G. Gerdin

The lock-on effect in electron-beam-controlled gallium arsenide switches

The authors analyze the lock-on effect, which is the inability of photoconductive or electron-beam-controlled semiconductor switches to recover to their initial hold-off voltages following the application of the laser or electron-beam pulse, if the applied voltage exceeds a certain value. For GaAs this threshold voltage corresponds to average electric fields in the range from 4 to 12 kV/cm. Experimental results on semi-insulating GaAs switches indicate that the corresponding lock-on current after e-beam irradiation is identical with the steady-state dark current. The highly resistive state of the switch before e-beam irradiation is shown to be a transient phase towards the much lower steady-state dark resistance, with a duration which depends on the impurity content of the switch material and the applied voltage. The irradiation of the GaAs samples with electrons or photons causes an acceleration of this temporal evolution; at sufficiently high laser or e-beam intensities, lock-on of the dark current after termination of the driving ionization source is observed. Based on the experimental results, a model is developed which describes the lock-on effect in terms of double injection and carrier trapping in deep intraband levels. The model explains the major characteristics of the lock-up effects and is supported by the qualitative agreement of the calculated current-voltage curves with the experimental data. >

The recovery behavior of semi-insulating GaAs in electron-beam-controlled switches

Semi-insulating GaAs is experimentally studied with respect to its application in electron-beam-controlled switches. The dark current through the switch is measured both before and after electron-beam irradiation. A lock-on effect, similar to that seen in photoconductive switches, is observed after the electron beam is terminated. This effect is characterized by the switch current continuing to flow, locked to a certain voltage, as long as the voltage is applied across the switch. A possible explanation for this effect, based on the process of electron and hole injection at the contacts, is presented. A method to minimize double injection is offered to make the electron-beam-controlled switch, along with the photoconductive switch, practical for use as both an opening and closing switch. >

Development of a digital acquisition system for partial discharges

A data acquisition system for partial discharges (PD) is being developed, which can simultaneously measure the phase /spl phi/ and apparent charge magnitude q of PDs occurring on a given power line voltage and display the number n distribution of PDs n(q,/spl phi/) on a three dimensional plot. The two-channels (one for the PD signal and one for the reduced power line voltage) data acquisition can sample at a rate of 20MSamples/sec (MSa/sec) and store up to 8Mega bytes of data per channel, which represents 20 power line cycles (PLC) at 50Hz. This system has been tested using simulated PD pulses, of constant q and random /spl phi/ and vice versa. The results show good agreement between the range of /spl phi/ and q in the simulated input PDs and the resulting n(q,/spl phi/) plots. The analysis part of the system has been further tested using experimental PD data obtained from an RLC detection circuit and the corresponding PLC both taken at a sampling rate of 1.2MSa/sec. Good agreement is obtained between the apparent ranges of q and /spl phi/ on the raw input data and the resulting n(q,/spl phi/) plot.

The Effect of Pendel Electrons on Breakdown and Sustainment of a Hollow Cathode Discharge

Although the basic properties of hollow cathode discharges have been known for over 70 years [1], the applications of these discharges have been rather limited. They have been used mainly as a spectroscopic light source of high emission efficiency, low power consumption and small doppler broadened line width [2]. Recently, there has been considerable renewed interest in hollow cathode discharges because of their possible application as high power switches. They utilize a cold cathode and under various operating conditions have switched over 100 kA, demonstrated over 2x1012 A/s dI/dt, and achieved subnanosecond jitter [3].

Development of a digital acquisition and processing system for partial discharges

A 2-channel data acquisition and processing system (DAPS) for partial discharges (PD) using a fast (20MSa/sec) digitizer (National Instruments 5102) has been developed which can acquire up to 16 million samples per channel (This represents 48 power line cycles (PLCs) at 60Hz), and simultaneously measure the phase /spl phi/ and q of PDs occurring on a given power line voltage and represent on a 3D plot of n(q,/spl phi/). This system has been tested, using artificially generated PD pulses at constant/variable q and constant/variable /spl phi/ and experimental data. Good agreement is obtained between the ranges of /spl phi/ and q in the artificially generated PDs, the raw input data and the resulting n(q,/spl phi/) plots, which is the number of PDs occuring in a given q and /spl phi/ window in a given period. The acquisition and processing programs developed in LabVIEW are optimized to handle 32 million (16 million/channel) samples of data. The resolution of this DAPS system is 50 mV in voltage and 1/spl deg/ in phase, and is limited by the voltage resolution of the digitizer, and the resolution in delay timing of the artificial PD generation system.

