1. Effects of Endcap Staging/Descoping D.Acosta University of Florida

2. CMS Week, September 2001Darin Acosta2 Effects of Endcap Staging/Descoping Staged Muon scenarios: 1. No ME 4/2 (CSC) è No fourth CSC station for  < 1.8 2. No RPC for stations 2,3,4 è No endcap RPC trigger, but station 1 there to cancel CSC ghosts if needed 3. ME1/1 strips put into “OR” è Reduce channel cost by not segmenting ME1/1 in 

3. CMS Week, September 2001Darin Acosta3 ME4/2 Staging With RPC trigger in place, CSC trigger runs in “loose” mode for high efficiency: è Only 2 CSC stations are required: one must be ME1 and one must be ME2 or ME3 (relax ME1 condition for DT/CSC overlap) è In principle, ME4 cannot affect efficiency or rate by design è But P T assignment is slightly better for 3-station tracks than for 2-station tracks, so might expect some changes to CSC rate and efficiency above certain P T threshold without ME4 è GMT requires match between CSC and RPC for low quality CSC muons, and optimizes the P T assignment, so any effect is reduced after GMT Without RPC trigger in place, the CSC trigger must run in “tight” mode for improved rate reduction è Three CSC stations are required, including ME1 è No redundancy without ME4 means efficiency loss

4. CMS Week, September 2001Darin Acosta4 B.Scurlock ME4/2 Staging – Efficiency “tight”, ME4/2 out,  =73% “tight” ME4/2 in,  =79% “loose” (unchanged)  = 97%  ME4/1 in/out Single muons: 3< P T <100 GeV

5. CMS Week, September 2001Darin Acosta5 ME4/2 Staging – Rates GMT is unchanged CSC loose CSC loose, no ME4/2 U.Gasparini L=2  10 33  5X CSC loose, no ME4/2 CSC loose CSC tight L=10 34 CSC tight without RPC achieves about same rate as CSC loose with RPC in GMT

6. CMS Week, September 2001Darin Acosta6 ME4/1 Needed at High Eta GMT as in ORCA 5.1.2 è rate at 20 GeV/c: 3.1 kHz è L1 efficiency (*) : 96.6 % re-tuned GMT selection: è Only three-station CSC tracks used without RPC confirmation è rate at 20 GeV/c: 1.4 kHz è L1 efficiency (*) : 96.3 % (*) efficiency to find muon of any p T in flat p T sample  Rate from unconfirmed 2-station CSC tracks, since no RPC coverage GMT single muon trigger rates (p T > 16 GeV/c) ORCA 5.1.2 re-tuned GMT H.Sakulin L=2  10 33

7. CMS Week, September 2001Darin Acosta7 Mixed Mode CSC Trigger è Without RPC trigger and without ME 4/2, still allow “loose” CSC trigger for 1.2 <  < 1.8 (but “tight” everywhere else) è Gains efficiency here è Still want ME 4/1 at high  è DT/CSC overlap still a problem Rate from “mixed-mode” close to that from “tight”  ~ 85% L=10 34 loose

8. CMS Week, September 2001Darin Acosta8 ME1/1 Staging Recall that ME1/1 has split strips è Motivation was to reduce occupancy è Can we avoid it to reduce channel cost? From muon TDR: è Charged particle occupancies in either half of ME1/1 are about 0.5% per chamber per BX è Neutron hit occupancies are no more than about 2%, but as they are uncorrelated through the chamber, a much smaller fraction actually give trigger primitives. p So these numbers are not large, and in fact are smaller by about 50% than the numbers for the other CSC chambers (ME2/1, ME3/1, and especially ME4/1) p In principle there should not be any problem if we can OR the strips from top to bottom of ME1/1 High Lumi: L=10 34

9. CMS Week, September 2001Darin Acosta9 CSC LCT Rates in ORCA5 J.Mumford High Lumi: L=10 34 min bias pile-up only no neutrons ME 1/1: 7 MHz ME 1/A: 10 MHz For 72 chambers of each type, occupancy per BX is: ME 1/A: 0.35% ME 1/1: 0.24%

10. CMS Week, September 2001Darin Acosta10 CSC LCT Occupancy with Neutrons Similar study performed with min bias pile-up and neutrons at L = 10 34 using ORCA5 Chamber LCT occupancies: Anode+CathodeCathode only è ME 1/A: 0.46%ME 1/A: 0.61% è ME 1/1: 0.44%ME 1/1: 1.50% è ME 1/2: 0.05%ME 1/2: 0.10% è ME 1/3: 0.05%ME 1/3: 0.14% è ME 2/1: 0.29% è ME 2/2: 0.34% è ME 3/1: 0.21% è ME 3/2: 0.25% è ME 4/1: 0.25%ME 4/1: 0.81% è ME 4/2: 0.86%ME 4/2: 2.54% B.Scurlock

11. CMS Week, September 2001Darin Acosta11 Discussion of ME 1/1 Strip “OR” è Combining the two halves of ME 1/1 can’t increase the trigger rate if the track segments come from real muons (i.e. correlated hits) è Extra trigger rate can only come from random combinations forming ghost track segments, which in turn form additional triggers è If the current LCT occupancies are dominated by real correlated hits, expect ME 1/1 and ME 1/A occupancies to add è If neutron-induced random hits dominate, expect occupancy to scale non-linearly p “Guestimate”: order of magnitude increase è How to tell without detailed simulation of strip OR? è Study dependence of occupancy with background level è ORCA4 study with pile-up and neutrons: Occupancy for L = 10 34 L = 3  10 34 ME 1/A:0.3%0.9% ME 1/1: 0.3%0.9% è Therefore, occupancies scale linearly

12. CMS Week, September 2001Darin Acosta12 Summary ME 4/2 staging è With RPC: No change in GMT rate or efficiency è Without RPC: Approximately no change in rate requiring 3 CSC stations, but 25% acceptance loss in endcaps for “tight” trigger (12% loss for mixed-mode trigger) è In any case, ME 4/1 is useful to reduce rate by factor of two if efficient triggering at  >2.1 is desired ME 1/1 Staging è At high luminosity, the ME 1/1 and ME 1/A track segment occupancies are about 0.5 – 1.0% per BX è Combining the two, the occupancies add and would be of order 1–2% è At low luminosity, occupancies would be 5X lower è Therefore, expect no change in efficiency or rate if strips in ME 1/1 and ME 1/A are put into an “OR” è Exactly how to do this is up to chamber designers p Different strip pitch in ME 1/1 and ME 1/A in current simulation è Perhaps it could be possible to segment HV so that high  wires can be turned off if occupancies turn out to be higher than expected (as is possible in other CSC chambers) p Complicated by wire tilt, however