Date post: | 21-Jan-2016 |
Category: |
Documents |
Upload: | jared-sharp |
View: | 216 times |
Download: | 0 times |
1Matthias Liepe 08/02/2007
ERL Main Linac: ERL Main Linac: Overview, Parameters
Cavity and HOM Damping
Matthias Liepe
2Matthias Liepe 08/02/2007
Outline• Overview
– Layout– Parameters and optimization
•Gradient•Temperature
• Cavity and HOM damping– Overall concept– Cavity design– HOM beam line absorber
• Test and R&D plans
3Matthias Liepe 08/02/2007
LayoutLayoutParameters and OptimizationParameters and Optimization
4Matthias Liepe 08/02/2007
Linac Layout• Main Linac Tunnel length: 319 m• Cryo-module: 6 SRF cavities + 1 magnet package• CW cavity operation!• Linacs are in 2+2 sections to limit cryo-load per
linac
Linac A1 = 16 modules
Linac B2 = 16 modulesLinac B1 = 16 modules
Linac A2 = 16 modules
= cold-warm transition (0.25 m)
Cryo-connection (4 m warm section)
319 m
5Matthias Liepe 08/02/2007
Critical Parameters / Objectives
Parameter Cornell ERL
XFEL consequence
operation mode cwcw pulsed 250 * 2K load per cavity,
factor 3 larger total 2K load
linac energy gain
5 GeV 20 GeV
average current 0.1 A* 2 3· 10-5 A (IERL/IXFEL)2=4· 107
(PHOM,ERL/PHOM,XFEL)=400
bunch charge 77 pC 1 nC
bunch length 2 ps 80 fs - 1 ps f < 100 GHz for HOMs
emittance (norm.)
0.3 mrad· mm
1.4 mrad· mm
coupler ports
energy spread (rms)
2e-4 1.25e-4 Similar, but much higher beam current,
QL!
• Accelerate / decelerate 100 mA beam to 5 GeV• Minimize emittance growth• Low trip rate• Minimize cost (construction and operation)
6Matthias Liepe 08/02/2007
Objectives and Challenges
– Operate SRF cavities CWCW •Very reliable operation essential •Avoiding excessive cryogenic loads •Minimize RF drive power
– Accelerate a high beam currenthigh beam current•Avoiding beam instability and excessive
HOM losses•Dispose high HOM power safely
– Preserve beam emittance•Small wake fields•Good cavity alignment•Small transverse kick fields from beam
pipe asymmetries (input couplers, …)
7Matthias Liepe 08/02/2007
Cryomodules
• 6 SRF cavities per module (tentative)
• Modules connect directly (no cold-warm transitions)
• All cryogenic piping is inside of the modules
• Details: Eric’s talk…
8Matthias Liepe 08/02/2007
ERL Main Linac: Technical Parameters I
Parameter Cornell ERL Comments / justification
Total linac length 629 m for 5 GeV; fill factor lower due to larger number of
magnets and longer, larger diameter beam tubes for HOM damping by beam
pipe absorbers
Module length 9.8 m
SRF cavities per module
6 (tentative)
Total number of cavities
384
Geometric fill factor 49 %• Note: A very optimistic 65% fill factor (with XFEL type cavity beam tubes and a different HOM damping scheme) would reduce the total SRF linac cost (modules, RF, cryo, tunnel) by 2 % (assuming same cost for HOM (assuming same cost for HOM damping and same Qdamping and same Q00).).
