Search

Cylindrical reflex triodes (CRTs) driven by a pulsed power generator can produce relatively low-endpoint X-ray spectra (~250 keV) that can be suitable for certain radiation–matter interaction studies. CRTs have several advantages over other reflex triode configurations, the most significant being the ability to operate multiple concentric CRTs, connected either in parallel or in series. Such configurations can result in increased X-ray output without increasing the endpoint or changing the pulsed-power generator. This article describes results from, and modeling of, experiments employing a single CRT and two CRTs in a series configuration. The modeling includes an analytic, physics-based model of CRT electrical operation; measured and calculated radiation distributions and spectra; and predicted effects of several configuration variations.

The basic operation of a Marx generator is well known and simple: capacitors are charged in parallel through high impedances and discharged in series, thus multiplying the output voltage compared to the charging voltage. As a basic explanation, in a Marx generator using spark gap switches, triggering the first stage is sufficient to double the voltage on the second stage's switch and so on. All the switches are then switched on in an avalanche mode. However, the behavior is often more complex. The parasitic impedances of the geometry play an important role for the creation of overvoltages. This can make the design and the development of Marx generators quite challenging, especially when aiming for good reproducibility and precise timing. The simulation method presented here has been developed to easily compare the effects and performances of different parameters on the Marx erection. Thus, the designers can readily optimize and compare different triggering schemes or other parameters influencing the erection process. Secondly, by coupling LTspice and Python, a statistical approach is presented to also study the effect of the triggering circuit and other parameters on the overall jitter of the machine.

Laser megajoule (LMJ) is a laser facility being built by the French nuclear science directorate, Commissariat a l'Energie Atomique. This facility is a large flashlamp pumped laser (1.8-MJ optical output). The pulsed power conditioning system of LMJ is based on a modular design of 480 modules (400-MJ energy capacitor bank). A new capacitor bank module (CBM) for the LMJ power conditioning system has been developed by the Institute of High Current Electronics (IHCE, Tomsk) in collaboration with International Technologies for High Pulsed Power (France). Two such modules were produced and tested in IHCE. The main parts of the CBM are the following: high-voltage power supplies, command control system, air conditioning system, capacitor block, and cables delivering the energy to a load. The parameters of the module are the following: total rated capacitance of 2900 (+4%) muF, charging voltage up to 24 kV, peak current amplitude of 240 kA, peak stored energy up to 864 kJ, pulsewidth at 10% of peak power of 360 mus, dimensions of 2.3 times 1.5 times 2.5 m3, and weight of ~2500 kg. In nominal regime (24-kV charging voltage, 40-m length cables), the energy delivered to flashlamps in each channel is 74 kJ (86% efficiency). No failures were observed in more than 1000 test shots, and all parameters of the CBM are well within the requirements. Detailed description of the module design, performance, and test results are stated in this paper.

Results of the experiments carried out on the GIT-12 generator at the current level of 3.5 MA and the Z-pinch implosion times from 700 ns to 1.1 mus are reported. A gas-puff-on-wire-array (triple-shell) load configuration with the outer double gas puff (neon) and the inner wire array (aluminum) was used in the experiments. In the course of the research, implosion dynamics of the triple-shell Z-pinch was studied, and the radiation yield in the spectral range of neon and aluminum K-shell X-rays has been measured. Optimization of the inner wire array parameters (masses and initial diameters) aimed at obtaining the maximum aluminum K-shell radiation yield has been carried out.

AWE is replacing its current Mevex X-ray machine (∼800kV, 35kA, 60ns pulse) to improve reliability and maintainability. A design using a Linear Transformer Driver (LTD) insulated with dry air at atmospheric pressure has been developed by ITHPP. A testbed using 3 cavities, made from 2 bricks of two GA35460M 8nF, 100kV capacitors, a multigap multichannel switch and a magnetic core, has been developed and extensively tested [1]. This led to the development of the full generator that uses 17 cavities in series to produce an 800kV output voltage on a vacuum insulated line and a 23–25Ω diode. The ∼4.3m long inner stalk is cantilevered by the back of the pulser and is positioned by a 3 point adjustable support. This setting is completed by 2 internal tensioning cables to compensate for flexing and provide fine adjustment of the diode end. The assembly of the generator has been done on the 1st half of 2015. The machine can be operated with negative and positive polarity outputs to satisfy operational requirements given by AWE. The tests results of this generator in both polarities are presented in terms of output performances. The reproducibility and reliability of the full system is also analyzed.

The SPHINX facility (with 1-mus rise time and 5 MA) has been developed for soft X-ray production at the Centre d'Etudes de Gramat, Gramat, France. The first experiments were performed with aluminum nested wire arrays Z-pinches for K-shell production. A K-shell yield of up to 25 kJ and the full-width at half-maximum of an X-ray pulse of about 50 ns were obtained with an 800-ns implosion time and a 4-MA peak current. The specific features observed during these experiments are discussed. Inner-array dynamics highlights the fact that resistivity should not be neglected in the analysis. The effect of the return-current housing on the array dynamics, as well as the ways to mitigate it and control its impact on the radiation pulse, is also examined

