Kaon_LT Experiment

                Mar 12-13/26 PionLT/KaonLT Analysis Meeting Notes

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

                                (Notes by GH)

                    Today: KaonLT will be discussed first

Thursday: Present

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

Regina - Garth Huber, Nathan Heinrich, Alicia Postuma, Nacer Hamdi

Virginia - Richard Trotta

CUA - Chi Kin Tam, Sameer Jain

CSULA - Konrad Aniol

JLab - Dave Gaskell

Nacer

-----

KaonLT Low Q2 cross section checks before systematic studies

- KSigma0/KLambda sigT ratios for eventual comparison with Nachtmann's prediction

  - select t-range in common for both reactions: 0.11->0.14

  - t-dependence of both in this region is faiarly flat

  - used 2 approaches to estimate ratio

    - directly compare closest t-bin values, no correction for slightly

      differing Q2,W,t values

      1st bin: Lambda -t=0.112   Sigma0 -t=0.111

      2nd bin: Lambda -t=0.122   Sigma0 -t=0.128

      3rd bin: Lambda -t=0.140   Sigma0 -t=0.139

    - compare Lambda data values to model values for Sigma0 at same t

    - plot of results

      Red=Data Method       Blue=Model Method

      Red:  ~0.25 +/- 0.05  Blue: ~0.2 +/- 0.1

      - results are consistent within (large) errors

      - blue has much bigger errors than red, due to sigificant parameter

        uncertainties in Sigma0 model

      - Garth: your Lambda model probably has smaller uncertainties, can try

        comparing Lambda model to Sigma0 data as a 3rd method

- recheck of pi+ leakthrough after background subtraction

  - checking if steep rise in sigL at low -t is due to pion contamination, pi+

    cross section should be largest for high epsilon, low -t bin

  - MM plots do not indicate any significant evidence of pion contamination,

    region to left of Lambda peak looks clean

- this hypothesis is ruled out, what about other possibilities?

  - Richard: the effect of the t-shift will be significant since Nacer's t-bins

    are so narrow, it could be a significant effect

  - Nacer will investigate this next

- CoinTime Blocking

  - found that Autum2018 ref time cuts are better than Winter2018 cuts

Sameer

------

KaonLT CoinTime Blocking

- plot of correction for Q2=0.5 vs CoinRate for high and low epsilon data

  - both show a nice trend vs CoinRate up to 14kHz, values 0.95->0.99

- had a meeting w/Nathan

  - Nathan: you should use the same time window for a run period, not

    run-by-run as this leads to over-fitting

  - i.e. the time window should only change when the DAQ trigger configuration

    is changed, otherwise combine the runs together to get smaller

    uncertainties

- now has RedMine access, will upload slides for this and previous meetings

Richard

-------

KaonLT high Q2 LT-sep

- implementing t-shift as calculated with Garth's code

  - looking at how this changed things

  - taking this opportunity to refine and simplify analysis code

  - looking also at pion background fits, some of the background uncertainties

    were estimated incorrectly as too large

- Nacer: did you change the t-binning after the t-shift?

  - No

    - Garth: it might be needed to change the lower limit of the 1st bin, since

      events near -tmin will move to higher values of -t

      - yes, this could be an issue for Q2=3.0 W=2.32 setting

Chi Kin

-------

- has Garth's t-shift code (it's posted on KaonLT RedMine)

  - went through how the code worked

  - Dave: what assumption did you have to make?

    - Garth: assume no changes to electron beam and HMS, since that would

      change Q2

      - assume all changes are on meson arm, recalculate meson momentum and

        angle for the shifted MM in order to conserve energy, momentum and then

        recalculate t

Friday: Present

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

Regina - Garth Huber, Alicia Postuma, Nacer Hamdi, Nathan Heinrich,

   Muhammad Junaid, Nermin Sadoun

JMU - Gabriel Niculescu, Ioana Niculescu

Ohio - Julie Roche

FIU - Pete Markowitz

York - Stephen Kay

Glasgow - Rachel Montgomery

Virginia - Richard Trotta

CUA - Chi Kin Tam

JLab - Dave Gaskell (at end)

Alicia

------

Q2=3.0 W=3.14 u-channel omega analysis

- u (or t?) binning study

  - forms MM plots of background subtracted omega region

  - trying 3 u-bins: bin1 0<-u<0.15   bin2 0.15<-u<0.25   bin3 0.25<-u<0.50

  - about 2000 omega events in each bin

  - each u-bin is divided into 8 phi bins

Next steps:

- Bill had binned in t instead of u, need to have a meeting with him to discuss

  this

  - t-binning is not quite equivalent to u-binning, due to relationship between

    s,t,u

- will try finer u-binning at low -u

- will try binning in t instead of u (in progress)

