Abstract
After acquiring several reliable spare targets, SNS ramped the beam power from 850 kW to 1.4 MW, which required an increase in H- beam pulse length from 0.88 to 1.0 ms at 60 Hz. This increase initially produced slow 2-MHz power ramp-ups and, after several weeks of uninterrupted operation, it produced plasma outages every time the pulse length was raised above ~0.95 ms. Similar outages were previously observed towards the end of long service cycles, which were believed to indicate that the breakdown voltage of the high purity hydrogen started to exceed the induced electric fields. In 2011 the RF was reconfigured to start with 10 cycles of 1.96 MHz, which yielded the shortest H- beam rise times and apparently eliminated those plasma outages. The new, pulse-length dependent outages were eliminated by increasing the initial frequency to 1.985 MHz. However, careful frequency studies are unable to justify this frequency. In addition, the paper discusses the issues and solutions for the electron-dump voltage, which starts to sag and become unstable after several weeks of high current operation.