The Centre d'Etudes de Gramat is now studying long implosion time (about 1 mus) plasma radiation sources. Z-pinches are routinely used by other laboratories in the 100-ns regime with the well-known experimental result, in which an about 2-cm initial radius should not be exceeded to preserve the quality of the implosion for a single array. In a many hundreds of nanoseconds implosion time scale, the question is raised if a bigger radius could be shot. This paper discusses a theoretical approach, which shows that a 7-cm radius is acceptable in the 1-mus regime. It is based on an analysis of magneto-Rayleigh-Taylor instabilities of the accelerating sheath in the r-z plane. It will be demonstrated that all the spatial features of the implosion are determined by the magnetic skin depth, and so essentially depend on the square root of the implosion time. Experiments were carried out with SPHINX generator at the 4-MA level with 800-ns implosion time and confirmed these results

The rod-pinch diode is operated successfully at peak voltages of 2.4–4.4 MV for peak electrical currents of 55–135 kA delivered to the diode. At 4 MV, tungsten anode rods of 1 or 2 mm diam produce on-axis doses at 1 m of 16 rad(Si) or 20 rad(Si), respectively. The on-axis source diameter based on the full width at half-maximum (FWHM) of the line-spread function (LSF) is 0.9±0.1 mm for a 1 mm diam rod and 1.4±0.1 mm for a 2 mm diam rod, independent of voltage. The LANL source diameter, determined from the modulation transfer function of the LSF, is nearly twice the FWHM. The measured rod-pinch current is reproduced with a diode model that includes ions and accounts for anode and cathode plasma expansion.

AWE is looking to replace its current Mevex X-ray machine (∼800kV, 35kA, 60ns pulse) to improve reliability and maintainability. A design using an LTD insulated with dry air at atmospheric pressure has been proposed by ITHPP. A testbed using 3 cavities, made from 2 bricks of two GA35460M 8nF, 100kV capacitors, a multigap multichannel switch and a magnetic core, has been developed and extensively tested. The 3 cavities have been fired on a single low inductance resistive load made from a single volume of electrolytic solution. With a ∼4.2Ω load, the 3 cavities deliver under vacuum ∼145kV, 35kA, in a 84ns pulse (at FWHM). The system proves to operate with high reliability, high reproducibility (

We describe the design of an Arkad’ev-Marx generator with the capacitors discharging energy into the magnetic field within a time less than 1 μsec. We give the equivalent circuit of the generator and calculations for the transient for different firing circuits. We present the results of an experimental study of a generator consisting of 10 stages.

A series of collaborative experiments on complex plasma loads has been carried out on the large inductive energy storage generator GIT-4. The aim of the experiments is to explore the different configurations for the formation of ultrahigh-energy density plasmas in high-voltage pulsed-power systems by direct electromagnetic energy coupling. In this paper, we present some of the underlying philosophy on these experiments and the results obtained. Particular emphasis is placed on the pulsed-power aspects and the effect of source-load coupling for the different studied Z-pinch loads. Resulting radiative properties of the classical exploding wire and liner are experimentally compared with those of the composite pinch scheme in which an intermediate low-density shell is used for staged energy transfer onto a micron-sized wire.

International audience; SPHINX is a microsecond linear transformer driver located at Atomic Energy Comission (CEA) Gramat (France), which can deliver a current pulse of 6 MA within 800 ns in a Z -pinch load. Using the concept of the dynamic load current multiplier (DLCM), which was proposed by Chuvatin, we expect to increase the load current above 6 MA, while decreasing its rise time to ~ 300 ns. The DLCM developed by the CEA Gramat and International Technologies for High Pulsed Power (ITHPP) is a compact system made up of concentric electrodes (autotransformer), a dynamic flux extruder (cylindrical wire array), a vacuum convolute (eight post-hole rods), and a closing switch (compact vacuum surface switch). The latter is a key component of the system, which is used to prevent the current from flowing into the load until the inductance builds up due to the implosion of the wire array. This paper presents the design and testing of the DLCM surface switch, resulting from both electrostatic simulations and experiments on the SPHINX generator. These studies, carried out either with or without load (open circuit), were valuable for a first experimental evaluation of the DLCM scheme in a microsecond regime and provided detailed information on the surface switch behavior.

The linear transformer driver (LTD) technological approach can result in relatively compact devices that can deliver fast, high current, and high-voltage pulses straight out of the LTD cavity without any complicated pulse forming and pulse compression network. Through multistage inductively insulated voltage adders, the output pulse, increased in voltage amplitude, can be applied directly to the load. The usual LTD architecture [A. A. Kim, M. G. Mazarakis, V. A. Sinebryukhov, B. M. Kovalchuk, V. A. Vizir, S. N Volkov, F. Bayol, A. N. Bastrikov, V. G. Durakov, S. V. Frolov, V. M. Alexeenko, D. H. McDaniel, W. E. Fowler, K. LeCheen, C. Olson, W. A. Stygar, K. W. Struve, J. Porter, and R. M. Gilgenbach, Phys. Rev. ST Accel. Beams 12, 050402 (2009)PRABFM1098-440210.1103/PhysRevSTAB.12.050402; M. G. Mazarakis, W. E. Fowler, A. A. Kim, V. A. Sinebryukhov, S. T. Rogowski, R. A. Sharpe, D. H. McDaniel, C. L. Olson, J. L. Porter, K. W. Struve, W. A. Stygar, and J. R. Woodworth, Phys. Rev. ST Accel. Beams 12, 050401 (2009)PRABFM1098-440210.1103/PhysRevSTAB.12.050401] provides sine shaped output pulses that may not be well suited for some applications like z-pinch drivers, flash radiography, high power microwaves, etc. A more suitable power pulse would have a flat or trapezoidal (rising or falling) top. In this paper, we present the design and first test results of an LTD cavity that generates such a type of output pulse by including within its circular array a number of third harmonic bricks in addition to the main bricks. A voltage adder made out of a square pulse cavity linear array will produce the same shape output pulses provided that the timing of each cavity is synchronized with the propagation of the electromagnetic pulse.

