LHC Machine Protection B Dehning CERN ABBI 05

LHC Machine Protection B. Dehning CERN AB/BI 05. 2008 BIW 2008, B. Dehning 1

Contend n Damage, Quench, Risk n Protection Strategy n Collimators n Design approach n Particularities of Superconducting Magnets n Beam loss measurement System n System settings and database n Survey and tests n Calculation and Simulation of damage risk and false dump 05. 2008 BIW 2008, B. Dehning 2

Material Damage Experiment at the SPS V. Kain et al. 6 cm 25 cm n Proton beam, 450 Ge. V, Cu, Fe sandwich target n beam size σx/y = 1. 1 mm/0. 6 mm 2 1012 no damage n 8 1012 damage n 05. 2008 BIW 2008, B. Dehning Safe at 0. 6 % of full LHC intensity 3

Density Change in Target after Impact of 100 Bunches Cross sectional view copper solid state beam impact 2 dimensional hydrodynamic computer code, N. A. Tahir et al. Reduction of density by a factor 10 05. 2008 BIW 2008, B. Dehning 4

Magnet Quenches cross sectional view of coil beam D. Bocian et al. Energy deposition by the beam Temperature difference Tquench – T (steady case) Location of QUENCH Beams 05. 2008 BIW 2008, B. Dehning 5

Beam Loss Durations and Protection Systems LOSS DURATION Ultra-fast loss PROTECTION SYSTEM Passive Components 4 turns (356 s) Fast losses 10 ms Intermediate losses 100 s + BLM (damage and quench prevention) + Quench Protection System, QPS (damage protection only) Slow losses Steady state losses + Cryogenic System Since not active protection possible for ultra-fast losses => passive system ICFA HB 2006, Tsukuba, Japan Eva Barbara Holzer June 1, 2006 6

Collimators and Absorbers Experiment Octant 7 Protection devices Primary collimator Secondary collimators TCT Tertiary collimators Absorbers Tertiary halo Beam Secondary halo + hadronic showers § Distribution of collimators and absorbers along the ring to protect equipment against ultra-fast and up to steady state losses 05. 2008 hadronic showers Beam loss rate a. u. Primary halo particle Triplet magnets BIW 2008, B. Dehning 7

Stored Beam Energies LHC will be exceptional => High RISK 05. 2008 BIW 2008, B. Dehning 8

Safety System Design Approach Risk Safety Protection Availability Methods: Damage Failsafe (system integrity) Redundancy Survey Functional Check Quench (operational Efficiency) Scaling: frequency of events x consequence Mean time between failures 1 10 -8 to SIL ALARP 1 10 -7 1/h 05. 2008 Stop of next injection Extraction of beam Reduction of operational efficiency Systems: Design issues: Beam loss Monitors Reliable components Quench protection system Redundancy, voting Interlock system ¨ Dump system Monitoring of drifts BIW 2008, B. Dehning 9

Failure Rate and Checks Systems parallel + survey + functional check: 1. in case of system failure dump beam (failsafe) 2. verification of functionality: simulate measurement and comparison with expected result => as good as new 05. 2008 BIW 2008, B. Dehning 10

The Active Protection System HERA SOURCES of Tevatron, LHC Dump system Dump requests 1. User/operator 2. PC failures 3. Magnet failures 4. Collimators failures Interlock system 05. 2008 beam losses BIW 2008, B. Dehning 5. RF failures 6. Obstacles 7. Vacuum 8. … 11

LHC Bending Magnet Quench Levels (values proportional) DESY 05. 2008 2. 6 – 6. 6 E-03 LHC quench values are lowest BIW 2008, B. Dehning 12

Quench Levels and Energy Dependence n Fast decrease of quench levels between 0. 45 to 2 Te. V n Similar behaviour expected for damage levels 05. 2008 BIW 2008, B. Dehning 13

Beam Loss Measurement System Layouts LHC 05. 2008 BIW 2008, B. Dehning 14

Ionisation Chamber and Secondary Emission Monitor n Stainless steal cylinder n Parallel electrodes distance 0. 5 cm n Diameter 8. 9 cm n Voltage 1. 5 k. V n Low pass filter at the HV input Signal Ratio: IC/SEM = 60000 SEM: IC: 05. 2008 n Al electrodes n Ti electrodes n Length 60 cm n Components UHV compatible n Ion collection time 85 us n Steel vacuum fired n N 2 gas filling at 1. 1 bar n n Sensitive volume 1. 5 l BIW 2008, B. Dehning Detector contains 170 cm 2 of NEG St 707 to keep the vacuum < 10 -4 mbar during 20 years 15

