CEPC SRF System Jiyuan Zhai On behalf of
CEPC SRF System Jiyuan Zhai On behalf of the IHEP CEPC SRF design team 55 th ICFA Advanced Beam Dynamics Workshop on High Luminosity Circular e+e- Colliders – Higgs Factory (HF 2014) October 9 -12, 2014, Beijing
Outline • CEPC SRF layout • Cavity design • HOM damping • Power coupler & tuner • Cost and efficiency • Summary 2
CEPC SRF Layout 3
CEPC Top-level Parameters related to SRF System Unit Main Ring (14 -10 -08) Booster Injection (14 -10 -10) Booster Extraction (14 -10 -10) Beam energy Ge. V 120 6 120 Circumference km 54. 752 Luminosity / IP cm-2 s-1 2. 04 E+34 - - Energy loss / turn Ge. V 3. 11 1. 62 E-5 2. 814 Synchrotron radiation power MW 103. 42 - 2. 4584 s 1. 83 E-04 Revolution frequency k. Hz 5. 4755 Bunch charge n. C 60. 56 3. 35 3. 2 Bunch length mm 2. 65 0. 13 2. 66 - 100 50 50 Beam current (total) m. A 33. 2 0. 9197 0. 8737 Bunch spacing μs 1. 825 3. 65 MHz 0. 55 0. 274 RF voltage GV 6. 87 0. 213867 5. 12 RF frequency GHz 0. 65 1. 3 Harmonic number - 118800 237423 Synchrotron oscillation tune - 0. 180 0. 32076 Energy acceptance (RF) % 5. 99 17. 307 2. 091 Transverse damping time turns 77. 05 682122 (124577. 9 ms) 85. 2 (15. 56 ms) Longitudinal damping time turns 38. 52 341219 (62317. 8 ms) 42. 7 (7. 789 ms) Lifetime Bhabha min 51. 77 - - Lifetime Beamstrahlung (sim. ) min 47 - - Parameter Revolution period Bunch number Beam spectral line spacing
CEPC SRF System Layout Booster magnet field cycle 120 Ge. V extraction to main ring collider and booster in one tunnel one vacuum chamber for collider 6 Ge. V 4 IPs, ~ 1 km each 4 straights, ~ 850 m each RF section length ~ 175 m injection from linac 90 s if 30 min beam lifetime e- e+ • • Total SRF module length 1. 4 km, 12 Ge. V RF voltage 104. 5 MW beam power, 2 MW HOM power 124 MW installed RF power, > 200 MW RF AC power 126 k. W (4. 2 K equiv. ) installed cryogenic power (seven of the eight 18 k. W cryoplants, 30 -40 MW AC power, similar to LHC) • Booster RF and cryogenic duty factor 22 % for continuous injection. • Each collider or booster cavity is driven by an individual RF power source.
CEPC SRF System Design Criteria PS Efficiency ¥ ↑ RF AC < 200 MW PSR 100 MW Pcoupler < 300 k. W Impedance T. H. ↑ ↓ ↓ HOM Power Mode Propagate HOM Damping Ferrite Hook WG ρ = 6. 1 km qb = 60. 6 n. C U 0 = 3. 11 Ge. V f 0 = 5. 5 k. Hz I = 33. 2 m. A Nb = 50 Ncavity¥ Ncell Eacc ↓ ↓ Vrf 6. 87 Ge. V ↓ frf¥ C Lc H E p Bp HOM R/Q, k ↑ H ¥ Ncav/module ¥H Nmodule Cavity Shape Low loss Large iris N-dope Bulk Nb ↓ Rs Riris HOM QL Nb 3 Sn/Nb QL H ↓ Nb/Cu Thin Film H Vc H Cavity Material¥ η↑, τBS ↑, σz↓ Q 0 H ↓ T ↓ 2 K 4. 2 K R/Q H G ↑ ↑ ↓ Vc Dynamic Cryo Load¥ Cryo AC < 30 MW H: handling or LLRF control 6
CEPC SRF System Parameters CEPC-Collider CEPC-Booster LEP 2 Cavity Type 650 MHz 5 -cell (Nitrogen-dope) Nb 1. 3 GHz 9 -cell Nitrogen-doped Nb 352 MHz 4 -cell Nb/Cu sputtered Cavity number 384 256 288 Vcav / VRF 18 MV / 6. 87 Ge. V 20 MV / 5. 12 Ge. V 12 MV / 3. 46 Ge. V Eacc (MV/m) 15. 5 19 6 ~ 7. 5 Q 0 2 E 10 @ 2 K 3. 2 E+9 @ 4. 2 K Cryo AC power (MW) 25 2. 5 (22% DF) 6. 1 Cryomodule number 96 (4 cav. / module) 32 (8 cav. / module) 72 (4 cav) RF input power / cav. (k. W) 260 20 125 RF source number 384(300 k. W klystron) 256 (25 k. W SSA) 36 (1. 2 MW/8 cav) RF AC power (MW) 200 1. 4 (22% DF) 85
