The first two weeks for the long promised Top-Up tests have been completed successfully with participation of selected user groups by using only one insertion device U-41 and the associated beamlines. Top-Up injections were only possible if there was stored beam and the efficiency of the previous injection was better than 90 %. The corresponding and redundant interlock systems to guarantee radiation safety were implemented and tested before.
The required high injection efficiency was achieved by improving the whole chain of accelerators and components involved. Nowadays, the gun delivers a higher charge in a shorter and more homogeneous pulse. In the microtron the reliability of the RF was improved significantly by a software modification. Already since quite a while the synchrotron operates at a higher horizontal tune. This reduces the horizontal emittance of the accelerated beam. With the dynamical vertical steering during the energy ramp the extracted beam is free from the strong non-linear horizontal to vertical coupling and the beam shows the expected elliptical shape in the transfer line. The smaller emittance and the smaller emittance coupling increase the injection efficiency. With the help of an RF-knockout system installed in the synchrotron the length of the bunch train in the booster synchrotron is made exactly 100 ns long. This is important since RF-knockout in the storage ring as done usually would lead to 12.5 % particle losses which is intolerable in top up mode. Therefore, the beam is injected at 7 consecutive positions on the circumference of the ring, thus leaving an unfilled gap of 100 ns. In order to support reliable and high injection efficiency the stored beam in the storage ring is moved permanently towards the septum magnet with a local closed orbit bump. This permits operation of the fast pulsed kicker magnets at 50% of their nominal set values. At these relaxed settings a better closed orbit adjustment can be achieved for the fast bump and the perturbation on the stored beam during injections is dominated by the stray fields of the septum magnet. In the last shut down the transfer line was equipped with new diagnostics. Now the charge as well as the position of the injected beam can be monitored just before the septum magnet without perturbations by the nearby pulsed kicker magnets. Based on all of these efforts and the available diagnostics a fine-tuning of the transfer line optics finally resulted in an injection efficiency of better than 90 %.
During the tests the repetition rate of the synchrotron was administratively set to 1Hz instead of the usual 10 Hz. Top up injections were only permitted every 10 seconds. It turned out that at this reduced rate all components relevant for the high injection efficiency worked surprisingly stable during the whole test week. Fig. 1 shows the situation on the last day of the top up test. The test ended at 11 pm and there were two planned interruptions of the top up injections at 7:30 in the morning and at 3:30 in the afternoon. These interruptions were required in order to give access to the experimental hall for maintenance work.

Top Up operation on the last day of the test week. During the test the current was stable to better ±0.1mA at 295 mA – in the middle. The injection efficiency was 90% based on the current measured at the end of the transfer line – at the bottom. This monitor systematically overestimates the charge by nearly 5 % compared to the DCCT-measurement in the synchrotron. In reality the injection efficiency is roughly 5% higher. The above stability was achieved during the whole week.
The successful daily operation in Top-Up mode requires: