Kaon_LT Experiment

                 Sep 19/25 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, Muhammad Junaid, Alicia Postuma,

   Nacer Hamdi, Vijay Kumar, Nermin Sadoun

JLab - Dave Gaskell

CUA - BC Manoj, Tanja Horn, Sameer Jain, Chi Kin Tam, Carlos Yero

Virginia - Richard Trotta

York - Stephen Kay

Nathan

------

PionLT Coin Blocking Studies

- Trigger timing was different in 2021 and 2022 runs

  See Brad's Trigger Setup Documentation

  https://hallcweb.jlab.org/wiki/index.php/Trigger_History

- SHMS-3/4XHMS-ELREAL was used as the CoinTrigger, so remmade CoinTime plots

  using that info

  - get a long distribution of times to the right of the main peak

  - Dave: SHMS-3/4XHMS-3/4 is the highest resolution version of the CoinTime,

    HMS-ELREAL has extra jitter in it caused by the other elements in the

    ELREAL trigger, need to keep using 3/4X3/4

    - it is not the CoinTrigger that is lost (that's already taken into account

      in the deadtime) but rather that the CoinTime is wrong for some events

    - the problem is that the TDC looks for signals in the lookback window and

      some events can sneak in and give a bogus CoinTime that is removed by

      cuts.  In that sense it's not really a Coin blocking, but a CoinTime

      blocking

    - for the blocking correction, need to use a window based on the 3/4X3/4

      signals

- tried two cuts, wide (110ns) and narrow (80ns)

  - the narrow cut gives a flat NormYield vs Rate, but the BlockingCorr is 0.8

    (i.e. 20%) even at low rate, this is very difficult to justify

  - the wider cut gives a 6% reduction of NormYield @ 2kHz CoinRate

    - some discussion comparing to the BlockingCorr that Peter Bosted's

      correction, which is proportional to Total SHMS+HMS rate

    - plots made of the BlockingCorr vs Total SHMS+HMS rate and CoinRate

      - the CoinRate gives a much better correlation, vs Total SHMS+HMS rate

        the scatter is significantly larger

      - the BlockingCorr uses loose PID cuts of:

        HMS: Hcal_EtotNorm>0.7, Hcer>2

	SHMS: HGC,Aero,NGC,Cal to loosely select pi+

- Dave, Tanja, Garth agree that applying Nathan's correction makes the most

  sense for the PionLT data

  - the correction will be calculated Run-by-Run based on the #events

    outside and inside the window for each setting, rather than from the

    fit vs rate

  - the fits vs rate will be used to estimate the systematic uncertainty in

    the BlockingCorr, Nathan will fit each CoinLumi setting separately,

    determine the scatter in the slopes vs CoinRate and from that calculate

    an uncertainty for a given CoinRate

  - the uncertainty in the CoinRate for each setting will have to be

    propagated separately in the errors for each run

  - *NB* Nathan will make a presentation on this at next Wednesday's Quarterly

    Analysis Meeting

- we also need to determine an equivalent correction for KaonLT data

  - Sameer Jain will look into this, following instructions from Nathan

    - KaonLT high Q2: using Chi-Kin and Richard's ROOT files

      - CoinTime plot will be a bit different than for PionLT

        - the used RO1,2C may be different

	- the cut windows and offsets need to be checked and changed

    - it also makes the most sense that he also look at the low Q2 KaonLT and

      PionLT data, since the setup will be nearly identical

      - *NB* Vijay and Nacer needs to point Sameer similarly to the location of

        these data: run list and location of calibrated data

  - Alma9 Setup on ifarm - Pion-LT (E12-19-006) - Redmine for virtual envir

    https://redmine.jlab.org/projects/hall-c/wiki/Alma9_Setup_on_ifarm

Junaid

------

PionLT Q2=3.85 LT-sep functions

- tried fiferent fits, with different LT,TT params floated

  - obtained Data/MC ratios centered at 1, but with phi-oscillations

- Latest version:

