Search

Microparticles from electrode erosion would lead field distortion, microdischarge, or impact damage, which are convinced to be mainly responsible for degradation of insulation performance in switches. To investigate the splashing characteristics of microparticles from electrode erosion, this paper performs the discharge experiments between two cylinder electrodes, made of copper, molybdenum, or tungsten-copper. A statistical method based on laser confocal microscopy is introduced to visualize the distribution of microparticles on the electrode surface. Experimental results indicate that the number of microparticles increases and the acceleration tendency gets mild as the discharge shots accumulate. The Cu surface would lead more production of the $> 2~\mu \text{m}$ microparticles than other two materials after hundreds of shots, which would provide an explanation for its relative poor reliability and insulation performance. The electrode surface has formed two distinct zones for distribution of these microparticles: inner zone and outer zone after hundreds of discharge shots. Estimation of inner zone radius and splashing angles of microparticles indicates that the larger the microparticles are, the smaller the angle they would be splashed out with. Finally, the evolution of arc column in the first quarter discharge cycle is captured by a high-speed camera and its relation with the concentration of microparticles is discussed. [ABSTRACT FROM AUTHOR]

In this paper, a high-voltage unipolar pulse power generator is introduced for electroporation applications. Operating principle of the proposed structure is based on the buck–boost (BB) converter. The generator is comprised of parallel modules connected to the converter through an SCR and is supplied by a low-voltage dc source. Each module consists of an insulated gate bipolar transistor (IGBT), two diodes, and a capacitor. In fact, the proposed topology is a combination of the Marx generator and BB (MG-BB) converter. In comparison with other modular MG-BB structures, the proposed topology achieves interesting advantages including simpler structure, reduction of IGBTs voltage rating, and reducing number of components (inductors and switches). In order to clarify the features of the proposed structure, it has been compared with other unipolar MGs considering different aspects including the number of circuit components, number of required input supplies, amplitude of the output pulse, and the maximum voltage across switches and diodes. To validate the effectiveness of the structure, simulation and experimental results are presented. The results confirm that the proposed topology is capable of generating unipolar pulses with desirable amplitude and repetition frequency using a low-voltage dc supply. [ABSTRACT FROM AUTHOR]

The investigation of a high output power, arc lamp exciplex ultraviolet (UV) source for pulsed power applications is presented. The arc lamp generates up to 550 W from XeF* exciplex radiation at 351 nm, totaling to nearly 0.15-mJ total radiated UV energy over the duration of the UV pulse. With an ellipsoidal reflector, the arc lamp produces 400 W/cm2 and up to 0.1 mJ of UV light onto a 1-cm2 area. A complete experimental investigation of the arc lamp for both XeCl* (308 nm) and XeF* (351 nm) exciplex sources operated under varying excitation and pressure conditions is reported. As an application, the arc lamp is successfully utilized as an illumination source for an intrinsically triggered, wide bandgap SiC photoconductive semiconductor switch (PCSS), where a PCSS ON-state resistance of $500~\Omega $ is achieved. [ABSTRACT FROM AUTHOR]

Prefire is the unexpected breakdown of gas switch at triggered mode and have to be limited at extremely low level in fast linear transformer driver (FLTD). In this paper, the dispersion of breakdown voltage as well as prefire characteristics of high pressure gas spark switches under typical working conditions of FLTD was investigated and reported. Tested switches included a three-electrode field-distortion gas switch (TFGS) and a six-stage multichannel gas switch (SMGS), which had compact structure and low inductance. The results showed that the self-breakdown voltage of both TFGS and SMGS obeyed three parameter well-bull distribution and the dispersion of Ub at ±90 kV could reach 4% of TFGS and 14% of SMGS, respectively. The dispersion of Ub was remarkably influenced by the floating potential of middle electrode. By grounding the triggering electrode via a kilohm resistor, the dispersion of Ub decreased to 2% in TFGS and 6.5% in SMGS, respectively. When a voltage equalization method by cascade resistors was introduced, the dispersion of Ub in SMGS dramatically decreased to 4.4%. In repetitive operation at triggered mode, the prefire rate of TFGS reached 0.9% at $\gamma = 80{\%}$ working coefficient and decreases with the decrease of $\gamma $ , being 0.3% at $\gamma = 72{\%}$ and 0.2% at $\gamma = 65{\%}$ , respectively. The prefire rate of SMGS was much higher, reaching 1.7% at $\gamma = 72{\%}$ and 0.7% at $\gamma = 65{\%}$ , and decreased to 0.6% and 0.3%, respectively, after applying the equalizing resistors. Generally, the prefire rate of SMGS was somewhat higher than that of TFGS with the same working coefficient and the floating potential limiting was imperative for prefire inhibition. [ABSTRACT FROM AUTHOR]

Characterization of a corona-stabilized switch in the single-shot regime, including triggering range, delay times, and jitter, is reported over the pressure range 0–3 bar gauge, as a continuation of work from similar characterization with this switch filled with SF6 with different gap spacings. When filled with mixtures of 1,3,3,3-tetrafluoropropene (HFO-1234ze) and N2, the breakdown voltage can be increased by up to ~306% and ~191% under negative and positive polarities, respectively, of that using 100% N2. These results were achieved with gas mixtures consisting of 80% N2 and 20% HFO-1234ze, by pressure. The maximum negative polarity triggering range was 13.6 kV, comparable to that achieved previously using SF6. The measured delay time and calculated jitter were generally found to increase with increasing pressure, and with increasing percentage (from 5% to 20%) of HFO-1234ze in the gas mixtures. von Laue statistical analysis of time-to-breakdown data showed that both the formative time and statistical time increased with increasing pressure, and with increasing percentage of HFO-1234ze in the gas mixtures. The formative time under negative polarity was significantly longer than that for positive polarity. The results indicate that HFO-1234ze may be considered as a suitable candidate to replace SF6 for switching applications, although there are some operational observations that require further investigation. [ABSTRACT FROM AUTHOR]

