TRD Status Report general status full scale prototype production radiator readout chamber integration issues results from test beams gas system electronics analog – pre-amplifier shaper digital – ADC development, tracklet processor system integration – MCM, readout boards power detector control system Johannes P. Wessels – GSI Darmstad simulations environment trigger performance, tracking testbeam schedule milestones
TRD Status Report. Johannes P. Wessels – GSI Darmstadt. general status full scale prototype production radiator readout chamber integration issues results from test beams gas system electronics analog – pre-amplifier shaper digital – ADC development, tracklet processor - PowerPoint PPT Presentation
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
TRD Status Report
general status full scale prototype production
radiatorreadout chamberintegration issues
results from test beams gas system electronics
analog – pre-amplifier shaperdigital – ADC development, tracklet processorsystem integration – MCM, readout boardspowerdetector control system
November `01 – presented to the LHCC January `02 – recommended for approval by LHCC February `02 – approved by Research Board
because of the funding situation -> recommendation to build `short asymmetric´version (about 58%)
Actively seeking funds at the moment; Japanese have expressed interest to join the project and are in the process of requesting support.
Chamber design
pad plane with HEXCELL/CF backing
amplification region
drift region
radiator
First pad plane support Polish manufacturer (LHCb)
bottom side: flatness better than 100 microns
top side: flatness better than 300 microns
tolerances well within needs
First pad planes
size of ordered pieces: 1200 x 512 mm2 (200 µm thickness) need 3-4 pieces/chamber depending on length of chamber manufactured by OPTIPRINT, Switzerland still problems with handling
Bending of pad plane support
sandwich with aluminum frame taped to flat table
measurement of deflection of center and sides as function of overpressure
sandwich deforms exactly as calculated using a simple model for sandwich structures
maximum bending with respect to anode wires <220microns
Expected gas gain variation
gain varies linearly with bending of pad plane
for the maximum tolerated overpressure in the chamber about 15% gain variation is expected
additional wire sag not important for a large range in cathode wire tension
partially filled with Freudenberg fibers LRP375BK (ATLAS)
Full size radiator
laminated material purchased – radiator assembly by Muenster group
Detail during chamber assembly
assembly of first full scale prototype done in Heidelberg
Full scale prototype
largest chamber: 1200 x 1600 mm
Winding machine
machine currently under test at GSI
tensioning mechanism conditioned
full-size transfer frames delivered
will be moved to Heidelberg in April
Influence of wire tension on frames
Influence of wire tension on frames
deformation of sides about 130µm each
corresponds to maximum relaxation of wire tension by 13%
0
20
40
60
80
100
120
140
160
180
200
0 200 400 600 800 1000 1200 1400 1600 1800
position along z (mm)
de
fle
cti
on
(µ
m)
TDR +10%
TDR +30%
TDR -12%
red line denotes load according to TDR
Gas feed-through - design
pressure drop for 10 chambers< 0.05 mbar
2 feed throughs per chamber10 chambers per circuit
Gas feed-through
•leakage has been tested with small prototypes•misalignment of chambers was pushed to 4 degrees between faces
at 10l/h flow and 10ppm oxygen -> full TRD would loose 0.027l/h due to seals (TDR < 0.5l/h)approx. 1000 CHF/yr
Integration into space frame
Revised super module design
Rollers
rollers now attached to L-profiles welded to space frame
reduces space necessary for support
increases stiffness of thin supermodules
same gadgets needed on installation tool
Results from 2001 testbeam
resolution of the `good´ xenon - `bad´ xenon story
momentum scan to evaluate pion rejection for various compositions of radiator materials
position resolution and angular resolution for tracking
Testbeam setup
TRD signals: pions, electrons
Good Xenon – Bad Xenon
different bottles of Xe show very different average pulse height distributions in drift region
finally managed to resolve this
long-standing problem traced to minuscule
contaminations of VERY electronegative components
<0.1ppm CF4
0.08ppm C2F6
0.8ppm O2
0.68ppm SF6 (!)