9Matthias Liepe 08/02/2007
ERL Main Linac: Technical Parameters II
Parameter Cornell ERL Comments / justification
Cavity frequency 1.3 GHz ILC technology available
Cells per cavity 7Strong HOM damping; risk of
trapped modesActive cavity length
0.8 m
Impedance per cavity (circuit definition)
400 Ohm Iris radius optimized; trade-off between HOM losses and fundamental mode losses
Cavity Loss Factor
10 V/pC
E_peak/E_acc < 2.2 upper limit to reduce field emission
Average acc. gradient
16.2 MV/m optimization (cost, field emission)
unloaded Q0 > 21010 cost of cryogenic plant (largest contributor)
loaded Q 6.5e7 (2107 - 1108) optimized for 20 Hz peak
detuning and 10 Hz typical detuningCavity full
bandwidth20 Hz
Peak detuning < 20 Hz
Cavity offset tolerance
1 mm Similar to ILC
Cavity angle tolerance
1 mrad
Operating temp. 1.8 K Optimization
10Matthias Liepe 08/02/2007
ERL Main Linac: Technical Parameters III
Parameter Cornell ERL
Comments / justification
Average HOM power per cavity
154 W Overhead for resonance excitation of modes, dipole
lossesMax. HOM power per cavity
300 W
Average 1.8K Static load/Cavity
0.5 W
Average 1.8K Dyn. load/Cavity
10.5 W for Q0 = 2· 1010
Total 1.8 K static load 0.2 kW
Total 1.8 K dynamic load 4 kW for Q0 = 2· 1010
Total 5 K static load 2.3 kW
Total 5 K dynamic load 3.1 kW dominated by HOM losses; assumes that 5% of HOM
power goes to 5K
Total 80 K static load 5 kW
Total 80K dynamic load 60 kW dominated by HOM losses
11Matthias Liepe 08/02/2007
ERL Main Linac: Technical Parameters IV
Parameter Cornell ERL
Comments / justification
Ave RF Power/Cavity 2 kW for 20 Hz peak detuning and 10 Hz typical detuningPeak RF Power/Cavity 5 kW
Number of Cavities/RF Unit
1 vector sum control difficult
Bunch to bunch energy fluctuation
2· 10-4
Stability requirements similar to XFEL, but have to achieve this at much higher
QL and with higher beam currents; see talk on LLRF
RMS field ampl. stab. uncorrelated
5· 10-4
RMS field ampl. stab. correlated
1· 10-4
RMS field phase stab. uncorrelated
0.15 deg
RMS field phase stab. correlated
0.02 deg
12Matthias Liepe 08/02/2007
Optimal Operating Temperature
for 10 n residual resistance
1.8K (25% reduced AC power as compared to 2K)
Note: T<1.8K is might cause instability in the cryo-system.
temperature [K]1.4 1.6 1.8 2 2.2 2.4
0
1
2
3
4x 10
4
Cry
o A
C p
ow
er
/ ac
tive
len
gth
[a
rb. u
nit
s]
750 MHz1300 MHz1500 MHz
Bernd Petersen DESY
13Matthias Liepe 08/02/2007
Cavity Operation at 1.8KERL2005: Bernd Petersen, DESY
• Lowering the temperature seems to be effective as long as Q = Q(T) follows BCS and the temperature dependent dynamic loads dominate (reasonable lower limit 1.5 K)
• HeII cooling might become unstable below 1.8 K – tests required
• Another cold compressor stage is required for each 0.2 K temperature step to lower temperatures – investment costs and system complexity increase
• In view of pressure drops, critical gas velocities, work of compression and general sizing the lower gas densities at lower temperatures seem to be balanced by the lower cooling loads and the related lower mass flows
14Matthias Liepe 08/02/2007
Optimal Field Gradient I
10 15 20 250
200
400
600total cost
cost
[ar
b. U
nit
s]
field gradient [MV/m]
capitaloperationtotal
10 15 20 250
50
100
150capital cost
cost
arb
. un
its]
field gradient [MV/m]
tunnellinacRFcryo
10 15 20 250
50
10010 year operating cost
cost
[ar
b. u
nit
s]
field gradient [MV/m]
RFcryo
10 15 20 25200
300
400
500
600tunnel length
len
gth
[m
]
field gradient [MV/m]10 15 20 25
200
400
600
800number of cavities
#
field gradient [MV/m]10 15 20 25
109
1010
1011
cavity Q0
Q0
field gradient [MV/m]
10 15 20 250
5
10
15IOT peak power
po
wer
[kW
]
field gradient [MV/m]10 15 20 25
0
10
20
30cryo AC power
po
wer
[M
W]
field gradient [MV/m]10 15 20 25
0
5
10cryo power fractions
po
wer
[M
W]
field gradient [MV/m]
cav. dyn.HOMinput Cstatic
15Matthias Liepe 08/02/2007
Main Linac Cost Distribution for E=16.2 MV/m
• Cryogenic plant and module costs dominate
Tunnel RF system Cryomodules Cryogenic plant0
10
20
30
40
50
rela
tive
co
st [
%]
16Matthias Liepe 08/02/2007
Optimal Field Gradient II
• Q0-value has significant impact on cost (high impact parameter)
• Construction cost changes only moderately for gradients between 16 and 23 MV/m
• Operating cost / AC power increases with gradient• Select gradient at lower end: 16.2 MV/m 16.2 MV/m
10 15 20 250
100
200
300co
st [
arb
. un
its]
field gradient [MV/m]
10 15 20 2510
9
1010
1011
Q0
field gradient [MV/m]
construction10 years operation
case 1case 2
Less risk for same cost!Less risk for same cost!
