Evolution of High Power Klystrons

This paper will review the evolution of high power klystrons and their applications in supplying rf energy for linear accelerators. Other types of possible rf sourc2s were conside,.ed for SLAC. and the reasons for the selectioa of klystrons will be _ gcven. A brief revir-5 of klystr 1 typr in j :e fo-. various accelerators will also be given, but tht emphasis will be on the work done by Stanford and its subcontractors in developing klystrons capable of achieving peak powers in excess of 20 ?.%4 and average powers of 20 kin for use with the two-mile Stanford linear accelerator. Selection of RF Sources Three main types of rf sources were considered for use at SLAC: the magnetron, the amplitron, and the klystron. Even though Slater and his collee&ues were able to demonstrate the us2 of a large numbe: of magnetron oscillators in an operating machine, the system is complicated and inefficient. The main difficulty arises from the relative instability of heavily loaded oscillators and the problems of accurate phasing between oscillators. In addition , it was felt that the rather short life expectancy and reasonably low peak power of magnetrons would make the tube unsuitdole for use in a large accelerator. The amplitron has the advantages of extremely good efficiency and low phase sensitivity to input voltage. Its main disadvantages result from very low gain and low isolation between input and output , requiring the us2 of complicated amplifier chains and of many high power isolators. The added cost and complexity and the lowering of the efficiency by the additional amplifier chains requirement made the amplitron solution unattractive for use with the SLAC accelerator. The klystron amplifier has been used on almost all linear electron accelerators having multiple power sources and operating at microwave frequencies. It appears most suitable for the following reasons: (1) it is a high gain amplifier, which simnlifies phasing and driving problems, partic~J-lariy in a very long machine; (2) the necessary average and peak powers can be easily obtained; (3) the efficiency is reasonable, although not as high as that of other tube types; (4) the high degree of isolation between input and output makes it more stable under conditions of load mismatch; and (5) experience on the Stanford Mark III accelerator over the past ten years and use in many radar applications indicates the potentially long life of this device. The use of high …