An ATLAS wide detector control system (DCS) is used to monitor the status of each sub-detector. It uses the supervisory control and data acquisition (SCADA) software SIMATIC WinCC Open Architecture [78] to collect status information from each sub- detector. Furthermore, the DCS interacts with the central LHC control center. The information is presented to the user in a front panel as shown in Figure 4.1 on the following page.
The status can be seen by the color of the sub-detectors. There are more detailed views available for individual sub-detectors. The SCADA tool executes commands for control- ling the detector. Automated actions are defined additionally, which are implemented with a finite state machine (FSM) as explained in section 4.1.2.
4.1.1 LHC operation
During operation, the LHC provides collisions to the experiment as often as possible. Because the beam cannot be provided indefinitely, the machine goes through cycles shown in Figure 4.2. The individual steps of the ATLAS run are described in Table 4.1 and for the LHC in Table 4.2 on page 39.
During stable beams, the experiment is in the physics data taking mode where collisions are happening. This mode should run for as long as possible for a maximum of physics data. During the warm-start at the beginning of the data taking mode, the HV power supplies are ramped up to deplete the sensors (see section 2.1). The actual duration of
Chapter 4 Detector control system Development of a DCS Chip
Figure 4.1: Front panel of the ATLAS DCS [79].
Figure 4.2: Cycle of the LHC and ATLAS runs [80].
a run depends mainly on the LHC operation. Beams are dumped after a certain time when the luminosity is becoming low. However, some beams are lost unexpectedly in case of malfunctions, like a magnet tripping.
The machine and experiments require maintenance to operate reliably. Moreover, upgrades are planned to improve and enhance the detector for better results as described in section 1.2.2. Each winter is the year’s-end-technical-stop for maintenance and small upgrades. Every few years a long shutdown is done to install larger upgrades. Depending on the task at hand, parts of the accelerator or experiments are warmed up. If the
4.1 Control and monitoring of ATLAS
Table 4.1: ATLAS run cycles.
ATLAS status Description
Calibration period No data taking is done and the detector is calibrated or tests are performed.
Standby The detector is safe for beam but not yet taking data. The HV of
the tracker is off. The ATLAS run starts.
Warm start The HV is switched on and data collection starts.
Physics data taking
ATLAS is collecting and storing data for physics.
Warm stop The detector is brought back to standby after a beam dump. The
ATLAS data taking run is finished.
Table 4.2: Operation cycles of the LHC. stable beams is the most important operation mode.
LHC status Description
Setup The machine is prepared for beam injection.
Injection probe
and physics beam A first bunch is injected to configure the beam before the protonbunches are filled.
Ramp The beam energy is increased.
Flat top and
squeeze Max energy is reached and the beam parameters are configured forcollisions.
Adjust The magnets are adjusted to collide the beams
Stable beams The collisions are stable and good for physics data.
Beam dump and
ramp down The beam is dumped and the energy is then ramped down.
operation is done on the vacuum pipe, a bake-out is required afterward. During the bake- out process, the pipe goes through heating cycles to perform outgassing of the vacuum vessel [81]. Without this process, it would not be possible to achieve the required vacuum, which is thinner than in interstellar space [82].
Another important aspect of the operation, are unwanted power cuts which can shut down the system in an uncontrolled way. This happens a few times every year and requires a reboot of the experiments [83]. There are uninterruptible power supplies with back-up batteries for critical systems. However, not everything can be covered by them. They have a limited capacity and can only cover the time it takes to properly switch off the experiment. The detector has to be brought back into normal operation after such a power cut.
To guarantee the safety of operators and the experiment through all these kinds of situations is the task of the DCS.
Chapter 4 Detector control system Development of a DCS Chip
4.1.2 DCS state machine
A finite state machine (FSM) is used to control and operate the detector. It was originally developed for the LHC experiments [84]. At the lowest levels are the devices, which access the hardware e.g. power supplies. The devices are grouped into a control unit (CU), based on the powering groups, like a serial power chain or mechanical structures e.g. a ring in the end-caps. An example tree structure is given in Figure 4.3.
Each control unit has a status (OK, warning, error or fatal), indicating its health. The status is propagated upwards to the root, i.e. the full ATLAS detector. Normally the worst sub-unit state defines the state of the higher level. However, a user can define other propagation rules.
The control unit is further in a state, which defines what operation is ongoing. This can be Shutdown, Standby, Ready, Unknown, Transition or Not Ready. Commands to change a state, e.g. going from standby to ready in case of a warm start, are issued from top to bottom.
Users can look at different control units in the branch. It is also possible to decouple a branch from the tree for debugging or testing purposes.
Automatic operations are defined in the state machine. During operation the state changes depending on what is happening. If a failure occurs, like a module voltage going out of range, the module status changes to Warning or Error. Automated actions are performed to recover the module or error messages are generated for the operators, if human intervention is required.
Figure 4.3: Tree structure for the DCS finite state machine [84].