Gain Variation of SPS Chambers n Test with Cs 137 n n 30 years of operation Measurements done with installed electronic Relative accuracy n n Total received dose: ring 0. 1 to 1 k. Gy/year extr 0. 1 to 10 MGy/year 05. 2008 Ds/s < 0. 01 (for ring BLMs) Ds/s < 0. 05 (for Extr. , inj. BLMs) Gain variation only observed in high radiation areas Consequences for LHC: n No gain variation expected in the straight section and ARC of LHC n Variation of gain in collimation possible for ionisation chambers Reliable component BIW 2008, B. Dehning 16

Ionisation Chamber Simulation and Measurements Ionisation chamber response function Ionisation chamber top view M. Stockner, Ph. D thesis Beam scanned n Good knowledge of behaviour => Reliable component 05. 2008 Comparison simulation measurements Rel. diff % Error % Proton 13. 1 11. 4 Gamma 14. 3 12. 1 neutron 37. 4 13. 9 Mixed field 20. 5 11. 4 BIW 2008, B. Dehning 17

Ionisation chamber currents (1 litre, LHC) 05. 2008 Quench level ranges (min. ) 450 Ge. V 100 s 12. 5 n. A 7 Te. V 100 s 2 n. A Dynamic range min. , used for tuning 450 Ge. V 100 s 2. 5 p. A 7 Te. V 100 s 80 p. A BIW 2008, B. Dehning 18

The BLM Acquisition System Analog front-end FEE § Current to Frequency Converters (CFCs) § Analogue to Digital Converters (ADCs) § Tunnel FPGAs: Actel’s 54 SX/A radiation tolerant. § Communication links: Gigabit Optical Links. 05. 2008 Real-Time Processing BEE n FPGA Altera’s Stratix EP 1 S 40 (medium size, SRAM based) n Mezzanine card for the optical links n 3 x 2 MB SRAMs for temporary data storage n NV-RAM for system settings and threshold table storage BIW 2008, B. Dehning 19

Test Procedure of Analog Signal Chain Modulation Example HV ripple (pp 10 v) Basic concept: Automatic test measurements in between of two fills n HV supply current Modulation of high voltage supply of chambers HV induced signal n Check of cabling n Check of components, R- C filter n Check of chamber capacity n n n Check of stability of signal, p. A to n. A (quench level region) Measurement of dark current Not checked: gas gain of chamber (only once a year with source) Functional checks – Monitoring of drifts 05. 2008 BIW 2008, B. Dehning 20

Digital Transmission Line Check Secondary B Signal (256 bits) Primary A Signal (256 bits) Reception _______ _ _ Check CRC validity Tx Check & Signal Choice Only CRC (4 byte) Signal Select (A or B) Error Status 10 -bits Error _______ _ _ A B Error Format Data Error CRCs Error OK _______ _ _Error Tunnel Status Check CRC 32 check Comparison _______ _ of_ 4 Byte Only CRC Check CRC validity (4 byte) Compare CRCs S/W & TTL output Signal Select Truncate. Table extra/redundant bits (leave 160 bits) Output Dump. De. Mux 1 Dump 2 3 … Remarks Both signals have error S/W 8 trigger (CRCgenerate or check wrong) Signal B S/W trigger (error at CRC detected) Error OK OK Signal B S/W trigger (error at data part) OK Error Signal A S/W trigger (error at CRC detected) OK Error OK Signal A S/W trigger (error at data part) OK OK Error Dump S/W trigger (one of the counters has error) OK OK OK Signal A By default (both signals are correct) 05. 2008 BIW 2008, B. Dehning At the Surface FPGA: § Signal CRC-32 § Error check / detection algorithm for each of the signals received. § Comparison of the pair of signals. § Select block § Logic that chooses signal to be used § Identifies problematic areas. § Tunnel’s Status Check block § HT, Power supplies § FPGA errors § Temperature 21

