Kaon_LT Experiment

                 Aug 21/25 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, Nathan Heinrich, Ivan Zhenchuk, Muhammad Junaid,

   Nacer Hamdi

Virginia - Richard Trotta

CUA - Chi Kin Tam

Nathan

------

PionLT Lumi analysis

- ran into some bugs when porting recent TDC cut changes to Lumi analysis

  scripts

  - confirmed that cut file changes are incorporated correctly, works now

  - ifarm changes: needed to add manual load of data from tape silo to analysis

    scripts, rather than being done automatically

  - should have new Lumi plots by end of week

- working on report of analysis changes

  - will have a meeting with Mark Jones next week on variables, e.g. not sure

    if cut on HT2 is needed or not

Junaid

------

Q2=3.85 LT-sep

- extended t-plotting range of functional form fits, as suggested by Dave last

  week

  - now see that sigL went very high at low -t, below fitting range, which

    explains why the ratios were bad even though the function fits looked good

- set LT,TT=0, also in sigL set p7=0 in exp(p6-p7*ln(Q^2))*(-t)

  but no change in other parameters

  - that means the subtraction term in the exponent is removed, ratios got

    worse, which makes sense

- trying a different functional form for sigL

  - Bosted parameterization uses

    sigL=(p5+p6/Q2)*-t/(t-mpi^2)*Q2*exp(p7*|t|)/(1+p8*Q2+p9*Q4)

    where p9 is small

  - using Bosted's input parameters for sigL, sigT stays as before from Tanja,

    got Data/MC ratios ~5 instead of ~30, with a phi-dependent ratio

    - first iteration with the new functional form is in progress

    - if it looks good, then will look more at LT,TT functions to use

Nacer

-----

KaonLT Q2=0.50 LT-sep

- trying different functional forms

  - previously was using Fpi-2 pi+ functions, now switching to K+ functions in

    sig_factorized_2007 (extension by Tanja of older Koltenuk model)

    - only using sigT, setting L=LT=TT=0 since the epsilon dependence is much

      less than for pi+n

    - Data/MC ratio using initial params from sig_factorized_2007 is ~40,

      puzzled since earlier SIMC studies had R~2

      - Garth: I believe your earlier SIMC studies used Koltenuk's

        sig_factorized, not sig_factorized_2007, so you should check this

      - doing some checks to confirm that SIMC and weight recalc scripts give

        same intermediate result

    - one good thing is that the kinematic and focal plane distributions have

      better agreement between data and MC with new functions than before

Ivan

----

KaonLT BSA analysis

- still writing Co-Op term refport, full draft by end of week

Richard

-------

KaonLT Q2=4.4 LT-sep

- fit studies for bkd underneath Lambda peak and Sigma0 region at higher MM

  - revised cut to 1.10<mK<1.14, back down from 1.16

    - thinks 1.14 cut is better as Sigma0 fit is then not needed

  - pi+n region is selected with 0.88<mpi<0.94

- with initial param fits (IT=0) with 1.14 cut, no Sigma0 fit

  - Data/MC ratios near 1 with some phi oscillation at low -t

- then do 3 iterations, sig_uns determined from Data/MC ratios changed hugely,

  particularly around phi~180deg

  - some high epsilon t-phi bins have very small errors, which could be driving

    the fits in weird directions

  - Richard will show some slides next week on the explicit calculation of

    uncertainties so we can see if they're done correctly

- starting on Q2=3.0, W=2.32 bkd subtraction checks

  - background underneath Lambda peak is more significant than at Q2=4.4

  - Chebyshev 2nd order polynomial appears to be doing a good job

- Junaid: when you do parameter fitting, do you apply fitting bounds?

  - Yes, but in a complicated sense.  The fitting script is automated, start

    with very wide parameter bounds then they get narrower in refits (in same

    iteration).

  - also applies a chisq<3 criterion

Cin Kin

-------

KaonLT Q2=2.115, W=2.95 LT-sep

- will start with bkd subtraction and fitting

- right now: running SIMC, applying general cuts (PID, CT, ...), applying MM

  shifts

- writing code for background subtraction

  - found an error in the charge uncertainty calculation

    - Dave Mack specifies 0.5% + 0.15/I in quadrature

    - erroneous code had a simple sum instead of quadrature and a decimal place

      error in the coefficient

    - Richard will propagate this correction to his code too

Garth

-----

pi+n BSA PLB changes

- shows new version of Fig 4, xB vs Q2 data points

  - made CLAS, CLAS-12 points smaller, changed shading of CLAS-12 bin regions

  - also added error bars (large) to CLAS-6 data, indicating the huge regions

    over which they are binned

  - Nacer found the plot to be confusing, the meaning of the shaded regions was

    unclear, also skeptical of the value of including any CLAS-6 data given the

    large binning regions

Next Meeting

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

- Thur Aug 28 @ 15:30 Eastern/13:30 Regina

  - KaonLT will go first

_ Note: we will need to do a poll soon, to pick a time for September