Hold-off Voltage Research Articles

Improved CuSO4 HV pulse divider

The response characteristics of a two-stage CuSO4 solution ‘‘tube’’ resistive divider placed in another CuSO4 solution ‘‘tube’’ in a concentric cylindrical configuration for measuring high-voltage nanosecond pulses have been studied. The test results showed that it was capable of maintaining pulse rise time less than 10 ns, and the hold-off voltage for pulses with rise times less than 10 ns was 600 kV.

X-ray emission and prebreakdown currents in plain and dielectric bridged vacuum gaps under DC excitation

X-ray activity in both plain and solid insulator bridged-vacuum gaps was investigated under DC stresses. The gap conditions under which X-ray emission takes place, and the correlation between X-ray activity and predischarge current were studied. Steady X-ray activity was observed for a plain gap with electrodes subjected to repeated breakdowns, and for gaps bridged with insulators having low secondary emission yields. The hysteresis in the V-I characteristics of both plain and bridged vacuum gaps is attributed to a regenerative feedback process between X-ray and photoemission processes. The nonresistive component of predischarged current for bridge vacuum gaps obeys the Fowler-Nordheim theory only in the high-field region (>60 kV/cm), and only in the absence of X-ray activity. Chromium oxide coatings on Wesgo AL-300 alumina ceramic gave a substantial improvement (>100%) in the voltage hold-off, but at the expense of increased prebreakdown current and X-ray activity. It has been shown that chromium oxide coatings do not improve the voltage hold-off unless a dense, low porosity, surface is produced by the coating.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

In-situ insulator surface charge measurements in dielectric bridged vacuum gaps using an electrostatic probe

Surface charge measurements on alumina and polymer insulators were carried out after stressing them with DC voltages in a high vacuum. The order of magnitude of surface charge density was found to be the same for materials with supposedly widely varying secondary-emission yields. Surface coatings on alumina insulators reduced charge accumulation because of increased surface conductivity and/or reduced secondary-emission yield, which led to significant improvement in voltage hold-off for alumina ceramics. Removing the cathode triple junction from the main body of the cylindrical insulator, reducing the X-ray activity in the gap, or relieving the stress at the critical junction did not significantly alter the surface charge characteristics of cylindrical insulators. Wet hydrogen firing of plain alumina reduced the voltage hold-off by 25% without altering the surface charge density. It is postulated that the charging of insulators in bridged vacuum gaps with DC stresses is due to internal secondary emission produced by ionization of the lattice in the surface layer of the insulating material, by primary electrons injected at the cathode triple junction. This mechanism of charge production differs from the current models where charging is believed to occur due to electrons hopping along the surface/vacuum interface.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Characteristics of a laser triggered spark gap using air, Ar, CH4, H2, He, N2, SF6, and Xe

A KrF discharge laser (248 nm) has been used to laser trigger, by volume preionization, a spark gap switch (38–65 kV, &amp;gt;10 kA, 100 ns pulse duration) filled with 20 different gas mixtures using various combinations of air, Ar, CH4, H2, He, N2 SF6, and Xe. A pulsed laser interferometer is used to probe the spark column. Characteristics studied include the internal structure of the column, the arc expansion rate, and evidence of any photoionization precursor effect. Our results show that the rate of arc expansion varies depending on the average molecular weight of the mixtures. In this experiment, pure H2 has the highest rate (≊9.5×105 cm/s) and air has one of the lowest (≊7×105 cm/s) for the same hold-off voltage. A computer model of the spark column formation is able to predict most of the structure observed in the arcs, including the effect of mixing gases with widely different molecular weights. The work suggests that, under proper circumstances, the spark gap switch performance may be improved by using gases lighter than conventional switch gases such as SF6.