CEPC Cryomodule Layout Euro-XFEL/ILC/LCLS-II type X Enlarge two- phase pipe and chimney. Omit 5 K shield while keep 5 K intercept for power coupler. As simple as possible. scaled from 1. 3 GHz cryomodule X 8
Cavity Design 9
RF Frequency Choice • • • Energy acceptance, bunch length (~ f -0. 5 ) HOM impedance and power (∥ ~ f 2, ⊥~ f 3) Cavity size and cost (~ f -2 ) Cryogenic heat load (~ f -0. 5 @ low T, ~ f @ high T) Booster current 1/20 of collider (higher f) Use mature tech and synergy with other projects (ILC, XFEL, ADS…) Thus • CEPC collider RF frequency = 650 MHz • CEPC booster RF frequency = 1. 3 GHz But, booster has lower energy and weaker radiation damping => beam instability caused by the 1. 3 GHz cavity (WANG Na’s talk in WG 7) • collider longer beam life time may reduce booster bunch charge (3 to < 1 n. C) 10
#Cavity, #Cell and Gradient Choice • Within the reasonable gradient, main ring cavity number is limited by the input coupler power capacity. • Main ring coupler power 260 k. W, 384 cavities, 18 MV - balance cryomodule capital cost, coupler operational risk, and cavity gradient and impedance. • Matched booster cavity BW 33 Hz, hard for transient LLRF. Over coupled, 20 k. W input power (50 Hz detuning microphonics & LFD). • More cells better for low gradient and high Q 0, but high cavity HOM power and impedance and low Qext of the HOMs. Booster low current and duty cycle. - collider 5 -cell 15. 5 MV/m (if 4 -cell 19. 5 MV/m) - booster 9 -cell 19 MV/m 11
Cavity Material, T & Q 0 Choice • 650 MHz sputtered Nb/Cu and bulk Nb operating at 4. 2 K are ruled out because of too high cryogenic power ( 4~5 x baseline) • 650 MHz cavity baseline: • bulk niobium operating at 2 K with Q 0=2 E 10 at 15. 5 MV/m • Q 0=2 E 10 at 20 MV/m for the acceptance vertical test. • BCS Q 0 limit is 4. 4 E 10 for 10 m. G ambient magnetic field • 1. 3 GHz cavity baseline: • nitrogen-doped bulk niobium operating at 2 K with Q 0=2 E 10 at 19 MV/m • Q 0=2 E 10 at 23. 5 MV/m for acceptance vertical test • New nitrogen-doping and flux expulsion technology for high quality factor SRF cavity will be used to reach these targets. • To avoid field emission, very clean cavity surface processing and string assembly is required. • Electro-polishing is also needed. • Thin film SRF cavity such as Nb 3 Sn (even HTS) with higher gradient and Q 0 operating at higher temperature will be studied as alternative. 12
Cavity Q performance with “ILC standard” surface processing recipe IHEP 1. 3 GHz 9 -cell Cavity test at 2013 theoretical limit 13
Nb N-Doping and Flux Expulsion Two major discoveries at Fermilab • Nitrogen doping to lower the BCS surface resistance − Lower than the previously perceived “theoretical limit” − Example at 16 MV/m: Q > 3 x 1010 by doping vs Q ~ 1. 5 x 1010 with the standard ILC/XFEL cavity processing − N-doping inhibits formation of Nb-H? Bring Nb to ideal behavior (B. P. Xiao) • Effective magnetic flux expulsion by fast/high thermal gradient cool down around Tc, drastically affects the Meissner effect and achieve record low residual resistances − Example: Q = 2. 7 x 1011 (1 n. Ohm Rres) in 27 m. G ambient field (0. 04 n. Ohm/m. G). Only 2 W is needed for 15 MV/m CEPC 5 -cell cavity − Relax very tight magnetic shield for high Q operation, otherwise high Q is not usable in real accelerator; demagnetize the module vacuum vessel − From 40 K, 2~3 K/min cooling down- high spacial thermal gradient within Nb to win the pinning force and sweep out magnetic flux 14
