accelerating (s)rf cavities - cornell universityhoff/lectures/08s_688/08s_688...matthias liepe;...
TRANSCRIPT
1Matthias Liepe; 03/28/2008
Accelerating (S)RF Cavities
Matthias Liepe
2Matthias Liepe; 03/28/2008
Outline
• RF Cavity Design– Design objectives– Numerical Eigenmode solver– Design examples:
• NC vs. SC• SC cavity center cell shape• Number of cells of SC cavities• SC cavity end cell optimization
• Higher order modes– Introduction: HOMs and excitation by a beam– HOM damping schemes– HOM damping examples and results
3Matthias Liepe; 03/28/2008
RF Cavity Design
• RF Cavity Design– Design objectives– Numerical Eigenmode solver– Design examples:
• NC vs. SC• SC cavity center cell shape• Number of cells of SC cavities• SC cavity end cell optimization
4Matthias Liepe; 03/28/2008
Cavity Design Objectives
Pulsed operation
CW operation
High beam current
High beam power
Beam quality (emittance) preservation
Low beam power
Heavy beam loading
Lorentz force detuning
RF power dissipationin cavity walls
Beam stability (HOMs)
Parasitic interactions(input coupler kick, alignment)
Low Qext
Availability of high-power RF sources
High Qext,microphonics
Mechanical design:stiffness,
vibration modes,tunability,
thermal analysis
RF design:frequency & operating temperature choice,optimal gradient,
cavity shape optimization,number of cells,
cell-to-cell coupling,HOM extraction,
RF power coupling
Input coupler design
HOM damper design
Tuner design
RF controls
Cryostat design
Cavity design
Machine parameters Effects/cavity parameters
from S. Belomestnykh
5Matthias Liepe; 03/28/2008
Cavity Design Parameters
• Choice of material (impacts losses and operating gradient)• Frequency (impacts size of cavity, cost, surface resistance,
assembly technique …)• Number of cells (impacts field flatness, tunability, HOM
extraction, pulsed operation …)• Aperture size (impacts beam stability, peak fields, HOM
damping …)• Cavity shape (impacts peak surface fields at a given Eacc, power
dissipation at a given Eacc, multipacting …)• …
6Matthias Liepe; 03/28/2008
Numerical Eigenmode Solver
The modes in real cavities (with beam tubes,…) cannot be calculated analytically.
⇒⇒⇒⇒ Use numerical codes (like MAFIA, Microwave studio, SLANS, Superfish,… ).
mesh modes
TM010 modes
ππππ / 2 mode
ππππ mode
7Matthias Liepe; 03/28/2008
Tools for RF Design
8Matthias Liepe; 03/28/2008
Examples - MAFIA
MAFIA is a 3D simulation code used for the design of RF cavities and other electromagnetic structures, includingelectrostatic and magnetostatic devices. It is an acronym for the solution of MA xwell’s equations using the FiniteIntegration Algorithm. MAFIA uses a rectangular mesh generation routine which is flexible enough to model even the most complex geometries. The routine allows the user to specify the"coarseness" of the mesh in a particular area of interest.
MAFIA User Guide, The MAFIA Collaboration: DESY, LANL and KFA, May 1988.
B-cell, CESR, Cornell U.
V. Shemelin, S. Belomestnykh. Calculation of the B-cell cavity external Q with MAFIA and Microwave Studio. Workshop on high power couplers for SC accelerators. Newport News, VA, 2002.
9Matthias Liepe; 03/28/2008
Examples - Microwave Studio
Injector Cavity for ERL
The program combines both a user friendly interface (Windows based) and simulation performance.
Perfect Boundary Approximation increases the accuracy of the simulation by an order of magnitude in comparison to conventional simulators.
The software contains 4 different simulation techniques (transient solver, frequency domain solver, eigenmode solver, modal analysis solver).
CST Microwave Studio, User Guide, CST GMbH, BuedingerStr. 2a, D-64289, Darmstadt, Germany.
