An experimental investigation of the voltage noise ("hash") and anode damage in a self-field, quasi-steady magnetoplasmadynamic thruster (MPDT) operating above onset is presented. "Onset" refers to an operating mode of the MPDT that occurs at high currents J and low mass flow rates m˙, or equivalently at high J2/m˙, which is characterized by noisy, unrepeatable operation and which represents a performance ceiling for this particular plasma thruster.;Careful measurements of the thruster voltage are made while operating the thruster over a range of J2/m˙ values and with three anode materials---lead, copper, and graphite. It is observed that the voltage hash is a random process, without a characteristic time scale, whose statistics evolve with J2/ m˙, without a significant dependence on the anode material. The statistics of the hash---particularly, the first four moments of the probability density---are Gaussian at low J2/ m˙, and evolve away from Gaussian until J 2/m˙ ∼ 110 kA2-s/g, after which they return to Gaussian. Using an anode spot model, the reversal of the statistics trends at this value of J2/ m˙ is interpreted as a result of the formation of many anode spots. It is suggested that many anode spots are formed due to a filamentation instability which causes the current to fragment into many channels, irrespective of the anode material.;The damage caused by anode spots on the MPDT anode is investigated by analyzing the marks left behind on anodes subjected to operation at high J2/m˙. It is found that no characteristic size of damage marks exists, but that the sizes self-organize into a power-law distribution, such that the probability of spot damage of size d ∼ 1/ddelta, where delta ≥ 1. The severity of the damage is observed to be primarily dependent on the anode thermal properties and not J2/m˙. Lead anodes, for example, show spot damage at all J2/ m˙ values, while graphite anodes show no significant damage at any J2/m˙. Two insights follow: one, that the current conduction to the MPDT anode is, to some extent, spotty at all current levels, though it is for the most part diffuse at low J2/m˙; and two, that the J2/m˙ value at which hash appears, and that at which significant damage appears, need not be the same if anode materials are properly chosen. The anode material is primarily a passive player in the discharge, responding to the action of anode spots but not significantly influencing their properties.;Optical diagnostics are used to study the effect of anode vapor on the voltage hash. It is shown that while hash of large magnitude can be accompanied by erosion, the hash is suppressed when the anode vapor is copious enough to bridge the gap between the anode and cathode. It is also shown that this effect is unsteady, which causes non-stationarity in the hash statistics. This concurs with the previous conclusions that the behavior of hash is not governed strictly by erosion, or vice versa.;Experiments with a constrained-attachment anode, which allows the arc to attach in only 0.5 mm discrete spots, show that the arc may be stabilized by forcing its attachment on small anode areas. It is found that the arc attachment stabilizes preferentially on the inner anode face, accompanied by a quiet voltage and significant anode erosion. Examination of the thruster behavior suggests that attachment stabilization precedes erosion. It is concluded that either voltage hash, or anode erosion, may be eliminated using the present constrained-attachment technique, and that eliminating both simultaneously may be accomplished with an appropriately refractory anode material.
展开▼
