Safe Injection into the LHC V.Kain AB/CO Contents and scope The need for machine protection at injection Interlocking – hardware and software Passive protection systems – collimators Protection level simulation results Commissioning aspects Conclusions Input from B. Goddard, M. Lamont, J. Wenninger, H. Burkhardt 1/19/2005 Verena Kain, AB-CO

Machine Protection at Injection 2.4MJ in 450 GeV injected beam Damage limit around 100 J/cc ~2x1012p+, ~5 % of injected batch 7.5s LHC aperture at 450GeV Tight aperture in transfer line (MSI ~7s) Holes in Cu : 450 GeV LHC p+ beam (from 2004 TT40 materials test) 8x1012p+ = ¼ of full batch 5.3x1012p+ = 1/6 of full batch Inside, damage visible over ~1m (melted steel) 1/19/2005 Verena Kain, AB-CO

What can go wrong: magnet trips, kicker flashovers, wrong settings… Magnet trips can move trajectory by many s in short time (MSE: 40s in 1ms) Kicker erratics, missings, timing etc. Operator error Corrupted settings Slow failures: 10s in > 2-3ms: relying on interlocking Fast failures: 10s in < 2-3ms: interlocking + collimators Ultra-fast failures: 10s in few ms: collimators 1/19/2005 Verena Kain, AB-CO

Machine Protection Strategy Avoidance: Procedures to avoid dangerous situations e.g. never inject intensity above damage level into empty LHC (beam presence condition) Protection systems: Active: surveillance + interlocks to inhibit injection Software – typical reaction time: ~seconds Hardwired – typical reaction time: normal power converter surveillance (PCS): ≥2-3ms (possible) fast current decay monitors (FCDM): ~ 1ms Passive: collimators for fast and ultra-fast faults Collimators must be robust enough to withstand full injected batch → low Z material (C, hBN). 1/19/2005 Verena Kain, AB-CO

Beam Presence condition Injection of beam above damage level only possible if beam already circulating in LHC (beam presence) Hardwired interlock ! SPS and LHC intensity must be monitored reliably Reliability and availability – 2 independent BCTs per machine/ring? LHC beam permit SPS beam “safe” LHC beam presence Injection permit NO X YES YES (low Intensity) YES (high Intensity) Protects against many of the failures in the LHC (trips, settings, …) 1/19/2005 Verena Kain, AB-CO

TDI injection stopper + TCDD Hardwired Interlocks LHC energy SPS intensity LHC intensity SPS beam SPS HW Extraction HW bumped BP MSE current, girder MKE bumper currents Settings Transfer line HW Fast beam losses LHC beam permit LHC Settings LHC HW Injection HW BTV screens Transfer line collimators SPS Beam Interlock Controller LHC Beam Interlock Controller + SLP Extraction permit TED position Injection permit TDI position Movable TED beam dumps (stoppers) +TBSE beam SPS extraction kicker LHC injection kicker Timing Timing TDI injection stopper + TCDD SPS Transfer Line LHC Power Convertor Surveillance (PCS) and Fast Current Decay Monitoring (FCDM) are critical parts of the HW interlock system ! 1/19/2005 Verena Kain, AB-CO

Software interlocks Software interlocks Mostly used for less critical parameters and for redundancy: beam quality, trajectory and losses in transfer line, etc. Much slower than hardware interlocks, not failsafe… Local reference values in front-ends for equipment surveillance compared with central reference values Reference values are machine mode dependent 1/19/2005 Verena Kain, AB-CO

Passive Protection for fast and ultra-fast failures Protection of LHC aperture and MSI aperture: TCDI collimators for failures upstream of injection regions “generic” protection system with full phase coverage Dedicated collimators for kicker failures: MKI (LHC) failure: TDI + TCLI are 90° downstream MKE (SPS) failure: TPSG is 90° downstream No dedicated collimators for septum failures Protection from MSE and MSI failures  interlocking 1/19/2005 Verena Kain, AB-CO

TCDI Transfer Line Collimators Close to LHC and MSI (last 300m matching section of TLs) Robust, based on TCS design (1.2m C jaw) FLUKA model of 300m of TI 8  local shield for each TCDI 3 collimators / plane (0-60-120°) Setting: 4.5s, tolerances: ≤1.4 s 2 motors/ jaw (angular control) Protection level 6.9 s (simulated with Monte-Carlo including all imperfections: b beat, mismatch from SPS, tolerances…) 0-60-120 degree collimators x’ 6s x 120o 60o amax 6.9 s LHC aperture to protect at 7.5 s 1/19/2005 Verena Kain, AB-CO

TDI – TCDD - TCLI Protects LHC (especially D1) against MKI failures 90° downstream of the MKI. ~4m long hBN jaws local protection of D1 with mask -> TCDD (1m, Cu) Auxiliary collimators TCLIs to complete system For MKI-TDI phase advance ≠90°, and for flexibility (halo load…) At nx180°±20° from TDI (1.2m long C jaws) 1-2 years after LHC start-up MKI 1/19/2005 Verena Kain, AB-CO

