Mircea Stirbet RF Conditioning: Systems and Procedures Jefferson Laboratory
General requirements for input couplers - Sustain RF power required for operation of accelerator with beam - Do not compromise cavity performance (vacuum integrity, cleanliness) - Control multipacting, which could limit performances at certain power levels Geometry ( RF window, impedance) Vacuum design and associated instrumentation (RF and UHV specific materials, fast gauges and vacuum controllers) Instrumentation and controls (DC bias, optical ports for arc detectors, temperature and electron probes) - Sustain mechanical stresses during manufacturing, assembling, transportation, cooldown and operation. - Minimize cryogenic heat load - Ensure the requested Qexternal, minimize VSWR and insertion loss to avoid reflections and heating - Ensure high reliability in machine operation. - Satisfy different safety requirements (RF, HV, electrical, UHV, cryogenics, pressure)
Factors that limit RF window performances - heating of the vacuum barrier (ceramic) - multipactor discharge - cleanness - violent arcing High average power windows are most susceptible to the first two items. Multipactor discharge is a precarious surface phenomenon which is driven by the electric fields in the RF environment, being dependent on geometry, RF frequency and surface quality. It is addressed by applying an anti-multipactor coating (Ti or TiNx), proper cleaning, baking and RF conditioning. The Ti coating is RF lossy and contributes to the other problem encountered in high average power windows, heating of the ceramic. In addition to surface heating by the anti-multipactor coating, bulk heating of the ceramic occurs due to its non-zero dielectric loss tangent. If arcing events are triggered (local bad vacuum), the ceramic window could be metalized with subsequent higher RF losses and heating. It is mainly the problem of heating of the ceramic and the induced mechanical stress that limit the RF windows performances.
Factors that limit RF coupler performances On the air side of the coupler : - surface finish - RF contacts - dust, humidity - arcing events On the vacuum side of the coupler: - material quality (bulk and surface) - multipactor discharge -arcing events - vacuum and vacuum instrumentation - cooling
Conditioning fundamental power couplers All existing high power couplers exhibit severe outgassing and multipacting barriers of vacuum and RF exposed surfaces, as a result of geometrical configurations and surface conditions (material, contaminants, finishing). These problems can be overcome by design and different procedures like cleaning, baking followed by RF Conditioning - application of RF power under various conditions and in various configurations: On waveguide test stands or cavities. Pulsing the RF with different duty factors In continuous wave mode (CW) as traveling wave mode as standing wave mode Sweeping of frequency or RF power amplitude RF processing in conjunction with bias voltages Objective: to clean the surfaces from contaminants (molecular or particulate) and minimize conditions for multipacting. During this process the QA of different coupler components and associated instrumentation is obtained.
Standard conditioning procedure No Standard conditioning procedure can be found. Each laboratory has developed or is in the process of developing an optimum procedure for its coupler and RF test conditions. Many different methods are applied in succession such as TW processing, SW (off resonance) processing, frequency sweeping, power sweeping, bias voltage processing, warm and cold processing, with vacuum interlocking at different vacuum levels. The objective is always to touch surface area with RF, burn particulates and induce controlled gas layers desorption (they enhance the secondary electron emission coefficient and cause local desorption outbursts which could facilitate arcing events).
Aging couplers In all cases the fundamental power couplers are conditioned in a dedicated test stand (or on a cavity) at power levels at least two times higher than the power specified for machine operation. RF conditioned is done also after coupler is assembled on the cavity, before machine operation and after a long shutdown of the machine. If a cavity quenches in operation or trips by some other reason, all the forward power in the coupler is reflected and the coupler see local peak power between a minimum value and 4 times the input power level, crossing all predicted multipacting levels for a certain geometry. As couplers demonstrate RF conditioning memory, it seems important to age couplers at higher RF powers levels than used in machine operation.
