CERN SPL Fundamental Power Coupler Progress report SPL

CERN SPL Fundamental Power Coupler Progress report SPL meeting CERN, 6 - 7 December 2012 Eric Montesinos, CERN BE-RF-PM

2 SPL meeting, CERN 5 -6 December 2012 Contents • Introduction ▫ CERN RF Fundamental Power Coupler team • SPL coupler project ▫ SPL FPC initial design proposals ▫ March 2010 Coupler review ▫ June 2010 SPL Coupler project ▫ Construction ▫ Assembly in DESY clean room ▫ Double walled Tubes (DT) ▫ LLRF tests ▫ High Power tests at CEA : TW ▫ High Power tests at CEA : SW • Next steps • Key R&D results : ▫ Mono bloc waveguide ▫ Test box ▫ Double walled Tube as support for cavity ▫ High Average power air cooled couplers

3 SPL meeting, CERN 5 -6 December 2012 CERN RF Fundamental Power Coupler team CERN Machine Design Prototype Series SPS 200 ü ü 2001 LHC 400 ü ü 2006 Linac 4 ü ü Under Way SPL cylindrical ü Under test SPL planar ü Under test HIE-Isolde ü Under test Crab Cavities To come SPS 800 To come LIU- SPS 200 To come SPS 200 MHz coupler LHC 400 MHz coupler Linac 4 352 MHz coupler SPL 704 MHz couplers HIE Isolde 100 MHz couplers

4 SPL meeting, CERN 5 -6 December 2012 CERN RF Fundamental Power Coupler team Machine Design Prototype Tests ESRF ü ü 300 k. W CW ANL-APS ü ü 100 k. W CW SOLEIL ü Under way ESRF 352 MHz coupler SOLEIL 352 MHz coupler design Argonne APS 352 MHz coupler

5 SPL meeting, CERN 2009 2010 2011 2012 5 -6 December 2012 2013 SPL FPC initial design proposals RF Characteristics f 0 704. 4 MHz Power levels 1000 k. W pulsed 0. 4 + 1. 2 + 0. 4 = 2. 0 ms 50 Hz (20 ms) 100 k. W average Cavity design gradient 19 -25 MV/m Qext of input coupler 1. 2 x 106 Input line Ø Waveguides Cylindrical Window CEA HIPPI (baseline) 100 / 43. 5 mm = 50 Ω (from the cavity design) WR 1150 Disk Coaxial window Disk WG Window

6 SPL meeting, CERN 2009 2010 2011 2012 5 -6 December 2012 2013 SPL FPC initial design proposals � Baseline: CEA Saclay coupler Window-Ceramic-Antenna 275 Series 4 -8 MCHF � To be upgraded for SPL cryomodule compatibility � Kept open the possibility to use another design, to be fixed in March 2010 � Comparison basis: � RF power capability � Low heat load � Tuning capability (fixed coupling, adjustable coupling) � Conditioning time � Contamination during beam part assembly � Easy installation � Integration with the cryomodule � Easy operation � Cost 15 k. CHF 4. 125 MCHF 14 k. CHF 3. 85 MCHF 30 k. CHF 8. 25 MCHF 15 k. CHF 4. 125 MCHF

7 SPL meeting, CERN 2009 2010 2011 5 -6 December 2012 2013 March 2010 Coupler review • Review committee members : Technical Choices Single window coupler Fixed coupler With a Double Walled Tube Mounted in clean room with its double walled tube horizontally in only one operation Vertically below the cavity and will be a support for the cavity (first time worldwide) With a HV DC biasing capacitor Air cooled ▫ ▫ ▫ Ali Nassiri (Chair) Wolf Dietrich Moeller Mark Champion Sergey Kazakov Mircea Stirbet • Presenters : ▫ Amos Dexter & Rama Calaga (Multipacting simulations) ▫ Miguel Jimenez (Vacuum) ▫ Sergio Calatroni (DT Coating) ▫ Ofelia Capatina & Vittorio Parma (Cryomodule integration) ▫ Eric Montesinos (FPC)

8 SPL meeting, CERN 5 -6 December 2012 Two designs have been validated Cylindrical ceramic window Coaxial disk ceramic window

