Run 13 BEMC Calibration : Spin PWG presentation-01/18/2016
Updated on Tue, 2016-05-10 15:00. Originally created by devika on 2016-01-19 13:05.
http://www.star.bnl.gov/protected/spin/devika/BEMC_calib_spin_PWG/run13-BEMC-PWG-summary-Final.pdf
Questions / comments from spin PWG meeting:
(1) Renee:
- Nice that calibration agrees between low energy and high energy probes
- Cluster correction: Slide 21
- Explain procedure concerning isolation ratio
- Slide 17: Are momentum cuts applied? No
- Has E/p behavior changed before? Issues still remain!
- Show more details that distributions on slide 17 are consistent (JP2 / BHT3):
Answer: As long as there is no BHT1 cut, average gain is the same between JP2 and BHT3 considering the full momentum range. With momentum cut, you introduce bias.
Is this the best way to estimate systematic uncertainties? How can we access an unbiased distribution!
Answer:
- MIN-BIAS distribution: What statistical precision do you have?
Answer: only MB trigger in the luminosity page is ,https://www.star.bnl.gov/protected/common/common2013/trigger2013/lumipp500GeV/lum_pertriggerid_pp2013_500GeV.txt
- Slide 47:
(3) Carl:
- Slide 21: Why does ratio not go to 1? Use 3X3 vs. 5X5?
- Why does the behavior look the way it looks for E/P (Cluster) vs. Tower Isolation Ratio!
2x2cluster_method_and_isolation_cuts
JInlong : https://drupal.star.bnl.gov/STAR/system/files/TriggerBiasandClusterMethod.pdf
First presentation of run13 BEMC Calibration to SPIN PWG
Presentation :http://www.star.bnl.gov/protected/spin/devika/BEMC_calib_spin_PWG/run13-BEMC-PWG-summary-Final.pdf
Questions / comments from spin PWG meeting:
(1) Renee:
- Nice that calibration agrees between low energy and high energy probes
- Cluster correction: Slide 21
- Explain procedure concerning isolation ratio
- Slide 17: Are momentum cuts applied? No
- Has E/p behavior changed before? Issues still remain!
- Show more details that distributions on slide 17 are consistent (JP2 / BHT3):
Answer: As long as there is no BHT1 cut, average gain is the same between JP2 and BHT3 considering the full momentum range. With momentum cut, you introduce bias.
Is this the best way to estimate systematic uncertainties? How can we access an unbiased distribution!
Answer:
- MIN-BIAS distribution: What statistical precision do you have?
Answer: only MB trigger in the luminosity page is ,https://www.star.bnl.gov/protected/common/common2013/trigger2013/lumipp500GeV/lum_pertriggerid_pp2013_500GeV.txt
BHT1*VPDMB 61.635 282.985 14077042 14161020 1.729 16.778 4.634 21.122and we alrady use this.. this is what we called BHT1-trigger
- Slide 47:
(2) Elke:
- Comparison to W/Z: Lower energy probes such as neutral pion decays or other known masses?
Answer:
- Result from tracking task force: How sure are you that features are not tracking related? Systematics of momentum reconstruction?
Answer:
TPC P resoultion simulation study ,, TPC global
- Comparison to W/Z: Lower energy probes such as neutral pion decays or other known masses?
Answer:
- Result from tracking task force: How sure are you that features are not tracking related? Systematics of momentum reconstruction?
Answer:
TPC P resoultion simulation study ,, TPC global
(3) Carl:
- Slide 21: Why does ratio not go to 1? Use 3X3 vs. 5X5?
- Why does the behavior look the way it looks for E/P (Cluster) vs. Tower Isolation Ratio!
2x2cluster_method_and_isolation_cuts
JInlong : https://drupal.star.bnl.gov/STAR/system/files/TriggerBiasandClusterMethod.pdf
(4) Will: http://www.star.bnl.gov/HyperNews-star/protected/get/emc2/4391/2.html
Yes, I agree that we haven't really solve this. But as you know we have done some studies to see that if this has something to do with the track momentum , hence the cause of this drop is at the tracking level. [simulation study for TPC P resolution , DCA cuts]. Those studies never support strong evidences as to why the E /p drops occur. If this is not caused by track momentum at least not the major contribution then it has to come from the calorimeter. By looking at the plots in backup slides 47, 48 isn't that obvious that this has to do something with the trigger bias (thresholds) ? [ since BHT3 has no lot of statistics above the trigger threshold , isn't this caused by BHT1 trigger ? ] Something happens when BHT1 trigger turned on. This is why the requirement place on BHT1 trigger , to be considered or not , play a major role and you see E / p drop vanish in slide 48 red points when asked not to include BHT1 trigger.