10Matthias Liepe; 03/28/2008
Examples-SuperLANS / CLANS
• SuperLANS (or SLANS) is a computer program designed to calculate the monopole modes of RF cavities using a finite element method of calculation and a mesh with quadrilateral biquadratic elements.
• SLANS has the ability to calculate the mode frequency, quality factor, stored energy, transit time factor, effective impedance, max electric and magnetic field, acceleration, acceleration rate, average field on the axis, force lines for a given mode, and surface fields.
• Later versions, SLANS2 and CLANS2, calculate azimuthally asymmetric modes, and CLANS and CLANS2 can include into geometry lossy dielectrics and ferromagnetics.
11Matthias Liepe; 03/28/2008
Design Example 1: Normal conducting vs. Superconducting RF
Sacc
S
sdiss RGQR
VdsHRP
⋅⋅⋅⋅======== ∫∫∫∫ /2
1 22rPower dissipated into the wall:
Example:Example:• Accelerating voltage: let’s take only 1 MV• Constant R/Q (depends on cell shape): 1000 ΩΩΩΩ• Geometry constant: 270 ΩΩΩΩ• Surface resistance: Rs,copper= 10 mΩΩΩΩ Rs,Nb= 10 nΩΩΩΩ
⇒ Pdiss,copper= 37 kW Pdiss,Nb = 37 mW
⇒ Copper is not the best choice for a ILC shape cavity…
12Matthias Liepe; 03/28/2008
Minimizing Losses
Sacc
S
sdiss RGQR
VdsHRP
⋅⋅⋅⋅======== ∫∫∫∫ /2
1 22r
Depend only on cavity geometry.Depend only on cavity geometry.⇒⇒⇒⇒⇒⇒⇒⇒ Maximize for copper cavities.Maximize for copper cavities.
Minimize surface resistance!Minimize surface resistance!⇒⇒ Superconducting cavities.Superconducting cavities.
mmΩΩΩΩΩΩΩΩ (copper) (copper) ⇒⇒⇒⇒⇒⇒⇒⇒ nnΩΩΩΩΩΩΩΩ
Less important for SRF cavities!
13Matthias Liepe; 03/28/2008
RF Cavities for Linacs
• TESLA• superconducting cavity• niobium • 1.3 GHz• 2 K (LHe)
• one cell from NLC• normal conducting cavity• copper• 11.4 GHz• water cooled
Fundamental differences due to difference in wall losses.
14Matthias Liepe; 03/28/2008
Superconducting Cavities: Advantages
Can operate at a higher voltage in cw operation orlong pulse operation because of low losses.
Power consumption is less. ⇒⇒⇒⇒ Operating costsavings, better conversion of ac power to beampower.
Power dissipation is not the primary concern! Can tailor design to a given accelerator application.
15Matthias Liepe; 03/28/2008
Superconducting Cavities: Advantages (cont.)
Freedom to adapt design better to the accelerator requirements allows, for example, the beam-tube and the cell iris size to be increased:
• Reduces the interaction of the beam with the cavity.(scales as iris radius2 to 3) ⇒⇒⇒⇒ The beam quality is better preserved.Important for, e.g., FELs.
• HOMs are removed more easily. Better beam stability.⇒⇒⇒⇒More current accelerated. Important for, e.g., B-factories.
• Reduce the amount of beam scraping.⇒⇒⇒⇒ Less activation in,e.g., proton machines.Important for, e.g., SNS, Neutrino factory.
• Allows more coupling between cells in multicell structures.⇒⇒⇒⇒ Better energy exchange between cells.Important for e.g., high-energy machines.
16Matthias Liepe; 03/28/2008
Design Example 2: Center Cell Shape of an SRF Cavit y
- 16 -
The cell length(L) determines the cavity geometrical beta value.
The cell iris radius (Riris) is mainly determined by the cell-to-cell coupling requirements and cavity impedance limitations.
The iris ellipse ratio (r=b/a) is primary determined by the local optimization of the peak electric field.
The cell radius (D) is used for the frequency tuning without modifying any electromagnetic or mechanical cavity parameter.
The side wall inclination (αααα) and B and A can be use to minimize relative surface peak fields.
