Diapositiva 1

Download Report

Transcript Diapositiva 1

Time Calibration
with
Optical Beacons
Pylos, 16-19 Apr 2007
C.Bigongiari IFIC (CSIC – Universitat de València)
KM3NeT WP3 – Calibration Session
Outline
• Time Calibration
• Optical Beacons
– LED Beacons
– LASER Beacons
• What we have learnt in ANTARES up to now
– Time resolution
– Optical Beacon illuminating same line OMs
– Optical Beacon illuminating other line OMs
• Laser OB development
• Conclusions
Pylos, 16-19/04/2007
C. Bigongiari
2
Time Calibration
The reconstruction of muons
trajectories in a neutrino
telescope heavily relies on the
measurements of Cherenkov
photons arrival times.
μ
ν
Pylos, 16-19/04/2007
A precise relative time calibration of
the detector is of utmost importance
to achieve a good angular resolution
and a high reconstruction efficiency
C. Bigongiari
3
Time Calibration
 Different systems are needed to measure time
delays and time jitters in different parts of the
electronic chain, from PMTs to DAQ.
 A redundant time calibration system is very
useful to disentangle different effects
 In-situ calibration systems are mandatory to
monitor the time calibration after the
deployment
Pylos, 16-19/04/2007
C. Bigongiari
4
ANTARES Optical Beacons
Optical Beacon  Well controlled pulsed light source,
(LEDs or LASERs)
LEDs
LASER
LASER Beacon
LED Beacon
Pylos, 16-19/04/2007
C. Bigongiari
5
ANTARES LED Beacon Features
LED Beacon without container
TOP
 6 faces with 6 LED each = 36 LED.
 3 groups (top, centre, four)
FOUR
CENTRE
 Internal PMT Hamamatsu H6780-03
(RT=0.7 ns) to know the actual time
emission of the light.
 Wave length emission = 472 nm (blue)
 Flux per shot @ max INT = 4 x 108
photons per LED.
 Intensity: variable
 Light emission: isotropic in
50º < θ < 120º range
 Location along the line
(storeys: 2, 9, 15, 21).
Pylos, 16-19/04/2007
C. Bigongiari
6
ANTARES LASER Beacon Features
Glass rod to avoid biofouling
 Internal fast photodiode (jitter ~50 ps) to
measure the actual time emission of the light.
 Wave length emission = 532 nm (green).
 Flux per shot = 1012 photons.
 Intensity: fixed. Adjustable in new design
 Light emission: Lambertian
 Location at string bottom
Pylos, 16-19/04/2007
C. Bigongiari
7
LED Optical Beacons
• Advantages:
– There are Blue LEDs
(472 ± 15) nm
Absorption length ~ 60 m
Effective scattering length > 200 m
– LED light yield is tunable
– LEDs are cheap
Pylos, 16-19/04/2007
• Disadvantages:
C. Bigongiari
– Long rise time 2 ns
– Expensive containers
– Low yield => Many LEDs
•
•
•
•
Cumbersome mounting
Source spatially spread
Hazy proper T0
Dull and tricky synchronization
procedure needed
8
LASER Optical Beacons
• Disadvantages:
• Advantages:
– Extremely coherent light
source
– High light yield
• No synchronization needed
• Well defined proper T0
– Green light 532 nm
Absorption length ~ 28 m
– The emitted light is not
isotropic
– Expensive (~ 12K€ )
– Very good rise time (<0.5 ns)
– Light yield is tunable
(see following slides)
Pylos, 16-19/04/2007
C. Bigongiari
9
Time Calibration with Optical Beacons
Shore Station
Photons emitted by Optical Beacons propagate through sea
water and, if they reach OMs, can produce electronic signals
in pretty the same way as Cherenkov photons do.
Time calibration with Optical Beacons is therefore a very
comprehensive system:
– Many different aspects are involved
•
•
•
•
Sea-water properties
Detector positioning (Rotation included)
Other timing systems
OM response
• Readout system
Junction box
Pylos, 16-19/04/2007
C. Bigongiari
10
What we have learnt from ANTARES
We studied the distribution of T1- T0
– T0 = Light emission time by OB
– T1 = Arrival time of light on OM
Optical Beacon OB
Optical Module OM
• In the following results about:
• Electronic chain time resolution
• OB flashing same line OMs
• OB flashing other line OMs
• LASER beacon flashing
Pylos, 16-19/04/2007
C. Bigongiari
11
Floor 21
Floor 15
Floor 9
Floor 2
Sector 5
Sector 3
Sector 2
Sector 1
LED
ANTARES
Optical Beacons
Layout
~87 m
~87 m

7+1 Lines deployed so far

5+1 Lines connected

20+9 LED Optical Beacons.

