Oct 30-31/25 PionLT/KaonLT Analysis Meeting Notes ------------------------------------------------- (Notes by GH) Today: PionLT will be discussed first Please remember to post your slides at: https://redmine.jlab.org/projects/kltexp/wiki/Kaon_LT_Meetings Thursday: Present ----------------- Regina - Garth Huber, Muhammad Junaid, Alicia Postuma, Nathan Heinrich, Nermin Sadoun JLab - Dave Gaskell Virginia - Richard Trotta CUA - Tanja Horn, Chi Kin Tam Ohio - Julie Roche FIU - Pete Markowitz NOTE: The JLab furlough will start EOB on Nov 14. Dave recommends that you have everything you need off the iFarm by EOB on Nov 13. Nathan ------ PionLT CoinLumi analysis - very nearly done first draft of technical report - still to do: use standard deviation of Normalized Yield around fit value using the EDTM to estimate the systematic uncertainty Junaid ------ PionLT Q2=3.85 W=2.62 LT-sep - fixed phi conversion issue identified last week - now: if phi<0 add 2pi; if phi>0 do nothing - using same L/T/LT/TT parameterizations as before - for sigL, all parameters float except those in the Fpi parameterization Fpi=1/(1+p7*Q2+p8*Q4) - Data/MC ratios now much better behaved - small oscillations in ratio for higher -t bins at low epsilon - ratios flatter at high epsilon than low epsilon - this implies the model is over-estimating the low epsilon LT/TT terms - kinematic and focal plane comparison plots - generally see shifts between data and MC for Q2,W that are in opposite directions to each other, due to the kinematic correlations - however, there are some settings where the shift is much less - *NB* Garth: suggests to look at these comparison plots each iteration, to see if the differences are getting less with iteration - *NB* Garth: also good to compare sig_UNS %-change between iterations, our criteria to stop iterations is when sig_UNC shifts only 1-2% - aerogel tray cuts - confirmed using same cuts in both data and SIMC Next steps: - will start writing report on the iterations - will start setting up for Q2=3.85, W=2.02 LT-sep Chi Kin ------- KaonLT LT-sep - compared yields w/o random subtraction with Richard, got same results - then applied MM cut, finding count differences with Richard of up to 300 - still investigating Alicia ------ KaonLT u-channel LT-sep - last week: showed that Q2=3.0, W=2.32 Pythia generator had extra bumps compared to higher W settings - compared to data: seems to more closely match Pythia than Xphasespace generator, however, it seems there would be no room for rho MC in this case - wondering whether exclusive rho is included in Pythia, investigating - Garth: notices that both the Pythia and Xphasespace end at MM~1.3 GeV, while the data extendes to MM~1.4 GeV. Is there some acceptance cut difference between data and MC? - discussion about the significantly worse MM resolution for this setting, compared to others, omega and pi+n peaks not cleanly resolved - double checked what SHMS optics were used, this was okay, but p_HMS=6.59 GeV/c, so need to replay with newest high momentum HMS matrix elements, this was an older replay - Q2=2.1, W=2.95 - MM resolution looks good, still double checking MM offsets - Q2=4.4, W=2.74 - very nice omega peak - only high epsilon replayed so far Friday: present --------------- Regina - Garth Huber, Alicia Postuma, Nathan Heinrich, Nermin Sadoun, Muhammad Junaid, Nacer Hamdi, Vijay Kumar Virginia - Richard Trotta CUA - Tanja Horn, Chi Kin Tam, Sameer Jain CSULA - Konrad Aniol JMU - Gabriel Niculescu Glasgow - Rachel Montgomery (joined at very end) Richard ------- KaonLT LT-sep - rechecked output of SIMC variables - when adding a variable, need to modify both NtupleInit.f and results_write.f - NtupleInit.f sets the ROOT branch names, it's vital to confirm that this list definition is consistent with what's output in results_write.f - found an issue, at some point (when converting to ALMA9?) a duplicate entry 29 in results_write.f was made, which caused thetacm to be over-written - this explains why LT/TT iterations were messed up, fixed now - will put a list of SIMC variables and their order in the ROOT tree on RedMine - the fix will affect LT:sin(theta), TT:sin(2theta), t-dep of L,T will be unchanged - interations after thetacm fix in SIMC ROOT tree - Data/MC ratios now have a sizable phi-dependence, R varying from 0.3 to +2.0 in a smooth single oscillation, pointing to LT term - *NB* Garth: the Ratio indicates the Data and MC model are out-of-phase from each other by 180deg, it's important to double check that the sign of the LT term is handled correctly in both data and MC Chi Kin ------- Follow-up to yesterday's KaonLT yield difference - met with Richard, now get same yields per setting - the issue was that