Conclusions and Questions: Did we really load new ALCT firmware? No difference in results for the 0 degree angle of incidence. When nph_pattern was changed only for ME2/2 and ME3/2, why did the ghost rate drop >10X for ME1/2 and ME1/1????? Meanwhile, chambers ME2/2 and ME3/2 stayed the same. Constants loaded for all chambers? The efficiency did drop for ME2/2 and ME3/2 (96% -> 86%) indicating that more layers were required in coincidence. For the 28 degree angle of incidence, the ghost probability is high for trig_modes = 0, 3 (collision and accelerator patterns on) This confirms earlier study done during 9/03 beam test, where 40-45% probability was found for trig_mode=3. We see the same this year, but not in ME1/2 or ME1/1 (Smaller wire group spacing?). Collision patterns fire efficiently even for 0 degree incidence, even for 6/6 pattern (85%). Accelerator patterns fire efficiently even for 28 degree incidence, even for 6/6 patterns (87%). Thus, we don't have much discrimination between the two types of muon trajectories, which limits how useful the accelerator muon patterns are. For 28 degree incidence, the ghost rate is lower (20-30%) for trig_mode = 0 (accelerator or collision patterns allowed), than for trig_mode = 3 (accelerator kills collision pattern) where it is 40-50%, for chambers ME2/2 and ME3/2. However, the situation is opposite for ME1/2 and ME1/1, where the ghost rate is 60-75% for trig_mode=0 and 5-7% for trig_mode = 3. Why? Why is the efficiency lower for trig_mode=0 (collision or accelerator patterns) than for trig_mode = 2 (collision only)? The lowest ghost rate configuration seems to be for trig_mode=2 (collision patterns only). For 28 degree incidence, it is 5% for 4 layer coincidence and 3% for 5 layers. Corresponding efficiency is 99.2% and 97% respectively, for ME2/2. I would recommend staying with our defaults: trig_mode = 2 (collision patterns only) nph_pattern = 4 (4 layer coincidence) To get the best efficiency and lowest ghost rate. Regards, Darin