May 25/23 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 Present ------- Regina - Garth Huber, Ali Usman, Muhammad Junaid, Nathan Heinrich, Portia Switzer, Alicia Postuma, Love Preet, Vijay Kumar FIU - Pete Markowitz Ohio - Mark Mathison, Julie Roche CUA - Richard Trotta JLab - Dave Gaskell CSULA - Konrad Aniol, Jonathan Conrad Richard Updates --------------- Spend most of last week on thesis writing, so only a short update - tried fixing UpRoot script issues - converting lists to NumPy arrays, these should be faster and more compact and address some memory issues - working on Lumi scans with Coin physics data, needed to modify some scripts, still waiting for jobs to finish running Ali Updates ----------- OOP offset studies - since SIMC xptar are non-zero, now subtracting Data-SIMC xptar values - Left: same as last week, Right: Data-MC - only slight movement of points, clustering slightly tighter - include 3 Heep points from Summer 2019 run - these have Pp/Pe ratio near 1 - fortunately, the 3 new points continue the main linear trend as the others - now it's much more clear that is one outlier at highest Pp/Pe~1.3 - this point is Heep with 4.9 GeV beam - the other low Q2 KaonLT run at 3.8 GeV is on the correlation, with Pp/Pe=0.83 - actually the two points have nearly the same xptar computed y-axis values, but they are at very different Pp/Pe ratios, so one point is on the correlation and the other is far off it - curious to see if *BPM* will explain the outlier - some discussion about the fact that the linear trend is with a positive slope, while the plot in Tanja's thesis is with a negative slope - DG: need to keep in mind the role of the HMS has switched from Fpi-2. There, HMS was the proton arm, here it's the electron arm This means you need to compare the slope of HMS in TH plot to offset of HMS in AU plot, since the spectrometer roles are reversed - some discussion about how to check the BPM value to see if the outlier can be discarded or not - DG: hcana already corrects for the beam position in the optics, so you just need to be sure BPM position is held constant - if the BPM value (extrapolated to target) is much different, then can't use this run for OOP offset study - VK: found a variable in hcana, is it the correct one? H.rb.raster.fr_ybpm.tar - DG: yes, it sounds like the right variable, would need to look at the hcana to be 100% sure - this is in BPM coordinates: y-vertical, x-beam left - NH: even better, simply compare the variable values with a hand calculation of the beam position at target, extrapolated from BPMs 3H07Ay, 3H07By, 3H07Cy - started looking at in-plane offsets using Heep-Check program, will present results next week Vijay Updates ------------- HMS Calorimeter Efficiency studies - physics data with new cuts, ~99% - Heep data, ~99.9% - DG: please re-check if using Binomial Errors on the Eff Calc, as the errors seem a bit large - DG: need to be careful with the Calorimeter Efficiency at low HMS momentum - at very low P, the shower peak is broader - for a constant Calorimeter cut, expect the efficiency to drop off at low momentum - VK: lowest momentum for Heep runs is P_HMS=1.73 GeV/c lowest momentum for physics runs is P_HMS=0.544 GeV/c So clearly this will be an issue, and we can't use only the high momentum Heep runs to determine the efficiencies for low momentum physics settigns - KaonLT should also check, there is probably a physics setting with P_HMS<1 GeV/c - Suggestions: - for next week, please show a HMS Calorimeter distribution at low momentum, it is probably broader - VK comments that yes this is an issue, he's using a much lower calorimeter cut for these low momentum runs (>0.2) - DG: please pick a fixed cut and plot all efficiencies versus P_HMS, you should get a smooth curve showing the drop off at low momentum - also, confirm that Binomial errors are being used Alicia Updates -------------- BSA analysis - looking for a newer version of VGG code, ours doesn't calculate LT' - JR: DVCS no longer uses VGG as it doesn't fit our data very well there is a new model from CEA called PARTON3D, I'll send information on it - RT: suggests that GH send an email to Meson WG - simluated pi+Delta0 channel, to see if it contributes the tail seen in high -t MM spectra - pi+Delta0 contributes significantly to bins 4,5 but at higher missing mass, does not contribute to the pi+n tail - there is clearly also room for 2pi-n phase-space before the pi+Delta0, but this has a threshold of MM~1.08, and also should not contribute significantly to the tail - DG: agrees this is not the cause of the tail - suggests to make a *plot* of MM vs delta for the high -t data, to see if there are any weird *correlations* - GH: thinks the issue is that there is an aspect in the radiative corrections model that doesn't work so well for the higher -t data. This deficiency was probably not visible at 6 GeV, where the momentum transfer was lower - at any rate, AP will proceed with the cut-dependence studies, clearly the higher -t bins will have a larger systematic due to the tail and we will just have to live with it - coming tasks: - Ali and Alicia need to meet to discuss t-binning - re-extract sig_LT' for all settings - MM and CT cut dependence studies - need to go through the info on Beam Polarizations from Steve Wood Nathan Updates -------------- HMS Cherenov calibrations - did more calibs on runs around the strange behavior periods, so that each calibration goes up to the end of a given kinematic setting. It's probably not a good idea to change calibrations part-way through a setting - 4 regions chosen, giving 4 different calibrations - error-weighted averages used to determine the calib for each one - updated standard.database and pushed to GitHub - looked at effect of new calibration for various runs, compared to online - 2D plot of HMS Cherenkov vs Calorimeter doesn't show much difference - probably not surprising, since many of the selected runs have online calibrations that are close to the new ones - will look for a run where the calibration difference is larger, as a check - JR: the #events seems lower for offline than online - NH: yes, the cuts are a bit different now, tighter Working on NGC calibrations - found a bug in code, alerted Cameron Cotton, NH already fixed it - looking to implement Petr Stepanov's multi-Gaussian peak fit that was used in Aerogel calibs - for the Aerogel, we found that this gave a better calibration than a fit with a simple Poisson distribution Junaid Updates -------------- - no report, working on reports for Proton Structure class - next week: SHMS DC calibs Next Meeting ------------ - Thur June 1 @ 16:00 Eastern/14:00 Regina/13:00 Pacific - PionLT will go first - DG has agreed to take notes on June 1,8 as GH will be unavailable in Europe