Usual design goal: Maximize R/Q*G / Usual design goal: Maximize R/Q*G / minimize peak magnetic surface field for minimize peak magnetic surface field for given wall angle, maximum peak electric given wall angle, maximum peak electric field and minimum iris radiusfield and minimum iris radius
17Matthias Liepe; 03/28/2008
JLab’s optimized shape was designed under restriction that the angle of the wall slope is not less than 8 deg. This angle is useful to let liquid easily flow from the surface when chemical treatment or rinsing are performed.
This shape is also more mechanically strong.
If one rejects the limitation of the angle we can further improve the value of G*R/Q
JLab
LL Cell
from V. Shemelin
G*R/Q, relative units:
0.90 (75deg) 1.00 (82deg) 1.02 (90deg) 1.042 (105 deg)
18Matthias Liepe; 03/28/2008
Cavity Cell Shape and R/Q*G and Loss Factor
Comparison of 1-Cell Geometries
25
35
45
55
65
75
85
95
105
115
0 50 100
Z (mm)
R (
mm
)
30mm_10%
30mm_20%
35mm_10%
35mm_20%
39mm_10%
39mm_20%
43mm_10%
43mm_20%
Baseline
1.3 GHz center-cell:
• Cells optimized for fixed side
wall angle (82 deg) and electric
peak field (E/Eacc=2.2)
3 3.5 4 4.5-12
-10
-8
-6
iris radius [cm]long
. los
s fa
ctor
[V
/pC
]
3 3.5 4 4.51.2
1.4
1.6
1.8x 10
4
iris radius [cm]
R/Q
*G [
ΩΩ ΩΩ2]
3 3.5 4 4.511.5
12
12.5
13co
olin
g po
wer
[kW
]
iris radius [cm]
Total cooling power (fund. mode + HOM at 80K)16 MV/m, 2*100 mA
7-cell cavity
For single cell
TTF shape
19Matthias Liepe; 03/28/2008
Wall Angle, Peak Field and R/Q*G
from V. Shemelin
20Matthias Liepe; 03/28/2008
Impact of Cell Shape on HOM Impedances (I)
from Bob Rimmer
21Matthias Liepe; 03/28/2008
Impact of Cell Shape on HOM Impedances (II)
from Bob Rimmer
22Matthias Liepe; 03/28/2008
Design Example 3: Number of Cells per SC Cavity
• Risk of trapped modes increases with number of cells5 6 7 8 990
95
100
105
110
115
cells per cavityrela
tive
cons
truc
tion
cost
[%
]
F. Marhauser at al. PAC 1999
1500 2000 2500 3000 3500 4000
102
104
106
Q
frequency [MHz]
7 cell8 cell9 cell
Monopoles
23Matthias Liepe; 03/28/2008
Trapped Modes
24Matthias Liepe; 03/28/2008
HOM strength (number of Cells)
from Bob Rimmer
25Matthias Liepe; 03/28/2008
Design Example 4: 7-Cell SRF Cavity End-Cell Design
• End cell shape has significant impact (example HOM) :
• Can fine-tuning of end cell to– Increase damping of most dangerous HOM(s)– Avoid strong monopole modes at beam harmonics– Note: Can optimize end cell shape only for a few selected modes!!
changed end-cell shape
26Matthias Liepe; 03/28/2008
End Cell Optimization: ILC Cavity
27Matthias Liepe; 03/28/2008
7-Cell SRF Cavity End-Cell Design
from V. Shemelin
28Matthias Liepe; 03/28/2008
Cavity Examples: SRF High beta Cavities
29Matthias Liepe; 03/28/2008
RF Cavity Design
• Higher order modes– Introduction: HOMs– HOM excitation by a beam– HOM damping schemes– HOM damping examples and results
30Matthias Liepe; 03/28/2008
HOM Excitation by a Bunch
time
The bunched beam excites higher-order-modes (HOMs)= wakefields = electromagnetic fields in the cavity.
bunch
bunch
bunch
31Matthias Liepe; 03/28/2008
Beam-Cavity Interaction
• Bunch traverses a cavity• ⇒⇒⇒⇒ deposits electromagnetic
energy, which is described as wakefields (time domain) or higher-order modes (HOMs, frequency domain)
• Subsequent bunches are affected by these fields and at high beam current one must consider instabilities
from S. Belomestnykh
32Matthias Liepe; 03/28/2008
Single Bunch Monopole Losses:Wake Potential of a Point-Charge
The beam excites higher-order-modes (HOMs) in a cavity:
bunch
• When a charge passes through a cavity, it excites HOMs.