Multipacting Electron Multipacting is a significant problem in vacuum and RF exposed surfaces and requires in most cases extensive RF conditioning. As in cavities, certain conditions have to be satisfied to generate multipacting events: An electron emitted from a wall of the line is under the influence of the EM fields returning to its origin within an integer number of RF cycles. The impacting electrons produce more than one electron, if the impact energy is high enough. Because in coaxial lines standing, traveling and mixed wave patterns can exit depending on the load conditions, multipacting is very complex in these systems and control methods shall be applied.
Multipacting simulations SNS FPC DC bias +1.9 / +2.4 kv Pasi Ylä-Oijala and Marko Ukkola, Multipacting Studies, Rolf Nevanlinna Institute, University of Helsinki, Finland
Resonator for multipacting studies MATERIAL N I[mA] N T[s] Copper 18 2.92 1 2080 Copper baked at 400 C 2 3.52 1 1223 Cu, stored one week in PE bag 2 3.30 2; 1 >6500 TiN on Copper 1 3.00 1 552 Titanium on Aluminum 2 2.88 1 885.5 Titanium on Copper 5 3.03 1 993.1 Cu Ti stored one week in PE bag 2 3.21 3; 1 >6500 Aluminum 7 4.30 3 >6500 Stainless Steel 18 3.37 2; 1 2871 SS electrochemically Cu plated 7 3.33 1 2571 All samples show multipacting of at least first order. They differ in magnitude of multipacting current and in processing time. One interesting result is the dramatic deterioration of Cu and Ti coated Cu samples after storage in a plastic (PE) bag. Nevertheless, the common practice to store or transport RF components in plastic bags should be avoided, if multipacting is of concern. D. Proch et al., Measurement of multipacting currents of metal surfaces in rf fields, PAC 1995, p.1776
KEKB Coupler The window of the input coupler has: A choke structure to decrease the electric field at the gold braze of the ceramic. The surface of the window on the vacuum side is coated with TiNxOy. The inner conductor is made of electro polished copper and is water-cooled. The waveguide transformer air-cooled. The outer conductor is copper plated (electrochemical) stainless steel and form. To control multipacting events, the doorknob inner conductor transition is separated via two layers of 0.125 thick polyimide film. Instrumentation ports are located near the ceramic window (arc detector, vacuum gauge and electron pick up). For KEKB the input coupler is conditioned up to 300 kw or full-reflected power.
KEKB test stand
KEKB - HER ring Each coupler is processed with DC bias +2000 V and - 2000 V up to 300 kw. Each time the cavity is warmed up, coupler conditioning is repeated. Power levels exceeding 380 kw with beam of 870 ma (950 ma) have been reported S. Mitsunobu, Operation Experience of Superconducting Cavities for KEKB, in : http://conference.kek.jp/srf2001/sa002, and HPC Workshop
KEKB conditioning with DC Bias Y.Kijima*, S.Mitsunobu, T.Furuya, T.Tajima, Input Coupler of Superconducting Cavity KEKB, EPAC 2000, Vienna, Austria, http://accelconf.web.cern.ch/accelconf/e00/papers/thp1a11.pdf
KEKB - ARES cavity High-power test of the coupler with RF power in excess of 800 kw (twice as the ARES cavity requires). A dedicated coupler test stand consisting of a small cylindrical cavity connecting two couplers to the RF klystron and to a dummy load was built. Different design couplers have been tested up to 950 kw in this configuration. F.Naito et al., The Input coupler for the KEKB ARES cavity, in : APAC98, http://accelconf.web.cern.ch/accelconf/a98/apac98/6d040.pdf
KEK - 972 MHz 792 MHz window assembly Components for 972 MHz couplers
KEK - 972 MHz Room temperature test stand for 972 MHz couplers
KEK - 972 MHz JAERI instrumentation for high RF power tests Vacuum pumping system Vacuum gauges and controllers Residual Gas Analyzer