Based on the self-breakdown experiments of gas switches, the relationship between the electrode area and the switch static performance has been studied. In the experiments, the structure of the switch is fixed, but the disc electrodes of the switch are replaceable to change the electrode area. Diameters of the electrodes are, respectively, 30, 35, 40, and 45 mm. The gas pressure, gas type, and discharge circuit are constant. As the electrode diameter has grown, distribution curve of the self-breakdown voltage has been regularly changed and both average and deviation of the voltage decreased. In order to describe the experimental results, a semiempirical model based on mathematics has been proposed. The calculated self-breakdown voltage distribution curves agree well with those in experiments. [ABSTRACT FROM AUTHOR]

Based on the commercial ANSYS workbench simulation software, the fluid--structure interaction method is used to calculate the air arc pressure distribution in arc chamber, and its impact effect on the shell during interrupting the high current is simulated and analyzed. The stress and deformation distributions of the arc chamber shell with different materials and thicknesses are obtained. It is confirmed that the maximum stress of the shell with 6-mm thickness decreases by ~25% compared with that of 3-mm thickness shell, and the maximum deformation decreases by ~20%. Four models for thickening different local regions are proposed to optimize the shell structure based on the above analysis, including thickened middle regions of the wall inside/outside and corresponding edge regions of the wall. The simulations results indicate that the structure of the arc chamber shell could be optimized very well to achieve the most uniform stress/deformation distribution using the model with thickening the middle region of the outside wall. [ABSTRACT FROM AUTHOR]

Explosive current opening switches (EOSs) operate based on the principle of the conductors mechanical destruction of the inductive energy storage system by detonation products of the high explosive charge. The EOSs are used for current pulse sharpening of the magnetocumulative generators (MCG). These explosive single-acting devices constructively match each other; both operate in the physical limit and provide maximum specific characteristics. This paper describes several basic schemes of the EOS constructions that were experimentally checked in different laboratories of the world. The investigation results on the flat and cylindrical models and also full-scale experiments on the current pulses sharpening of the helical and disc MCG are presented. [ABSTRACT FROM AUTHOR]

Studies have been taken into the breakdown voltage jitter of corona-stabilized switch at high repetition rate up to several kHz. It is thought that the gas recovery affect the jitter of breakdown voltage directly at high repetition rate, because the gas switch will breakdown at a low voltage if the gas cannot fully recover. The breakdown voltage jitter is called voltage jitter in the following paper. However, based on the experimental results, the voltage jitter still exists at low repetition rate, even at single-shot mode. This is not explained by the previous studies. In this paper, voltage jitter below 100 Hz of a corona-stabilized switch is investigated experimentally. A corona-stabilized switch is introduced, using a cylindrical electrode as the cathode and a plane electrode as the anode. The gas used between the electrodes is SF6. A negative pulse with a maximum width of 30~\mu \texts is used to charge this switch. The tested factors include gas pressure, gap space, repetition rate, and switch geometry. Experimental results show that voltage jitter is mainly affected by the gas pressure and switch geometry. It is seen that voltage jitter decreases as gas pressure increases and a suitable cathode geometry presents a low voltage jitter. Critical volume is a region close to the highly stressed electrode, where the presence of an electron will lead to an electron avalanche which attains a critical size and hence leads to a stabilizing corona. Critical volume is introduced to explain the experimental results. It is found that less voltage jitter can be achieved with a smaller and more concentrated critical volume, as well as efficient corona stabilization. The results indicate that voltage jitter may be reduced by optimize the critical volume at low repetition rate. [ABSTRACT FROM PUBLISHER]

Pulsed electric field (PEF) technology is a promising nonthermal processing techniques that can be utilized to inactivate microorganisms in liquid food with high-voltage PEF. Herein, a high-voltage solid-state switch consisting of 64 insulated gate bipolar transistors (IGBTs) connected in series was designed and developed for the PEF treatment. Regarding the unbalanced sharing of voltage in series-connected IGBTs, the resistor–capacitor–diode snubber circuit was specifically used and investigated in terms of model of parameters. Furthermore, using gate drivers and optic fiber, the driving circuit and protection circuit were designed and validated. A 50-kV isolation level power supply was built in order to provide 16 independent IGBT stacks with 24 V of power each. The results show that the developed switch works adequately a delay time of 380 ns with 35.8-kV voltage and 44.8-A current capacity. Moreover, the response time of the short-circuit protection is acceptable as well with a reaction time of under 7~\mu \texts . In conclusion, the switch designed for PEF treatment of liquid food performs within set parameters and is ready for pilot-scale processing capability. [ABSTRACT FROM PUBLISHER]

The performance of an all solid-state linear transformer driver (LTD) is evaluated based on experimentally verified behavior of a single stage. The single-stage LTD utilizes a low-profile design with robust thyristor switches and high-energy-density mica capacitors to minimize overall system inductance. Subnanosecond jitter is achieved with simultaneous thyristor triggering. The stage is magnetically coupled to a secondary winding through a central nanocrystalline core. A dc current source, decoupled with a large inductance, actively resets the core between pulses. The overall result is a low-impedance ( $<1~\Omega $ per stage) pulse generator that rivals the performance of traditional Marx systems with the improved reliability, increased lifetime, and fast rep-rate capabilities of solid-state switches. The stage is tested with charging voltages up to 8 kV into various loads and compared with simulations based on an analog behavioral thyristor switch model previously developed at Texas Tech University. The simulation is expanded into a full-scale, multistage LTD simulation and compared with a previously constructed Marx generator. [ABSTRACT FROM AUTHOR]