A 1 MV linear transformer driver (LTD) generator has been manufactured by ITHPP and SEIV, and delivered to CEA/DIF/Polygone d'Experimentation de Moronvilliers, in order to demonstrate the capability of this technology for flash X-ray radiography applications. This generator is composed of 10 cavities; each cavity includes 16 bricks representing two capacitors (GA 35436, 100kV, 8nF) and one multi-gaps air pressurized switch. To reduce the resistive losses in the switches, the inter-electrode gaps have been reduced from 6 mm in initial design to 4 mm, and the dry air pressure has been slightly increased. Each cavity has been tested individually on an automated testbed using a low inductive resistive load (∼1.2nH) and extensive diagnostics, including a set of sixteen capacitive probes in front of each switch to control their timing sequence. The cavities deliver more than 100kV at ±100kV charging voltage into a matched load of 0.55–0.6Ω in a ∼70-ns FWHM pulse. Around 300 shots have been made on each cavity (to condition the switches) before their integration into the generator.

The usual LTD architecture [1, 2] provides sine shaped output pulses that may not be suitable in flash radiography, high power microwave production, z-pinch drivers, and certain other applications. A more suitable driver output pulse would have a flat or inclined top (slightly rising or falling). In this paper, we present the design and first test results of an LTD cavity that generates this type of the output pulse by including within its circular array some number of the third harmonic bricks in addition to the standard bricks.

ASTERTX is a 6MV Marx+Blumlein X-ray generaton, dedicated to the testing of electrical components. For our experiments, the vacuum diode has been modified in order to set up a flash radiographic diode. The negative polarity rod pinch diode (NPRPD) has been tested on this generator since 2004. The polarity has been reversed from previous experiments2, in order to take advantage from the maximal X-ray emission toward the anode holder at such a voltage. The main goals of these new experiments are a better understanding of the physical phenomena, by the use of new diagnostics as a 4 images high speed camera, a streak camera and vacuum voltage probes.

In this paper, we present the most recent test of the multigap gas switches which are one of the key elements of the fast LTD stages with oil insulation [1, 2]. Inside these switches, the charge voltage is distributed between the multiple gaps with the use of a corona discharge [3]. The evident advantages of this technique are low cost and extreme compactness of the voltage divider because it just consists of few needles that are soldered to the switch electrodes. At the same time, for proper operation of the switch, special features of the corona discharge have to be taken into account. These features include the nonlinear volt-ampere characteristics of the corona discharge that must be the same in each serial gap; the influence of UV radiation and free charged particles, that appear during the corona discharge, on the breakdown voltage of the gaps; and the variation of the shape of the needles that may burn during the switch operation thus limiting its life time. Each of these features (and probably some others that might be not as evident at this moment) may influence the voltage distribution between the gaps and therefore be a reason for the switch prefire. Below we present the design of the switch type Fast LTD, the breakdown voltage of the switch gaps, the volt-ampere characteristics of the corona discharge, and the statistics on the switch jitter and prefire probability.

In this article we present the design and test results of the most powerful, fast linear transformer driver (LTD) stage developed to date. This 1-MA LTD stage consists of 40 parallel RLC (resistor R, inductor L, and capacitor C) circuits called ``bricks'' that are triggered simultaneously; it is able to deliver $\ensuremath{\sim}1\text{ }\text{ }\mathrm{MA}$ current pulse with a rise time of $\ensuremath{\sim}100\text{ }\text{ }\mathrm{ns}$ into the $\ensuremath{\sim}0.1\mathrm{\text{\ensuremath{-}}}\mathrm{Ohm}$ matched load. The electrical behavior of the stage can be predicted by using a simple RLC circuit, thus simplifying the designing of various LTD-based accelerators. Five 1-MA LTD stages assembled in series into a module have been successfully tested with both resistive and vacuum electron-beam diode loads.

Besides the desired K-shell X-rays produced by Z-pinches, there may be unwanted emissions of "hot", >50 keV electrons . We used two types of diagnostic to look for such hot electrons produced by the Sphinx generator . Sphinx uses LTD technology to produce >4 MA pulsed power, driving large diameter (>=14 cm), long implosion time (>500 ns) Z-pinch implosions. Our diagnostics comprised stacks of radiochromic film (RCF) and filtered Faraday cups (FC). The RCF provided time-integrated data on both the X-ray and potential electron outputs of the Z-pinches. With a range of filtrations, the FC let us establish the possible time correlations between the X-rays and the electrons. These sensors were complemented on some Sphinx shots with "sweeping" magnets to confirm the presence of the electrons. On aluminum wire array pinches, there was no evidence for hot electrons; the signals seen on our diagnostics were due solely to the free-bound K X-ray continuum. We will discuss the design and use of the electron diagnostics along with more recent data on the possible presence of hot electrons from the argon gas puff loads.

