LEP Status and Performance in 2000

Transcription

1 LEP Status and Performance in 2 R. Assmann, SL/OP for the SL Division Outline: Operational strategy Overview on luminosity and energy performance Energy reach Luminosity performance Other issues Further improvements/options Conclusion RA R. Assmann, LEPC, 2/7/2 1

2 Operational strategy: Traditional: 1) Select a working point for beam energy 2) Optimize luminosity production 3) Collect all required luminosity 4) Select a new beam energy LEP before 2: LEP in 2: Not more than ~3 energies per year Unscheduled change of beam energy discouraged (e.g. not possible for energy to follow available RF voltage) Optimize for ultimate discovery reach - Unconstrained number of beam energies - Simultaneous luminosity production at different beam energies up to limit Change discussed and promoted by P. Janot et al LEP operation and performance in this mode RA R. Assmann, LEPC, 2/7/2 2

3 Understanding the choice of beam energy E: Energy loss U per turn: 4 E U For example: At 14 GeV ~ 3% of beam energy lost per turn Limitation: RF voltage to compensate synchrotron radiation losses Minimal accelerating RF voltage U min required: U min > U RF system with N klystrons (simplified): U RF = N U k Some probability for klystron unavailability (klystron trip rate): Klystron trips occur mainly on statistical basis (LEP every ~ 2 minutes) Finite recovery time of 2-3 minutes Available RF voltage regularly reduced with 1 or 2 klystrons off RA R. Assmann, LEPC, 2/7/2 3

4 Assuming fill at constant energy (traditional strategy): Energy such that U min = (N-2) U k U min = (N-1) U k U min = N U k Fill length set by dump ~ 1.5 h ~ 2 min Fill at highest energy would be short and efficiency would be very low. Fill length ~ 2 min Overhead per fill ~ 69 min Good efficiency requires: Fill length >> Overhead For high energy LEP in 2: Ramp beam energy during physics fill with colliding beams RA R. Assmann, LEPC, 2/7/2 4

5 Typical fill in 2: 22 GeV Injection 12 GeV Set-up, colliding beams, golden orbit, BFS, 12.7 GeV Luminosity production (2 klystron overhead) 13.4 GeV Luminosity production (1 klystron overhead) 14.1 GeV Luminosity production, ended by RF trip Mini-ramps: Used for polarization up to 1994 Revived for high energy Beams ramped in collision with collimators closed Possible due to strong radiation damping RA R. Assmann, LEPC, 2/7/2 5

6 Overview of 2 performance: (14-Jul-2) So far: 558 physics fills Compare: 436 physics fills 1998 (in ~3 months) 653 physics fills Energy: Percentage of fills < 12.5 GeV Start-up Fall-back Mainly: GeV More than 1%: Several energies per physics fill Beam energy [GeV] Physics energy as function of RF voltage. Many different values RA R. Assmann, LEPC, 2/7/2 6

7 Beam energy versus time: Beam energy [GeV] /4 22/4 6/5 2/5 3/6 17/6 1/7 15/7 Energy 3 Energy 2 Energy 1 Date Many physics energies. Usually three energies per fill ( mini-ramp ) RA R. Assmann, LEPC, 2/7/2 7

8 Delivered luminosity versus beam energy: Integrated luminosity [nb -1 ] Overhead (klystrons): Beam energy [GeV] RA R. Assmann, LEPC, 2/7/2 8

9 Luminosity production in 2: Integrated luminosity [pb -1 ] < 12 GeV 8/4 22/4 6/5 2/5 3/6 17/6 1/7 15/7 Date Best slope Raise of beam energy on cost of luminosity production 2 klystrons overhead GeV 1 klystron overhead GeV klystron overhead > 13.6 GeV RA R. Assmann, LEPC, 2/7/2 9

10 Beam current and luminosity per fill: Total initial beam current Produced luminosity per fill Average initial beam current [ma] Luminosity/fill [nb -1 ] Beam energy [GeV] Beam energy [GeV] Higher energies with lower beam currents Higher energies without margin are soon lost with RF trips RA R. Assmann, LEPC, 2/7/2 1

11 Nevertheless, luminosity production in 2 better than in 1998: RA R. Assmann, LEPC, 2/7/2 11

12 Energy increase of LEP from 1999 to 2: LEP 2 preparation: 15 GeV (optics, power supplies, etc checked) Gain from 1999 physics to 2: Improvements: 11 GeV 14.4 GeV GeV 8 additional Cu RF units +.14 GeV Higher RF gradient +.96 GeV Less RF margin GeV Reduced RF frequency +.7 GeV Bending length +.2 GeV Total GeV RF system Operational procedures Reduced luminosity production, potentially higher backgrounds RA R. Assmann, LEPC, 2/7/2 12

