Filtering Volume & Simulated Cosmic Truth Studies

  • this shows some stuff i see when looking at the different filtering volume cosmic sim samples

How the truth is done

  • for the truth i use Thijs FillTrack CBNTAA alg (from InnerDetector/InDetTestBeam/InDetTestBeamAlgs/InDetTestBeamCBNT) which finds the truth track as it crosses a Layer, and then extrapolates this track to the beamline to calculate d0, z0
  • his version uses the CaloEntryLayer - but this only gives truth values for tracks with d0<~1000mm so i have changed this to be the MuonExitLayer, this now gives truth params for tracks with all d0
  • but something strange happens with this volume. we now get mostly 2 mc truth tracks per event (where obviously there should be 1) the top left plot in the pictures below shows the number of mctruth tracks per event). for the CaloEntryLayer this was mostly 1 truth track / event - anyone know why?
  • looking at the truth params (d0, z0) of the 2 tracks in each event these seem to be basically the same (see bottom right and bottom middle plots of pictures below which show d0 of track1 versus d0 of track2 (and same for z0) where you can see a large correlation).
  • so for now i just use the first mc truth track for my studies.

  • for some studies it would also be nice to have access to the true z0, d0 which are just the truth values at the vertex (100m above the beamline) extrapolated as a straight line to the beamline (this is what we used to have). can i get this from some existing CBNTAA alg?

Muon Volume

  • truth plots for muon volume:
  • so the muon volume size seems to be ~ |z|<10m, radius ~4m (funny hole at z~+/-5m??)

TRT Volume

  • truth plots for TRT barrel volume:
  • so the trt volume size seems to be ~ |z|<1.2m, radius ~1080mm (no holes)

SCT Barrel Volume

  • truth plots for SCT barrel volume:
  • so the trt volume size seems to be ~ |z|<1.m, radius ~500mm (no holes)

Pixel Volume

  • truth plots for pixel volume:
  • so the pixel volume size seems to be ~ |z|<3.5m, radius ~25cm (no holes)

Some comments

  • Muon volume seems to have a strange (to me) hole at
  • Pixel volume is way to big in z (+/-~3.5m)
  • would be good to have an SCT barrel volume
  • rate(s) (no TRG eff, or requirements that there was any reco - just rate that there was a simulated hit in the volume)
  • Muon vol: 3139 Hz
  • TRTBarrel vol: 27Hz
  • Pixel vol: 34Hz
  • SCTBarrel vol: 12Hz
  • Comments:
  • 1. TRT vol and Pixel vol having same rate seems weird - but actual x-sec area of TRT vol is 2*1.2*2*1 = 4.8m^2, whereas for pixel vol it is 2*3.5*2*0.25 = 3.5m^2, which are qute similar.
  • 2. looking at subset of events in muon vol that have d0,z0 compaticle with being in TRT vol gives a fraction of event of something like 0.046 - 0.078 => rate = 143 - 244Hz (this is much higher than 27Hz - why???)

effects of TRG (LVL1 RPC only with 900Gev/Cosmic config)

  • Muon vol truth params w and w/o passing LVL1 trigger (eff ~ 62%) :

  • trt vol truth params w and w/o LVL1 trigger (eff ~92%) :

* SCT barrel vol truth params w and w/o passing LVL1 trigger:

  • Pixel vol truth params w and w/o LVL1 trigger (eff ~ 92%):

Fraction of events with pixel hits with different volumes

  • For TRT volume we have 646 events with at least 1 pixel hit out of 8971 events => 7.2%
  • For SCT volume we have 328 events with at least 1 pixel hit out of 1910 events (low stats) => 17.2% (low stats)

Charge asymmetry of cosmics that arrive at detector

  • Looking at the signed mtm of the truth tracks at the perigee i see
  • So you see significantly more -ve muons than +ve ones. (the plot above shows the different filtering volumes (from left to right, top: Mu, TRT, TRTwithField,, btm left SCT, btm middle pixel) which all show the same effect.
  • counting the number of -ve tracks versus +ve tracks i see
 vol = mu +ve = 6968 (0.379769) -ve = 11380 (0.620231)
 vol = trt +ve = 722 (0.391328) -ve = 1123 (0.608672)
 vol = trtField +ve = 3616 (0.403122) -ve = 5354 (0.596878)
 vol = sct +ve = 724 (0.379058) -ve = 1186 (0.620942)
 vol = pix +ve = 1213 (0.403661) -ve = 1792 (0.596339)
  • so all volumes see ~60% -ve tracks and ~40% +ve. im not sure about this.
  • Looking in the PDG and its references i found this paper on measured muon asymmetries in cosmic rays - which says that an asymmetry is expected but you expect more mu+ than mu-
  • Looking at the truth 1./qoverp at the generator surface (ie. at the surface) i see the plot on the left (below) (for fixed 50GeV cosmics) whereas the plot on the right (below) is for truht 1/qoverp at the perigee
  • So this shows that we see the same charge asymmetry at the surface - which i think means there are 2 options for what could be going on
  • 1. the generator or interface from generator to athena is somehow mixing up the charges
  • 2. the charge dependent material interactions in the rock are such that the events which get pass the Filter volume are more likely to be -ve than +ve (this in spite of the fact that (probably) the generated spectrum will have more +ve than -ve tracks in it.

  • Further investigation i found a bug in the code that passes the muons from the generator into the simulation. the code (CosmicGenerator/src/CosmicGenerator.cxx) says:
483       // Get the mass of the particle to be generated
484       int charge = gun->GetMuonCharge();
485       m_pdgCode.push_back(charge*13);
  • and then m_pdgCode is used to construct the muon candidate. but for charged leptons the sign of the PDG code is opposite to the charge so this code flips the charge of the produced muons.
  • we have fixed this privately and Rob McPherson will make a tag which officially fixes it.

Looking at energy loss

  • to study the energy loss of the cosmics from the surface to the detector we generated a sample with fixed energy 50/100GeV (on the surface) and then look at the mtm of the particles at the beam line.
  • both these datasets were generated without field and using the TRT barrel volume
  • truth spectrum for cosmics generated with fixed 50GeV energy:
  • so it looks like quite a lot of energy is lost on average (mean is 13 GeV)
  • also notice the shape of the truth particles in theta is different for these high energy cosmics (the effect of the shaft is much less visible)
  • same thing for 100GeV fix energy cosmics (generated at surface with 100GeV) shown below
  • This shows a loss of mostly >30GeV. also can see the hole in the muon system in the triggered (red) plots at z=0, as now these cosmics are coming much more straight down (i think). and effect of the shaft seems to be not really visible at all.
  • also looking at the mtm versus theta can see that cosmics that go down the shaft (i think) have higher mtm which is what i would expect:
  • singleE_100.cosmics2.gif:
  • singleE_50.cosmics2.gif:
  • the charge assymetry for the single energy cosmics are:
 100GeV +ve = 1774 (0.394924) -ve = 2718 (0.605076)
   50GeV +ve = 661 (0.369274) -ve = 1129 (0.630726)
  • looks pretty similar...
  • below is for single energy 25GeV and 10GeV cosmics.
    singleE_25.cosmics.gif * singleE_10.cosmics.gif:

  • can see that only have cosmics of 25GeV getting to the detector down the shaft!

-- JamieBoyd - 17 Jun 2007

Edit | Attach | Watch | Print version | History: r11 < r10 < r9 < r8 < r7 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r11 - 2007-06-26 - JamieBoyd
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    Main All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2019 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding TWiki? Send feedback