Sharing similarities with the GEPI pulser which is dedicated to Isentropic Compression Experiments (ICE), VELOCE, an even more compact electrical pulser, has been designed and built in duplicate for SNL and WSU. This type of machine complements gun and laser facilities in the study of material response. In order to achieve a broad loading capability and fast turn around, the design is built around a solid dielectric transmission line to couple current from low inductance capacitors and electrically triggered switches. Peaking capacitors enhanced by a low inductance, multi‐channel sharpening switch reduce the quarter period of the pulser to about 470 ns (0–100%). Gas mixtures in the switch cavity and inductances in parallel allow modification of the shape of the induced pressure wave. At 80 kV of charge voltage, the peak current can reach 3.5 MA. Design of the pulser, range of pressures and velocities, as well as potential applications are presented. A consistent numerical tool developed for pulsers design based on a circuit code coupled to a 1D MHD code is also introduced.

For future flash radiographic needs, an 8 MV radiographic machine, IDERIX, will be developed for the CEA / PEM. This machine will be composed by ∼80 super fast LTD (Linear Transformer Driver) stages. The output voltage of each of these stages (100 kV − 75 ns) will be inductively added along a ∼20 m stepped magnetically insulated transmission line to deliver the power up to the beam diode. In each stage, 16 bricks, made with two 8 nF capacitors (that can be charged up to +/− 100 kV) and one multi-channels multi-gaps switch, are arranged in parallel (with a star pattern). The number of bricks is chosen to adapt the stage impedance to the diode impedance and operate the LTD generator close to matched mode. Moreover, new magnetic cores using a thinner ferromagnetic tape (50μm) allow reducing the losses and improving the performances of the generator. The insulation inside the stage will be done using dielectric oil.

Summary form only given. Within the context of the pulsed power system's development undertaken by CEA for the megajoule laser (LMJ) facility, new capacitor bank module (CBM) is developed by ILICE (Tomsk) and ITHPP (France). Two CBMs were produced and tested in Tomsk. Large number of modules (480) in the complete LMJ installation imposes strict requirements to CBM's parameters. These include reliability of the elements (30 years estimated LMJ lifetime), cost minimization, ability' to withstand failure regime, stability from shot to shot in controlled parameters, limits on size and weight, easy maintenance, and a row of other technical requirements. Main parts of the CBM are: HV power supply, control system, air conditioning system, capacitor block, cables delivering the energy to the load. The capacitor block includes 1 preionization and 10 main capacitors, 1 preionization and 1 main switches, charging and damping resistors, pulse shaping inductors, high voltage relays and safety system. Control system provides operation of the capacitor block in selected mode, monitoring and recording currents as well as voltages, command for dumping at failure and some other functions. CBM's parameters are: total capacitance 3000 muF, charge voltage up to 24 kV, peak current amplitude 240 kA, peak stored energy 864 kJ, pulse width at 10% of peak power 360 mus, dimensions 2.3times1.5times2.5 m3, weight ~2500 kg. In nominal regime (24 kV, 40 m cables), the energy delivered to the flash lamps in each channel is 74 kJ (86% efficiency). No failures were observed in more than 1000 test shots and all the CBM parameters are well within the requirements. Detailed description of the CBM and test results is given in the report.

Summary form only given. For future hydrodynamic experiments, an 8 MV radiographic machine, IDERIX, will be developed for the CEA/PEM. This machine will be composed by ~80 super fast LTD (linear transformer driver) stages. The output voltage of each of these stages (100 kV -75 ns) will be inductively added along a ~20 m stepped magnetically insulated transmission line to deliver the power up to the beam diode. In each stage, 16 bricks, made with two 8 nF capacitors (that can be charged up to plusmn100 kV) and one multi-channels multi-gaps switch, are arranged in parallel (with a star pattern). The number of bricks is chosen to adapt the stage impedance to the diode impedance and operate the LTD generator close to matched mode. Moreover, new magnetic cores using a thinner ferromagnetic tape (50 mum) will allow reducing the losses and improving the performances of the generator. The insulation inside the stage will be done using dielectric oil. Downstream, the vacuum transmission line will be connected to a diode. The negative rod pinch diode geometry has been studied and tested at 6 MV on the ASTERIX generator since 2004. In 2006, a 2 meters MITL with the same geometry than the foreseen end of the future LTD generator, has been added to ASTERIX to mimic the power feed. Thus a dose of 70 rads at 1 m and a spot size of 2 mm (LANL/CEA definition) have been obtained. Simulations predict that those performances should increase at 8 MV and should satisfy die specification asked to IDERIX. A global description of the machine followed by an update on generator and diode studies will be proposed.

The use of magnetic fields to isentropically compress materials for equation-of-state studies has been first demonstrated on the Z machine at SNL [1]. Sharing similarities with the GEPI pulser [2], a compact pulser has been designed and built, focusing on Isentropic Compression Experiments. In order to achieve high compacity and fast turn around, the design is built around a solid dielectric transmission line to couple current from eight low-inductance capacitors that are switched with ultra-low-inductance multi-channel gas switches operating in dry air at atmospheric pressure. A peaking stage made of 72 capacitors enhanced by a low inductance, multi-channel sharpening switch brings the fundamental rise time of the pulser down to 350 ns (10-90%). A set of inductances in parallel with the sharpening switch as well as using various gases into this switch allow us to modify the current wave shape. The pulser delivers a peak current of 4 MA at a charge voltage of 80 kV on a short circuit. The rise time can be lengthened to around 650 ns for a current of 4.2 MA. The use of post-holes convolutes in a solid dielectric insulation design makes that pulser unique as well as its compact size, ease of use and ease of access to the load.

