Aug 31/22 PionLT/KaonLT Analysis Meeting Notes ---------------------------------------------- (Notes by GH and SJDK) Please remember to post your slides at: https://redmine.jlab.org/projects/kltexp/wiki/Kaon_LT_Meetings Present: Regina - Stephen Kay, Muhammad Junaid, Vijay Kumar, Garth Huber, Ali Usman CUA - Richard Trotta FIU - Pete Markowitz Richard Updates --------------- - Luminosity study Yield calculations - Total Charge - some EDTM-LT still don't make sense, discussing with Jacob - EDTM live times slightly different between KaonLT and PionLT - Due to event type changes? - SHMS = Event Type 1 and 3 - Potentially can't disentangle types 1 and 3 uniquely - HMS = Event Type 2 - How do prescales work in combination with this? - GH suggestion: for the case where EventType 1 and 3 have same tdcTimeRaw, divide up these events according to the relevant prescale ratios and see how well this works - CPULT - Using this rather than EDTM-LT - DaveG suggested just using this across the board - Assume what electronic LT should be based on total LT - Extrapolate - Tracking efficiency - Lumi uncertainties being updated - Full uncertainty calculations for HeeP and luminosity studies still underway. For example, need to include uncertainty in BCM calibration, particularly the offset at low current - Lumi Scan Results of Yield vs Beam Current/Rate - Low current setting improved a lot. - Overall, carbon lumis looking better although still not flat (need errors on them yet) - Tracking is also an issue - Some cuts need adjusting still - Etracknorm - Beta - No consistent trends in the Lumi, which is also still a problem - all of them use TLT, not CPULT Ali Updates ----------- - High Q2 HeeP-COIN settings w/Richard - All resuts include efficiencies - Need to give eqn next time, so we can see exactly what was done - 6.2 GeV Results - HMS and SHMS xfp/yfp, slight offset - HMS and SHMS xpfp/ypfp look ok for both spectrometers - Pmiss looks very different, EXP wider and flat compared to SIMC - Emiss also looks very different - Seems to be PMz that's the big issue, PMx, PMy look more similar - This difference would filter through to the Emiss and Pmiss plots - PMy also looks a little flat? - Replay comparisons to investigate the source of the Exp/MC PM, EM issue Test #1: - asked Vijay to replay an 8.2 GeV setting and process through python analysis - Results look basically the same - So doesn't seem to be something tied to the replay - Rules out calibrations, param files etc -> Not quite - It rules out a DIFFERENCE in the files being used as the cause Test #2: - DaveG suggested to look at delta vs PM/EM to investigate saturation effects - PM/EM should be independent of delta if optics are correct - 6.2GeV: pHMS=3.571, pSHMS=3.486 - not surprising no saturation effect - 10.6GeV: pHMS=6.390, pSHMS=4.840 - Looks pretty flat across delta, no obvious correlation - Because these distributions are just raw replay quantities, this implies *strongly* that this is an issue with the replay Test #3: - Low Q2 analysis. Try to reproduce Vijay's analysis in Ali/Richard framework - Full Analysis including replaying the data and generating SIMC files etc - Results show EM/PM deviation from Vijay's analysis! - Ali's Distributions are wider than Vijay's - Difference in replay files - Vijay had better agreement with SIMC - PMy and PMz are the two that look very different - Most likely explanation: Vijay seems to have some different (but correct) set of param files, *specifically* for low Q2 - Files for High Q2 seem to match (test #1) - Stephen - My opinion is that it's only really the replay step itself that could be causing these issues given that it's a difference in shape So now the problem is that we need to hunt down the param file that's the issue. - A *quick suggestion* on trying to figure out what's going on with the Emiss/Pmiss difference: Vijay and Ali should pick some low Q2 run number, and then replay it The output file itself isn't really that important, but we might as well make it to compare too - What's more important is the initial info dump from hcana. You should copy *all* of this into a text file and do a diff between the two - Hunt through carefully and identify any different param files in use - You should also verify that any param files that *look* like they're the same actually are between your two repositories. Again, just do a diff between the files - You should also compare the std.kin values for the run in question Test #4: - send high Q2 data to Vijay for him to process - waiting for results - GH: This seems like a useful check, but predicts you won't see any difference, due to what was already learned in Test #1 Vijay Updates ------------- - No updates - Prepping for comittee meeting even though that's not for another month - Working on setup of LT separation software - Will start Lumi study after progress report is done - kinematic offsets discussed at last meeting not yet adjusted - will have new values by time of committee meeting - Discussion on diamond cuts - Check Jacob's new code - Fits 4 lines to the low epsilon diamond, applies cuts to data - Preliminary t-phi binning of data - Important, but need to finalise other studies first - Spending too much time on issues that can be left until later. Need to concentrate more on the needed preliminaries for the data analysis. - These diversions are actually costing you more time, not speeding things up! - Please give a *high priority* to tracking down the discrepancy from Test #3. It is good news that the discrepancy was identified, as it gives a tool to investigate what is wrong. It is VERY IMPORTANT to fix this soon, as it is delaying many other studies Junaid ------ - No updates - Working on the lumi scan - Goes live tonight! Next meeting ------------ - GH proposes a meeting next week, as the following week will be difficult due to our travel back to Canada, etc. - Thursday September 8 @ 11:30 Eastern/09:30 Regina time - Hoping to see at this meeting: - improvements in Lumi from Richard - PM/EM debugging progress from Ali+Vijay