Comparison with Pascal Pralavorio's Studies
Single electron efficiencies and jet rejection
Following the work of Pascal Pralavorio (see ATL-PHYS-99-015), I reproduced his efficiency and jet rejection values with additional timing information. For these studies I used the official ATRECON CVS/SRT release 0.0.33 with modifications to include ATRIG.
For these studies I use single electrons and dijet events with and without pileup. The electronic noise is switched on with the CNOI datacard, where I use a noise threshold of 0 sigmas. The efficiencies are specified as percentage of reconstructed electrons to generated electrons on the given trigger step. In parenthesis the efficiency compared to the previous step is given. The background rejection is calculated in respect to the generated jets with abs(eta)<2.5 and pT>17GeV. In parenthesis the rejection compared to the previous step is specified. The rejection does only take into account the hadrons and gammas, while the expected trigger rate also includes electrons originating from W, prompt photon production, quark bremsstrahlung, etc, generated according to their cross-sections. The reconstruction time is measured in seconds on the ATLAS shift machines using the extended timestamps. The three times given are needed to reconstruct 50%, 80%, and 95% of the events which passed LVL2.
The timing distributions can be obtained by clicking on the timing values. For the ECAL all events and the events passing LVL2 are shown. The quantiles are given for the events passing LVL2. For xKalman in the upper plot all events and the events passing LVL2 are shown. The quantiles are calculated for all events, while in the lower plot the quantiles (and distribution) of events passing LVL2 are given.
Lolumi: 20GeV pT electrons vs. dijets with pT>17GeV
For the low luminosity studies, i.e. without pile-up, I used 1250 single electrons from tape Y20337.295 produced with DICE 98_2, and 18561 jet samples from tapes Y00341.1-44, Y00384.1-49 (DICE 97_6).
| Trigger step | Electron efficiency | Jet rejection | |||||
|---|---|---|---|---|---|---|---|
| Pascal's value | my value | Recon. time 50/80/95% | Pascal's value | my value | Trigger rate (Hz) | Recon. time 50/80/95% | |
| LVL1 (EM20I) | 94.0% | 92.6% | n/a | 80 | 90 | 5850 | n/a |
| LVL2 (e20i) | 82.5% (87.7%) | 82.0% (88.5%) | n/a | 3500 (44) | 3900 (42) | 140 (42) | n/a |
| EM calo Etmin=5GeV | 80.9% (98.1%) | 80.4% (98.0%) | 0.44/0.46/0.49s | 9800 (2.8) | 7400 (1.9) | 85 (1.6) | 0.54/0.66/0.89s |
| ID (xKalman) ptmin=5GeV | 77.4% (93.8%) | 76.6% (95.3%) | 0.04/0.07/0.15s | 16800 (1.7) | 13400 (1.8) | 57 (1.5) | 0.6/1.5/7.7s |
| Matching | 75.4% (97.5%) | 75.2% (98.1%) | n/a | 40000 (2.4) | 25300 (1.9) | 41 (1.4) | n/a |
Hilumi: 30GeV pT electrons vs. dijets with pT>17GeV
For the high luminosity studies I used 2622 single electrons with 30GeV pT where the pileup is added with the new pileup method which are found on tapes Y01565 to Y01568 (DICE 98_2). The 18012 background events were produced with the DICE 97_6 and have been redigitised.
| Trigger step | Electron efficiency | Jet rejection | |||||
|---|---|---|---|---|---|---|---|
| Pascal's value | my value | Recon. time 50/80/95% | Pascal's value | my value | Trigger rate (Hz) | Recon. time 50/80/95% | |
| LVL1 (EM30I) | 96.1% | 94.4% | n/a | 90 | 90 | 21600 | n/a |
| LVL2 (e30i) | 82.5% (85.6%) | 80.7% (85.5%) | n/a | 3700 (41) | 4200 (46) | 470 (46) | n/a |
| EM calo Etmin=5GeV | 81.1% (98.3%) | 79.0% (97.9%) | 0.45/0.51/0.64s | 8400 (2.2) | 6700 (1.6) | 315 (1.5) | 0.62/0.79/1.02s |
| ID (xKalman) ptmin=5GeV | 77.2% (93.6%) | 75.1% (95.0%) | 8.2/17.0/39.4s | 22700 (2.7) | 17500 (2.6) | 157 (2.0) | 20.0/60.1/114.9s |
| Matching | 75.3% (97.4%) | 73.4% (97.8%) | n/a | 35800 (1.6) | 22700 (1.3) | 134 (1.2) | n/a |
This table uses the cuts by Pascal. However, there is an error in the numbers he gives for LVL2 ecal (see email of Monika Wielers).