Functional Tests Overview Detector Ph. D thesis G. Guaglio Tunnel electronics Surface electronics Combiner Functional tests before installation Barcode check Current source test Radioactive source test HV modulation test Beam inhibit lines tests Threshold table data base comparison 10 p. A test Double optical line comparison System component identity check Inspection frequency: Reception Installation and yearly maintenance Before (each) fill Parallel with beam Functional checks – Monitoring of drifts ICFA HB 2006, Tsukuba, Japan Eva Barbara Holzer June 1, 2006 22

Software Overview, Management of Settings Safety given by: n Comparison of settings at DB and front-end n Safe transmission of settings Surveyed 05. 2008 BIW 2008, B. Dehning 23

Data Base Structure n Two layers n entry layer (stage tables) n validated layer (final tables) n Concept of Master and Applied table – Comparison of Threshold values (Applied < Master) n Master: less frequent changes n thresholds possible with Failsafe 05. 2008 Applied: change of user interface BIW 2008, B. Dehning 24

Reliability Study – Fault-Tree Approach Relative probability of a system component being responsible for a damage to an LHC magnet in the case of a loss. Highest damage probability given by the Ionisation chamber (80%) because: 1. Reduced checks 2. Harsh environment 05. 2008 Relative probability of a BLM component generating a false dump. by G. Guaglio Most false dumps initiated by analog front end (98%) because: 1. Reduced check 2. Quantity 3. Harsh environment BIW 2008, B. Dehning 25

Modelling of Machine Protection System BEAM DUMP BLM S. Wagner et al. Laboratory for Safety Analysis, ETH Zurich 470 BLM connected to each BIC (pair) Combined model: Fault Tree & Monte Carlo BLM Beam interlock 05. 2008 BIW 2008, B. Dehning 26

First Modelling Results Mission Distribution n n contribution of the components to false dumps by triggering false dump requests. n n n fraction of early ended missions triggered by beam loss event 11. 3% false dump due to a false dump request by a component failure 1. 7% Front electronics and BIC contribute with 40 % BLM system analysis reveals ARC power supply contribute most to FEE failure VME crate failure contribute significantly Comparison between simulation and installed system => survey 05. 2008 BIW 2008, B. Dehning 27

Safety System Design Approach Risk Safety Protection Availability Methods: Damage Failsafe (system integrity) Redundancy Survey Functional Check Quench (operational Efficiency) Scaling: frequency of events x consequence Mean time between failures 1 10 -8 to SIL ALARP 1 10 -7 1/h 05. 2008 Stop of next injection Extraction of beam Reduction of operational efficiency Systems: Design issues: Beam loss Monitors Reliable components Quench protection system Redundancy, voting Interlock system ¨ Dump system Monitoring of drifts BIW 2008, B. Dehning 28

Literature n n http: //cern. ch/blm LHC n n n Reliability issues, thesis, G. Guaglio Reliability issues, R. Filippini et al. , PAC 05 Front end electronics, analog, thesis, W. Friesenbichler Front end electronics, analog-digital, E. Effinger et al. Digital signal treatment, thesis, C. Zamantzas Balancing Safety and Availability for an Electronic Protection System, S. Wagner et al. , to be published, ESREL 2008 05. 2008 BIW 2008, B. Dehning 29

Reserve slides 05. 2008 BIW 2008, B. Dehning 30

Beam Loss Display 05. 2008 BIW 2008, B. Dehning 31

Intensities n n n n Intensity one “pilot” bunch 5⋅109 Nominal bunch intensity 1. 1⋅1011 Batch from SPS (216/288 bunches at 450 Ge. V) 3⋅1013 Nominal beam intensity with 2808 bunches 3⋅1014 Damage level for fast losses at 450 Ge. V 1 -2⋅1012 Damage level for fast losses at 7 Te. V 1 -2⋅1010 Quench level for fast losses at 450 Ge. V 2 -3⋅109 Quench level for fast losses at 7 Te. V 1 -2⋅106 05. 2008 BIW 2008, B. Dehning 32

Strategy for machine protection § Definition of aperture by collimators. Beam Cleaning System § Early detection of failures for equipment acting on beams generates dump request, possibly before the beam is affected. Powering Interlocks Fast Magnet Current change Monitor § Active monitoring of the beams detects abnormal beam conditions and generates beam dump requests down to a single machine turn. § Reliable transmission of beam dump requests to beam dumping system. Active signal required for operation, absence of signal is considered as beam dump request and injection inhibit. Beam Loss Monitors Other Beam Monitors § Reliable operation of beam dumping system for dump requests or internal faults, safely extract the beams onto the external dump blocks. § Passive protection by beam absorbers and collimators for specific ICFA HB 2006, Tsukuba, Japan failure cases. Eva Barbara Holzer Beam Interlock System Beam Dumping System Beam Absorbers June 1, 2006 33

