HOM coupler design for CEPC

Size: px

Start display at page:

Download "HOM coupler design for CEPC"

Erick Day
5 years ago
Views:

1 HOM coupler design for CEPC Hongjuan Zheng, Fanbo Meng Institute of High Energy Physics Beijing, CAS 6th IHEP-KEK SCRF Collaboration Meeting. July 15, 2017, IHEP, Beijing, China.

2 Outline Overview Design goal Damping scheme Thermal analysis Multipacting Tolerances analysis Mechanical structure design Others 2

Overview Design contents mainly includes RF

analysis, heat loss, mechanical designs 0 RF

Design schemes for HOM coupler Multipacting

600 800 1000 1200 1400 1600 1800 Frequency(MHz)

3 Overview Design contents mainly includes RF design, power dissipation, multipacting analysis, heat loss, mechanical designs 0 RF design Double notch/s21(db) Single notch/s21(db) Design schemes for HOM coupler Multipacting analysis S21(dB) Frequency(MHz) Heat loss Mechanical designs Single notch coupler double notch coupler 3

4 Design goal Cut off TE11: 1126MHz Cut off TM01: 1471MHz R/Q (ohm) f (MHz) monopole R/Q(ohm) dipole R/Q(ohm/m) Dangerous monopole: aroud1200mhz Dangerous dipole: 800MHz~900MHz, 1200MHz HOM absorber: operating frequency>1400mhz R/Q (ohm/m) Beam stability: HOM Qe: 10 4 ~10 5 HOM power: If beam spectrum coincide with HOM, the requirement for Qe should less than (For TM011, if resonance happen, Pt~924W (Qe=10 4 )) Maximum power: 1 kw HOM coupler: operating frequency 800~1400MHz

Higher-Order RF power (Watts) Frequency distribution of HOM power 140 120 100 80 60 40 20 HOM power spectrum (H-LP) 425 bunches per beam 0.

75 V/pC Single 2-cell cavity 0 2 3 4 5 6 7 8 9 10111213141516171819202122232425262728293031323334353637383940 RF frequency range (GHz) f

5 Higher-Order RF power (Watts) Frequency distribution of HOM power HOM power spectrum (H-LP) 425 bunches per beam ma av. in each bunch σz=2.9 mm toal HOM power: kw/cavity for f> 1.47 GHz: kw loss factor ~ 0.75 V/pC Single 2-cell cavity RF frequency range (GHz) f (GHz) 1~10 10~20 20~30 30~40 P (W) percentage 72.9% 22.17% 4.45% 0.48% Compared with Q0: Ea=20MV/m Q0=1.2E+10 Pc=33.5W Bandwidth for absorber: 1 GHz~20 GHz

6 Design methods To design the coupler from equivalent circuit concept. Then transform to transmission line models (TLM). To simplify the HOM coupler design by ABCD matrix. RLC TML 3D model design [1] Frank Gerigk, CERN, Studienarbeit. [2] K. Papke, U. Van Rienen, and F. Gerigk, HOM Couplers for CERN SPL Cavities, [3] W. XU. HOM coupler design for High current SRF cavities

7 Matrix methods for coupler design θ Circuit units: 1 0 y y 1 z 1 z z cos jsin c =1 01 jsin cos Combining all components: a1 a2 a3 Note: a is the normalized scattering matrix. Atotal=a1*a2*a3 S 21 20lg 2 Atotal[1,1] Atotal[1,2] Atotal[2,1] Atotal[2,2] Relationship between the normalized scattering matrix and scattering matrix is, Z c2 A11 A12 / Zc2Zc 1 a11 a 12 Zc 1 a21 a 22 Zc 1 A21 Zc2Zc 1 A 22 Z c2

8 Transmission line model for the coupler design Single notch Double notch l1 l2 C2t l3 l1 l2 C2t l3 l4 Ln L1n L2n I0 Cn Z1 M Z2 Z3 Ct Z I0 C1n Z1 M Z2 Z3 C2n Ct Z 8

