Main Injector Cavity Simulation and Optimization for Project X

Transcription

1 Main Injector Cavity Simulation and Optimization for Project X Liling Xiao Advanced Computations Group Beam Physics Department Accelerator Research Division Status Meeting, April 7, 2011

2 Outline Background - Project X Conceptual Design - Current FNAL MI RF Cavity - New Project X MI RF Cavity I and II Designs MI Cavity I & II Simulations MI Cavity II Optimization Summary 2

3 Project X Conceptual Design Experimental Hall Project X will use FNAL existing RR/MI complex, but require upgrading MI RF system. 2GeV 1mA CW H - Recycler Storage Ring 8GeV RCS 8GeV 190kW 120GeV 2MW Neutrinos Main Injector Accelerator Single turn 8GeV 3

4 Current FNAL MI RF Cavity Using ferrite tuners to tune cavity frequency to the velocity of the circulating beam starting from low to the speed of light. Changing μ of ferrite to tune the cavity frequency by applying a variable external B-field to the ferrite. Project X requires: 4 X the current number of particles 3 X the current beam intensity 6 X the current beam power Current 30+ yrs MI RF cavity does not have enough power to accelerate 1.6e14 protons to 240GeV/sec for a total power of 2.3MW for Project X even with additional PA. 4

5 New Project X MI RF Cavity Designs Cavity I Cavity II 5 New MI RF Cavity I & II: 1. A quarter wave coaxial resonator with a single accelerating gap using a perpendicularly biased ferrite tuner results in a compact cavity design. 2. A bias field perpendicular to the RF field reduces losses in the ferrite. Cavity II Advantage: 1. Requires only a single vacuum ceramic. Tuner, driver, and HOM dampers will be at atmospheric pressure for easy installation and repair. 2. Conical shape gives support for the inner conductor lever arm.

6 New Project X MI Cavity RF Parameters Parameter Value Units R/Q 50 Ω Q Max. Voltage 240 kv Harmonic number 588 Frequency MHz Number of Cavities 20 FNAL & SLAC signed a MOU to perform the new MI cavity I & II simulations and optimization to meet the requirements for Project X. 6

7 Outline Background MI Cavity I and II Simulations - Operating Mode RF Parameters - Tuning Range vs. Tuner Coupling - Maximum Surface E/B-Fields - Power Distributions - HOM modes MI Cavity II Optimization Summary 7

8 MI Cavity I & II RF Simulations Finite element parallel eigen-solver Omega3P code for the cavity I&II RF simulations. Tetrahedral meshes with curved surfaces and 2 nd order basis functions. About 400k mesh elements for converged results. Cavity I Cavity II 30 ferrite cores with 5mm separation: εr=13.5, tan( )=0.0002, μr=2.5 ~ 1.2, tan( )= (external B=300~2250Gauss) Ceramic window: εr=12, tan( )=0.0001, μr=1, tan( )= Copper coated wall: σ=5.8e7s/m Meshed Computational Models 8

9 Operating Mode RF Parameters Tuner intrusion is 55mm. Required F= ~ MHz, R/Q=50Ω, Q0=10000 MI Cavity μr F(MHz) Q0 R/Q (Ω) ΔF (KHz) I II E B E B The cavity II has a slightly higher R/Q and f than the design values that can be easily adjusted by changing the cavity coaxial line radius and length. 9

10 Tuning Range vs. Tuner Coupling Adjusting the tuner intrusion can change cavity tuning range. Required tuning range is 490KHz with 50 mm tuner intrusion (d). Intrusion d 0 mm: Tuner center conductor is positioned at the cavity outer surface for both cavity I & II. 85/95 mm: The maximum tuner intrusion for cavity I/II, respectively. 10

11 Maximum Surface Field - Es Maximum Surface Vgap=240kV μr=2.5 and μr=1.2 with 20mm gap rounding Cavity I μr=2.5 and μr=1.2 with 20mm gap rounding Cavity II For comparison: RHIC 28MHz Cavity: Es=7.8MV/m The cavity II has higher maximum peak Es than the cavity I. The magnetic permeability of the ferrite won t affect the maximum surface E-field. 11

