Kaon_LT Experiment

                 May 4/23 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 - Nathan Heinrich, Ali Usman, Alicia Postuma, Love Preet, 

   Muhammad Junaid, Portia Switzer, Garth Huber, Vijay Kumar

Ohio - Jacob Murphy, Julie Roche

CUA - Richard Trotta, Tanja Horn

CSULA - Konrad Aniol

FIU - Pete Markowitz

JLab - Dave Gaskell

Nathan Updates

--------------

HMS Cherenkov Calibrations

- sees two peaks in PE spectrum, behavior is puzzling

   - selects low calorimeter energy, low peak doesn't change

   - selects electrons with calorimeter, only low peak present

   - both peaks are at very small amplitude <20, most of the histogram up to

     100 is empty

- electrons appear to be in lower peak and pi- in upper peak, but this makes no

  sense

   - we are confident that the detector is working properly, because the 50k

     detector replays during the experiment made sense, something else must be

     wrong with the analysis

- GH had earlier suggested looking into timing cuts

   - cuts are a little wide, could be tighter

- HMS Cherenkov detector time cuts

   - PMT#1: 3 peaks

   - PMT#2: 2 peaks

- DG: could Cermode-0 be causing issues?

   - NH: hasn't looked yet at mode-10 data, but could

- DG would like to look at the data

   - root trees are in Nathan's volatile directory

Later in the meeting, we return to this topic

- DG makes a plot of H.cer.goodAdcPulseInt with 

  H.cal.Etotnorm>0.7 & abs(H.gtr.dp<8) cut

   - sees a more sensible distributino where pi- have 1PE peak ~25 and e-

     distribution is ~60-160

   - DG: please look at run interactively first, not a fan of running a sript

     until you know what the script needs to do

   - somehow NH is not selecting the electrons

Junaid Updates

--------------

- Nathan, Jacob and Junaid had a discussion earlier this week on PionLT

  analysis timeline

   - the updated schedule is posted on the Wiki at:

     https://redmine.jlab.org/projects/hall-c/wiki/Analysis_Tasks

HMS DC Calibrations

- looking now at LH+, LD+ runs.  For LD- runs, is it better to use pi- or e-?

   - DG: e- are probably easier to calibrate, drift maps should be equally

     valid for both

   - GH: also probably more e- than pi- for many LD- runs

- HMS Drift Distance distribution for different Planes

   - new calibrations are a significant improvement, residuals looking much

     better

   - sharp peaks at both ends of drift distance spectra

   - DG: Cameron Cotton needed to place Multiplicity Cut=1 to avoid extraneous

     peaks at end

      - the calibration algorithm needs clean single track events so this makes

        sense, and should probably also improve the calibration result

- still working on SHMS DC calibrations, will need same Multiplicity Cut

Jacob Updates

-------------

Results of delta-optimizations from 2022 run

- 6.8 GeV/c HMS delta-scan

    - delta_true-delta_recon sees improvement

    - W_calc vs X_fp, after corrections the slope is largely gone, centered on

      right W value

    - still have a correlation for X_fp>35

    - GH: since these are a small portion of the events, maybe try placing a

      cut and fitting only the large X_fp separately

- 5.9 GeV/c HMS delta-scan

   - delta_true-delta_recon still not quite lined up at zero

   - W_new vs X_fp, boomerang shape near X_fp~0 after calibration, but overall

     W peak is narrower than before

- working on optics work chapter of his thesis

   - Nathan asked for an operation manual on the optics, will base one on his

     thesis chapter

- Next Plans

   - once done writing chapter, then help with Calorimeter calibrations, since

     there are a lot of them to do

Richard Updates

---------------

Proton leakthrough issue that we discussed last week

   - some cuts were probably not applied correctly, a cut-and-paste issue

   - no new results yet

Luminosity studies

- met with Jacob on PreScale Calc to be sure it's calculated correctly

   - TLT calculation look fine

   - scaler read code, discrepancies noted by CarlosY at Hall C software

     meeting last week, couldn't find anything wrong

   - Jacob noticed that Carlos' T3-scaler rates and BCM currents nearly

     identical, which seemed suspicious

      - T3 = ELREAL

      - meeting with CarlosY tomorrow about this

Small update on carbon 10.6 Lumi Run #1 w/ SHMS positive polarity

- HMS carbon, LH2 scans

   - restored some previously "bad runs", current cuts were too tight, only 6

     sec data in some cases

   - +2% carbon anti-boiling @ 65uA

   - same value for LH2

   - still investigating a few issues

   - DG: the highest current/rate point is driving the fit, probably a flat

     line would work, in which case no significant rate dependence

   - GH: the fit doesn't look like it corresponds to minimum chi-square, is

     this an error-weighted fit?