Degradation Due to Wet Hydrogen Firing on the High-Voltage Performance of Alumina Insulators in Vacuum Applications

The HV characteristics of plain and doped alumina ceramics fired under air and air plus wet hydrogen conditions were tested in a uniform field vacuum gap using dc excitation. Wet hydrogen firing of doped alumina ceramics considerably degraded voltage hold-off as compared to air firing. Baking hydrogen-fired samples in vacuum did not improve their performance and the V-I curves remained unaltered. Gap currents measured for plain and doped air-fired samples were essentially the same (~1 to 10 nA), however, for wet hydrogen-fired samples the currents were much higher for doped than for undoped alumina samples. In general, the non-resistive current component obtained from the V-I curves was found to follow the Fowler-Nordheim equation only in the high-field region (> 60 kV/cm) for air-fired alumina samples. Processing the wet-hydrogen-fired samples by subsequent firing in an atmosphere of dry hydrogen revived the voltage hold-off of the doped alumina ceramics; the currents being comparable to that of wet-hydrogen-fired samples in the low voltage range (< 50 kV).

Application of a particle simulation to modeling commutation in a linear thyratron

Commutation is the period during which a thyratron makes the transition between an open state and a conducting state. During commutation the voltage across the thyratron decreases from the holdoff voltage to the conduction value; 10’s of kV to &amp;lt;100 V. The time for commutation, as measured by the anode voltage fall time, is usually 10–100 ns, a value which depends on the holdoff voltage, internal gas pressure, and grid geometry. In this paper, a model for commutation in a thyratron is described and its results are compared to experiment for a thyratron having a linear geometry. The model uses a Monte Carlo particle simulation for electrons and a continuum fluid representation for ions. A particle multiplication and renormalization scheme is used in the model to simulate electron avalanche so that only a moderate number of particles (4000–12000) need to be used. A modified null cross-section technique is used to account for large changes in the density of electron collision partners as a function of position or time. A model for the external circuit enables simulation of current and voltage. Results from the model for these quantities agree well with experiment, and indicate that commutation occurs in two stages. A survey study contrasts the tradeoff between high-voltage holdoff and switching speed.

Field-enhancement calculations for a field-distortion triggered spark gap

We present the results of numerical field calculation which supplement a recent article in which we described a new design concept for field-distortion triggered spark gaps. The calculations verify the shielding and field enhancement assumptions made in the article, and they provide insight into the interaction of the design tradeoffs associated with simultaneously maximizing the holdoff voltage and the triggering capability of the gap.

The Effect of Doping on the Voltage Holdoff Performance of Alumina Insulators in Vacuum

Alumina ceramics doped with various combinations of Cr, Mn, and Ti were fabricated into hollow cylindrical insulators, then assembled into vacuum diodes and tested with 25 μs voltage pulses. Four doped alumina ceramics with total dopant levels of 2.6 to 4.7% wt and partial Cr plus Ti dopant levels of 1.0 to 2.4% performed very well. They had voltage holdoffs significantly better (25 to 40%) than plain alumina ceramics. Ceramics with total dopant levels of 5.3 to 5.6% performed only slightly better than plain alumina. Ceramics with partial Cr plus Ti dopant levels of 3.8 to 4.0% had significantly improved voltage holdoff capabilities, but undesirably high permanent failure rates (19 to 33%). Ceramics with total dopant levels greater than 7% performed very poorly, nearly all suffered permanent damage from breakdowns. These results suggest that desirable upper limits for doping are 5% for total dopants and 3% for Cr and Ti combined. Thus, doping a 94% Al203 alumina ceramic with appropriate amounts of Cr, Mn, and Ti can significantly improve the voltage holdoff capability, of insulators fabricated from such ceramics. These doped ceramic insulators are suitable for use in ultrahigh vacuum applications.

KrF-laser-triggered switching of a multi-line pulsed power system

Describes the switching aspects of a 200 GW pulsed power generator which is used to power a high-energy gas laser system. Four 5 Omega water-filled pulsed-forming lines are synchronously switched with a jitter of approximately 1 ns using a few mJ from a high-brightness KrF laser. The switches are high-pressure SF6-insulated spark gaps which achieve a hold-off voltage of 1 MV, a peak current of 105 A and a peak di/dt of 1013 A s-1.

Subnanosecond high-voltage two-stage resistive divider

A simple, easily constructed, two-stage CuSO4 solution ‘‘tube’’ resistive divider for measuring high-voltage ns pulses has been developed. The test results for the divider showed that it was capable of maintaining pulse rise times less than 1.5 ns, and the hold-off voltage for pulses with rise times less than 5 ns was above 120 kV.