If the main cause of E / p drop is trigger effect then one possible way to understand this is to a do a simulaton study of triiger effect.
--we have talked alot about this but the new approach
No I havent did any weighting, only took average. Namely fill histograms using whole momentum region. One could see E /p drop start at after P = 5 GeV. But if you staty away one could say what matters is P > 6 GeV. Now there is no impact from E / p drom as lon as the low mometum region [ P < 6 ] is used. So I would say there is no impact from E /p drop. There is no change in <E /p > including or not including the statistics from P > 6 GeV, so no impact from E /p drop , again as long as no special reguirement from BHT1 trigger.
---I'd guess mechanism drives selected
events toward the 2x2 cluster of choice. Thus one has
on pg. 23, a strong dependence on TDR cut ... how does
one justify picking the "right" value (add systematic)?
Answer:
We need to add the variation to the systamtic. It will be < 1%.
---pgs 25-28 (and re: also pg. 18) ... we haven't discussed much the 2 outer rings and how to handle. Given the variation of E/p with p, how can one justify selection of a given p cut?
Yes I agree. but also need to consider the fact that low S / BG of outer rings. One need to put some P cut to bring S / BG higher in the forward eta.
1) as brought up by Renee at yesterday's SPWG: triggered spectra & E/p ... gross behavior in 2D plts pg 17, but really haven't addressed the strong E/p vs. p behavior e.g., as seen in backup pg 47.
AnswerYes, I agree that we haven't really solve this. But as you know we have done some studies to see that if this has something to do with the track momentum , hence the cause of this drop is at the tracking level. [simulation study for TPC P resolution , DCA cuts]. Those studies never support strong evidences as to why the E /p drops occur. If this is not caused by track momentum at least not the major contribution then it has to come from the calorimeter. By looking at the plots in backup slides 47, 48 isn't that obvious that this has to do something with the trigger bias (thresholds) ? [ since BHT3 has no lot of statistics above the trigger threshold , isn't this caused by BHT1 trigger ? ] Something happens when BHT1 trigger turned on. This is why the requirement place on BHT1 trigger , to be considered or not , play a major role and you see E / p drop vanish in slide 48 red points when asked not to include BHT1 trigger.
If the main cause of E / p drop is trigger effect then one possible way to understand this is to a do a simulaton study of triiger effect.
--we have talked alot about this but the new approach
to take an "average" (I guess of E/p slice extracted
pts -- or are you effectively/artificially weighting
by the momentum distribution) must lead of course to a
fair uncertainty on the central value because of the
variation of E/p points w.r.t. momentum.
Answer: No I havent did any weighting, only took average. Namely fill histograms using whole momentum region. One could see E /p drop start at after P = 5 GeV. But if you staty away one could say what matters is P > 6 GeV. Now there is no impact from E / p drom as lon as the low mometum region [ P < 6 ] is used. So I would say there is no impact from E /p drop. There is no change in <E /p > including or not including the statistics from P > 6 GeV, so no impact from E /p drop , again as long as no special reguirement from BHT1 trigger.
---I'd guess mechanism drives selected
events toward the 2x2 cluster of choice. Thus one has
on pg. 23, a strong dependence on TDR cut ... how does
one justify picking the "right" value (add systematic)?
Answer:
We need to add the variation to the systamtic. It will be < 1%.
---pgs 25-28 (and re: also pg. 18) ... we haven't discussed much the 2 outer rings and how to handle. Given the variation of E/p with p, how can one justify selection of a given p cut?
Yes I agree. but also need to consider the fact that low S / BG of outer rings. One need to put some P cut to bring S / BG higher in the forward eta.
»
- devika's blog
- Login or register to post comments