Analysis Meeting, Jan 27, 2003
Implementation of Alignment in
Alignment is implemented through TkrAlignmentSvc;
first iteration is in cvs:
- There are six alignment constants (dX, dY, dZ,
RotX, RotY, RotZ) for each wafer in the tracker. A wafer is specified by its
tower, layer, view, ladder and wafer numbers.
- There are two lists of alignment constants
(which may be the same):
- one for simulation
- one for reconstruction
- jobOptions specify the names of the files
containing the lists, or special test lists may be generated:
- currently, offset every wafer during sim and/or
recon by a fixed (dX,
- others, as desired.
- At digitization, each entry and exit point are
translated in the plane of the detector to the position determined by the
alignment constants. Thus, the z location of the transformed hits is
unchanged. This means that geometry still deals with the ideal detector.
- The transformation depends on the position
(X, Y) of the hit and the slope of the track (alphaX, alphaY).
- During reconstruction, a second set of
constants is used:
- During clustering, only the x and y shifts
are applied to the clusters, since neither the orthogonal coordinate nor
the track slope is known.
- During track propagation, the remaining
transformations can be applied, as the track is fully characterized at
this time. (Yet to be done.)
The following details from the event display show
the operation of the Alignment service. The alignment constants are in test
mode, which specifies an offset of 1 strip in X and two in Y, for both sim
In each case, the black line is the MC
track, which doesn't move; the green "X" is the at the center of the
hit cluster; the blue line is the reconstructed track; and the red line is a
"thermometer" at the
point of energy deposit. (The red line is not yet corrected. But
see Late-Breaking News!)
1) Original Event...
3) Digitization transformed, Recon not...
4) Both Digitization and Recon transformed using
the same alignment constants.
5) Late-Breaking news! Event thermometer fixed!
not yet in cvs...
display could use some work...
- Apply the alignment transformation to "tracks."
- Validate the transformation formulae for deltaZ and rotations.
- Modify the code to interface to actual alignment
- Start work on calibration database format.
- Reorganize hit display.