CSCTF: News

1. CSCTF: News By Anna Kropivnitskaya ● CSCTF DDU status and test request ● Firmware and PTLUTs changes ● New data format test, chain 1 22/11/11, see Matt’s presentation CSCTF working meeting, 24 January 2012

2. CSCTF DDU status and test request CSCs only have 5 DDUs in the world with the "TF type" modifications.  - DDU numbers 1-4 and now also 7 (this is at OSU with an extender card, will be sent to Cern): - DDU001 – is at UF without cutout of S-Link- DDU002 – is at 904 in peripheral crate  with cutout of S-Link- DDU004 – is at 904 at cosmic test stand without cutout of S-Link- DDU003 – is running at Point 5 as a consequence ✓ So we have one real spare DDU002. CSCTF DDU expert (Janos Gilmore) says that it is enough. ✓ Firmware status for CSCTF DDU: - at least now possible upload only at 904! - good to have such possibility at Point 5 too. - Jason would like to test new DDU firmware and DDU002 (stress test at Point 5) for this he needs swap DDU boards DDU003 and DDU002… - If CSCTF group has no objection then all CSCTF expert should be at Point 5 that Jason train us how to change S-Link in DDU Request S-Link for DDU002 from DAQ group

3. New data format test We start tested new firmware chain 1 22/11/11: Take 1st local run -> it seems good, need more test For details see Matt’s presentation

4. Firmware and PTLUTs changes ✓ Code now in C++ version only thanks to Bobby and Alex ✓ Vasile about new CSCTF emulator: - submit it to 52X which is the development release: dead line is not defined yet - If 50X/51X will go on Tier-0 and online, he will ask the backporting from 52X ✓ Fast Preliminary study on new PTLUT2012 shows good results: - use Commissioning data with HLT_MuOpen with good vertex reconstruction and no scrapping “+” improvements in resolution for all Quality tracks “+” significant drop in rate, more then 30% in CSC region “-” some extra tracks are reconstructed -> halo muons with mode = 15 -> needs to understand “-” some plateau efficiency suppression in at pT ~ 10 GeV Still needs more study but look very promising For details see next slides

5. New PTLUT2012 golden region, Quality = 2& 3 Could be that plateau is a little bit less in new PTLUT2012 -> should be study with high pt muon data sample/ Monte Carlo

6. New PTLUT2012 golden region, Quality = 2& 3 Could be that plateau is a little bit less in new PTLUT2012 -> should be study with high pt muon data sample/ Monte Carlo

7. New PTLUT2012 overlap DT-CSC, Quality = 3 Strange drop in efficiency in new PTLUT2012 up pT ~ 15 GeV? -> should be understand and fixed -> did we change anything at overlap region?

8. New PTLUT2012 high eta |eta|> 2.1, Quality = 3 new PTLUT2012 suppress more low pT muons

9. New PTLUT2012 |eta| = 1.2-1.6, Quality 3 Improvement in resolution

10. Rate comparison for new PTLUT2012 (dash) with 2011 Significant suppression in rate

11. New PTLUT2012 golden region, Quality = 2& 3 Significant suppression in rate

13. Extra track are generated in new PTLUT2012/Firmware Some extra tracks are reconstructed: halo muons with mode = 15 Only 2 difference compare official version (compare python file configuration) 2011 new PTLUT2012 PtMethod = 28 29 FirmwareSP = 20220204 20120110

14. Back up

15. CSCTF Resolution study • Trigger: HLT_L1SingleMuOpen or HLT_L1SignleMu10 • CMSSW_4_2_8_patch6 with official Promt Analysis code: /UserCode/L1TriggerDPG/ • Commissioning data 2011 /Commissioning/Run2011A-PromptReco-v5 • /Commissioning/Run2011A-PromptReco-v6 • /Commissioning/Run2011B-PromptReco-v1 • GOODCOLL + Good Vertex + CSC skim • Use only global muon which pass standard tight muon selection: • - Track numberOfValidTrackerHits > 10 • - GlobalTrack norm. χ2 < 10 • - GlobalTracknumberOfValidMuonHits > 0 • - Track numberOfValidPixelHits > 0 • - GlobalTracknumberOfMatchedStations > 1 • - innerTrack dB < 0.2 • and matching with csctf tracks with: • dr = sqrt{dη2 + dϕ2} < 0.1 • ✓ Resolution definition • if res < 0 -> CSCTF overestimate (possible lost in ε ) • if res > 0 -> CSCTF underestimate (contribute to rate)

16. CSCTF Quality definition Tight definition Q_CSCTF_tight: Quality definition till May, 10 2011: <- this definition is used here for resolution  - Overlap DT-CSC region:   mode 11, 12, 14 -> Quality 3 (any 2 or 3 station track with DT stub)   for |eta| <1.2 any other modes -> Quality 1 no Quality 2 in this region   CSC region |eta| > 1.2:  mode 2, 3, 4 -> Quality 3 (any 3 station track with ME1)  mode 6, 7, 13 -> Quality 2 (any 2 station track with ME1)  mode 1, 5, 8, 9, 10, 15 -> Quality 1 (other tracks)Loose definition Q_CSCTF_loose: Quality definition after May, 10 2011   Overlap DT-CSC stay the same without changes   CSC region |eta|> 1.2 changes:   mode 5 -> Quality 3 (3 station track without ME1)   for |eta|: 1.6-1.8 mode 8, 9, 10 -> Quality 2 (2 station track without ME1) We are consider the possibility to return to the tight definition in the 2012

17. Firmware and PTLUTs changes • Firmware changing: make non-linear Δϕ12 and Δϕ23 for 2 and 3 • station tracks without DT stubs • 2. PTLUTs changing: • - max Likelihood • - non-linear Δϕ12 and Δϕ23 for 2 and 3 station tracks with ME1 (without DT stubs) • 3. Add Δη info to any 2 station tracks (without DT): • - 2 station track with ME1: • now we have Δϕ23 (CLCT pattern) – 4 bits • -> CLCT – 1 bit: if(CLCT = 8, 9, 10) -> 1 • otherwise -> 0 • Δη – 3 bits -> eta window could be up to 7 (1 unit = 0.0125) • we even could consider possibility • to roll back eta window cut to 6 (now 4 -> 0.05) • - 2 station track without ME1: • now Δϕ12 – 12 bits • -> Δϕ12 – 9 bits, Δη – 3 bits • ✓ Code now in C++ version only • ✓ Need test it and produce firmware which will be tested too -> still a lot of work to do • ✓ We expect reduction in the rate -> under study

18. GMT Triggering at |eta| = 0.9-1.2 GMT triggering at SingleMu after September 2011: - Q_GMT = 7: any CSC/RPC tracks -> Q_CSCTF_loose = 1, 2, 3 - Q_GMT = 6: unconfirmed CSC candidate (no RPC) if - Q_CSCTF_loose = 3 - Q_CSCTF_loose = 2 and {|η| < 1.2 (no such tracks at CSCTF) or |η| = 1.5-1.8} - Q_CSCTF_loose = 1 and |η| < 1.3 - Q_GMT = 5: any unconfirmed RPC candidate (no CSC)