Jan 8-9/26 PionLT/KaonLT Analysis Meeting Notes ----------------------------------------------- (Notes by GH) Today: PionLT will be discussed first Thursday: Present ----------------- Regina - Garth Huber, Muhammad Junaid, Alicia Postuma, Nathan Heinrich, Nermin Sadoun CUA - Sameer Jain, Tanja Horn, Chi Kin Tam FIU - Pete Markowitz Junaid ------ Q2=3.85, W=2.02 analysis - PID: should an Aerogel cut be used or not? - tried cuts >0, 0.5, 1.5, 3.0, 5.0 npe - low epsilon: no real difference in events away from pi+n MM peak - high epsilon: a bigger difference, looks like losting more events in peak than background though - conclusion: won't apply an Aerogel cut - *NB* W=2.62 data set had a >1.5 npe cut, Nathan should recheck if this is needed - MM offsets and Diamond cut finalized - working next on t-binning Nathan ------ PionLT Lumi report - finished tracking systematic uncertainties - plotted HMS, SHMS effs vs S1X rate - found mean & stddev of residuals from linear fit to get scale and random systematics - random syst: ~10-6 so can safely ignore - scale syst: ~0.1% HMS, ~0.2% SHMS - a one -sided uncertainty, means we might be over-estimating the tracking effs - *NB* Garth: looking more carefully at the scatter of the points in the plots, suggests to take the uncertainty as +/- anyways, there are more points below than above the linear fit, but the scatter of the points on both sides of the line (particularly around 300 kHz) is about the same - info added to Lumi report - will send Lumi report to email list for comments, and eventually post on RedMine and Hall-C docDB Nermin ------ PionLT LD2 analysis - first step: reviewing Das' PID study for a single LD+, LD- setting from last year - his study was based on 4 runs of each SHMS polarity - CoinTime plots: no offset is applied, will do so - SHMS-CAL: will use <0.8 cut - SHMS-NGC: will use <2.4 npe cut, particularly important for pi- settings Next steps: - will reply all LD2 data - then do PID study, find detector efficiencies and estimate pi- purity Sameer ------ KaonLT CoinTime blocking - was using 3/4xELREAL for CoinRate - switched to 3/4x3/4, as suggested by Nathan - needed to adjust time in Rate calculation for beam-off periods, as otherwise the rate is underestimated (sometimes significantly) - *NB* however this is not yet implemented for all Rate plots - *NB* Nathan: the Q2=5.5 plot of CoinBlock vs Rate has some significant outliers that need investigation, likely something in the analysis failed that needs fixing - should use EDTM-LT for Coin data, CPU*ELLT for Singles Friday: Present --------------- Regina - Garth Huber, Nathan Heinrich, Muhammad Junaid, Alicia Postuma, Vijay Kumar, Nermin Sadoun CUA - Chi Kin Tam, Tanja Horn, Sameer Jain JMU - Gabriel Niculescu CSULA - Konrad Aniol Chi Kin ------- KaonLT Q2=3.0, W=3.14 LT-sep - investigated MM resolution discrepancy Data vs SIMC - plot ssdelta vs ssxpfp for SHMS center, low and high epsilon - plot x'tar vs MM - in both cases, looked for correlations, didn't find anything significant - also checked optics matrix element files - HMS: hms_recon_conf_opt2018 - Junaid says this is the correct one - HMS: shms_newfit_xptar_april2020 - studied MC resolution parameter sigma_res=3.5 to 4.5 - 3.5 seemed best one for left-high epsilon setting - center-low epsilon: 4.0 seems better - center-high epsilon: tried 1.0 to 10.0 - data is much wider than MC, but the effect is t-dependent - decision: will use modified resolution factor only for Lambda background subtraction, not for the MC yield calcs - updated Lambda background fit procedure - when resolution is not a problem: - fit Gausian to both data and SIMC to find peak and apply MM offset - calculate separate chi-square for peak and sideband regions, and optimize chi-square separately for each region - seems to work, but needs some individual attention that everything is working correctly - when resolution is a problem - looks more objectively at the sideband regions and change fit by hand - LT-sep results: - sigL shows a nice trend with t, but the 2nd t-bin is a bit low compared to 3rd t-bin - looking at background subtraction to check - Data/MC ratios look fairly resonable at low, high epsilon, generally 1.0+/-0.25 with some mild phi-oscillations - however, when adjusting background subtraction to move the 2nd t-bin to the 1-3 bin trend, the 1st t-bin moves by ~50%(!) - *NB* this points to a large background subtraction systematic uncertainty that needs to be understood - Data vs MC kinematic comparison plots - some significant deviations noted, particularly for HMS-delta, yptar - reran SIMC for all 5 settings - need to make sure diamond cut is correct Alicia ------ - BSA press release - some discussion on modifications to the figures to be used Vijay ----- PionLT Low Q2 LT-sep - reanalyzed unseparated cross sections with CoinTime cut width varied +/-10%, and computed the difference from the nominal cut: - low epsilon: 3.25% - mid epsilon: 1.4% - high epsion: 1.5% - Garth: suggests a scale syst unc of 1.5% for all data and an epsilon-dependent uncertainty of 1.8% that will affect primarily sigL (with delta-epsilon) magnification Next Meetings ------------------ - we will continue with the current meeting schedule for now, can adjust once people better know their 2026 timetables - Thurs: Jan 15 @ 16:00 Eastern/15:00 Regina - KaonLT will go first - Fri: Jan 16 @ 11:00 Eastern/10:00 Regina - we will continue where we left off