TDI – TCDD - TCLI TCT design, half jaw, low Z two beams in one pipe MKI +90˚ TCDD TCLIB TDI +340˚ TCLIA TDI +200˚ Kicker MKI LEFT OF IP2 RIGHT OF IP2 TCLIM TCT design, half jaw, low Z two beams in one pipe TCS design, beams in separate pipes 1/19/2005 Verena Kain, AB-CO

Protection level simulations Safe LHC injection  losses on aperture below 5% damage limit during injection Extensive tracking simulations to check performance MKI flashover scanned with TDI, TCLI setting Full Monte Carlo of single failures at injection All magnet and kicker families (SPS extraction, Transfer Line, LHC injection) for LSS4 - TI 8 – IR8 Full TL + LHC injection region aperture model All imperfections and errors included 1/19/2005 Verena Kain, AB-CO

MKI flashover simulations TDI, TCLI setting of 6.8s, to guarantee max. 5% above 7.5s Increasing opening increases risk of damage N/N0 of particles with amplitudes >7.5 sy 1/19/2005 Verena Kain, AB-CO

Single failure tracking Monte Carlo results (1000 runs per failure) Family Tolerable error [Dk/k0] required reaction time [ms] Covered by LHC TL MPLH (LSS4 bumper) 0.185 201.0 TCDI PCS MKE (LSS4 kicker) 0.125 - MSE (LSS4 septum) 0.005 0.1 MBHC (TT40 H bend) 5.1 PCS / FCDM MBHA (TT40 H bend) 0.015 39.4 MBI (main TI 8 bends) 0.003 2.7 FCDM MCIBH (start TI 8 H bend) 0.630 389.0 MBIAH (end TI 8 H bend) 13.2 MBIBV (end TI 8 V bend) 39.5 3MCIAV (end TI 8 V bend) 0.225 124.0 MSI (LHC injection septum) 3.0 n/a PCS = standard Power Convertor Surveillance (≥3ms) FCDM = Fast Current Decay Monitor (~1ms), dedicated new system 1/19/2005 Verena Kain, AB-CO

Discussion of results LHC protection looks OK, provided FCDM for MSI Detect and react to 0.3% change in 2.5ms TL needs FCDM for MSE septum (reduce risk window), MBI and MBHC MKE and MSE faults can still cause damage to the TL… possible solutions being studied All other single failures are covered for TL & LHC TCDI system gives full protection from upstream failures Failures at end of line: slow enough for PCS and FCDM More complicated scenarios not yet studied Grouped powering failures: e.g. general power cut Combined failures Fault of machine protection system + other failure e.g. wrong setting of passive absorber + trip of magnet family 1/19/2005 Verena Kain, AB-CO

Commissioning aspects Limited role of protection systems for early stage of LHC commissioning. (SPS extraction, SPS safe beam, TI 8 interlocks and LHC injection BIC needed !) Injection protection must be fully operational for injected intensity above damage level ≈ 15 bunches Many parts can be commissioned without beam: LHC injection BIC + matrices + data exchange; collimator movements + interlocks; FCDM, PCS, logging, software, failsafe behavior, etc. etc. Beam is essential for commissioning of: SPS & LHC safe beam and beam presence (BCTs) TDI/TCLI/TCDI jaw alignment Tests of full interlock system (EMC, …) 1/19/2005 Verena Kain, AB-CO

Commissioning of TCDI system Setting up collimators: 1) finding beam axis 2) align jaws with beam envelope 3) set jaws to required gap Setting up with beam: Transmission Measurement Angular jaw alignment Beam size measurement Use BCTs in SPS and downstream of TCDI (in LHC injection region - not in layout yet) Move TCDI jaws through beam, measuring transmission… Required beam intensity ~1010 depending on BCT resolution Impact on LHC: ”inject – TDI dump” or “inject and dump” mode

TCDI setting up with beam - tests First Results of transmission measurement for TCDI (TCS) setting up. Measured transmission vs. angular misalignment – looks promising Simulated data Methods tested in simulation with realistic parameters and tolerances Attainable accuracy can be predicted – expect able to align to ~100mrad. Simulated transmission vs. angular misalignment: 1/19/2005 Verena Kain, AB-CO

Commissioning of TDI /TCLI Setting up TDI and TCLI with beam Beam axis measurement : circulating pilot Transmission Measurement for jaw alignment similar to TCDI inject-dump required, intensity depending on BCT resolution (max. ~1010) Beam size measurement : circulating pilot (max. ~1010) jaw moved through beam reconstitute transverse beam profile from surviving beam intensity vs. position 1/19/2005 Verena Kain, AB-CO

Conclusions Injection protection has to ensure only “safe” beam is injected into LHC during commissioning is absolutely mandatory if injected intensities exceed 2 1012 p+ LHC protected against many failures by “beam presence” condition Comprehensive tracking simulations extensively used to define protection systems and to determine protection levels LHC fully protected with foreseen active and passive protection Require Fast Current Decay Monitor to measure 0.3% DI in 2.5ms Hardware and software interlocking is being specified → InjWG At present the protection systems cannot fully exclude TL damage Simulations will be extended to grouped and combined failures Commissioning of the injection protection system is being prepared Methods for setting up protection collimators being worked out 1/19/2005 Verena Kain, AB-CO