In this paper, measurement results of the arc erosion limitation in low voltage (LV) switches are presented. The reduction of the arc erosion in the analyzed electromechanical switches was achieved through an application of the hybrid switching in the tested alternative current circuit during interruption of the inductive current. The presented method is based on a commutation of the current during its interruption into a semiconductor branch that is connected in parallel to the operated switch. Thus, this approach is able to limit the electric arc energy efficiently, and as a consequence the arc erosion of the electrical contacts is also significantly reduced. For this reason, the application of the hybrid switching has a positive influence on a lifespan and a reliability of the operated switch. In order to evaluate the influence of the hybrid switching application on an effectiveness of the limitation of the arc erosion in the analyzed LV switches, 50 000 current interruptions were conducted in the tested circuit by each of four analyzed switches. Researches were performed for two standalone switches during a standard current interruption, as well as, for two switches with applied the hybrid switching system. In order to compare the arc erosion of the contact surfaces in both considered cases, scanning electron microscope method coupled with energy dispersion spectroscopy was applied. Moreover, photos of the electric contact surfaces are presented. Additionally, frames registered by high-speed camera during current interruption are presented for both considered approaches. [ABSTRACT FROM PUBLISHER]

Pulsed-power generator based on linear transformer driver (LTD) technology has been used to demonstrate waveform control for pulsed atmospheric discharge. This kind of discharge load is known for its time-varying impedance where constant voltage usually does not result in constant current. In this paper, an LTD system has been combined with a field-programmable gate array in order to generate a flexible voltage waveform which, in turn, dictates the discharge current. The experimental results have proved that the current of a discharge load can be arbitrarily controlled, to a certain extent, by properly adjusting the applied voltage. This kind of current waveform control should have significant implications for various applications of discharge-generated plasmas. [ABSTRACT FROM PUBLISHER]

Drift-step-recovery diodes (DSRDs) are fast-opening switches capable of delivering nanosecond-scale high-voltage (HV) pulses into a load. The HV capability is achieved by stacking DSRD dies in series. In this paper, we characterize a DSRD die based on silicon epitaxial layers, which was designed and manufactured at the Soreq Nuclear Research Center. In the static characterization, we have measured the diode’s forward- and reverse-blocking voltages, and the junction capacitance. In the dynamic characterization, we have measured the peak voltage and its rise time for a single die, and up to a stack of 32 dies in series, where the stack was operated at current densities of up to $\sim \!\!1.3$ kA/cm2. The shortest rise time was 0.65 ns from a stack of five dies. An HV increase of 250 V per die was obtained. The maximum measured peak voltage was 6.09 kV with a rise time of 2.2 ns, and these results being limited by the setup capability. [ABSTRACT FROM AUTHOR]

The development and construction of self-scanning capability short-range ultrawideband (UWB) radar is forthcoming for the French Military for Defense. This paper is devoted to the presentation of the self-scanning radar principle and of the development, construction, and tests of the different blocks needed to implement the elementary part of the forthcoming radar. The elementary radiation source is composed of a microstrip line with photoconductive switches, an optical system to trigger the switches, a pulsed modulator used as bias voltage to load the line, and a UWB antenna to radiate high electric fields in a broad frequency band. The main proposed innovations are devoted to the integration of the photoconductive switches within the antenna and to the use of the pulsed modulator as bias source instead of a dc voltage source. This modulator allows considering the increase in the yield of the whole source as well as the increase in the radiation bandwidth. Experimental results are presented in order to compare the switching efficiency with a dc bias source against this pulsed modulator bias source. Finally, photoswitches triggered by an 80-ps/20- \mu \textJ laser source are able to generate monopolar or bipolar pulses in the kilovolt range with subnanosecond duration and a 33-MHz pulse repetition frequency. First radiated electric field results of the whole device are also reported. [ABSTRACT FROM AUTHOR]

Megavolt-class-triggered switches play a key role in large-induction voltage adder accelerators. This paper introduces a 2.8-MV, low-inductance, low-jitter, electrical-triggered, gas SF6 switch. The main section is a $V/n$ switch and the pretrigger section is a trigatron switch. Based on electric field simulation, experiment, and analysis, the structure and spacing of gaps, location of trigger disk, diaphragm, and switch inductance are all discussed. The primary test indicates that the switch can be operated in the range of 0.85–2.76 MV with the pressure of the SF6 changing from 0.1 to 0.9 MPa. The self-breakdown voltage is linearly related to the logarithm of the pressure, as the gap spacing and the voltage rise time are fixed. Three working modes were observed: self-breakdown, self-triggered, and ex-triggered. In the triggered modes, the switch jitter is <3 ns and the switch inductance is usually below 150 nH. [ABSTRACT FROM AUTHOR]

Bipolar junction transistors (BJTs) have been widely studied and used in nanosecond high-voltage pulse generators due to the advantages of fast switching speed and high repetitive frequency. Usually the Marx-type circuit is used to raise the peak value of output pulses. In this paper, a traditional BJT-Marx circuit using charging resistors is analyzed theoretically and experimentally. Based on the results, the structure of a circuit is changed to decrease the isolation voltage across the resistors, and diodes are used to replace charging resistors to block capacitors discharging to resistors. Experimental results showed that the three Marx generators proposed in this paper output pulses with higher voltage amplitude, and the last improved circuit outputs pulses with faster fall times at a higher efficiency. [ABSTRACT FROM AUTHOR]

For semiconductor-based pulsed power generators, short turn-ON and turn-OFF times of the employed switching elements are especially in hard switching condition of advantage. In order to enhance switching speeds for standard devices under hard switching conditions, we propose a capacitively coupled gate-boosting circuit with gate-drive voltages of up to 80 V. A low inductance test circuit has been set up to evaluate different switching elements under varying driving conditions. Beside single pulse operation, the setup allows for burst mode and continuous repetitive switching. The flexible design of the setup allows for testing metal–oxide–semiconductor field-effect transistors (MOSFETs) as well as insulated-gate bipolar transistors (IGBTs). Exemplary, we show results for a SiC MOSFET and a standard IGBT. By applying 80 V gate-drive voltage, the current rise rate of the SiC MOSFET could be increased by a factor of 3.5 up to 2.6 kA/ \mu \texts . For standard IGBT devices, we achieved an eightfold increase in current rise rate up to 4.2 kA/ \mu \texts . We verified a reproducible pulse shape for over \textrm 10^\textrm 8 pulses. In addition, the gate-drive circuit is able to drive the device in 3-MHz burst mode at 250 A peak collector current. [ABSTRACT FROM AUTHOR]

