|
|
DIRC Prompt Reco QA Histograms
The plots that are shown here as an example were done from run 27025.
Hits are associated to a track if they are in a reasonable time and theta window.
-
Occupancy, associated hits
The occupancy highly depends on the background level, so we expect the
corresponding plot to vary significantly over time. After adding several layers
of shielding, in general we observe the highest occupancies in the region
around sector four.
The plots for associated hits should be less dependent on background levels.
-
Timing
The first plot is the time of the bunch crossing as it is given
by the Drift Chamber. This is not a DIRC plot, but the DIRC
reconstruction uses this time as t0 of the event. So if the other
three plots on this page indicate timing problems in the DIRC, this
first plot could give hints to understand them. Small shifts of the
mean are no problem for the DIRC (calibration).
The time for associated hits should be in a +/- 100 ns window with two
clearly visible peaks due to signal photons (reflected by a mirror or not).
The meantime of an event is some timing correction calculated from the
photons in the event. It should be centered around zero.
The quantity that is plotted in the fourth plot is delta_t = time(hits) -
time(propagated photon) - t0(Dch) +/- various corrections. The mean
of the central gaussian should be close to zero. Photons within +/-
10 ns only are used in the reconstruction. The timing resolution is
important for the reconstruction. The width of the central gaussian
should be less than 2 ns.
-
Theta C resolution.
The momentum distribution of tracks with th-C measurement should be
relatively insensitive to background conditions. Note, e.g., the
"Bhabha-Bump".
The nex two plots show the resolution per photon and per track. The
central gaussian should be centered at zero and its width should be
comparable to the example given here.
-
Number of Photons, theta residual.
Check that the mean number of photons is similar from one sector to an
other.
A shift in the residual plots from 0 may indicate a problem of alignment.
There should be no major changes between run for a longer time period.
-
Particle Identification
The hypothesis frequencies histograms show the result of identification:
-2 = no identification; 0, 1, 2, 3, 4 = electron, muon, pion, kaon, and
proton, respectively.
The second plot is the threshold plot. One should see the contributions
from electron and muon, pion, kaon, proton distinctively.
-
Bhabha and DiMuon control sample
All of the plots on the next two pages also appear on previous pages, but here
we make them for a well defined set of events (Bhabhas and DiMuons). We therefore
expect the plots to be less sensitive to different running
conditions. It seems like a good idea to spend some extra time on these
pages to have a closer look at the fits (resolution of central
gaussian, centered at zero).
-
KS, proton control samples
These two control samples can be used to obtain rough estimates of the
PID performance.
Comments to : J.
Stark, L. Roos.
Revised by Thomas Hadig March 20th, 2002
|
|
