PEP-II hardware reliability

Hardware reliability takes on special importance in large accelerator facilities intended to work as factories; i.e., when they are expected to deliver design performance for extended periods of time. The PEP-II "B-Factory" at SLAC is such a facility. In this paper, we summarize PEP-II reliability statistics from the first four years of production running. The four running periods extended from January 12 through October 31, 2000, from February 4, 2001 through June 30, 2002, from November 15, 2002 through June 30, 2003, and from September 9, 2003 through July 31, 2004. These four periods are designated Runs 1, 2, 3, and 4 in the discussion and tables presented in the paper. The first four runs encompassed 30,359 hours. During this time, PEP-II was delivering luminosity to the BaBar detector 57.9 percent of the time. In addition, 5.3 percent of the time was used for scheduled dedicated machine development work, and 4.5 percent was scheduled off for maintenance, installations, or safety checks. Injection and tuning accounted for 19.9 percent. The remaining 12.4 percent was lost due to malfunctions. During this time period, a total of 9701 malfunctions were reported, but most did not interrupt the running program. The unscheduled down time, a total of 3883 hours, was attributed to 1724 of these malfunctions. Mean time to fail (MTTF) and mean time to repair (MTTR) are presented for each of the major subsystems, and long-term availability trends are discussed.