This is the second part of a two-part paper in which we present a new technique for the design of coaxial switched oscillators (SWOs). In this part, we present a model for the analysis of the signal radiated by the SWO, in both frequency and time domains. Use is made of a semianalytical two-port model relating the field radiated by the antenna and the discharge at the spark gap. Experimental validations of the model are presented and discussed. [ABSTRACT FROM AUTHOR]

The field-reversed configuration heating experiment (FRCHX) was designed to form closed-field-line magnetized target plasmas for magnetoinertial fusion and other high energy density plasma research. These plasmas are in a field-reversed configuration and are formed via a reversed-field theta pinch on an already magnetized background plasma. To extend the duration and temporal uniformity of the pinch, the capacitor bank driving the reversed-field discharge is crowbarred near the current peak. Four parallel rail-gap switches are used on the FRCHX for this application to ensure a low-inductance crowbar discharge path and to accommodate the large magnitude of the discharge current (often greater than 1 MA). Historically, parallel operation of spark gap switches in a crowbarring arrangement has often proved to be difficult due to the very low voltage present on the bank and across the switches at the time of peak current. In a low-inductance design, triggering can be further complicated by the rapid collapse of what little voltage there is across the switches as soon as the first spark gap begins conduction. This paper reports on the efforts that were made to develop a low-inductance crowbar switch for the FRCHX and to ultimately enable successful triggering and operation of the four parallel rail-gap switches used in the crowbar. The design of the low-inductance parallel switch assembly is presented first, followed by a description of the triggering scheme employed to ensure conduction of all four switches. [ABSTRACT FROM AUTHOR]

The process of triggering thyristors by an overvoltage pulse with a short rise time was investigated. Low-frequency commercial thyristors of tablet design with diameters of silicon wafers of 32–56 mm and an operating voltage of 2–2.4 kV dc were used in the experiments. An external overvoltage pulse was applied across the thyristor main electrodes, which ensured a voltage rise rate from 0.5 to 6 kV/ns within a few nanoseconds. Under such conditions, the thyristor closing process occurred due to initiation and propagation of a fast ionization front across the semiconductor structure. The time of switching the thyristor from the blocking state to the conducting state was within 200–400 ps. The thyristor-based switches operated in this mode were tested in two discharge circuits. In the first circuit, an assembly of nine 2-kV and 40-mm thyristors connected in series switched a 2- \mu \textF capacitor, which was charged to a voltage of 20 kV and to a resistive load of 0.17~\Omega . The following results were obtained: discharge current amplitude of 45 kA, maximum current-rise rate of 134 kA/ \mu \text{s} , current rise time (0.1–0.9 level) of \sim 0.4~\mu \text{s} , pulse duration (FWHM) of \sim 1~\mu \texts , and switching efficiency of 0.85. In the second circuit, the switch contained two 2.4-kV and 56-mm thyristors connected in series. A 400- \mu \textF capacitor, which was charged to a voltage of 5 kV, was switched to a resistive load of 0.026~\Omega . The following discharge parameters were obtained: discharge current amplitude of 114 kA, maximum current-rise rate of 32 kA/ \mu \texts , current rise time (0.1–0.9 level) of \sim 5~\mu \texts , pulse duration (FWHM) of \sim 20~\mu \texts , and switching efficiency of 0.96. [ABSTRACT FROM AUTHOR]

The spark gap switch is the key element in any pulsed power system, which directly determines the pulse repetitive frequency of the system. The dielectric recovery data of the spark gap play a basic role, but they have not yet been adequately studied for repetitive pulse charge conditions. This paper describes some of the factors that affect the recovery of spark gap switch for pulse-charged operation. A two-pulse technique has been used, in which the first pulse is applied to the spark gap to overvolt the gap and initiate the breakdown and the second pulse is used to determine the recovery voltage of the gap. The effects of electrode size, pulse rise time, and conduction current on the spark gap recovery performance are presented. Experiments were performed by using a double-pulse generator capable of generating 40-kV pulses with a delay of 1 ms–1 s. Spark gap electrodes of 34-, 40-, and 68-mm diameters and a point type were used in the study. The effect of voltage rise rate was studied by changing the charging capacitor to 100, 500, and 1000 nF in the circuit. An intermediate plateau was observed in the spark gap recovery curves. In addition, an increase in the output impedance resulted in a decrease in the time taken for the recovery of the spark gap. Also, the switch recovery appears to be largely dependent on the electrode shape and size compared with the pulse rise time. [ABSTRACT FROM AUTHOR]

Impact and shock wave events represent typical extreme dynamic loads of interest for accessing the response of materials and structures. Here, we focus on an intense electric current pulse as an extreme dynamic load. Experimentally, metallic samples were electromechanically loaded with currents up to 400 kA. For this purpose, a test rig containing a high-voltage pulsed power supply and high-performance switches was built. In this paper, we present an approach for modeling the response of thick metallic wire samples, with diameters on the order of millimeters, to extreme electromechanical loads sufficient for deformation and fragmentation, but not for material disintegration by wire explosion. The essential electrodynamic, thermodynamic, and mechanical aspects are considered, and a way to couple those physical regimes is suggested, which allows the use of numerical simulation based on the finite-element method to determine the material response. Simulations are employed to describe the time-dependent process of the structural and material behavior of thick wires. The focus here is on the structural mechanical behavior including bending or buckling before the onset of the wire explosion. The new approach for a simulation model is able to capture the basic experimental observations. [ABSTRACT FROM PUBLISHER]

