As a design and analysis aid for the development of an experimental TWT, a computer program is written which allows the small-signal gain to be computed for various operating conditions, such as various conditions of tube bias (beam voltage and current) and frequency. In order to arrive at a value for the gain, a number of parameters need first to be defined or calculated. Using the method (Approach II) of Jain and Basu [17] which is applicable to a helix with a free-space gap between it and circular dielectric support rods surrounded by a metal shell, the dielectric loading factor (DLF) for the structure is found and the dispersion relation then solved to obtain the radial propagation constant y and axial propagation constant B. The method is tested for a helix with measured data and found to be acceptably accurate. Helix losses are calculated for the low-loss input and output sections of the helix, using the procedures developed by Gilmour et al [14,18], from which values are found for the helix loss parameter d. Another value for d, obviously much larger, is also found for the lossy attenuator section of the helix. Here measured data for the attenuator is used as a basis for a polynomial which models the attenuator loss as a function of frequency. The Pierce gain parameter C is found using the well-known equations of Pierce [21,22,26], and then the space-charge parameter Q. Here knowledge of the space-charge reduction factor F is required to find Q, and a simple non-iterative method is presented for its calculation, with some results. From the other parameters already calculated the velocity parameter, b, is then found. since sufficient information is now available, the electronic equations are solved. These equations are in a modified form, better accounting for the effects of space-charge than the well-known standard forms. Results are compared and slight differences found to exist in the computed gain. Now that the x's and y's (respectively the real and imaginary parts of the complex propagation constants for the slow and fast space-charge waves) are known the launching loss can be calculated. Launching losses are found for the three space-charge waves, not just for the gaining wave. The gain of the TWT is not found from the asymptotic gain equation but from a model which includes the effects of internal feedback due to reflections at the ports and attenuator. Values of reflection coefficients are modelled on the results of time-domain measurements (attenuator) and found by calculation (ports). This model permits the unstable behaviour of the tube to be predicted for various conditions of beam current and voltage and anticipates the frequencies at which instability would be likely. Results from simulations are compared with experimental observations. The need to pulse the experimental tube under controlled conditions led to the development of a high-voltage solid state pulse modulator providing regulated output pulses of up to 5000V and 200mA directly, without the use of transformers. The pulse modulator design embodies two unusual features a) its operation is bipolar, delivering positive or negative output pulses, depending only on the polarity of the rectifier input, and b) the use of multiple regulating loops and stacked pass elements to achieve high-voltage operation. Some results are presented. / Thesis (Ph.D.)-University of Natal, Durban,1994.

The stresses and deformations in cross-ply composite tubes subjected in circumferential temperature gradients are studied. The motivation behind the study is the anticipated use of composite tubes in space structures where the tube is exposed to the heat of the sun on one side and the cryogenic temperatures of space on the other. Experiments were performed to measure the functional form of the temperature gradient and the displacements. It was found that the form of the temperature gradient, T(Ɵ), can accurately be represented by T(Ɵ) = A + BcosƟ¸ and that the displacement of the tube is parabolic in the axial coordinate. Two types of analytical solutions were developed: an exact elasticity U solution and an approximate solution. The approximate solution includes a linear variation of the material properties with temperature and uses the principle of complementary virtual work in conjunction with a Ritz approximation on the stress field. The elasticity solution predicts that high tensile stresses could crack the matrix. The effect of including temperature-dependent material properties is to reduce the circumferential dependency of the stresses. / Master of Science

To provide a basis for the development of interaction curves for tubular beam-columns of annular cross section, a general purpose beam-column computer program is developed, and used to determine ultimate load capacities. The paper presents the analytical model and the computer method. The analytical results are compared with published test data as well as experimental data obtained as part of this project.