2 x higher Q leads to a factor of 2 decrease in required refrigeration => 1 cryoplant less => 50 -100 M$ in capital costs + 10 s of M$ in operational costs 15
Cavity RF Design Parameters Parameter Unit Main Ring Booster Cavity frequency MHz 650 1300 Number of cells - 5 9 Cavity effective length m 1. 154 1. 038 Cavity iris diameter mm 156 70 Beam tube diameter mm 170 78 Cell-to-cell coupling - 3% 2% R/Q Ω 514 1036 Geometry factor Ω 268 270 Epeak/Eacc - 2. 4 2 Bpeak/Eacc m. T/(MV/m) 4. 23 4. 26 V/p. C 1. 8 3. 34 V/p. C/m 2. 4 35. 3 MV/m 20 23. 5 - 2 E 10 Cavity longitudinal loss factor k∥ HOM Cavity transverse loss factor k⊥ Qualified gradient Qualified Q 0 16
HOM Damping 17
HOM Power • Very small beam spectral line spacing: main ring 0. 55 MHz, booster 0. 275 MHz. • Impossible to detune HOM modes away from beam spectral lines with large HOM frequency scattering from cavity to cavity − by fabrication tolerances and RF tuning of the fundamental mode − DESY TESLA cavity measured spread: TM 011 9 MHz, TE 111 5 MHz, TM 110 1~6 MHz, TE 121 2 MHz • Average power losses calculated as single pass excitation. HOM power damping of 3. 5 k. W for each 650 MHz 5 -cell cavity and 21 k. W for each cryomodule of the CEPC main ring. • Resonant excitation for the low frequency modes below cut-off. For the main ring cavity. TM 011 power at Qext = 1 E 4 is 937 W (310 W of the single pass case). 18
HOM Spectrum 80 % above cut-off frequency , propagate through cavities and finally absorbed by the two HOM absorbers at room temperature outside the cryomodule. Each absorber has to damp about 10 k. W HOM power, can’t be in the cryogenic region. KEKB and Super. KEKB ferrite HOM absorbers 20 -40 k. W. IHEP 4. 5 k. W tested. XFEL/ILC type HOM coupler and absorber can easily handle the booster HOM power. Ceramic HOM absorber at 70 K in the cryomodule beam line. LEP/LHC-type HOM coupler for the k. W level power capacity. BNL also developing k. W class coaxial HOM couplers. 19
LEP 2 HOM coupler
Waveguide Damping R. Rimmer, HOM 10 Waveguide at the cavity beam pipe is also a possible solution for the main ring cavity HOM power extraction, but with large size waveguides, more complicated structure and interfaces of the cryomodule, and large heat load. 21
HOM Heat Load Main Ring Booster HOM power / cavity 3. 5 k. W 5. 3 W HOM power / module 21 k. W 56 W HOM power / HOM coupler 330 W 1 W HOM power / HOM absorber 9. 2 k. W 47 W HOM 2 K heat load / module 26 W 7. 2 W HOM 5 K heat load / module 78 W 3. 2 W HOM 80 K heat load / module 780 W 52. 8 W Percent of total cryogenic load 30 % 15 % Estimated upper limit, design goal, enough margin • Main ring cryomodule: RF shielded bellows (copper plated) and gate valves, flange connection with gap-free gaskets • Booster heat load scaled from ILC TDR. Duty factor of booster HOM power is 60 % for continuous injection (much lowered, e. g. 1 -5%, for longer beam lifetime => long injection cycle and/or low charge). 22
