LTRX HomeResearch MissionLTRX DevicePhysics TopicsDiagnosticsLTRX GalleryLTRX People

CPLA Home Directory Publications Links
University of Wisconsin Physics Department Department of Energy National Science Foundation

The LTRX is a cylindrical plasma device with a screw pinch magnetic geometry. The experimental volume is bounded by a 2.1m long and 20 cm diameter cylinder, as shown:
The cylinder itself is interchangeable and experiments have been performed with Pyrex and various combinations of stainless steel, copper, and MuMetal. At each end are large bell-shaped enclosures that house solenoid magnets, vacuum pumps, diagnostics, and the plasma source array. Including these bells the experiment is 4.2 m long and 1.5 m wide, and sits on a rigid stainless steel frame.

Vacuum Vessel and Pumping System

Vacuum is established and maintained at a base pressure of about 0.1 microTorr using Aligent Triscroll roughing pumps in conjunction with a Leybold turbomolecular pump with a capacity of 1100 litres/sec and a CTI-Cryogenics cryopump with a capacity of 2000 liters/sec. The end bell sections are much larger than the diameter of the plasma column to allow for the expansion and escape of unionized neutral gas particles, thus minimizing neutral gas build-up during the discharge.

Interchangeable Vacuum Walls

Cross section of inflatable bladder seal assembly, with sealing surfaces shown. Overlap regions in the drawing indicate compression of the flexible inflatable bladder, both by the flange and by the vacuum vessel. The expansion of the bladder allows a range of vacuum vessel sizes to be accommodated.

To facilitate the study of different boundary conditions on MHD activity, it is of great practical importance to be able to switch the cylindrical vacuum vessel wall or liner with relative ease. However, the central solenoids and end bell sections, shown here, preclude radial extraction. To deal with this problem commercially available inflatable bladders, shown: here, are used for both mechanical support and vacuum sealing. The bladders are compressed by a pair of large flanges and expanded by pressurized air at approximately 60 psi, forming two sealing surfaces. This expansion offers the added benefit of allowing the system to accommodate inner walls which vary in outer diameter by +/- 1 cm. %A drawback of this method is a relative inability to fine-tune the alignment of the tube, though uniform pressure in the bladder assures rough alignment.

External Magnet Coil System and Power Supply

The axial guide field for the screw pinch magnetic geometry is provided by eight discrete solenoids, shown in
here. Solenoid power is provided by steady-state SCR-fired DC supplies capable of generating a 1 kG field on-axis. Both the power supplies and the magnets are actively water cooled, allowing for shot cycle lengths not limited by magnet heat loads. In a typical discharge, the solenoids are energized several seconds before the plasma source becomes active, yielding a temporal decoupling of the magnetic fluxes arising from the plasma and from the external solenoids. Multiple power supplies yield independent control of the end and central solenoids, allowing variable mirror ratios to be achieved. It should also be noted that the solenoids cover a large fraction of the experimental volume, in contrast to other linear (or toroidal) devices. This dramatically limits the field ripple for most of the experimental volume. Nonetheless, ripple of about 3.6% is experienced on-axis at the midplane gap and at the end gaps.

Plasma Source

The UW-RWM employs a high current, high density source developed for helicity injection on the Madison Symmetric Torus. The source is a cylindrical molybdenum washer-stabilized arc plasma.

The arc inside the plasma source (gun) is maintained at 1.2 kA across 100 V for 20 ms by a pulse forming network (PFN), as shown:

Such large current injection is possible due to the very high density plasma formed inside the gun itself which allows for large space-charge limited emission. To inject the current into the experimental volume, the gun discharge must be biased relative to the anode of the device. This is achieved by independent capacitor banks, shown:

A hexagonal array of nineteen such sources are installed on the UW-RWM, though usually only the central seven are used. The guns give a degree of control over spatial current profiles in the device. A typical series of traces for a single plasma source is shown

Cross section of plasma gun: Gas is first puffed through the inlet, then a bias is established between the cathode and anode, yielding a high density arc plasma. Alternating boron nitride and molybdenum washers provide stability to the discharge. Dimensions are in mm.

Typical plasma source time traces, illustrating source operation. The $V_{gate}$ signal controls a gate turn-off thyristor (GTO), yielding in this case a flat top followed by a slowly ramping current.

Illustration of the UW-RWM plasma generation circuit for each plasma source. High voltage banks strike the discharge and quickly bring the current to the desired initial level and are then disengaged. During the discharge, the GTO timing circuit decides whether the low or mid voltage banks are active and thus controls the current.

Shot Timing and Current Control

RWMMasterCircuit illustrates the effective circuit of each plasma source on the UW-RWM. The discharge begins by puffing gas into the plasma source. After 2 ms, the gun power supply PFN is discharged (box 1 in RWMMasterCircuit) thus establishing a high density plasma within the plasma source. After 1 ms, a high voltage capacitor bank is discharged to strongly bias the source with respect to the anode and to breakdown the gas in the experimental volume (box 2). After 2 ms, a pulse width modulation system (box 3) begins to control the bias current by alternating between the mid voltage and low voltage capacitor bank. The signal is determined by a comparator circuit which makes a decision based on the reference and actual waveform. If the measured current is above (below) the requested current, the low (mid) voltage capacitors will be engaged, thus dropping (raising) the current. This system proceeds for the duration of the discharge, and after 20 ms a shorting SCR is triggered (box 4) which terminates the discharge. As the signal is unique to each plasma source, current profiles controlled in both space and time can be generated by the UW-RWM. Switching is achieved on the microsecond time scale through the use of gate turn-off thyristors (GTO).

Data Acquisition and Machine Control

Data acquisition is achieved through a set D-TACQ brand digitizers connect to a Linux PC running Python scripts. Digitization of fast signals is carried out by LeCroy 6810 CAMAC modules capable of digitizing 12 bits at 1MS/s. A total of 480 channels are digitized at 500ksps producing ~20MB of data per shot. The digitizers are equipment with an embedded Linux operating system and MDSplus which automatically uploads data to the Linux PC.

Though the discharge timing cycle and digitization is controlled by the Linux PC, a Windows PC running LabView controls much of the machine operations, such as water cooling, capacitor charging, magnet and temperature control using National Instruments FieldPoint modules. A robust machine clean grounding system is also maintained separate from the building ground to keep stray currents and voltages from contaminating measurements. This is especially important in a linear machine where up to 7 kA of current enters and exits the machine and could find unexpected paths back to its source.