Ionisation chamber SNS n n n 05. 2008 Stainless steal Coaxial design, 3 cylinder (outside for shielding) Low pass filter at the HV input Ar, N 2 gas filling at 100 mbar over pressure Outer inner electrode diameter 1. 9 / 1. 3 cm n Length 40 cm n Sensitive volume 0. 1 l n Voltage 2 k V n Ion collection time 72 us BIW 2008, B. Dehning 34

Approximation of Quench Levels (LHC) n n Dump level tables are loaded in a non volatile RAM Any curve approximation possible n n Loss durations Energy dependence 05. 2008 BIW 2008, B. Dehning Relative error kept < 20 % 35

Drift times of electrons and ions (II) 05. 2008 BIW 2008, B. Dehning 36

Drift times of electrons and ions (I) 05. 2008 BIW 2008, B. Dehning 37

neutron proton pi+ pi- u+ u- Gamma e+ e- Response of ion chambers for different particle species Due to attenuation of shower => increase of non linearity of chamber response 05. 2008 BIW 2008, B. Dehning 38

Quench and Damage Levels n Detection of shower particles outside the cryostat or near the collimators to determine the coil temperature increase due to particle losses Quench level and observation range BLMS* & BLMC 450 Ge. V 7 Te. V Damage levels Special & Collimator 1 turn 05. 2008 Dynamic Arc: 108 Collimator: 1013 Arc 2. 5 ms BIW 2008, B. Dehning 39

Energy spectrum of shower particles outside of cryostat • Number of charged particles and energy deposition simulated: • Energy spectrum: 1 bin = 5 Me. V Energy [Ge. V] 05. 2008 BIW 2008, B. Dehning 40

Ionisation Chamber Time Response Measurements (BOOSTER) Chamber beam response Chamber current vs beam current slength proton= 50 ns FWHMe-= 150 ns 80 % of signal in one turn Intensity discrepancy by a factor two Intensity density: - Booster 6 109 prot. /cm 2, two orders larger as in LHC 05. 2008 BIW 2008, B. Dehning 41

Current to Frequency Converter and Radiation Quench 7 Te. V n n Variation at the very low end of the dynamic range Insignificant variations at quench levels 05. 2008 BIW 2008, B. Dehning 42

LHC cycle and stored beam energy 7000 Energy [Ge. V/c 6000 energy ramp 25 5000 MJ 360 MJ circulating beam 2808 bunches 4000 coast 360 MJ circulating beam 3000 beam dump injection phase 3 MJ 25 MJ beam transfer circulating beam 2000 1000 360 MJ via transfer line 0 -4000 12 batches from the SPS (every-2000 20 sec)0 2000 one batch 216 / 288 bunches 3 MJ per batch 05. 2008 BIW 2008, B. Dehning 4000 time from start of injection (s) 43

FNAL beam loss integrator and digitizer VME Control bus n n 05. 2008 FNAL LHC channels 4 16 Time resolution 21 s 40 s # of running sums 3 11 windows 21 s to 1. 4 s 80 s to 84 s thresholds 4 12 Synchronized to yes machine timing no post mortem buffer 1 k values 4 k values Independent operation form crate CPU (FNAL, LHC) Thresholds managed by control card over control bus (LHC combined) BIW 2008, B. Dehning 44

LHC tunnel card n Not very complicated design “simple” n Large Dynamic Range (8 orders) n V out n Current-to-Frequency Converter (CFC) n Analogue-to-Digital Converter Radiation tolerant (500 Gy, 5 108 p/s/cm 2) n Bipolar n Customs ASICs n Triple module redundancy 05. 2008 Threshold Comparator I- I+ Reset time Integration time 100 ns to 100 s BIW 2008, B. Dehning 45

FNAL abort concentrator n n 05. 2008 BIW 2008, B. Dehning Measurements and threshold are compared every 21 s (fastest) (LHC 80 s) Channels can be masked (LHC yes) Aborts of particular type are counted and compared to the required multiplicity value for this type (LHC: single channel will trigger abort, channel can be masked depending on beam condition) Ring wide concentration possible (LHC no) 46