KEK - 972 MHz Results Specification Preparation External Q value : 5 x10 5 Required rf input power : 250 kw (Beam current = 30 ma) Pulsed operation Cleaning Assembling Baking Vacuum pressure High Power Test : ~3.0 msec, 25 Hz : rinsing with ultra-pure water : in the clean room : around 120 o C for 24 hours : less than 10-6 Pa Number of couplers : 2 couplers Result of processing : 340 kw (2.45 msec, 25Hz) Limitation Problem : nothing : 1.0 MW (0.6 msec, 25Hz) : available rf power of Klystron
KEK RF conditioning The coupler was newly designed being based on that for the 508 MHz APS cavity of TRISTAN ring. - initial vacuum 5 10-8 Torr - start RF conditioning in CW at ~100 W and increase RF power keeping the vacuum better than 5 10-6 Torr. The RFL power was used as interlock - 60 kw are reached in 60 hours. - with RF OFF vacuum is 5 10-8 Torr - baking is done at 150 C for 24 hours and vacuum is improved at 3 10-9 Torr - RF processing by pulsing RF 10% duty cycle allows reaching 100 kw in 12 hours. - Continue RF conditioning in CW and 150 kw are obtained without problems after a total of 60 hours. T. Koseki et al., High-Power Conditioning of an RF Cavity for High Brilliant Synchrotron Radiation Source, EPAC 1996, http://accelconf.web.cern.ch/accelconf/e96/papers/wepl/wep040l.pdf
DESY
DESY
DESY Results RF conditioning computer controlled, starting with small pulse duration. Interlocks on electron activity and arcing. Electronic module to control RF amplitude as function of vacuum outburst made but not yet in use.
LEP2-50 ohm variable coupler
LEP2 - components for the 75 ohm fix coupler LEP2 Power Coupler Air inlet antenna RF screen and HV protection HV connection Air inlet doorknob Doorknob Kapton foil Capacitor for DC bias Waveguide UHV gauge port Ceramic window Assembly body He gas outlet Insulating vacuum Antenna (copper) Extension (copper plated stainless steel) 75 Ω Coaxial line 4.2 o K He gas inlet Cavity vacuum Liquid helium CW max 125 kw
LEP2 - waveguide and alternative DC bias system
LEP2 - doorknob and capacitor for DC bias
LEP2 - RF conditioning of couplers All main power couplers have been preconditioned and qualified on a room temperature test stand. RF conditioning was always preceded by cleaning, assembling in a clean room and by baking for 24 hours at 200 C, then RF processing was started in pulse mode, with small pulse duration and low amplitude, followed by CW processing by cycling the RF power between two power levels or performing constant power tests. Necessary time: 75-100 hours for RF processing if assisted by fast vacuum feedback loop. On cold test modules (with five or single cell(s) cavity) - similar RF conditioning was used and and results regarding procedures and different coupler components (long term high RF power tests, baking in situ, He pulse processing of cavities, DC bias, copper plating methods for the outer conductors) were obtained. LEP2 couplers have been conditioned cryo modules in String area before installation the machine. LEP2 couplers have been re-conditioned after shut-downs. In String area, RF conditioning was done as function of cavity field (not RF power) and the time needed to qualify a module for installation in the machine was about three weeks.
LEP2 - room temperature test cart Room temperature test cart for LEP couplers
LEP2 - LHe module for testing cavities and couplers
LEP2 - single cell LHe module setup
LEP2 - Procedure for RF conditioning All main couplers mounted in LEP2 on SC modules have been RF processed according with the following procedure: Pulsed RF (1-10 ms pulse length and 50-200 ms repetition rate). The use of a fast analog vacuum feedback loop was essential. Cycling the RF power 10-250 kw, with different rise times (controlled by a computer program in addition to the fast analog vacuum loop). Cycling the RF power 10-150 kw for about 1 hour, followed by a test at constant of several hours at 150 kw, then by another cycling 10-150 kw. Cycling again 10-250 kw for several hours, during this time the DC bias efficiency in suppressing multipactor events was tested. After RF processing at room temperature, couplers have been stored in stainless steel containers, under dry, dust free nitrogen.