9 Design approach l1 l2 C2t l3 l4 L1n L2n I0 Z1 M Z2 Z3 Ct Z C1n C2n design approach: equivalent circuit (transmission line model) optimize each part according to S21 curve change to 3D model electromagnetic optimization 9

10 Double notch coupler design motivation 0 Double notch/s21(db) Single notch/s21(db) S21(dB) Hook v.1 Snotch 5.28 MHz Hook v.2 Dnotch 88.5 MHz Frequency(MHz) Single notch design is easier to fabrication but it is more sensitive and difficult to tuning the fundamental mode. Hook type design has a better coupling to dipole modes. Transmission behavior optimized according to the HOM spectrum of the cavity by adjusting all design parameters. 10

11 Double notch design TE11-TEM transmission TE11-TEM transmission TM01-TEM transmission cut off: TE MHz cut off: TM MHz Double notch Port2 Port1 Port3 Beam tube model used to get transmission characteristic. TM01-TEM: 0-3GHz

12 TM011 Mode at GHz Due to mode transition Need to optimize the antenna structure for this mode cut off:te MHz cut off: TM MHz TE11-TEM transmission TE11-TEM transmission TM01-TEM transmission Double notch Source port Min. S21=-66.99dB 12

2-cell cavity impedance 1E+09 1E+08 H Z CST W 1E+09 CST H Z W 1E+08 Z(ohm) 1E+07 1E+06 TM011 Z(ohm/m) 1E+07 1E+06 1E+05 1E+04 TM021 TM020 TM012 0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 f

13 2-cell cavity impedance 1E+09 1E+08 H Z CST W 1E+09 CST H Z W 1E+08 Z(ohm) 1E+07 1E+06 TM011 Z(ohm/m) 1E+07 1E+06 1E+05 1E+04 TM021 TM020 TM f (MHz) 1E+05 1E+04 TE111/ TM f (MHz) TM120/ hybrid TM111/ TE121 Impedance threshold: H~336 cavity, W&Z~168 cavity Feedback~3.3 ms Transverse beta~30 m TM010: Qe=1.7e+12 Frequency spread effect not include 13

14 Frequency spread Considering the whole RF system, there will be finite tolerances in the cavity construction. To find the total effects of all the RF cavities, we need to take into account the spread in the resonance frequencies of different cavities. For small frequency spread, this will result in an effective quality factor Q of the whole RF system. A. Hofmann and J. R. Maidment, LEP note 168, R ( Z / R ) R = th s L s max L 2( E0 / e) s N f I cav L 0 p z f R = GHz, Q= , 336 RF cavities [1] N. Wang et al. Impedance and collective effects of CEPC, 55th ICFA Advanced Beam Dynamics Workshop on High Luminosity Circular e+e- Colliders Higgs Factory (HF2014) 14

Motivation : HOM propagating among cavities Higher order modes are not required to be necessarily confined in the individual resonators (multi-cavity modes, inter-cavity modes etc.

15 Motivation : HOM propagating among cavities Higher order modes are not required to be necessarily confined in the individual resonators (multi-cavity modes, inter-cavity modes etc.) Consideration of entire cavity chain for a reasonable RF analysis is needed Thomas Flisgen Methods: Direct Computations-----using software and supercomputers State-Space Concatenations algorithm HOM propagating properties will be one important issue for further research. [1] Thomas Flisgen, Johann Heller, and Ursula van Rienen, Generation of a Compendium of Resonant Modes in the Chain of 3 rd Harmonic TESLA Cavities for the European XFEL, (HOMSC16), Warnemünde, Rostock, Germany, 22 th 24 th of August

16 Thermal analysis - Procedure Static thermal analysis results Dynamic thermal analysis procedure Initial static temperature distribution electric conductivity iterative procedure Power dissipation (calculated based on Rs) Preliminary calculation According to the temperature distribution T, calculate the surface resistance Rs According to the dissipation power Ploss, load surface heat source No According to the surface heat source, recalculate the temperature distribution T ΔT <convergence factor δ Yes Get the steady state temperature distribution Materials thermal conductivity change with temperature Thermal conductivity iterative procedure 16