12 Maximum Surface Field - Bs Maximum Surface Vgap=240kV μr=2.5 μr=1.2 with fully rounded the tuner loop μr=2.5 μr=1.2 with fully rounded the tuner loop Cavity I Cavity II Max. Bs location For comparison: RHIC 28MHz Cavity: Bs=8T (without ferrite tuner) Both the cavity I & II have strong maximum peak Bs at the edges of the tuner loop at ur=2.5. When ur=1.2, the maximum surface B-fields locate at the rounding area between the tuner tank and the cavity. 12

13 Power Distributions Power dissipation in the ferrite cores can cause a temperature rise which will increase the μ-value of the ferrite. Tuner intrusion is 55mm. Vgap=240KV. Ferrite and ceramic are lossy MI Cavity μr F(MHz) R/Q (Ω) Q0 (wall) QL1 (ferrite) QL2 (ceramic) P (kw) (wall) P (kw) (ferrite) P (kw) (ceramic) I II The cavity II has more power dissipation in the ceramic window and less in the ferrite cores than the cavity I. Moving the ceramic window away from the accelerating gap couldn t reduce power dissipation in the window much. 13

14 Monopole HOMs Monopole modes will induce longitudinal coupled-bunch instability. <300MHz μr F (MHz) R/Q (Ω/cavity) Q0 Rs (kω/cavity) E-field μr= Cavity I Cavity II The cavity II monopole HOMs have lower shunt impedances than the cavity I, but they still need to be damped by HOM couplers. 14

15 Dipole HOMs in Cavity I MI-Cavity I (<300MHz) F(MHz) R/Q_T (Ω/cavity) Q0 Rsh_T (kω/mm/cavity) H-dipole Ur= V-dipole Ur=1.2 H-dipole V-dipole E-field μr=2.5 There are additional dipole modes in the cavity I at μ=2.5 due to the coupling of the cavity to the ferrite tuner. 15

16 Dipole HOMs in Cavity II MI-Cavity II (<300MHz) F (MHz) R/Q_t (Ω/cavity) Q0 Rs (kω/mm/cavity) Ur=2.5 H-dipole V-dipole Ur=1.2 H-dipole V-dipole Horizontal Dipole Vertical Dipole E-field μr=2.5 R VT ( ) T Q U 2 V z 2 /( r 0 U R R _ t ( ) _ t * Q*( / c) Q * / c) 2 The cavity II has only one pair of dipole modes below 300MHz, and smaller transverse shunt impedances than the cavity I. 16

17 Vertical Dipole HOMs Due to the ferrite vessel, the vertical dipole modes are all off-center from the ferrite vessel. The vertical dipole modes can be excited and generate transverse instability even the beam is on beam axis. Cavity I: μr=1.2 off center=57mm Cavity II: μr=2.5 off center=29mm The cavity II vertical dipole modes have smaller off-center shifts than the cavity I. 17

18 MI Cavity I & II Vgap=240kV, tuner intrusion=55mm MI-Cavity-I MI-Cavity-II Operating Mode R/Q (Ω) F (MHz) Tuning range Δf (KHz) Max. Es (mv/m) Max. Bs (T) Power Distributions P(kW) (wall/ferrite/ceramic) 117/39/ /28/7 Monopole HOMs Max. R/Q (Ω) Horizontal Dipole HOMs Max. R/Q_T (Ω) Vertical Dipole HOMs Max. R/Q_T (Ω) Max. center shift (mm) The cavity II has better RF performances than the cavity I, and is chosen for the Project-X MI cavity design. 18

19 Outline Background MI Cavity I & II Simulations MI Cavity II Optimization - Optimal 1MHz tuning range - HOM damper - High-Pass Filter - Input coupler Summary 19

20 Optimal 1MHz Tuning Range The tuning range of 487KHz is required for 6GeV to 120GeV operation. Optimal design for the new MI cavity would be 1MHz tuning range. 1MHz tuning range with larger tuning intrusion may cause vacuum break down. There are additional three ways to increase the cavity II tuning range. 3. Moving tuner tank away from the rear end of the cavity (5mm->20mm) 1. Using shorter tuner tank (495mm->460mm) 2. Using narrower tuner loop (120mm->60mm) All the three ways are used for optimal 1MHz tuning range. 20