• If it passes exactly an axis, it will only excite monopole modes.
• For a point charge, the HOM excitation depends only on the bunch charge and the cavity shape.
• The excited field can be described by the wake potential.
33Matthias Liepe; 03/28/2008
Higher-Order-Modes (HOMs)
transverse kicktransverse kick--fieldsfields
Dipole modes, quadrupole modes,Dipole modes, quadrupole modes,……
Monopole modesMonopole modes
electric field magnetic field
electric field
z
longitudinal electric longitudinal electric field on axisfield on axis
34Matthias Liepe; 03/28/2008
Monopole, Dipole and Quadrupole Modes…
Monopole Dipole
Quadrupole
from R. Wanzenberg
35Matthias Liepe; 03/28/2008
Methods of HOM Calculations: Frequency Domain
36Matthias Liepe; 03/28/2008
Time-Domain Method (I)
37Matthias Liepe; 03/28/2008
Time-Domain Method (II)
38Matthias Liepe; 03/28/2008
RF Cavity Design
• Higher order modes– Introduction: HOMs– HOM excitation by a beam– HOM damping schemes– HOM damping examples and results
39Matthias Liepe; 03/28/2008
HOM Excitation
The excited HOM power of a single bunch depends on:
the HOMs of the cavity (i.e. their shunt impedance),
the bunch charge (PHOM ∝∝∝∝qb2),
the bunch length (i.e. the spectrum of a bunch).
0 20 40 60 80 1000
20
40
60
80
100
HOM frequency [GHz]
% o
f HO
M p
ower
loss
abo
ve f
HO
M
Example:
σσσσbunch = 0.6 mm ⇒⇒⇒⇒ Short bunches excite very high frequency modes!
40Matthias Liepe; 03/28/2008
Single Bunch Monopole Losses: The Bunch
-5 -4 -3 -2 -1 0 1 2 3 4 50
200
400
600
800
1000
z [mm]
char
ge d
istr
ibut
ion
[arb
. uni
ts]
Longitudinal charge distribution for a 600 µµµµm bunch:
41Matthias Liepe; 03/28/2008
Single Bunch Monopole Losses:The Bunch Spectrum
0 200 400 6000
0.2
0.4
0.6
0.8
1
f [GHz]
bunc
h sp
ectr
um
Spectrum of a 600 µµµµm bunch:
42Matthias Liepe; 03/28/2008
Beam-cavity interaction: Wave Function
Lorentz-Forces on test charge:
The integrated field seen by a test particle travel ing on the same path at a constant distance s behind a point charge q is the longitudinal wake (Green) function w(s).
43Matthias Liepe; 03/28/2008
Single Bunch Monopole Losses:Wake Function of a Point Charge after a TESLA Cavit y
-2 0 2 4 6 8-5
-4
-3
-2
-1
0x 10
wak
e fu
nctio
n [V
/pC
]
distance s [mm]
44Matthias Liepe; 03/28/2008
Single Bunch Monopole Losses:Wake Potential of a Point Charge after a TESLA Cavi ty
The fft of the wake function gives the cavity impedance Z(ωωωω):
1 10 100 100010 -1
10 0
10 1
10 2
10 3
real
par
t of Z
(ωω ωω
) fo
r a
TE
SLA
cav
ity [
ΩΩ ΩΩ]
f [GHz]
45Matthias Liepe; 03/28/2008
Single Bunch Monopole Losses:Wake Potential of a Bunch after a TESLA Cavity
-5 -4 -3 -2 -1 0 1 2 3 4 5-3
-2.5
-2
-1.5
-1
-0.5
0x 10 13
z [mm]
wak
e po
tent
ial [
V/C
]
bunch
sdsswsqsWs
′′−′= ∫∞−
)()()(
The wake potential W is a convolution of the linear bunch charge density distribution q(s) and the wake function w
46Matthias Liepe; 03/28/2008
Single Bunch Monopole Losses:Loss Factor
∫∫∫∫∞∞∞∞
∞∞∞∞−−−−==== dssWsqk )()(||
beambunchIQkP |||| =
This is the total energy lost by a bunch divided by the time separation of two consecutive bunches.