17Matthias Liepe 08/02/2007
Field EmissionGamma radiation measured at DESY/FLASH from cavity field
emission:
• Exponential growth in FE with gradient
• Serious problem in cw cavity operationcw cavity operation
• Low trip rate Low trip rate essential for light essential for light source!source!
• Favors lower gradients
• High reliability: don’t push gradient and RF power to limit
16.2 MV/m16.2 MV/mFor ERL : 10Gy/h * 200 (for cw)= 2 mGy/h = 0.2 rad/h2 mGy/h = 0.2 rad/h
10 years of operation: 100 Gy = 10,000 rad100 Gy = 10,000 rad (at 5000h/year)
18Matthias Liepe 08/02/2007
Cavity Performance Goals
– For gradient overhead: Require average cavity performance in linac: 18 MV/m at Q = 218 MV/m at Q = 2101010 10 with with 2 2 MV/m spreadMV/m spread
Min. cavity performance in linac: 16 MV/m at Q = 21010
– Average operating gradient: 16.2 MV/m16.2 MV/m 384 cavities!
– This gives This gives 12.5 % overhead12.5 % overhead for initial performance for initial performance risks and failures (tuner, IOT, power supply…) risks and failures (tuner, IOT, power supply…)
– Individual cavities can operate at gradients up to 20 MV/m
• Cryogenic system can support 20 MV/m at Q = 11010 for individual cavities
• RF power sufficient for 20 MV/m with <20 Hz peak detuning
19Matthias Liepe 08/02/2007
Cavity and HOM dampingCavity and HOM damping
20Matthias Liepe 08/02/2007
Overall Concept
• 7-cell, 1.3 GHz SRF cavity
• HOM damping via beam line absorbers– Relative simple and quite effective concept
– Avoids kicks from beam line asymmetries
– Deals with high HOM power
– Works well at high frequencies
– Supports high Q0 operation
• Fill factor is not a strong cost driver
1.8K 80K80K1.8K 1.8K
21Matthias Liepe 08/02/2007
Design Approach• Center cell shape optimized for low cryo-
losses
• Optimize mechanical design for low microphonics
• Input coupler with opposite stub to minimize transverse kick fields
• End cells and tubes optimized for good HOM power extraction
• All higher-order monopole, and dipole modes propagate in beam tube
• Cold beamline absorbers between cavities
22Matthias Liepe 08/02/2007
Cavity Cell Shape and R/Q*G
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)
•Selected iris radius = 35 mm
3 3.5 4 4.5-12
-10
-8
-6
iris radius [cm]lo
ng
. lo
ss f
acto
r [
V/p
C] 3 3.5 4 4.5
1.2
1.4
1.6
1.8x 10
4
iris radius [cm]
R/Q
*G [
2
]
3 3.5 4 4.511.5
12
12.5
13co
olin
g p
ow
er [
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
23Matthias Liepe 08/02/2007
Number of Cells per Cavity
• Risk of trapped modes increases with number of cells
5 6 7 8 990
95
100
105
110
115
cells per cavityrela
tive
co
nst
ruct
ion
co
st [
%]
F. Marhauser at al. PAC 1999
1500 2000 2500 3000 3500 4000
102
104
106
Q
frequency [MHz]
7 cell8 cell9 cell
Monopoles
24Matthias Liepe 08/02/2007
Coupler Kick
Symmetrizing stub helps to reduce transverse kick fields and resulting emittance growth.