1 Laser Optical Beacon.
~101 m
~94 m
~114 m
Line 5
Line 1
MILOM
Line 2
Line 4
Line 3
Laser
Pylos, 16-19/04/2007
C. Bigongiari
12
TOM – TOB (ANTARES Data)
The distribution is clearly asymmetric due to the photon scattering. We consider only the
rising part, which is less affected by scattering, early photon effect, and fit it with a
Gaussian. Better results can be obtained with a two-steps fit:
Second step: Gaussian from 10% of
First step: Gaussian fit from 20% of
the mean bin content to the 90% over
maximum to the 100 %
passed bin.
100%
90%
20%
T100 = Gaussian mean
Pylos, 16-19/04/2007
10%
Sigma = Gaussian sigma
C. Bigongiari
T50 = Time @ 50%
13
Time Resolution
Flashing OMs in the storey above the OB at high intensity we can measure the
electronic chain contribution to the overall time resolution.
The σ of the distribution is well below the requested resolution, 0.5
ns, for all the OMs
Pylos, 16-19/04/2007
C. Bigongiari
14
OB illuminating same line OMs
OB
OB
OB
OB
OB
OB
1. A LED OB is able to illuminate its own storey OMs and even the ones in the storey below
2. The statistic is enough to perform the fit up to 8 storeys above (about 116m far away)
Pylos, 16-19/04/2007
C. Bigongiari
16
OB illuminating same line OMs
Number of events [a.u.]
The T100 grows as function of the distance/storey (up to 10 ns) more or less
linearly with slope ~1.7 ns/storey
The observed slope is due to the
‘early photons’ effect
high
intensity
~ 9 ns
low
intensity
The RISE TIME of the light source
must be smaller that the
requested time resolution
Or
Illuminate OMs at 1 phe level
Dt [ns]
Pylos, 16-19/04/2007
C. Bigongiari
17
Sigma as function of OB-OM distance
The sigma of the fit is a measure of the direct photons peak width. This is the result of
the convolution of source time spread (~1.7 ns) and phototube TTS (~1.3 ns).
TTS depend on the number of photoelectrons Npe, and therefore on the OB-OM distance.
At large distance we reach the phe level -> Sigma =√(1.7ˆ2 + 1.3ˆ2) = 2.1 ns
Pylos, 16-19/04/2007
C. Bigongiari
18
OB illuminating another line OMs
OMs in storeys 1 to 13 of line 3 illuminated by the beacon of storey 2 lines 5.
There is the same linear
dependency on the OB-OM
distance of the previous case
due to early photon effect.
There is an anti-correlation
between OMs in the same
storey due to storey rotation
We can notice some wrong T0s
We need an independent measurement of storey position/rotation
Pylos, 16-19/04/2007
C. Bigongiari
19
Laser OB – Correcting for position
Not Corrected
σ = 2.3 ns
Corrected
300
m
60 m
σ = 0.6 ns
Pylos, 16-19/04/2007
C. Bigongiari
20
LASER Beacon Development
The light emitted by the LASER
can be varied using a Voltage
Controlled Optical Attenuator, a
linear polarizer followed by a
liquid-crystal retarder and
another linear polarizer.
Varying the voltage applied
to the retarder the
polarization of outgoing light
changes.
In this way the transmission
of the attenuator can be
varied.
Pylos, 16-19/04/2007
C. Bigongiari
21
Variable intensity LASER Beacon
Variable
Voltage
Laser Head
Liquid Crystal
Head
Polarizing cube
beam-splitter
Schematic view of the
Variable Intensity Laser
Beacon.
The amount of outgoing light
can be changed by varying
the voltage V applied to the
liquid crystal retarder
Liquid Crystal Retarder
Polarizing Beam-Splitter
Pylos, 16-19/04/2007
C. Bigongiari
22
Variable Intensity LASER Beacon
We measured the energy per pulse emitted at
different pulsing frequencies as function of the
applied voltage. The maximum output is
above 1 μJ for all frequencies considered.
