January 2006Beam dump and injection inhibits / J.Wenninger1 Beam Dump and Injection Inhibits J. Wenninger AB-OP & D. Macina TS-LEA LHC experiments signals to the Beam Interlock System Technical issues Acknowledgments : B. Puccio, R. Schmidt & MPWG

January 2006Beam dump and injection inhibits / J.Wenninger2 Introduction In 2005 interlock channels and signal exchange related to machine operation and protection have been defined and specified. The work was coordinated by the LEADE working group. The specifications have been based on best knowledge about LHC operation to ensure sufficient flexibility for both machine and experiments. Obviously the final answers will be given when the LHC starts up ! Had to make assumptions on LHC machine modes… Assumed the existence of a collimator control system. Tried to guess operational aspects

January 2006Beam dump and injection inhibits / J.Wenninger3 Specification The approval procedure of the specification was launched in Nov./December It was approved with smaller comments by everyone except one party ! Final release expected for february 2006 !

January 2006Beam dump and injection inhibits / J.Wenninger4 Signals

January 2006Beam dump and injection inhibits / J.Wenninger5 Hardware signals Beam dump requests Fast interlock signal based on radiation monitors Concerns all experiments Injection inhibits Slow interlock signal to prevent injection when detectors are powered… Concerns all experiments Interlocks of movable devices Protection against uncontrolled movements, inconsistent positioning… Concerns Roman pots and LHCb VELO  potentially up to 3 signals per experiment ! Interlocks of experimental magnets Trigger a beam dump when solenoids or spectrometer dipoles fail The following interlock categories were identified :

January 2006Beam dump and injection inhibits / J.Wenninger6 LHC Beam Interlock System Up to 20 User Systems per BIC: 6 x Beam-1 8 x Both-Beam 6 x Beam-2 Beam-1 / Beam-2 are Independent! Connected to injection IR2/IR8: -In case of an interlock (=NO beam permit), the beam is dumped & injection is inhibited. - It is not possible to inhibit injection ALONE. - fast reaction time (~  s) - safe (SIL3) - limited no. of inputs - Some inputs maskable for safe beam intensity

January 2006Beam dump and injection inhibits / J.Wenninger7 LHC mode The LHC mode informs the world what operation is TRYING to achieve at a given moment. Modes: ACCESS, FILLING, RAMP, ADJUST, STABLE-BEAMS…. The mode names are not ‘officialized’. The following names were adopted for the interlock specifications: ADJUST is used during  -squeeze and for other perturbing operations. STABLE-BEAMS is the data-taking mode for experiments. Beams should be colliding, collimators in position and backgrounds OK. ‘Light’ tuning of beam parameters will be performed. UNSTABLE-BEAMS is used to inform experiments that the situation degrades and that more drastic tuning may be required (compared to STABLE-BEAMS). Such a mode is entered from the STABLE- BEAMS mode when ‘conditions’ degrade suddenly.

January 2006Beam dump and injection inhibits / J.Wenninger8 Mode Transitions RAMP ADJUST STABLE-BEAMS UNSTABLE-BEAMS BEAM-DUMP RECOVER Normal flow Slow ‘degradation’ Fast ‘degradation’ ‘Degraded’ beam conditions

January 2006Beam dump and injection inhibits / J.Wenninger9 Beam dump request The experiment beam dump request is based on data from various radiation monitors that indicates that there is an immediate danger of damage to the detector. The signal must NOT be used in cases when backgrounds are ‘high’. In such a case the experiment must contact the machine control room to request some ‘actions’. Beam dump will be done with a fail-safe system.

January 2006Beam dump and injection inhibits / J.Wenninger10 Injection inhibit The experiments have asked for the possibility to inhibit injection without dumping the beam: –The injection inhibit is based on the state of the detectors and does not depend on data from radiation monitors. –It indicates that the detectors are not in a safe state to cope with potentially rough conditions that will occur during injection. –Should one of the injection interlock become FALSE during injection, it is not so clear (from the experiment point of view) if beam already circulating should be dumped. –Injection will also be inhibited after a dump has been triggered (by the same experiment) pending assessment of the causes of the dump. But it is NOT possible to inhibit injection without dumping the beam dump with the Beam Interlock System ! A ‘simple solutions’ could be obtained by conditioning an injection inhibit request using the LHC mode (i.e. dump & inhibit only during filling). See later for possible hardware solutions…

January 2006Beam dump and injection inhibits / J.Wenninger11 Movable device interlocks Concerns roman pots (TOTEM & ATLAS) & LHCb VELO since their position (between  ) may directly interfere with beam OP. End-switches define the garage positions (out of beam). The interlock signal is based on the POSITION. It becomes FALSE when the garage position is left unless the machine mode is: –STABLE BEAM (to allow data taking) –UNSTABLE BEAM: on a transition STABLE  UNSTABLE-BEAMS movable device should move back to garage position automatically & asap ! This allows OP crews to intervene on the beam as soon as possible without waiting for VELO and Roman Pots being in their garage position (may take some minutes). OP crews must keep in mind that VELO and Roman pots are not in the safe position until they hit the end-switch. –If the conditions degrade slowly, OP crews should go to ADJUST mode after having informed the experiments. The mode must only change when Roman Pots and VELO have reached their garage position (if not a beam dump will be triggered).