Heat Load Estimation • Assume 10 k. W HOM power propagating through the beam tubes and bellows (thin copper film RRR=30, in abnormal skin effect regime), power dissipation < 2 W/m @ 2 K. − Even RRR=1 for copper plating (in normal skin effect regime), power dissipation < 10 W/m. Note 300 k. W power through input coupler connected to 2 K and 5 K. • HOM power loss in cavity at 2 K less than 0. 3 W (all modes Qext < 106) • Heat load at 5 K and 80 K dominated by HOM cable heating. Assume 0. 1 d. B/m power dissipation of the LEP/LHC-type rigid coaxial line (copper plated stainless steel) and 1 m length, total heat load is 10 W for the resonant excitation of TM 011 mode. • Further study: HOM propagating, heat load calculation, statistical approach to study heat load at 2 K (LCLS-II), trap mode … 23
HOM Q Limit 1. 3 GHz 9 -cell cavity 650 MHz 5 -cell cavity Monopole Mode TM 011 TM 020 Dipole Mode TE 111 TM 110 TE 112 TM 111 f (GHz) R/Q (Ω) Qext limit 1. 173 1. 350 84. 8 5. 5 5. 1 E+5 6. 8 E+6 f (GHz) R/Q (Ω/m) Qext limit 0. 824 0. 930 1. 225 1. 440 832. 2 681. 2 36. 2 101. 5 2. 3 E+4 2. 8 E+4 5. 2 E+5 1. 9 E+5 Monopole Mode TM 011 TM 012 Dipole Mode TE 111 TM 110 TM 111 TE 121 f Qext R/Q (Ω) (GHz) measured 2. 450 156 5. 9 E 4 3. 845 44 2. 4 E 5 f R/Q Qext (GHz) (Ω/m) measured 1. 739 4283 3. 4 E 3 1. 874 2293 5. 0 E 4 2. 577 4336 5. 0 E 4 3. 087 196 4. 4 E 4 • Large HOM frequency spread from cavity to cavity will relax the Qext requirement. However, some modes far above cut-off frequency may become trapped among cavities in the cryomodule due to the large frequency spread • Further work: enlarge the iris diameter to decrease loss factors while keep relatively high R/Q and low surface field of the fundamental mode, identify trapped modes within the cavity and cryomodule, reduce the cavity cell number or design asymmetry end cells to avoid trapped modes (e. g. TE 121) 24
Power Coupler and Tuner 25
Power Coupler Requirement very high power handling capability: CW, 300 k. W; two windows for vacuum safety and cavity clean assembly; very small heat load; simple structure for cost saving; high yield and high reliability. Parameters Main ring Booster 650 MHz 1. 3 GHz CW, 300 k. W Average < 4. 4 k. W (20 k. W peak) 2. 4 E 6 1. 0 E 7 Coupling type Antenna Coupler type Coaxial 2 2 Frequency Maximum power Qext Number of windows Static heat loss spec. 2 K: 0. 06 W; 5 K: 0. 6 W; 80 K: 6 W 2 K: 0. 02 W; 5 K: 0. 2 W; 80 K: 2 W Dynamic heat loss spec. 2 K: 1 W; 5 K: 10 W; 80 K: 50 W 2 K: 0. 1 W; 5 K: 1 W; 80 K: 14 W (22% duty cycle) Window type One waveguide or cylindrical warm window; One coaxial Tristan type warm window Two same size coaxial Tristan type windows: one warm, one cold 26
Power Coupler Type For main ring 650 MHz cavity, Tristan/KEKB/BEPCII type as baseline. KEK c. ERL injector coupler (one window) as reference. reduce the distance between the window and the coupling port for putting the window into the cryostat profile to realize the window and cavity assembled in Class 10 cleanroom; add one waveguide or cylindrical type warm window for vacuum safety; redesign RF structure for better matching at 650 MHz; redesign the mechanical structure for higher power capability and lower heat load. For booster 1. 3 GHz cavity, average power < 4 k. W, KEK STF 1 -type coupler developed by IHEP (test to 800 k. W with DF 0. 75% = 6 k. W) can be used. for higher power, KEK c. ERL main linac coupler as reference (gas colling of inner conductor, variable coupling). H. Sakai, ERL 11 27
