Transcript eppley_bnl_Dec1_2005..

BNL
December 1, 2005
The STAR MRPC TOF Project
Geary Eppley, with
Jo Schambach,
for the TOF Group
Design considerations
• The integration space
available for TOF in STAR is
that currently occupied by the
CTB.
• The space is about 2.2m from
the beam line: requires
precision time measurement.
• The clearance between the
TPC and EMC is about 10 cm.
• The space has a uniform 0.5T
magnetic field parallel to the
beam.
• There is limited space for
“nearby” electronics outside
the magnet steel.
Magnet
Coils
TPC
Endcap &
MWPC
Time
Projection
Chamber
Silicon
Vertex
Tracker
FTPCs
ZCal
ZCal
Endcap
Calorimeter
Vertex
Position
Detectors
Barrel EM
Calorimeter
Central
Trigger
Barrel or
TOF
RICH
MRPC development
•
•
Multigap glass RPCs developed in
the late 1990s at CERN, C.
Williams, et.al.
STAR TOF group designed STARspecific MRPC modules that
would fit existing CTB trays and
tested them at CERN in 2000 &
2001.
MRPC prototypes in STAR
•
•
•
•
•
One tray of MRPC modules has
been installed in STAR for Runs 3,
4, & 5.
An improved tray design each
year but some modules have been
installed for all 3 years.
Runs 3 & 4 used CAMAC
electronics.
Prototype HPTDC based
electronics used in Run 5
NINO/HPTDC based electronics
for Run 6/7.
Electronics integral to tray
•
•
95% of the electronics resides on
the tray inside the cover
4 external electronics boxes
located on the magnet steel
connect the 23k-channel system
to STAR trigger and DAQ
Tray specifications
•
•
•
•
•
•
•
There are 32, 6-channel modules
per tray.
The active area covers about 87%
of -0.9<eta<0.9
Occupancy in central AuAu is
~12%.
HV provides +-7 kV.
Gas mixture: 95% r134a, 5%
isobutane.
Average noise rates are ~15
Hz/channel with greater than
100% correlation.
Detection efficiency is ~95%.
Time measuring
•
•
•
•
•
Absolute times are recorded and buffered by the HPTDC for all hits above
threshold.
A single 40 MHz oscillator provides the clock for the 2882 HPTDCs in the
system.
Both start and stop detectors use the HPTDC to record times eliminating
systematic uncertainty from the choice of counter.
Each HPTDC has a 21 bit clock counter providing a 51 us clock period. The
HPTDCs are reset by a signal with a common origin at the beginning of
each run giving each HPTDC a different phase that may be learned from
the data.
Since this phase is always the same, the phase becomes a powerful tool to
monitor the integrity of the data through buffering, trigger matching, event
building, transmission to DAQ, and global event building.
Total time resolution from Run 5:
200GeV CuCu, 105ps
62GeV CuCu, 124ps
Time resolution by channel
Timing results, Runs 3, 4, &5:
Time Resolution (ps)
Operation conditions
Run III
Run IV
Run V
pVPD
TOFr
(total)
TOFr
(stop-side)
200GeV d+Au
~85
~120
~85
200GeV p+p
~140
~160
~80
62GeV (Au+Au)
~55
~105
~89
FF/RFF, w/o
E pVPD
~40
~96
~86
FF/RFF
~27
~86
~82
HF
~20
~82
~80
200GeV Cu Cu (ToT)
~ 55
~105
~89
(ToT)
~ 85
~125
~92
200GeV
(Au+Au)
62GeV Cu Cu
Manufacturing overview
•
•
•
•
•
•
•
•
Module production and test, electronics production and test, and tray
assembly and test each have a duration of ~2.2 years.
Module production begins ~3 months before the start of tray assembly and
electronics production ~1.5 months before the start of tray assembly.
Modules are thoroughly tested in China with integral HV wire and signal
cables.
Electronics cards are tested in groups of 17, 1 tray, with their integral
cables and sent to the tray assembly site as a set.
Modules are assembled into trays, the electronics added, and tested for 2-4
weeks as a complete detector unit.
Tested trays are shipped to BNL as complete detector units ready for
insertion into STAR.
Trays are re-tested at BNL: HV current, noise rates, gas leak test
Test beam at BNL for selected trays?
STAR TOF publications:
•
Physics
– Open charm yields in d+Au collisions: STAR Collaboration, Phys.Rev.Lett. 94
(2005) 062301.
– Cronin effect of identified particle at RHIC: STAR Collaboration, Phys.Lett.B 616
(2005) 8.
•
Technical
–
–
–
–
–
–
–
B. Bonner et al., Nucl.Instr.Meth.A 478 (2002) 176.
M. Shao et al., Nucl.Instr.Meth.A 492 (2002) 344.
W.J. Llope et al., Nucl.Instr.Meth.A 522 (2004) 252.