Field Shaping to Improve Vacuum Diode Voltage Holdoff

In our work we need a small vacuum diode which will reliably hold off a pulse voltage in excess of 150 kV. These diodes typically have an interelectrode spacing of 6 to 10 mm with a glass insulator length 3 to 4 times the interelectrode spacing. We found that the overall holdoff voltage for a diode with a fixed insulator and interelectrode gap length was strongly affected by the position of the interelectrode gap. The highest voltage holdoff was obtained when the interelectrode space was at the negative end of the insulator. Experiments were made with two sizes of diodes, varying the location of the interelectrode space from the negative end to the positive end. The large diode high voltage holdoff ranged from 240 kV with the interelectrode space at the negative end to a minimum of 50 kV in the middle and returning to a 180 kV with the interelectrode space at the positive end. The smaller diode high voltage holdoff ranged from a high of 200 kV with the interelectrode space at the negative end to 130 kV with the interelectrode space at the positive end. The minimum at the middle was not as pronoucned in this diode. Since most of the HV breakdowns were across the glass insulator, electron avalanching was suspected. Computer-generated field plots confirmed this possibility.

Voltage Recovery Time of Small Spark Gaps

A two-pulse method is used to determine how fast and to what degree a small spark gap can recover its voltage holdoff capability after breaking down. The first pulse is used to overvolt and break down the gap. The second pulse is used, after a time delay, to determine the voltage recovery of the gap. By varying the time delay to the second pulse, a recovery voltage versus time plot can be obtained. Time delays from 10 ?s to 100 ms have been recorded. The spark gap discharges millijoules of energy with a gap spacing of less than 1 mm. Recovery has been measured at breakdown voltages of up to 10 kV in argon, hydrogen, and a mixture of the 2 gases. The experimental setup, pulse circuits, and data collection methods are described. Percent voltage recovery versus time plots for various parameters (gas species, gap spacing, and pressure) are discussed.

Pulsed power switching using saturable core inductors

We have investigated some of the problems involved in switching fast, high-power pulses using single-turn saturable core inductors constructed of Metglas 2605SC. These include hold-off voltages and times, output pulse risetimes, B-H curves of the material under fast pulse conditions, and propagation of the saturation field through the core. Experiments were done at Sandia on the RAMSES I accelerator which uses a 200-kV Marx generator to provide a 200-ns pulse to a 1 Ω coaxial water transmission line.

Improving the voltage holdoff performance of alumina insulators in vacuum by quasimetallizing or doping

Treatment of the surface of an alumina insulator with coatings incorporating varying amounts of Cr, Mn, and Ti can significantly increase the vacuum voltage holdoff capability of the insulator (up to 25%). During processing (quasimetallizing) the coating penetrates into the alumina, so it is insensitive to mechanical damage. This quasimetallizing treatment is also compatible with subsequent metallizing and brazing of the alumina insulator. Quasimetallizing with appropriate formulations decreases both the surface resistivity of the alumina and the secondary electron emission yield of the alumina. Each change improves the voltage holdoff characteristics of the alumina. Similar improvements in voltage holdoff capability can be obtained by doping the plain alumina to the 5% (by weight) level with additives of 4 parts Mn/1 part Ti or 2 Mn/1 Ti/1 Cr. Such doped ceramics have the advantage of requiring fewer processing steps than do the quasimetallized ceramics. Both doped and quasimetallized ceramics are suitable for use in ultrahigh vacuum apparatus.

Energetic breakdowns and voltage hold-off between copper electrodes in vacuum

In high power neutral beam sources the accelerator structure must hold voltages as high as possible. On the other hand, there is always in the sources a stray capacitance driving energetic breakdowns that can affect the voltage hold-off of the system. This work is devoted to the study of these two points, with copper as a material for the electrodes.

