Star-tpc-l mailing list

["Fisyak, Yuri V" <fisyak AT bnl.gov>]

Hi Rongrong, by “electron is a special beast” I mean comparison measured (log) dE/dx for e+/- identified by unique kinematically reconstructed γ conversions from the model. There is momentum dependence which you converted in eta one.

https://www.star.bnl.gov/~fisyak/star/PiD/dAu200_P23id/mudEdxqeg3sGraph_pdEdx_dAu200_P23id.dEdx.png

“What puzzles us is that nsigma_e is more Gaussian in the old (SL21c) production than in the new production (SL23d). “    My guess is that in SL23d the resolution is better than in SL21c, and this is the reason why you do see more non Gaussian.

 

 

 

 

                                                              Yuri Fisyak

STAR                                           Phone: +1 631 344 3913
Brookhaven National Laboratory  Fax:    +1 631 344 4206
510A/1-161
http://www.star.bnl.gov/~fisyak       E-mail: fisyak AT bnl.gov

 

 

 

From: Ma, Rongrong <marr AT bnl.gov>
Date: Tuesday, December 5, 2023 at 12:01 PM
To: Fisyak, Yuri V <fisyak AT bnl.gov>
Cc: skradhakrishnan AT lbl.gov <skradhakrishnan AT lbl.gov>, Van Buren, Gene <gene AT bnl.gov>, geurts AT rice.edu <geurts AT rice.edu>, Barbara Trzeciak <barbara.trzeciak AT gmail.com>, 张炜 <wzhang AT m.scnu.edu.cn>, Mooney, Isaac <isaac.mooney AT yale.edu>, Yi Yang <yiyang AT ncku.edu.tw>, nihar AT rcf.rhic.bnl.gov <nihar AT rcf.rhic.bnl.gov>, xuyue <xuyue AT impcas.ac.cn>, subhash <subhash AT rcf.rhic.bnl.gov>, Tribedy, Prithwish <ptribedy AT bnl.gov>, Zhenyu Chen <zhenyuchen AT sdu.edu.cn>, Yiding-Han AT rice.edu <Yiding-Han AT rice.edu>, Star-tpc L <Star-tpc-l AT lists.bnl.gov>
Subject: Re: Some issues with SL23d and SL23e productions

Hello Yuri

 

When Wei looked at the eta dependence of the mean nsigma_e, he saw a non-flat distribution, which prompted him to shift the nsigma_e distribution by hand to make <nsigma_e> sit at 0 in all eta bins. It will make it easier to cut on nsigma_e for PID and calculate its efficiency. This is what "corrected" means, which is essentially a shift of nsigma_e distribution vs. eta. 

 

What puzzles us is that nsigma_e is more Gaussian in the old (SL21c) production than in the new production (SL23d). 

 

Sure, we can discuss this at the TPC meeting this Thursday. 

 

Best

Rongrong

On Dec 5, 2023, at 11:49 AM, Fisyak, Yuri V <fisyak AT bnl.gov> wrote:

 

1.       Non Gaussian nsigma_e distributions (https://drupal.star.bnl.gov/STAR/system/files/nsigmaeCorr.pdf):

a.       I don’t know what do you call “corrected” nsigma_e. 

b.       I know that with new dE/dx model I have momentum dependence overestimation for electrons with respect to the model up to 10%. The electron dE/dx is a special beast and it requires a special treatment. 

c.        Log(dE/dx) has much better match with Gaussian than dE/dx (which is Landau distribution with infinite dispersion) due to the finite dispersion and the Central Limit theorem, but it is not exactly Gaussian. How close it to Gaussian depends on resolution and statistics.