RF Design of Normal Conducting Deflecting Cavity

Transcription

1 RF Design of Normal Conducting Deflecting Cavity Valery Dolgashev (SLAC), Geoff Waldschmidt, Ali Nassiri (Argonne National Laboratory, Advanced Photon Source) 48th ICFA Advanced Beam Dynamics Workshop on Future Light Sources March 1-5, 2010 SLAC National Accelerator Laboratory Menlo Park, California

2 Motivation Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 2

3 Simplified Configuration RF source Klystron Hybrid Hybrid Hybrid 1 st and 2 nd structures Undulator 3 rd and 4 th structures Storage ring Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 3

4 Outline Design considerations Thermal stability Wakefield damping Cavity pulsing X-band deflector Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 4

5 Design requirements for APS deflecting cavity, assuming 4 cavities in the system Frequency Deflecting Voltage Available power Repetition rate GHz 2 MV per structure 4 MW per structure ~1000 Hz The main constraints on the rf design are set by high average power loss in the cavity and heavy wakefield damping. Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 5

6 Design Considerations Available power ~ 25 MW for whole system Large aperture radius 21 mm => Low shunt impedance Pulsed heating < 100 o C => maximum surface magnetic field < 300 ka/m for 4 us pulse Maximum surface electric fields < 100 MV/m No field amplification on edges of input coupler => Fat lip coupler Heavy loading of LOM / HOM s High average power operation Do no harm the set of the transverse cavities should not degrade existing operation modes of the APS ring Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 6

7 Evolution of the Deflecting Cavity Design Input coupler not shown (a) 9-cell symmetric cavity with center-fed input coupler and no wakefield damping (b) 9-cell with wakefield damping and external loads (c) 9-cell with heavy wakefield damping and internal loads (d) 3-cell with heavy wakefield damping and internal loads Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 7

8 3-Cell deflector fed from middle cell, 2 MeV vertical kick with 2.83 MW of input rf power x beam axis z input waveguide damping waveguides y Surface electric fields for 2 MeV Surface magnetic fields for 2 MeV transverse kick. Maximum surface electric transverse kick. Maximum surface fields is 60 MV/m. magnetic fields 240 ka/m. Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 8

9 3 cell deflector fed from end cell, transverse kick 2 MeV, input power 2.86 MW 232 mm Surface electric fields for 2 MeV kick, maximum fields 60.5 MV/m Surface magnetic fields for 2 MeV kick, maximum fields 240 ka/m With an end-cell coupling the loaded Q of the next-to-working mode is reduced from to for critically coupled cavity and can be reduced more for over-coupled cavity. Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 9

10 Deflecting Cavity with Damping Waveguides and Loads Beam axis Input coupler LOM-HOM loads in ridged waveguides HOM load in rectangular waveguide Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 10

11 Deflecting Cavity Parameters Frequency Cavity Length GHz ~23 cm Deflecting Voltage 2 MV Peak Power 2.86 MW Working mode Q o R t / Q 117 Beam pipe aperture radius 21 mm Iris radius 22 mm Phase advance per cell π Structure length w/o beam cm pipes Iris thickness 18 mm Duty Factor 0.147% Kick / (Power) 1/ MeV/MW 1/2 Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 11

12 Transverse Electric and Magnetic Fields on Axis Electric Field [MV/m ], Magnetic Field * Zo [MV/m] Ey Hx Ex Hy z [mm] Transverse electric and magnetic fields on axis vs. z for 2MeV transverse kick Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 12

13 High Average Power Operation Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 13

14 Iris Optimization Large >21 mm radius aperture necessary to avoid radiation heating Klystron pulse width reduced to 1.3 μs to reduce thermal load => duty factor is 0.147%. Cavity iris thickened to 18 mm to reduce peak thermal gradients and increase cooling efficiency. Example: Optimization of iris thickness T Iris thickness (mm) R Iris radius (mm) Peak power density on cavity s iris surface T12R22: 0.94 MW/m 2 T15R23: 0.83 MW/m 2 T18R24: 0.68 MW/m 2 Elliptical iris with thickness of 18 mm and radius of 22 mm was chosen, for which peak power density was reduced by ~30% while reducing shunt impedance by ~15% Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 14