Improving the Voltage Holdoff Performance of Alumina Insulators in Vacuum Through Quasimetallizing

Treatment of the surface of an alumina insulator with coatings incorporating varying amounts of Cr, Mn, and Ti can increase the vacuum voltage holdoff capability of the insulator significantly (up to 25%). During processing (quasimetallizing) the coating penetrates into the alumina, making it insensitive to mechanical damage. This quasimetallizing treatment is also compatible with subsequent metallizing and brazing of the alumina insulator. A 7/1 Mn/Ti mix performed very well, being found to be as effective on a 94% A12O3 alumina as on the previously investigated 95% A12O3, 1% Cr2O3 alumina. Mixes of 6/1/1 Mn/Ti/Cr and 6/3 Mn/Cr performed about as well as the 7/1 Mn/Ti mix, but no better. Quasimetallizing with pure Mn improved the voltage holdoff capability of alumina by about half as much as when using the 7/1 Mn/Ti mix. Mixes with relatively high titanium content (4/3 Mn/Ti and 3/3/2 Mn/Ti/Cr) significantly increased the voltage holdoff capability of the alumina, but unfortunately were much more prone than the 7/1 Mn/Ti mix (or plain alumina) to suffer severe and permanent damage when a breakdown did occur. Quasimetallizing with appropriate formulations. has been shown to change the surface characteristics of alumina in two ways: (1) it decreases the surface resistivity of the alumina, and (2) it decreases the secondary electron emission yield of the alumina. Each change improves the voltage holdoff characteristics of the alumina.

High-Pressure Surface-Discharge Plasma Switches

Surface-discharge plasma switches operating at high gas pressures with gas discharge laser and resistive loads have been the subject of empirical investigations in recent years. The particular interest has been in the low-inductance uniform multichanneling evident in these switches, in conjunction with the high hold-off voltages that have been observed at multiatmosphere operating gas pressures. This paper will review the progress to date and present a theoretical model of the surface-discharge plasma switch that explains the observed data.

Multichannel surface spark gaps

A study has been undertaken on high-pressure surface discharge switches potentially capable of moderate repetition rate operation. The parametric experiments reported were carried out utilizing the gaps as transfer switches, under pulse charging conditions, between several types of low-impedance transmission lines and a high-pressure rare-gas halide laser discharge, acting as the load. The effects of spark gap internal geometry, gas composition, and controlled changes in the laser load, upon gap multichanneling, closure simultaneity, and peak holdoff capability are discussed. These surface gaps, of length 66 cm, reliably close 19 channels per side (29 per meter) with a holdoff voltage greater than 120 kV and a closure simultaneity of ∼ 2 ns for the first 500 shots, increasing to ∼ 5 ns and remaining there for 10 000 shots, the test limit to date. Preliminary results at higher charging voltages have yielded intense multichanneling with holdoff voltages in excess of 210 kV.

Development of a 100-kV multimegawatt repetition rate gas switch

A 100-kV gas switch has been developed and tested which is capable of controlling 5-MW average power when operated up to 250- pps repetition rate. Recovery of the switch voltage holdoff capability after each discharge was accomplished by providing both a 1-ms grace period during which no voltage is reapplied, and by continuously purging the switch with 40-psig pressurized air at flow rates up to 60 SCFM. The switch was tested using a simulation technique in which the switch was subjected to the same repetitive peak voltage and current as it would in controlling several megawatts of average power. Limits of switch performance as a function of air flow rate and peak voltage have been established.

Pulsed ferrite core tests for 50-ns linear induction accelerator

The Lawrence Livermore Laboratory undertook an investigation of the properties of ferrite materials to be used in a 5-MeV, 50-ns linear induction accelerator. The investigation, on a part-time basis, lasted about one year and had the cooperation and helpful suggestions of several manufacturers: TDK of Japan, Phillips of Holland, and Stackpole of the U.S.A. Ferrites have been widely used as tuning cavities for proton synchrotron accelerators at radio frequencies. In such an application, the μQf factor is used in describing the figure of merit for ferrites where a high duty factor requires low loss ferrites. In our linear induction accelerator with an average rep-rate of 5-Hz, the ferrite losses are negligible and the concept of complex permeability in describing the losses will not be introduced, but a large signal ΔB/ΔlH will be used to describe their properties. The properties of interest in designing the accelerating cavity were: a) flux swing ΔB = B r + B m > .5T b) a residual flux density B_{r} \geq .15T with a reset no greater than 2 Oer. c) a relatively high incremental μ > 400 to keep the excitation current small in relation to the load current, d) a high resistivity for the 250-kV voltage hold-off.