Automated BWO design using iterative learning control

Summary form only given, as follows. In our earlier works we reported on experimental results showing how finite length variations in a high power backward wave oscillator (BWO) will result in sinusoidal variations in both frequency and power outputs. By manually shifting the slow wave structure with respect to the cutoff neck by one-quarter of a wavelength we were able to achieve maximum frequency agility (large bandwidth). Automated control of the shifting, beam current, and cathode voltage will allow the device to perform specific tasks and will ultimately lead to a "Smart Tube." Such tasks may include the ability to maximize power of efficiency for a given frequency, or to achieve maximum frequency agility at a given constant power. Accomplishing these goals requires the development of a learning control system in conjunction with directed hardware. In this paper, we report on various completed subsystems of such a smart tube. In particular, we report on the implementation and testing of a vacuum-compatible step motor assembly and the corresponding motor control, for shifting the slow wave structure. We also report on the design and implementation of a pressure control device for the nitrogen in the spark gap, in order to adjust the cathode voltage. We finally report on the design and simulation of an iterative learning controller which automatically adjusts the pressure and shifting in order to achieve frequency agility.