Kaon_LT Experiment

                 Aug 29/24 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, Alicia Postuma, Ali Usman, Nathan Heinrich, 

   Muhammad Junaid, Vijay Kumar, Zach Sullivan, Nacer Hamdi

York - Stephen Kay

JLab - Dave Gaskell

FIU - Pete Markowitz

CUA - Casey Morean, Tanja Horn

Virginia - Richard Trotta

Nathan & Zach

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

PionLT Lumi studies

- last week it was discussed using TLT from CPULT & ElLT instead of EDTM

  - ElLT calc from 55ns*(Total Rate)

- Plots  Red: EDTM  Blue: CPU method

  - SHMS Carbon Runs 16738-16746

    - w/EDTM, a small yield rise w/current, 2 outlier runs

    - CPU method: LT used if it disagrees w/TLT >1 sigma

      yield is flat within errors

  - HMS LH2 Runs 16703-16712

    - CPULT~80%, must be an issue with how it is calculated

    - Dave: *NB* you can infer the CPULT from other info in the report file

      - it should be #ADC-events/PreTrig-Scaler-events

      - maybe the calc is pointing to the wrong PreTrig

      - can calculate by hand to check

- quickly runs through slides for next week's Quarterly Analysis Mtg on BCM

  calib, looks good

Junaid

------

PionLT SHMS-3/4 Hodo Effs

- some runs have low eff ~85-87%

- plots of single plane eff vs 3/4-Trig-Rate and S1X-Hodo-Rate

  - 1st plane: ~99%

  - 2nd plane: one set at 95-96%, and other at ~98%

  - 3rd plane: two trends: flat at 78-82%, exp drop 98-80%

  - 4th plane: two trends: flat at 80-90%, lower trend 65-70%

- Dave: *NB* suggest plot of Eff vs Momentum

  - quartz bar effs are lower at low momentum

    - makes it difficult to test 3rd plane eff when 4th plane inefficient

- Dave: *NB* check the slop parameters in hcana on track projection to planes

  in param file

  - assumes there is a valid trigger and a vaid track

  - looks at each plane in succession and uses track selection to find out

    which hodo bar should have been hit, and then checks if that hodo fired

  - maybe the track is pointing to the wrong hodo due to multiple-scattering

  - if the track slop parameter is too tight, then the hodo is not counted,

    leading to a too-low efficiency

- Dave: *NB* it would be helpful to see if there is a version of this

  efficiency calculation with PID cuts

  - i.e. in the 6-GeV era, we had multiple calculations of efficiencies:

    overall, pions only, protons only, electrons only

  - not sure if this was carried over into the 12-GeV era code or not

- Dave: *NB* checks the code and finds there is a hardwired Cherenkov cut in

  the hodoscope efficiency calculation, CERCUT

  - it might be inappropriately set

  - if this cut value is changed, hcana needs to be recompiled

- SHMS_Hodo-S1XY, S2XY effs are much higher, >90%

  - Dave: these were for an STOF OR test, not so relevant anymore

PionLT HeeP

- data have pi0 contribs

  - calculate Exp/MC ratio w/ and w/o pi0 cut, MM<0.1 GeV

  - 8.5 GeV data: Ratio goes from 1.136 (w/o cut) to 1.028 (w/ cut)

  - implemented Ali's Drift Chamber resolution param in SIMC for this analysis

Nacer

-----

KaonLT HeeP

- looking at 8.2 GeV data, which have worst Exp/MC ratio

- ElLT calc:

  - using 55ns*Rate >99.9%, where HMS-ELREAL+SHMS-3/4 ~ 4kHz

  - using ElLT from EDTM and CPULT: 95.6-99.6%

- EDTM: R=1.126+/-0.006  w/ElLT: R-1.111+/-0.006

- MM comparo

  - Change resmult=1.0 to 3.5 (Ali's resolution smearing parameter adjustment)

  - delta vs MM for SHMS, HMS

    - data shows a wider region (delta<0 SHMS, delta>0 HMS) than SIMC

- pi0 background, place conservative cut >0.11 GeV, could have used >0.10

  - with resolution tuning but no MM cut, still get an extra bump in delta

  - add MM>0.11 cut, the data-MC overlap is much better

  - R-1.027+/-0.006

- Ratios for the other Heep runs:

   3.8: 1.000 +/- 0.004

   4.9: 1.001 +/- 0.004

   6.2: 1.019 +/- 0.004

   6.8: 1.027 +/- 0.006

  10.6: 0.930 +/- 0.005  w/o MM cut: R=1.043

- Overall, very good news, things much better than before!  Well done Nacer!

- GH: *NB* should investigate some MM cut dep

  - *NB* should also look at applying a proton absorption correction 

    (John Matter's spreadsheet, but need to adjust cross section for specific

    momentum)

  - *NB* finally, need to estimate the systematic uncertainty in the Ratio, it

    won't change the Ratios themselves, but it will change the significance of

    the deviation from 1, and help establish our overall systematic uncertainty

    - this would include not only the systematic in the Exp yield, but also the

      systematic in the SIMC yield (Nacer already determined this is higher at

      10.6, from elastic model dep)

- Discussion about using old HMS Matrix Elements

  - Dave has found a systematic improvement between Data and MC comparison in

    XEM data when using 6-GeV era matrix elements (ME)

    - apparently these ME are not in the standard distribution, Dave had to

      find them from older code

  - it sounds worthwhile to check this with the Heep data, and if it is better,

    then we should also use it for the physics data (both KaonLT and PionLT)

  - *NB* Dave will send more information on what file to use

Richard

-------

Q2=4.4 LT-separations

- in a separate meeting, Dave had suggested to recheck recon_hcana calculations

- plots of xptar vs yptar, all cuts applied, for SHMS, HMS

  - also: delta vs delta

  - agreement of shapes between data and MC seems reasonable, although there

    are some binning issues with the data which hopefully will be fixed with

    improved plotting

- Some weird results last week were due to SIMC using a different

  parameterization than the iterations, so params in one set of eqns did not

  work in the other set

  - that is fixed now, things starting to improve

  - the issue, however, is that ratios get stuck ~0.4-0.6 for all Q2

    - Garth: strongly feel there is still an inconsistency in the code

      somewhere, Ratios must converge to 1, if the function is wrong, there

      will be weird shapes to the ratios centered around 1

    - the good news appears to be that the bug seems to be the same for all Q2,

      so if Richard finds it once, it should be good for all settings

  - Vijay does not recalculate SIMC weights, he reruns SIMC every time, so

    checking against Vijay's version of this code is not very meaningful

    - *NB* Vijay can meet with Richard later next week, probably Thursday

  - will do a full replay of all data, including HeeP, to reproduce Nacer's work

Ali, Alicia, Vijay

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

- no report this week

 Next Meeting

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

- Thur Sept 5 @ 16:30 Eastern/14:30 Regina

  - KaonLT will go first

  - *NB* Please note the new time and new Zoom ID!