PRISM RF C Ohmori KEK 200972 PRISM FFAG

PRISM RF C. Ohmori （KEK) 2009/7/2 PRISM FFAG workshop@Imperial College

Contents PRISM RF Introductions Present status, RF for a beam RF for 6 cell ring Upgrade plan EMMA RF RF system for PRSIM 2009/7/2 PRISM FFAG workshop@Imperial College

Requirements for RF High voltage at 3. 8 MHz Total 2 -3 MV 200 k. V/m 8 straights for RF 2009/7/2 PRISM FFAG workshop@Imperial College

Requirements for RF • Saw-Tooth RF –Linear RF bucket –Composed of 3 harmonics 2009/7/2 PRISM FFAG workshop@Imperial College

MA cavity for PRISM High field gradient at low frequency Wideband (low Q) Thin cavity (about 30 cm / cavity ) Use the maximum size for MA cores (1. 7 m X 1 m) Very low duty RF system To reduce the cost Small tetrodes for the end stage Small APS (anode power supply) 2009/7/2 PRISM FFAG workshop@Imperial College

High Field Gradient : around 200 k. V/m few MV RF for quick phase rotation (around 1. 5 us) Dedicated system for pulse operation (low duty : 0. 1%) 2009/7/2 PRISM FFAG workshop@Imperial College

Characteristics of Magnetic Cores Ferrites シ ャ ン ト イ ン ピ High Loss Effect ー 200 V/div, 5 ms/div ダ ン ス に 比 例 電圧に比例 2009/7/2 2000 Gauss Magnetic Alloys PRISM FFAG workshop@Imperial College

m 7. 1 1 m m 0. 1 2009/7/2 PRISM FFAG workshop@Imperial College

Thin RF cavities surrounded by RF amplifiers 2009/7/2 PRISM FFAG workshop@Imperial College

Dedicated system for low duty AMP Use small tubes Works for short moment; 1 -2 us X 1 k. Hz For 1 k. Hz repetition, need to minimize RF-ON time 99 % of time: zero anode current, 99. 9%: zero RF output Cavity loss : few k. W Tube loss : few ten k. W APS Old fashion to minimize cost: Crowbar, 3 -phase Full -wave rectification J-PARC : 1 MW system, no crowbar, switching with IGBT Supplies power to 4 AMPs, several MW in total. 2009/7/2 PRISM FFAG workshop@Imperial College

Cathode current ＲＦＯＮ Tube ON by modulation of CG voltage 2009/7/2 PRISM FFAG workshop@Imperial College

Dedicated RF system for low duty 100 k. W tube AMP, >1 MW output 1. 4 X 0. 7 X 0. 8 m J-PARC 600 k. W tube AMP 500 k. W output 1. 4 X 1. 0 X 2. 4 m Multi-MW APS, 1 X 1. 5 X 2. 0 m 1. 2 MW APS for J-PARC, 4. 5 X 2 X 2. 7 m 2009/7/2 PRISM FFAG workshop@Imperial College

STATUS of PRISM RF RF frequency 5 -> 3. 8 MHz (larger circumference)->2 MHz for a beam AMP has modified for low frequency operation. Achieved 30 k. V/gap, 100 k. V/m. 100 W/core @2 MHz 128 W/core @3. 8 MHz 244 W/core@ 18 MHz Number of cores: 4 instead of 6 (design : 6 cores, total 1 k. W) Core impedance : 2009/7/2 PRISM FFAG workshop@Imperial College

6 cell ring 2009/7/2 PRISM FFAG workshop@Imperial College

2009/7/2 PRISM FFAG workshop@Imperial College

Gap voltage 2009/7/2 PRISM FFAG workshop@Imperial College

6 cell PRISM Test using a beam has been carried out. At 2 MHz, 100 k. V/m was achieved Saw-tooth will be tried. 2009/7/2 PRISM FFAG workshop@Imperial College

Hybrid RF system Proposed by A. Schnase. Combination of MA cavity with a resonant circuit composed by inductor and capacitor. Developed for J-PARC RCS cavities. f=1/2 p√LC 1/L=1/Lcore+1/Lind Q=Rp/w. L Rp: shunt 2009/7/2 J-PARC: add C and L to control Q and f PRISM : add L to control f PRISM FFAG workshop@Imperial College

Parallel inductor for J-PARC 2009/7/2 Inside of PRISM AMP PRISM FFAG workshop@Imperial College

Expected impedance with parallel inductor Total C =180 p. F Hybrid (+40 u. H inductor) Hybrid (+ 8 u. H inductor) 2009/7/2 Hybrid ( +18 u. H), 3. 8 MHz PRISM FFAG workshop@Imperial College

Saw-Tooth : RF Cavity will be a wideband cavity. But, bandwidth of AMP is still limited (1/RC). To obtain high RF voltage, a large drive voltage is still required for CG-Cathode. Solutions Low duty high power DAMP based on CERN/J-PARC DAMP. Drive from both CG and Cathode is possible in case of short pulse operation. Narrow bandwidths are enough for both CG and Cathode. -> save the cost for Driver AMP Both need test. 2009/7/2 PRISM FFAG workshop@Imperial College

