Other bug fixes
Silicon alignment tables in Simulation, for PASS 17
- Default simulation is set to generate realistic MC with
misalignment and beamline from DB after setting a run number in
run_cdfSim.tcl. The beamlines are taken from DB based on the run
number, but the silicon alignment table is hadwired in the
setup_simulation.tcl due to a generic problem in passing the
mis-alignment information to geant3 through DB. Recently we have found
out that silicon alignment table was not matching the runs number for
run >190000.
The correct values are given below and have to replace
the lines 21-29 in
http://cdfkits.fnal.gov/CdfCode/source/mcProduction/tcl/mcprod_setup_simulation.tcl?v=6.1.4
and
the lines 21-29 in
http://cdfkits.fnal.gov/CdfCode/source/SimulationMods/test/setup_simulation.tcl?v=6.1.4
When generating MC data, check which files you use to setup *ALIGN_TABLE environment variable.
- The fix is available for 6.1.4 MCProd (mcProduction)
when building mcProduction tarball using
mcProduction/scripts/build_v6_1_4
which automatically picks up
SimulationMods/test/setup_simulation.tcl -r1.42
mcProduction/tcl/mcprod_setup_simulation.tcl -r1.13
through
Release/patches/6.1.4_mcprod.patches -r1.3
- The correct talk-to is shown below:
###### Specify alignment table
#set ALIGN_TABLE default
if { $RUN_NUMBER < 149717 } {
set ALIGN_TABLE "ofotl_prd_read 100045 2 GOOD"
} elseif { $RUN_NUMBER < 158636 } {
set ALIGN_TABLE "ofotl_prd_read 150045 1 GOOD"
} elseif { $RUN_NUMBER < 174656 } {
set ALIGN_TABLE "ofotl_prd_read 160045 1 GOOD"
} elseif { $RUN_NUMBER < 179056 } {
set ALIGN_TABLE "ofotl_prd_read 170050 1 GOOD"
} elseif { $RUN_NUMBER < 190000 } {
set ALIGN_TABLE "ofotl_prd_read 180050 1 GOOD"
} elseif { $RUN_NUMBER < 212430 } {
set ALIGN_TABLE "ofotl_prd_read 190050 1 GOOD"
} else {
## Need to add whatever table will be used after shutdown
## The code will freeze if no alignment table is given!!!
set ALIGN_TABLE ""
echo "********************************************* "
echo "Severe Error: missing silicon alignment table "
echo "********************************************* "
}
- Validation of the fix.
We produced a ttbar sample (0.5 M) events and checked that d0 vs phi plot looks fine. Thanks Daniel and Clark for help.
Mircea Coca
Last modified: Tue Mar 7 14:25:08 CST 2006