- will check u-u_true vs t-t_true in SIMC

- effect of omega radiative tail in SIMC

  - Bill had removed all t-phi bins from analysis where the RadOn/RadOff omega

    ratio was large

  - will investigate using SIMC and see which bins should be excluded

- then need to extend shape study to each (u,phi) bin

Gabriel:

- there is a statistical method on how to deal with bins where the background

  subtracted yields are negative

- please see https://arxiv.org/abs/physics/9711021

- ROOT can handle this method (asymmetric error bar, with lower limit zero),

  but need to select the correct option

Junaid

------

PionLT Q2=3.85 W=2.62 LTsep

- implementing t-shift offsets into analysis scripts

  - added an extra column to CSV file of corrections to data

  - will look at t-binning statistics after shift is applied

Nathan

------

PionLT Q2=3.85 W=62 cross-checks w/Junaid

- got data yields from Junaid for direct comparison

- then will do LTsep fit using his fit parameters and compare results

- Garth: suggests to start replay of next setting in parallel

  - expects to start on that next week

Nermin

------

PionLT LD- analysis

- checked run-dependence of NGC calibration

  - grouped runs together in different ways, generally consistent

- NGC after calibration with PCAL>0.85 and PCAL<0.85 cuts show no significant

  difference in NPE response

  - clearly not effective for this setting

  - Nathan and Garth: we recall low NGC gains early in PionLT and amplifiers

    were added at some point

    - Junaid looks at the 2021 Run Status Page:

      https://hallcweb.jlab.org/wiki/index.php/PionLT_2021_Physics_Status

      - the Misc Issues for Analysis section has the note:

        NGC not useful Run < 12066

	https://logbooks.jlab.org/entry/3906264

	https://logbooks.jlab.org/entry/3906271

   - Nermin is looking at runs before this, so it explains why NGC not useful

     here

   - Note: there is also the 2022 Run Status Page:

     https://hallcweb.jlab.org/wiki/index.php/PionLT_2022_Physics_Status

     *NB* it is important for Nermin and Kathleen to pay attention to what is

     posted on these pages

Rachel

------

PionLT LD+ analysis

- Kathleen could not attend, but sent slides that Rachel discussed

- SHMS HGC cut

  - *NB* Garth: if an HGC cut is required to get a clean pion sample, then it

    is needed to place a cut around the inefficient light collection region

    neaar the middle of the detector (where the 4 mirrors come together)

  - need to make a 3D plot of X_hgc, Y_hgc vs NPE

  - Junaid: shows some example plots showing the inefficient region

  - Stephen: here are some old histo definition I found quickly, just filled

    with that X/Y and NPE, that will be useful for Kathleen and Nermin

    TTreeReaderArray<Double_t> P_hgcer_xAtCer = {fReader, "P.hgcer.xAtCer"};

    TTreeReaderArray<Double_t> P_hgcer_yAtCer = {fReader, "P.hgcer.yAtCer"};

    h3SHMS_HGC = new TH3F("SHMS_HGC","SHMS HGC Distribution of NPE in X-Y Mirror Plane;X Position (cm);Y Position (cm);NPE",100,-50,50,100,-50,50,100,0.1,35);

- SHMS NGC cut

  - this is for a run number where NGC should be seful

  - *NB* don't use a cut on pi+ data, can see pi+n MM peak extending to high

    NPE that would have to be corrected for if you cut them out

    - pi- runs will likely have no choice but to apply such a cut to remove

      electrons, in this case you will need to use the corresponding pi+ data

      to determine (under the assumption that pi+n and pi-p tails are the

      same) the correction to apply to pi-p data because of the NGC cut

- MM plots shown for progressive cuts applied

  - *NB* Gabriel: you need to apply the most effective cut (removing the most

    events) first, and the other cuts later

  - Junaid: first apply acceptance cuts, then CoinTime cut

    - other cuts after these (RF, Aerogel, HGC) should be applied in different

      orders, so you can see their effect on the plots, it's difficult to see

      the effect of a given cut if it's masked by another cut

- question about what acceptance cuts to apply

  - *NB* Nathan: the acceptance cuts are tied to the HMS, SHMS matrix elements

    that are used

    - each ME set should indicate the region of reliable magnetic optics, you

      must use that as the criterion, not the edges of the distributions

    - the values in the Hall C manual are generic, not specific to the used ME,

      this is particularly important for settings where the HMS or SHMS have

      high momentum

    - Vijay's acceptance cuts should be okay for his data, but his settings

      were at low momentum, the region of reliable optics can be different at

      high momentum

Next Meetings

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

- Thurs: Mar 19 @ 16:00 Eastern/14:00 Regina

  - PionLT will go first

- Fri: Mar 20 @ 11:00 Eastern/9:00 Regina

  - we will continue where we left off