May 21-22/26 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 Thursday: Present ----------------- Regina - Garth Huber, Alicia Postuma, Nathan Heinrich, Dex Yadlowski, Vijay Kumar, Muhammad Junaid Virginia - Richard Trotta CUA - Chi Kin Tam, Sameer Jain Glasgow - Kathleen Ramage JLab - Dave Gaskell Ohio - Julie Roche FIU - Pete Markowitz Richard ------- KaonLT Q2=4.4 W=2.74 revisiting the background subtraction - added some checks to empirical fit for systematics check - shows plots of Lambda region yield after various subtractions: - Randoms - Dummy target - Pion leakthrough - Empirical Fit1 - Empirical Fit2 (following fit 1, as per last week) -- termed FINAL fit - there is a large drop in yield going from Fit1 to Fit2, even for low epsilon data, which has smaller background (with available RFtime cut) - the high epsilon data will be even MORE sensitive, due to significantly more background to subtract - try to quantify the background subtraction fraction by comparing after pion subtraction to after Fit 2 - also compare high-eps and low-eps data - find that lower t-bins are better behaved than higher -t (due to larger background there) - conclusion: empirical fits by themselves are not sufficiently reliable - will try empirical fits again as a last step after using Chi Kin's SIMC background subtraction method - saving all of the cross-check info and will let everyone know where the documentation is kept when complete Chi Kin ------- KaonLT Looking at effect of Ybeam on xptar - tried old HMS100 matrix elements from Tanja - worked out effect on xptar, delta - also tried forward and backward SHMS matrix elements - Dave: spoke with Mark Jones about this. He says that this procedure works, but that the first order raster matrix elements include this effect automatically - vertical position component is listed at end of reconstruction matrix elements used by hcana - shows shms_newfit_xptar_apr2020.dat where raster vertical sensitive contributions are listed (lines 213-218) - information on the file structure is available from the SIMC routine shared/transp.f, admittedly this is not very friendly documentation - *NB* CKT will comare to what he found to see if they're consistent - OOP angle offset from HeepCoin data - last week, used distribution width as the uncertainty (too large) - now using the centroid fit uncertainty plus a systematic uncertainty - evaluated the systematic by changing the fit range and functional forms - result is a bit different than what GH found, perhaps as he used a different number of data points in his fit - looking into correct recipe to apply offsets - they need to be in SIMC input files and recon_hcana - Richard: recon_hcana is set up to read these automatically from the relevant SIMC input file - shows plot of BPM xpos for HeepCoin runs - sometimes there are too few values - Dave: it depends on how often the BPM info was read out, either 2sec or 30sec. If it was set to 30sec and the run was short, there could be a small number of reads for that run - a more likely scenario, however, is that the xpos histogram ranges are too small - some variables have no entries at all, need to widen the histogram scale a lot Next steps: - will verify beam positions in production data and rerun SIMC to recheck the MM shifts Vijay ----- Low Q2 PionLT systematic study update - radiative corrections - as discussed last week, following method in Blok et al. paper - look at Data/MC ratio for different MM cuts to see if Data and MC have same tail shape, particularly noting if there is an epsilon dependence - difference from last week: plots for different t-bins, averaged over phi-bins - shows plots of change in ratio for different MM cuts at low, medium, high epsilon - t-bins 3-7: behave similarly, deviation of low and high epsilon from medium epsilon is within +/-1.5% band - deviation is quite a bit lower for bins 1-2, nearly zero for bin 1 - will apply a bin-by-bin systematic - model dependence - looked at effect on L,T,LT,TT by changing L(+/-10%), T(+/-10%), LT(+/-14%), TT(+/-36%), where the variation is consistent with 1sigma statistical error band of data - the resulting separated cross sections are not sensitive, variations of order of +/-1% Alicia ------ KaonLT u-channel - follow up on Aerogel knock-on correction for protons - using a combination of PionLT, KaonLT HeepCoin and KaonLT physics data to get a wider momentum range - since different aerogel trays are used, scaling the interaction probability by the aerogel density proportional to (n-1) - last week: binomial errors were too small, Dave had suggested the SHMS-delta dependence of the different HeepCoin runs could be enough to justify an additional systematic uncertainty - looked at the illuminated Aero-X,Y positions for different HeepCoin runs, and indeed find a variation in both the width of the illuminated area as well as the centroid, some settings have smaller regions illuminated than others - added a 1.4% systematic in quadrature w/ binomial errors - plot of knock-on corr vs momentum shows a nice linear relationship w/ reasonable error band - correcting the momentum plotted for each Heep run according to delta_SHMS, not simply using P0_SHMS - from Heep data: P=3.15 GeV/c: corr=91.5% +/- 0.9% 3.76 94.0% 1.1% - from