  - added Q2-dep to LT,TT: both have param/Q^2 dependence

  - in addition: LT has p12/t^2 and TT has p15/t^3 dependence

  - all LT,TT params can float

  - this gives Data/MC ratios with much reduced phi-dependence at high epsilon

  - feels on the right track, now that fatios are fairly flat

Vijay

-----

PionLT Q2=0.375, 0.425 LT-sep

- modified LT,TT functions

  - removed ft=t/(|t|+m_pi^2)^2 from LT equation

  - LT now has 1/t^2 and TT has 1/t^3 dependence, similar to Junaid

  - W-factor unchanged

- added offsets to all SIMC input files

- rechecked Diamond cuts, SIMC diamond was slightly off, one edge of diamond

  needed a small adjustment

- Q2=0.375 Data/MC yield ratios

  - high eps: fairly flat at low -t, slightly bigger osc at higher -t

  - mid eps: similar

  - low eps: oscillations a bit bigger for higher -t bins

- Q2=0.425 ratios have had less work, general trend is similar, so promising

- sig_UNS

  - Q2=0.375 now has sigL bigger than sigT for all except the highest -t bin,

    which is closer to expectations

    - also sigTT is now much smaller at low -t, also closer to expectations

  - Q2=0.425 sigL has more fluctuations, will need more work

  - *NB* Garth: notes that none of the functions have any Q2-dependence.  This

    was fine for earlier (more approximate) work, but now that we're getting

    closer to final results, we should add in a small Q2-dependence, to take

    into account the variation of the cross section across the diamond

Alicia

------

pi+ BSA paper modifications

- summarized changes to the paper and the draft Response to Referees

  - *NB* please send comments within a week.  Goal is to resubmit the paper by

    the end of September

Chi-Kin

-------

KaonLT high Q2 LT-sep

- checked Diamond Cut for Data and SIMC for his Q2=2.1, W=2.95 and Q2=3.0,

  W=3.14 replays

  - huge difference between Data and SIMC diamonds, with the Data diamonds

    being significantly smaller than the SIMC diamonds at both Q2

    - *NB* clearly there is something wrong, Chi-Kin and Richard will discuss

      and recheck

- for now, just using the Data low epsilon diamond

  - formed NormYields for Data and MC per t-phi bin

  - for the MC yields, tried both Richard's latest work and CK's corrected

    parameterization (with positive weight, discussed last week)

Richard

-------

KaonLT high Q2 plans

- looking at Q2-scan at fixed xB:

  Q2=3.0, W=2.32; Q2=4.4, W=2.74; Q2=5.5, W=3.02

- until mid-Oct: looking at refinements to LTsep model

- November: preliminary systematics

- by Thanksgiving: final data replay

- then CK and RT will look at the Q2-dependence study, depending on how much

  time RT has (due to A1n analysis)

- also working on a Fitting Algorithm paper, likely for Comp.Phys.Communic.

Nacer

-----

Low Q2 KaonLT LT-sep model optimizations

- how using a model inspired by Tanja's sig_factorized_2007 report

- before any iterations: Data/MC ratios ~7 for both low and high epsilon,

  reasonably flat vs phi

  - sigUNS eseentially has low eps=high eps, i.e. sigL nearly zero

  - keeping LT=TT=0 in parameterization, although of course the Rosenbluth fit

    gives some non-zero values from the phi-dependence

- 1st iteration: Data/MC ratios ~1, except at low -t, where sigL diverges

  - later iterations: ratios change only a little

- one issue is the diverging sigL at low -t

  - *NB* Richard had the same issuesuggests to double pion contamination, which

    is particularly a problem at low -t

  - will also set p5=9 in the exponential vs t in sigL and let LT,TT params to

    float

    Richard's parameterization

    https://redmine.jlab.org/attachments/2819

  Next Meeting

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

- Fri Sept 26 @ 11:00 Eastern/9:00 Regina

  - KaonLT will go first

** NOTE THE NEW TIME **

- this time will be in effect until USA daylight savings time change, after

  which it will switch to 10:30 Eastern/9:30 Regina