This does not include any interaction between bunches (i.e. resonant mode excitation)!!!
Once the longitudinal wake potential is known, the longitudinal loss factor , which tells us how much electromagnetic energy a bunch leaves behind in a structure can be defined as:
2q
Uk
∆=
AverageAverage power loss:
47Matthias Liepe; 03/28/2008
Single Bunch Monopole Losses:HOM Power Frequency Distribution
0 100 200 3000
50
100
150
200
250
f [GHz]
pow
er [W
]
integrated power up to frequency f
Most of the HOM powerMost of the HOM power
is well below 100 GHz.is well below 100 GHz.
The frequency distribution of the HOM losses is determined by the bunch spectrum and the cavity impedance Z(ωωωω):
[[[[ ]]]]2)(~)()( ωωωωωωωωωωωω qZP ∝∝∝∝
48Matthias Liepe; 03/28/2008 48
High current and short bunches
0 0.5 1 1.5 2 2.5
x 104
0
50
100
150
200
250
300
350
beam
cur
rent
[mA
]
1/(bunch length) [m-1 ]
CESR
JLAB FEL CEBAFTTF II
TTF II
ERLs:Storage ring currents and linac bunch length⇒ Significant HOM power up to 100 GHz!⇒ Where does the high frequency power go?
49Matthias Liepe; 03/28/2008
50Matthias Liepe; 03/28/2008
Bunch Trains:The more Complex Picture
Tb
The HOMs excited by a bunch are decaying due to losses,
but: still significant field present in the cavity when the next bunch enters the cavity!
⇒⇒⇒⇒ Resonant excitation of a HOM, if
bHOM T
Nf1≈
51Matthias Liepe; 03/28/2008
HOM Excitation
The excited HOM power of a bunch train depends on:
the HOM losses of a single bunch,
the beam harmonic frequencies and the HOM frequencies (resonant excitation is possible!),
the bunch charge and the beam current (PHOM ∝∝∝∝QI),
and the external quality factor, Qext of the modes. Lower Qext means less energy deposited by the beam:
PHOM ∝∝∝∝ Qext
52Matthias Liepe; 03/28/2008
Bunch Trains:The more Complex Picture
In averagethe total HOM losses per cavity are given by the single bunch losses (77 pC bunch charge, 2.6 GHz
bunch repetition rate, σσσσb= 600 µµµµm):
W160 A0.2 77pC V/pC4.10|||| ====⋅⋅⋅⋅⋅⋅⋅⋅======== beambunchIQkP
But: If a monopole mode is excited on resonance, the loss for this mode can be much higher:
2beamQI
QR
P
====
⇒ To stay below 200 W200 W: • achieve (R/Q)Q < 5000(R/Q)Q < 5000, • or avoid resonant excitation of the mode.
53Matthias Liepe; 03/28/2008
Bunch Trains:The more Complex Picture
Example: Cornell ERL:
injector the in GHz 3.1⋅⋅⋅⋅==== Nf HOMlinac main the in GHz 6.2⋅⋅⋅⋅==== Nf HOM
… so most of the monopole modes in the ERL will not be excited resonantly.
1000 2000 3000 4000 5000 6000 70000
0.5
1
1.5
2
f [MHz]
modes in a 9-cell cavity
beam harmonics
…up to xx GHz
54Matthias Liepe; 03/28/2008
Bunch Trains:HOM Frequencies Spread
Can one design the HOM frequencies such, that non of the modes are excited resonantly?
