Preliminary plans for LVL2 e ESDAOD

1
Preliminary plans for LVL2 e/? ESD/AOD

• Monika Wielers
• Simon George
• Ricardo Gonçalo
• Note We just try to understand the ESD/AOD
  business, so this is premature and should be seen
  as starting point

2
Aims and constraints

• LVL2 e/? variables currently available in the
  CBNT
• The aim is to store this information at the
  AOD/ESD level need persistified objects
• AOD level
• Should be only minimal info about trigger menus
  satisfied (EF only or L2EF?)
• AOD should be obtainable from ESD
• Worry about it later
• ESD level is the top priority
• LVL2 e/? information would allow trigger studies
  using large data samples produced for Rome
• How to do it?
• First approximation is to use available offline
  objects TrackParticle, EMShowerMinimal
• T2Cluster needs to be clarified, just need ET,
  eta, phi
• Tracks can be converted by existing
  TrigToTrkTrackTool switched on through job
  option must check it out (thanks to John Baines)
• Real data is a separate matter due to
  space/bandwidth need L2/EF fragments of 1kByte
  for real data (i.e. 1MByte/s for 1kHz accept
  rate)

3
CBNT variables tracking - I

• First create TrkTrack and then use
  TrackParticleCreator to produce TrackParticles
• TrkTracks are vectors of TrackStateOnSurface
  which contain TrackParameters, RIO_OnTrack, etc
• author data member (enum) in TrkTrack needs
  to be extended to take LVL2 algos
• However, if we want to join the official
  productions we will have to limit ourselves to
  one (Ketevi) !
• Parameters at closest approach to origin ?
  TrkPerigee
• Must persistify level-2 vertex as a different
  class? TrackParticle has link to VtxRecVertex

Can TrackParticle be used for LVL2 variables?
(Thanks to Ed Moyse)
T2IdAlgo(T2IdNtracks) algorithm that created
track t2IdAlgo1 SiTrack t2IdAlgo2
IDscan t2IdAlgo3 TrtLUT t2IdAlgo4 TrtXk
T2IdPhi0(T2IdNtracks) Phi at closest approach
to origin T2IdZ0(T2IdNtracks) Z at closest
approach to origin T2IdD0(T2IdNtracks) R of
closest approach to origin
4
CBNT variables tracking - II

• Some variable transformations needed to fill
  TrkTracks
• Pt and ? of track need to be transformed into ?
  and q/p
• Must transform errors at closest distance to
  origin into covariance matrix on
  TrkMeasuredPerigee
• TrkTrack has TrkFitQuality class with members
  ?2 and NDoF
• Extrapolator tool exists should we keep T2IdPhic?

T2IdPt(T2IdNtracks) PT T2IdEta(T2IdNtracks)
eta
T2IdErrPt(T2IdNtracks)error on
PT T2IdErrEta(T2IdNtracks) error on
eta T2IdErrPhi0(T2IdNtracks) error on phi0
T2IdChi2(T2IdNtracks) ?2 of track T2IdNDoF(T2IdN
tracks) degrees of freedom
T2IdPhic(T2IdNtracks) phi at calorimeter face
5
CBNT variables tracking - III

• Should these go in TrkTrackSummary ? What if
  wwe need info, which cannot be stored in
  TrackSummary ?
• Still more questions than answers well have to
  look more in detail what offline track offers

• T2IdNstraw(T2IdNtracks) number of TRT straws
• T2IdNtime(T2IdNtracks) number of TRT straws
  with time (not in TrkSum)
• T2IdNtr(T2IdNtracks) number of high threshold
  TRT hits
• T2IdLastPlane(T2IdNtracks) first TRT
  straw-plane on track (not in TrkSum)
• T2IdFirstPlane(T2IdNtracks) last TRT
  straw-plane on track (not in TrkSum)

• Truth information should normally be kept
  separate from tracks, using a link from
  TrackParticle

T2IdNkineHit number of space-points on the
track associated with kine T2IdKineRef GEANT
KINE number for the track T2IdKineEnt entry
number for this particle in the CBNT KINE Block
6
CBNT variables calorimetry

• Store EMShowerMinimal, cluster ET, eta, phi to be
  clarified
• forget about energy per sampling/calo ?

• T2CaNclus Number of LVL2 e.m. clusters
• T2CaEmE(T2CaNclus) Total energy in LAr EM
• T2CaEmES0(T2CaNclus) Energy in Samp. 0 of LAr EM
  
• T2CaEmES1(T2CaNclus) Energy in Samp. 1 of LAr EM
  
• T2CaEmES2(T2CaNclus) Energy in Samp. 2 of LAr EM
  
• T2CaEmES3(T2CaNclus) Energy in Samp. 3 of LAr EM
  
• T2CaHadE(T2CaNclus) Total energy in LAr HEC and
  Tilecal
• T2CaHadES0(T2CaNclus) Energy in Samp. 0 of Had.
  calorimeters
• T2CaHadES1(T2CaNclus) Energy in Samp. 1 of Had.
  calorimeters
• T2CaHadES2(T2CaNclus) Energy in Samp. 2 of Had.
  calorimeters
• T2CaHadES3(T2CaNclus) Energy in Tilecal
  scintillator gap.
• T2CaRcore(T2CaNclus) E3x7/E7x7 in samp.2, with
  Enxm energy in nxm cell window around refined
  LVL1 position.
• T2CaEratio(T2CaNclus) (E1st - E2nd)/(E1st
  E2nd) in samp.1, where E1st and E2nd are the
  energies of the two highest maxima.
• T2CaWidth(T2CaNclus) Width of the e.m. cluster
  in samp. 2 in 3x5 cell window.
• T2CaF73(T2CaNclus) (E7 - E3)/E3 in samp.1, with
  En Energy of n cells around highest maximum in
  samp.1.
• T2CaEta(T2CaNclus) Weighted eta position of the
  e.m. cluster in samp. 2 in a 3x7 cell window.
• T2CaPhi(T2CaNclus) Weighted phi position of the
  e.m. cluster in samp. 2 in a 3x7 cell window.

7
Conclusions and outlook

• Trying to use existing persistifiable offline
  objects to store level 2 e/? info at the ESD
  level
• Concentrating on ESD level for now
• TrackParticle class looks appropriate for ESD
• EMShowerMinimal good for calo shower shapes,
  T2Calo cluster to be clarified
• If we want something for Rome we will need to
  iterate with Alan on LVL1 calo. First bits and
  pieces there (RoIBuilder result), but it wont
  allow to tune Level-1 calo. Missing pieces
  forseen for release 10, so for testing we should
  come up with an interim solution where the final
  persistified class looks like what we expect in
  release 10.
• Well discuss this with Alan before x-mas
• Later, it may be useful to consider lighter,
  LVL2-specific, persistent objects
• Advice welcome! (thanks to Tadashi Maeno, John
  Baines and Ed Moyse)