sumption is prohibitively expensive, given the number of tubes A number of experimental klystrons have been con-in the machine. structed and evaluated at SLAC, KEK and 1NP, aiming toward The magnetic focusing issue is not serious. Superconducoutput power objectives of 100 and 120 MW at 11.4 GHz. tivity provides a solution, and beyond that, there is reason to (SLAC and KEK respectively) or 150 MW at 14 GHz (INP), hope that periodic permanent magnet focusing (PPM) can be with pulse lengths on the order of 1 ItS. Since rf breakdown is made to work for these tubes. considered to be the principal mechanism limiting power for The principal problem is the pulse length in combination such tube.s, most of the effort has been concentrated on the with high peak power. In a well-designed klystron with good design of output circuits that reduce rf gradients by distributing beam optics and little beam interception, the vulnerable areas fields over a longer region of interaction. Another klystron will be the windows and the output circuit. Windows have in component receiving emphasis has i_..en the output window, fact accounted tor most of the experimental klystron failures at where the approach for future tubes may be to use a circular SLAC. The weakness is in the metallized and brazed joint TE01-mode, half-wave window, between the ceramic window and its copper sleeve. There is a Best results to date in this continuing international effort solution, however, described later. The major remaining hurdle are: 50 MW with 1 Its pulses, using a traveling-wave output is in the output circuit which must develop high gradients in circuit (SLAC and INP), and 85 MW wilh 200 ns. pulses order to extract energy from the bunched beam and may cause (SLAC), using two conventional re.entrant, but uncoupled, out-breakdown in vacuum. put cavities. At KEK a klystron with a single, but not reentrant, To avoid very high gradients, interaction between circuit ca city has produced 80 MW in 50 ns pulses. Finally, Haimson and beam must take place over an extended space, rather than has dcmonstrated 100 MW at 50 ns with a traveling-wave out-the usual short gap in a reentrant klystron cavity. Such put. This paper addresses primarily the work performed at extended interaction circuits, however, whether of the standing SLAC during the last two),ears. or traveling wave variety, are large compared to wavelength and capable of supporting unwanted modes or resonances. I, THE PROBLEM Some of these may be trappe.d if they do not couple well to the output waveguide. When excited they can develop very high A future electron-positron collider, if it operates at X-fields and cause o_illations or, if they are of the dipole variety, Band, will require approximately 2000 klystrons at the steer the beam into the drift tube. The_ effects may not mani-100 MW level, for a center-of-mass energy of 0.5 TeV. For 1 or rest themselves unless the beam pulse is long enough, due to 1.5 TeV, which are top energies that physicists would like to high Q's and long filling times, in .so...