13 LEP RF system: Eight additional Cu units installed Clean-up on reliability (tuner power supplies changed) Condition to higher fields (hardware limit w/o beam) Active damping of field oscillations Fast diagnostics of RF trips Automatic adjustment of trippy RF units for mini-ramps Optimization of RF voltage ramp for cryogenics stability RA R. Assmann, LEPC, 2/7/2 13

14 RF voltage (design and actual): 4 35 Available RF voltage O. Brunner Beam energy 125 [GeV] RF voltage [MV] Nominal RF voltage Cryogenics upgrade Jul-95 Feb-96 Aug-96 Mar-97 Sep-97 Apr-98 Nov-98 May-99 Dec-99 Jun- Date Beam energy follows available RF voltage Beam energy Improvements: Install additional RF cavities (8 new CU units in 2) Increase accelerating gradient RA R. Assmann, LEPC, 2/7/2 14

15 Progress with RF conditioning: Condition to higher fields (to hardware limit without beam). Maximum gradients after 2 conditioning (Nb/Cu SC units) Design O. Brunner Average cavity field (MV/m) Average: 7.4 MV/m Unit number RA R. Assmann, LEPC, 2/7/2 15

16 RF stability: 36/8 klystrons (SC/Cu) 288/56 cavities (SC/Cu) 53 kw cooling power (He 4.5K) ~ 1 interlocks RF trips reduce the available RF voltage: Equipment failures (a few % of trips) Running at performance limit (acceptable trip rate) - Mainly field emission (He pressure rise/level) - Arcing in RF distribution system Trip event Voltage reduction Occurancy 1 klystron loss 1 MV ~ 2 min 2 klystrons loss 2 MV ~ 1-2 hours Beam dump (Statistical processes, fast recovery ~ min) 1 MV ~.8 GeV RF voltage Beam energy Energy determined by RF voltage and trip rate RA R. Assmann, LEPC, 2/7/2 16

17 Transient effects on RF voltage: Example: Loss of one half-unit (1MV) at 13 GeV Eff. RF voltage [MV] Vrf eff Vrf 36 MV Effective short-term Vrf following one RF Unit trip Vs. Idc. 1 MV lost with trip Vlq 347 MV Us 329 MV Energy 13 GeV Idc ma Total Beam Current [ma] E. Ciapala Total beam current RF voltage Lost RF voltage ma 35 MV - 1 MV 2 ma 346 MV - 14 MV 4 ma 342 MV - 18 MV 6 ma 333 MV - 27 MV Additional RF voltage reserve for transients required (or lower beam current) RA R. Assmann, LEPC, 2/7/2 17

18 Hardware damage in RF system: 1) Damage in waveguides (Transport of RF accelerating fields from klystrons to cavities) Origin: Damage: Empirical limit for total beam current: ~ 5 ma Beam-induced electro-magnetic fields (HOM) Heating, deformation, holes High energy operation of LEP leaves its marks RA R. Assmann, LEPC, 2/7/2 18

19 2) Corrosion of cables in solid Niobium units Beam induced electro-magnetic fields (HOM) are guided out with cables to avoid excessive heating/damage Solid Niobium RF units: 1) Cable feed-through cooled 2) Condensation of water too much 3) Corrosion 4) Feed-through is destroyed (Hole between insulating vacuum and atmosphere) Fix: Remove cable, plug connector. HOM power stays in 1-3: All solid Niobium 4: Solid Niobium unit 273. Repair: Requires opening cryostat (can be done in situ?) 3) Loss of single cavities 3 cavities lost in 2 RA R. Assmann, LEPC, 2/7/2 19

20 Choice of RF frequency: Damping partition number J x used to reduce horizontal beam size σ x : rms x x x x x x / J D E Increase with beam energy. Good for luminosity and backgrounds in experiments J x controlled with RF frequency f RF. f RF = Hz J x = 1. f RF = 1 Hz J x = 1.55 E max = -.7 GeV Pay with reduction of maximum beam energy. In 2: Keep RF frequency shift small (~ -2 Hz). RA R. Assmann, LEPC, 2/7/2 2

