Optimizing compact Marx generator networks for driving capacitive loads

Summary form only given. Small linear Marx generators are low-energy systems that are frequently constructed in close-fitting metal housings. In these compact systems the parasitic capacitance between network components and the enclosure can be relatively large and play an important role in the Marx network performance. The parasitic capacitance and inter-stage inductance form the components of a lumped-constant transmission line, which facilitates proper sequential firing of the spark switches.When the total parallel parasitic capacitance exceeds the Marx capacitance, two thirds or less of the stored energy may be transferred to an external load capacitor if driven directly by the Marx generator. In addition to this reduction in transfer efficiency, the remaining energy induces high frequency oscillations in the Marx circuit that will lead to component heating and early failure. These problems can be overcome by suitable choices of Marx-network component values, switch timing, and external network components. In fact, even with relatively large parasitic capacitance values, ideal theoretical solutions are found for which energy transfer efficiencies of unity are achieved with lossless network components. Essential to finding ideal solutions are theorems governing waveform symmetries, reciprocity, and restrictions on component values derived from principles of conservation of energy and charge. This intrinsically time-domain problem is then recast in the frequency domain where the network resonant frequencies must be arranged with prescribed harmonic relationships. These methods are extended to more complete network models than those reported earlier. Examples of ideal solutions are presented with PSpice simulations.