We have achieved phase-locked operation of a module consisting of up to seven relativistic magnetrons. The magnetrons are connected by waveguide in a peer/peer configuration, with no one magnetron acting as a master oscillator. The signals used for coupling are each a substantial fraction of a magnetron’s radiated power. Total extracted power, at 2.8 GHz, was 2.0 GW with four magnetrons and 2.9 GW with seven. Several interconnection geometries were explored. One particular geometry was found to produce qualitatively better phase-locked operation than any other. This was in agreement with the predictions of a numerical model of the magnetrons as coupled van der Pol oscillators.

No abstract

Until the present work, experience with repetitive relativistic magnetrons was sparse, consisting of -1 Hz cooled devices and 100 Hz bursts in uncooled resonators (13)• The highest average power reported is approximately 500 W. We have modified one of our existing magnetrons to include water cooling for repetitive operation. The system is designed for five second bursts at 200 Hz. Earlier single shot operation produces RF energies exceeding 50Joules per micropulse. We have achieved 40 Joules r micropulse at a repetition rate of 100 Herz, or an average power of4 kW for 0.5 seconds.In this paper we discuss issues in interfacing magnetrons with repetitive pulsed power and the operating characterisflcs during a burst of pulses. L INTRODUCTIONWe previously designed and built a magnetron operating near 1 GHz with the expectation of much higher power operation than our S band magnetrons have shown. The magnetron has been used on CAMEL and CAMEL X, which are low impedance, single shot machines. This paper describes the modification and operation of this L band magnetron to operate with a repetitive pulse power system.The experiments have proceeded in two phases. The first was limited to single and double pulses, and the recent phase has used trains of one to one hundred pulses. In the first phase, pulse intervals of 10 msec have been reliably obtained, and pulse intervals of 5 msec have been obtained. In a single pulse mode, extracted power at 1.2 GHz has been up to 700 MW in one waveguide, and the pulse duration is as great as the current pulse, with a FWHM of about 60 nsec. An RF energy of 70 J has been produced from a single pulse using two output waveguides. Efficiencies up to 18% have been inferred.In the latest experiments, pulse trains with variable spacing have been used with repetition rates from five to one hundred Hertz. Reproducible operation has been ahieved, and the individual micropulses are nearly identical within a burst and are smoother than the pulses hieved on our low impedance drivers Camel and Camel X. SYSTEM CONFIGURATION Pulsed Power ConfigurationThe power supply for both phases of the experiment is CLIA (Compact Linear Induction Accelerator), which produces 750 kV at 10 kA into a matched load of 75( for 60 ns at 100 Hz.The CLIA system consists of a ten-cell accelerator with a cathode stalk to sum the voltage up to a single diode load, ten magnetically switched water insulated PFL's, a single two-stage Magnetic Compression Unit (MCU) to charge the PFL's, and thyratron switched Intermediate Energy Store (IES) and Command Resonant Charge (CRC) units to drive the MW.A linear induction accelerator system permits all pulse compression to be done at moderate voltage (40 to 150 kV) and then uses the accelerator structure to add parallel voltage pulses into a single high voltage output (750 kV). This allows the switching to be done at moderate voltage and the use of hydrogen thyratrons and magnetic switches is possible. A more complete description will be published later.For the initial tests, a Marx bank...

We discuss design of a multiple magnetron module and issues related to it, including modelling of the phase-locking process with various geometries via a Van der Pol analysis. We also describe two important "confidence building" experiments; one determined if the couplers load the magnetron output, which would then restrict the number of couplers we would use. Another determined the effect of operating a magnetron with the axis of the cathode displaced from the axis of the magnet coils, and showed we could use a single coil set for the entire module rather than individual coil sets for each magnetron.

No abstract