1. Speed: update measured
beam position every 1-10 seconds
2. Accuracy: < 10
microns
3. Robustness: can survive
any kind of bad data
4. Running all the time:
results updated continuosly on ACNET and Web page
5. Written in ANSI-C
for portability
VxWorks vs. Unix or Linux
Definition of d: impact parameter
Minor problem: there
is a discrepancy of sign between how d is currently defined in SVT and
in the rest of CDF. I propose that we conform to the CDF standard.
At phi = 0 (track is parallel to the x axis) the impact
parameter is equal to the coordinate of the intersection of the track with
the y axis
Major problem: parameters
of tracks output by SVT must be measured in a fixed well defined coordinate
system common to all CDF. This implies that impact
parameter must be defined with respect to the CDF coordinate system (which
does not change from run to run) and not with
respect to the beam axis (which may change from run to run).
At the beginning of each shot we will try to steer the
beam to the origin (goal is < 10-20 microns), so that the two definitions
of d are as close as possible. If we want to apply a further correction
to d based on continuosly monitoring the position of the beam spot, this
must be done somewhere else (in the Level 2 processors?)
Math is simple:
d' = d + x*sin(phi) - y*cos(phi);
where (x,y) is the beam position and trig functions come
from look-ups.
Beam Alignment - Milestones
Feb 5, 2001: working
fit algorithm
testing with CR data
timing
Unix
Linux
VxWorks
documentation of algorithm and test results
Mar 1, 2001:beam
position on ACNET and Web
demo using simulated data (no VME readout)
April ?, 2001:process
data from TF's
requires data flow through TF's (real or simulated)