Overview of cavity reliability at Diamond Light Source 3 GeV 300 mA third generation light source Operating for users since 2007 NC linac and full energy booster, SC storage ring RF

Storage Ring RF Space for up to three CESR-B cavities in the RF straight Usually operating with two cavities at any on time Voltages 1.1 MV and 1.4 MV, powers 180 kW and 240 kW Cavity pressure ~5x10-10 mbar with beam

Storage ring RF history Diamond has four CESR-B cavities Two in operation at any one time Cavity failure is a major disruption Mean Time Between Failures for RF system has improved 20 hours in 2007 200 hours in 2017 Cavity Failure date Detail A none B 2009, 2014 UHV leak C 2006 Insulation vacuum leak D 2015 Window failure

Cavity “fast vacuum trips” Sequence of events Discharge in high field near coupling tongue Discharge is fed by desorbed gas Electrons are detected Arc crowbars cavity Reflected power trips amplifier Pressure spike follows

Reliability and cavity voltage Trips initially dominated by fast vacuum trips MTBF is strongly dependent on cavity voltage Each cavity has a “safe” operating voltage below which it is unconditionally stable Trip rate is independent of power Reliable operating voltages Cavity A 1.1 MV Cavity B 1.2 MV Cavity C 1.4 MV Cavity D 0.8 MV Trips eliminated by reducing voltage Effective Why can’t we operate at higher voltage? Not always free to drop voltage Impact on lifetime Special modes of operation, e.g.. Low α SRLEm3ps: 1.7MV, 3.5ps, stable SRLETHz: 1.7MV, 3.5ps, bursting SR21: 1.7 MV, 16ps, stable

Cavity warm-ups Cavities have been subjected to warm-ups to room temperature in the past Occasional full warm-ups tried in early days and with problematic cavities Gas is cleared from surfaces and identified by RGA More frequent partial warm-ups have been used with all cavities Every warm-up carries the risk of disaster! Very little benefit at Diamond Cavities are no better in run following warm-up Cavities are no worse in run following no warm-up No degradation of reliability if partial warm-ups are stopped for several runs Short term improvement apparent following partial warm-up No trips in first few days Trips resume when vacuum returns to normal

Cavity conditioning Cavity “probe blips” Pulse conditioning every week in run-time 2.3 MV peak voltage,10% duty cycle (10 ms/100 ms) Detune angle scanned to sweep standing wave Carried out when work is going on elsewhere in SR X-ray emission reduced after conditioning Two cavities simplify “with-beam” conditioning Sweep cavity phases to move power between cavities Beam is restored after violent conditioning events Never tried an extended period without conditioning Cavity “probe blips” Probe blips Appear as fall in cavity voltage LLRF increases power and power level trip follows Add filter to reduce bandwidth 220 kW 270 kW Before After 1µs Pfwd Cavity Probe Prev Blip recognition and handling to be addressed in new digital LLRF Development of ALBA µTCA-based system First DLLRF has been tested with NC cavity in booster and RF test bunker Blip recognition not yet implemented

Recent cavity failures 2014: Leak from helium can into cavity UHV Failed during cool-down from room temperature No more warm-ups unless absolutely necessary Indium seal at waveguide flange Returned to RI in December 2014 Cavity returned to DLS in 2016 Failed acceptance test with leak at indium seal on FBT Repair is to resume shortly 2015: Failure of ceramic-metal braze at window During normal operation after standard conditioning Repaired on-site at RAL in February 2016 Installed spare window assembly Used RAL Space satellite assembly cleanroom ISO class 5 cleanroom with 5 tonne crane Tested to 2.1 MV operation in RF test facility

Normal conducting cavities for resilience Cannot risk interruption to operations arising from a lengthy repair of a superconducting cavity The first of two normal-conducting cavities has been installed in the storage ring Installed outside the RF straight Will allow operation with only one superconducting cavity in ring Will support SC cavities, not replace them Keeping the cavities permanently cold Cavities are now only warmed up when absolutely necessary Cryogenic plant is regularly serviced We must warm up the cold box Can we keep the cavities cold? One cryostat of helium lasts longer than one day Pass cold helium gas through cryostat until Dewar is depleted How long will this give us? How long does a service require?

Summary and conclusions Cavity reliability has improved enormously since operation started in 2007 Cavity fast vacuum trips have been eliminated Largely down to a reduction in operating voltage Different cavities have different safe operational voltages Fundamental studies are ongoing Cavity warm-ups have been discontinued A risky operation with no obvious benefit Can we keep the cavities cold permanently? Regular pulsed conditioning continues Scheduled around weekly machine interventions Cavity failure has long-term impact Normal conducting cavities are being introduced to ensure machine continuity The Diamond RF Group Chris Christou Shivaji Pande Pengda Gu Adam Rankin Matt Maddock David Spink Peter Marten Anton Tropp