Mar 9/22 PionLT/KaonLT Analysis Meeting Notes --------------------------------------------- Participants: - Richard Trotta, Vladimir Berdnikov, CUA - Pete Markowitz, FIU - Dave Gaskell, JLab - Jacob Murphy, Ohio - Nathan Heinrich, Garth Huber, Muhammad Junaid, Stephen Kay, Vijay Kumar, Love Preet, Ali Usman, Regina Please remember to post your slides at: https://redmine.jlab.org/projects/kltexp/wiki/Kaon_LT_Meetings Nathan ------ 1) Mode-10 Data Analysis Update - Shows waveform output (equivalent to oscilloscope trace) for one event - turn ON parameter to get trace for every event in a data file (would take a lot of space) - Plots of mV vs PMT# for both hcana and FADC readout - hcana has a software threshold applied while FADC does not - overlay of mV plots for individual HGC PMTs shows good agreement between FADC and hcana - *Suggestions* - Gaskell: one idea would be to add a new variable which replicates an integrating ADC, i.e. integral over a fixed time range, so see how this compares with the FADC algorithm * DG will discuss this with Mark * - Huber: we need to keep track of #events where we get waveform data instead of firmware info, so we can easily monitor which channels have more errors - Stephen says this would be easy to add to the DEF file - right now, the waveform data replaces the FADC data if there is an error, but there is no monitor of how often this was done - *Another task for DG* - DG will talk to Mark/Brad on whether we can get the FADC firmware fixed to return BOTH firmware and waveform data when there is an error in time for the upcoming run - *Goal for Next Meeting* - There is additional HGC waveform data that NH needs to look at Richard ------- - brief update on Heep-COIN analysis - reproduced Vijay's Heep-COIN data and SIMC plots for one 3.8 GeV setting - still no efficiencies or livetimes implemented - *Suggestions* - Gaskell: it would be advisable to do your first livetime calculation with just Comp-LT instead of EDTM-LT - Electronic-LT should be close to one anyways since we used buffered readout, so this would indicate problems separately with Comp-LT (expected to be the largest contribution) - *Goal for Next Meeting* - Goal is to have 10.6 GeV Heep-COIN plots (data and simulations) at next meeting (hopefully with efficiencies, livetimes?) Stephen ------- - ifarm batch code updates - looking at hcswif. Seems nice for literally submitting a replay, but it would take a lot of work to make it as flexible as the scripts we already have - *Suggestions* - Huber: it would be useful to advertise our job system more broadly, see if others adopt it. Stephen suggests a name: ltswif Ali --- - setting up replay scripts on ifarm, had problems - *Goal for Next Meeting* - Goal is to have 8.6 GeV Heep-COIN plots (data and simulations) at next meeting (hopefully with efficiencies, livetimes?) - tracking cuts still need to be implemented, tracking efficiencies are otherwise implemented, but won't be accurate until cuts are set/checked Vijay ----- - simulating p(e,e'K+)L and comparing to data - getting much higher data yield than SIMC, so SIMC will be re-scaled by an arbitrary factor for now - background from piDelta (underneath) Lambda peak not yet subtracted - shows a COINTIME plot - getting two sets of e-pi and e-K coincidence peaks, the main batch at -4 and a secondary set near zero - *Suggestions* - Kay: it looks suspiciously like the setting has a change in offset for some runs. Please check if there is any run-dependence in the peak positions. - Later note: there does not appear to be any such run dependence, so the double peaks have some other (more troubling) origin - Huber: look at epi-COINTIME in addition to eK-COINTIME. The main background is pion leakthrough, which is tilted on a 2D plot of eK-COINTIME vs MM. It would be better to remove the leakthrough with a "vertical cut" on epi-COINTIME vs MM, as otherwise a titled cut is needed. - after this cut is applied, then look at eK-COINTIME for the remaining events - tight COINTIME cuts are okay for investigating issues of data purity, but they are difficult to use in the physics analysis, as then we have to determine what fraction of (e,e'K+) events are removed by the cut and correct for them. Better to use a looser cut that doesn't throw away an appreciable number of (e,e'K+) events and subtract the leakthrough by some other means - identifies some problem with low Q^2 hodoscope calibration - Dave Mack and Carlos Yero are working on a new hodoscope calibration code - *Task for Jacob* - Jacob will follow up with DM and CY about the status of the calibration code. It would be great if we could test it on the PionLT and KaonLT data, and see if it resolves any of the hodoscope calibration issues we have - *Goal for Next Meeting* - will look first at the COINTIME offset issue - then look at data/simc comparisons for central Q2=0.50 setting Jacob ----- - just got back from holiday, so no progress report - *Goal for Next Meeting* - will do some testing of new hodoscope calibration code - continue to look at EDTM-LT calculations - will meet with Richard and Brad on this Upcoming Run ------------ - we will start with the identical target configuration to what we had in the fall run Next Meeting ------------ Wednesday March 23 at 13:00 Eastern/11:00 Regina