17 Thermal conductivity data Nb-RRR300 Cu-RRR50 Al % Nb55Ti [1] Fermilab. Material Properties for Engineering Analyses of SRF Cavities ES , Rev. [2] I. E. Jensen et al. Brookhaven National Laboratory Selected Cryogenic Data Notebook. BNL R, Vol. II. 17

18 Ceramic properties AL 300, a top quality alumina ceramic of 97.6% Al2O3 [1] Morgan TechnicalCeramics Wesgo Ceramics. 18

19 With He tank 1W ~8.67 K 1W Al gasket~2.55k ~8.63 K ceramic~8.67k 5K anchor He tank ~2.50 K 19

20 EM field E field E field H field EM model for fundamental mode: E acc =20 MV/m, f=650 MHz RF loss from surface magnetic field H field EM model for HOM: P hom =1 kw, f=1700 MHz RF loss from surface magnetic field 20

21 Power dissipation analysis Ceramic Probe Loop P c = 1 2 R surf H s n 2 ds P D = πftan(δ)ε0εr R BCS 1 T f2 exp T Gasket E 2 dv (T < Tc 2 ) Copper ~ anomalous skin effect T Coupler parts Power dissipation from fundamental mode [mw] Power dissipation from 1kW HOM 1) [mw] Total [mw] Loop (Nb) e T (Cu) Probe (Cu) Gasket (Al) e ceramic Al2O3 1) 1700 MHz 2) To be adjusted according to the resulting temperature. Notes Rs=22.8nΩ 2) for fundamental mode (20 MV/m) Rs=97.5 nω for HOM Rs=1.18 mω for fundamental mode (20 MV/m) Rs=2.1 mω for HOM Rs=1.18 mω for fundamental mode (20 MV/m) Rs=2.097 mω for HOM Rs=0.3 mω for fundamental mode (20 MV/m) Rs=0.48 mω for HOM 21

Dynamic results A B I D C E H Boundary conditions: A

553E-3 W Heat: 1.108E-3 W Heat: 0.271E-3 W ~8.

22 Dynamic results A B I D C E H Boundary conditions: A B C D E F Heat: 1 W Heat: 1 W Heat: 0.03 W Heat: 0.553E-3 W Heat: 1.108E-3 W Heat: 0.271E-3 W ~8.72 K ~8.69 K F G H I Heat: 0.217E-3 W Temperature: 2.5K Temperature: 5K ~2.502 K G Power dissipation caused by the fundamental mode and 1 kw HOM is in the mw range. A helium tank outside the HOM coupler is needed. No active cooling by liquid helium inside the loop. 22

23 Multipacting analysis Ar discharge cleaned No multipacting SEY--Ar discharge cleaned Ar discharge cleaned No multipacting f=650 MHz, Eacc=20MV/m, Ar discharge cleaned 23

24 Tolerance parameters (1) (4) + (3) (2) (1) Probe offset on horizontal direction. (2) Gap (between probe and loop part) offset on vertical direction. (3) Loop offset on vertical direction. (4) Loop offset on horizontal direction. (5) Loop rotation around x axis. (6) Loop rotation around z axis. (6) (5) Tolerances used for vertical and horizontal direction is ±0.5 mm. Tolerances used for rotations around x and z axis is ±0.5 deg. 24

25 Tolerance parameters + (1) Probe offset on horizontal direction Only affect the bandwidth of fundamental mode Little effect on HOMs

26 Tolerance parameters + (2) Gap offset on vertical direction Little effect on fundamental mode For HOMs: affects the frequency of the first and second peak on S21 curve

27 Tolerance parameters 0-20 loop-up=-0.5mm loop-up=0mm loop-up=+0.5mm S21(dB) (3) f (MHz) Loop offset on vertical direction For fundamental mode: only affects damping intensity For HOMs: affects the frequency of the first and second peak on S21 curve

28 Tolerance parameters + (4) Loop offset on horizontal direction Little effect on fundamental mode Little effect on HOMs

29 Tolerance parameters 0 xrot=-0.5deg xrot=0.5deg xrot=0deg S21 (db) (5) Frequency(MHz) Loop rotation around x axis Little effect on fundamental mode Little effect on HOMs

