Publications

Abstract not provided.

Pulsed-power systems operating in the terawatt regime must deal with large electron flows in vacuum transmission lines. In most parts of these transmission lines the electrons are constrained by the self-magnetic field to flow parallel to the conductors. In very low impedance systems, such as those used to drive Z-pinch radiation sources, the currents from multiple transmission lines are added together. This addition necessarily involves magnetic nulls that connect the positive and negative electrodes. The resultant local loss of magnetic insulation results in electron losses at the anode in the vicinity of the nulls. The lost current due to the magnetic null might or might not be appreciable. In some cases the lost current due to the null is not large, but is spatially localized, and may create a gas and plasma release from the anode that can lead to an excessive loss, and possibly to catastrophic damage to the hardware. In this paper we describe an analytic model that uses one geometric parameter (aside from straightforward hardware size measurements) that determines the loss to the anode, and the extent of the loss region when the driving source and load are known. The parameter can be calculated in terms of the magnetic field in the region of the null calculated when no electron flow is present. The model is compared to some experimental data, and to simulations of several different hardware geometries, including some cases with multiple nulls, and unbalanced feeds. © 2006 American Institute of Physics.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Abstract not provided.

Study of the triggered plasma opening switch (TPOS) characteristics is in progress via an ion current collection diagnostic (ICCD), in addition to offline apparatus. This initial ion current collection diagnostic has been designed, fabricated, and tested on the TPOS in order to explore the opening profile of the main switch. The initial ion current collection device utilizes five collectors which are positioned perpendicularly to the main switch stage in order to collect radially traveling ions. It has been shown through analytical prowess that this specific geometry can be treated as a planar case of the Child-Langmuir law with only a 6% deviation from the cylindrical case. Additionally, magnetostatic simulations with self consistent space charge emitting surfaces of the main switch using the Trak code are under way. It is hoped that the simulations will provide evidence in support of both the analytical derivations and experimental data. Finally, an improved design of the ICCD (containing 12 collectors in the axial direction) is presently being implemented.

Abstract not provided.

Abstract not provided.

Abstract not provided.

The Z machine at Sandia National Laboratories (SNL) is the world's largest and most powerful laboratory X-ray source. The Z Refurbishment Project (ZR) is presently underway to provide an improved precision, more shot capacity, and a higher current capability. The ZR upgrade has a total output current requirement of at least 26 MA for a 100-ns standard Z-pinch load. To accomplish this with minimal impact on the surrounding hardware, the 60 high-energy discharge capacitors in each of the existing 36 Marx generators must be replaced with identical size units but with twice the capacitance. Before the six-month shut down and transition from Z to ZR occurs, 2500 of these capacitors will be delivered. We chose to undertake an ambitious vendor qualification program to reduce the risk of not meeting ZR performance goals, to encourage the pulsed-power industry to revisit the design and development of high- energy discharge capacitors, and to meet the cost and delivery schedule within the ZR project plans. Five manufacturers were willing to fabricate and sell SNL samples of six capacitors each to be evaluated. The 8000-shot qualification test phase of the evaluation effort is now complete. This paper summarizes how the 0.279 x 0.356 x 0.635-m (11 x 14 x 25-in) stainless steel can, Scyllac-style insulator bushing, 2.65-{micro}F, < 30-nH, 100-kV, 35%-reversal capacitor lifetime specifications were determined, briefly describes the nominal 260-kJ test facility configuration, presents the test results of the most successful candidates, and discusses acceptance testing protocols that balance available resources against performance, cost, and schedule risk. We also summarize the results of our accelerated lifetime testing of the selected General Atomics P/N 32896 capacitor. We have completed lifetime tests with twelve capacitors at 100 kV and with fourteen capacitors at 110-kV charge voltage. The means of the fitted Weibull distributions for these two cases are about 17,000 and 10,000 shots, respectively. As a result of this effort plus the rigorous vendor testing prior to shipping, we are confident in the high reliability of these capacitors and have acquired information pertaining to their lifetime dependence on the operating voltage. One result of the analysis is that, for these capacitors, lifetime scales inversely with voltage to the 6.28 {+-} 0.91 power, over this 100 to 110-kV voltage range. Accepting the assumptions leading to this outcome allows us to predict the overall ZR system Marx generator capacitor reliability at the expected lower operating voltage of about 85 kV.