15 High Average Power Operation Thermal Stability Cavity will deform due to the rf losses and the deformed cavity may change field profile and result in increased losses. Deflecting field profile will be perturbed and may require additional power in order to maintain prescribed deflecting voltage. If this continues, the cavity will absorb more and more heat and exceed material stress limits. π-mode in the cavity is backward standing wave and it may be more susceptible to rf thermal issues. We need to show that the cavity is thermally stable. Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 15

16 RF / Thermal Flowchart with ANSYS Start of rf / thermal analysis Self-contained analysis performed entirely in ANSYS Loop continues until power level converges Undeformed cavity Deformed cavity Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 16

17 Multi-Pass Magnetic Field Distortion Minimal field distortion after 43 iterations Black: Original Field Magnitude Red: Deformed Cavity Field Magnitude Original / deformed fields are nearly identical Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 17

18 Multi-Pass Frequency Shift Maximum frequency shift from iterationto-iteration is 2.4 khz Frequency Shift Study done on the most recent slightly modified geometry Frequency (khz) Frequency converges after the third iteration Slater Method Slater: End Cell Slater: Center Cell Ansys HF Frequency shift shown for a full end / center cell Iteration Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 18

19 Multi-Pass Power Loss Maximum power variation from iterationto-iteration is 10.0 W 2098 Power Loss Study done on the most recent slightly modified geometry Power iterations simulated Iteration Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 19

20 Wakefield Damping Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 20

21 Low-Order-Modes/ Higher- Order-Mode analysis LOM / HOM monopole and HOM dipole modes were analyzed using the frequency-domain finite-element code HFSS up to 5-6 GHz. MAFIA and GdFidL time-domain wakefield solvers were used to evaluate the monopole and dipole mode impedances and compare with HFSS as a verification. GdfidL parallel simulations calculated mode impedances > 12 GHz with λ/10 resolution in the high permittivity (ε r = 30) damping material. 3-cell structure was adopted since the heavily-loaded 9-cell cavity could not be adequately damped for beam stability. Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 21

22 LOM/HOM Damping Considerations Mode coupling from cavity to damping waveguide was optimized to increase the damping of HOM / LOM s and reject the operating mode (while keeping thermal issues under control) Monopole and horizontal dipole modes are heavily damped with a loaded Q of less than 200 for the majority of modes. Vertical dipole modes are not easily damped since their frequency and field configurations may be close to the operating mode. Ridge waveguide has been integrated into the design to improve damping of the vertical modes. Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 22

23 Suppression of Long-Range Wakes x Monopole wake Horizontal wake Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 23

24 Monopole Modes R s 2 V = 2P l 2022 MHz 1956 MHz Q l ~ 130 Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 24

25 Horizontal Dipole Modes R t = V 2 r= r o 2 2Pl kro 2620 MHz 2663 MHz Q l ~ 20 Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 25

26 R t = V 2 r = r o 2 2Pl k ro Vertical Dipole Modes 2815 MHz MHz 2827 MHz mode has a large R t /Q and is damped minimally. It is the greatest HOM contributor to the vertical long range wake Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 26

27 Z-Impedance with Inter-Cavity-Coupling Effects 1956 MHz 2022 MHz o HFSS GdFidl f p *R S < 0.8 MOhm-GHz Monopole Impedance Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 27

28 Horizontal Transverse -Impedance with Inter-Cavity-Coupling Effects 2.5 MOhm/m 2620 MHz o HFSS GdFidl Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 28

29 Vertical Transverse Impedance with Inter-Cavity-Coupling Effects 2815 MHz 2827 MHz 2.5 MOhm/m o HFSS GdFidl 2930 MHz Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 29

30 LOM/HOM Damper Loads Simplified Model Load Y Load X Center Load X Outer Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 30

31 Beam Induced Damper Load (W) Total losses due to monopole modes is 2.03 kw Loss factor: Mafia: 7.802*10 11 GdFidl: 7.82* ma maximum beam current, 24 singlets fill pattern generates greatest damper losses Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 31