Physics data: P=2.93 92.7% 1.2% 3.33 92.7% 1.2% - Corr to be applied to data: Q2=3.0 W=3.14 P=6.04 83.4% +/- 1.6% Next steps: - working on results to show at Hall A/C meeting in June - will include knock-on corr in Qeff - setting up binned kinematics for LT-sep framework - plans to circulate draft slides by June 5 Later steps: - determine HGC and RFcut corrections to apply to data - verify flat-dependence of KaonLT CoinLumi scans as verification of boiling and CoinBlocking corrections Sameer ------ PionLT Q2=1.6 PID cuts - initially used Junaid's Q2=3.85 cuts - these cuts appear to work for SHMS-center, need to be revised for left, right - also looking at run-dependence - some runs have small #counts - suggestions: check hclog and run sheets to see if shift workers noted anything - Nathan: don't need the NGC for pi+ analysis - the HGC in principle can be useful for K/p rejection, but needs more study to see this setting is a good one for that - *NB* will have to recheck RFtime and CoinTime cuts - shows RFtime histo for Run 16408 - extra blob could be due to e+p coincidences, need to apply PID cuts first to this plot before deciding cuts Friday: Present --------------- Regina - Garth Huber, Alicia Postuma, Nathan Heinrich, Dex Yadlowski, Nacer Hamdi, Vijay Kumar, Nermin Sadoun Virginia - Richard Trotta JMU - Gabriel Niculescu, Ioana Niculescu Glasgow - Kathleen Ramage JLab - Dave Gaskell CUA - Chi Kin Tam Nacer ----- KaonLT Sigma0/Lambda sigT ratios - used common t-overlap region of 0.11<-t<0.14 - Sigma/Lambda ratio ~0.4 with little variation acros 3 t-bins - there are slight (Q2,W) shifts between Lambda and Sigma0 for the 3 t-bins - there is a slight difference in t-center for the middle t-bin, t-centers are aligned for bins 1,3 (although the integration regions are different) - when comparing to VGL and CKY models, GH will evaluate the models at the exact kinematics of the data, so any mismatches between Lambda and Sigma0 kinematics are the same between data and model - Vijay: asks questions about forming ratio for sigL as well, to investigate K-pole dominance - this is a good idea, can be done - Gabriel: inquires to check that thetaCM was calculated correctly for Sigma0 analysis, using m_Sigma rather than m_Lambda Kathleen -------- PionLT LD2 cryotarget boiling study - 6.4 GeV data, Runs 16727-37 (carbon) and RUns 16716-26 (LD2) - these runs have singles 1 arm at a time, so EDTM should be reliable - shows Carbon plots of scaler, notrack, tracked yields vs rate and vs current - both HMS and SHMS are in agreement with what Nathan and Nermin saw, indicating that the code is working correctly - HMS CPULT looks weird even though EDTM is reliable (i.e. TLT is good) - estimation of CPULT from scaler tree gives ~80% - estimation from ELLT and EDTM give 120% - Nathan was unable to figure out what is wrong, Kathleen is welcome to look at it in more detail if she wants - LD2 boiling plots of scaler, notrack, tracked yields vs rate and vs current - SHMS: last 2 boiling points are too low, giving an unrealistic slope, and needs double checking - some other scans look more reasonable - Nathan: it's important to combine plots of different settings together vs rate and vs current, these can give clues as to what is wrong with outliers - if the different settings are truly measuring the LD2 boiling, they should be consistent with each other - *NB* will resume weekly meetings with Nermin to compare analyses and discuss issues Nermin ------ PionLT LD2 cryotarget boiling study - went through 3 different lumi scans, checking the current cuts and PID cuts for some runs - this improved the tracked, untracked yields for some runs - sees same issues @ 6.4 GeV as Kathleen - planning to combine all sets together - also will look at 7.4 GeV lumi scan Nathan ------ PionLT Q2=5.0 W=2.95 t-binning - applyed t-shifts before deciding binning, so we have some discussion first about MM offsets - mid-eps Right-SHMS setting has a ~10 MeV MM shift, plots of data and MC indicate they really do differ by that much - *NB* it would be worthwhile to confirm the correct optics offsets are used in the replay for this setting - many other MM offsets are much smaller, two high-eps settings have opposite sign than the others - t-shift is typically ~0.0015, but with varying sign as noted above - 7 t-bins selected: low-eps: 3600-4100 counts, mid-eps: ~5000, hi-eps: ~4500 - compared normaized data yields to physics_pion simulation: Data/MC ratio about 5.8 Next steps: - calculate average kinematics per bin - start iteration procedure - while that's going on, will then set up Q2=6.0 W=3.09 setting - Q2-scan at x=0.39 also needs Q2=2.12 data - low epsilon data were taken in 2019, high eps in 2022 - *NB* Vijay offers to replay 2019 data, Nathan will send a run list - Nathan will also need the HeepCoin offset values Vijay used Next Meetings ------------- - Thurs: May 28 @ 13:00 Eastern/11:00 Regina - PionLT will go first - Fri: May 29 @ 12:00 Eastern/10:00 Regina - we will continue where we left off PS: Gabriel and Ioana have been patiently waiting for some time for their presentation and discussion - *NB* people should expect next Friday's meeting to go at least 30 minutes longer than usual, and try to have their slides for Thursday if possible