Tuner Requirement • Total number of tuner for CEPC is 640. • Highly reliable and maintainable tuners are required. 650 MHz cavity tuner 1. 3 GHz cavity tuner 500 k. Hz 900 k. Hz Coarse tuning resolution 1 Hz Fine tuning range 1 k. Hz 1. 8 k. Hz Fine tuning resolution 0. 1 Hz Motor / tuner 1 1 Piezo / tuner 2 2 Piezo work environment Vacuum, 2 -5 K Motor work environment Vacuum, 2 -5 K 1 atm, 300 K > 2000 kgf > 1000 kgf Parameters Coarse tuning range Max load 28
Tuner Type • Lever end tuner (e. g. Fermilab) for 650 MHz 5 -cell cavity (but may interfere with beam tube HOM coupler and rigid line). Access port on the cryomodule. • KEK slide jack tuner for 1. 3 GHz 9 -cell cavity (based on IHEP experience). Blade tuner and lever tuner also suitable. Access port. • I. Gonin, PAC 13 • • • Helium vessel with larger diameter: tuner in the mid not suitable (Ti bellow, flange, size, force. . . ) Small bellow to minimize df/d. P Two piezo stacks Motor and piezo maintainable 29
IHEP SRF Key Technology Experience 1. 3 GHz 9 -cell cavity vertical test 20 MV/m, Q 0=1. 4 E 10 650 MHz β=0. 82 5 -cell cavity vertical test soon 1. 3 GHz test cryomodule horizontal test soon 500 MHz coupler 420 k. W CW TW HOM absorber ferrite 4. 5 k. W 12 m 1. 3 GHz cryomodule for Euro-XFEL 500 MHz cavity module horizontal tested 30
Cost and Efficiency 31
What is power used for? • Beijing Subway (Metro) − 2014: 500 km, 200 MW, 3 billion passengers − 2020: 1000 km, 400 MW ( = CEPC? 14 BCNY for 10 yr) • High speed train ( Beijing <=> Shanghai ) − 5 MW / train @ 300 km/h (40 trains on-line: 200 MW) • Nuclear plant reactor unit − 1000 MW • Beijing city peak power load (2013) − 20, 000 MW (1 k. W / person) (2% for CEPC, 20 W / Beijinger). − How much power is being wasted on the street, in your office, at home, and in this hotel? (TV: 150 W, PC: 400 W) • LEP & LHC − 120 MW? (CERN 200 MW, 30% of Geneva city) • ILC (250 Ge. V center-of-mass) − TDR number: 122 MW (163 MW for 500 Ge. V). ILC is already a green machine. We knew it 30 years ago. And we need integrated luminosity. • Be Green for each Higgs, not total power. − Remind: our goal (and what the gov. and taxpayers want, we wish, ) is to keep on doing good physics. Try to create demand.
Main Ring Cost and Efficiency CEPC Main Ring SRF System Cost Estimation CCP ~ 1/PFPC(~#CV. & FPC, HOM), Eacc 2, PFPC, 1/Q 0(Eacc) => curve CCM = 400/PFPC BCNY (~/#, ~ Eacc(CV. FPC)? ) CPS = 3. 23 BCNY (~#) (384 300 k. W klystron, LLRF, PD) (if SSA 40 CNY/W, 50% higher cost) PCP = 10. 4*CCP MW = 47. 6*CCP k. W 4. 2 K equiv. (700 W/W @ 2 K? ) PAC-Beam = 100/(ηRF(85%)* ηPS(60%)) = 200 MW (~#? ) (50% => 70% (e. g. CPD klystron)? SSA 47% * 85% = 40% (future Ga. N to 70%? ) Risk (yield, reliability, R&D time) 33
Coupler power capacity impact on main ring cost Coupler Power (k. W) BEPCII level Eacc (MV/m) Cavity # Module Cost (Billion CNY) 156 9. 3 640 2. 6 260 15. 5 386 1. 5 330 19. 7 302 1. 2 4 M CNY / cavity (including coupler, tuner, cryomodule, etc) KEKB 380 k. W, LHC 300 k. W, ~ 10 -20 pieces 34
