ATLAS Upgrade R&D and Plans
July 7, 2008 SLAC Annual Program Review Page 1
Meeting the Challenge of superLHC
gives ~40% higher
effective energy reach
for new physics
2013: Pixel b-layer replacement
2013-15: Improved L1 Trigger
New Si tracker for SLHC
Trigger DAQ upgrade
LAr and Muon upgrades
Intense radiation environment: ~2x1016 P/cm2 (10yr@1035) at R=4cm
Dramatic pileup background: ~400 interactions/crossing
July 7, 2008 SLAC Annual Program Review Page 2
SLAC Involvement in superLHC Upgrade
* superLHC is probably an inevitable path for HEP:
– Early discoveries would imply the effective extra energy reach at
high lumi could uncover additional new particles. Given the present
investment, SLAC should continue to sLHC, with complementary
physics to Linear Collider.
– If early phase of LHC not revealing new physics, it would be hard to
argue for ILC as is or other new facilities. sLHC will be taking on an
even central focus of hope for new physics through its effective
additional energy reach.
* The time to be involved is Now. Schedule for upgrade design and
construction is already not optimistic. With the end of B-
factory operation and launch of GLAST, this becomes a natural
interest among many staff members.
* The upgrade activities at SLAC will be a key addition to better
utilize the SLAC resources and expertise to complement current
ATLAS efforts. The joint activities of physics, operations and
upgrade will combine to a role appropriate for a national lab and
encourage growth of user presence at SLAC.
* Important detector training ground for postdocs and students.
July 7, 2008 SLAC Annual Program Review Page 3
sLHC Tracker Upgrade: Introduction
Long (from a early
strip SLAC study of
* The entire inner detector of pixel+SCT+TRT will be replaced
with an all silicon tracking system for sLHC:
– Inner most pixel layer(s) need new sensor technology
– Outer tracking with short and long silicon strips
– Increased granularity for higher particle density demands new
approaches to cooling, power deliver, data transmission.
– Aiming at lower material budget.
* SLAC has silicon experience from MK-II, SLD (pixel), GLAST
and future silicon detector design of SiD.
July 7, 2008 SLAC Annual Program Review Page 4
Tracker Upgrade: 3D Sensor for pixel
* 3D Sensor technology pioneered by Sherwood
Parker et al is a primary candidate for inner
most pixel layer:
– Radiation hardness
– Active Edge
* The 3D R&D devices are produced at the
Stanford Center of Integrated Systems. Future ATLAS 3D
sensors designs are also expected to have test samples from
Stanford first for testing before production at SINTEF.
* SLAC is participating in the Jun-Jul/08 3D beam test at CERN,
and performing 3D sensor simulation.
* SLAC is developing upgrade pixel teststand with compatibility to
current and future ATLAS pixel readout electronics. Ideally
located for first testing of 3D devices from Stanford.
* Preparing proposal to join the 3D sensor collaboration.
July 7, 2008 SLAC Annual Program Review Page 5
Tracker Upgrade: CO2 Cooling (I)
* The commissioning of the current evaporative C3F8 cooling
system has been a rather difficult experience. Cooling becomes
a central focus for the larger and higher granularity silicon
system for sLHC with bigger challenge to avoid thermal runaway.
* CO2 cooling is widely perceived as a better alternative to the
current C3F8 cooling system and already chosen as baseline for
strip detector upgrade. High latent heat and high heat and high
vapor pressure allow efficient heat transfer with smaller pipes
for reduced material. Also a more environment friendly solution.
* Despite the priority and broad interests, very little has
happened on CO2 cooling. More practical for a national lab.
* SLAC engineer Marco Oriunno visited NIKHEF, CERN LHCb CO2
cooling system to engage in cooling system design for ATLAS.
* CO2 cooling is also the new trend in industry and synergy with
July 7, 2008 SLAC Annual Program Review Page 6
Tracker Upgrade: CO2 Cooling (II)
* SLAC’s plan for CO2 cooling involvement:
– Step 1: Build an open loop small refrigeration system to study
basic design features
– Step 2: Serve the detector mechanical prototype cooling
tests to iterate on design choices for piping/fitting and the
– Step 3: A small pilot cooling plant at SLAC to serve stave
tests and develop/validate the overall cooling system design.
* SLAC is geographically well situated between LBNL
and Santa Cruz as a center for the various cooling
related tests for both pixel and strip detectors.
* It’s a natural gateway to broader involvement in
mechanical system designs.
July 7, 2008 SLAC Annual Program Review Page 7
Tracker Upgrade: Data Transmission
* The high hit density and high radiation at sLHC poses
new challenges to data transmission.
* Current pixel optical data transmission elements will
not survive sLHC radiation dose, and even more
difficult to work for the colder operating
temperature of -350C expected from CO2 cooling.
* SLAC is leading the alternative technology R&D with
multi-Gb/s electrical transmission through microCoax
cables. Already demonstrated 4Gb/s on a rad-hard
microCoax with bit error tests on FPGA test boards.
* Doing radiation tests on better dielectric sample to
validate a custom design with more favorable material
budget and even better transmission bandwidth.
=> See Martin Kocian’s breakout session talk for details.
July 7, 2008 SLAC Annual Program Review Page 8
Tracker Upgrade: Teststand/DAQ
* Teststand will be a central point of integration to
investigate designs and execute production. A
versatile and performant test system is essential.
