Status of the Test System of the MWPC for the LHCb Muon System
description
Transcript of Status of the Test System of the MWPC for the LHCb Muon System
1
Status of the Test System of the MWPC for the LHCb Muon
System
Andre’ Massafferri
(Universita’ Tor Vergata – INFN sezione II - Roma)
V. Bocci, R. Nobrega (Universita’ La Sapienza – INFN sezione I - Roma)
G. Carboni, E. Santovetti (Universita’ Tor Vergata – INFN sezione II - Roma)
IEEE San Diego 2006
2
5 Muon Stations
Calorimeters
Tracker RICH-2
Vertex Locator
Magnet
RICH-1
Muon Detectors
Iron Filters
Chamber
MWPC in Muon System
MWPC task Fast muon triggering Muon identification
5 Stations divided in 4 Regions 19 different geometries 1368 chambers More than 120K channels
LHCb: an experiment for precise measurements of CP violation and B
mesons rare decays
3
GoalFriendly and automatic procedure for diagnose
of errors of the Chamber + Electronics system
Check of our 1368 MWPC(~ 128000 channels) before installation
Database for Online Monitoring
1 - Chamber & Front-End Electronic (FEE)
2 - Test setup and description
3 - Results
4 - Cross check using cosmics
5 - Conclusion
Topics
4
MWPC Design 40 pF < Cdet < 220 pF (geometry) Wire (Anode) & Pad (Cathode) readout 4-gaps MWPC gap size: 5 mm (wire plane centered) gas mixture: Ar/CO2/CF4 (40:55:5) wire spacing: 2 mm, mechanical tension: 65 gr HV = 2.650 KV gas gain: G ≈ 50 000 gain uniformity: ≤ 30%
Not wired panel
Wired panel
Side bar
Wire fixation bar
Spacer
MWPC Sandwich
M5R3 full equipped
M3R3 panel
6 FEEsSingle padFaraday cage
5
Chamber & Front-End
Electronics (FEE)
CARIOCA: 8 chs current-mode Ampl • signal amplification• tail cancellation• base line restoration• digitalization into LVDS lines
DIALOG: 16 chs control chip
• 8-bits DACs for threshold voltage• width and delay adjustment• masking• 24-bits scaler • pulse injection feature• access via LVDS-based I2C protocol
The acquisition basic unitChamber: 2
bigapsLVDS
READOUTCh(AB)OR
ORSPB
CONTROL
Ch ACONTROL
Ch B
Single padlogic
Spark protection board
2 CARIOCAs + DIALOG
= FEE (CARDIAC)
6
Band-width: 10 / 25 MHz depending on Polarity Sensitivity: 16 to 8 mV/fC ENC: 0.3 to 2 fC Min Detectable charge: about 40
mV/sens
FEE characteristic I
charge Anode Q > 14 fCCathode Q > 8 fC
Qth
µ signal
time
dead region
noise
Min Det
charge
AFTER TEST FEE DATABASE
Capacitance (pF)
Capacitance (pF)
Capacitance (pF)
7
The Discriminator Differential Threshold Voltage
o Less sensitive to Noiseo Vth Nominal: 0 to 1.3 V o Vth Effective Abs(VrefA – VrefB) VrefA & VrefB obtained from Vth Nominal
o OffSet
(Vth Nominal @ Min Vth Effective)
740 to 860mV
FEE characteristic II
offset Vth Nominal
VrefA
VrefB
Vth Eff
Vth Nominal
Noise Rate
Analysis of
Noise X Threshold
Rate method
8
Test Setup Control FEE via I2C (Service Board &
CANopen) Internal Counters (Dialog feature) External Counters (ACQ & Gate Boards and USB-VME) BarCode Reader PC (WIN, Visual C++ & ROOT) Barcode and Test Program
Full Equipped Chamber (UNDER TEST)
FEE
ACQ
LVDS
TTL
Gate Board Task
input
output
UNDER TEST
USB
Service Board (SB) controls FEE in the experiment.
ACQ is a 64-channels VME module used as external counter.
Gate Board translates the SB gate signal sent to the FEE to be used also by the ACQ.