The integration of photoconductive semiconductor switches (PCSSs) into pulsed power systems may have improvements over conventional switching, but failure modes occurring near the illuminated edge of PCSS devices between switching operations present a serious limiting factor. While these failure modes have previously been difficult to characterize, this paper proposes a theory for a potential contributing source of failure supported by observations derived from the results of an FDTD simulation on an optically excited semiconductor switch. A computational analysis combining both solid-state and electromagnetic physical models reveals what appears to be an unintentional waveguide-like behavior in the optically excited plasma-filled PCSS device in strong agreement with the documented observations in the experiment. The net effect of nonuniform charge carrier generation as a result of waveguide-like mode formation is theoretically assessed, and the additional areas of investigation into this pivotal problem area in the PCSS design are suggested. [ABSTRACT FROM AUTHOR]

Liquid dielectrics as switching media are of special interest because of their higher breakdown field strength than that of compressed gases, especially in the case of ultrafast nanosecond charging, better thermal management, and removal of eroded electrode material while circulating through the switch. The pulse-forming line (PFL) presented here is designed to operate together with our own Marx generators. It is comprised of a peaking switch with, compared with the previous design, an increased high-pressure operability up to 4000 kPa and an additional annular crowbar electrode. The geometry of the peaking switch electrodes was optimized with computational fluid dynamics simulations, in order to ensure a laminar liquid flow through the switch with flow rates up to 55 mL/s. Galden-type HT270 perfluoropolyether fluid was chosen as the liquid dielectric. The breakdown field strength showed a significant pressure dependency, which is due to the high gas solubility of Galden. At a pressure of 3950 kPa and a gap distance of 0.25 mm, a maximum breakdown field strength of 14.3 MV/cm has been reported, an increase of 360% compared with the breakdown field strength of 3.8 MV/cm at atmospheric pressure. The minimum peaking switch rise time was 220 ps; the corresponding peak voltage rise time amounted to 1.2\times 10^15 V/s. With the addition of the crowbar electrode, an ultrashort monopolar output pulse with a characteristic pulsewidth of 210 ps was achieved. First, the results of a repetitive operation of the PFL driven by an inductive Marx generator will be discussed. [ABSTRACT FROM PUBLISHER]

In this paper, the mathematical lifetime models of capacitors and spark switches are investigated and a linear transformer driver circuit model consisting of five 100 kV/MA cavities is built. The effect of random failures, switch jitter and prefire probability, on the module lifetime is investigated, respectively, by observing the influence on the capacitor lifetime and module output. Sensitivity analysis of time delay jitter of switches and prefire probability of switches on the module lifetime is carried out. The module lifetime prediction model is built by the combination of the lifetime model of key components, random failure model, and the module circuit model. Examples are taken to illustrate the process of module lifetime prediction and maintenance cycle determination. [ABSTRACT FROM PUBLISHER]

In this paper, an adjustable magnetic switch (AMS) based on splice winding is investigated numerically and experimentally. The proposed AMS has advantage of high power level, high repetitive rate capability, long lifetime achievability, and electrical parameters tunability, which meets the growing requirements of the military and industrial application of pulsed power technology. Adjustment of the electrical parameters of magnetic switch is achieved by altering the number of windings via the modification of the joint points between the AMS and the main circuit. Specifically, basic principle of the AMS is analyzed based on characteristics of the magnetic switch. Detailed design considerations are described for achieving modification of the winding’s number ranging from 1 to $M$ (positive integer larger than 1), which is limited by the insulation density, easily and reliably. In order to verify the design, two typical electrical parameters of an AMS are numerically calculated and simulated by the circuit simulation software. Accordingly, a test experiment platform was setup in our laboratory and the primary experimental study demonstrates that the AMS can successfully alter the electrical parameters with results similar to the calculation. A typical ( 1-\cos \omega t)$ pulse with rise times of 4 and 5 \mu \text{s} , a peak voltage of approximately 20 kV, and a peak current of over 1.0 kA was achieved on the capacitor load. The experimental results show reasonable agreement with the theoretical and numerical analyses. [ABSTRACT FROM PUBLISHER]

A new design of powerful high-voltage high-current impulse source for nanosecond megavolt charging of short forming lines (SFLs) is proposed. The first stage of the device is based on inductive energy storage. It accumulates energy supplied by the current of magnetic flux compression generator (MCG), whereas the electro-explosive opening switch is used for the energy output into intermediate SFL. In the second stage, we use a Tesla transformer with oil insulation. The results of the earlier experimental study of submegavolt charging of intermediate forming line as well as the results of calculations of the two-stage facility given in this paper demonstrate the feasibility of the device based on MCGs for nanosecond charging of SFL up to about 3 MV. [ABSTRACT FROM PUBLISHER]

The basic structure and work mechanism of laser-triggered vacuum switch (LTVS) are introduced briefly in this paper, and the processes of primary plasma production and expansion are analyzed as well. A series of experiments has been carried out with different gap voltages, laser wavelengths, laser energies, polarity configuration modes, switch structures, focal lengths, and laser waveforms. The switch structure is modified to improve the heavy-current capability of the switch, and the gap distance is increased to improve the voltage capacity. The special design of the trigger structure can reduce the dependence on the laser energy of the switch and the time delay. A new laser waveform of LTVS multiple wave crests could further decrease the laser energy, and it is the research basis of laser trigger miniaturization and its application in the LTVS. [ABSTRACT FROM PUBLISHER]

High-performance trigger generators with high amplitude, fast rise time, low jitter are essential for large-scaled pulsed-power systems. Most of the trigger generators are also tiny multistage pulsed-power systems with complicated structures. In this paper, a high-performance trigger generator was proposed by combining the advantages of low-jitter semiconductor switch and high-power gas switch. An avalanche transistor Marx circuit was developed. It was used to trigger a two-stage Marx generator based on trigatron switch. The output characteristics of the Marx circuit based on avalanche transistors and the two-stage Marx generator were studied. The experimental results show that a subnanosecond jitter, nanosecond rise time, high-voltage trigger generator can be achieved by triggering trigatron switch with an avalanche transistor circuit. [ABSTRACT FROM PUBLISHER]

