In 2003, a total of 680 shifts, representing 5441 hours, were dedicated to the Users (see Table 6 for statistics). This year once again, high figures have been maintained. From this year onwards, special care must be taken in interpreting the 'refill time', availability and Mean Time Between Failures (MTBF) figures. This is due to the fact that since 5th February 2003 the Front End shutters remain open during refills. This year, contrary to previous years, we therefore consider the X-ray source availability based on the assumption that the beam can be used during the refill. With this definition in mind, the beam availability reached 97.95% in 2003. If the refill time were to be considered as non-available time, the beam availability would have been 96.95%. However in order to be consistent, we have to take into account all the failures which occur during refills. These failures, which were transparent to the Users with the injection 'Front-End closed', are now visible and contribute to a reduction in the MTBF. With this revised way of counting, the MTBF reached 39.4 hours. This figure remains good and reflects the fact that on average, the beam is interrupted only 48 minutes every 39.4 hours. Note that 28 failures out of 138 occurred during injection and would not have been accounted for in the MTBF computation of previous years. Removing the failures during injection would give an MTBF of 49.5 hours in 2003, compared to 58 hours in 2002.


Table 6: Statistics of shifts, dead time, mean time between failures and beam availability in 2003.


This lower MTBF in 2003 is essentially due to the partial unavailability of HQPS. The HQPS system consists of 10 x 0.8 MW Diesel Engines coupled to a rotating accumulator and an alternator. It is capable of eliminating the electron beam losses originating from the voltage and phase drops which occur during stormy weather (somewhat frequent in the Grenoble area during spring and summer due to the surrounding mountains). The HQPS system has been successfully operating for 7 years and has recently suffered some damage. Repairs were carried out and the system was operated in its normal fashion from June to November, once again protecting the machine infrastructure. Unfortunately it had to be stopped again following the appearance of abnormal wear on the bearings. Detailed investigations into the problem are being made with the help of the companies who manufactured the various components, and under the leadership of an Expert appointed by the "Tribunal de Grenoble".

Finally, let us mention some excellent periods of delivery with 14 periods of uninterrupted beam greater than 79 hours (including one period of 156 hours and two periods of 144 hours!)