Muon reconstruction in release 14.2.10.1

D. Adams
July 2008

Release 14.2.10.1 is for production.

For other releases, see the complete list of results.

Issues in 14.2.10, 14.2.20 and 14.2.21

Issues were last reported for 14.2.10.0.
Issues are next reported for 14.2.22.1.

1. Hole counts are absent for muid and staco combined.
  • Hole counting is not enabled in 14.2.22 reconstruction. This was the also the case in all past releases.

    2. Muid combined fit chi-square does not include the inner track. Staco includes inner, outer and match.
  • Users should add Muid inner and combined chi-squares (and DOF's) to get comparable values.
  • This is a lon-standing feature, rather than a bug.

    3. Calorimeter-tagged muons have very low efficiency with displaced vertex.
  • There is apparently a cut at 100 microns.
  • Bug report: 39658.
  • Expect the same in 14.2.20.
  • This is fixed in 14.2.21.

    4. Moore/Muid truth match is degraded.
  • Mean truth chi-square is now 2.7 instead of 2.5 (mean probability is 1.3 instead of 1.1). Result is independent of geometry.
  • The momentum resolution is slightly improved. Degradation is in other parameters.
  • Could this be a change in the vertex constraint?
  • Degradation is half as big in 14.2.20 but combined is now also degraded by 10%.

    5. Muid energy loss is decreased.
  • Mean was 4.8 Gev in 14.1 and is now 4.3.
  • Bug report: 39674.
  • Same in 14.2.20.
  • Problem was that trigger was overriding reco job options so that a parametrization was alway used to estimate the energy loss. This was fixed in som release (14.2.21?).

    6. The calorimeter isolation (sum ET) is drastically reduced in 14.2.10.
  • In 14.1.0.3 csc500 T1, the mean is 2.5 GeV; in 14.2.10.1 it is 1.0.
  • There is another 10% drop with the new geometry geo200 and a few percent increase with the displaced vertex (geo210).
  • Bug report: 39661.
  • Same in 14.2.20.
  • Fixed in 14.2.21.1. See combined muon talk at the 30 Sep physics validation meeting.
  • Still about 20% lower than 14.1.0.4 and 10% lower than 13.0.40.5.

    7. The Moore (and Muid) RPC and TGC phi hit counts have increased dramatically in 14.2.10.
  • This may be related to the degradation in the Moore truth quality (item 4).
  • Bug report: 39666.
  • In 14.2.20, there is a patch to put these back to the 13.0.40 level--much more than 14.1.0 and 14.2.0 but much less than 14.2.10.
  • Fixed in 14.2.21.1. See combined muon talk at the 30 Sep physics validation meeting.

    8. MuGirl show some efficiency loss at high eta.
  • Problem may be that she is not finding CSC segments.
  • Fixed in 14.2.20, MuGirl has high efficiency at all eta.

    9. Craig Wigglesworth reported an inconsistency between the track isolation count and pT sum in 14.2.10 data.
  • See muon combined performance hypernews thread 154.
  • Ketevi recommended Craid use IsolationTool-00-05-04 for 14.X data. I use this for muperf results generated after October 25.

    10. The efficiency for Moore/Muid falls by 20% for |eta|>2.0 when dead channels are flagged in the conditions data (OFLCOND-SIM-03-00-00).
  • Bug report: 41695.
  • Problem comes from CSC.
  • The Fall 2008 production shows this effect.
  • The problem is fixed in 14.2.22.2.

    11. There is poor standalone resolution for |eta|>2 when data simulated with ATLAS-GEO-02-01-00 is digitized in 14.2.20.1 is reconstructed in 14.2.20.3 or later.
  • Bug report: 43672.

    General issues for the MC08 production (using 14.2.20.3) are recorded on the PhysVal10TeV wiki.

    Graphical summaries

    T1 validation
    Z->mumu validation
    pt=100 single muon validation

    dladams@bnl.gov