30 Tolerance parameters 0 zrot=-0.5deg zrot=0.5deg zrot=0deg S21(dB) (6) Frequency(MHz) Loop rotation around z axis Little effect on fundamental mode Little effect on HOMs

31 Mechanical structure and material ceramic Stainless steel copper Nb55Ti Nb55Ti Nb Nb Is copper coating on stainless steel part necessary? Welding procedure: The structures which are made of niobium and Nb55Ti will be e-beam welded. Ceramic with copper and ceramic with stainless steel is brazing. 31

32 Processing and assembling steps Gasket (Al) used for sealing. The coupler mainly include two parts. The full 3D structure. 32

33 test in a test bench HOM coupler test plan Text box ~ circular waveguide or coaxial waveguide Test included the characteristics of the transmission curve and fundamental mode frequency vertical test with cavity (HOM excitation) horizontal and beam test with cavity in cryomodule 33

34 Other considerations Multi-bunch induced HOM power with the real fill pattern. Thermal analysis by ANSYS. HOM propagating among cavities. Influences of beam tube modes. 34

35 Conclusion Finish the RF design. The coupler design can meet the impedance requirements. Finish the thermal analysis. Determined the mechanical structure design. We hope to collaborate with KEK in the HOM coupler design as well as HOM analysis. 35

36 Thank you! 36

37 Backup slides 37

Normalized time-average HOM power T b T 0 Time structure for H-pole: Bunch uniformly distributed Bunch number/beam=412 Time spacing between bunches: T b =406.2 ns Bunch revolution time: T 0 =333.

38 Normalized time-average HOM power T b T 0 Time structure for H-pole: Bunch uniformly distributed Bunch number/beam=412 Time spacing between bunches: T b =406.2 ns Bunch revolution time: T 0 =333.6μs 2.47MHz TM011 mode f= MHz [1]Sang-ho Kim, Marc Doleans, HOM findings and HOM induced power in the superconducting linac of the intense pulsed proton accelerator, SNS/AP Technical Note No.10, August 29, [2] Haipeng Wang s, Beam Induced HOM Power Spectrum in JLab 1MW ERL-FEL, April. 28, 2005

39 Parameters of CEPC double ring Higgs W Z Number of IPs Energy (GeV) SR loss/turn (GeV) Half crossing angle (mrad) Piwinski angle N e /bunch (10 11 ) Bunch number Beam current (ma) SR power /beam (MW) Bending radius (km) Momentum compaction (10-5 ) IP x/y (m) 0.171/ / /0.002 Emittance x/y (nm) 1.31/ / / Transverse IP (um) 15.0/ / /0.125 x / y /IP 0.013/ / /0.054 RF Phase (degree) V RF (GV) f RF (MHz) (harmonic) (217800) 650 (217800) Nature z (mm) Total z (mm) HOM power/cavity (kw) 0.41(2cell) 0.36(2cell) 1.99(2cell) Energy spread (%) Energy acceptance (%) 1.5 Energy acceptance by RF (%) n Life time due to beamstrahlung (min) 52 F (hour glass) L max /IP (10 34 cm -2 s -1 )

CEPC Main Ring SRF Parameters (1) ZJ20170410 Main Ring parameter: WD20170306 H W Z Luminosity / IP [10 34 cm -2 s -1 ] 2 5 12 SR power / beam [MW] 32 32 16 RF voltage [GV] 2.1 0.41 0.

40 CEPC Main Ring SRF Parameters (1) ZJ Main Ring parameter: WD H W Z Luminosity / IP [10 34 cm -2 s -1 ] SR power / beam [MW] RF voltage [GV] No staging (except super-z and higher energy). Same RF cavity for H, W, Z. No cavity push-pull. Beam current / beam [ma] Bunch charge [nc] Bunch length [mm] Cavity number in use / beam (650 MHz 2-cell) Gradient [MV/m] (with margin for HV-H) Input power / cavity [kw] (with margin for HL-H) Klystron power [kw] (2 cavities / klystron) HOM power / cavity [kw] Cryomodule number (6 cavities / module) Q 2 K at operating gradient (long term) 1E10 1E10 1E10 Total wall 4.5 K eq. [kw] Challenging input power, low heat load, input coupler short to reduce module diameter Cavity acceptance Q 0 > 4E10 (N-doping), Module horizontal test > 2E10 (clean assembly and magnetic hygiene) 40