Computation of ac and dc electric field around a wet polluted insulator

Insulators in actual service conditions are polluted due to ambient operating conditions. Due to leakage currents dry bands can be formed on the surface of the insulator. Distortion of the potential distributions would occur by the existence of both a wet polluted band and a dry band in the presence of air. In our previous paper, a benchmark problem for the dc case of a cylindrical insulator with a wet and dry band was formulated. Substantial agreement was found between our finite-difference numerical results, the analytic solution, and the commercially available finite element package. In the present paper, the commercially available package is used to extend the calculations to include ac effects. Two configurations are considered. In one, the dry band is maintained next to the upper high voltage electrode; in the other, the dry band is next to the grounded low voltage electrode. The dry band length is varied as a percentage ratio of the length of the insulator.

Time‐resolved temperature measurement of a pinched plasma using the dispersive x‐ray analysis of the continuum emission

The dispersive analysis of the x‐ray continuum is employed to determine the electron temperature of a dense pinched plasma. A curved crystal spectrometer was used to obtain the desired dispersion, and two or more p‐i‐n (doubly diffused silicon) diodes were used to monitor the time‐resolved emission. The dispersive technique is inherently more accurate than filtering techniques and simpler to implement than laser scattering techniques. In an experimental verification of the technique, the time‐resolved temperature of a plasma focus pinch was measured under various conditions, and was found to be consistent with density observations and MHD calculations. It was possible to observe the cooling effect of radiation emission from seeded plasma focus discharges. The elliptical crystal spectrometer is shown to be well suited to the analysis of the x‐ray emission from a pinched plasma source.

Charge-State Equilibrium and Nonequilibrium Modeling of the Carbon-Pellet Plasma Interaction

Self‐consistent equilibrium and nonequilibrium charge‐state models are formulated for the spherical expansion of low‐Z pellet vapor as an inviscid perfect gas of constant ratio of specific heats being heated volumetrically by the incident electrons of a thermonuclear plasma. The two models are found to be in agreement in the region where the ratio of the ionization length ζj to pellet radius rp is less than unity, but a single parameter, such as the magnitude of this ratio on the sonic surface, is insufficient to determine whether an equilibrium model will be valid for all regions of the ablatant for carbon pellets. Thus a nonequilibrium model is necessary to model the outer regions of the ablatant cloud even for thermonuclear plasma conditions when the cloud is very dense. Also, the effect of the ionization of the ablatant by the incident plasma electrons is found to be 10% or less for even the C+3 region in the thermonuclear regime. Finally, although the model used for the healing of the ablatant by the...

Application of the results of carbon pellet modeling to the problem of plasma penetration

The assumptions of the evaporation model for low‐Z pellets interacting with magnetic fusion plasmas developed by Parks are tested. These assumptions are that the vapor density profile in the region adjacent to the pellet surface falls off with radial distance as r−α, where 5<α<6, and that the ionization time for the transition between charge states τzi is much less than a flow time for the vapor in this same region τf (i.e., for r<∼3 sonic radii). The first assumption is tested by solving a two‐parameter eigenvalue problem for the evaporation cloud in the region interior to the sonic radius; the results are found to be consistent with the low‐Z evaporation model. The second assumption, that τzi ≪ τf, is tested at the sonic radius using the results from atomic physics and the low‐Z evaporation model. It is found that indeed τzi ≪ τf for plasmas with parameters close to thermonuclear conditions (e.g., CIT), but not for those of smaller Tokamaks such as TEXT. The results of pellet penetration calculations fo...

The lock-on effect in electron-beam controlled gallium arsenide switches

A current lock-on at a constant voltage was observed in electron-beam controlled GaAs switches. Experimental results indicate that the lock-on current of the system is actually identical with the time-asymptotic dark current under double-injection conditions. It is shown that the pre-illumination of the sample with an ionization source does not influence the amplitude of the current but causes only a reduction in the time necessary to reach its final value. In particular, it is demonstrated that the initial highly resistive state is not an actual steady state but rather a transient phase characterized by a nonequilibrium distribution in the electron and hole trap occupation. Based on these experimental results, a scenario is developed which describes the lock-on effect in terms of current injection through the contacts and carrier trapping in deep intraband levels.<<ETX>>