LEP2 - pulse modulation Vacuum feedback loop RF pulse modulation
LEP2 - RF conditioning
LEP - Results More than 300 main power couplers have been qualified (in CW, traveling mode up to 250 kw ) on test room temperature test stand for assembly on the LEP SC cavities. After RF conditioning, machine operation have not been hindered by couplers problems - less than 5% of RF trips have been due to Coupler vacuum during machine operation. DC bias had a major contribution in ensuring stable machine operation (multipacting free) for more than 5 years. The reliability of the main power couplers have allowed to transfer higher RF power values and so to increase the cavity fields from the design value of 6 MV/m to more than 7.5 MV/m. None of the 288 power couplers broke during the whole life of LEP2 and only a small fraction of the machine down time was due to superconducting RF - a much smaller fraction than due to conventional equipment faults.
LEP final
LHC main power coupler Variable input couplers, providing a remotely controlled change of external Q by an order of magnitude under power, are required for the 400 MHz LHC superconducting cavities. These couplers must handle a forward power of 120 kw average and 180 kw pulsed (with about 50 ms pulse duration) under a large variety of load conditions up to full reflection.
LHC variable main coupler Air cooling inlet Air inlet antenna Displacement mechanism (60 mm stroke) RF screen and HV protection 7 Ω Coaxial line (under vacuum) Air pressure interlock Bellows Motor drive Capacitor for DC bias Waveguide Ceramic window Antenna (copper) 75 Ω Coaxial line (under vacuum)
LHC baking and room temperature test devices
LHC variable couplers - RF conditioning RF conditioning was done as function of coupler vacuum (a fast vacuum feedback loop controlling the RF power). On room temperature test cart: - In traveling wave mode started with pulse (small duration, low repetition rate) up to 500 kw, then in CW from 1 kw up to 500 kw (klystron limit). After conditioning, long term RF power tests have been performed in traveling wave : at 400 kw for more than 150 hours and at 500 kw (maximum klystron power) for about 50 hours. - In standing wave mode, conditioning at any phase and any coupling has been achieved up to the maximum klystron power. Up to 500 kw forward power (2 MW equivalent peak power) has been reached with pulses of 50 ms duration and of 10% duty cycle. On LHC cold bi-module: Each coupler was RF conditioned using our standard procedure starting with loose coupling, small pulses and low duty factor, then changing to CW mode and continued with to stronger coupling by changing the penetration of the antenna in steps up to the cavity field limit or other RF components (elbows, terminating loads).
LHC RF - vacuum feedback during conditioning
LHC variable main power couplers on a bimodule
LEP and LHC controls for room temperature qualification
LHC variable coupler performance On room temperature test stand: In travelling wave mode: - 500 kw (limited by klystron) for more than 50 hours - 400 kw for more than 150 hours In pure standing wave mode (any phase, any coupling) 500 kw forward power (2 MW equivalent peak power) with pulsed RF (50 ms, 10% duty cycle). On superconducting cavity: More than one order of magnitude of the Qext has been achieved by changing the position of the antenna. Only several hours of RF conditioning were needed to reach 300 kw (limit by the terminating load on the circulator) or 3 MV/m (cavity field limit).
