J.M. Elizondo

Development/tests of 6-MV triggered gas switches at SNL

Gas switch development for application in Z refurbishment (ZR) has continued since PPPS Conference 2001. Several iterations have been tested both in oil and water dielectrics. The first switch tested was an evolved version of the Sandia designed HERMES III switch [G. J. Denison et al., 1985]. The 6 MV ZR baseline switch consists of a self-breakdown (cascade) section in which the discharge current flows in several parallel channels, and a trigger section where the current flows through a single spark channel. The second switch tested is a cantilevered switch. The entire cascade section is cantilevered on a plastic support rod allowing for a single continuous insulator housing but having the same characteristics as the baseline. The third switch tested is a Sandia hybrid switch. The laser triggered section of the hybrid switch includes 5 cascade like electrodes connected in parallel to a triggered gap via a /spl sim/2 /spl mu/H series isolation inductor. The discharge current in the hybrid switch trigger section flows in several parallel channels eliminating the single channel flow as in the baseline switch trigger section. The design iterations of these switches and results of these tests are presented.

The ZR refurbishment project

ZR is a refurbished (R) version of Z aiming to improve its overall performance, reliability, precision, pulse shape tailoring and reproducibility. Z, the largest pulsed power machine at Sandia, began in December 1985 as the Particle Beam Fusion Accelerator II (PBFA II). PBFA II was modified in 1996 to a z-pinch driver by incorporating a high-current (20 MA, 2.5 MV) configuration in the inner /spl sim/ 4.5 meter section. Following its remarkable success as a z-pinch driver, PBFA II was renamed Z in 1997. Currently Z fires 170 to 180 shots a year with a peak load current of the order of 18-20 MA. The maximum z-pinch output achieved to date is 1.6 MJ, 170 TW radiated energy and power from a single 4 cm diameter, 2 cm tall array, and 215 eV temperature from a dynamic hohlraum. ZR in turn will, operating in double shift, enable 400 shots per year, deliver a peak current of 26 MA into a standard 4 m /spl times/ 2 m Z-pinch load, and should provide a total radiated X-ray energy and power of 3 MJ and 350 TW, respectively, achieve a maximum hohlraum temperature of 260 eV, and include a pulse-shaping flexibility extending from 100 s to 300 s for equation of state and isentropic compression studies. To achieve this performance ZR will incorporate substantial modifications and upgrades to Marx generator, intermediate store capacitors, gas and water switches, water transmission lines and the laser triggering system. Test beds are already in place, and the new pulsed power components are undergoing extensive evaluation. The Z refurbishment (ZR) will be operational by 2006 and cost approximately

Circuit modeling for ZR

60M.ZR is a refurbished (R) version of Z aiming to improve its overall performance, reliability, precision, pulse shape tailoring and reproducibility. Z, the largest pulsed power machine at Sandia, began in December 1985 as the Particle Beam Fusion Accelerator II (PBFA II). PBFAII was modified in 1996 to a z-pinch driver by incorporating a high-current (20-MA, 2.5-MV) configuration in the inner ∼ 4.5 meter section. Following its remarkable success as z-pinch driver, PBFA II was renamed Z in 1997. Currently Z fires 170 to 180 shots a year with a peak load current of the order of 18–20 MA. The maximum z-pinch output achieved to date is 1.6-MJ, 170-TW radiated energy and power from a single 4-cm diameter, 2-cm tall array, and 215 eV temperature from a dynamic hohlraum. ZR in turn will, operating in double shift, enable 400 shots per year, deliver a peak current of 26 MA into a standard 4cm × 2cm Z-pinch load, and should provide a total radiated x-ray energy and power of 3 MJ and 350 TW, respectively, achieve a maximum hohlraum temperature of 260 eV, and include a pulse-shaping flexibility extending from 100ns to 300ns for equation of state and isentropic compression studies. To achieve this performance ZR will incorporate substantial modifications and upgrades to Marx generator, intermediate store capacitors, gas and water switches, water transmission lines and the laser triggering system. Test beds are already in place, and the new pulsed power components are undergoing extensive evaluation. The Z refurbishment (ZR) will be operational by 2006 and will cost approximately

Water switches impedance from Screamer circuit model and experimental waveform match

60M.

High-voltage, high rep-rate, low jitter, UWB source with ferroelectric trigger

In the Z Refurbishment (ZR) project, the capabilities of the Z accelerator is expanded [D. H. McDaniel et al., June 2002]. Most of the pulsed power hardware is being redesigned for operation with higher reliability, a greater shot rate, and increased pulse energy. The topology of the ZR pulsed power circuit is similar to that of Z with a few significant changes. The initial baseline circuit design for ZR was developed using Screamer [M. L. Kiefer and M. M. Widner, 1985] which has been the primary pulsed power circuit design tool employed at SNL for years. Screamer is, however, limited in the topologies it can simulate. For ZR there are several circuit features which cannot be modeled adequately in Screamer. The most significant of these is the need to allow for a high degree of independence in module firing times so that a prescribed temporally shaped pulse can be generated for isentropic compression experiments (ICE) [3]. For this reason, the baseline circuit design for ZR is now being developed using Bertha [W. N. Weseloh, 1989]. Bertha is a NRL developed T-line circuit simulation code that is not topology limited and runs fast. In this paper, a discussion of the circuit modeling strategy for the ZR project along with a discussion of the current status of the ZR baseline circuit model is presented.