MWPC
FEE
FEE
FEE
FEE
8X16X
9
Association of Chamber to FEE boards LOCATION
DATABASE
Extraction of Chamber and FEE parameters
The Test
Protocol
Read the Chamber and FEE barcodes
USER TASK
In dressing area(1)
Preparation
SOFTWARE GOAL
Complete Test
Initialize
Press Check LVDS output line
Check Chamber-FEE association
Search for Dead Channels
Estimate Cdet (Rate Method)
Search for Short-Circuit
Search for Open Channels
Check Level Noise @ 3 specific
thresholds
Read Chamber barcode
In test area
(2) TestProcedure
10
Check if FEE is working properly (pulse injection, internal counters, output lines)
Inject 500 pulses to all channels
Reading by dialog & ACQ counters
Comparison
First Step
Check LVDS output line
Cable CheckingCheck if cables are swapped
It can be useful to check cables in pit
Pulse Injection Test
11
Check the Position of FEE comparison of OFFSET values
obtainedhere to the ones found on FEE
database
Check dead FEE channelsby the simple existence of Noise (min
3 pts)
Second Step
Threshold Scan Test
Check Chamber-FEE association
Search for Dead Channels
12
Estimate Cdet (Rate Method)
Search for Short-Circuit
Search for Open Channels
Third Step: Cdet
The detector capacitance determines the noiselevel since it acts as aseries noise source
SPB 2 nF
bigap
Threshold Scan Test
Rate=
Nexp 2 ENV2
-Vth2
offsetMin Det signalOffset &
Min detect signalENV ENC(Cdet)
Sensitivity(Cdet) Cdet
functionsNoise fit
Short-Circuit : Noise (2nF)~Noise (300pF) saturation due CARIOCA band-width
Open Channel : Cdet ~ 0pF easy to discriminate even for M2R2 40pF
13
<Cdet>
Third Step: Cdet
Criteria
channels
Cdet
+-
> 2.5
> 3 or
ERROR
Mostly associatedto setup itself
Very important Alarms !
Cdet > 300 pFor
Cdet < 20 pF
14
Criteria for Noise Rateat 3 specific Thresholds
OK / WARNING / ERROR
if Rate > 1 KHz
Cathode Readout: 6 7 8 fC
Anode Readout: 10 12 14 fC
Check Level Noise @ 3 specific
thresholds
Fourth
Step
Electronic Noise can be ~100 Hz / channel
1 KHz in Noisy channels is acceptableChamber Eff >99%
@ Th = 8 fC (cathode) & 14 fC (anode)
threshold
15
Results Format: DATABASE
Diagnostic file
Output file
PLots & Histograms
• Detailed description of all steps through the pre-defined ranges of: OKs , Warnings, Errors messages
• List of 33 relevant parameters, including FEE database.
• Allow further comparisons
Leds & Buttons
OKWarningERROR
Info: Where is the problem !
• Histograms of the main output parameters
• Graph of the Noise X Thr of all FEE/channels
• OffSet signature
• Vertex position (cross of each combination of two channels)
• Acquisition facility: any threshold/gate-time/ bigap-logic
• PLOT
Raw-Noise file• All points in Threshold Scan
• Can be used for further re-analysis and comparisons
16
Results: First Chambers tested
@ LNF
Cdet
Alarms !
Those tests were performed using
a MaxNoiseRate cut = 2 MHz
Recently we have observed a FEE patology that distorted the Th-
Scan at high Noise Rate
We upgrade
MaxNoiseRate = 20 MHz
to detect it
About 10% needed some intervention
pointed out by the Test System Mostly Change FEE (dead,noisy)
Also bad connection, SPB, I2C (chain)
100 - M5R4 Cdet (sim) = 220 pF22 - M3R3 Cdet (sim) = 140 pF06 - M5R2 Cdet (sim) = 120 pF
128 - chambers
17
More Plots:
Noise @ th = 10 fC & 6 fC
Noise @ th = 12 fC & 7 fC
Noise @ th = 14 fC & 8 fC
10 KHz
1 KHz
100 Hz
Noise Rate of 4 M5R4 chambers
12 CARDIACs @ th = 14fC
M5R4 & M3R3 & M5R2
OffSet
18
Cross-Check using Cosmics• Cosmic Acquisition is being done at CERN on chambers tested in LNF
Low Eff channel
2 chambers without PLATEAUHV scan, gas, Operational Threshold & signal
coincidence between 2 Bigaps
15 M5R4 analysed up to now
100% of the defects found were successfully
detected by the Upgraded Test System
High Cdet signature
More Statistic (mainly with other chamber types) must be
collected
19
Conclusions
• Automatic and fast (5-10 minutes) system has been implemented to be used also for non-experts
• System has shown to be very effective on 130 INFN chambers tested: fundamental guide to fix problematic chambers (10%)
• Cosmic Acquisition pointed an unexpected kind of problem in 0.5% of the channels. A retest using the Upgraded version of Test System has been able to discriminate all those channels
• Now we have 4 systems operating: CERN (2), LNF (1) and in the pit (1)
• We aim to minimize drastically the installation of chambers with problems