9 SPL meeting, CERN Same Waveguide with integrated matching step (instead of a doorknob) And same DC capacitor Same Double Walled Tube (Outer line of a coaxial transmission Line) With same interface flange to cryomodule 5 -6 December 2012

10 SPL meeting, CERN Cylindrical window • Advantages ▫ LHC window with high power capability due to its solid copper collars ▫ Simple to cool down with air ▫ Absolutely free of mechanical stress onto the antenna ▫ “plug and play” waveguide and DC capacitor, no stress to the ceramic • Drawbacks ▫ Ceramic is part of the matching system, imposing the waveguide position 5 -6 December 2012

11 SPL meeting, CERN Coaxial Disk window • Advantages ▫ Very simple and well mastered brazing of ceramic onto a titanium flange ▫ Simple to cool down with air ▫ “plug and play” waveguide and DC capacitor, no stress to the ceramic • Drawback ▫ Ceramic is part of the matching system, fixing the waveguide position 5 -6 December 2012

12 SPL meeting, CERN 2009 2010 2011 2012 5 -6 December 2012 2013 SPL coupler project (June 2010) • Four vacuum lines: ▫ 4 cylindrical window couplers ▫ 4 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)

13 SPL meeting, CERN 2009 2010 2011 Couplers construction • All parts have been produced by May 2011 (10 months) • All components have been individually vacuum leak free tested before being sent to DESY for assembly 2012 5 -6 December 2012 2013

14 SPL meeting, CERN 2009 2010 2011 Couplers construction • All parts have been produced by May 2011 (10 months) • All components have been individually vacuum leak free tested before being sent to DESY for assembly • Specific transport boxes with springs as per tetrodes have been designed to avoid any shocks 2012 5 -6 December 2012 2013

15 SPL meeting, CERN 2009 2010 2011 5 -6 December 2012 Assembly in DESY clean room • In June 2011, the four cavities were assembled in DESY clean room • Use of a specially designed helicoflex seal 2013

16 SPL meeting, CERN 2009 2010 2011 5 -6 December 2012 Assembly in DESY clean room • In June 2011, the four cavities were assembled in DESY clean room • Use of a specially designed helicoflex seal • Thanks to DESY colleagues who have performed a very good job and to a very good preparation job of the jointing surfaces all four cavities were leak free 2013

17 SPL meeting, CERN 2009 2010 2011 2012 Double walled Tube (DT) • Copper sputtering removed with simple water Ultra Sonic cleaning process • Mistake in the machining process : ▫ An additional machining step not included in the lists has erased all the care put into the preparation of DT • Decision to continue the first two vacuum lines with NOT copper sputtered DT : ▫ Lead into a very important delay onto the schedule ▫ Will limit average power during tests 5 -6 December 2012 2013

18 SPL meeting, CERN 2009 2010 2011 2012 5 -6 December 2012 2013 LLRF measurements at CERN (summer 2011) • S 11 with 2 x N/WG adaptors + 2 couplers + 2 DT + test box : ▫ Cylindrical windows = -26. 5 d. B ▫ Coaxial planar disk = -18 d. B • Pfwd/Prev (S 11) at CEA premises with 2 couplers + 2 DT + test box : ▫ Cylindrical window = -17 d. B ▫ Coaxial planar disk = -40 d. B • Even with only -16 d. B, 1 MW will reverse 25 k. W, acceptable with CEA premises, thanks to their circulator

19 SPL meeting, CERN 2009 2010 2011 2012 5 -6 December 2012 2013 RF power tests at CEA cylindrical window •

20 SPL meeting, CERN 2009 2010 2011 5 -6 December 2012 2013 RF power tests at CEA disk windows • In the meantime, second test box with two disk window couplers has been baked out at CERN • Pulse mode process • No helicoflex leak • Ultimate goal of 1000 k. W 2 ms - 50 Hz was not possible due to losses in uncoated DT limitation ▫ Very slow heating up and heating down ramps 10 C / hour ▫ Maximum temperature during 48 hours was only 150 C ▫ Nitrogen onto copper rings to avoid any oxidization • Very good static vacuum after the process ~ 1 x 10 -10 mbar (vs 5 x 10 -7 mbar before starting the bake out) • After two weeks : ▫ 1000 k. W 2 ms - 20 Hz

