Aug 31/23 PionLT/KaonLT Analysis Meeting Notes ---------------------------------------------- (Notes by GH & AH) 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, Nathan Heinrich, Muhammad Junaid, Vijay Kumar, Nacer Hamdi, Alicia Postuma FIU - Pete Markowitz CSULA - Konrad Aniol CUA - Richard Trotta, Tanja Horn, Casey Morean JLab - Dave Gaskell Nathan Updates -------------- Corrections to standard.kinematics went well. Finished! - pushed changes to GitHub yesterday - no huge errors, only small corrections - a few notable issues: - Target Mass: Corrected entries to use free proton mass rather than hydrogen atomic mass (in AMU). - LD- runs, same for deuteron (use free neutron mass) Next steps: - will meet soon with Junaid to coordinate strategy - need to double check Jacob's Hodoscope calibration, to be sure its set up in the database correctly - if it was not entered correctly, it would affect Drift Chamber calibs, and some work would unfortunately need to be redone - started writing batch scripts for full replay, which will be done after Junaid finishes calorimeter calibs - Garth: are the report files ready for the new replay? Nathan: need to add in NGC info - may want to add PID efficiency info - can probably do a full replay, then do 1st pass studies and optimize cuts for 2nd pass replay - Richard: will clean up his Lumi scripts, particularly the EDTM calcs, so they're ready for Nathan and Junaid's analysis Junaid Updates -------------- Calorimeter calibs - HMS 2021,22 calibs done - now working on SHMS calorimeter calibs (negative polarity runs), should be done in a few days - Cuts used for HMS CAL calibration: H.cent.npeSum > 7 -8 < H.gtr.dp < +8 - Cuts used for SHMS CAL calibration: H.cent.npeSum > 7 -10 < H.gtr.dp < 20 - the high SHMS npe cut is needed to select electrons (neg polarity runs only) Nest steps: - checking the hodoscope calibs (mentioned above) is the most important thing - will probably do a few DC calibs and compare, as a check - Found Jacobs's hodoscope calib slides. He sent them to Garth and Nathan, and will post them on RedMine Richard Updates --------------- Luminosity analysis - looked more carefully at getting a clean electron sample - Solved software bug which allowed bad events not passing cut to still get included in the analysis - fixing this helped a lot, much less scatter in the data - Carbon HMS relative yield vs current: looks good - compare to Carlos Yero's Carbon and LH2 boiling factors vs current - Carbon HMS Boil factor vs. ELREAL rate: looks flat without correction - (unweighed fit) - Obtained corrections (error-weighted fits): Carbon: 1.85% +/- 2.1%/100uA (slight boiling) LH2: 4.86% +/- 2.8%/100uA - results are nicely consistent with Carlos' - DG: the correlation for Carbon vs Current is tighter than the flat correlation vs Rate, this suggests a correction to the BCM calibration is needed, based on the Carbon vs Current correlation of 1.85%/100uA - then re-check Carbon vs Rate to see if the BCM correction introduces a residual rate dependence that needs to be removed (there might not be one, as the points get re-ordered on the rate plot) - finally, check the corrected LH2 correlation vs Current to get the final cryotarget boiling correction Next steps: - update SIMC for LT-sep analysis, needs to update the function in physics_iterate.f - was planning to start with the function in Marco's thesis - GH suggests to instead use the same function that Vijay has already tried, as it has more free parameters and is more flexible - Marco's analysis had many fewer statistics, and so could only use a simple function with fewer free parameters - GH suggests to get the subroutine from Vijay, as he already has it working - the only change that should be needed is to replace mpi with mK in the pole term part Vijay Updates ------------- - DaveG asked for hodoscope efficiencies per plane, needed to add to report file, still waiting for replay Next steps: - working on Luminosity Summer-2019 data Alicia Updates -------------- BSA Binning MC requested by DaveG - the question to be answered is "how much difference does phi-binning make to the BSA fit results?" - Since some sets have limited number of bins, trying to find the adequate number in bins for beam asymmetry fits - generate random pseudodata with known