32 Cavity Pulsing Critically matched coupler requires ~2.8 MW peak power pulsed at 1 khz rep rate for a 2 MeV kick of the 16 ma bunch. Net deflecting voltage must be reduced below 13 kv for the following 86 ma bunch train (while cavity is emptying). Variations in cavity Q due to manufacturing or contamination creates a voltage differential Overcouple cavity to reduce fill time constant Reverse klystron phase to empty cavity more quickly Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 32

33 Pulse Shape and Approximate Timing Diagram Klystron power (100 ns ramp) 16 ma enters Cav. #1 (Sector 6) t = 0 Cavity power Timing pictorial in APS SR 86 ma enters Cav. #1 Duty factor 0.147% V t = 0.2 MV Red: Ideal square input Black: 100 ns ramp Q l ~ 6000 for critically coupled input coupler Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 33

34 Deflecting Voltage Differential for 86 ma Bunch Single cavity in cavity set #2 with Q 1-5% lower than nominal value (12,000) All cavities have Q s ranging from Q u = 12,000 ± 600. Cavities optimally positioned based on Q s. β = 1.7 1% 2% 3% 4% 5% Peak input power at cavity Net difference in deflecting voltage between cavity set #1 and cavity set #2 P=3.0 MW Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 34

35 APS Deflector Summary The deflecting cavity design has evolved to satisfy strict requirements in beam stability and high average power losses. A 3-cell standing wave structure will produce 2 MeV kick with ~2.86 MW input power for a critically coupled cavity. Four structures will produce a 4 MeV initial kick and a 4 MeV recovery kick. A single, commercially available 25 MW klystron is sufficient. A set of two structures will occupy less then 50 cm of beam space. Low-order and high-order modes are heavily loaded by six ridged and four rectangular waveguides with internal loads. Cavity power coupling has been characterized for various parameters affecting parasitic voltage kicks to the beam. Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 35

36 X-band Deflector Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07

37 6 cell SW GHz deflector Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07

38 Periodic cell of Pi standing wave deflector, 0.25 MW/cell, deflecting gradient 26 MV/m Maximum surface magnetic fields 410 ka/m, Pulse heating 23 deg. C for 100 ns pulse. Maximum surface electric fields 105 MV/m. a = 6 mm t = 2 mm, round iris Q=7,792 Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07

39 Waveguide coupler for 6 cell SW X-band deflector, 1.5 MW of input power, deflection 2 MeV Maximum surface magnetic fields ~420 ka/m, Pulse heating 24 deg. C for 100 ns pulse. Maximum surface electric fields ~105 MV/m. Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07

40 Parameters of 6 cell X-band SW deflector Frequency GHz Beam pipe diameter 10 mm One cell length mm Phase advance per cell One cell kick Structure kick (6 cells) π 0.34 MeV/Sqrt(0.25 MW) 1 MeV/Sqrt(375 kw) Unloaded Q 7800 Loaded Q 3800 Maximum Electric field 53 MV/m / Sqrt(375 kw) Maximum Magnetic field 210 (ka/m) / Sqrt(375 kw) Structure length (with beam pipes) 12 cm Near mode separation 13.6 MHz Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07

41 Summary of RF Power Considerations for X-band option SLAC 11 GHz XL-4 klystron can produce 50 MW of power at 120 Hz repetition rate and pulse length 2 μs. We expect that with some development a modified klystron can work in low power mode ~5 MW at 1 khz repetition rate. A pair of 6-cell deflectors bracketing one undulator will need about 1 MW, so one such klystron is capable of powering 5 (~20 with SLED) short x-ray pulse stations at ~1 khz repletion rate. For 1kHz operation, average power loss in 6 cell deflector would be manageable 200 W. With lower beam energy (say for SPEAR III) the deflectors could be driven by commercial 100 kw klystrons, like CPI s VKX-7876E. Deflecting Cavity RF Design and Analysis G. Waldschmidt 8/23/07 41