User name R. Michaels
Log entry time 14:48:42 on October 6,2002
Entry number 87471
This entry is a followup to: 87468
keyword=helicity disaster (NOT)
We analyzed a few other runs and found a pattern -- once the helicity algorithm
successfully locks into the correct bit pattern, it predicts the helicity for many
thousands of events. We cannot think of any hardware failure (e.g. wrong bits)
which would only happen at the start of a run; therefore, the problem must be
software. Other points about this:
1. Xiadong agrees with this based on his analysis. He's working on it.
2. There is an ADC which measures helicity, Xiadong says the data from it are ok.
So, this is a backup measurement in case there really is a problem with the I/O
register that flags the helicity info.
3. When we went from COINC to UNBUFF mode, the deadtime increased and
so did the frequency of helicity "disasters". More evidence of a software problem,
i.e. a NON-problem.