Kaon_LT Experiment

                 Feb 9/23 PionLT/KaonLT Analysis Meeting Notes

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

                           (Notes by GH and SJDK)

                    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 - Stephen Kay, Ali Usman, Alicia Postuma, Nathan Heinrich, Garth Huber,

   Vijay Kumar, Muhammad Junaid, Love Preet

JLab - Dave Gaskell

CSULA - Konrad Aniol

CUA - Tanja Horn, Richard Trotta

Nathan Updates

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

Result of detector time and reference time cut analysis

- Q2 = 3.85, W = 2.62, Left, Low eps (LH2+), whole setting

   - With new cuts, fine ~200 more pions compared to online

   - ~2.5% more events (8261 vs 8070)

   - Distributons look slightly better, very subtle changes

- DG - Probably worth checking a setting with a poor real/random ratio

   - May see a more dramatic difference

   - Try a pi- run

   - Nathan has it, but no online plots for comparison

      - Q2 = 3.85, W= 2.62, low eps, left

      - https://logbooks.jlab.org/files/2022/09/4039344/Q3p85W2p62left_lowe_LD-_Pion_Analysis_Distributions.pdf

      - 4806 pions

      - Looks like we have less now

         - Could be that some runs were missed (setting was split in two)

         - Please do a more thorough comparison for a pi- run

         - REMEMBER: All data are guilty until proven innocent!

- Pushed changes to param files, std.kin

   - not everything is replayed yet

   - ref time cuts stable so this part should be okay now

   - Next step is Cherenkov calibs

Junaid Updates

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

Back from Pakistan

- Trying to catch up on classes/get back on track with analysis

- DC and calorimeter calibration is next step

Richard Updates

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

Changes to PM/EM calcs in SIMC

- For SIMC/hcana comparison, complicated doing it in SIMC

- Use a script to convert root file

   - Made this and ran it

   - Nothing looks different?

   - New variables look the same as the old ones

      - DG - Seems very strange

      - RT will meet with DG tomorrow about this

Weird Right setting from last week

- tightened up CoinTime cuts

- CT peaks slightly offset, will correct for at some point

- Dramatically imrpoved MM distributions

- Black is right, green centre, red left

   - Black looks much improved

   - Still a bit high, but now only by ~50% instead of 10x

   - -t distribution no longer going through 0

      - SK - Thought, we never plot 2D t vs anything other than phi

      - Could be interesting to check and remove junk events

      - t vs CT/RF?

- For black (right) CT dist, tails are being removed - real events removed

   - GH - Do we need to correct for this? Cutting real events?

   - DG - Is this a final cut or just a diagnostic?

      - In principle, spectra should look the same with tight cut or with

        random subtraction, not the same indicates a problem that needs to

        be investigated

        e.g. if peaks left/right of real peak are actual physics events from

        other process

- Still need to optimize PID cuts

   - only temporary cuts

   - will have to investigate where e-pi, e-p coins are compared to e-K and

     adjust cuts accordingly

- Maybe cuts not being applied correctly for right setting?

   - Something is clearly dramatically different for the right setting

   - GH: Right setting probably has high SHMS rates, at very forward angle

     Please *look up* running conditions of right setting comp to cent, left

Ali Updates

-----------

HMS Cherenkov and calorimeter efficiency study

- slide nicely lists ALL cuts applied

- Wide range of kinematics investigated in KaonLT data

- Run 4871: 

   - Cherenkov efficiency is very low (94.82%)

   - Cal eff=98.66%

- Run 4919:

   - Cher eff=93.96%

   -  Calorimeter efficiency v.low (90.75%)

   - A lot of pi- contamination present, not a good setting for efficiency study

- Run 4987:

   - Cherenkov efficiency ~97.15%

   - Cal 98.93% - Calorimeter a little low?

- Run 5222:

   - Cherenkov 96.16%, calorimeter 99.18%

   - DG: this looks like a classic case of FADC-DT effect due to HMS-Cer light

     leak

     This is a *real effect* that will have to be included in efficiency

     The complication is that this FADC-DT is in priciple rate dependent and it

     may be tricky to separate out the rate dependence from other inefficient

     effects

   - Looks like we have a light leak in the Autumn 2018 data

   - Could've been introduced when the HMS mirror was fixed

   - Will be there for *all* KaonLT data and Summer 2019 PionLT data, so Vijay

     needs to be careful about this in low Q2 analysis as well

Spring 2019 data efficiencies

- Run 7909:

   - 75.71% Cherenkov efficiency, 96.98% calorimeter efficiency

   - Low total efficiency in low epsilon data - it's all the pions!

- Run 7979:

   - Cer 94.92%, cal 99.09%

   - Calorimeter efficiency is pretty good

   - DG: I would expect a cal eff closer to 99.5%

   - AU: will try applying a neutron MM cut to further clean the sample,

     expecting a further increase of 0.5%, i.e. closer to what DG expects

- Run 8255:

   - Cer 71.66%, Cal 95.58%

   - low stats in a single run, this is high Q2 low epsilon

   - need to add more runs to get better #events

- DG - Suggest to Look at HeeP singles to get a handle on these efficiencies

   - Cut tightly on W2 dist, eliminate pions

      - Should have a nice clean sample

   - GH - Hopefully singles fill more of the focal plane than HeepCoin

   - Most of the physics settings it's hard to get a usable efficiency number

     due to pi- contamination in HMS

   - GH - Focus on 1-2 settings that look promising, otherwise, focus on the HeeP

      - Don't do a run by run efficiency, just use result from HeeP to guide this

Vijay Updates

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

Analysing Summer 2019 Q2=0.38 data with 3 epsilon settings

   - fixed issue with cut file, weird random peak to left of prompt peak is gone

   - Can probably take a lot more peaks on each side of the prompt peak for

     random subtraction

      - At least 2 more on each side

- Diamond cut implemented, needs some small adjustment on top edge in

  comparison to low epsilon data

- Still finalising t-binning

   - ~ 45k events per bin (summed over all SHMS angles and epsilon)

   - 8 t bins

   - GH for first t bin, should move left edge of the bin closer to the data

      - For low Q2 FF test, the dependence at low -t is crucial

      - Should go very narrow on binning at low -t

     - Could go up to 10 bins

   - DG - What is the -t resolution?

      - Can't bin finer than this or t0bin migration will be a big issue

      - GH will discuss with VK how to get this info from SIMC

Alicia Updates

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

- Assignments/presentation updates

- No new updates, will have something after WNPPC

Next Meeting

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

Thur Feb 16 @ 17:00 Eastern/16:00 Regina/15:00 Mtn/14:00 Pacific

- KaonLT will go first

- SK, NH, GH, AP, LP will connect from our hotel room in Banff

- GH, AP, LP might be a bit late