21 Increase average bending radius ρ: (BFS) Energy loss U per turn: 4 E U With larger ρ a higher beam energy E gives the same energy loss. How to increase bending radius? Bending with length L installed for 2π total bending. Add additional bending length L: For LEP: Increase of beam energy to get 2π Less bending in original bends Larger bending radius in original bends Use horizontal correctors and quadrupoles as additional bends Average bending radius increased by.7%.4% of total bending from correctors (2/3) and quadrupoles (1/3) Net gain in energy:.19 GeV RA R. Assmann, LEPC, 2/7/2 21

22 Dipole correctors and quadrupoles as bending magnets : x (mm) Run Run - with BFS J. Wenninger Kick (µrad) 1 5 Monitor No. 1 5 Monitor No Corrector No. Corrector No. RA R. Assmann, LEPC, 2/7/2 22

23 Luminosity performance: Best slope 2 Integrated luminosity [pb -1 ] klystrons overhead 8/4 22/4 < 12 6/5 GeV 2/5 3/6 17/6 1/7 15/7 Date 1 klystron overhead klystron overhead Year Av. rate [pb -1 /day] Raise of beam energy on cost of luminosity production Production rate below 1999 value, but better than 1998 (same period) RA R. Assmann, LEPC, 2/7/2 23

24 Reduced luminosity rate due to trade-off: Luminosity Energy! Factor 4 luminosity ~ 1 GeV increase of beam energy Important trade-offs: Increase J x for small hor. beam size Increase beam current Run with RF voltage reserve Stable energy for tuning, experiments No fills lost with RF trips Decrease J x for highest energy reach Decrease beam current (better RF stability) Run without any reserve in RF voltage Energy follows available RF voltage All fills lost with RF trips Trying to counteract luminosity reduction, but there are limits RA R. Assmann, LEPC, 2/7/2 24

25 Trade-off reflects in key parameters: Average length of physics fills 2: 1.82 h (16-Jun-2) Average coast length [hours] Beam energy [GeV] Overhead per fill (re-cycling, injection, ramping) very important: 1998: 11 min 1999: 93 min 2: 69 min RA R. Assmann, LEPC, 2/7/2 25

26 Optimization of turn-around time: Year Recover [min] Filling [min] Ramp / Squeeze [min] Adjust [min] Total [min] # fills Data: 1/4-16/ Difference Faster degauss, optimize procedure Less current Twice the ramp speed BFS Average turn-around time improved by ~ 24 minutes! Typical 2 turn-around: ~ 45 minutes RA R. Assmann, LEPC, 2/7/2 26

27 We profit from beam behavior at high energy: Strong transverse damping (τ ~ 1/E 3 ) Reminder: Particles perturbed at time t. Consequences for LEP: E.g. orbit oscillation around closed orbit. Oscillation amplitude reduced by e after the damping time τ. Second beam-beam limit (tails, resonances) is overcome Higher beam-beam tune shifts with higher beam-beam limit 1/3 resonance can be jumped Beams can be ramped in collision RA R. Assmann, LEPC, 2/7/2 27

28 Vert. beam-beam parameter: * re me y ib L y 2 e f E i Observed in LEP (1994-2): Energy ξ y (max) Damping [GeV] per IP [turns] σ x σ y from 45.6 GeV to > 98 GeV: Beam-beam limited Beam-beam limit not reached rev x y b 1/ E y Strong damping 3 naively Beam-beam limit pushed upwards Reduced by factor ~ 1.6 (factor ~2 reduction in vertical beam size) RA R. Assmann, LEPC, 2/7/2 28

29 Background in the experiments: RA R. Assmann, LEPC, 2/7/2 29

30 Other issues: Hardware performance - Vacuum system - Magnets - Power supplies - Instrumentation - etc Effects from LHC civil engineering LEP Cryogenics excellent without major worries. RA R. Assmann, LEPC, 2/7/2 3

31 Large radiated power at high energy: NORMALISED MEAN PRESSURE INCREASE (PILOT SECTORS) NORMALISED MEAN PRESSURE INCREASE (ARCS) NORMALISED MEAN PRESSURE INCREASE (INJECTION) NORMALISED RADIATED POWER (E^4) N. Hilleret 1.5 Poly. (NORMALISED MEAN PRESSURE INCREASE (ARCS)) BEAM ENERGY (GeV) Consequences: 1) Higher vacuum pressure (no problem) 2) Possible damage to vacuum system (leaks) RA R. Assmann, LEPC, 2/7/2 31