Electrode erosion is a notable issue influencing the performance and restricting the lifetime of gas switches. In the erosion process, electrode materials heated by spark channel will melt or evaporate and be driven to depart from electrode surface, contaminating the gas switch insulator, and potentially inducing flashover accidents. In this paper, sputtering spots are observed near the erosion crater in a two-electrode gas switch, which indicates the existence of molten droplet ejection. In addition, a set of polymethyl methacrylate rings are inserted in the gas switch to detect the ejection of molten droplet as well as to simulate the switch insulator for investigating the degradation of surface insulation strength. The results show that the ejection of molten droplet is along the tangent surface of electrode with a small divergent angle. After repeated discharges, the insulator surface is bombarded by the droplet ejection and forms dense cracks and embedded metal particles in a narrow band. The spatial varied droplet ejection causes a decrease in both flashover electric field and surface resistance. It also leads to different surface resistances in different regions, which can result in an uneven electric field distribution. Thus, droplet ejection increases the probability of surface flashover accidents in gas switches. [ABSTRACT FROM PUBLISHER]

One of the important issues of concerns for gas-insulated switchgear (GIS) is the generation of very fast transient overvoltage (VFTO) during switching operations, especially at operational voltage levels of 330 kV and beyond. It becomes extremely important to build an arc model for VFTO simulation, since the arc determines VFTO characteristics to a great extent. However, existing arc models for VFTO simulation are too simple to provide simulation results that agree well with the experimental ones. As the arc resistance is often taken as a function of the very fast transient current (VFTC), it is necessary to get the experimental data on the VFTC. Therefore, a 252-kV GIS platform has been built as well as the VFTO and VFTC measuring system. On the basis of the experimental investigation, a novel arc model for VFTO simulation was proposed, in which the time dependent characteristics of the arc resistance is revealed. The model may be used for VFTO simulation for extra high voltage or ultrahigh voltage GIS. [ABSTRACT FROM PUBLISHER]

This paper describes the design and implementation of a three-phase resonant converter with low ripple and high control accuracy. Based on a three-phase LCC-type resonant converter—which has advantages of low ripple, high-efficiency, and high-power density compared with a single-phase converter—a high-voltage power supply with low ripple (<0.1%) was designed. In addition to the general merits of an LCC-type resonant converter operating at continuous conduction mode—including soft switching, low conduction loss, and current source characteristics—the proposed scheme uses only one phase under a light-load condition by having different leg designs of the gate drive circuit and snubber parameters. This allows the design to overcome the operational constraints of the general LCC-type resonant converter. The distinctive design of the three-phase converter structure provides high efficiency and low ripple not only during rated operation, but also under light-load conditions. In order to analyze the high performance of the proposed scheme from no load to rated load, a PSPICE simulation was carried out. Comparison results with a conventional LCC-type resonant converter based on a single-phase structure are analyzed from the viewpoints of output ripple, losses, and operable load range. Using the proposed converter, a 20-kV, 20-kW high-voltage dc power supply design and implementation was presented with a superior gate drive circuit. Finally, the superiority of the proposed converter was verified through a simulation and experimental results. It was experimentally confirmed that the developed power supply achieves high performance in terms of efficiency (98%), operable load range (0.5–20 kV), and low ripple (0.05%), with a high power density. [ABSTRACT FROM PUBLISHER]

In this paper, an inductive cell driven by four pulse forming lines (PFLs), capable of operating at two modes to create a single pulse or four separate pulses, is proposed. A single hardware configuration of azimuthal lines must meet the different requirements of the two operating modes. The optimizations and tradeoffs of azimuthal lines for these two modes are presented. Four candidate azimuthal line configurations that are compatible with either mode are proposed. The cell output voltage and azimuthal uniformity of feed currents are simulated, respectively. The simulation results indicate that the cylinder azimuthal line with four equidistant tabs connected to the cathode palates is the most suitable configuration. As the input pulses are 1000 kV and 25-ns rise time, at the mode with four PFLs driving simultaneously, the cell would produce a pulse with a peak of 865 kV and a rise time of 46 ns into 1.5- \(\Omega \) load. Meanwhile, it could create four separate pulses of 980 kV and 32-ns rise time into 5- \(\Omega \) load with each PFL driving separately. In this case, the azimuthally asymmetric coefficient is calculated to be 25.6%. In addition, the voltage of undriven ports, i.e., the voltage applied across the separating switches, is estimated to be about 1150 kV. [ABSTRACT FROM PUBLISHER]

In this paper, a compact circuit topology is presented for pulsed power generation with a semiconductor opening switch (SOS). Such circuits require the generation of a fast forward current through the diode, followed by a reverse current that activates the recovery process. In general, magnetic switches are used to switch the diode current from forward to reverse. Therefore, adjustment of the output voltage of an SOS pulser is difficult. This paper introduces a circuit that allows an adjustable output voltage of the pulser. A solution is realized with a premagnetizing circuit on the primary winding of a pulse transformer (PT), so that a third winding is not needed. The primary capacitor voltage can now be varied while maintaining proper saturation of the PT. [ABSTRACT FROM PUBLISHER]

The virtual cathode oscillator (vircator), which is one of the microwave sources, has been widely studied for about 40 years, since it has advantages, such as its simple structure and high-power radiation. To implement the behavior of the vircator, a simulation model of the axial vircator using circuit elements is studied in this paper. There are two steps in simulation of this paper: 1) obtaining the optimized case in the fixed input power that has a maximum efficiency and 2) at the case, conducting the circuit modeling. Passive elements and switches are used to model the power losses, emission threshold, etc. This circuit analysis is simulated by the Electro Magnetic Transient Program. Through the simulation model of the axial vircator using circuit elements, the high-power electromagnetic system can be handled as an integrated system, which contains a prime power part, a pulsed power part, a microwave source part, etc. This means that the circuit analysis can be carried out from the prime power to the vircator. [ABSTRACT FROM PUBLISHER]

