Kaon_LT Experiment

                 Apr 4/24 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 - Garth Huber, Nathan Heinrich, Nacer Hamdi, Ali Usman, Alicia Postuma,

   Muhammad Junaid, Vijay Kumar

FIU - Pete Markowitz

Nathan

------

PionLT Lumi update

- still waiting on Dave Mack re BCM calib, will prompt him tomorrow

- have calibs that are valid for 2 lumi scans @ 9.2 GeV, show first results

9.2 GeV Carbon Scan #1

- sees a small trend in HMS TrkEff over 0-15 kHz range, 99.8-99.75%

- SHMS TrkEff shows similar trend, but up to 200 kHz

- sees a 5% drop in RelYield at low beam current, corresponding to a 5%

  over-estimate of beam current at 3uA, a 0.15uA offset

9.2 GeV Scan #2

- similar trends seen

Some questions about corrections to the beam current in the scripts

- first there is an unusual logarithmic correction to the beam current

  attributed to Peter Bosted

- after this, a second correction is applied

- both are applied to the values used in Nathan's analysis

- more investigation is needed to understand these

- would like to discuss them with Richard

- it would be helpful also if Vijay could check what's in his code

Junaid

------

PionLT Heep Coin settings

- list of cuts shown for both data and SIMC

- data are normalized for Q and TrkEff

  - DetEff and EDTM are not yet applied

  - BCM calibs not yet correct

- Data MC comparos shown for a variety of variables for all 9 Heep settings

  - overall the shapes agree nicely

  - rebinned plots for Heep offsets

  - PMY will be used to determine OOP offsets

  - Ali: notices that PMZ, EM resolution is much worse than predicted by SIMC

    - could be due to a BPM calib error

    - will discuss this further with Junaid

Ali

---

piDelta BSA study

- results shown for Q2=3.0, W=3.14 high statistics setting

  - 2 t-bins, 8 phi-bins

  - bins with very low statistics or Delta fit fails are excluded

- Procedure used for MM fit:

  1) fit pi+n MC by matching integrals over 0.92-0.96

  - MC is narrower than data, so matching integrals results in the peak being

    too high, which is offset by being below data in the width

  2) fit SIDIS MC in region with minimal Delta contrib, 1.4-1.5

  - for bins where the data don't extend all the way to 1.5, the fit region is

    slid slightly lower, keeping the fit region width the same

  3) then fit piDelta MC to Bkd subtracted data

  4) sum of MC is formed after all fitting

- Result of fit shown for low -t, all phi bins, L,C,R SHMS

  - this is first time fits are shown with longer SIMC data runs

  - resulting MC distributions are smoother, resulting in better fit overall

  - MC consistently under-predicts 1.0-1.06 MM region

    - possibly the SHMS collimator scattering that GH is implementing will help

      fill in this region

  - there is a slight over-prediction of data around 1.1 in -90 to -45 phi bin,

    can be included in systematic

  - Right SHMS: 4 phi bins

    Left SHMS: only 2 phi bins, 2 bins removed due to failed fits, will show GH

    later

  - BSA shown vs phi, weighted average of L,R,C

    - BSA calculated by integrating Delta MM from 1.12-1.4 GeV

      - will do a MM cut dependence study later

    - resulting asymmetry is very nicely sinusoidal

Vijay

-----

Low Q2 pi+ LTsep iterations

- getting some weird results fitting interference terms

  - new params from fitting process not behaving as expected

  - going through code again to see if there is a bug

  - if checks don't find anything, will try setting LT,TT to zero and try

    improving from there

Alicia

------

pi+n BSA error analysis Part 2

- first show of Q2=3, W=2.32 results

  - BSA are similar to before delta-correction

  - t=0.25-0.30 still has a wonky phi distribution

- Q2=2.1, W=2.95

  - first bin (-t=0.01-0.04) full fit is over-fitting, but still gives results

    consistent with other bins

- Some small changes to Systematics Calc from Tuesday's discussion

  - will not include (A-A') simple vs full fit comparison in RMS of

    cut-dependence study

  - (A-A') is much larger than the others, so lumping it into RMS calculation

    would result in under-estimate of uncertainties

  - delta(CUTS)=RMS[ (A-CN), (A-CW), (A-MN), (A-MW) ]

  - shows table of errors for each point at one Q2, STAT, CUTS, BPOL, FIT

  - then takes SYST as sum of CUTS, BPOL, FIT in quadrature

  - shows LT' results vs -t for each setting, with double error bars

  - we should be able to finalize the selection of CLAS data to compare to ours

    for Q2-scan at fixed X_B soon

Nacer

-----

Low Q2 KaonLT RFtime correction

- plot of P.gtr.th vs RFtime shows correlation, data is for 4-5 runs

- fits a straight line: 0.64-3.05*xptar

- applies this as an adhoc correction, RFcorrected

- plot of MM vs RFcorrected shows even better separation between K+L, K+S and

  piDelta than before

  - width of piDelta band is narrower than before

  - will redo cut study using RFcorrected

  - can clearly see some K+L events wrapped around the bottom of plot

    - could not see these before due to wider pion background

    - will need to place a lower limit on pion cut at ~0.1 to keep them

    - alternately, can adjust RF offset by ~0.5ns to better center the K+L

      events in the RF window

Next Meeting

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

- Thur April 11 @ 15:00 Eastern/13:00 Regina

  - KaonLT will go first