Sept 5/24 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 Present ------- Regina - Garth Huber, Alicia Postuma, Muhammad Junaid, Nathan Heinrich, Ali Usman, Nacer Hamdi Virginia - Richard Trotta JLab - Dave Gaskell CUA - Tanja Horn Ohio - Julie Roche CSULA - Konrad Aniol Junaid ------ PionLT Heep Coin Hodo-3/4 effs - new plot of 3/4 effs vs SHMS momentum - planes 1,2,3 show a linear drop of eff with increasing momentum - plane 4 shows a low eff, but very large variation (50-75%) at P~2.4 GeV/c - GH: the drop off at low momentum is expected, but the linear drop at high momentum is puzzling - Dave: *NB* confident that there is an issue in the PID selection cuts used to calculate these efficiencies, that needs to be sorted out first, and then we can come back to this - finds the 3/4 effs are sensitive to the ptof_tolerance value - PionLT value=10, apparently set by C. Yero at start of PionLT run - changing to 15 results in slightly LARGER efficiencies - Ali: please recall the beta vs. cointime issue KaonLT encountered last year, this was because the tolerance variable was set wrong (was 100, it was supposed to be 2) - a value of 2 should be applied to ALL data - Junaid had tried this as well, results in even LOWER effs More discussion about that to do about the 3/4 effs: - Dave: NCER cut needs to be removed entirely, makes sense for negative polarity SHMS, but no sense at all for positive polarity - Junaid tried varying the CERCUT parameter, all the way down to zero (default is 2pe) and it made NO difference to the efficiencies - note Nathan did calibrate the NGC, so the pedestals should have a reasonable value when applying npe=0 cut that Junaid tried - Nathan thinks the CERCUT is legacy code that is not called, since for example no NGC was installed for KaonLT. What did the code do then? - we spend some time looking at the code, it seems that the HGC is being called here instead of the NGC - however, protons also never Cherenkov in the HGC either, so this cut also should not be applied to Heep data effs - Dave suggests to remove this code completely from hcana and see what happens - Dave looks a bit more at the code, and sees that in THcHodoEff.cxx both Cherenkov and Calorimeter cuts are used - *NB* Junaid will investigate the code more thoroughly Nathan ------ PionLT Lumi studies - new plots using Zach's code applied to 2021 data - SHMS Carbon 12143-12154 - Scaler analysis shows a small linear rate dep - ElLT not yet applied - will look at fixing slopes on Carbon Scaler/NoTrack/Track analyses - Mild antiboiling seen for LH2 12192-12198 - Dave and Tanja expect Coin data are not as reliable for determining boiling corr than the 2022 dedicated singles runs - One issue is what to use for TLT - right now, Nathan picks the higher TLT from either EDTM or CPU*ElLT methods - GH: *NB* this is "cherry picking", much better to decide which method is more reliable for a given run period, and then use it consistently for both Heep and physics data Richard ------- KaonLT Ratio debugging for Q2=4.4, W=2.74 - everything is looking pretty good, except that Ratios converge around 1/3 instead of 1 - ratios much flatter than before - comparisons of data and MC distribution shapes are reasonable - Richard reports that this factor is consistent across all settings, not just the Q2=4.4 shown today - GH and Tanja agree there seems to be a missing factor of ~pi somewhere in one section of the code that is consistent with the others. When this is identified, things should look fairly good - various studies to try to find the missing factor - reran SIMC from scratch instead of recalculating weights -- ratio unchanged - looking at histo binning in Exp Yield calc - also will check SIMC w/o recon_hcana - should change the distribution shapes, but not the normalization - Tanja suggested to run for only a small t range (~0.1) to check if it's a functional issue w/t-parameteriation - Nacer: try applying your code to Vijay's data and hunt for the difference between your plots and his - Vijay should be back and available for a meeting with Richard next week Nacer ----- KaonLT Heep studies - proton absorption correction - reviewed John Matter's thesis and spreadsheet - they compared calculated and experimentally determined proton absorption results: A_pred=8.56% spreadsheet A_exp=9.03+/-0.07% by looking for missing protons other spectrometer in Heep singles data - modified their table, removed NGC, need to adjust aerogel density for actual values used during KaonLT - 0.2 density should correspond to n=1.03 aerogel - Tanja: 1.015 density=0.08, 1.011 density=0.04 all trays should have the same thickness aerogel. The area will be different though - the correction factor ~7.5%, applying this factor changes the ratio from near 1.0 to ~1.08, except for 10.6GeV data where R=1.00 after factor - tried different MM cuts to exclude pi0 - ratios get a bit smaller for MM<0.10, 0.09, 0.08 cuts, but all above 1.0 except 10.6GeV - HMS matrix element check, hcana default vs old 6 GeV matrix - data vs MC shape difference gets a bit worse (8.2 GeV data) - ratios also get a few % worse, which is a bit surprising, only expected events to move around but not get lost - clearly the default matrix elements are better, this issue closed More discussion about that to do about the ratios with absorption corr included: - Ali: asks about the momentum dependence to the interaction length in spreadsheet - Matter's thesis: cross section indep of momentum >1 GeV/c - confirmed with their data for 4-10 GeV/c - our data are 2-8 GeV/c should be fine - Dave: the proton absorption correction is applied by other analyses, it is not optional as there are no proton absorption effects in SIMC - Ali: there is a cut applied on aero tray in the replay, irrespective of whether an aero NPE cut is applied to the data - *NB* need to be 100% sure the same cut is applied also to SIMC - GH: *NB* the MM distribution clearly has a much larger tail in data compared to MC, independent of the pi0 contamination. This tail is ~5%, so seems to be the cause of the Ratio problem - Dave: *NB* please try looking at Emiss, W for Data & MC, not so used to looking at MM for Heep data, maybe it will show something - Dave: *NB* it is important to look at some Single Arm Heep to see if there are similar issues - GH agrees this is an excellent idea. Since no protons are detected, no absorption corr is needed to be applied, so it will give important information - *NB* GH's suspicion of what's wrong (added after meeting): - the spreadsheet assumes that interacting protons are completely lost and no trigger is made - if the protons scatter but still give a trigger, they end up in the Heep tail, and the spreadsheet correction is an over-estimate - if so, we would need to apply a much tigher MM cut to remove the interacting protons not included in the SIMC tail - hopefully the W and EM plots will shed light on what cut to apply - the Heep singles Data/MC ratios will also be critical to this - Henk came up with a modified method to evaluate the absorption correction, looking at the proportions of events in peak and tail, to accont for this - *NB* GH will review this info and see how to apply it here Alicia ------ Update from PRL: - referee reports received Aug 12,14,31 - queries sent to 2 referees after receiving 3rd referee report, 1 response received so far - we infer that one referee gives a report disagreeing with the other 2, and they have been asked to comment on it Garth ----- Hall C Winter Workshop Jan 13-14/25 - at last meeting, discussed a session on "nearly final results" and what KaonLT might have ready by then - 3 talks seem likely: - Vijay, Richard, Ali - Richard and Vijay should coordinate e.g. on how best to show the various systematic studies - Tanja: what about general experiment updates? - GH: this was not explicitly discussed. Since there are starting to be lots of results nearing publication (unlike a few year ago), priority will be given first them. - once that part of the program is fleshed out, we will likely go back and ask for more general updates Next Meeting ------------- - Thur Sept 12 @ 16:30 Eastern/14:30 Regina - KaonLT will go first - *NB* Please note the new time and new Zoom ID!