A variable intensity LASER Beacon
has been already installed on line 7
Pylos, 16-19/04/2007
C. Bigongiari
23
Conclusions (1/2)
• OB allow an in-situ time calibration and
monitoring of the detector
• From ANTARES data we have learnt:
– The overall time resolution is below 0.5 ns
– LED_OB-OM time difference depends on the distance
due to early photons effect
• Very short rise time light sources are needed
• Otherwise operate at 1 phe level to avoid this effect
• The Optical Beacon system is very
comprehensive
– Sensitive to Optical Module position (Rotation
included)
– Sensitive to sea-water conditions
Pylos, 16-19/04/2007
C. Bigongiari
24
Conclusions (2/2)
• This is an advantage:
– We can get information on other aspects of the detector
• Attenuation and scattering lengths of water
• Efficiency of bright point reconstruction
• Cross-check of position measurements
• But it has also drawbacks:
– What you get is the convolution of many different phenomena
– It’s hard to get precise results without additional information
• Optical Beacons are not cheap -> Reduce cost.
– Containers make a non negligible fraction of the cost. Study
alternative solutions.
– Mass production should help Km3Net to reduce the cost
• A variable intensity LASER beacon already realized,
tested and installed
Pylos, 16-19/04/2007
C. Bigongiari
25
Optical Beacon Cost
• LED Beacon Cost
–
–
–
–
• Laser Beacon Cost
Container
4000 €
Mounting
600 €
Faces (6 x 40) 240 €
Motherboard
300 €
– PMT
600 €
TOTAL
5740 €
– Laser
– Container
– Optical attenuator
– Electronics
TOTAL
5000 €
4500 €
2000 €
500 €
12000 €
Pressure and Climatic internal tests not included
Pylos, 16-19/04/2007
C. Bigongiari
26
Overdrive Mode
~33Hz
The LED OB flashing frequency
has been recently increased by a
factor ten
~330Hz
Everything worked as expected
Pylos, 16-19/04/2007
C. Bigongiari
27
Correcting for position
•
•
•
•
•
•
Fixed geometry takes, for the position of the OMs in a storey, the point in the centre
of the OMs plane.
Correction for position takes:
Geometric constants from ANTARES-CALI-2006-002 (G. Lelaizant)
Rotation matrix from ANTARES-SLOW/1999-001 (F. Cassol)
Euler angles (A1,A2,A3) from table “ALIGNMENT_VALUES2”. They are referred to the centre of
the OMs plane (0,0, 0.576)
We took the ALIG_VALUES which are closer in time w.r.t. the start of the OB run.
rOM_0=(0.437, 0, 0)
rOM_2=(-0.218, 0.378, 0)
rOM_1=(-0.218, -0.378, 0) rLOB=(0, 0, 1.003)
rcorr = R * ri + r
Y
X
Pylos, 16-19/04/2007
C. Bigongiari
28
Effect of scattering
ANTARES-Cali/2000-005
From MonteCarlo we know that scattering has a minor effect in the rising edge of time
difference distributions, although systematic shift of in T100 peaks of +0.5 ns is expected.
However delays larger than ~3 ns are unlikely due to scattering.
~1 ns
The T100 delay depends on the increasing fraction of scattered photons as we move
away from the light source. Therefore T100 depends on the water properties. However a
delay greater than 3 ns is not expected.
Pylos, 16-19/04/2007
C. Bigongiari
29
Early Photons Effect
•
Order statistic:
– Theorem:
• If X1,…,Xn are r.v. following with density f and distribution function F,
then the minimum has density function:
n(1  F ( x)) n1 f ( x)
Naive Monte-Carlo simulation
(No real data):
The time measured in the OM is
given by the early photons
There is a backward shift in the arrival
time distribution which is function of
the Npe
All Monte-Carlo (Calibob) simulations
were done at phe level, hence this
effect was not considered
Pylos, 16-19/04/2007
C. Bigongiari
30