      - RT: no.  Error bars are plotted, but not used in the fit

      - it's important to ALWAYS do error-weighted fits, otherwise you get

        artifacts like this, where the high rate point with larger error bars

	drives the fit

   - GH: to improve the statistical significance of the high rate region,

     suggests to combine both sets of data

      - if the rate/current effect is real, both will behave consistently

      - first for each set of data extrapolate the fit to R=1 at zero

        rate/current

      - then use the extrapolated point to separately normalize both sets of

        data

      - then do a combined fit using both sets of normalized data

      - that should give 2 high rate points instead of one

      - if you truly see "carbon anti-boiling" then it indicates a rate

        dependence that needs to be applied to all the data

      - only after that step can the cryo boiling vs current be determined

Ali Updates

-----------

BPM calibration issues

- was using only one file for all KaonLT runs

- there was BPM girder work in Feb 2019, so different calibs are needed for

  10.6,3.8,4.9 GeV compared to 6.2,8.2 GeV and Summer 2019 data

- DG provided a different calib file for 2019 data

Comparison with no offsets - 8.2 GeV data

- Left=Old, Right=New

- PMdata has shifted significantly

   - PMZ: shift larger, PMY: shift closer to SIMC, PMX: shift away from SIMC

- Target variables

   - xptar: HMS almost exact overlap w/SIMC now while there was a shift before

   - SHMS also looks pretty good

- HMS delta almost unchanged

- 6.2 GeV data are similar, W distribution now worse

DG: follow up on Vijay's email

- DG's file assumes use of EPICS xpos variable from survey

- Discussed with MJ, hcana uses the RAW variables, not the survey offset

  corrected variables

   - either set of variables will probably give a good result, but 

     MarkJ says it's better to use the RAW variables if you want to compare

     your results with those from other experiments

   - GH: it's important to have the ability to cross-compare our offsets to

     those from other experiments, so we can have greater confidence in our

     result

- DG: will generate new files early next week

   - both the fall 2018 and spring 2019 files need to change, all data will

     have to be replayed with the new calibs

- summer 2019 data has a dedicated calibration by MarkJ

   - in principle that calib should be the same as for spring 2019, will be a

     useful cross check

- DG will also look again at the 2021 BPM calibs

   - slopes should not change in 2022, but offsets will, these can be

     determined from the harp scans

Plot of Raster_Y vs W

- looks like a slight left pointing slope

- DG: change sign on RasterY gain and see if it improves, hopefully it gets

  more straight

DaveG has to leave at this point.  He has agreed to take notes at next week's

meeting in GH's absence

Discussion on determination of OutOfPlane offsets in Tanja's thesis

- what is beam offset D?

   - the offset needs to be projected to the target, in a manner similar to

     what was done in the counting house GUI used in PionLT

   - use the closest BPM (x,y) values, and the (x',y') slopes between the last

     two BPMs, then extrapolate to z=0

- Ali shows plot of xptar_SOS vs formula from Tanja's thesis

   - numbers are from means of xptar from both spectrometers, and errors on

     means

   - one offset comes from slope, the other from the offset

- Ali will endeavor to make a similar plot from KaonLT data, using both the

  3.8,4.8 GeV data to get more range

Vijay Updates

-------------

Verbal update

- working on HMS calorimeter and Cherenkov efficiency studies, and setting up

  LT-sep software for second Q2

- will also be starting low Q2 Lumi study

Alicia Updates

--------------

BSA progress

- 5 (Q^2,W) settings with full phi-coverage @ 10.6GeV

   - submitting jobs for 2 (Q^2,W) that were not analyzed yet

Systematic study on MM cut

- looking at HGC and how it affects MM dist vs t

   - magenta=data w/o HGC cut, K+ leak through

   - blue=data w/ HGC

   - grey=SIMC (weighted and scaled to data w/HGC cut)

   - lowest 2 t-bins have good agreement between blue and grey, but significant

     difference seen for highest 2 t-bins

   - data much wider than SIMC at high -t

   - for full data, without t-binning, the agreement fairly good, looks like

     MM cut ~0.96 okay

   - Julie: is this due to bad description of radiative tail at high -t?

   - Tanja: gives link to 2010 document on the radiative corrections in SIMC

     https://hallaweb.jlab.org/data_reduc/AnaWork2010/mkj_simc_mceep_radcor.pdf 

      - radiative corrections not updated since then, probably not sufficient

        for some regions of 12GeV data

   - Ali: Alicia's data of course have no diamond cuts, since there is no

     LT-separation

      - hopefully the discrepancy between data and SIMC will be less at high -t

        when the diamond cuts are applied

- the good news is that no significant BSA variations are seen with HGC cut

   - fluctuations are consistent with BSA error bars

   - w/HGC Asin(phi) error bars ~60% larger than without

   - preferring to not apply an HGC cut to the BSA data since the K+

     leakthrough can be removed entirely with MM cut and then keep full pi+

     statistics

Next Meeting

------------

- Thur May 11 @ 16:00 Eastern/14:00 Regina/13:00 Pacific

- KaonLT will go first