In the use of spark gaps as switching devices in pulsed power systems, spark resistance is a crucial factor influencing the voltage rise time and the efficiency of energy delivered to the load. However, current oscillations make it difficult to accurately measure the spark resistance. Therefore, we present a spectrographic method to measure the spark radius and the conductivity of spark channel according to its emission spectrums. A series of experiments has been carried out to investigate the time-dependent properties of spark resistance under oscillatory damped current pulses. The current and voltage waveforms have been recorded to calculate the energy deposited in the spark channel. Furthermore, the influences of gas pressure and gap length on spark resistance and the deposited energy have been discussed. The results indicate that lower resistance, and energy deposition is obtained with shorter gap lengths operating at higher pressures. [ABSTRACT FROM PUBLISHER]

Present research in widebandgap semiconductors is concentrated predominantly in continuous-wave applications. While their applicability for pulsed power is recognized, little work has been done in taking advantage of the individual characteristics, such as the large drift velocities, of the candidate materials. Analysis of the relative merits of the key semiconductor technologies—Si, GaAs, SiC, and GaN—in the pulsed-power regime is presented, including limitations associated with majority carrier operation and thermal transients. We will also discuss a GaN photoconductor with a vertical topology, which is particularly well suited to higher current applications and also is geometrically more applicable for integration to transmission lines without degrading the high blocking field (2 MV/cm) now available. The relation between the blocking voltage and conduction current typical of majority carrier operation is analyzed and we will show how optical carrier generation provides the much higher switching power associated with bipolar operation. [ABSTRACT FROM PUBLISHER]

In this paper, the charging precision of a 40-kW 3-kV soft-switching boost converter for capacitor charging is investigated. At the beginning, an overview on the topology and the control of the converter is given. Then, two different feedback controllers are presented and compared analytically to determine their sensitivity on noisy input variables. An algorithm to investigate the charging precision is illustrated. This algorithm considers the signal-to-noise ratio (SNR) of all measured signals, the controller output limitations, quantization related errors, the finite resolution in the digital domain, and the switching signal jitter to calculate the charging precision. All effects mentioned before are analyzed separately. This analysis shows that the output voltage measurement SNR is the key parameter for the charging precision. Then, measurements are performed to verify the model used to calculate the precision. These measurements show that the measured precision is only 6% higher than the precision predicted by the model for the used setup. Finally, the charging precision of the presented converter is analyzed including all considered effects. The repetition accuracy of the converter is determined to be 8.3 ppm at an output voltage of 3 kV, neglecting the load connection and the input voltage variation during one switching cycle. [ABSTRACT FROM PUBLISHER]

Trigatrons have been widely used in many fields. However, the performance of trigatrons typically degrades as operational time increases. Among many influence factors, electrode erosion is a primary cause of trigatrons failure. In this paper, the peak current and pulsewidth of trigatrons are in the order of kiloampere (kA) and microsecond (\mus) , respectively, with an electrode size in the order of millimeters (mm). The discharge images show that the arc appearance looks filamentary in shape during the early period of the discharge. As the discharge process progresses, the arc expands and moves radially to the side surface of the anode. Simultaneously, electrode erosion experiments are conducted on the trigatrons in an atmospheric environment. Experimental results show that the erosion level becomes more serious with an increase of gap distance and discharge current. The weight loss of Mo and W electrodes with erosion times are very close to a linear distribution, however, weight loss rate of W-Cu electrodes increases more dramatically with an increase of erosion times. The erosion of WCu electrode is more pronounced than Mo and W electrodes the volume erosion rate of W is smaller than Mo, but the mass erosion rate of W is higher than Mo. Many cracks always appear on the W and Mo electrode surfaces, which are due to frequent and intensive interaction between the expansion and shrinkage of the electrode material. [ABSTRACT FROM PUBLISHER]

This paper discusses the development and analysis of a transfer function model of the peaking capacitor and peaking switch for an input from Marx generator and an output across load. The effect of variations of gap distance, pressure, peaking capacitor, and the input pulse rise time on the rise time of the output pulse is simulated using this model. For a given geometry, the modeling gives an optimum input pulse rise time and the effect of variation of peaking capacitor on various output parameters. The results obtained from modeling are validated with the experimental results. [ABSTRACT FROM PUBLISHER]

When vacuum interrupters are connected in series, some merits of this structure will be seen and a lot of improvements are expected. This paper describes our investigations on the reignition problem and interruption characteristics of series connected vacuum interrupters. In vacuum interrupters, CuCr (Cr: 25%) is not sufficient at the voltage level >120~kV because of saturation of material. Thus, we should add some material or use completely new material to replace it. By connecting two interrupters in series, the reignition problem sufficiently reduced. Simultaneously, compared with spiral electrode, axial magnetic field has an advantage on voltage sharing because it has a smaller fluctuation of postarc current. In addition, this was confirmed by the experiments with practical ac vacuum interrupters and nuclear fusion experiment dc interrupters. For dc circuit breakers (CBs), because of special characteristic, insulation recovers instantly, and this can be used for dc current breaker for subway and Tokamak nuclear experimental devices. For the application of ac CBs, as with series connection, no reignition happens, and it can be used in the Japanese market for the section separator in super express train. [ABSTRACT FROM AUTHOR]

This paper is concerned with the conduction process of field-breakdown triggered vacuum switches (TVS). The TVS with long gap spacing is fabricated to enlarge the time scale for the investigation of the physical phenomena. According to the image analysis of the conduction process, an approach based on introducing a concept of triggering charge quantity is put forward, which can be used to describe the conduction condition of TVS. Finally, the forming reason of the time delay and the jitter and the influence of the current growth rate and the peak value of trigger current on them are discussed on the basis of test results and come to the conclusion that the injection form of energy plays a prominent part in the conduction process. [ABSTRACT FROM AUTHOR]

