Download - Coupler Multipactor Studies F. Wang, B. Rusnak, C. Adolphsen, C. Nantista, G. Bowden, Lixin Ge
Coupler Multipactor Studies
F. Wang, B. Rusnak, C. Adolphsen, C. Nantista, G. Bowden, Lixin Ge
SLAC Coupler Test Stand - Evaluate Parts Individually
• An SRF coupler is a complex integration of many components
• When a coupler is conditioned, it necessarily responds in aggregate
• This makes it difficult to know which components dominate the behavior and limit performance
• As multipacting and RF conditioning are largely surface phenomena, modeling effectiveness is limited
• We developed a test stand to evaluate individual components of the TTF coupler design
Schematic Layout of Test Stand with Parts to be Evaluated
RF load
window
window WG to coax
Device Under Test
RF in
More Detailed Cross Section of the Test Fixture
RF in RF out
Parts to be Tests
– straight 40 mm diam. coaxial line
– SST and copper plating by different recipes
– straight 40 mm coax with bellows
– cold window assembly with window
– cold window assembly without window
Specialized Vacuum WG-to-Coax Transitions
Designs accommodated existing high-power vacuum windows at SLAC
UCRL-CONF-2319326
Detachable Center Conductor
Significant effort went into ensuring the center conductor could easily disconnect upon opening the test stand, but still maintained solid electrical and thermal contacts with the anchors on each side. This was done by screw-attaching the center-conductor length on the right, and having the center slide into a 4-jaw BeCu receiver on the left.
Test Stand Set UpWR650 vacuum window
custom waveguide-to-coaxial transitions
coaxial step transitions from 62 to 40 mm
straight coax Device Under Test
electron pickup probe
The coupler component test stand at SLAC can test components up to 5 MW at 1.3 GHz at pulse lengths up to 1 msec.
Artist View of our Setup (Viewed Through Plastic Screen)
Plot Showing Vacuum Activity Consistent with Expected Multipacting Bands for a Straight Coax
0 200 400 600 800 1000 1200 14000
0.5
1
1.5
2
Forward RF Power (kW)S
econ
dary
yie
ld
third order
fourth order
fifth ordersixth order
seven order
While observed bands generally agreed with simulations, there were difference in the power level and width. The variations are being further investigated, but are suspected to be related to either onset delays in MP emission relative to the RF wave or space charge effects, neither of which are present in the simulations.
0 200 400 600 800 1000 1200 14000
50
100
150
Cu
rre
nt o
f Io
n P
um
e: u
A
Forward power: kW
0 200 400 600 800 1000 1200 1400
-100
-50
0
Ele
ctro
n s
ign
al:m
V
electron signal amplitude
vacuum level
600mm long straight SS coax section test results
600mm long straight Copper plated SS coax section test results
Strong MP around 300kW!
Magic Simulations
3/
4
~ss
sU
UUf
Secondary electron energy:
is the work function of material (4.31 for Fe and 4.98 for Cu)
The peak of secondary electron is at
24 cm coax line, OD 40mm, ID 12.5mm, 1.3 GHz RF, TW
0 500 1000 1500 2000 25000
0.5
1
1.5
2
Primary Electron Energy: eVS
eco
nd
ary
Yie
ld
[a]
[b]
[c]
[d]
[a,d] SEY based on the measurements[b c] scaled to better match data
Simulation results depends on the Secondary electron distribution
Red for Copper and blue for SS
0 200 400 600 800 1000 12000
0.2
0.4
0.6
0.8
1
1.2
RF Input Power (kW)
Ave
rage
Del
ta
third order
fourth order
fifth ordersixth order
seven order
SS(1.3@310)
Cu(1.3@410)
MAGIC Multipacting Simulation and ‘Resonant Finder’ Results for a 40 mm Coax Line
The electron signal turn-on time is unstable after initial processing, likely due to insufficient electrons available to seed the onset of multipacting.[1]
[1] D. Raboso, A. Woode, “A new method of electron seeding used for accurate testing of multipactor transients”, Vol.1, Oct. 1995.
Behavior of the Electron Pickup Signal
E-signal
RF
.
[a] The RF pulse with a 1us high power (3MW) spike;[b] The multipacting electron signal.
0 200 400 600 800 1000 1200 1400 1600 1800 2000
0
0.2
0.4
0.6
[a]
1
Time: ns
Forward RF power level normalized to 300kW
Electron signal amplitude normalized to 30mV
[b]
[a] 200ns RF pulse Spike[b] Multipacting electron signal.
Multipactor Seeding and Time Response
0 50 100 150 200
10-6
10-5
10-4
10-3
10-2
10-1
Time: us
Am
plit
ud
e: a
rb.u
.
DC1st-1.3GHz W/RF2nd-2.6GHz3rd-3.9GHz4th-5.2GHz1st-1.3GHzWO/RF
0 1 2 3 4 5-1.6
-1.4
-1.2
-1
-0.8
-0.6
Time: Ry cycle
Am
plitu
de (
Nor
mal
ized
to
DC
com
pone
t)
klystron power 7001
Harmonics of the Electron Probe Signal
Reconstructed electron signal
Signal Recorded with 5 GHz Scope
TTF3 Coupler Multipacting Simulation Using Parallel Code
Track3P
Lixin Ge
Advanced Computations Department
SLAC
Warm Window cold
TTF3 Multipacting Simulation Components
Taper Region
Input Power: 0-2MW, Scan interval: 50KW.
Cold Coax
300mm 6.25mm
20mm
Simulation Cases:1. with/without reflection
2. with/without external magnetic field
3. with variable center conductor DC biases
Conclusions:Our multipacting simulation results are in excellent agreement with
theoretical calculations and experiment measurements at SLAC.
40 mm Diameter Coax Simulations
Delta as a function of RF input power and Multipacting order
Samples of particle’s resonant trajectory
Reflection: 0.4
Input power level: 160KW
Order: 5th order
Impact Energy region: 542-544 eV
Reflection: 0.8
Input power level: 580KW
Order: 2nd order
Impact Energy region: 1100-1200 eV
Cold Bellows
Model Mesh
Cold Coax with Bellows
NO multipacting activities in the bellows region
Particles Distribution after 20 impactsParticles Distribution at 2nd Period
Particles Distribution at 100th Period
400 kW input power
5mm
5mm
1mm
0.4mm
Ceramic Window Region
Distribution of two points resonant particles’ impact energies, orders versus Input Power between the ceramic window and cold cavity
Model and mesh Field
One resonant particle’s trajectory at 0.68 MW input power
2000eV
3800eV
0 2 4 6 8 10 12 140
0.2
0.4
0.6
0.8
1
1.2
Time: Hrs
RF
Po
we
r in
to C
ou
ple
r P
air
: MW
0 2 4 6 8 10 12 14
10-7
10-6
10-5
10-4
10-3
Time: Hrs
Ion
Pu
mp
Cu
rre
nt:
A
RF-In 10BCPC 10CRF-Out 10D
200s
800s
1100s
100s
400s
50 s
vacuumdebug andRF leak checkRF Leak
Check
Example of RF Processing History – See Smooth Ramp-Up in Power (Slight Dwell at 300 kW)
Similar Results at DESY/Orsay
- D. Kostin, XFEL EIFast XFEL mtg 2006
Overall Summary• Predicted MP is not always observed due to
smearing effects, especially for high order resonances
• Due to absorbed gases on WG surfaces, SEY initally high, producing noticeable MP – however it diminishes as gas is removed from electron/rf interactions
• TTF3 coupler shows no noticeable MP ‘barriers’ during processing.