Doyeong
Summary:
- Compared the distributions of individual cell time and average cell time in bins of pion E ([1, 5, 10, 20]) using charged pion (pi- for now) and neutral pion.
- First look of pi0 sample with PU.
Raised questions:
- Saw clear double peaks in pi- w/o PU (ex. slide 4). Also, cell time distributions somehow look like Cauchy distribution rather than Gaussian.
- In pi0 w/o PU, there are enhanced contributions in cell time distribution on the right-hand side tail (slide 18).
- Average cell time distributions don't agree with the naive expectation from 1/sqrt(N).
- By looking at the cell time distributions of pi-,0 w/o PU without any cuts (slide 23), they don't really look like Gaussian.
What we've seen can be real physics (pi → n, and n decays later), or some anomalous things going on due to the particular phase space region of the detector which mess up the entire average distribution. Therefore, we'll study more by looking at the following.
- RMS vs #cell plots in different truth pion pT windows ([1, 5, 10, 20])
- Check kinematic distributions in different detector geometry (different eta, layer, etc.) and also different cell time ranges (tail (non-Gaussian) and center (Gaussian)).
- Check if E resolution behaves the same:
- if E resolution is getting better → issue is in the measurement of timing
- if E resolution is getting worse → can be something from a particular region of phase space (need to check...)
- In addition, the general comments are the following:
- Add significance to the Ntuple (URGENT) and replace reco cell E > 0.1 GeV cut
- truth pion E cut → truth pion pT cut
- Use the same x-axis range & binning for both individual cell time and avg. cell time
- slide 4: check if I can see the double peak in individual cell time plot as well
- About PU samples, we need to increase reco cell E cut (currently 0.1 GeV, but we can probably replace this to significance very soon). W/ and w/o PU samples will be studied together.
- High E cut motivated from w/ PU sample can be tested in w/o PU sample.
Zahra
Studied the response with different cuts on cell level and pion level. Also checked response (1. truth E vs reco E 2. reco E/truthE vs reco E) in bins of different timing cuts. It was hard to tell how the response changes in the 2D distributions, so we'll check 'reco E - truth E' in different bins of time.