January 2006Beam dump and injection inhibits / J.Wenninger12 Spectrometer dipole interlocks The Magnet Safety System that surveys the magnets of all LHC experiments is under the responsibility of PH/DT1: -The Magnet Safety System will be provide an interlock signal to the BIS. In case of a magnet fault an interlock will be generated a few milliseconds before the power converter is switched off. -The magnet interlock signal is foreseen as a MASKABLE input channel to the BIS. -Conforms to our general interlock philosophy : ‘beam out before equipment off !’ The ALICE and LHCb spectrometer magnets are normal conducting dipoles with circuit time constants of ~ 10 seconds. A powering failure leads to orbit movements of ~ 1  in 750 ms at 7 TeV. The spectrometer dipoles are not the most worrying elements for machine protection, but operation must stop if the magnets fail: the interlock signals will be connected and active for LHC startup.

January 2006Beam dump and injection inhibits / J.Wenninger13 Solenoid interlocks Effects of the solenoids: -They are not expected to provoke major perturbations on the beam at 7 TeV. -At injection the CMS solenoid will induce non-negligible coupling. -The former-L3 / ALICE solenoid used to induce large orbit perturbations (~ 10 mm peak) on the LEP orbit at 22 GeV. At 450 GeV the orbit effect may still be visible at the level of 0.5 mm, i.e. 0.5  – at the limit of being a problem ?! Time constant are long for CMS and ATLAS, but not for ALICE (0.1 s). For the solenoids an input is reserved in the BIS for the interlock signal but the decision if the signal must actually be connected will be taken later (possibly after LHC startup) based on practical experience. Important note : this signal is NOT an interlock on the magnet state, but an interlock on the beam, derived from the magnet state.

January 2006Beam dump and injection inhibits / J.Wenninger14 ‘Software’ signal exchange Clearly in the initial phases of LHC operation communication between experiments and control room will be done mostly by ‘voice contact’. To anticipate the implementation of automatic procedures, the following signals have been defined : -Each experiment will provide a READY-FOR-ADJUST signal to indicate that it is ready for a procedure with increased risk (of background) like squeeze… The signal will be send as a reply to a ADJUST-REQUEST from the machine control room. -Each experiment will provide a READY-FOR-BEAM-DUMP signal to indicate that it is ready for a scheduled beam dump. The signal will be send as a reply to a IMMINENT- BEAM-DUMP signal from the machine control room. -A time-out will be foreseen for both signals.

January 2006Beam dump and injection inhibits / J.Wenninger15 Technical issues

January 2006Beam dump and injection inhibits / J.Wenninger16 BIS connections The experiments presently request NON-MASKABLE connections that must dump BOTH beams for dump requests and movable device interlocks. –The experiments will apply their own masks internally. –The non-maskable interlock signals must be operational and fully reliable already during the LHC machine checkout period. Each BIC module has only 7 NON-MASKABLE inputs – shared between clients that dump a single beams (1 OR 2) or both beams. -There are 2 inputs available per IR for the experiments ! -ATLAS, LHCb & ALICE are OK. -In IR5 we have: -1 channel for CMS dump reques, -2 channels for TOTEM dump request and movable device interlock. -In IR5, or for any new experiment, we must: -share channels, i.e experiments combine their signals, -accept a MASKABLE input, -have a ‘signal concentrator for exp. interlock signals’, for example an independent BIC crate ! Reminder : Inputs are only maskable by the control room when the beam intensity is SAFE !

17 Injection inhibit implementation options -Build an injection interlock loop around the ring. The loop interfaces to the injection elements and to the SPS extractions in IR2 & IR8. ‘high’ price per channel. same HW than BIS. easy to accommodate further ‘clients’. eases maintenance. -Add dedicated point-to-point links between each client experiment and the IR2 and IR8 injection BIC modules (PLC & optical links). potentially lower price per channel. different HW than BIS. -Additional ‘input’: the Beam Dumping System needs a similar link to IR2 & IR8 injection BICs to perform a safe boot- strapping. Decision is pending… Today there are no resources and no budget ! Injection Interlock System X

January 2006Beam dump and injection inhibits / J.Wenninger18 Safe LHC mode The LHC mode plays a critical role for the interlocking of movable devices. It must therefore be send to the experiments with high reliability ! The LHC mode will be part of the Safe LHC Parameters that will be distributed around the machine in a failsafe and highly reliable way: Within the Safe LHC Parameters STABLE-BEAMS will only be sent when the energy is 7 TeV (or whatever the physics energy will be !). The present concept uses the General Machine Timing system for data distribution, with special HW module to enter and extract the data. Details have still to be finalized. In addition the mode will also be distributed over the DIP system for data exchange machine-experiments.

January 2006Beam dump and injection inhibits / J.Wenninger19 Conclusions The experiments will provide up to interlock 3 signals for beam dump request, injection inhibit and for movable device positions. A number of technical issues have to be solved in 2006: Implementation of the injection inhibit. Signal numbers. Cables & rack locations. Distribution and reception of the Safe Parameters (mostly mode) by the experiments.