Results of the experiments carried out on the GIT‐12 generator at the current level of 3.5 MA and the Z‐pinch implosion times from 700 ns to 1.1 μs are presented. A multi‐shell (triple‐shell) load configuration with the outer gas puffs (neon) and the inner wire array (aluminum) was used in the experiments. In the course of the research, implosion dynamics of the triple‐shell z‐pinch was studied, and the radiation yield in the spectral range of neon and aluminum K‐lines have been measured. Optimization of the inner wire array parameters aimed at obtaining the maximum aluminum K‐shell radiation yield has been carried out. As a result of optimization of the gas‐puff‐on‐wire‐array Z‐pinch load, the aluminum K‐shell radiation yield (hv> 1.55 keV) up to 4 kJ/cm in the radiation pulse with FWHM less than 30 ns has been obtained. Comparison of the experimental results with the results of preliminary 1D RMHD simulations allows a conclusion that at least 2/3 of the generator current is switched from a gas puff to a...

Sphinx is the new name of the facility developed for soft X-ray production at Centre d'Etudes de Gramat (CEG), France. This facility is based on the ECF2 generator which was designed as a 1-3 MJ, 1 mus rise time, 5-10 MA current driver. In this paper, we present a detailed view of the technology used for the capacitor bank. Some details are also given for the central vacuum part and for the Z-pinch load. The capacitor bank is based on the 1 mus LTD technology; it now stores 1 MJ with 12 parallel branches and 8 LTD stages per branch (each stage charged under 50 kV). We plan to have within the next 2 years, 3 MJ stored energy with 16 branches, 10 LTD stages per branch (charged under 70 kV). Improvements made on the LTD stages are presented. The modularity of this driver and the design of central vacuum part allow to test different schemes: flux compression up to 2003 [1] and direct drive Z-pinches now. We give details on the mechanical arrangements and some examples of experimental results

Summary form only given. This paper discusses preliminary experiments carried out with the SPHINX generator at CEG and analyzes the potential of these sources with implosion time increased up to 800 ns. SPHINX generator (1MJ stored energy, 1 ms current rise time) is used to implode single or nested aluminum wire arrays in direct drive mode with 3.8 MA maximum current. Different designs of double-shell Z-pinches were tested: outer radius 6 to 7 cm, inner radius 3 to 3.7 cm, outer mass 0.2 to 1.7 mg, inner mass 0.3 to 0.8 mg, outer wires number 16 to 216, inner wires number 48 to 72. The benefit of using nested arrays configuration was proved for the stabilization and reproducibility of the source. The dynamic of the implosion is studied with 1D and 2D visible camera and X-UV time-integrated pin-hole cameras. The X-ray emission is qualified with time resolved resistive bolometry, filtered detectors (diamond PCDs, XRDs) and time integrated crystal spectrometer.

Summary form only given. Sphinx is the new name of the facility developed for soft X-ray production at Centre d'Etudes de Gramat (CEG), France. This facility is based on the ECF2 generator which was designed as a 1-3 MJ, 1 /spl mu/s rise time, 5-10 MA current driver. In this paper, we present a detailed view of the technology used for the capacitor bank, for the central vacuum part and for the Z-pinch loads. The capacitor bank is based on the 1 /spl mu/s LTD technology; it now stores 1MJ with 12 parallel coaxial lines and 8 LTD stages per line (charged under 50 kV). We plan to have within the next 2 years, 3 MJ stored energy with 16 lines, 10 LTD stages per line (charged under 70 kV). Improvements made on the LTD stages (8 /spl mu/F, 16 nH, 7.5 m/spl Omega/) and on the trigger system are presented. The modularity of this driver and the design of central vacuum part allow to test different power amplification or direct drive schemes: flux compression and now direct drive Z-pinches are the main experiments. Simple and nested aluminum wire arrays designed for K-shell radiation with implosion time as long as 800 ns were tested. We give details on the mechanical arrangements and some experimental results obtained in 2003 and beginning of 2004.

Summary form only given. Centre d'Etudes de Gramat (CEG) is now studying long implosion time (about 1/spl mu/s) plasma radiation sources which are of great interest because they allow to reduce the complexity and cost of generators. Single or double shell Z-pinch sources are routinely used by other laboratories in the 100 ns regime with the well known experimental result that an about 2 cm initial radius should not be exceded to preserve the stability of the implosion. In a many hundreds ns implosion time scale, the question is raised if bigger radius could be shot. This paper discusses a numerical and theoretical approach which shows that a 6 to 7 cm radius is acceptable in the 1 /spl mu/s regime. It is based on a shell like implosion approximation and on an analysis of Rayleigh-Taylor instabilities in the r-z plane. It will be demonstrated that all the spatial features of the implosion are determined by the magnetic skin depth and so essentially depends on the square root of the implosion time. Some critics of this result will finally be done taking into account the 3D array structure. Some experiments carried out with the 4MA SPHINX generator at CEG confirm that a 7 cm initial radius can be used with 800 ns implosion time, especially with double nested shells. With aluminum wire arrays, this allows to obtain more than 10 kJ of energy radiated above 1 keV, with pulse widths of 30-50 ns, for a total radiation yield around 100 kJ.