This paper presents the commutating over-voltage snubber circuit design method of a thyristor converter that is applied to the International Thermonuclear Experimental Reactor poloidal field-converter module. In the course of poloidal field ac/dc converter normal operation, large transient surges called commutating over-voltages may be generated internally during the commutating process of thyristor, which poses a severe threat to system reliability and needs to be suppressed. In this paper, based on the analysis of reverse recovery transient of the parallel thyristors, the turn-off model of the PF converter thyristor is created and analytical expressions are derived, including the over-voltage crest value and the maximum over-voltage rise rate. Utilizing the analysis results, a systematic approach to the RC snubber circuit design is described. [ABSTRACT FROM PUBLISHER]

For the purpose of repetitive pulsed power applications, specially designed triggered vacuum switches (TVSs) were tested under different high-frequency currents in this contribution. Two TVSs named as TVS-1 and TVS-2, respectively, were designed to investigate the high-frequency current interrupting characteristics with or without an axial magnetic field (AMF) aid. The TVS with AMF at the cup-shaped anode can generate a magnetic field parallel to the arc column by the flowing current itself. TVSs show promise for use in high-current and high-frequency conversion and switching circuits. The V/I characteristic results show that the arc voltage increases with increasing the arc current. The maximum interrupting main current has a sharp decreasing trend with increasing the current frequency from zero to 1 kHz, whereas the curves tend to be flat after 1 kHz. The AMF in the anode gives great aid to decrease the arc voltage and increase the high-frequency current interrupting capacity of the TVS. [ABSTRACT FROM PUBLISHER]

A triggered vacuum switch (TVS) shows good performance when used in pulsed power systems. The time delay and lifetime are the important features of the TVS. We carry out several experiments to measure and analyze the time delay and lifetime characteristics of TVS with a multirod system. The results show that the time delay is strongly influenced by the polarity effect. The lifetime experiments are repeated for 1000 shots under two different operational modes. After repeated 1000-shot experiments, the TVSs are not damaged. The results prove that the TVS with a multirod system can work under a long lifetime in excess of 1000 shots. After the first 200 shots, the trigger voltage maintains between 40 and 1000 V, and the rise time and collapse time are around 2.5 \mus. [ABSTRACT FROM PUBLISHER]

The French-German Research Institute of Saint-Louis obtained essential parts of a 30-MJ energy supply formerly installed at Rheinmetall in Unterlüß, Germany. This paper describes the original state of the energy supply and the refurbishment of one rack with a stored energy of 1.28 MJ and its testing, taking into account worst-case conditions. The rack was split into two subunits of 675 and 608 kJ, which were upgraded with newly developed high-power thyristor and crowbar diode assemblies. The subunits were successfully tested individually and in the pulse shaping mode by connecting them to a common load. Several tests with voltages up to 22 kV, currents up to 267 kA, and a different trigger delay proved the reliability of the modified rack and its suitability as a prototype for the remaining 23 racks. [ABSTRACT FROM PUBLISHER]

Inductive pulse power supplies are of interest because their energy densities are one order of magnitude higher than those of capacitive power supplies, based on the same power-output capacity. The Institute of Advanced Technology puts forward an inductive pulse power topology called the slow transfer of energy through capacitive hybrid (STRETCH) meat grinder. In this topology, an active switch is used to cut off the charging current. Limited by the current turning off ability of integrated gated commuted thyristor, it may encounter difficulties in extending to higher-energy systems. Other than that, the cost of the system is very high. In this paper, the principle of the STRETCH meat grinder is presented. Then, the STRETCH meat grinder with inverse current commutation with semiconductor devices (ICCOS), which is proposed by the Institute of Saint Louis, is presented. The process of ICCOS in the STRETCH meat grinder is illustrated in detail. Finally, simulations are made based on Simplorer 8 to explain how ICCOS works clearly. A preliminary discussion of how to design ICCOS is also presented. We conclude that the STRETCH meat grinder with ICCOS removes the limitations of the original one and has potential applications for high-energy systems in future and can cut down the cost greatly. [ABSTRACT FROM PUBLISHER]

The authors designed the voltage withstand test of the photoconductive semiconductor switch (PCSS) in which the electrode gap was 18 mm in the condition of the pulsed laser triggered. During the experiments, the flashover quenching the photon-activated charge domain (PACD) has been observed, namely, when the flashover occurred, the pulse current amplitude of semi-insulating GaAs PCSS was lower than the normal value without flashover and the rise time of the output current increased. The experiment indicates that the current output characteristics of PCSS contain the information of the flashover, when the flashover occurs with the discharge of the PCSS. The analysis denotes that the secondary electron emission in the flashover quenches the PACD and ultimately quenches the PCSS. The external electric field of the PACD modulated by the secondary electron emission determines if the PACD is to be quenched or not. The Gunn domain is imported to describe the critical state where the flashover quenches the PACD and the threshold condition has been given. [ABSTRACT FROM PUBLISHER]

Pseudospark switches have been successfully employed in a variety of pulsed power applications including high-power electron beam sources, high-power microwaves, and transient plasma ignition for pulsed detonation engines. For applications such as these, advances that improve and simplify triggering are important. Here, magnesium was investigated as a photocathode candidate for optically triggered pseudospark switches, also called back-lighted thyratrons (BLTs). A Mg foil on the back surface of a copper BLT cathode reduced trigger delay and jitter by one order of magnitude when a 5 ns, 3 mJ light pulse at 266-nm wavelength was incident at the back of the cathode at a constant switch voltage of 4.1 kV. Delays less than 150 ns were obtained. An increased photoemission due to the lower work function of magnesium is considered as the main reason for the improvement in trigger delay and jitter for the BLT system. Mg-based cathodes are promising for development of compact optical triggering units with low optical energy sources for BLTs. [ABSTRACT FROM PUBLISHER]