Source for the production of ions of deuterium extracted from a radio frequency plasma (SPIDER), the full-scale prototype of the ion source of the ITER heating neutral beam (HNB) injector, started the operation in June 2018. During the first experimental phase, some issues concerned its novel and complex power supply system. Dedicated analyses and studies, carried out to understand and address them, resulted in beam source and plant modifications to be implemented during the SPIDER major shutdown, started in late 2021. To assess the validity of the conceived countermeasures, reducing the risks of introducing additional issues, partial modifications, or mitigating actions were adopted in SPIDER in 2020, as an intermediate step to confirm each intervention with experimental evidence. This article deals mainly with two problems: the stray RF currents, flowing on plant components not intended for that purpose and the mutual interaction of the oscillators supplying the RF driven plasma source. A modification of the output stage of the oscillators allowed the reduction of the stray currents to at least 25% of the original value, while the RF circuits layout improvement resulted in the reduction of the mutual inductance of the nearby circuits to 25% of the original value. The introduced improvements and the obtained experimental results are presented and discussed.

In SPIDER the negative ions are extracted from a plasma generated with RadioFrequency drivers, each fed with a power up to 100 kW at the frequency of 1 MHz. One of its distinguishing characteristics is the source fully installed within a vacuum vessel and operated in the residual gas pressure, including the matching networks for the driver impedance installed on the source backside. These features lead to more challenging design and integration also of the RF system, and in fact during the experimentation some issues arose: the operation of the RF generators is affected by the magnetic coupling among the RF circuits on board the source; breakdowns are observed on the source backside while powering the RF circuit and common mode RF currents flow through the diagnostics and dc power circuits. An overall system design review was launched with the aim to identify the root cause of these issues, feasible solutions and a plan for their fix, which are presented and discussed in this paper.