Summary form only given. Recent experiments show potential for driving long-implosion-time plasma radiation source (PRS) loads directly coupled to a capacitor bank. The LC generator inductance, L/sub 0/, is typically several times greater than that of the load, L/sub d/, limiting the energy efficiency in those direct drive systems. We suggest a simple modification to improve the generator-to-load coupling, increasing the load current in the case of interest, where L/sub 0//spl Gt/L/sub d/. An intermediate current multiplier circuit element, representing an N:1 transformer, is considered for N=2. It consists of an additional large vacuum volume having inductance L and is connected through a single convolute to both generator and load. Analysis of the modified electrical circuit with constant load inductance demonstrates the benefits of the proposed hardware: having possibly one additional unknown, the convolute, the load current could be multiplied by the factor of I/sub d//I/sub 0/=2L/(L+L/sub d/), where L can be easily made much greater than L/sub d/. The value of inductance added to the generator inductance is 4LL/sub d//(L+L/sub d/) and can be much smaller than the original generator inductance, L/sub 0/. Therefore, the load current could be increased without increasing the current rise-time very much. For existing multi-MA LC generators, circuit simulations coupled to the 0D/MQK PRS model show potential for substantial K-shell yield increase. The PRS load height necessary for a given radiated energy could be decreased, facilitating further potential applications. The absence of plasma armatures in the scheme allowed preliminary experimental tests with decreased size hardware, scaled 1:4 with respect to the full scale design for the ECF2 generator. The tests are performed in two regimes, where the current multiplier volume electrodes are solid or transparent to the magnetic field, and they provide information of geometry optimization at low power.

Summary form only given. Self Magnetic Pinch diode experiments carried out at voltages of 4-6 MV using a variety of different pulse power drivers. Experiments were carried out using the single pulse forming line Asterix driver at CEG Gramat in France and both the high and low impedance versions of the Inductive Voltage Adder (IVA) RITS3 driver at Sandia National Labs in the USA. The experiments showed successful operation of the diode on all the drivers with doses in excess of 150 Rads @ 1 m and spot sizes of less than 3 mm (AWE definition). Record current densities on the target for a radiographic diode were produced at approximately 350 kA/cm. Results of the experiments are presented and compared to code predictions. The Self Magnetic Pinch diode has been used at lower voltages (

Summary form only given. We present the results obtained during the ASTERIX generator experiment at CEG. B field generation is the main technological issue. A 400 kJ capacitor bank connected to different solenoid designs allowed 27 T to be achieved (maximum on axis) inside a 40 mm diameter vacuum cell. An effort has been made to calibrate the electrical diagnostics to evaluate diode impedance. X-ray diagnostics measured the horizontal and vertical spot size, dose on the axis and dose rate. Simulations using the Lsp code were used to design the geometries and results were compared to the experimental data. CEA/DAM is interested in a magnetically immersed diode for radiography dynamic experiments.

GEPI is a pulsed power generator developed by ITHPP for Centre d'Etudes de Gramat (CEG), devoted to isentropic compression experiments in the 1 Gpa to 100 Gpa (IMbar) range, and to non shocked high velocity flyer plates in the 100 m/s to 17 km/s range. The main idea is to generate a high magnetic pressure in a strip line where the samples are located. The whole design is based on low inductance technologies. Depending on the load, the current reaches between 3 and 4 MA in 600 ns. The entire design has been done in a cost effective way and in order to achieve an easy-to-use capability. A description of the generator is shown and typical results of the studies led by CEG are presented. The matters of concern are equations of state, phase transitions and impact of high velocity flyer plates.

This work presents the highlights of high-voltage, positive-polarity experiments on the ASTERIX generator characterizing the rod-pinch diode at up to 6 MV as a source for high-resolution, flash radiography. The paper reviews experimental results and analyses including conversion of ASTERIX to positive polarity, rod-pinch electrical and radiation characteristics, comparisons to numerical simulation, and composite rod-pinch performance. A minimum LANL source diameter of 1.32 mm with a dose of 23.7 rad(air) at 1 m was achieved at a peak voltage of 5.9 MV, leading to a radiographic figure-of-merit of 13.6 rad(Si)/mm/sup 2/. The results demonstrate the utility of the rod pinch for high-power pulsed radiography at up to 6 MV.

The angular distribution of radiation from a rod-pinch diode is examined with coupled PIC-Monte-Carlo simulations using the LSP code (D.R. Welch, et al., 2001) and compared with angular dose measurements on the ASTERIX pulsed-power generator. For the 4- to 6-MV voltage range accessible in the experiments, both the measured and simulated doses at 1 m were 25 to 38 rad-LiF in the forward direction (0/spl deg/). The calculated dose at 90/spl deg/ is 1.2-1.8 times larger than the forward-directed dose depending on both the voltage and rod diameter. This is also in agreement with the measured angular variation (F.C. Young, et al.,). At 6 MV, the simulations show that the dose near 180/spl deg/ is more than twice the forward-directed dose. At these high voltages, electrons approach the anode rod primarily at angles close to 180/spl deg/ resulting in peak radiation in the backward direction. This suggests that negative polarity should be considered for voltages at or above 6 MV to maximize the extracted dose. Simulations show that the 0/spl deg/ dose is higher with tapered rods and increases with rod diameter. Simulations at 10 MV show that the dose near 180/spl deg/ is more than ten times higher than the 0/spl deg/ dose. At 10 MV, the dose near 180/spl deg/ at 1 m from a 2-mm diameter, blunt rod is calculated to be /spl sim/800 Rad (LiF).