Abstract not provided.

To meet or exceed the 26-MA goal for ZR, the refurbished upgrade to the Z machine at Sandia National Labs, the existing Marx generator capacitors must be replaced with identical size units but with twice the capacitance. Before the six-month shut down and transition from Z to ZR occurs in late 2005, most of the 2500 capacitors must be delivered for acceptance testing and installation. We chose to undertake an ambitious vendor qualification program to reduce the risk of not meeting ZR performance goals, to encourage the pulsed-power industry to revisit the design and development of high energy discharge capacitors, and to meet the cost and delivery schedule within the ZR project plans. Five manufacturers were willing to fabricate and sell Sandia samples of six capacitors each to be evaluated. In addition, four more samples of modified or alternate designs were submitted for testing at the vendors' expense, giving us a total of 45 capacitors to test. The 8,000-shot qualification test phase of the effort is now complete. This paper summarizes how the 0.279×0.356×0.635-m Scyllac-style 2.6-μF, <30-nH, 100-kV, 35%-reversal capacitor lifetime specifications were determined, briefly describes the nominal 260-kJ test facility configuration, presents the test results of the most successful candidates, and provides procurement strategy and acceptance testing protocols that balance available resources against performance, cost, and schedule risk.

Optimizing the design of the upgrade to the Z pulser at Sandia National Laboratories renewed interest in the ubiquitous Scyllac-cased capacitor. For the Z upgrade, the desired capacitance value in each case is different than those built before, and double that of the existing units in Z. The cost and fundamental importance of the Marx capacitors in pulsers like Z prompted the decision to build a test facility that could evaluate sample units from capacitor manufacturers. The number of interested vendors and the expected lifetime indicated about 350 thousand capacitor-shots for the capacitors in a plus-minus configuration. The project schedule demanded that the initial testing be completed in a few months. These factors, and budget limitations, pointed to the need for a system that could test multiple pairs of capacitors at once, without a full-time attendant. The system described here tests up to ten pairs of 2.6 μF capacitors charged to 100 kV in 90 seconds, then discharged at 150 kA and 35 percent reversal. Unattended operation requires sophisticated fault detection, and so much attention has been paid to this. This paper will describe the system, and the key components including the control system, the switches, and the load resistors. The paper will also show some lifetime and performance data from commonly used 200 kV spark gap switches.

This research consisted of testing surface treatment processes for stainless steel and aluminum for the purpose of suppressing electron emission over large surface areas to improve the pulsed high voltage hold-off capabilities of these metals. Improvements to hold-off would be beneficial to the operation of the vacuum-insulator grading rings and final self-magnetically insulated transmission line on the ZR-upgrade machine and other pulsed power applications such as flash radiograph and pulsed-microwave machines. The treatments tested for stainless steel include the Z-protocol (chemical polish, HVFF, and gold coating), pulsed E-beam surface treatments by IHCE, Russia, and chromium oxide coatings. Treatments for aluminum were anodized and polymer coatings. Breakdown thresholds also were measured for a range of surface finishes and gap distances. The study found that: (1.) Electrical conditioning and solvent cleaning in a filtered air environment each improve HV hold-off 30%. (2.) Anodized coatings on aluminum give a factor of two improvement in high voltage hold-off. However, anodized aluminum loses this improvement when the damage is severe. Chromium oxide coatings on stainless steel give a 40% and 20% improvement in hold-off before and after damage from many arcs. (3.) Bare aluminum gives similar hold-off for surface roughness, R{sub a}, ranging from 0.08 to 3.2 {micro}m. (4.) The various EBEST surfaces tested give high voltage hold-off a factor of two better than typical machined and similar to R{sub a} = 0.05 {micro}m polished stainless steel surfaces. (5.) For gaps > 2 mm the hold-off voltage increases as the square root of the gap for bare metal surfaces. This is inconsistent with the accepted model for metals that involves E-field induced electron emission from dielectric inclusions. Micro-particles accelerated across the gap during the voltage pulse give the observed voltage dependence. However the similarity in observed breakdown times for large and small gaps places a requirement that the particles be of molecular size. This makes accelerated micro-particle induced breakdown seem improbable also.

Abstract not provided.

Abstract not provided.