- I took a closer look at the 3-hit tracking under the new
tracker, and found a few bugs.
- The z estimation in the BMU_DtoE has been repaired, and now
seems to make sense. In the upper left
plot (due to Camille: it is a plot based on reconstructed muons)
we see x vs y. The gap at the bottom is where we lack chambers
because of the toroid supports. Notice that most of the stub positions
are exactly where we expect--in a ring. However, 7.5% of them show up
in various places inside the toroid region, including quite a number
at about 1.3 meters radius. I do not yet understand why these nominal
stub positions appear here. When I plot the position of the stub
at the time of stub creation I see only the outer ring.
All but 3 of these curiously placed stubs are
3-hit stubs; which suggests
a place to start looking. See below for solution
The upper right plot
is that of the z of the stubs, which looks
sensible. The spike at the rear of the chambers is due to accidental
matching of a COT track and one of the rear stubs that bedevil us.
The lower left plot shows that the odd x/y positions don't correlate
with z. The lower right shows the chi-squared of the stubs, which are
looking sensible again since the alignment corrections.
- The bogus hits are now understood, and solved. The problem was
that the position of the stub was the average of the hit positions,
and thanks to an error in the hit discarding algorithm a hit on the
opposite side of the detector could masquerade as one of the set of
good hits. All is still not bliss,
since this now picks up a lot more 3-hit stubs than it used to, and
the chi-squared mystery is back. The left column is 3-hit stubs, and
the right is 4-hit stubs.
Modified 19-Sep-2002 at 14:37
Previous meeting notes
Next meeting notes
Main slide directory
if you have trouble accessing the information on this page.