The rod-pinch diode is evaluated as a pulsed radiographic source for diode voltages up to 6 MV on the ASTERIX generator. Peak currents of 105 to 135 kA are delivered to the diode for peak voltages of 5.2 to 6.3 MV. The diode couples efficiently to the generator for 1.6-, 2.0-, and 3.0-mm diameter tungsten (or gold) anode rods producing a 40-ns radiation pulse. The coupling is degraded if the anode-cathode gap (Rc - RA) is too small (7.5 mm), if the anode rod extension is too short (10 mm), or if the aspect ratio Rc/RA is too large (20). For the standard anodes considered here, the figure-of-merit (FOM) is largest [11 rad(air)/mm:] for a 1.6-mm diameter anode, corresponding to an on-axis (0/spl deg/) dose of 28 rad(air) at 1 meter from the source and a 1.55-mm source diameter (LANL measure). For a 2-mm diameter anode, the largest FOM [9.2 rad(air)/mm/sub 2/] is obtained for a 1.95-mm source diameter and a 35-rad(air) dose. Blunt-end anodes have larger on-axis source diameters and smaller FOMs than tapered-tip anodes. Measured angular distributions indicate that the dose increases monotonically from 0/spl deg/ to 80/spl deg/ by a factor of 1.6 to 1.8 and suggest that the electron distribution incident on the anode rod peaks at a large angle from the forward direction.

Composite-rod-pinch loads on Asterix consisting of hollow aluminum tubes supporting either 1-cm-long, 1-mm-diam blunt-end or tapered gold slugs, or 1.5- to 2-mm-diam gold spheres are characterized. Composite-slug loads have slightly-lower doses than the 1.6- or 2-mm-diam standard rod pinches reported elsewhere and smaller spot sizes, leading to higher measured radiography figures-of-merit (FOM). The FOM for the gold-sphere loads is substantially-smaller than for the slug loads.

Summary form only given, as follows. The Magnetic Flux Compression scheme is studied as a pulse shaping scheme for very high pressure and isentropic compression in material and as a power amplification scheme for X ray production. Experiments have been done in two different regimes: 100 ns, 15 MA pulse on the Z generator at Sandia and 1 ps, 3 to 8 MA pulse on the ECF generator at Centre d'Etudes de Gramat. A summary of the shots on Z, as well as on ECF is given. For example, 5 Mbar as a shock, more than 2 Mbar in isentropic compression in a copper sample have been achieved on the Z generator using Flux Compression. Comparisons between simulations and experiments are presented. They allow us to tune the codes in order to predict the experimental results and improve our physical understanding of the scheme.

Summary form only given. The stabilization of the implosion of a 15 cm long, 5 cm initial radius, 15 mg tungsten wire array liner, by a B/sub z/ magnetic field is studied. This liner is driven by a 3 MA, 600 ns rise time current, and the B, field is created by a twisted return current. The stability of the implosion itself will first be studied with no twist in the return current, and then, different twist angles will allow the study of the influence of the amplitude of the magnetic field on the stabilization of the liner. Some theoretical results and numerical simulations will be presented, as well as the early experimental results.

The future of PRS programmes is oriented towards very high pulsed power generators of few tens of MJ. At the present moment SATURN and mainly PBFA2-Z are at the edge of this range. It may happen that their technologies, even if they produce excellent results, are not easily extrapolable to higher power, in terms of cost, volume and efficiency. From this investigation, one may conclude that meeting requirements of future powerful plasma radiation sources needs to develop new technological concepts, in order to deliver the appropriate power to a Z-pinch load at affordable cost. This is the main motivation of the SYRINX program which was initiated by the French Defense Department in 1993. The Centre d'Etude de Gramat was asked to realize this research and development program in order to prepare the construction of future large pulsed power facilities for X-rays production which could be used by several customers. The technical specifications and program strategy for SYRINX are outlined as are the main components (HAEFELY-TRENCH capacitors, multigap multichannel coaxial switch and linear transformer driver prototype). The GSI generator prototype and plasma opening switch approach are discussed, as are diagnostics and numerical simulations.

The main advantage of the primary storage based on linear transformer scheme is the ground potential on the capacitor bodies during the shot, allowing exclusion of the total output voltage insulation of the highest stages, and to trigger all the stages simultaneously by using an external trigger pulse. The problem was to build a fast linear transformer driver (LTD), providing both high efficiency of energy transfer into the secondary turn and the current rise time below /spl sim/1 CLs, because of relatively high inductance of the caps and the switches. The key elements of the LTD described below are the HAEFELY capacitors (75 kV, 5.65 /spl mu/F, 13 nH, 40 mOhm, or 90 kV, 3.95 /spl mu/F, 10 nH, 13 mOhm) and the multi gap, multi channel spark switches developed at HCEI for SYRINX project (90 kV, 0.7 MA, 8.5 nH). This LTD was designed as an alternative to the Marx generator in order to improve the performance of the IES technology in the direct drive approach where the only unit providing power multiplication on the load is the POS. The LTD was designed as a single stage with two HAEFELY caps in parallel. The authors describe a 75 kV LTD stage with oil insulation and a 90 kV LTD stage with SF/sub 6/ insulation.

