Post on 16-Feb-2016
description
2nd harmonic RF perpendicular biased cavity update
C.Y. Tan, W. Pellico, G. Romanov, R. Madrak, and D. Wildman
04 Apr 2014
04 Apr 2014; C.Y. Tan2
Proposed cavity.
Ferrite disk: 380 mm outer diam., 230 mm inner diam., 25 mm thicknessBeO disk: 380 mm outer diam., 230 mm inner diam., 5 mm thickness
490
7040
200
220
390
Ferrite
BeO
solenoid not shown here
Some possible parameters
• Tuning range 76.7 − 107 MHz.• Gap voltage. 100 kV per cavity.• Ramp profile determines losses in the garnet.
04 Apr 2014; C.Y. Tan3
CST Model (done by G. Romanov)
04 Apr 2014; C.Y. Tan4
Complete cavity model with magnetic field generated by solenoid
Solenoid coil
04 Apr 2014; C.Y. Tan5
R11
0R
205
190 mm
Yoke, steel 1008
Coil, 12 turns
Water cooling channels, 10x5 mm
Ferrite G810, R=190 mm, r=115 mm, l=25 mm
Ceramic AlN, l=5mm
230 mm90
mm
20 m
m
This is old picture, not properly scaled. But the marked dimensions are current.
Ferrite tuner details
Static field distribution in ferrite
04 Apr 2014; C.Y. Tan6
Separate solenoid model
Complete cavity model
Field non-uniformity is about 25-30%
Suppression of fields with steel insert.
04 Apr 2014; C.Y. Tan7
Yoke
Coil
Ferrite
Copper drift tube
Steel 1008 insert
Radial distribution of H field at tuner median plane. H field on the
axis can be as high as 85 kA/m. With steel drift tube it is
suppressed.
RF magnetic field distribution in ferrite and losses
04 Apr 2014; C.Y. Tan8
f=75.6 MHz
These power losses spikes are not real. They are due to the singularity of low frequency mesh that is used for thermal simulations
Tuning curves
04 Apr 2014; C.Y. Tan9
0 5000 10000 15000 20000 25000 30000 3500070
75
80
85
90
95
100
105
110
115
Solenoid current, Ampere·turns
Freq
uenc
y, M
Hz
Conversion of the solenoid current to the equivalent uniform field. We can continue to use uniform magnetization – the results are very close.
Thermal analysis
04 Apr 2014; C.Y. Tan10
AlN cooling disks. Thermal losses in the ferrite are 14 kW for V=100 kV. Max T ≈ 75°C with cooling water temperature of 25°C.
020
040
060
080
010
0012
0014
0016
0050
100
150
200
250
300
Saturation magnetization, Gauss
Cur
ie te
mpe
ratu
re, °
C
AL-400-30
Curie temperature
Mulitpacting with maximum solenoidal field
04 Apr 2014; C.Y. Tan11
Range: up to 2.5 % of nominal field level
Location of multipacting Exponential growth of particle number
Mulitpacting without solenoidal field
04 Apr 2014; C.Y. Tan12
Range: up to 1 % of nominal field level
Exponential growth of particle number
Location of multipacting
Magnetic permeability (Gyrotropic model)
04 Apr 2014; C.Y. Tan13
Measuring AL400 (R. Madrak and D. Wildman)
04 Apr 2014; C.Y. Tan14
Measured losses
04 Apr 2014; C.Y. Tan15
0 20 40 60 80 100 120-1
0
1
2
3
4
5
6
7
8 Measured Losses Vs Solenoid Bias
76 MHz
106 MHz
53 MHz
solenoid bias (A)
Loss
es (-
dB)
method looks at s11 and from there calculate the loss in the garnet.
This number will scale with the length of the garnet.
Model in ADS used to calculate μ’ from s11 phase data
04 Apr 2014; C.Y. Tan16
Fits to the s11 phase data
04 Apr 2014; C.Y. Tan17
Measured μ
04 Apr 2014; C.Y. Tan18
15 35 55 75 95 1151.001.502.002.503.003.504.004.505.005.506.00
-0.1
0.1
0.3
0.5
0.7
0.9Mu and Losses, 76 MHz
Mu Losses
solenoid bias (A)
Mu
Loss
es (-
dB)
15 35 55 75 95 1151.001.502.002.503.003.504.004.505.005.506.00
-0.1
0.1
0.3
0.5
0.7
0.9Mu and Losses, 106 MHz
Mu Losses
solenoid bias (A)
Mu
Loss
es (-
dB)
recall μe = μ’ – iμ’’.
Back of the envelope requiresμmax/μmin = (fmax/fmin)2 = (106/76) 2≈ 2.
Sims say ratio is 2.5, then if μmin=1.5, then μmax=1.5×2.5 = 3.75
μ’
prop to μ’’
3.75
24
-0.4 dB loss @ μ’=3.75
s11 = (Pin - Pref )/Pin = 1-Pref/Pin.
In dB = 10 Log10(1- Pref/Pin)≈10(-0.43 x – 0.21 x2)
Tan[δ] = μ’’/μ’=1/Q
Todo (partial list) to be discussed
• Check that solenoid can be built. PS for solenoid -- kA
• Coupler?• PA?• Dimensions OK?• Impedance of cavity? Shunt impedance > 1
MΩ?• How to incorporate μ measurements into
simulations?
04 Apr 2014; C.Y. Tan19