A 5-Megavolt, 600-Kiloampere Laser-Triggered Gas Switch for use on Z-R: Comparison of Experiments and Simulations

Experimental waveforms obtained from the Z-20 water switch test bed at SNL, and PITHON at PSD, are matched with their respective circuit models to determine the equivalent inductance and conduction resistance of water switches. These circuit models and three water switch models, available in SCREAMER, were used to obtain a match to the experimental water switch breakdown current waveforms. In both cases the waveform match is very good and the resulting inductance and conduction resistance values consistent. The equivalent conduction resistance obtained is in the range: 0.064/spl les/R (/spl Omega//cm) /spl les/0.080, single switch per unit length of switch gap. This match was obtained, with a series of experimental waveforms, out of shots with positive and negative enhanced electrodes from both machines. Excellent experimental to model match was found, for a full series of waveforms, all within the range of resistance values. A series of waveforms are shown together with the details of the switch model and the technique used to determine the resistance values.

Requirements for optimal performance and the consequences of using toroidal shaped electrodes in multichanneling switches

The author consider here a novel trigger for high-voltage spark-gap switches that uses a ferroelectric material as an electron source. The new device is called a ferratron because it is in some ways reminiscent of a trigatron, and because it uses a ferroelectric material to inject electrons into the gap to trigger the breakdown. The device presently under advanced development uses a ferroelectric ceramic as the electron source, coupled with a high gas flow rate. This configuration allows reliable triggering at low trigger voltages, even at high repetition rates. The trigger system is integrated into a switch chamber design capable of sustaining 500 kV at a pressure of 1500 psig, The switch chamber is tailored to reduce field stresses and to provide a low inductance current path with a very compact geometry. The high gas flow rate allows the replacement of the gas in the discharge region within the time scale necessary to sustain a repetition rate of 100 Hz to 1 kHz. The switch has been successfully operated during preliminary tests using N/sub 2/ at low pressures with a transmitted risetime of 500 ps. The jitter was found to be less than 70 ps at a rep rate of 1 Hz. Faster risetimes, along with higher repetition rates and higher voltages, will be attained as development progresses.

Recent measurements of the ferratron: a high-voltage, fast risetime gas switch with a ferroelectric trigger

The Z refurbishment project is designed to increase the peak current to the load on Z to ~26 MA in a 100-ns wide power pulse. This current is achieved by summing the current from 36 independent pulse-power modules. To meet these requirements, we have designed and constructed an SF6-insulated gas switch that can hold off 5.5 MV and conduct a peak current of 600 kA for over a hundred shots. The gas switch is charged by a Marx generator in ~1 microsecond and transfers about 200- kilojoules of energy and 0.25 Coulombs of charge to a pulse-forming line in a ~150-ns-wide power pulse peaking at 2.5 TW. The gas switch consists of a laser- triggered section holding off 15% of the voltage followed by 25 self-breakdown gaps. The self-breaking gaps are designed to provide multiple breakdown arcs in order to lower the overall inductance of the switch. The gas switch is submerged in transformer oil during operation. In this work, we show how simulation and experiment have worked together, first to verify proper operation of the switch, and then to solve problems with the switch design that arose during testing.

Recent progress in ferratron measurements

In many multi-channel switch designs, toroidal shaped electrodes are used to form the cascade breakdown section of a closing switch. Advantages of using toroidal electrodes include providing evenly graded electric fields from one gap to the next during charging and shaping the electric fields away from the mechanical connections that hold the electrodes in place. Often the inductance of arc channels is considered, but the inductance of electrodes must be considered when evaluating a multi-channeling switch. In typical geometries, electrode inductances are one to two orders of magnitude larger than of the inductance of the arcs themselves and may not be neglected. It is generally assumed that achieving more channels ad infinitum is desired to produce a low switching inductance. Calculations suggest this is not necessarily the case. Calculations also suggest that channel distribution, not simply the number of channels, strongly affect the operating inductance of the switch. For example, if one gap in the cascade section does not multi- channel, the bottleneck caused by this will dictate switch inductance, even if every other gap closes with many channels. Suggestions and supporting calculations are presented for minimizing arc inductance and electrode inductance of a toroidal arrangement. There must be the same number of channels from gap to gap to minimize the effects of arc inductance. The distribution of channels must be identical from gap to gap to minimize electrode inductance and azimuthal current flow on the electrodes with respect to the axial direction of the cascade section. Future work on optimizing fast, multi-channeling switch electrodes at the University of Missouri terawatt test stand (MUTTS) will be introduced.

Progress in the development of the ferratron: a novel repetitively rated triggered spark-gap switch with ultralow jitter

We provide here an update on the status of ferratron development. The ferratron is a high-voltage, fast risetime, gas switch with low-jitter and high repetition rate. It is triggered by the emission of electrons from a ferroelectric device. It may be suitable for phased arrays due to its low jitter. We have designed and built a test chamber for measuring the output of the ferratron. The test chamber is designed to launch a fast-risetime wave into an electrically large coaxial structure. The transition from gas to oil is accomplished with a polyethylene lens with one surface being a hyperbola of revolution. The goal of the transition is to preserve the risetime of the wave as it transitions from gas to an electrically large oil coax. We provide here preliminary data on the test chamber characteristics, in terms of both TDR (time domain reflectometry) and TDT (time domain transmission). We have included in our test chamber a method of connecting the gap to a feed cable, in order to drive the gap with a known source for calibration. We demonstrate that the risetime of the test chamber in time domain transmission (TDT) mode is around 50 ps, which is sufficiently fast to measure the anticipated 100 ps risetime of the pulses to be generated by the ferratron. We also provide measurements of the switch in a hybrid trigatron ferratron mode.