inputs: A=3; B=0.8; C=0.3 - Fit breaks at: 5 phi bins, and at <= 100 entries - 9 phi bins fit OK w/ good statistics - 11,13 phi bins OK except when stats poor - second study with A=6; B=0.2; C=0.1 - similar results - Deltapi data have poor statistics, study shows probably fitting only A is best - real data Q2=5.5, W=3.02: - vary # phi bins and compare Full Fit to Approximated Fit - Number of bins do not change the beam asymmetry outcome up to 1 sigma - recommended binning: Q2=5.5: 4 t-bins, 13 phi-bins Q2=4.4: 5 t, 13 phi Q2=3.3, W=3.14: 7 t, 15 phi Q2=3.3, W=2.3: 7 t, 15 phi Q2=2.1, W=2.95: 8 t, 15 phi - Dave: Use same number of phi bins for all the plots, study suggests that the minimum number of bins, maintaining good fit in beam asymmetry, is 13 bins. Makes it a lot easier to explain in the paper Discussion of new CLAS BSA paper: S. Diehl, et al., Phys Lett B 839 (2023) 137761 1-7 - analysis of sigma_LT'/sigma_0 for similar kinematics to ours - Our results are still important: expecting higher precision than CLAS-12 - combining our data with theirs, we can do Q2-scans of LT' at fixed x=0.255+/-0.01, 0.415+/-0.01 - Diehl's conclusion: JML Regge better at low Q2, GK GPD better at high Q2 - however, Regge comparison to their data is actually never very good, while GK is equally good at low Q2 - our Q2 scan at fixed x,t should provide a more definitive conclusion - Dave: Hall C should have significantly better t-resolution than CLAS-12, so t-bin migration will be much less of an issue - can do finer t-binning to better see rise in LT' at low -t - considered whether we can reliably extract LT, TT from BSA A,B,C fits - used binning MC w/ known inputs and plot outputs to see how well we can accurately measure B->LT, C->TT - A=3; B=C=0, Nevts=1000: full fit does give a slight non-zero B,C and slight shift to A - leaning to not extract B,C values, as they are too uncertain - leaning to quoting weighted average for A fits as most reliable - the MC study has been very useful, definitely we can make use of these - results in the paper - Diehl's paper used approximate fit and assign 3% systematic error from full fit - Alicia's MC study suggests that their errors largely underestimated Garth Updates ------------- In plane global offsets - wrote a simple Fortran code to do a global analysis, took only a few days - code includes the Heepcheck derivatives and Vijay/Ali's input data, with global minimization - unconstrained fit gives unrealistically large offsets to the beam energy, and this then drives unrealistically large HMS, SHMS momentum offsets - then tried a constrained fit, keeping the beam energy offsets to not larger than +/-0.7E-3, based on the arc energy measurement uncertainties - Constrained fit works much better - Dave: obtained global offsets look plausible - SHMS momentum offset: MarkJ put in ad-hoc 2% correction, another 0.18% correction to this is reasonable - GH can do a further study, including the three higher beam energy Heep data, will ask Richard/Ali for the shift data - it's not obvious whether the same set of offsets will work for the higher energy data, or not, as the HMS momentum gets quite high for some of them - will decide what to do after seeing initial fit results Out of plane global offsets - Vijay provided xptar info for three Heep coin settings from Summer-2019 data - with these data included, offsets follow linear trend, with one exception of one data (3.8 GeV) - Can Richard/Ali take another look at the replay of the 3.8GeV Heep data, there is clearly something wrong with it, maybe a wrong BPM calibration, or an error in standard.kinematics? - to extract the final offsets from the linear fit: - need to locate the first order SHMS expansion coefficient in Matrix Element file - need plots of BPM data for several Heep coin settings - will differ from online results, due to newer calibration - offsets of up to several mm are reasonable - in principle, only 1 setting is needed, but best to do at least several, to check for consistency Next Meeting ------------ - Thur Sept 7 @ 16:00 Eastern/14:00 Regina/13:00 Pacific - KaonLT will go first - Poll to see if this meeting time will continue to work for the fall https://www.when2meet.com/?21121667-zMYFn - so far, the best time appears to be Tuesdays at 16:00 Eastern/14:00 Regina