For recon 12 roughly 20% of all PHCHRG tracks have SWMSTAT = 30, i.e. the swim to the beamline is not attempted. For those tracks, the data in HLXPAR does NOT correspond to the track parameters at the point of closest approach to the beamline. It should be noted that none of these tracks have associated VXD hits.
In recon 12A all tracks with SWMSTAT = 30 were successfully swum to the beamline using Homer's BBEXTRK routine (see also PHSWIM and ZXSWIM).
- The systematic error for hit reconstruction in the gas radiator was reduced from 4.5 to 1.5 mrad (CRDRECP.GASYSERR). This reduction is the result of improved resolution with the new alignment constants. This significantly improves the particle identification at largish momentum (p > 12 GeV/c).
- One more mirror is flagged as "missing", thereby setting the BADID bit in the gas reconstruction.
- The TPCbad bit is set less often in the liquid radiator reconstruction (this bit does NOT set the BADID bit).
- For the data, the gas index of refraction was measured using SONAR measurements on a run-by-run basis. This SONAR-based index is now used in the hit reconstruction.
Other improvements that affect only the Monte Carlo are as follows.
- For recon 12 the CRID high voltage for the summer 94 period was marked OFF for far too many events. This was fixed by updating the HV information in CSTATUS based on CRID event sizes.
- Too many hits were tossed for the 45 degree mirrors. So the hit tossing was readjusted to better reproduce the number of hits seen on the data.
- The gas index of refraction used in MC reconstruction was adjusted slightly to reflect a small offset in the average generated index due to an asymmetric photon energy acceptance.