21 SPL meeting, CERN 2009 2010 2011 5 -6 December 2012 2013 RF power tests at CEA cylindrical window • We also baked out the first test box with two cylindrical windows at CERN • No helicoflex leak • Very good static vacuum after the process ~ 1 x 10 -10 mbar (vs 2 x 10 -7 mbar before starting the bake out) • Pulse mode process • After 1 week: ▫ 1000 k. W 2 ms - 20 Hz • Ultimate goal of 2 ms - 50 Hz was not possible due to losses in uncoated DT limitation

22 SPL meeting, CERN 2009 2010 2011 5 -6 December 2012 2013 TW tests ok (May 2012) • Both coupler versions have reached TW maximum values of 1 MW 2 ms - 20 Hz • Limitation of average power to 20 Hz instead of 50 Hz due to uncoated TD • Successful TW tests

23 SPL meeting, CERN 2009 2010 2011 2012 5 -6 December 2012 2013 SW tests with cylindrical window • A variable short circuit has been designed and constructed at CERN for SW test • Steps of 20 mm to have enough accuracy performing tests within all phases

24 SPL meeting, CERN 2009 2010 2011 2012 5 -6 December 2012 2013 SW tests with cylindrical window • We started SW tests in pulse mode • With 500 µs – 8 Hz : ▫ Up to 500 k. W, no trouble ▫ Arcing when reached 575 k. W • It seemed to be air side as there was no vacuum activity • There was no arc detector to stop the test, a photomultiplier modulated RF, and it was possible to hear arcing in the WG system

25 SPL meeting, CERN 2009 2010 2011 2012 5 -6 December 2012 2013 SW tests with cylindrical window • Air side arcing were confirmed when dismounting WG • Input coupler presented some impressive arcing traces • Output coupler was without any trouble

26 SPL meeting, CERN 2009 2010 2011 2012 5 -6 December 2012 2013 SW tests with cylindrical window • Analyzing black deposit, we only identified : ▫ Gold coming from upper gold platted copper collar ▫ Copper coming from lower copper collar • There without any doubt electrical arcing (>1’ 000) between two copper brazed collars • The only positive point is that ceramic remained vacuum leak free

27 SPL meeting, CERN 5 -6 December 2012 Improvement of cylindrical window • A trick given by Michel Langlois (ex-Thales tube designer) is to improve air flow in the critical area in order to avoid ionization of the air around the ceramic itself • improve air cooling around ceramic with a short circuit as air inlet • Inserting PEEK screen all around ceramic Air inlet through WG short circuit PEEK Screen all around ceramic

28 SPL meeting, CERN 2009 2010 2011 2012 5 -6 December 2012 2013 SW tests with coaxial disk windows • We then decided to start SW tests with coaxial disk windows • Up to date, we performed: ▫ 1000 k. W 1. 5 ms – 4 Hz • Some RF leaks were observed • CEA was finally not sure if these RF leaks were coming form couplers

29 SPL meeting, CERN 2009 2010 2011 2012 5 -6 December 2012 2013 SW tests with coaxial disk windows Arcing traces on inner line • Last week, while preparing test box with copper coated DT, we observed some arcing traces on the outer line in front of the inner contact • These traces were not present after TW test • There were no arc detector fault during the whole SW test (interlock active) Arcing traces on outer line

30 SPL meeting, CERN 2009 2010 2011 2012 SW tests with coaxial disk windows 5 -6 December 2012 2013 • Still waiting items to find an explanation • Possible candidates : ▫ Not enough springs ensuring specific 1 MW RF contact : 1. Inner ceramic is brazed with a stainless steel nut 2. Air cane is built with a specific bolt 3. Air cane is screwed compressing springs which when relaxed ensure 1 MW RF contact ▫ Contact surfaces were not flat enough 2. 3. 1.

