I think we can add an expert histogram to YMon that will help us distinguish sparking from oscillation. Oscillations typically have short widths, and stand out clearly in plots such as this: (PostScript here)
More recently Run 224480 had some problems, which we can view below
You can see the differences in width distributions fairly easily for both the spikes. There appear to be 3 distributions: one peaking at about 60 nsec width, one at about 10, and a spike at 0. The first distribution with the 60 nsec peak width is missing/suppressed in the case of oscillations. I had thought that oscillations tended to have more 0-width hits, but this seems not to be the case.
Alternatively, we can look at the difference between times of hits. In the following plot the upper left peak is the distribution of differences between start times for the hot channel seen above on the East side. The lower left is the distribution for all other channels on the East side. There is still a peak at that time. If I look at the stave distribution for hits with difference in start time in that peak, I get the upper right distribution, which isn't that different from what was seen before, though more pronounced. The lower right is the West side distribution, which is again similar to that found by cutting on widths, but more pronounced. Naturally a cell has to have more than one hit to appear in this plot.
Pierre Savard confirms that the remaining Muon and MuonMods changes will go in 6.1.4int5. The BMU/scintillator timing mismatch was solved in 6.1.4int2
Modified 16-October-2006 at 08:41
http://hep.physics.wisc.edu/~jnb/imu/16Oct2006