GSI Forum
GSI Helmholtzzentrum für Schwerionenforschung

Home » PANDA » PandaRoot » Analysis » entries at unphysical mass after vertex fit (lambda vertex fit with PndKinVtxFitter results in better fwhm but large tails in Lambda mass)
entries at unphysical mass after vertex fit [message #17625] Thu, 04 December 2014 10:17 Go to next message
Albrecht Gillitzer is currently offline  Albrecht Gillitzer
Messages: 49
Registered: March 2004
continuous participant
From: *ikp.kfa-juelich.de
Hi all,

When comparing the Lambda mass spectrum evaluated before and after the vertex fit using PndKinVtxFitter, one can see that the fwhm of the Lambda mass distribution improves by the vertex fit but in contrast larger tails in the mass distribution are created. On the low mass side the tails extend far into the unphysical region at masses below the sum mass of the daughter particles M_sum = m_pi + m_p.
The attached plot has been generated from 10000 events 4 GeV/c pbar p --> Xibar+ Xi- (PHSP) --> Lambdabar pi+ Lambda pi- --> pbar pi+ pi+ p pi- pi- with the oct14 release. Upper row: Lambda mass before vertex fit / Lower row: after vertex fit // Left: lin y axis / Right: log y axis.
The corresponding plot for Lambdabar shows the same feature. Simulation and analysis of the same reaction with the scrut14 release also shows the same feature.

My interpretation: Can it be that in the vertex fit the 3-momenta of the daughter particles are modified according to the modified track parameters as a result of the fit, but their energies are left unchanged? As a consequence the energy momentum relation would be destroyed and particles no longer would have the correct mass value.

Albrecht
Re: entries at unphysical mass after vertex fit [message #17630 is a reply to message #17625] Thu, 04 December 2014 11:49 Go to previous messageGo to next message
Ralf Kliemt is currently offline  Ralf Kliemt
Messages: 507
Registered: May 2007
Location: GSI, Darmstadt
first-grade participant

From: *gsi.de
Hallo Albrecht,

Firstly I suggest you do a crosscheck with the other vertex fitter PndVtxPRG. From my understanding there you don't alter the momentum magnitude and energy of the daughter particles.

Secondly, you may want to study the fit quality (chisquare) for the unphysical candidates.

Kind regards.
Ralf
Re: entries at unphysical mass after vertex fit [message #17631 is a reply to message #17625] Thu, 04 December 2014 11:59 Go to previous messageGo to next message
Ralf Kliemt is currently offline  Ralf Kliemt
Messages: 507
Registered: May 2007
Location: GSI, Darmstadt
first-grade participant

From: *gsi.de
Hi again,

Please plot the masses of the daughters (pions/protons) before and after the fit. Those should be exact needles.

Cheers
Re: entries at unphysical mass after vertex fit [message #18422 is a reply to message #17625] Mon, 17 August 2015 15:16 Go to previous message
Jennifer Pütz is currently offline  Jennifer Pütz
Messages: 47
Registered: April 2015
Location: FZ Juelich
continuous participant
From: *ikp.kfa-juelich.de
Hi everyone,

I found the same probelm as Albrecht. I ploted the difference of the masses of the daughters before and after the fit and saw that they are not exact needles. So I had a closer look into the PndKinVtxFitter and found that the four-momentum of the dauhgters after the vertex fit is set to (p_vtx, E_vtx). p_tx and E_vtx are the 3-momentum-vector and the energy of the daughter particle comming out of the vertex fit.
But E_vtx is not taking the mass hypothesis of the dauhgter particle into account. So I replaced E_vtx with sqrt(p_vtx^2 + m^2). (This was already implemented in the code but commented out.)
After doing that, the unphysical masses for Lambda0 have disappeared and the mass differences show now exact needles (see attached files).

Best,

Jenny


[Updated on: Mon, 17 August 2015 15:16]

Report message to a moderator

Previous Topic: Invariant Mass Distribution Widening after 4C Fit
Next Topic: Using Rho without MC info
Goto Forum:
  


Current Time: Thu Nov 30 01:59:51 CET 2023

Total time taken to generate the page: 0.00907 seconds