Transcript psd_Feb2013 - Indico

Projectile Spectator Detector:
Status and Plans
A.Ivashkin (INR, Moscow)
• PSD performance in Be run.
• Problems and drawbacks.
• Future steps.
1
PSD – Projectile Spectator Detector
Compensating calorimeter
PSD in 2011
•Pb/scintillator (4/1) 60
sandwiches in one module
•Modules 10 x 10 x 120 cm3 –
central part
• Modules 20 x 20 x 120 cm3outer part
• 10 longitudinal sections with
10 MAPDs readout
Precise measurement of the energy
of projectile spectators.
•Centrality selection (on trigger level)
• Measurement of event-by-event
fluctuations (to reduce Npart
fluctuations)
• Reconstruction of the reaction plane
PSD modules in
2011 Be-run
16 central modules
in trigger
2
Photo of PSD after full construction (Feb.2012)
Rear side
Front
before FEE
installation
44 modules
440 MAPDs and readout
channels
17 tons
3
7Be
spectra in PSD at low energies
13 GeV
20 GeV
30 GeV
Good identification of beam ions.
But the resolution is still ~20% worse of
expected numbers.
Needs the improvement of calibration.
Contribution of electronic noise is maximum at
low energies.
4
Temperature dependent 7Be peak position
Energy in PSD
25.01 – 26.01
27.01 – 28.01
29.01 - 04.02
MAPD gain ~3%/0C
Temperature
Time dependent calibration is needed!
A few variants are considered.
What actions are needed for the improvement of PSD
performance?
1. Improvement of temperature stabilization
2. Modification of trigger signal
3. Control on MAPD’s voltages
4. Full upgrade of readout
5. New approach in calibration with cosmic muons
6
Temperature stabilization
At present: The flow of underground air trough the distributor box
and pipes in each PSD module.
Problems:
• the temperature of underground air is not constant.
• ~6 0C difference between summer/winter.
• ~2 0C spread during the run.
• Temperature gradient ~2.5-3 0C between top/down modules due
to different length of pipes and air pressure in modules.
What can be done:
1. Improvement of present cooling system: Chiller before the
distributor box (accuracy ~1 0C ), optimization of the pressure
and pipe length for air. (The simplest and cheapest solution)
2. Full modification – installation of Peltier elements in each
module (accuracy ~0.1 0C ).
7
Modification of PSD trigger signal
MAPD signal
M~5x104
integrated signal M~107
ADC signal
300 ns
60 ns
Plans: trigger
signal after MAPD
– fast signal, no
problem with time
walk and delay.
At present: the PSD trigger
signal comes after integrators
with rise time ~60 ns: rather
slow and large time-amplitude
walk. Needs careful
adjustment for each beam
energy and each beam ion.
This action requires the replacement of present FEE!
New amplifiers, new adders!
8
Control of MAPD voltages
•
•
MAPD voltages must be stable within 0.05V accuracy.
At present two power sources in one chain (reference and variable)
are used for all PSD modules.
• What can be done:
1. Slow control of output voltages after power sources. – Only 2
numbers, Rather simple solution. But the real voltage at MAPD would
be not measured.
Or
2. Control of voltages in each individual MAPD.
Installation of ADC in each of 440 MAPDs.
Full upgrade of slow control, replacement of FEE.
9
Approach in calibration with cosmic muons
Test of one PSD section reveals
peak in energy spectrum from
cosmic muons
Can we apply this
approach for NA61?
10
Full replacement of FEE?
• Modification of trigger or control of individual MAPD voltages
requires full replacement of PSD FEE.
• In this case one needs to develop new amplifiers, adders, slow
control (hardware and software).
• The readout would be done in another way (DRS, TRB3…)
• The resources and manpower are needed.
• In present FEE one needs to repair two sections in one outer
modules (short circuit during intervention), to repair one adder
in central module (result of power cut at CERN).
Description of present FEE in:
https://twiki.cern.ch/twiki/bin/viewauth/NA61/PsdHardware
11
Some inputs for new FEE
•
•
•
•
•
•
•
•
•
•
Number of modules - 44
Number of channels – 440 ( 10 in each module)
Geometrical size of one PCB 95x95 mm2 with holes in the corners.
Could be a few PCBs or mezzanine boards.
Vop of MAPD: 62-68 V
Accuracy of Vset ~0.05V
Signal – rise time 8-10 ns, full length 60 ns.
Slow control must write and readout the MAPD voltages.
Amplifier: gain~100-120, bandwidth~20 MHz, dynamical range 5mV1.5V. Two outputs – 9/10 parts for DRS and 1/10 for adder. Would be
nice to have the gain switch from 100-120 to 20-30. 10 amplifiers in one
module.
Analog adder – 10 channels with weight ~1. Output – LEMO
connector.
The electronic must ensure the gain/amplitude stability of about 1% in
the temperature range 15-30 0C.
12
The last item to be fixed – remote control of
PSD moving table
Thank you!
13