The higher the frequency, the more sensitive is the frequency of a HOM to small perturbations in the cavity shape:
How large is “const”? Example: 2.4 GHz modes at TTF
Simple approximation: .constff
HOM
HOM ====∆∆∆∆
Mode #
σf[H
z]
σf = 10 MHz⇒ const = 0.4 %
i.e. σf = 20 MHz at 5.2 GHzσf = 31 MHz at 7.8 GHzσf = 42 MHz at 10.4 GHz
…
16 MHz
2 MHz
55Matthias Liepe; 03/28/2008
Bunch Trains:A Simple Model: 10000 Monopoles with random f’s
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 100000
20
40
60fr
eque
ncy
[GH
z]
mode #
0 10 20 30 40 50 600
10
20
30
40
50
frequency [GHz]
# of
mod
es
frequency of the modes
frequency distribution
56Matthias Liepe; 03/28/2008
Bunch Trains:A Simple Model: 10000 Monopoles with random f’s
0 10 20 30 40 50 6010
0
105
1010
frequency [GHz]
loss
fac
tor
k [
V/(
As)
]
0 10 20 30 40 50 600
50
100
150
frequency [GHz]
inte
grat
ed p
ower
[W
]
assumed loss factor of the modes
calculated single bunch losses
57Matthias Liepe; 03/28/2008U N I V E R S I T YCORNELL
Bunch Trains:A Simple Model: 10000 Monopoles with random f’s
0 10 20 30 40 50 600
1
2
3
4
5
6
7
8
pow
er [W
]
frequency [GHz]
Example: all HOMs have Q = 1000,one set of frequencies
58Matthias Liepe; 03/28/2008
A Simple Model: 1000 Monopoles with random f’sTotal HOM Monopole Power for random Sets of Frequen cies
95 100 105 110 115 120 125 130 135 140 1450
50
100
#
QL=100
0 50 100 150 200 250 3000
50
100
#
QL=1000
0 200 400 600 800 1000 1200 14000
200
400
#
QL=10 4
0 2000 4000 6000 8000 10000 120000
2000
4000
6000
#
power [W]
QL=10 5
All HOMs have Q = 100
All HOMs have Q = 1000
All HOMs have Q = 104
All HOMs have Q = 105
59Matthias Liepe; 03/28/2008
Higher-Order-Modes (HOMs)
Parasitic modes excited by the accelerated beam may lead to:
degradation of the beam quality (transverse emittancegrowth due to dipole modes, BBU, energy spread),
additional cryo-losses (wall losses, heating ofcables and feedthroughs), mostly due to monopole modes.
⇒ Requirements on the Requirements on the external quality factor,external quality factor,QQext ext of the modes.of the modes.
Without additional damping the HOMs can haveWithout additional damping the HOMs can havevery high quality factors (Q>10very high quality factors (Q>101010)! )!
60Matthias Liepe; 03/28/2008
RF Cavity Design
• Higher order modes– Introduction: HOMs– HOM excitation by a beam– HOM damping schemes– HOM damping examples and results
61Matthias Liepe; 03/28/2008
Solution (for SC Cavities):HOM Couplers and Absorbers
The parasitic e-m fields can be kept below the threshold by means of HOM couplers and HOM
absorbers, usually attached to the beam tubes of a s.c. cavity.
HOM coupler
HOM couplerHOM absorber
62Matthias Liepe; 03/28/2008
Higher-Order-Mode Couplers and Absorbers
f/GHz1 10 100
trapped and quasi trapped modes
propagating modes
HOM couplers HOM beam-pipe absorber
to room temperatureload
facc
absorber between cavitiesat temperature level with good
cryogenic-efficiencyThe frequency where the HOMs start to propagate
depends on the beam tube diameter: ωωωωc ∝∝∝∝ 1/diameter!
waveguide couplersantenna couplers
63Matthias Liepe; 03/28/2008
HOM Extraction/Damping Schemes
Several approaches are used:
• Loop couplers (several per cavity for different modes/orientations)
• Waveguide dampers
• Beam pipe absorbers (ferrite or ceramic)
JLab proposalBNL ERL cryomodule B-cell beam line components (TLS)
TESLA cavity loop coupler
64Matthias Liepe; 03/28/2008
Broadband Beam Pipe RF Absorber
propagating modes
• High frequency modes propagate out the beam pipe.