41 CEPC Main Ring SRF Parameters (2) ZJ Main Ring machine parameter: WD H W Z Optimal Q L 1.0E6 2.7E5 1.2E5 Relative optimal Q L (to H) Extra power (if fixed optimal coupling for H) % 155 % Cavity bandwidth [khz] Optimal detuning [khz] Cavity time constant [μs] Cavity stored energy [J] Max relative voltage drop for 1 % beam gap 0.9 % 3.3 % 23.2 % Max phase shift for 1 % beam gap [deg] Max relative voltage drop for 4+4 APDR 10 % 77 % decelerate Max bunch train phase shift for 4+4 APDR [deg] decelerate Variable coupler needed for W&Z, and even for Higgs itself to save power. Heavy beam loading damp acceleration mode CBI instability during injection and top-up operation Even 1 % beam gap will have large bunch phase shift change fill pattern from one long gap to many small gaps Alternative APDR (4+4 trains) scheme may not work unless phase shift corrected by beat cavity or other method. 41

42 CEPC Booster SRF Parameters (preliminary) ZJ GeV injection. Booster machine parameter: CX H W Z Extraction beam energy [GeV] Bunch charge [nc] Beam current [ma] Extraction RF voltage [GV] Extraction bunch length [mm] Cavity number in use (1.3 GHz TESLA 9-cell) Gradient [MV/m] Q L 2E+07 2E+07 2E+07 Cavity bandwidth [Hz] Input power per cavity [kw] (remained detuning 10 Hz) SSA power [kw] (one cavity per SSA) HOM power per cavity [W] Cryomodule number in use (8 cavities per module) Q 2 K at operating gradient (long term) 2E+10 2E+10 2E+10 Total wall 4.5 K eq. [kw] (assume CW) Narrow bandwidth, microphonics Voltage ramp 12 times in 1 s LLRF challenge 42

43 2 W loss from coaxial line 1W 1W ~9.4K Ceramic Probe T 5K anchor ~9.72K Loop ~7 K Gasket Thermal model used ~9.16K 43

44 Power dissipation analysis Ceramic Coupler parts Power dissipation from fundamental mode [mw] Power dissipation from 1kW HOM 1) [mw] Total [mw] Notes Probe T Hook (9.16K) Rs=1.41m Ω 2) for fundamental mode (20 MV/m) Rs=2.27 mω for HOM T (9.72K) Rs=1.18 mω for fundamental mode (20 MV/m) Rs=2.097 mω for HOM Loop Gasket Probe (9.72K) Rs=1.18 mω for fundamental mode (20 MV/m) Rs=2.097 mω for HOM P c = 1 2 R surf H s n 2 ds P D = πftan(δ)ε0εr R BCS 1 T f2 exp T E 2 dv (T < Tc 2 ) Copper ~ anomalous skin effect R s = ωμ 2σ = πfμ 0μ r σ Gasket (Al) (7K) 5.18e e e-3 Rs=0.3 mω for fundamental mode (20 MV/m) Rs=0.48 mω for HOM ceramic Al2O3 1) 1700MHz 2) To be adjusted according to the resulting temperature. 44

Dynamic results 1W 1W Ceramic Probe T ~30 K Loop 15.

45 Dynamic results 1W 1W Ceramic Probe T ~30 K Loop W Gasket ~230K Nb is near the critical temperature, the magnetic field near loop part is large, the RF power is large, so it quenches soon. 45

Advance on High Power Couplers for SC Accelerators

Advance on High Power Couplers for SC Accelerators Eiji Kako (KEK, Japan) IAS conference at Hong Kong for High Energy Physics, 2017, January 23th Eiji KAKO (KEK, Japan) IAS at Hong Kong, 2017 Jan. 23 1