31 SPL meeting, CERN 2009 2010 2011 2012 5 -6 December 2012 2013 SW tests conclusion • Cylindrical window couplers have performed 575 k. W 500 µs - 8 Hz full reflection all phases -> very strong arcing, interlock not operative • Coaxial disk window couplers have performed 1000 k. W 1. 5 ms - 4 Hz full reflection all phases, no interlock (operative), but few arcing traces (to be understood and explained) • SW tests are not conclusive, work still have to be done

32 SPL meeting, CERN 2009 2010 2011 5 -6 December 2012 2013 Coated DT • 6 DT with new copper sputtering have been processed : ▫ Test with 150 Bars water have been performed ▫ TD qualified ▫ 2 have been provided to cryomodule team • Unfortunately DESY clean room was not anymore available • CEA proposed to assemble couplers and copper sputtered DT onto test boxes in ‘L’orme les Meurisiers’ premises : ▫ One Test box with two copper plated DT and two coaxial disk window has been assembled last week

33 SPL meeting, CERN 2009 2010 2011 2012 5 -6 December 2012 2013 Next steps • Second test box will be assembled beginning 2013 • Cylindrical or Disk windows ? Decision after cylindrical arcing analysis • Beginning 2013, final tests at CEA with : ▫ 1000 k. W TW 2 ms – 50 Hz ▫ 1000 k. W SW limited to 200 µs – 50 Hz

34 SPL meeting, CERN 2009 2010 2011 Next steps • EB welding of cryostat flanges will be done afterward • Study of clean room tooling for assembly onto SPL cavity has started • Assembly of couplers onto cavities in CERN clean room goal remains mid-2013 • RF power tests on cavity asap 2012 5 -6 December 2012 2013

35 SPL meeting, CERN 2009 2010 2011 2012 5 -6 December 2012 2013 Next steps • Provide a CERN clean assembly premises for couplers • Continue all remaining couplers power tests with CERN amplifiers when available • Longer term basis : ▫ After qualification of SPL cavities, make a test with ‘dirty’ couplers (ISO 7 instead of ISO 4 for example, tbd), to quantify coupler cleanliness impact onto cavity field

36 SPL meeting, CERN 5 -6 December 2012 Key R&D results Mono bloc waveguides 1. 2. • ‘Plug and Play’ waveguide with matching step and DC capacitor included : 1. 2. 3. Connect the waveguide to the body line Insert the contacts ring Final assembly of air cooling system • No doorknob, reduced height, with no mechanical stress to the ceramic 3.

37 SPL meeting, CERN Key R&D results Test box • All in one, only two covers conditioning test box • Pros : ▫ Easier (not easy) copper sputtering ▫ Self supporting shape ▫ Easily cleanable for SRF needs, can be used for several sets of coupler (if large series : SPL, ESS, …) • Cons : ▫ Helicoflex faces to be very well prepared ▫ Self-supporting structure : heavy weight 5 -6 December 2012

38 SPL meeting, CERN 5 -6 December 2012 Key R&D results 3 D printing & RF • Successfully made available 3 D printing for RF tests with a special silver paint Machined 3 D printed @ 400 MHz S 21 = 0. 01 d. B S 11 = -40 d. B S 21 = 0. 03 d. B S 11 = -25 d. B Delay [days] Cost [CHF] Machined 3 D 15 5 9000 3800

39 SPL meeting, CERN 5 -6 December 2012 Key R&D results High average power air cooled couplers • Cylindrical window : ▫ TW : 1000 k. W 2 ms - 20 Hz ▫ SW : 550 k. W 500 µs - 8 Hz • Coaxial disk window : ▫ TW : 1000 k. W 2 ms - 20 Hz ▫ SW : 1000 k. W 1. 5 ms - 4 Hz

40 SPL meeting, CERN 5 -6 December 2012 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 • CERN : ▫ Mechanical & Material Engineering group : � Francesco Bertinelli, Ramon Folch, 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 ▫ Vacuum, Surface & Coating group : � Miguel Jimenez, Sergio Calatroni, Wilhelmus Vollenberg, Marina Malabaila, Nicolas Zelko ▫ Magnets, Superconductors & Cryostats : � Vittorio Parma, Arnaud Van de Craene, ▫ RF group : � Sebastien Calvo, Antoine Boucherie, all FSU-BE 03 members • DESY : ▫ Wolf-dietrich Moeller, Axel Matthaisen, Birte Van der Horst, and local team • CEA : ▫ Stephane Chel, Guillaume Devanz, Michel Desmond, and local team • ESRF : ▫ Jorn Jacob, Vincent Serriere, Loys, Jean. Maurice Mercier, Didier Boillot • APS : ▫ Ali Nassiri, Doug Horan, Gian Trenko, Dave Brubenker, and local team

41 SPL meeting, CERN 5 -6 December 2012 Thank you very much for your attention