Mar 28/24 PionLT/KaonLT Analysis Meeting Notes ---------------------------------------------- (Notes by GH) Today: KaonLT will be discussed first Please remember to post your slides at: https://redmine.jlab.org/projects/kltexp/wiki/Kaon_LT_Meetings Present ------- Regina - Garth Huber, Ali Usman, Alicia Postuma, Muhammad Junaid, Nathan Heinrich, Nacer Hamdi York - Stephen Kay JLab - Dave Gaskell CUA - Richard Trotta, Tanja Horn CSULA - Konrad Aniol Richard ------- RF TOF correction - discovered an issue in xptar, xptar distributions for physics data - Heep data had good similarity between data and MC xptar, yptar distributions - p(e,e'K+)L had similar MC xptar, yptar dists than Heep, but physics data dists were different than expected - investigated whether this was optics issue, etc. and finally tracked the problem to originate from the RF cut - the issue is that there is no TOF pathlength correction to the RF time, giving a correlation between xptar, yptar and RF time - the RF cut (for low epsilon K+ data) was too tight, and removed events at edges of TOF distribution - the RF distributions for K+ and pi+ overlap strongly for higher Q2 data, so the RF cut is not all that effective for these events - relying on pion subtraction anyways, so probably will not use RF cut for these data Nacer ----- Low Q2 KaonLT PID analysis, Q2=0.5, W=2.40 THshms=9.79(left) - for these data, the RF time cut is remarkably effective - plot of RFtime vs MMk shows a large separation between K+Lambda, K+Sigma and pi+Delta events - plots of K+Lambda yield and K+Lambda/\sqrt(sig+bkd) vs (>RF cut) show a stable plateau for some ranges of cuts, ~1.05-1.2 - plot of MMk for various types of cuts: - using all other PID cuts but no RF cut, still lots of pi+ bkd underneath K+Lambda and K+Sigma peaks - using only RF cut, no other PID, pi+ are highly suppressed - comparison of K+Lambda yields for various cuts indicates that we can get rid of all SHMS PID cuts and keep only RF cut, reduction in K+Lambda peak counts is minimized in comparison to bkd, with additional SHMS PID cuts the K+Lambda peak yield also is decreased - Dave: is the SHMS momentum low enough to rely only on SHMS_beta? - NH shows beta distribution plot, and MMk using beta cut, clearly not effective for distinguishing pi+ from K+ - Clearly for Nacer's analysis we need to deal with the RF TOF correction identified by Richard - Stephen had tried implementing the TOF correction from the CoinTime, unfortunately it was not successful (the TOF correction seemed to go the wrong direction, made the RF worse instead of better), and not pursued further - Stephen has two blocks of code where he was trying various things: One block is at: https://github.com/sjdkay/hcana/blob/d4c646ecc46b99bfd610669f57047c4e8a83e457/src/THcRFTime.cxx#L348 Another big block of stuff in THcCoinTime - CT corrected RF values: https://github.com/sjdkay/hcana/blob/d4c646ecc46b99bfd610669f57047c4e8a83e457/src/THcCoinTime.cxx#L170 Nacer can access some of Stephen's RF testing work on lark at: /d8tb3/sjdkay/JLab/RF_Class_Update_21_10_22_Tests - might be some pdfs with testing there. Unfortunately his notebooks are in York. will check when he gets back next week and forward anything on Ali --- pi+n Radiative Tail contribs in piDelta BSA Region - pi+n normalizing SIMC to data peak vs data integral Option 1: match pi+n peaks - use for systematics studies Option 2: match pi+n radiative tails - too extreme, won't be used Option 3: match pi+n integral - chosen default for analysis - integral range is 0.92-0.96, seems to approximately balance overheight at peak with underwidth lower down - Other fits planned for systematic studies: - integrate Bkd subtracted data - integrate piDelta MC fit to data - MM-cut dependence Alicia ------ pi+n BSA results w/ delta_SHMS correction now implemented - Q2=3.0, W=3.14, 7 t-bins, data looks very nice, maybe a bit smoother than before (did not yet do a detailed comparison) - Q2=5.5, W=3.02, 5 t-bins - Q2=4.4, W=2.74, 2nd t-bin still has some phi-oscillation as before - hopes to have rest of BSA results ready by early next week PARTONS GPD model update - previously found weird statistical fluctuations in PARTONS output - uses a MC module, needed to increase MC calls by 10x to remove fluctuations - PARTONS now gives roughly the same shape as LT' data vs t, but ~3x too high - will rerun PARTONS again with GPD H_Tx2, as Stefan Diehl did, to improve agreement BSA uncertainty discussion - for systematics studies, will have 2 BSA fits, and cut adjustments to CT, MM - separate discussion needed to incorporate all this into errors - Tues Apr 2 @ 15:00 Eastern/13:00 Regina/12:00 Calif Zoom ID: 991 7509 1967 Passcode: 225211 Nathan ------ PionLT Lumi analysis - modified RT's scripts to get them working for PionLT - added some dummy=-1 flags for HMS when doing SHMS only - getting some weird results: TLT~4000% for Carbon Runs 16738-16746 - running times are very nearly zero - most events apparently are excluded by beam current cut - BCM not calibrated for these runs yet - Dave Mack can do BCM calib for this one run next week - NH can do other BCM calibs if DM gives instructions - other aspects do look good: - CPULT looks reasonable - SHMS track-eff ~99.88%, seems suspiciously too good - Richard: there are some hard coded corrections and cuts in the script you need to watch for Junaid ------ PionLT Heep Coin study (9 settings) - applied same cuts to Data & SIMC - BCM calibs not yet done, as noted by Nathan - made plots of PMX,Y,Z,EM,W for in-plane offsets - all beam energies show same trends, unlike KaonLT - hopefully this means that all 9 settings can use the same offsets, will see - GH asks for some plots to have finer binning so that peak centroids are more accurately defined - next steps: - HMS Calorimeter and Cherenkov efficiency studies Stephen ------- Reminder to send comments for York NSTAR* conference abstract Next Meeting ------------ - Thur April 4 @ 15:00 Eastern/13:00 Regina - PionLT will go first - Also don't forget about the dedicated BSA uncertainties discussion - Tue April 2 @ 15:00 Eastern/13:00 Regina