Frascati test beam analysis

1
OUTLINE

• Frascati test beam analysis
• Trigger analysis
• Muon rate
• Q distribution
• Baseline
• Pulse shape
• Z measurement
• Att measurement

2
Result from June Frascati test beam
3
Analysis topics

• PMT transfert function
• Attenuation length measurements
• Position Measurements with ln(Q1/Q2)
• Effective velocity veff measurement

4
PMT transfert function
5
Attenuation length
The effective bar attenuation length ?eff can be
estimated with the log of charge ratio of
opposite PMT.
Q1G1E0 exp(-(L/2-z)??eff?
Q2G2E0 exp(-(L/2z)??eff?
ln(Q1/Q2)ln(G1/G2) z???eff?
?eff is not the bulk ?140cm from data
sheets ?????eff / ltcos ?gt ltcos ?gt
(11/nsc)/20.81
6
For the bulk value ? ?eff/ltcos ?gt 103.5
14.8 cm Different from the data sheet value and
with a large spread
7
Reflection loss and ?
Reflection losses at the surface below the
critical angle affects ?eff If the reflection
efficiency is R2 1 ??there is a reflection
absorption length
?Ra/(tan????ln (1/R2)) a/(tan?????)
1??eff1????cos????? 1??R
From ?eff to R2
Bar 15d 16d 17d
19d ???? 0.60 1.21
1.63 0.43 Bar 20
2d 8d Ave ???? 1.56
0.40 1.76 1.0
A large spread in ? depends on surface quality
and may be more credible that the same spread in
bulk property
8
Position measurement
Using ?eff we can obtain a position measurement
from
ln(Q1/Q2)ln(G1/G2)fit z???eff,fit?
Averaging all positions ?(z) 2.6 cm Error is
larger for z close to PMT where the linear
relation breaks down because of increased
direct light
9
Effective velocity measurement
veff defined by
t1t0 (L/2-z)?veff t2t0 (L/2z)?veff t2-t1
2z?veff
Important t1 t2 depends on algorithm e.g.
amplitude normalized thresholds timing fires at
? of the maximum (?10,50,90)
veff is faster when the algorithm selects the
fastest photons (small ???small ?) For ?????veff
-gt c/nsc 18.87 cm/ns For ?????veff -gt c/nsc2
11.87 cm/ns
10
The trend of increasing veff,??with?decreasing
??is confirmed More studies with test beam MC to
gain better understanding
11
Trigger DataCharge vs Amplitude PMT
Key for the Trigger System using pulse height
with no information about pulse charge (RUN
236,237,239)
Charge (a.u.)
Amplitude (a.u.)
12
Charge Inner vs Outer
We observe that charge distributions are
different between inner and outer PMTs (RUN 236,
237, 239).
Charge Inner PMTs Charge Outer PMTs
Q (a.u.)
13
Comparison with MC Data
Measured charge distribution for inner PMTs the
Montecarlo one are in agreement, as well as the
outer ones
DATA MONTECARLO
Q (a.u.)
14
Low energy events on PMT spectrum
RUN 236, 237, 239
Bar 16 (first of upstream TC) Outer PMT Bar 19
(fourth of upstream TC) Outer PMT
Q (a.u.)
Data Montecarlo
Q (a.u.)
15
TRG vs DRS height pulse correlation
Key for Type3 boards and LXe acquisition, we can
see a good linearity until 0.6 mV (0.1 V on DRS ?
0.4 V on TRG) (RUN 236, 237, 239).
(V)
(V)
16
Muon rate vs beam intensity
17
Charge distribution (1)
Calibrated by superimposing charge peak position
of MC events on experimental data
18
Charge distribution (2)
Charge
Run 154 Low Intensity
Charge
Run 236 High Intensity
Data wider probably for difficulties of
individual software calibration
19
Number of hit bar per event
Run 154 Low Intensity
20
Number of hits on each bar
Run 236 High Intensity
10 KHz
Down stream
Up stream
KHz
Down stream
Up stream
Run 154 Low Intensity
21
Position reconstruction by means of charge ratio
PSI setup
In COBRA, with and without Beam acquisition
with oscilloscope of waveform from PMs with
single bar triggering, and evaluation of charge
ratio for each pulse, with no selection of impact
position distribution of log(Qin/Qout) for each
bar reflects event distribution on each bar
width of the distribution is inversely
proportional to attenuation length of bar gt
clue to the applicability of position
reconstruction by means of charge division
charge distribution on a bar from cosmic rays
22
waveform analysis - continue
Beam turned on event distribution on 4 US bars
most of the events are on the inner part of the
detector needed calibration of each bar to
determine impact point.
Inner PMT
Outer PMT
23
CAMAC results position correlations
correlation between adjacent bars, distribution
of charge ratio on bar 1 for selected impact
position on bar 0
24
CAMAC results position correlations
correlation between adjacent bars US TC (left)
and DS TC(right) curves are fit to the data
25
Baseline noise analysis 1
yes
no
low threshold
no
yes

• Threshold ?
• Range ?
• How stable ?
• How often ?

26
Baseline noise analysis 2 Stability

• Noise level 0.6 1mV
• Baseline fluctuation in a run 0.3mV
• Baseline shift
• with change of beam intensity
• with change of trigger type
• Baseline is stable among same condition runs
• We need baseline value in database each run
  condition.

27
Pulse shape of PMT signal

• Made one template waveform for all channel.
• Fitting seem to work well

• Not yet making template for each channel
• Feed back to simulation parameter.

28
First Look at Beam data

• Michel positron from beam, TC trigger
• 2GHz sampling data, high intensity beam
  (79,80,81)

• Trigger related pulses are located 120150 nsec
  in DRS window.
• Trigger latency 370 nsec, enough for baseline
  calculation.
• Pulse height peak 60mV, tail with Landau
  Distribution.
• about 2.2 pulses are saturated.

29
Z reconstruction

• Reconstruct z-position by time difference of both
  side PMTs

• Roughly estimate
• Effective light velocity (width)
• Relative time pedestal (center)

30
Z position correction

• Effective velocity 15 cm/ns
• Time pedestals in database

31
Relative gain estimation

• Using spectra, estimated relative gain.
• Pulse height after correction by position

• Fit with Landau distribution
• Compare the MPV

32
Height ratio

• If pulse height ratio is mainly determined by
  attenuation length,

• Roughly extracted effective attenuation length
• Att 96 cm