Conference Paper

Status on the Sphinx Machine Based on the 1 Microsecond LTD Technology

Centre d 'Etudes de Gramat, Gramat;
DOI: 10.1109/PPPS.2007.4345549 In proceeding of: Plasma Science, 2007. ICOPS 2007. IEEE 34th International Conference on
Source: IEEE Xplore

ABSTRACT Summary form only given. The SPHINX machine developed at Centre d'Etudes de Gramat is based on the 1 microsecond LTD technology. Before 2006 it operated with 16 branches of 8 LTD stages each and was used for implosion of Z-pinch loads in direct drive mode. At a charging voltage of 50 kV the stored energy was 1 MJ and the level of reliability of the machine allowed for many experiments on 4 MA, 800 ns implosion of nested wire arrays. Parallel to these experiments, improvements on LTD stages have been achieved on test-beds up to 70 kV charging voltage, leading to a new design of LTD stage, named LTD05. The SPHINX machine has been upgraded with 144 such LTD05 stages. Details are given on LTD05 design, on electrical and electrostatic simulations and on experimental results. Several sub-systems modifications or developments are also described: premagnetizing and triggering systems, prepulse current generator for wire-array pre-heating, and a status on improved performance and reliability of the complete SPHINX system is given.

  • [Show abstract] [Hide abstract]
    ABSTRACT: Summary form only given. The Sphinx generator uses LTD technology to produce >4 MA pulsed power to drive large diameter (>14 cm), long implosion time (>600 ns) aluminum wire array Z-pinch implosions. Recently, an argon gas puff load was developed for use on Sphinx. Specifically, we prepared a triple plenum, triple valve system that drives (1) a central 1 cm diameter gas column, (2) a wide "shell" flow between diameters of 1 and 9 cm, and (3) an outer wide "shell" over the diameter range 9 to 20 cm. A large initial diameter is needed to get a long implosion time. Then one can take advantage of the current drive of the generator while still achieving a high enough implosion velocity to excite argon. 0D modeling shows that the optimum implosion time will be about 500 ns, more than a factor of two larger than previous experience with argon at many megamp currents. Initial, limited testing on Sphinx produced remarkably tight pinches (~2 mm in the K-shell) with <= 8 ns overall pulse widths. Corrected for zipper, the intrinsic K-shell pulse width was under 4 ns. As seen in optical framing images, the implosions were not hopelessly unstable. However, there clearly remains a need for additional refinements to the gas puff system to make it easier to operate and to optimize for implosion time (less than 540 ns) and radial gas distribution (more on-and near-axis mass). The un-optimized system produced ~3 kJ of argon K emission.
    Plasma Science, 2007. ICOPS 2007. IEEE 34th International Conference on; 07/2007
  • [Show abstract] [Hide abstract]
    ABSTRACT: Summary form only given. The current multiplier (CM) concept was proposed to increase the driver-to-load energy transfer efficiency. The suggested CM requires additional volumes with high self-inductance (magnetic flux extruders) connected through vacuum convolutes prior to the load and they extrude the magnetic flux toward the load magnifying the load current. We present the design criteria allowing to achieve high extruder self-inductance at low parasitic inductances added to generator and load in the modified circuit. Two configurations of this new device with one extruder are considered for GIT 12 microsecond MA generator having inductance of ~100 nH. The extruder inductance was either a large vacuum volume or a smaller volume with magnetic core. The discussed design procedure allowed to define optimum coreless and cored CM hardware configurations at conservative values of the AK gaps in CM vacuum lines (15-25 mm). The optimum coreless CM had 80 cm height and 170 cm diameter. Operating on a 8 nH inductive load in experimental tests on GIT 12 it allowed to increase the load current from 4.7 MA @ 1.7 mus without CM to ~6 MA @ 1.5 mus<sup>2</sup>. A more compact cored CM configuration with 36 cm height and 85 cm diameter operating at a 4.6 nH inductive load allowed further load current increase up to ~8 MA @ 1.7 mus. Further experimental tests with a Ne gas-puff z-pinch load showed that the peak load current at ~1 mus was increased from 3.5 MA (no CM) to 5.2 MA and that the energy-delivered to load at implosion was increased from ~170 kJ to ~330 kJ. No considerable energy losses in the CM vacuum gaps and CM convolute were recorded. Therefore, it is experimentally confirmed that the proposed new device is applicable for improving characteristics of existing and future pulse-power generators.
    Plasma Science, 2007. ICOPS 2007. IEEE 34th International Conference on; 07/2007
  • [Show abstract] [Hide abstract]
    ABSTRACT: Summary form only given. The SPHINX machine based on microsecond LTD technology is used in direct drive mode to implode nested aluminum wire arrays with outer diameter up to 140 mm and maximum current from 3.5 MA to 4.5 MA. Wire array implosions were usually done with no prepulse, only current from the LTD lines was switched to the load. Best results obtained in this configuration were 1-3 TW radial total power, 100-300 kJ total energy and 20-30 kJ energy above 1 keV; axial total power reaches around 0.1 TW. The physical processes of z-pinch wire array implosion in this 800 ns regime were shown to be very similar to those observed on shorter implosion time (60 to 300 ns) machines. SPHINX machine configuration has been modified to allow a multi-microsecond current to be injected through the wire array load before the start of the mam current. Amplitude and duration of this current prepulse is adjustable, with maxima ~10 kA and 50 mus. This prepulse dramatically changes the ablation phase and the implosion leading to a delayed implosion and an improvement of the axial homogeneity of both the implosion and the final pinch. First shots were done without any optimization of this prepulse. However, radiation pulse was strongly improved: the full width at half maximum (FWHW) went from 40-50 ns to less than 20 ns with a reproducible behaviour and only one single peak: the total radiated energy was multiplied by a factor of 1.6 and the total power by a factor of 4. Optimization of the prepulse current amplitude and duration was then performed showing that total power can be multiplied by 6 to 7 and power above 1 keV by a factor of 2. Experimental results as well as 2D and 3D simulations are shown to analyze the phenomenology of this improvement and first explanations are proposed.
    IEEE International Conference on Plasma Science 01/2007;