Feb 15/24 PionLT/KaonLT Analysis Meeting Notes
----------------------------------------------
(Notes by GH)
Today: KaonLT 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, Ali Usman, Muhammad Junaid, Nathan Heinrich,
Alicia Postuma, Nacer Hamdi
JLab - Dave Gaskell
York - Stephen Kay
CUA - Casey Morean, Richard Trotta, Tanja Horn
FIU - Pete Markowitz
CSULA - Konrad Aniol
Richard
-------
Updated Efficiencies
- needed to fix used HMS Cherenkov and Calorimeter Effs in Effective Charge
script
Now using: hcer eff=0.9730 unc=0.0006
hcal eff=0.9961 unc=0.0005
Fixed issue with Data/MC average ratios script
- was over-writing ratio values w/ just ones from Center Setting in
average_ratios.f
- this means Left and Right settings were not included
- of course this makes a big difference to the phi-coverage at higher -t
- a short discussion followed on the optimal number of t,phi bins to use, how
many is too many?
- the important issue is to have sufficient statistical significance to
distinguish a potentially small sig_L from zero
- From the online KaonLT statistics, GH had estimated #t-bins (K+Lambda):
Q2=0.50 x=0.09 5 t-bins
2.115 x=0.21 5 t-bins
3.00 x=0.25 5 t-bins
3.00 x=0.40 2 t-bins
4.40 x=0.40 2 t-bins
5.50 x=0.40 2 t-bins
This is only an approximate guide, might need fewer bins to have good
L/T-separation
- using 10 phi bins now. Seems appropriate. pi+ channel can probably use 16
bins.
Corrections to calculation of statistical uncertainties
- had a long discussion w/ Ali, a second meeting with GH
- reviewed equations from T.Horn and W.Li theses
- for Q2=2.115 delta-R_bin: 5-10%, and dY_data_bin~dY_simc_bin
- surprised SIMC errors aren't smaller, given #evnts generated, will
investigate further
Aerogel/HGC cuts in SIMC
- aerogel tray cut applied in recon_hcana, while HGC cut is applied in
lt_analysis
- the HGC cut is arranged this way to be more consistent w/ data analysis
- comparing data & SMC x,y_HGC plots w/ hole cut
- sees a 10cm X-shift in HGC cut compared to Vijay
GH: perhaps RT and VK are projecting to different z_HGC locations?
Pion/proton leakage subtraction script
- GH: need to use pi+ data instead of simulation, since the K+L peak lies above
the Delta0pi+ and inclusive region, pi+n simulation would only include
radiative tail and none of these other contributions, would then have to run
multiple SIMC generators to get all of the contributions, making for a much
more complicated fit
- This is even more important for K+Sigma extraction!
Unseparated XSect systematic studies
- met with TH to discuss systematic error table
- nice table listed on what cuts to vary for each systematic study
Tracking: compare pruning and chi-square methods
Acceptance: vary acceptance cuts
PID: compare high, low epsilon HGC hole cuts
Kinematics: vary angle, momentum, energy offsets
MC Model: vary fit model parameters from final iteration values
Radiative Tail: SIMC w/ and w/o rad flag
Issues w/ generating new parameterization
- now that L,R,C ratios issue fixed, things starting to look good
- initial indications is that t-dependence in the model is wrong
- will try removing t-dependence to sigT
Ali
---
Optimizing SIMC MM resolution
- high epsilon 10.6GeV data have worst resolution mismatch
- using data dists from most recent full pass replay, incl TOF-CT leakage
correction and offsets
- SIMC also offset corrected
- initial plots only for SIMC resolution tuning
- Dave's delta correction not included
- shows plots for central settings only, expect L,R to be very similar
- one SHMS resolution tuning parameter used for all kinematics
- Q2=3.0 W=2.32 HMS=6.58 - mismatch is fairly large here
- Q2=5.5 W=3.03 HMS=3.26 - best resolution agreement
- shows t-binned Q2=5.5 plots from Alicia
- overall, the agreement is better now
- definitely see a t-dependence to the discrepancy, so the t-binned plots
will be vital for determining a final correction
- added HMS DC resolution correction from Abhyuday
- didn't see any significant improvement for Q2=3.0 W=2.32 center
- DG: how much is the SHMS resolution changed by?
- a multiplicative factor is used in SIMC, 9x
- seems very high!
- Nacer: asks some questions about the pi+n SIMC model
- GH: it would be valuable to check 1-2 low epsilon settings where P_HMS is
lower
- Nacer: any final discrepancy between data and MC agreement should end up
being quoted as part of the Model systematic error
- GH agrees with this
Alicia
------
Recap of BSA analysis status
- need to implement Dave's delta-correction and troubleshoot
- then rerun BSA calc, mean kin, cut dependence scripts
- will arrange a separate future meeting to discuss error propagation
- then finalize Q2-scans at fixed xB including Hall B data
- goal of distributing draft paper by end of March
Troubleshooting delta-correction
- delta_corr=hsdelta+6.0*hsxpfp
- Problem: finding that MM resolution w/o correction is worse than from hcana
- DG: to do a quick check, used an old script which calculated MM like old
6GeV ENGINE
- should use Richard's recon script instead, which has same MM formulas as
hcana
- the issue is that hcana has improved corrections for xtar dependence
compared to ENGINE, since longer target lenghts are used in 12GeV running
compared to 6GeV
- sees an improvement when the delta-correction is applied, but will need to
redo the study due to different MM calc than hcana
- Ali will also have to retune SHMS resolution study after Alicia is done her
study
- Alicia's t-binning plots at Q2=4.4 shows that Ali's 9x resolution factor is
too big
- would like SIMC to never be wider than data, for any t-bin
Nacer
-----
DC calibs
- pushed changes to GitHub
- Junaid should check things over before accepting
- maybe Ali can look things over as well, as he's more familiar with KaonLT
layout
- hoping to submit full replay over weekend
Nathan
------
- catching up on things after being RC
- BCM param map now set up
- hoping to have some Lumi plots next week
- will start with the robust singles study from end of PionLT run
Junaid
------
Tracking Eff study
- plots of EDTM, Elec_sing_TrkEff, Elec_COIN_TrkEff, Hodo3/4Eff for HMS
- a small number of outlier runs (eff=0) that need checking, but overall
looks good, with reasonably tight, monotonic trends vs rate
- similar plots for SHMS
- also a few outliers that need checking
- pTRIG6 runs for HMS also shown
- these were for the ELREAL trigger efficiency study
- HeepCoin data plots
- plot of SHMS >1MHz Hodo rate shows sharper drop off of eff at very high
rate
Next Meeting
------------
- Thur Feb 22 @ 15:00 Eastern/14:00 Regina
- PionLT will go first