32 Vacuum leaks and related downtime: N. Hilleret DOWNTIME DUE TO LEAKS (HOURS) DOWNTIME DUE TO VACUUM LEAKS NUMBER OF LEAKS DURING LEP OPERATION NUMBER OF LEAKS Data 2 up to 15/ Year Vacuum system performs very well at highest LEP energies Same true for magnets, power converters, instrumentation, etc RA R. Assmann, LEPC, 2/7/2 32

33 Further improvements/options: RF system Optics RF frequency 2-on-2 bunches - RF voltage at limit of system capability - Slower mini-ramp for better beam stability? - RF stability with lower beam current (2-on-2)? - 18/9 and 132/9 optics? Does not look hopeful. - Run with lower RF frequency (larger beam size)? (lower luminosity, higher backgrounds) - Can be worth for lower beam currents - Better RF stability with lower bunch currents? - RF stability at > 14 GeV looked promising during 4 test fills 8.5 GeV - Higher luminosity production RA R. Assmann, LEPC, 2/7/2 33

34 Summary: LEP runs in Higgs discovery mode: Push beam energy on cost of luminosity Reduce beam current Run with small J x, large σ x Mini-ramp to quantum lifetime limit (zero margin in RF voltage) Loose all fills with RF trips Luminosity production still excellent: (1999 and 2 better than 1998) Highlights: I max 6.25 ma # bunches Beam-beam par..83 per IP Max. luminosity cm -2 s -1 Vert. emittance.1 nm Emittance ratio <.5% Max. beam energy 14.4 GeV Lumin. spread ~1-2 % Turnaround time ~ 45 min Improvements in vertical emittance tuning (dispersion-free steering, luminosity observation, tune working point, turnaround time, ) Higher beam-beam limit with strong damping (infer limit ~ ) Higgs 3σ sensitivity at 112 GeV/c 2. Hope: 114 GeV/c 2. RA R. Assmann, LEPC, 2/7/2 34

35 The end Reserve slides to follow RA R. Assmann, LEPC, 2/7/2 35

36 Outline: Operational strategy Overview on luminosity and energy performance Energy reach Contributions RF system Damping partition number Increase of bending radius Luminosity performance Summary Trade-off luminosity / energy Overhead per physics fill (turn-around) Background (tune jump, RF trips) Other issues Further improvements/options Conclusion Hardware performance LHC civil engineering LEP cryogenics system RA R. Assmann, LEPC, 2/7/2 36

37 Quadrupoles contributing to bending: J. Wenninger X (mm).6 QF QF QD QD QD s (m) RA R. Assmann, LEPC, 2/7/2 37

38 Vertical emittance: rms /2: β * y = 5 cm y C Dy E K Initial tuning of coupling, chromaticity, orbit, dispersion, Vertical orbit to get smallest RMS dispersion Coupling to get smallest global coupling Local dispersion, coupling, β-function at IP x E Peak luminosity Luminosity balance (solenoids) Golden orbit strategy for optimization: Trial and error! Complement with: (Lumi. measurements: MOP6B4) Dispersion-free steering (DFS): 1) Measure orbit and dispersion MOP6B3 2) Calculate correctors to minimize both Note: Global correction generally also improves local dispersion/coupling! RA R. Assmann, LEPC, 2/7/2 38

39 Measured single beam performance of DFS in LEP: ORBIT DISPERSION CORR. KICKS y [mm] D y [cm] θ y [µrad] BPM number BPM number Corrector number DFS: Simultaneously optimize orbit, disp., corr y [mm] D y [cm] θ y [µrad] BPM number BPM number Corrector number RA R. Assmann, LEPC, 2/7/2 39

40 Vertical optimization: Reduction of RMS dispersion (DFS + change of separation optics) εy [nm] (simulated) 98 GeV Vertical emittance [nm] (Data 5-55 µa) E [GeV] Fill number D y (rms) [cm] Reduction of vertical emittance Emittance ratio:.5% RA R. Assmann, LEPC, 2/7/2 4

41 Vertical beam-beam blow-up: Simple model used to fit unperturbed emittance and beam-beam limit: 1 y i A Bi 2 Two fit parameters A and B: * 2 e f x A * x re y B y 2 1 ( i ) b b b y Vertical beam-beam parameter GeV 25 Limited gain 2 Poster TUP6B1. 15 in luminosity 1 ξ y (asymp) =.115 with ε y : 5 ε y (no BB) =.1 nm No BB blow-up Luminosity [1 3 cm -2 s -1 ] Bunch current [µa] No BB limit Vertical emittance [nm] With BB limit RA R. Assmann, LEPC, 2/7/2 41