650 MHz cavity Q 0 impact on main ring cost and risk Q 0 Installed 4 K Cryo installed Cryo-system equiv. (k. W) power (MW) Cost (Billion CNY) 3. 0 E+10 97 21 2. 0 E+10 116 25 2. 5 1. 0 E+10 177 38 3. 8 5. 0 E+09 293 64 6. 4 Eacc = 15. 5 MV/m, 384 cavities, 21 M CNY / k. W 4 K equiv. cryo-system Note: LHC installed cryo AC power 40 MW, ILC 500 Ge. V 45 MW 35
Booster Cost and Efficiency CCP ~ 1/Eacc (~#CV), Eacc 2, 1/Q 0 (Eacc) − For constant Q, nearly linear with Eacc CCM = 14. 6/Eacc BCNY CPS = 13. 6/Eacc BCNY (~Eacc, Qext, df) − smaller BW & df for higher Eacc to keep same cavity power; otherwise CPS = 0. 034*Eacc BCNY − SSA cost 80 CNY / W (decreasing). Baseline 256 25 k. W, Qext 1 E 7, df 50 Hz, PRF 20 k. W (similar for low charge), SSA cost / cavity: 2 M + LLRF 0. 4 M + distribution 0. 4 M = 2. 8 M; − If Qext 4 E 7(BW 33 Hz, LCLS-II), df 10 Hz, PRF 10 k. W, 12. 5 k. W SSA, 1 M + LLRF 0. 4 M + distribution 0. 4 M = 1. 8 M, save 256 MCNY and power 36
Booster Cost and Efficiency • If klystron for many cavities (220 k. W for 32 cav. ), half cost, no feedback, open loop, piezo feedback? • LLRF control difference between low current SRF booster ring and SRF linac and ERL? • If fixed cavity power, optimized gradient 19 -25 MV/m, even 30, depend on Q 0 (Eacc) − very high Q 0 at very high gradient with narrow BW and microphonics => tech. frontier • If fixed Qext 1 E 7, df 50 Hz, 19 MV/m is nearly the optimized point • Efficiency (SSA total 35 %~ 40 %) 37
Operational Cost Saving with Ga. N Transistors Nobel Prize 2014 Si LDMOS Ga. N HEMT Power per Transistor Pair 160 W 400 W Transistor efficiency 43 % 60 % Combination efficiency 86 % 90 % AC-RF efficiency 35 % 51 % Annual power cost (280 units at 3. 8 k. W) 910 k$ 620 k$ There should also be a ~ 30% cost/unit savings given less modules are needed Chris Adolphsen, LCWS 2014, Belgrade 38
Summary and Outlook • CEPC SRF system: unique challenge and opportunity for China & world SRF accelerator community. • Three main challenges (baseline feasible, require more solid designs, evolve accordingly; in-house demonstration; push technology frontier) − Huge HOM power extraction with low heat load (efficient with low cost) − Very high power CW coupler (robust, clean assembly and low heat load) − Cavity with very high Q 0 at 15 -20 MV/m (use state-of-the-art technology) • Cost & power consumption model for optimization. Make the machine as cheap and green as possible, but not cheaper and greener. • Cooperation or synergy with other projects (R&D and industrialization) − CADS & PIP-II (650 MHz, RF source and coupler, high Q cavity), Euro-XFEL, LCLS-II, ERLs & ILC (1. 3 GHz, high Q, HOM), FCC-ee (HOM). − SRF R&D and pre-production planned for extensive development of key technology, personnel, infrastructure and industrialization. 39
Special thanks to Carlo Pagani, Sergey Belomestnykh and Eiji Kako for their valuable suggestions and help. Thank you for your attention. 40
Back up 41
