Z's is around 5/10,000.] ---------------------------------------------------------------------- Eric James -- Trigger Studies =============== 1. Trigger Status. The system is functioning pretty well: the trigger simulation agrees well with the actual events. (Note: the `simulation' is really an emulation -- the raw data are fed into code which emulates what the hardware does and comes up with an accept/reject decision.) CMU & CMP look good. There is some structure for CMX. BMU looks very good. Note that the BMU muons are volunteers. Jonathan Lewis is reluctant to put the BMU into the trigger because he anticipates high rates. Camille Ginsburg has been trying to convince him to try it. 2. Scintillator Status. Work needs to proceed on the scintillators, as people expect they will be needed to control trigger rates, especially for CMX. A lot of progress has been made on the boards, FPGA's, and timing calculations. Dan Cyr has studied the BSU timing data and this aids the timing studies. The goal is to have some hardware configurations for testing at a later date, ie., to be prepared for when scintillators are required in the trigger. 3. Trigger Efficiencies. Eric worked with a sample of pre-shutdown data, taken with an 8 GeV pT threshold. This sample contains volunteers, which allowed him to observe the threshold at 8 GeV. The CMU curve plateaus at about 90%, but since he does not have good control over the purity of his sample (due to lack of statistics), this can only be a lower bound. He found that the CMP trigger was generally 100% efficient, with some problems at the corners which has has fixed. Eric has also used a sample of high-pT electrons to obtain a sample of 471 unbiased muons. With these he can examine the several steps to form a trigger decision. He can see the turn-on around 4 GeV, and the plateau efficiency appears to be about 85%. He needs to run on more data as the statistical uncertainties are large. ---------------------------------------------------------------------- Andreas Korn -- the G3X Problem =============== Andreas has employed Slava's fixes to the geometry, and performed more tests with multiple scattering turned off. He explained that the earlier problems were caused by an inconsistent implementation of delta-ray simulation. The details are given in the GEANT manual, but there are several options and it is easy to get it wrong. The crux of the matter is the connection between the upper limit in the energy loss simulation and the starting point for simulation of actual delta rays. Andreas shows that with the full simulation, everything comes out OK. G3X is, once again, ready for 4.5.0. ---------------------------------------------------------------------- Slava Krutelyov -- More on G3X =========== Slava made some interesting comparisons of G3X and MExtrapolator. He shows that at low momenta, G3X is slightly more efficient than MExtrapolator (at the level of a couple percent). The physical thresholds vs pT are very clear in this comparison. However, Slava discovered that there will be significant losses if a track is dropped because G3X sets the `stop' flag, indicating that the candidate muon has ranged out. In the current release, the track will not be dropped, but we clearly have to revisit this issue later and decide how to handle it. ---------------------------------------------------------------------- M.Schmitt