42 Luminosity decay due to vertical orbit drifts: Lum. [1 3 cm -2 s -1 ] Orbit correction :3 :35 :4 :45 :5 :55 1: 1:5 Time BCT EXPT cm s per minute L.2 nm per minute Measurement illustrates great sensitivity useful for fast online tuning Luminosity stabilized with the vertical orbit feedback ( autopilot ) every 7-8 minutes (3% effect). Both visible from experiments and beam lifetime BCT (faster)! (new operational tool in 1999) RA R. Assmann, LEPC, 2/7/2 42

43 Fast luminosity monitoring from LEP lifetime (BCT): Different regimes: 1) Without collision: Compton scattering on thermal photons, beam-gas scattering. τ = 32 h. 2) In collision: Radiative Bhabha scattering or beam-beam bremsstrahlung. Lifetime [h] Lifetime without collision Putting into collision 18: 19: 2: 21: Time Electrons Positrons Lifetime during collision (increase with current decrease) Observe rate particle loss (BCT) Cross section Calculate luminosity RA R. Assmann, LEPC, 2/7/2 43

44 Reduction in design vertical dispersion: DFS 1998 tests successful. Residual dispersion measured: Single beam: Colliding beams: 1. cm 3.5 cm Difference explained by separation bumps in odd IP s optics: 2.5 cm 1999 modified: 1.6 cm Used for start-up 1999 optimized:.3 cm Tested for 3 physics fills in 7/99 New solutions required change of separator polarities WHY the difference? Trade-off: Small separation bumps Large separation bumps (reduce dispersion from (reduce dispersion from bumps) residual beam-beam kicks) RA R. Assmann, LEPC, 2/7/2 44

45 New working point for horizontal tune: Strategy from 1998: Put Q x as high as possible (~.3) Lower Q y to ~.18 Limits for Q x : Third integer resonance at 1/3 Sensitivity to background storms closer to 1/3 June 1999: Jump the 1/3 resonance with Q x to ~.36 Observation: Higher luminosity No background storms with J x = 1.5 RA R. Assmann, LEPC, 2/7/2 45

46 Details of vacuum leaks: 3 25 N. Hilleret 12 DOWNTIME (HOUR) NUMBER OF LEAKS ION PUMP FEED THROUGH BPM FEEDTHROUGH BELLOWS LEP AL FLANGES Ø 225 CF FLANGES Ø113.5 WELD ON Al SPECIAL CHAMBER RF CAVITY WINDOW RF FIELD PROBE EXPERIMENT VAC. CHAMBER SECTOR VALVE GASKET DIR. COPPLER Be WINDOW QUARTZ WINDOW ST.STEEL TRANSITIONS Al TRANSITIONS (GASKET) PENNING GAUGE COMPONENTS RA R. Assmann, LEPC, 2/7/2 46

47 Nb/Cu SC units - Maximum field after conditioning (2): 12. O. Brunner Unit field (MV) Unit number RA R. Assmann, LEPC, 2/7/2 47

48 Understanding the choice of beam energy: Beam energy E Synchrotron radiation losses U ~ E 4 Minimal accelerating RF voltage U min required with : U min > U RF system with N klystrons (simplified): U RF = N U k Some probability for klystron unavailability (klystron trip rate) Klystron trips occur mainly on statistical basis (LEP every ~ 2 minutes) Finite recovery time of 2-3 minutes Energy such that U min = (N-2) U k U min = (N-1) U k U min = N U k Fill length set by dump ~ 1.5 h ~ 2 min Fills at highest energy would have very low efficiency (69 min overhead) RA R. Assmann, LEPC, 2/7/2 48

49 Horizontal beam size: / J D E rms x x x x x x Compensate increase with energy (smaller luminosity, larger background): 1) High Q x optics with smaller D rms x (D. Brandt et al, PAC99) 2) Smaller β * x (2. m m m) 3) Increase damping partition number J x via RF frequency Automatic control J x = function (U RF ) RF frequency [Hz] GeV :4 1: 1:2 1:4 2: 2:2 2:4 3: Time J x = 1.6 (smaller σ x ) J x = 1.4 Safety setting J x = 1.4 (larger σ x ) RA R. Assmann, LEPC, 2/7/2 49