Techniques for maintaining design efficiency when operating klystron amplifiers at levels below the maximum output power

Some accelerator designs require that the amount of RF power supplied to the accelerating cavities be varied along the accelerator length. In order to minimize operating expense for a high-power, CW system, it is also desirable to maximize RF-generator efficiency. Klystrons are the traditional source used for RF-accelerator applications, but a high-efficiency klystron design is optimized to provide high efficiency at only one operating point, usually at saturation. Different klystron designs could be used for different power levels to maintain high efficiency; however, this approach would increase the capital costs of the accelerator. We discuss several methods of varying the output power of a klystron and show semi-empirically a way to preserve klystron efficiency. We derive two semi-empirical methods to predict the variations in power and efficiency under nonoptimum operating conditions. Experimental data from a 1.25 MW klystron are also shown. These data support our hypothesis and our models