I might suggest a minor modification to the monitor to ignore uVAX3 until
you trust the calibration constants. All the bursts that have been coming
lately seem the product of bad calibration constants. (Box 134 has appeared
in most bursts, more than once in many bursts... Obviously a badly calibrated
box.) I've tried to stress this point before, but for the online monitor
it is absolutely crucial to have good, current calibrations. If only a small
number of boxes are giving most of your hits-above-threshold, you are dealing
with a very different statistical beast than if the boxes are more uniform.
I think you need a way of doing "quick and dirty" calibrations that don't
depend on the laser/LED runs. I've written software to do the required
calibrations, although I'll admit it has a less-than-user-friendly interface.
I explained my calibration method in some detail at Cape May, but the short
list of advantages includes:
1) Independence from calibration system problems.
2) Weeds out boxes that can't be calibrated. (VERY important because it
can raise red flags about detector problems and also prevents bad boxes
from contaminating your burst search. ~10% of boxes have some serious
problems.)
3) UNIFORM rate-above-threshhold for all boxes. This ensures that all boxes
are contributing equally to the burst search. As is, you are much more
sensitive in some boxes than in others. The issue of whether a "reconstructed
energy" corresponds well the the actual energy deposition is important, but
I've resolved it to my satisfaction. I believe the energies I reconstructed
much more than I trusted the reconstructions from standard calibrations. At
least with my constants, I didn't get weirdness like box 134 giving 100x the
average rate of 10MeV+ hits compared with other boxes.
Well, that's just the short list.... :)
Guess I should worry about my current job too. Say hi to everyone in Boston,
Rich
--
Alec Habig, Boston University Particle Astrophysics Group
habig@budoe.bu.edu
http://www.astro.indiana.edu/home/ahabig/