RUN 2B L1CAL MEETING MINUTES
			     5 September, 2002


All			      Rate Estimates
---
o Two problems are currently seen at Nevis
  1) Jet trigger rates calculated using QCD files generated using
     (sim-reco) mp07-p08 are a factor of 2-3 lower than those
     calculated using mp10-p10.
  2) Jet trigger rates calculated using the version of the sliding
     windows simulation from spring 2002 are also significantly lower
     (factor 2-5) than those using the latest version in CVS.
     . the main difference between these versions appears to be the
       treatment of rounding and truncation.
     . but this shouldn't cause such large effects
  -  We will continue to investigate both these issues at Saclay and
     Nevis. 
o Toy Trigger List
  - Mike Hildreth has estimated rates (L=2 10**32 @ 396ns) for a toy
    trigger list.

    Term                            Rate

    CEM(1,10)                       1338
    CEM(2,5)CEM(1,7)                 539
    mu1pt4attx                     37741
    mu2pt2wttx                       358
    CEM(1,7)CJT(2,5)                 762
    mu1pt2wttx.CJT(1,5)                9
    mu1pt2wllx.CEM(1,5)               10
    CME(10)CJT(2,5)                 2126
    TTK(2,3)TTK(1,4)CEM(1,5)TIS(4)   643

    Total                          43526

  - For the review, we should try to estimate the rate reductions the
    new trigger will give us for jet (CJT) triggers.
    . EM and missing-Et are not yet well-enough understood in the new
      L1Cal to get such estimates.


Jiri			      Data Plots
----
o Jet turnon curve with data
  - tried different ways of matching Trigger jet w/ reco jet
    . end up using 0.5 cone around jet direction
  - curves don't plateau nicely as would be expected
    . jets at eta~1.5, phi~pi, Pt(reco)~30 seem to be the problem
      > seem to be coming from low Et TTs
    . all come from a small set of runs
    . when these runs are cut out, things look much better
o Et(trig)/Et(jet)
  - using new jet-RoI matching
  - note that adding more statistics to the dimuon sample used to make
    these plots for the August Temple Review doesn't improve the look
    of the plots at all
    . enough stat's with the old data sample
o Jiri will make new plots available
  - turnons for two threshold sets
  - Et(trig)/Et(jet) for Moriond dimuon sample


All			      Lehman Preparation
---
o Two main things left to do
  1) Improve Basis-of-Estimate notes describing labor and cost
     estimates for each task
     . Patrick and Hal will look into this
  2) Come up with "global gain factors" from using the new L1Cal in
     Run IIb (as suggested by the Temple Committee)
     . Gain in Higgs sensitivity: this will be of order 15-20%
       (i.e. not extremely impressive)
     . Reduction in rate of global trigger list and important triggers
       (see Rates discussion above)


Patrick                       Trigger Meeting in France
-------
o The current best guess for the dates for this meeting is 5-7 November.
  - hopefully this will be finalized in the very near future