* SLAC has current pixel Turbo DAQ system running
and helping on improving the LBNL strip stave
* Investigating a teststand system with more on board
processing power to speed up tests. Multi-channel
readout for stave tests and cosmic telescope.
* Thinking more generally to serve pixel and strip tests
with the evolution to new Read Out Driver in mind.
* Based on extensive DAQ design experience at SLAC.
July 7, 2008 SLAC Annual Program Review Page 9
Upgrade Tracker: Strip Stave Design
* The development of the strip detector barrel stave is one of the
most central tasks in tracker upgrade.
* The Jun/08 stave review at CERN selected the stave concept
pursued by Carl Haber et al at LBNL based on CDF runIIb
upgrade with lower mass and more advanced electrical
* Dave Nelson from SLAC has been helping on teststand DAQ
improvements and other electrical integration issues.
* SLAC is interested in becoming a significant partner in the strip
stave effort. Besides the electrical integration and testing,
another possible area of contribution is mechanical design and
testing in concert with the CO2 cooling development.
* Based on SiD and past silicon experience and well located
between LBNL and Santa Cruz.
July 7, 2008 SLAC Annual Program Review Page 10
Pixel B-Layer Replacement
* Before superLHC, pixel b-layer is expected to reach
its radiation dose limit by ~300fb-1, so that a
replacement needs to be planned for 2012.
* Can no long do simple b-layer replacement with <9
months shutdown. The b-layer task force reached
recommendation for an insertion of new b-layer inside
present detector on a smaller beam pipe.
* Requires some of the new technology for sLHC
upgrade to be ready in time for this to be viable.
* SLAC may naturally become part of the project if
seriously moving ahead.
July 7, 2008 SLAC Annual Program Review Page 11
Tracker Upgrade: Simulation and Layout
* The tracker design layout is a central activity which
critically relies on simulation. The local design choices
also inevitably rely on simulation for selecting
* The upgrade simulation has become a universal source
of frustration due the inflexibility for introducing
new geometry and heavy CPU demand.
* The LCsim framework developed at SLAC for Linear
Collider Detector designs, has very flexible and
simple geometry input for producing fast design
* We are investigating a strategy to significantly
improve the simulation utilities for ATLAS upgrade
leveraging the strong simulation expertise at SLAC.
July 7, 2008 SLAC Annual Program Review Page 12
Trigger/DAQ: Current system
July 7, 2008 SLAC Annual Program Review Page 13
Trigger/DAQ Upgrade: Introduction
* Trigger/DAQ upgrades/improvements are inevitable
for ATLAS like every other experiment, but people
are very busy with current system so far.
* Known prospects at Phase 1 (by 2013)
– some L1 improvements Calo, Muon, Central trigger
– Level 1.5 Fast Track Trigger
* Known prospects for Phase 2 (by 2017)
– Rebuild of many frontend electronics to allow more L1
improvements with longer latency
– ROD/ROL/ROB/ROS rebuild to improve bandwidth
– L1 track trigger ?
– Allow L1 rate to go >100KHz beyond the baseline ? Must
decide soon or upgrade tracker design may prohibit this.
July 7, 2008 SLAC Annual Program Review Page 14
Trigger/DAQ Upgrade: New Readout System
* Current Read Out Drivers (ROD) have 7 different flavors for
different subsystems, which no one is proud of. They cannot last
* The Read Out Link (ROL) limits L1 rate to <100Khz.
* Mike Huffer (SLAC) has a draft proposal to condense the entire
ROD/ROL/ROB/ROS system into one layer of new ATCA
(Advanced Telecommunications Computing Architecture) based
– Reduces 932 RODs-> 234 ROMs
– Can absorbe L2supervisor and ROIB as well
– Improving bandwidth to L2/EF from 8GB/s -> nx45GB/s.
* More intriguingly, the improved local density and high bandwidth
of inter-module data availability is what one looks for in an ideal
* Based on extensive DAQ experience at SLAC and the Petacache
project for fast random data access.
July 7, 2008 SLAC Annual Program Review Page 15
Trigger/DAQ Upgrade: New Readout Crate
July 7, 2008 SLAC Annual Program Review Page 16
Trigger/DAQ Upgrade: Petacache system
Prototype Petacache system
Module Reconfigurable Cluster
Element (RCE) module
July 7, 2008 SLAC Annual Program Review Page 17
Synergy between Projects
The choice of projects had the efficient use of lab
resource in mind to best utilize our expertise:
* CO2 cooling, data transmission, teststand/DAQ are
relevant for both strip and pixel detectors.
* Teststand/DAQ, Gb/s transmission, Trigger/DAQ
upgrade are based on the general Trigger/DAQ
* We strongly believe in the need to open up L1
bandwidth and working on the two key enabling
aspects: tracker data transmission and DAQ upgrade.
July 7, 2008 SLAC Annual Program Review Page 18
Summary and Outlook
* We have identified many interesting directions
making significant contributions to the ATLAS
upgrades and started some detailed R&D and design.
* We believe these directions speak to the real needs
to complement existing upgrade effort.
* There are strong synergies among the investigated
projects to maximize utilization of SLAC expertise.
* We intend to keep the broad vision of the overall
ATLAS upgrade needs and pay attention to system
design issues. We believe SLAC can play a major role
in the ATLAS upgrade.
July 7, 2008 SLAC Annual Program Review Page 19