Kaon_LT Experiment

                 July 17/25 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, Muhammad Junaid, Ivan Zhenchuk, Nathan Heinrich,

   Nacer Hamdi, Alicia Postuma

FIU - Pete Markowitz

JLab - Dave Gaskell

Virginia - Richard Trotta

Ohio - Julie Roche

CUA - Tanja Horn

Nathan

------

Diagnosing CT shift vs HMSxfp

- checked if the timing cuts make sense by looking at

  T.coin.pTRIG4_ROC1_RawTime, etc., comparing versions w/ and w/o cuts

  - two main peaks are seen, puzzled by their origin

- went through all timing cut parameters one by one, comparing online version

  (w/ no shift) to offline version (w/shift), doing 500k replays for each

  - changed adc offset, changed set of triggers used to calculate CT

    t_coin_trignames changed from pTRIG4(HMS-ELREAL) to pTRIG3(HMS-3/4)

  - now understands why the offset changed, ELREAL to 3/4 have different cable

    lengths so the CT value has to be different

  - can reproduce shift or no shift, and changed offset

  - still need to change from pTRIG4_ROC1,2 to pTRIG3_ROC1,2

    and TDC coin windows back to online version

    - will have to change time windows accordingly

  - changed all windows all at once, still need to identify exactly which cut

    is causing the problem, likely it's related to pTRIG3,4 change

Next step will be to determine the Coin Blocking correction

- will look at CTime.ePiCoinTime_ROC1

  - will compare the contribution of the shoulders in these plots to the good reals

    and randoms region

  - *NB* Dave: it will be easier to see the blocked events in the uncorrected

    Raw CT, rather than the one with pathlength corrections added

    - try to see if it looks clearer

    - expecting you will see a time block of main coin events, and a separate

      time block of blocked coincidences

  - *NB* Garth: suggest to look at plots for runs with high and low rates,

    the low rate one should have minimal coin blocking that grows with rate

Junaid

------

PionLT LT-separations

- completed the SIMC weight recalculation script

  - met with Richard on Tuesday about it

  - testing the script now

    - shows output of 3 events, identical between SIMC and weight script

    - would like Nacer also to check the script

Next step

- will rerun SIMC w/new params for 1 setting and compare output to SIMC w/old

  params and weight recalc

- should have results of first few iterations by next week

Nacer

-----

KaonLT Low Q2 LT-separations

- zeroth iteration w/Tanja's functions and thesis params give R~0.35

- investigated fit param sensitivity to initial values in fit

  - did both ROOTfit and pythonfit

  - pythonfit is more sensitive to initial params, ROOTfit less sensitive

- data & MC comparo plots

  - more deiations w/Tanja model compared to Koltenuk model

Richard

-------

KaonLT LT-separations

- using simpler functional forms shown last week

- changed Lambeda MM selection cut from 1.08<MMk<1.14 to 1.10<MMk<1.14

- pion background is using 0.88<MMk<0.94

- no further iterations done since last week

  - ratios are near one, but show a decreasing trend with t-bin

  - clearly there's a t-dependence difference between data and model

- finds that adding sin(thetacm) to LT,TT screws up the t-dependence compared

  to without it

- Q2,W plots shown for Q2=4.4, W=2.74

- modified Wfac denominator in model

  - Wset changed from Wcentral to Wbin, the distributions are improved

- *NB* MMk low epsilon data are slightly shifted compared to MC, offset needs

   adjustment

- at high epsilon, the Lambda radiative tail is not described well by SIMC

- *NB* Dave: suggests to apply a different MMk offset for each t-bin, there may

   be a residual optics effect in the data that hasn't been corrected for

  - Nacer: Junaid and him have different MM offsets setting by setting

  - Tanja: please double check by t-bin, as suggested by Dave

- still has an issue with sig_loeps>sig_hieps for some t-phi bins, but the

  diference is getting smaller now

- first priorities will be to finish Q2=3.0,W=2.32, Q2=4.4, 5.5

  - still planning on a full data replay in August

Alicia

------

u-channel analysis

- finding strategies to fit large bkd for MMp>1 GeV

  - two options under consideration are modifying Ambrozewicz Xphasespace model

    and Pythia generator

- Xphasespace

  - Bill had suggested to modify this to include 3-pion phasespace

  - reviewed Ambrozewicz thesis

  - it is a uniform sampling of MX<W-mp, with weight taken from

    Lorentz-invariant phasespace of a 2-body process in CM frame

  - has identified where in SIMC this is calculated

  - will look up the phasespace sources cited in thesis

- Pythia

  - had a problem reading in Pythia events into SIMC, zero successes

  - to help diagnose, tried to read SIMC events back into SIMC, which also had

    some intial problems

  - traced this to a mismatch in Pythia output compared to what SIMC wants

    - SIMC wants spectrometer frame, Pythia outputs in LAB frame

    - to use LAB frame input, need to set spectrometer angles to 0 in input

      file

  - wrote a python script to investigate mean momentum and angles of Pythia

    output

    - need to adjust Pythia kinematics so they correspond to what is needed by

      SIMC

- u-channel physics discussion w/Bill @ JLab

  - comparison of DEMP over range of xB at JLab and EIC may give verton insight

  - cross-section scaling may be different, depending on if verton is exchanged

    or not in the u-channel handbag diagram (where 3q are exchanged instead of

    q-qbar as in t-channel handbag diagram)

  - Garth cautions that baryon number is not only conserved globally, but also

    locally in every part of the diagram

    - in the Standard Model, baryon current is expected to be continuous, just

      as lepton current is known to be continuous

    - this means that the 3q exchange must carry baryon number, irrespective of

      whether it includes a verton or not

Ivan

----

KaonLT Low Q2 BSA analysis

- refining cuts for final results

- plot of CoinTime vs MMk shows 4 distinct bands

  - one band is clearly pi+n and pi+Delta

  - another band is clearly K+Lambda and K+Sigma

  - what are presumably protons are divided into two bands though, one

    containing p+omega events and the other containing p+DIS events that

    Alicia is trying to model

  - *NB* Nathan: probably the RF cut is creating a gap between the omega and

    p+DIS events

    - suggest to make a version of the plot w/o RF cut

- SIMC doesn't fully reproduce MMk radiative tail, similar to what Richard

  showed

  - Garth: unless this creates a big problem, we will use it as is

- *NB* pion subtraction: need to adjust it so the region to left of Lambda peak

  is not negative

  - it appears to be negative in a few cases, need to investigate for each

    t-phi bin

  - pion background is very small in higher -t bins

Next steps

- setting up for systematics studies

- Alicia: the BSA results will not be finalized in advance of the systematics

  checks

  - we are using the checks to confirm that all cuts used in the analysis are

    correct and that the binning is appropriate

Next Meeting

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

- Thur July 24 @ 15:30 Eastern/13:30 Regina

  - KaonLT will go first