• RF absorbing material can damp these modes.
• Dissipated power will be intercepted by cooling (water, GHe, LN2).
• Candidate absorber materials: ferrites (used in CESR HOM load) Zr10CB5 CERADYNE (used for
CEBAF HOM load) Mo in AL2O3 …facc is below the cut-
off frequency of the tube
65Matthias Liepe; 03/28/2008- 65 -
Broadband RF Absorber
Fundamental mode: f=1.300 GHz ferrite absorber
16.5 cm
Example: dipole mode: f=3.9 GHz
•Low field at absorber ⇒⇒⇒⇒ no significant damping of the fundamental mode
but:• Propagating modes have higher fields at the absorber
⇒⇒⇒⇒ damping and power extraction!
ferrite absorber
Q < 104
Q > 1010
66Matthias Liepe; 03/28/2008
Higher-Order-Mode Couplers
Coaxial Coupler Waveguide Coupler
Rejection filter suppresses coupling to the accelerating mode.
Waveguide cutoff suppresses coupling to the accelerating mode.
HOM out
HOM out
67Matthias Liepe; 03/28/2008
TTF HOM Loop Coupler (1)
HOM coupler at each side of the cavity close to end cell
to damp HOMs
68Matthias Liepe; 03/28/2008
HOM Loop-Couplers
Coupler model: superconductingpick-up antennasuperconducting
pick-up loop
capacitivecoupling
output toroom temp.
load
capacitor of the1.3 GHz notch
filter
Important to reduce Q of non-propagating dipole modes. Can only handle a few 10 W.
Will work up to a few GHz but not above. Cooling / heating from fundamental mode issue in cwcavity
operation.
69Matthias Liepe; 03/28/2008
Methods for HOM Damping
from Bob Rimmer
70Matthias Liepe; 03/28/2008
Methods for HOM Damping: Effectiveness
from Bob Rimmer
71Matthias Liepe; 03/28/2008
HOMs
• Higher order modes– Introduction: HOMs– HOM excitation by a beam– HOM damping schemes– HOM damping examples and results
72Matthias Liepe; 03/28/2008
Example 1: CESR HOM Ferrite Absorber (1)
Flute beam pipe ⇒⇒⇒⇒ guide out the first two defecting modes.
Total HOM power: several kW!
Qext < 103
73Matthias Liepe; 03/28/2008
CESR HOM Ferrite Absorber (2)
ferrite
water cooling
74Matthias Liepe; 03/28/2008
CESR HOM Ferrite Absorber (3)
Use a single cell (no reflection by irises between cells) Open beam tubes so that all modes propagate out the beam tubes!
Use material with very high RF losses.
75Matthias Liepe; 03/28/2008
Example 2: The Cornell ERL Injector
beam
4.34 m
ferrite #1
ferrite #4ferrite #5ferrite #3
ferrite #2ferrite #6
HOM damping concept: Make all TM monopole and all dipole modes propagating by increasing the beam tube diameter (as in
CESR).
