Veto shield timing calibration

1
Veto shield timing calibration - A usefulness
study -
Calibration workshop September 9th, 2005 Pedro
Ochoa

• Getting timing info. from the shield
• Can we use the T0s?
• Approach 1
• Approach 2
• Summary

2

1. Getting timing info from the shield.

For most tracks we are able to extract some
usable information from the shield.
i. found where expected to be ii. with adc gt
200 iii. not in a PMT where other hit was
expected to hit first
A good digit in the shield is one
Using tracks which cross 10 planes or more, we
find

• For 79.9 of tracks we are able to extract at
  least 1 good shield digit.
• ? For each track, we find in average 1.74 good
  digits in the shield.

I will not use the noisy shield cut described
my last talk. Even if it almost completely
eliminates early hits produced by other particles
coming in with the muon, it eliminates about 70
of our sample.
(ns) vs. adc
(ns) vs. adc
Before noisy shield cut
After noisy shield cut
I want to see if even those hits whose timing is
not very exact can be useful
3
The shield may have a considerable impact in the
direction determination of atmospheric events
(for instance, in the partially contained upward
going PCUP analysis)
In order to distinguish upward going events from
downward going, it is useful to make time vs.
pathlength (distance along track) graphs for
each track, measuring the pathlength with respect
to the tracks lowest point in the body of the
detector. When doing so we get the two distinct
cases
UPWARD GOING
DOWNWARD GOING
Time(ns) vs pathlength(m)
Time(ns) vs pathlength(m)
track body shield (without calibration)
shield (calibrated)
4
4 more typical examples
track body shield (without calibration)
shield (calibrated)
Time(ns) vs pathlength(m)
Time(ns) vs pathlength(m)
Time(ns) vs pathlength(m)
Time(ns) vs pathlength(m)
5
II. Can we use the T0s to determine
direction? Approach 1 including the shield in
the calculation of 1/
Took a sample of cosmic ray muons, and calculated
1/ for each by fitting a time vs
pathlength plot with a straight line. The fit
was done with shield hits and without shield
hits separately. The RMS of the obtained 1/
distributions are then compared, as a function of
the ADC cut made when selecting the appropriate
veto shield hits
(without shield RMS) (with shield
RMS) (without shield RMS)
Tracks cross 30 planes or more
between 20 and 30 planes
between 10 and 20 planes
The use of the veto shield information
contributes to a better velocity determination
for cosmic ray muons.
Typically 9-18 Improvement
6
Approach 2 use some measure of upness
downiness

• Took a similar approach as the atmospheric
  neutrino group
• Fit the time vs. pathlength graphs with two
  straight lines, one of slope c (hypothesis 1,
  downward going) and with a line of slope c
  (hypothesis 2, upward going).

-c
c
Then take the RMS with respect to each line, and
compare them. Ideally, (RMS_up
RMS_down) gt 0 ? downward going
(RMS_up RMS_down) lt 0 ? upward going
7
Applying this to tracks that cross between 10 and
20 planes and using 1 month of R1.14 data we
obtain
WITHOUT veto shield
WITH veto shield
Mean 7.42 RMS 5.57
Mean 8.59 RMS 4.63
peak reduction (good)
Zoom ( log scale)
Zoom ( log scale)
Mean -0.59 RMS 0. 85
Mean -0.78 RMS 1.25
more background (bad)
8
Similar principle results are obtained for tracks
that cross between 20 and 30 planes
WITHOUT veto shield
WITH veto shield
Mean 9.10 RMS 4.17
Mean 10.07 RMS 4.14
Zoom ( log scale)
Zoom ( log scale)
Mean -0.78 RMS 1.33
Mean -1.1 RMS 1.52
9
It seems we have effectively resolved some of the
dubious cases into downward going events
(good!), but weve also introduced more
background (bad!)
Part of this background is due to crazy guys in
the track body
crazy guy !!
Time(ns) vs pathlength(m)
These two examples were flagged as upper-going by
our method because of these crazy guys
10
But, as it could be expected, we also have crazy
veto guys
crazy veto !!
Time(ns) vs pathlength(m)
Time(ns) vs pathlength(m)
These two examples were flagged as upper-going by
our method because of these crazy veto guys
11
We can try to remove some of this background by
requiring the winning hypothesis to not have an
RMS larger than a certain amount (otherwise, it
means both hypothesis are bad and most likely
there are outliers). We obtain
WITHOUT veto shield
WITH veto shield
Mean -0.54 RMS 0.9409
Mean -0.65 RMS 1.31
bacground reduction !!
12
Clearly, when doing a formal analysis, approach 2
has to be made more robust so that it wont be
fooled by outliers, whether they are in the veto
or in the track. However, when theres no
outliers, approach 2 still allows us to pick up
some interesting (besides the obvious ones)
upward going candidates after using the shield.
EXAMPLE 1
Time(ns) vs pathlength(m)
? Maybe wouldnt bet on it
13
Heres another dubious example
EXAMPLE 2
14
In this case the shield confirmed the up-ness of
the event. It appears to be a clear upward going
muon.
EXAMPLE 3
4 hits !!
? This is the kind of events where the shield
would have a bigger impact
15
Summary Ongoing Work

• The shields timing appears to be in good shape.
• Overall, it does seem that the veto shield will
  improve the direction determination of
  atmospheric events
• Approach 1 showed a 9-18 improvement in the RMS
  of the velocity distribution of cosmic rays after
  using the veto shield timing information.
• Approach 2 showed an increase in downiness
  after applying the veto shield.
• The background that was added by the shield in
  approach 2 will have to be removed by smarter
  algorithms.
• Need to put the constants/functions in the
  database.
• Need to show this to the atmospheric neutrino
  group.