This work is part of technological developments carried out within the SYRINX project in order to implement future multi megajoules drivers, using the multi module inductive energy storage (IES) concept. One of the two 640 kJ IES modules developed for that purpose requires a compact, sub-microsecond Marx generator. This paper presents first results on a 10 stages, one capacitor per stage Marx module whose parameters are as follows : 90 kV charging voltage, 160 kJ stored energy, 670 nH internal inductance, 470 kA peak current, and a 700 ns current risetime. Specific capacitors, gas switches and connections have been developed in the scope to reach stringent requirements in terms of low inductance value, high current switching, voltage insulation and reliability. Performances of those components and of the complete Marx including erection process and voltage and current diagnostics for various loads (short circuit, 0.5 /spl Omega/ and 2.1 /spl Omega/ resistances) are presented. Those results are correlated with numerical simulations using the circuit analysis code PSPICE.

The research module GSI (Generator a Stockage Inductif), a 640 kJ inductive energy storage generator, is one of the two prototype candidate to bring the future MAG3 z-pinch load driver into operation. The GSI is made of 4 capacitor banks, a water transmission line, a vacuum storage line and a plasma opening switch. Results of parametrical circuit simulations that have been performed on SABER and SCREAMER to choose its electrical features are given. Also, design details of each element and schedules are presented.

First experiments with exploding wires (EW) showed that the time of the energy input in a wire, t/sub E/, was of crucial importance for elimination of different pinch inhomogeneities. At the same time, practically realized current drivers (t/sub E//spl Gt/10 ns) still do not allow avoidance of coronal expansion and recompression of the wire plasma and, hence, instabilities formation. In order to sharpen the energy input time it was proposed to use intermediate conversion of the electromagnetic energy into kinetic energy of an accelerated plasma shell. This approach evoked considerable progress in maximizing the radiation power of Z-pinches. Further development of the liner systems was seen in fast collisions of accelerated liners onto targets, such as in double-liner or plasma-on-wire (POW) configuration. In particular, the POW scheme was shown to be rather perspective in obtaining more homogeneous plasma radiation source (PRS). Stable external shell implosion was initially thought to be necessary but hardly achievable point in the POW. The paper aims to advance a concept of the Composite Pinch (CP, or C-pinch) in which a controlled shell instability leads to successive decrease of t/sub E/ that results in improved PRS properties. Discussed model describes general principles of C-pinch formation and defines the limits of its application. The model correlates with the experimental results from GIT-4 generator. The main results demonstrate higher hard X-ray power to be produced by a high-temperature/density uniform plasma.

Summary form only given, as follows. ASTERIX is a high-voltage (6 to 7 MV), /spl sim/60-/spl Omega/, /spl sim/50-ns FWHM pulse generator located at the Centre d'Etudes de Gramat (CEG) in Gramat, France. It has been recently used for the first evaluation of the rod-pinch electron-beam diode as a source for high-resolution, penetrating flash radiography at up to 4 MV (on a 40 to 50-/spl Omega/ load). For this evaluation, the generator was operated in positive polarity, i.e., with positive high voltage applied to the inner coaxial electrode. This had never been done before on ASTERIX. Also, it was necessary to eliminate pre-pulse with a vacuum pre-pulse switch to achieve the desired diode behavior. To evaluate the diode physics, the current and voltage at the diode load must be determined. The electrical length of the vacuum section, the front-end architecture, and the design of the voltage probe located on the oil side of the vacuum/oil interface present difficulties in determining an equivalent circuit for inferring the load-voltage time history. In this paper, the modifications made to the ASTERIX generator for this rod-pinch evaluation will be discussed. The limitations encountered will be reviewed and ideas for improvement will be suggested. Also, the design of the rod-pinch hardware will be presented. We will describe the approach taken to determine the load-voltage time history and present comparisons between the data and circuit modeling using a physics-based rod-pinch model. The effects of the pre-pulse on the diode behavior and the performance of the pre-pulse switch will also be presented.

Summary form only given, as follows. The rod-pinch electron-beam diode as a source for high-resolution, penetrating flash radiography at voltages between 2 and 4 MV has been evaluated on the ASTERIX generator (6 to 7 MV, /spl sim/60 /spl Omega/, /spl sim/50-ns FWHM) located at the Centre d'Etudes de Gramat (CEG) in Gramat, France. This generator was modified for positive polarity operation, and a rod-pinch diode load was fielded. Diagnostics for this load include current and voltage probes to measure the diode electrical behavior, TLDs and RPLs to measure the radiation dose, pin diode and scintillator/photodiode detectors to determine the radiation time-history, two pinhole-cameras to image the radiation source, and a I-m radius tungsten "rolled edge" to measure the source diameter. Successful rod-pinch operation was achieved at 4 MV. An intense pinch and radiation emission was produced at the tip of the anode rod. Doses of up to 20 R were produced at I in from the rod-pinch tip in the forward direction. The dose/charge scales quadratically with the diode voltage from 1.5 to 4 MV. These data can also be fit by a radiographer's equation. For a 1-mm diameter tapered anode rod, the source diameter in the forward direction is 0.8 mm (FWHM of the derivative of the measured edge spread function). The source diameter scales linearly with rod diameter for rods from 0.5 to 2.0 mm in diameter. The source distribution appears to have a Bennett profile. The angular distribution of the radiation was measured with three pin diode detectors and will be discussed. Finally, results from this experiment will be compared with other rod-pinch results at lower voltage.