76Matthias Liepe; 03/28/2008
Cornell ERL Beam Line HOM Loads
Flange to Cavity
Flange to Cavity
RF Absorbing
Tiles
Cooling Channel
(GHe)Shielded Bellow
TT2, Co2Z, CeralloyRF absorbing tiles
He GasCoolant
80 KOperating temp.
1.4 – 100 GHzHOM frequencies
26 W (200 W max)Power per load
77Matthias Liepe; 03/28/2008
Cornell ERL Beam Line HOM Loads: Damping Calculations
1000 1500 2000 2500 3000 3500 4000 450010
-2
100
102
104
106
108
1010
1012
f [MHz]
Qfe
rrite
accelerating mode, almost undamped, as it should be
strongly damped HOMs
78Matthias Liepe; 03/28/2008
ERL Main Linac HOM Damping Simulations
• CLANS calculations (started 3D Microwave Studio mod els)• Modes are sufficiently damped for 100 mA operation
1000 1500 2000 2500 3000 3500 400010-2
100
102
104
106
frequency [MHz]
R/Q
*Q/2
f [O
hm/c
m2 /
GH
z]
2400 2600 2800
102
104
106
Q
frequency [MHz]
MonopolesDipoles
5000 5200 5400
102
104
106
Q
frequency [MHz]
≈ Factor 5 below BBU limit
7-cell TTF shape
7-cell low loss
79Matthias Liepe; 03/28/2008
Example 4: ILC Cavity with HOM Loop Couplers
1.E+03
1.E+04
1.E+05
1.E+06
1 2 3 4 5 6 7 8 9
#D41 #S32#S29 not measured #S30#D39 #D40#S28 #D42
TM 011 mode #
Q
measured by G. Kreps, DESY
TTF module #3
80Matthias Liepe; 03/28/2008
TTF HOM Coupler:Measured Damping of Dipole Modes
1.0E+03
1.0E+04
1.0E+05
1.0E+06
1 2 3 4 5 6 7 8 9
#D41 #S32 #S29 #S30
#D39 #D40 #S28 #D42
TE111 mode #
Q
1b 2b 3b 4b 5b 6b 7b 8b 9b
1.0E+03
1.0E+04
1.0E+05
1.0E+06
1.0E+07
Q
1 2 3 4 5 6 7 8 91b 2b 3b 4b 5b 6b 7b 8b 9bTM 110 mode # measured by G. Kreps, DESY
low R/Q
high R/Q modes
high R/Q modes
low R/Q
81Matthias Liepe; 03/28/2008
0.00E+00
1.00E+09
2.00E+09
3.00E+09
4.00E+09
5.00E+09
6.00E+09
7.00E+09
8.00E+09
9.00E+09
1.00E+10
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Series1
Series2
Series3
Series4
Series5
Series6
Series7
Series8
9-cell CavitiesTE111 Dipole Modes: TTF Module 2
(R/Q)Qf [ΩΩΩΩMHz]
Mode #
82Matthias Liepe; 03/28/2008
0.00E+00
1.00E+09
2.00E+09
3.00E+09
4.00E+09
5.00E+09
6.00E+09
7.00E+09
8.00E+09
9.00E+09
1.00E+10
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Series1Series2Series3Series4Series5Series6Series7
Series8
9-cell CavitiesTE111 Dipole Modes: TTF Module 3
(R/Q)Qf [ΩΩΩΩMHz]
Mode #
83Matthias Liepe; 03/28/2008
9-cell CavitiesTM110 Dipole Modes: TTF Module 2
0.00E+00
1.00E+09
2.00E+09
3.00E+09
4.00E+09
5.00E+09
6.00E+09
7.00E+09
8.00E+09
9.00E+09
1.00E+10
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Series1
Series2
Series3
Series4
Series5
Series6
Series7
Series8
(R/Q)Qf [ΩΩΩΩMHz]
Mode #
84Matthias Liepe; 03/28/2008
0.00E+00
2.00E+09
4.00E+09
6.00E+09
8.00E+09
1.00E+10
1.20E+10
1.40E+10
1.60E+10
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Series1Series2
Series3Series4
Series5Series6
Series7Series8
9-cell CavitiesTM110 Dipole Modes: TTF Module 3
(R/Q)Qf [ΩΩΩΩMHz]
Mode #
85Matthias Liepe; 03/28/2008
Experimental Setup for Beam Based HOM Measurements at TTF/FLASH
86Matthias Liepe; 03/28/2008
Beam Based HOM Measurements at TTF/FLASH
87Matthias Liepe; 03/28/2008
Example 5: BNL ERL Cavity
88Matthias Liepe; 03/28/2008
Example 6: TJNAF 1A Cryomodule Design
89Matthias Liepe; 03/28/2008
TJNAF 1A Cryomodule Design