21 Optimal 1MHz Tuning Range * * Ori.@90mm Cavity d 1MHz tuning range F (MHz) Max. d allowed R/Q (Ω/cavity) P (kw) (wall) P (kw) (ferrite) P (kw) (ceramic) original Optimal The new design will increase the power dissipation in the ferrite cores without affect other RF parameters significantly. 21

22 HOM Damper w/o Filter There are two dangerous monopole HOMs to be damped by HOM dampers. HOM dampers should be equipped with a fundamental mode rejection filter. HOM dampers can handle a certain amount of power. HOM damper is constructed within 1.0/2.3 coaxial line for a larger power capacity. HOM coaxial damper with a large loop located at the rear end of the cavity can damp many HOMs. HOM damper with 45 degree orientation can damp both monopole and dipole modes. Rounded loop design can suppress MP activities at HOM damper. Cavity II with two mirrored HOM dampers w/o filter Bended HOM coaxial damper can take less space for installing two more RF cavities (18- >20). 22

23 Mode Damping w/o Filter ur=2.5 Monopole Modes F=167MHz, R/Q=19.3Ω, Qext=112, Q0=14476 F=280MHz, R/Q=11.8Ω, Qext=92, Q0=18757 Horizontal Dipole Vertical Dipole F=250MHz, R/Q_t=1.6Ω, Qext=203, Q0=21566 F=249MHz, R/Q_t=1.5Ω, Qext=215, Q0=20264 The two prominent monopole modes and dipole modes can be heavily damped with two mirrored HOM dampers. 23

24 High Pass Filter A high-pass filter is a better option for the HOM damper filter design in a cavity with a swing. Higher element filters can provide a sharper rejection response at the fundamental frequency. Filter constructed within larger coaxial lines can handle more power extracted by HOM dampers. 5-element high-pass filter 7-element high-pass filter Cu cavity: Rejection ~60dB 15.2mm X 25mm SRF cavity: X Rejection ~80dB RHIC 28 MHz Cavity HOM Damper (Courtesy: J. Rose) RHIC 56 MHz SRF Cavity HOM Damper (Courtesy: Q. Wu) 24

25 High-Pass Filter Transmission Curve L2 L3 L4 L1 MI cavity HOM 7-element high-pass filter First strong HOM HOM output C1 C2 C3 Fundamental mode L1 L2 L3 and L4 are formed by rods, whose lengths can be adjusted by tuning stubs at their ends on the outer can. C1 C2 and C3 are coaxial low loss Sapphire rings with copper spacer. Sapphire ring: εr=9.5 The fundamental frequency at ~50MHz can be strongly attenuated by the filter (~-80 db). The signal frequency can be transmitted with a less than -7 db of attenuation over 140MHz to 600MHz. 25

26 Power Input Coupler Eimac 8973 power tetrode is chosen as the power source for the new MI cavity. It can operate at both 53MHz and 106 MHz with more than 1MW output power. E-coupling with larger disk at the end of the antenna is used to enhance the input coupling. 50Ω coaxial line (D230/100mm) Eimac 8973 Power Tetrode d 26 Using existing MI RF PA

27 Input Coupling The input coupling should be matching when the maximum beam current is accelerated. Required input coupling β=q0/qload=10000/4000=2.5. (β=2.5, d=35mm) 27

28 MI Cavity II Conceptual Design Tuner HOM dampers w/ filter Window Input Coupler Finalizing the MI cavity design is ongoing to meet the requirements for Project X. 28

29 Cavity I and II Simulations Summary Evaluated FNAL cavity I and II baseline designs Cavity I Cavity II Cavity II Optimization Achieved optimal 1MHz tuning range. Realized a HOM damper design with high-pass filter. Designed input power coupler. Finalizing the design is ongoing. SLAC will continue to collaborate with FNAL on a second-harmonic cavity design. 29

30 Acknowledgments Thank Uli Wienands for helping us prepare this MOU. Thank Cho Ng, Ioanis Kourbanis, and Joseph Dey for their helpful discussions and advices during our monthly phone meetings. 30