Recent Fundamental RF Power Coupler developments at CERN

Recent Fundamental RF Power Coupler developments at CERN

First SLHiPP meetingCERN, Geneva, 8 & 9 December 2011

First SLHiPP meeting, 8 & 9 December 2011

2

Acknowledgements• SPL coupler review committee :

▫ Ali Nassiri, Wolf Dietrich Moeller, Mark Champion, Sergey Kazakov, Mircea Stirbet, Amos Dexter, Rama Calaga, Miguel Jimenez, Sergio Calatroni, Ofelia Capatina, Vittorio Parma

• DESY :▫ Wolf-dietrich Moeller, Axel

Matthaisen, Birte Van der Horst, and team

• CEA :▫ Stephane Chel, Guillaume Devanz,

Michel Desmond, and team• ESRF :

▫ Jorn Jacob, Vincent Serriere, Jean-Maurice Mercier, Didier Boilot, and team

• APS :▫ Ali Nassiri, Doug Horan, Gian

Tenko, Dave Brubenker, and team

• CERN :▫ Mechanical & Material

Engineering group : Ramon Folch, Francesco Bertinelli,

Serge Mathot, Agostino Vacca, Thierry Tardy, Thierry Calamand, Thierry Renaglia, Ofelia Capatina, Marc Polini, Laurent Deparis, Philippe Frichot, Jean-Marie Geisser, Jean-Marc Malzacker, Pierre Moyret, Alain Stadler, and team

▫ Vacuum, Surface & Coating group : Miguel Jimenez, Sergio Calatroni,

Wilhelmus Vollenberg, Marina Malabaila, Nicolas Zelko, and team

▫ Magnets, Superconductors & Cryostats group: Vittorio Parma, Arnaud Van de

Craene, and team▫ RF group :

Sebastien Calvo, Antoine Boucherie, Nicolas Jurado, Charles Julie, all FSU-AB03 team members

Eric Montesinos


3

Contents

• SPL couplers

• ESRF/SOLEIL/APS couplers

• Linac 4 window

Eric Montesinos


4

SPL coupler requirements

Eric Montesinos

Technical ChoicesSingle window couplerFixed couplerWith a Double Walled TubeMounted in clean room with its double walled tube horizontally in only one operationVertically below the cavity and will be a support for the cavity (first time worldwide)

With a HV DC biasing capacitorAir cooled

RF Characteristicsf0 704.4 MHz

Power levels

1000 kW pulsed0.4 + 1.2 + 0.4 = 2.0 ms50 Hz (20 ms)100 kW average

Cavity design gradient 19-25 MV/m

Qext of input coupler 1.2 x 106

Input line Ø 100 / 43.5 mm = 50 Ω(from the cavity design)

Waveguides WR 1150


5

Two proposed designs

Eric Montesinos


6

SPL coupler project• Four ‘vacuum lines’:

▫ 4 cylindrical window couplers▫ 4 planar disk window

couplers▫ 8 Double walled Tubes▫ 4 test boxes

• DESY clean process assembly▫ (Jlab also proposed to help)

• CERN LLRF measurements

• CEA RF power tests▫ (BNL also proposed to help)

Eric Montesinos


7

Couplers construction• All components have been

constructed by May 2011▫ 4 cylindrical window couplers▫ 4 planar disk window couplers▫ 8 Double walled Tubes▫ 4 test boxes

• All components have been individually vacuum leak free tested

• Specific transport boxes with springs as per tetrodes have been designed to avoid any shocks

Eric Montesinos


8

Test box• Test box is made with only two parts

with a helicoflex seal :▫ Upper cover with included machined

vacuum knife flanges included▫ Two covers with inner copper plated

• Advantages :▫ Easier (not easy) to copper plate▫ Easily cleanable regarding clean

room requirements for high gradient field cavity

▫ Can be used for several sets of couplers (if large series : SPL, …)

• Drawbacks :▫ Helicoflex faces have to be very well

prepared▫ Self-supporting structure : heavy

weight

• Holes for vacuum port :▫ without sharp edges▫ L = 5 x D to allow

correct RF shielding (P peak max = 1’000 kW !), helped with additional vacuum tombac

Eric Montesinos


9

Waveguides• Six parts waveguide

• Easy mounting

• Easily exchangeable shims and allows post machining of internal step for matching if needed

• Knife contacts all along to suppress any gap at intersections in order to avoid any arcing (DESY difficulties with brazed WG)

• No mechanical stress given to the ceramic, fixed point onto the body only

• Capacitor self centered around the ceramic and designed to symmetrically air cool the ceramic

Eric Montesinos


10

Planar disk inner line• As the coupler must be air

cooled to avoid any water leak

• Specific trick to ensure inner RF contact :▫ Inner line on top of the

ceramic▫ Springs are compressed▫ Air cane inside the inner line▫ Screw to the inner ceramic▫ Release springs to ensure

