Occupancy and associated hits
The postscript file for this page can be found here.
The plots drawn on this page are the following.
- Top left
This plot shows the mean occupancy (in %) of each sector, i.e. the average
number of PMT hits per event per sector, divided by the number of PMTs per
10,752 (in total) / 12 (number of sectors) = 896
This histogram should be roughly flat, with perhaps some excess of hits due
to background in sectors 3, 4 and 5. The average occupancy depends on the
luminosity and on the background conditions for the run. When lumi is high
and background reasonnable it should be around 4-5%.
To ease the reading of the plot, the 'average' occupancy and its 'dispersion'
are written above it. These values are simply the mean and the RMS of the
12-bin histogram. The same convention holds for all DIRC QA plots which have
such red statements written above the graph area.
- Top right
This histogram shows the number of associated hits per event, i.e. the number
of hits which are found to be compatible with tracks. Note the log. scale on
The huge peak at 0 is always present (several events have no track radiating
Cerenkov light in the DIRC). After a bump around 30 associated hits, the
distribution usually goes down 'log-linearly'. Its RMS is around 40 and its
mean value -- which used to be around 40 as well -- is now (Run 5b onwards)
closer to 35 after a change in the trigger (the tau filter is now accepting
- Bottom left
This histogram shows the average number of associated hits per event in each
sector. It should be pretty flat with an average value around 3. Holes in
this distribution which are not visible in the occupancy plot trigger a
reconstruction problem, for instance a bad timing correction in the missing
sectors. In that case, the culprid is likely to be a rolling calibration.
- Bottom right
This plot shows the distribution of the number of associated photons per track.
The mean value should be around 28 -- it's slowly decreasing with time due to
the DIRC ageing -- and the RMS around 16.
Top right and bottom left plots are computed using the same sample of tracks
(cuts on pt and on the track quality in DrcOprMon). So, summing the 12 bins of the bottom
left histogram should give something compatible with the mean value of the top
Back to the DIRC QA plots page