Excess RF power required for RF control of the Spallation Neutron Source (SNS) linac, a pulsed high-intensity superconducting proton accelerator

A high-intensity proton linac, such as that being planned for the SNS, requires accurate RF control of cavity fields for the entire pulse in order to avoid beam spill. The current design requirement for the SNS is RF field stability within /spl plusmn/0.5% and /spl plusmn/0.5/spl deg/. This RF control capability is achieved by the control electronics using the excess RF power to correct disturbances. To minimize the initial capital costs, the RF system is designed with 'just enough' RF power. All the usual disturbances exist, such as beam noise, klystron/HVPS noise, coupler imperfections, transport losses, turn-on and turn-off transients, etc. As a superconducting linac, there are added disturbances of large magnitude, including Lorentz detuning and microphonics. The effects of these disturbances and the power required to correct them are estimated, and the result shows that the highest power systems in the SNS have just enough margin, with little or no excess margin.

1 Overview of SNS

The SNS is a 1 GeV Iinac, followed by an accumulator ring and neutron spallation target. The SNS is being built at Oak Ridge National Laboratory by a 6 laboratory collaboration. Los Alamos has responsibility for the room temperature portions of the linac, as well as the RF systems for the entire linac [2], [3]. SNS is very concerned with beam spill because of the resulting activation of the hardware that can come from the spilled beam. Numerous discussions between LANL, Brookhaven and ORNL lead to a specification for the cavity field stability of 0.5% 0.5°. Excess power capability leads to excess costs, and the entire collaboration has been very interested in minimizing the costs for the project construction. The decision was made to buy the minimum power level in the klystron that is needed in the highest power portion of the superconducting sections. That turned out to be 550 kW. Follow-on changes and discussions by SNS engineers and physicists and by review committees questioned whether there was sufficient, or perhaps, too much excess margin. The questioning came from two directions. One was a concern that there was not enough margin, and the system would not function properly. The other is that there was too much margin, and the system could accelerate more beam, or power cavities at higher gradients. This paper summarizes the analysis that was done. Figure 1: Klystron saturated power setting, beam power, and min. klystron power per superconducting module