RF contacts▫ Absolutely no stress to the

ceramic

Eric Montesinos

1 MWRFcontact


11

Clean room assembly (DESY)• Test box assembly not easy

because of specific surfaces roughness needed for helicoflex

• Couplers assembly was also not easy because :▫ Couplers are heavy▫ Last connection has to be

done manually▫ Ok for few prototypes, not

for a large series

• Two cavities were sent to CERN, but…

Eric Montesinos


12

Double walled Tube (DT)• Despite a previous long

experience with LEP and LHC DT, a difficulty occurred in the construction process :▫ An additional machining has

erased all the care put into the preparation of DT

• Copper sputtering was removed with a simple water Ultra Sonic cleaning process

• Decision to condition the two first vacuum lines with NOT copper coated DT

Eric Montesinos


13

Low Power measurements (CERN)

Eric Montesinos

• S11 with 2 x N/WG adaptors + 2 couplers + 2 DT + test box :▫ Cylindrical = -26.5 dB▫ Planar disk = -18 dB

• Even with -16 dB, 1 MW corresponds to 25 kW reverse power

• Ok for first tests


14

RF tests (CEA premises)• Tests started with cylindrical

window couplers

• Not baked out, static vacuum~ 2 x 10 -7 mbar▫ Wanted to check RF▫ Size of the test box 250 mm x

600 mm▫ Helicoflex

• Pulse mode process

• After 3 weeks,50 kW 20 us / 20 Hz

• We stopped the test

Eric Montesinos


15

RF tests• In the meantime, second test box

with two planar disk window coupler has been baked out

• No helicoflex leak▫ Very slow heating up and heating

down ramps 10 C / hour▫ Maximum temperature was only

150 C during three days▫ Nitrogen onto copper rings to

avoid any oxidization

• Very good static vacuum after the process:▫ ~ 1 x 10 -9 mbar▫ Was 2 x 10-5 mbar before

starting the bake out

Eric Montesinos


16

RF tests

• Restarted the tests with the planar disk window couplers

• Pulse mode process starting with 20 µs pulses / 8 Hz

• After 10 days : 1000 kW 2 ms / 8 Hz

Eric Montesinos


17

RF tests• Still some out gazing around

three power levels :▫ 20 kW▫ 140 kW▫ 300 kW

• No more vacuum activity above 300 kW

• Processing around these values is under way and vacuum decreases ramp after ramp

• Tests ongoing to reach 1000 kW 2ms / 50 Hz (uncoated DT limitation)

Eric Montesinos


18

Next steps• While testing the planar disk

window couplers, we baked out cylindrical couplers (already seen RF) :▫ Initial static vacuum

~ 1.0 x 10-6 mbar▫ After bake out, static vacuum

~ 1 x 10-10 mbar

• Once planar disk window coupler tests finished, restart cylindrical window coupler tests

• These tests are scheduled beginning 2012

• Corrected DT copper coating under way

• Cryostat flanges to be added prior to new clean room assembly

• Mounting of chosen design at DESY clean room:▫ Two tests cavities▫ Four copper coated DT with

cryostat flanges▫ Four couplers

• RF tests at Saclay premises of the four chosen couplers

Eric Montesinos


19

Key RF items

• Ceramic and waveguide as a whole matching system

• Waveguides:▫ ‘Plug and Play’ ▫ No doorknob▫ Reduced height▫ Screwed waveguide▫ No mechanical stress to

the ceramic

• Test cavity in two parts, easier to copper plate and easy to clean

• Inner line of planar disk window with RF contacts ensured by compression of springs

• Only air cooled couplers

Eric Montesinos


20

ESRF/SOLEIL/APS coupler requirements

RF Characteristics

f0 352.2 MHz

Power levels 200 kW cw

Technical choices Single window, fixed coupler

Waveguides WR 2300 half height

Mechanical constraintsMust match with previous design without any infrastructure modification

Eric Montesinos


21

Kovar vs copper window• LEP Kovar® brazed window :

▫ Kovar ® rings on both ends▫ Body and Antenna with Stainless Steel

rings brazed onto copper▫ TIG welding under mechanical

pressure ensuring RF contact▫ Easy Kovar ® brazing process▫ Weakness of the design : RF contact▫ High field between two Kovar ®

‘antennas’, arcing

• LHC copper brazed window :▫ Massive copper ring for high power▫ No RF contact, fully continuous RF

path▫ Spherical ends of copper rings for no

arcing▫ Very difficult brazing process (6 years

to be finalized)▫ Advantage of the design: extremely

powerful

Eric Montesinos

LEP kovar collar

RF contact via compression

while TIG welding

LHC copper collars

EB welded

Metallic continuity

Ceramic

Ceramic


22

Kovar vs copper window• APS coupler with kovar/copper

misalignment :▫ RF contact lost▫ kovar has metalized ceramic (grey)▫ Thermal losses▫ Vacuum leak

• CERN Kovar vs CERN copper with a LHC test :▫ kovar RF losses▫ Thermal overheating▫ Burnt kapton and peek (> 250 C)

• Copper ceramic window is not only a simple upgrade, it is a very new design really stronger than kovar window ceramic

Eric Montesinos


23

ESRF coupler project• LEP I style coupler became ESRF

trouble maker (five unexplained vacuum leaks in 2008)

• Proposal of a new coupler to fit into existing premises with NO infrastructure modification