LEP2 NEG coating
LEP2 NEG coating - 2
LEP2 NEG coating - 3
JLAB - SNS - FPC design parameters Design parameters Parameter in Operation in Processing Q ext 7.3 / 7.0 x 10 5 (20%) NA Impedance 50 Ohm Peak power 550 kw 1 MW(TW).5MW (SW) Pulse length 1.3 ms 1.3 ms Repetition rate 60 pps 60 pps max average power 48 kw 60 kw Bias ± 2.5 kv ± 2.5 kv
JLAB - SNS low RF power measurements
JLAB - SNS baking FPCs Baking at 200 ºC for 24 hrs temperature ramping with 10 ºC/h
JLAB - SNS - FPCs Instrumentation
JLAB - SNS - RF conditioning Start RF conditioning in traveling wave mode with small pulse duration, amplitude and low duty cycle, under the control of the feedback vacuum loop ant of the computer program (expect explosive vacuum outbursts for 6-8 hours),. Change pulse characteristics as function of vacuum and electron activity. If arcing in the vacuum side of the coupler, change vacuum threshold. After reaching 1 MW, cycle the RF amplitude for several hours between two power levels (1-800 kw) and/or keep the RF constant at a certain power level for at least 24 hours. Test DC bias efficiency in controlling multipacting events Perform RF conditioning in standing wave mode, using pulse duration smaller than the nominal value. For a certain position of the short circuit increase the input RF power from a minimum value up to a designated value, conditioning multipacting levels.
JLAB - SNS vacuum RF amplitude modulation During RF conditioning a complex series of events occurs on the vacuum side of the coupler. Safe RF processing: limiting pressure rise to 2 10-7 -5 10-7 mbar and responding to explosive vacuum events with fast vacuum gauges and controllers. Fast gauges, vacuum controllers and RF amplitude feedback loops are indispensable for safe RF processing.
JLAB - SNS RF conditioning Cycling, keeping constant, and again cycling the RF power between 10-800 kw (1 ms, 60 Hz).
JLAB - SNS RF conditioning - 2
JLAB - SNS RF conditioning - 3
JLAB - SNS RF conditioning RF Power through the Jlab Coupler August 21st - August 30th, 2001 750 700 650 600 550 500 450 400 350 300 Aug 21, 2001 Aug 22, 2001 Aug 23, 2001 Aug 24, 2001 Aug 27, 2001 Aug 28, 2001 Aug 29, 2001 Aug 30, 2001 250 200 150 100 50 0 0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00 45.00 50.00 Test Time (Hr)
Conditioning and testing - multipacting Multipacting at 120 and 480 kw
JLAB - SNS - DC bias A B C D Cycling RF: A: - no DC bias B: +2.5 kv C: - 2.5 kv D: - no DC bias (10-800 kw, 60 sec on top, pulse 1 ms, 60 Hz)
JLAB - SNS - RGA during RF conditioning A B RGA while cycling 10-800 kw with DC bias: - A with - 2.5 kv - B with +2.5 kv
JLAB - SNS - Results RF tests performed at LANL at room temperature: In TW mode reaching 2 MW with 0.6 ms pulse width at 60 Hz repetition rate, transferring an average power of more than 70 kw. In SW mode: local RF peak power in excess of 2 MW. RF tests performed at JLAB 1 MW room temperature test facility: In TW mode up to 1 MW (1 ms, 30 Hz repetition rate) (long term tests in CW 550 kw, 1 ms with 60 Hz repetition rate) In SW mode more than 2 MW local peak power RF tests at JLAB on cryomodule: on the first coupler tested on the cryomodule, 16 mv/m were obtained after only 5 hours of RF conditioning. Later, 700 kw in full reflection with 1 ms pulses have been reported without problems.
Guide d onde (magic T)
Conclusions Efficient RF conditioning and multipacting control: Assembly of components with surfaces exposed to vacuum and RF in a clean room. Avoid moisture and plastic bags during storage. Titanium coating of ceramic on the vacuum side. Apply outbaking procedure before RF processing. Efficient vacuum and electron current measurements near critical components. Avoid RF overheating of critical components. Conditioning should start with pulsed (low amplitude and duty factor) RF followed by cycling the RF power, or keeping the power constant for at least 24 hours, under the control of a fast vacuum feedback loop and/or of a computer program (allow for out gassing of less than 2 10-7 mbar). Keep the RF processed components under dry, dust free nitrogen. Start RF conditioning of couplers on SC cavity with warm outer conductors. Apply nominal value DC bias only after RF conditioning.