F. Geurts et al., Nucl.Instr.Meth.A 533 (2004) 60.
J. Wu et al., Nucl.Instr.Meth.A 538 (2005) 243.
Y. Wang et al., Nucl.Instr.Meth.A 538 (2005) 425.
Y.E. Zhao et al., Nucl.Instr.Meth.A 547 (2005) 334.
TOF electronics schematic
Electronic board status
Board
Current Status
Future Plans
Timeline
TPMT
Used in Run 5
Rev design: add
discriminators
Ready for production by
Q4 FY06.
TINO
Prototype designed and
built. Testing with
cosmics.
Test in test beam.
Decision milestone by Q2
FY06. Production by Q4,
FY06
TDIG
Used in Run 5.
Rev design by Q3
FY06: delete
comparators & 1 TDC,
add L0-mult.
Production by Q4 FY06
TCPU
Used in Run 5
Rev design: add TCPUTHUB I/F, delete SIU,
TCD I/Fs
Production by Q4 FY06
THUB
Prototype currently in
preliminary design
stage
Prototype with 3 I/F
channels, redesign
after bench test
Production Q1 FY07
Interface to L0 trigger
•
•
•
•
•
•
Provides multiplicity at 9.4 MHz with <~700 ns latency.
The multiplicity range is 0-12 for each half-tray, where a bit is added to the
sum if any of the 8 TOF channels in a NINO chip is above threshold.
The multiplicity sum is formed asynchronously and sent to L0 where it is
gated and readout either by the current CDB or by the DSMI directly.
The point of this is that the RHIC clock is not used anywhere by the TOF
electronics (other than to readout trigger commands).
It is not clear that the noise rates will be low enough for the TOF system to
be a useful trigger for UPC. If it does work well enough for UPC, it will still
take an enormous amount of manpower to calibrate and commission such a
trigger.
The barrel EMC is already in L0 and has a granularity of 300 compared to
240 for the CTB. It could serve as the complement to the ZDC for centrality
triggers.
Interface to L2 trigger
•
•
•
•
The intent is to provide a 23K bit map of the TOF hits to L2 for each “L0
accept” command
At the STAR review of TOF in April 2004, the L2 connection was to be
implemented by sending this information to the TOF DAQ receiver which
would pass the information to L2 over a network connection. This plan has
since been dropped by STAR DAQ.
The trigger group is now implementing a custom-designed fiber connection
(STP) from L0 to L2, and TOF could send its information to L2 over the
same type of connection.
There is no budget in the TOF project to implement this connection,
however, the THUB card has not yet been prototyped and it might be
relatively straightforward and low-cost to add the necessary components to
THUB for an STP fiber connection
L2, continued.
•
•
STAR Trigger could use the CERN/ALICE DDL link that will be the fiber
interface between all STAR detector subsystems and DAQ in the upgraded
STAR detector to receive L2 trigger data from detector subsystems. A
similar scheme will be implemented in ALICE and just requires splitting the
fiber. The software to receive data using the DDL protocol is already
implemented in STAR DAQ and could easily be implemented in trigger as
well.
The existing trigger STP fiber concentrator is already full just from L0 so a
second custom concentrator would need to be installed to accommodate
TOF, and any other subsystems.
Interface to DAQ
•
•
•
•
•
The TOF system needs to be faster than the upgraded TPC so as not to
introduce any additional dead time. The TOF information is only useful in a
STAR event if the TPC is also readout in that event.
The system will handle L0 accept commands at >10 kHz.
The system will handle L2 accept commands at >2 kHz.
A design issue with a large expense consequence is the required size of the
pre-L2 buffer. Or, how many tokens are allowed in the system for events
with TPC/TOF readout? The ALICE/CERN ALTRO has buffering for 3
events. Is STAR DAQ planning extra buffer space for TPC events
somewhere to allow for asynchronous L2 trigger processing? TOF is
currently thinking of allowing for a 256-event buffer and this costs ~$10k for
memory.
Is there a minimum dead time between L0 accept commands because of
TPC readout limitations?
Current project status:
•
•
•
•
•
•
•
•
The official closeout report from the DOE review of the TOF project in
August has been received and all the issues dealt with.
A revised management plan has been accepted by the DOE.
Cosmic ray testing of the TINO front-end board is ongoing.
Automated TINO board production has not been achieved. Trying a new
vendor.
Upgraded 38-channel start detector under construction.
Based on a recommendation from the review, the China TOF project built a
few new modules with highly resistive electrodes. Cosmic testing is
underway.
Layout of the first prototype THUB is underway.
Waiting for the DOE go-ahead to start the project.