CEPC SRF Injector Linac Parameters Unit Value R/Q Ω 1036 1. 3 Geometry factor Ω 270 n. C 3. 2 Iris diameter mm 70 Bunch repetition rate for CEPC Hz 100 Operating temperature K 2 Bunch charge for XFEL p. C 100 Cavity quality factor Bunch repetition rate for XFEL MHz 1 Aver. beam current for XFEL m. A Parameter Unit Value RF voltage GV 6 RF frequency GHz Bunch charge for CEPC Cavity CW gradient Cavity effective length (9 -cell) k. W 40 0. 1 Total cryogenic AC power MW 8. 5 MV/m 20 External Q of input coupler m 1. 038 289 Cavities in one cryomodule 8 m Number of cryomodules SRF Linac length 2× 1010 4 K equiv cryogenic installed Number of cavities Cryomodule length Parameter m 2 E+07 RF power per cavity k. W 6. 4 Solid State Amplifier power k. W 8 Number of SSA 289 Total RF source power MW 2. 3 36 Total RF AC power MW 4. 6 500 RF&Cryo AC power MW 13. 1 12. 2 Alternative to normal conducting linac. Simultaneous XFEL user facility? 42
650 MHz 5 -cell Cavity Need further shape optimization, especially for the HOM properties internal Riris(mm) Alpha(deg) A(mm) B(mm) a(mm) b(mm) L(mm) D(mm) flatness coupling Ep/Eacc Hp/Eacc [m. T/(MV/m)] 77. 96 2. 24 94. 4 20. 03 22. 09 115 206. 6 Left end half cell 84. 46 16. 727 92. 1 13. 76 21. 14 114 206. 6 2. 17% 3. 04% 2. 43 4. 23 Right end half cell 84. 46 16. 39 91 91 13. 71 20. 29 113 206. 6 43
1. 3 GHz Cavity Impedance Budget Longitudinal Impedance Threshold 900 800 Zl(MΩ) 700 600 500 400 300 200 100 0 0 R per cavity is 267 Ω Transverse impedance threshold 4 8 12 16 20 Frequency (GHz) Cavity Impedances close to the budget upper limit, more serious at low energy. But HOM frequency spread will relax it. 44
HOM Damping Scheme CESR KEKB Ferrite Outside LEP 2 & SOLEIL, BNL Hook + Ferrite Another possible solution is hybrid module with water cooled HOM absorber inside.
Parameter LEP 2 F. Zimmermann, HF 2014 FCC-ee Z Z (c. w. ) W H t 80 120 175 100 10% HOM 1431 power 152 100 of CEPC 100 100 30 6. 6 100 29791 4490 1360 98 4 MHz 1. 0 0. 7 0. 46 1. 4 0. 14 3. 3 0. 94 7 n. C 1 1 2 2 0. 5 1. 0 Ebeam [Ge. V] 104 45 45 circumference [km] 26. 7 100 current [m. A] 3. 0 1450 PSR, tot [MW] 22 100 no. bunches Nb [1011] ex [nm] ey [pm] b*x [m] 50 MHz 20 times 4 16700 HOM power 4. 2 1. 8 of CEPC. 28. 8 n. C 29 ~100 k. W per 22 cavity if 800 250 60 MHz cavity! 1. 2 0. 5 2 b*y [mm] 50 1 1 1 s*y [nm] 3500 250 32 84 44 45 sz, SR [mm] 11. 5 1. 64 2. 7 1. 01 0. 81 1. 16 sz, tot [mm] (w beamstr. ) 11. 5 2. 56 5. 9 1. 49 1. 17 1. 49 hourglass factor Fhg 0. 99 0. 64 0. 94 0. 79 0. 80 0. 73 L/IP [1034 cm-2 s-1] 0. 01 28 212 12 6 1. 7 tbeam [min] 434 298 39 73 29 21
Open Loop Cavity Stability Range Klystron approach costs half as much per cavity than using SSAs. However open loop (no FB) operation can be unstable due to Lorentz force distortion of the Lorentzian cavity frequency response. Realistic operating point 10 to 20 Hz higher than peak Will test whether piezo-actuator feedback eliminates instabilities as rf feedback does. Input Predictions Gradient Squared vs Detuning Chris Adolphsen, LCWS 2014, Belgrade 47
SSA Si Transistor Trends No Scale Chris Adolphsen, LCWS 2014, Belgrade Scott Blum, NXP, CWRF 2012 48
Example Operating Curves: NAUTEL 3 k. W, 650 MHz SSA for PX AC-RF efficiency = 54% Adjust drain voltage depending on operating power range to maintain high efficiency Chris Adolphsen, LCWS 2014, Belgrade 49
S. Posen, LINAC 14 50
- Slides: 50