suppliers now aware of this
Status of gas system
test system with purification and recuperation operational
molecular sieves retain CO2
because of small volumes used so far, recuperation only performed off-line
test system modified; now includes buffer vessel for regulating small volumes (compressor can handle full detector -> 27m3)
system including recuperation will be used in the stack test in Fall 2002
Pion efficiency
CERN measurements scaled to reproduce GSI data at 1 GeV/c for identical radiator
pion sample still contains some (soft?) electrons (next slide)
pion rejection shown to not deteriorate for higher momenta
thick carbon fiber laminate 300 microns decreases rejection by factor 2-3. Still at 1% in relevant momentum range
new production uses only about 100 microns
Electron pion spectra
Simulation + Data - GSI
Data (CERN T10 + GSI)
Local track reconstruction
Tracking resolution
ALICE operational point at S/N = 30-40 difference in angular resolution is an artifact due to slightly
different length of the drift region in the two test chambers
Tracking resolution (B-field)
test setup at GSI with large coils
soft delta electrons curl up in magnetic field
improved point and angular resolution
Lorentz angle for Xenon
exact knowledge of Lorentz angle necessary for precise track reconstruction
data in literature not consistent
within limits posed by the precise knowledge of the CO2 content values agree with MAGBOLTZ
TRD electronics chain
Preamplifier status
Parameter TDR Simulation
No. channels 1 200 000 --
Noise (ENC) 1000 e 720 e (20pF)
Conversion Gain
12.2 mV/fC 12.5 mV/fC
Shaping time about 120 ns about 130 ns
Non-linearity < 1% < 0.25%
Power consumption
< 20 mW/ch 12.8 mW/ch
output variationswith T, Vdda
-- <0.73% (20 deg)<0.09% (200mV)
Chip area -- 22 mm2
3rd chip submitted end of January
expected back end of March
possibly last iteration
increased size due to fully differential design
Footprint of preamplifier
18 fourth order pre-amp shapers with differential outputs (21)
external references to adjust to range of ADCs digital test structure for chip verification size of chip: 3030 µm x 7280 µm
Evaluation of PASA
complete evaluation prepared; chip expected back in two weeks
some results from simulationsfull signal in center channel50% signal in adjacent pads
-> produces cross talk on <10-5 level in neighboring channels
full corner analysisonly small yield expected outside specifications (mainly leading to too short shaping time)gain variations tolerable
Digital chip - recent changes & developments
Conceptual Design Review conducted in early January complete design now for 0.18µm UMC process tail cancellation network now baseline design link width reduced to D8+P+Spare double data rate
giving up on dynamic error correction (Hamming)implementation of one spare signal allowing to multiplex out stuck bits (maximum 2 per link)
readout tree simplified configuration tree fully operable in FPGAs (2 nodes tested;
currently implementing Linux interface) MIMD CPU (1 node) operable in FPGA FPGA CPU boots UCLinux for Slow Control LVDS Rx and Tx implement power down mode now (wake-up
time ~500 ps) use of ALTRO chip as 16-channel ADC from TPC (test/fall back)
ADC – Univ. Kaiserslautern
D Q D Q
D Q D Q
...
...
S hiftreg iste r + D ecoder
O utpu t-La tches
...
...
B 9 B 0
C on tro lle r(sha red w itho ther A D C s)
Load
D R
StrobeA Z(S am p le)
2 -P hase-genera tor,Leve lsh ift
+1
-1
+1
-1
-1
-1
A m p l i f i e r b l o c k C o m p a r a t o r s
I N + I N -
V o l t a g e r e f e r e n c e( s h a r e d w i t h o t h e rA D C s )
geometry follows latest design detailed geometry of MCMs and
cooling included supermodule support structure
included; rails, services soon to come
hit compression from TPC included
tilted pad option included adapted pad response functions time response function included digital filter tail cancellation
included
MCM detail
Detail of simulated geometry
New digits display available
allows to superimposedigits from different events
Tracking
trigger performance now checked for tilted pads
fast simulations -> parameterization of response will done
global tracking strategy for all detectors IST, TPC, TRD now using Kalman filter
still some problems with tilted pads for electrons special treatment of
Bremsstrahlung needs to be implemented
electron identification and pion rejection using transition radiation and dE/dx signals need to be provided for electron tracking strategy
Future use of testbeams
2002 – 12 days (performance)stack test (at least three layers)momentum dependence of pion rejectionfinal preamp and tracklet processor on MCMfinal gas systemcooling and DCS test systems
2003 – 9 days (1/3 performance, 2/3 quality assurance)
full stack test – multi-hit capabilitytrigger performance evaluationtest of readout systemcheck of cooling and DCS systems
2004 – 8 days (1/2 calibration, 1/2 quality assurance)
several full stack testshigh statistics measurement of pion and electron spectra up to highest momenta (6 GeV/c)test of global tracking unit (GTU)test of calibration scheme
2005 – 7 days (calibration)calibration of several stacks including full momentum scancommissioning of GTU
Milestones
12-01: digital chip engineering design reviewdone in January; concept considered appropriate yet challenging; some suggestions already incorporated for submission
01-02: 3rd MPW run submission of PASAdate has been met, chips expected back in about two weeks
04-02: 1st submission of digital chip TRAP1actual submission date will be April, 17; contains track let processor, 3 ADCs, configuration tree, separate inputs for use with external ADCs
05-02: engineering design review of chamber designforeseen for end of May; date can be met if all pending tests go well
Dates for production milestones will need to be revisited if change in overall schedule will be announced