X-ray Over 30 k. V anode voltage, soft X-ray was observed. Additional X-ray shields were add on vacuum tubes and AMP. Most sensitive X-ray detector was used. 2009/7/2 PRISM FFAG workshop@Imperial College

Upgrade Plan High Field Gradient Cost reduction 2009/7/2 PRISM FFAG workshop@Imperial College

Improvements of cavity cores X 2 by annealing under magnetic field for thinner ribbon Small cores : OK Large core ? 2009/7/2 ∝shunt impedance Improvements of cavity impedance PRISM FFAG workshop@Imperial College

How to improve MA consists of Fe, Si, B, Cu and Nb. Amorphous ribbon (<20 mm) is annealed and crystallized. Combination of magnetic field during annealing and thinner ribbon (13 mm) The small crystal has an axis magnetized easily. By the special annealing, the axis is equal. But relation between core impedance and this effect is not clear. Small cores : proved by Hitachi Metal Large core : need big magnet and special oven. => Appling JSPS grant to produce these special core in KEK. B-H curve of MA core produced by annealing with/without magnetic field. (by Hitachi Metal) 2009/7/2 PRISM FFAG workshop@Imperial College

N_backward Decayed positron N_forward Polarized μ finemet ‖cylinder ⊥cylinder Asymetry =(N_forward - N_backward)/(N_forward+N_backward) 2009/7/2 PRISM FFAG workshop@Imperial College

2009/7/2 PRISM FFAG workshop@Imperial College

2009/7/2 PRISM FFAG workshop@Imperial College

It clearly showed the effects on magnetic properties by applying the magnetic field during the crystallization process in production. It suggests that the magnetic axis of nano-scale crystalline in FT 3 L are aligned to the direction of the magnetic field during the annealing process. In the case that the initial spin direction of implanted muons is perpendicular to the assumed easy-axis of nano-crystalline FT 3 L, the polarization of muons showed a quite fast damping. In contrast, a slow relaxing time spectrum was obtained when the initial direction was aligned with the axis along which the magnetic field had been applied during the annealing, suggesting that the muon polarization is retained due to the local magnetic field. On the other hand, such a drastic change was not seen in the case of FT 3 M. It turned out, however, that an anisotropic behaviour against the initial muon spin direction in FT 3 M was still observed, in spite of the absence of the magnetic field during the production. The muon implanted in parallel to the ribbon surface depolarizes slightly faster than that implanted in perpendicular. This may suggest that the shape of MA, e. g. thickness, causes magnetic anisotropy. It hints that the characteristics of FT 3 L depends more on the thickness of ribbon than on an expected eddy current effect. 2009/7/2 PRISM FFAG workshop@Imperial College

High impedance core Further experiments using m. SR to confirm the effects of ribbon thickness. We will Make larger cores to confirm the impedance measurements. 27 cm size cores will be produced in this summer. These R&D are also important for high intensity accelerators (J-PARC RCS, MR, ISIS-upgrade, CSNS etc. ). To confirm finally, it is important to build a cavity structure. 2009/7/2 PRISM FFAG workshop@Imperial College

Cost issues So far, 6 cores were necessary to generate 50 k. V. However, 4 cores will be enough to generate 60 k. V in case of high impedance cores. Achieving 1 k. W impedance will make a system design similar to original one (6 cores, 5 MHz). The cavity cost seems to be larger than other cost in case of PRISM. Higher voltage per core is preferable. However, total cost to obtain 2 MV is still expensive. 2009/7/2 PRISM FFAG workshop@Imperial College

conclusions Beam test was performed by using PRISM rf cavity Demonstrate > 100 k. V/m. Also plan to test saw-tooth RF To reduce the rf cost, developments of high impedance cores are important. 2009/7/2 PRISM FFAG workshop@Imperial College

EMMA MA System * Many FFAG applications require slow acceleration * Non-scaling FFAGs cross many resonances - Nonlinear resonances - Imperfection resonances * Resonances damage beam more when you cross them slowly * There is thus a minimum rate at which you can cross resonances - May depend on magnitude of errors * Low-frequency RF to allow slow acceleration - EMMA as-is only allows very rapid acceleration - Primarily due to high-frequency RF system * Accelerate rapidly then reduce rate - Start with 100 turns to insure success - Reduce acceleration rate and study effects 2009/7/2 PRISM FFAG workshop@Imperial College

Parameters frequency 18 MHz frequency sweep 3% Total Voltage 100 k. V per turn Number of cavities 3 Voltage 33 k. V Length of cavity 10 cm Number of MA cores 2 per cavity Size of MA core 27 cm O. D, 10 cm I. D x 2. 5 cm MA core Cut core Q-value About 20 Cavity impedance 700 W (1. 4 k W ) Core material FT 3 M (FT 3 L) 2009/7/2 PRISM FFAG workshop@Imperial College =1. 3 GHz/72 100 turns/cycle

EMMA MA CAVITY 2009/7/2 PRISM FFAG workshop@Imperial College