25Matthias Liepe 08/02/2007
Multipacting
•Multipacting happens 3.5-13 MV/m.
•Location is the bend of the enlarged beam tube.
•Can be processed through and can re-appear.
• If required: can modify bend region to suppress multipacting
From ERL injector cavity:
26Matthias Liepe 08/02/2007
Mechanical Design for low Microphonics
• Cavity design:
– High mechanical vibration frequencies
– Low sensitivity to He-pressure changes 1 bar pressure
Courtesy E. Zaplatin
mode 1
mode 3
Stif. ring 0.7*req 0.4*req 0.65*req no ring
mode freq / Hz freq / Hz freq / Hz freq / Hz
1 131.03 85.34 115.15 54.62
2 131.04 85.33 115.15 54.62
3 315.52 191.3 268.39 133.34
4 315.52 191.3 268.39 133.34
27Matthias Liepe 08/02/2007
7-Cell Cavity End-Cell Design
• End cell shape has significant impact (example HOM):
• Will use fine-tuning of end cell to– Increase damping of strongest dipole mode(s)– Avoid strong monopole modes at beam harmonics
(2600 MHz, 5200 MHz, …)
changed end-cell shape
28Matthias Liepe 08/02/2007
Main Linac HOM Loads• HOM load based on injector HOM load design
• But: Higher current (factor 2) and longer cavities significant more power to be absorbed (150 W vs. 30 W)
• Resonant HOM excitation of high frequency modes can result in even greater HOM power in a few cavities
Design need to support > 200 W
• Future work:– Design optimization (3D models)
– Material studies
– Improved and simplified design for higher power handling and reduced fabrication cost (reduced number of absorber tiles, only one absorbing material …)
29Matthias Liepe 08/02/2007
80K Cornell Beamline Absorber
Flange to Cavity
Flange to Cavity
RF Absorbing
Tiles
Cooling Channel
(GHe)Shielded Bellow
•Baseline design finished
•Three RF absorbing materials selected to cover full frequency range
•Full beam test in Full beam test in injector module in injector module in 2008 2008
30Matthias Liepe 08/02/2007
Cornell Beamline Absorber II
• ANSYS simulations confirm 200 W power capability
• Need to determine optimal temperature of HOM loads (probably 100 K; balance of static losses vs. better cryogenic efficiency)
31Matthias Liepe 08/02/2007
HOM Damping Simulations
• CLANS calculations (started 3D Microwave Studio models)
• 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
/cm
2 /G
Hz]
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
32Matthias Liepe 08/02/2007
Test and R&D plansTest and R&D plans
33Matthias Liepe 08/02/2007
Cavity R&D Items– Finalize design
•end cells•number of cells per cavity•study multipacting in more detail
– Polarized cavity to further suppress BBU?
– Study QStudy Q00(E), microphonics level, FE, (E), microphonics level, FE, radiation and trip rates to finalize radiation and trip rates to finalize parametersparameters•Results from injector cryomodule•Daresbury / Cornell / LBNL Test Module•Main linac test cryomodule
– R&D program for high Q0 at medium fields
34Matthias Liepe 08/02/2007
Modified
Daresbury / Cornell / LBNL Test Module
•Collaborative effort to study high-Qhigh-Q00 cavity cavity operation operation
•Trip rate and QTrip rate and Q00 vs. gradient (long term operation vs. gradient (long term operation planned)planned)
•Microphonics levels and high QMicrophonics levels and high QLL operation operation
•Beam operation (ERLP@Daresbury)Beam operation (ERLP@Daresbury)
•Modified Stanford/Rossendorf cryomodule
Two 1.3 GHz 7 cell cavities
Cornell-style cold HOM load
Cornell-style input coupler (from ERL injector)
35Matthias Liepe 08/02/2007
HOM Load R&D Items
– Optimize and simplify HOM beam line load design
– Optimize operating temperature of HOM loads
– Explore waveguide HOM damping scheme
– Verify HOM damping for main linac cavity with beam