• Coupling loop :▫ Water cool to be sure there will be no

mechanical stress given to the ceramic

• Three vacuum ports for conditioning process:▫ Vacuum gauge close to the ceramic▫ Glass window for arc detection▫ e- antenna

• First prototype delivered September 2011

Eric Montesinos


24

ESRF coupler tests• Bake out onto cavity process

• CERN pulsed mode conditioning process under vacuum feedback

• Off tune, 200 kWpulsed 5 ms / 50 Hz in 22 hours

• On tune, 200 kW cw in 60 hours

• Qualified for ESRF needs

• Last week, on tune 300 kW cw

Eric Montesinos

20 us / 20 ms pulses

50 us / 20 ms pulses

500 us / 20 ms pulses CW

Pulse length of 20us - 50us - 100us - 200us - 500us - 1ms - 2ms - 5ms - 10mswith fixed repetition time of 20 ms and finally CW

time

Power


25

ESRF/SOLEIL next steps• Further tests to higher

power levels, up to maximum test bench capability, i.e. 350 kW cw

• Two additional ESRF ‘series couplers’

• Two SOLEIL couplers:▫ Electrical antenna, air

cooled▫ CERN LEP type test cavity▫ Outer coaxial line with

corrected length▫ Tests at ESRF

Eric Montesinos


26

APS coupler project• Collaboration requested during

CWRF workshop Barcelona in April 2010

• Present coupler also LEP I type limited to 100 kW cw

• Design similar to ESRF:▫ Coupling loop differs▫ Inches vs mm▫ Waveguide air cooling as

personal safety interlock

• Individual bake out of the coupler, reduced exposition to air (not under mobile clean room)

Eric Montesinos


27

APS ‘picture horizontal flip’• Loop angle does not allow any

vacuum ports to be useful

• Not needed because cavity is small enough, pumping perfect (1 x 10-9 Torr without RF)

• If one additional coupler, it will be with no more vacuum ports:▫ Reduces the costs to its

minimum▫ Reduces the risk of vacuum

potential leaks▫ Cavity small enough to ensure

a correct vacuum monitoring

Eric Montesinos


28

APS coupler tests and next steps• Tests started last week

• Conditioning in cw mode with a very low pressure:▫ 1 x 10-8 Torr maximum▫ interlock at 5 x 10-8 Torr

• 25 kW cw after two days

• Further tests to higher power levels, up to maximum test bench capability, i.e. 200 kW cw

• Implementation of CERN pulse mode conditioning system

Eric Montesinos

Vacuum Power


29

Linac 4 window requirements

RF Characteristics

f0 352.2 MHz

Power levels 1250 kW pulsed 2ms / 50 Hz (SPL cycle)

Waveguides WR 2300

Mechanical constraints As short as possible

Eric Montesinos


30

Ceramic assembly• As simple as possible :

▫ 1 : Ceramic Ø 400 mm x 25 mm▫ 2 : Inner copper ring Ø 400 mm x

1.25 mm▫ Outer copper ring▫ Stainless steel ring for Helicoflex

• Two first individually brazed sub-assemblies:▫ Brazed ceramic = ceramic + inner

copper ring▫ Outer ring for Vacuum seal = outer

copper ring + stainless steel ring

• Main ceramic = EB welding of two parts

• Ti sputtering of the vacuum side

Eric Montesinos


31

LINAC 4 window project• As simple as possible

▫ 1-2-3-4 : Ceramic assembly▫ 5 : spacer▫ 6 : Helicoflex seal▫ 7-8 : Stainless Steel flanges

• Massive flanges, not copper plated

• More difficult design than it look likes because of the two shapes : cylindrical and rectangular, with incorporated screws

Eric Montesinos

1 2 437

56

8

EB1EB2

FM


32

S11 < 32 dB over ~ 16 MHz

LINAC 4 window next step• First prototype with helicoflex

▫ Low Power RF ok

• Second prototype with no helicoflex but an Electron Beam Welding is under construction

• Tests of two prototypes face to face onto a vacuum waveguide line in CEA premises :▫ 1.25 MW 2 ms 50 Hz▫ Define the best design

(Helicoflex/EBW)

• Series construction of 30 windows▫ Ceramics already ordered

Eric Montesinos


33

Future coupler developments• HIE-Isolde, under way

▫ 100 MHz▫ 500 W (1 kW ?)▫ Operating at 2k▫ Inside vacuum vessel

• LIU-SPS 200 MHz▫ New concept with coaxial plain

disk and capacitive coupling▫ 900 kW pulsed 50 % (ms)▫ 450 kW cw power

• SPS 800 MHz consolidation▫ Scaling of Linac 4 to 800 MHz▫ 150 kW cw

• One for all Crab Cavities coupler▫ Same window for all

cavities▫ Compatible with various

coupling system

• Whatever new coupler needed

Eric Montesinos


34

Thank you very muchfor your attention

And many thanks again to all the persons involvedin all these projects

Eric Montesinos

Date post:	24-Feb-2016
Category:	Documents
Upload:	vivian
View:	53 times
Download:	0 times

Recent Fundamental RF Power Coupler developments at CERN

Documents