Speaker
Description
Summary
The Compact Muon Solenoid (CMS) detector is currently being assembled at the CERN
Large Hadron Collider (LHC), due to start operation in 2007. The LHC will collide
bunches of protons with a bunch crossing frequency of 40.0789MHz at an energy of
7TeV/beam. Given the high target luminosity of 10^34/cm2/s it is of paramount
importance that data coming from all parts of the CMS detector be assigned to the
correct (same) collision. The Timing, Trigger and Control (TTC) distribution system
that provides all front-end and back-end components in CMS with the bunch clock is
thus a key component in the overall system without which all others cannot take
beam-synchronous data. There are over 10 million individual detector channels spread
over 10 sub-detectors systems that rely on the TTC distribution system. The CMS TTC
system is based upon a common LHC-wide development that has been adapted to the
specific needs of the CMS experiment. The TTC system distributes the bunch clock,
First-level Trigger (L1A) and fast control signals from a central location in the CMS
underground counting room to the sub-detector electronics located in both the
counting room and on the CMS detector itself. Fast control signals are used to
synchronize the sub-detector systems with one another and with the LHC machine – an
example signal is the “Orbit” pulse which marks the start of a turn in the machine
every 3564 clocks and is a fundamental tool for synchronization to the LHC machine’s
bunch structure.
The installed TTC system must be able to transmit the bunch clock of the LHC that is
expected to be 40.0789MHz with a tolerance of at least ±3kHz to allow for beam path
variations. Requirements on the Jitter of the clock signal come from two major
sources: the timing accuracy intrinsic to the various sub-detectors of CMS and the
requirements of high-speed serial links that are widely used in the readout systems.
For the detector types used in CMS, the former requirement is rather loose – of the
order of 1ns. It is the requirement placed on the reference clocks for synchronous
high-speed serial links that read-out data at speeds above 1Gb/s across many
sub-detector systems in CMS which provides the most stringent constraint of maximum
jitter of 350ps pk-pk. While care has been taken during the design and testing of
the electronic ASICs and modules that make up the system to ensure that these
requirements are met, the installation and commissioning of the TTC system provide
the final in-situ verification.
The top of the CMS TTC distribution system is installed over two 52U-high racks in
ten 6U VME crates. These racks are located approximately 100m below the surface in
the CMS underground counting room in a central position to minimize the signal path
for time-critical signals such as the L1A Trigger. Nine of the ten VME crates house
the distribution electronics for the sub-detectors of CMS while electronics in the
tenth receives the beam timing signals (Clock and Orbit) from the LHC RF systems and
fans those signals out to the sub-detector TTC crates. Each sub-detector crate holds
one Local Trigger Controller (LTC) VME module, up to six TTC CMS Interface (TTCci)
modules and up to six TTC Encoder and Transmitter (TTCex) modules. The LTC is used
by sub-detectors to provide the sequences of Triggers and Fast Commands necessary for
testing and commissioning in the absence of a Central (Global) Trigger system. These
sequences are translated by the TTCci module into the commands specific to that
sub-detector system and these commands are then encoded and sent to the sub-detector
over single-mode optical fiber by the TTCex modules. Fibers are installed within the
counting room to reach the racks used by the individual sub-detectors, from where the
signals are split optically and further distributed to the on-detector electronics.
Once installed, the correct functioning of the system must be verified.
Commissioning of the TTC distribution in CMS will proceed in two phases: verification
of signal integrity on installed lines that is carried out before hand-over to each
sub-detector; followed by a period of signal verification and synchronization testing
that is carried out with the sub-detectors. Software tools based on CMS-wide Data
Acquisition framework XDAQ have been put in place to allow the installers to rapidly
check the quality of the connections that are made. These are augmented by detailed
measurements of the time stability of the central system in the first instance,
followed by measurements of the signals received at the ends of the distribution
system. Once the system is sufficiently stable the integration of sub-detector
systems with the central systems is done one system at time. In this phase the
synchronization of the entire system is built up step by step at the end of which the
whole of CMS will be ready to sample the LHC beams when they turn on. The overall
synchronization scheme will be described in detail for the individual systems and
then the bringing together of all the systems to form a coherent whole. Once CMS is
in this state a global phase shift of the master clock will be all that will be
required to bring the detector sampling into phase with the actual particle crossings
in the LHC.