Richard had a different definition of Yield than Chi Kin - Richard's was scaled by bin size in both Data and MC, the scale factor canceling in the ratio - this was an artifact from an earlier study, now removed Nacer ----- KaonLT Q2=0.5 LT-sep - had to correct phi in same way as Junaid reported yesterday - now: if phi<0 add 2pi; if phi>0 do nothing - with this correction, wiggles in Data/MC ratios are reduced after each iteration, instead of growing - ratios sag below 1 for high -t, high epsilon bins, which points to the t-dep of sigT being slightly wrong in the model - first iterations for K+Sigma0 data - using same functional forms in model as Lambda, given that sig_factorized2007 also had same functions (but different parameters) for both - LT=TT=0 initially in this model - using initial params from sig_factorized2007 gives Data/MC ratios ~40 (compare to ~6 for Lambda), with more wiggles in ratio for high -t, high epsilon bins (double bump so TT) - after 1 iteration, ratios move to near 1, but big double-bump oscillation at high -t - *NB* Garth: please make focal plane and kinematic plots of data and MC overlaid for the K+Lambda settings, so we can investigate systematic trends, just as we did for Junaid yesterday - it's also important to look again at Sigma0 region MM plot, this time with MC overlaid, so we can confirm the Lambda radiative tail isn't causing any issues Vijay ----- PionLT Q2=0.375 CoinBlocking correction - shows plot of CoinBlocking vs CoinRate - generally around 0.95, but interestingly some outliers 0.88-0.92 around 0.8 kHz - *NB* Nathan: suggest to double check the timing cuts for the outlier runs - Nacer: could this correction be momentum dependent? - Nathan: only in the sense that the timing could shift at low momentum, not otherwise - Vijay: wonders if the outliers are dummy runs - Garth: comparison with CoinBlocking vs RunNumber plot indicates the outliers are all near the beginning of the run, RunNumber ~8500, so not likely to be all dummy runs - *NB* Junaid will make a simlar plot of CoinBlocking vs CoinRate for PionLT physics data, to check for outliers - confirmed that no aerogel tray cuts were applied in either Data or MC, since the timing resolution is excellent and the momenta low, no Cherenkov cuts were needed for PID, but it was good to confirm that no tray cuts were erroneously applied To Do: - Q2=0.425 CoinBlocking correction - Garth also would like to see kinematic and focal plane plots of Data and MC overlaid, just as for Junaid and Nacer Sameer ------ KaonLT high Q2 CoinBlocking - investigating what raw time cuts apply in hcana when evaluating the CoinBlocking correction - Nathan: enforcing raw time cuts in the tcoin param file is important, as you'll recover some of the events that have early noise events. This is because the cut in tcoin.param is to force hcana to only form coin times with triggers in this range. - To fix this issue I would strongly recommend that you change the cuts on T.coin.pTRIG1_ROC1_TdcTimeRaw T.coin.pTRIG4_ROC1_TdcTimeRaw - in the tcoin.param file to very tight cuts based on T.coin.pTRIG1_ROC1_tdcMultiplicity == 1 cut and then replay the data with the new version of the tcoin.param file. - look at T.coin.pTRIG1_ROC1_TdcTimeRaw and select fairly tight cuts to eliminate the large tail to the left of the main peak - when replaying with the cuts, the recovered events will move from the tail to the peak. The cuts cannot be in the python script, they must be in hcana itself - to find the where the cuts are applied, Nacer suggests to look at: /u/group/c-kaonlt/USERS/nacer/hallc_replay_lt/DBASE/COIN/standard_KaonLT.database PARAM/TRIG/KaonLT_PARAM/tcoin_Winter18_Offline.param Gabriel ------- Read the paper on the CEBAF accelerator at: https://journals.aps.org/prab/abstract/10.1103/PhysRevAccelBeams.27.084802 which indicates a large beam energy variation for 4,5 pass beam to Halls A,C - there is a variable in hcana which tracks the variation, need to rescale this variable according to the arc energy measurement and the offsets - was erroneously using a fixed beam energy value instead, MM resolution and offsets are improved now - *NB* Richard: there is a flag in hcana which indicates whether to use the tracked beam energy value or only the central value, it's important for everyone to confirm that this flag is enabled in their analysis Next Meeting ------------ New 2-meeting/week structure - Thurs: Nov 6 @ 16:00 Eastern/15:00 Regina - KaonLT will go first - Fri: Nov 7 @ 11:00 Eastern/10:00 Regina - we will continue where we left off *NB* Regina group needs to note the changes in meeting time! - UK groups should be okay, after this week's confusion due to differing Standard Time conventions