The legacy ECEI system here shown setup at UC Davis for characterization. Note the use of large reflective optics as the plasma facing elements.
The new TEXTOR ECEI system comprised of simplified translatable lenses. The metal aperture at the left simulates the tokamak window during characterization. The forward lenses are shown in the wide zoom configuration.
Possible plasma coverage for both the narrow and wide zoom positions. Eight frequency channels divide the image horizontally, while 16 separate antennas divide the image vertically such that each single image identified above is comprised of 128 channels. The choice of single image is made by tuning of the LO source. Imaging of the high field side is limited not by the optical properties of the diagnostic, but rather by third-harmonic overlap which occurs at a major radius of approximately .
Antenna sensitivity is compared for signal coupling to the air and substrate (lens) sides. The nominal 15 dB improvement in sensitivity illustrates the benefit of “front-side” LO coupling as is made possible by the minilens array concept.
The antenna array box (cover removed to show minilens array structure) is compared to the experimentally measured beam pattern of the LO power. A single astigmatic beam is used and power is adequately coupled to all antenna elements.
New optimized antennas (bottom) exhibit improved radiation patterns compared to previous elements (top). Sidelobes are reduced as much as 12 dB and the main lobe is made to better approximate a Gaussian distribution.
The field patterns of the eight spatial channels in the upper half-plane of the image space are shown. The position of the focal plane is unchanged in the narrow (top) and wide (bottom) zoom configurations. Although mode competition perturbs the beam pattern to the low field side of the beam focus, beams are shown to be very clean and uniform about the focus and toward the high field side.
Horizontal (top) and vertical (bottom) slices of the field distribution in the focal plane are shown for the narrow zoom case. Note that -plane or toroidal patterns for central and marginal channels lie right on top of one another.
Translation of the focal plane is demonstrated. The focal plane location is adjusted from 20 cm to the high field side of the plasma center (top) to 30 cm to the low field side (bottom). This translation is performed with minimal change in spot or total image size, demonstrating the independence of the zoom and focus features. As in Fig. 7, only the upper eight channels are shown.
Imaging of the low field side region illustrates the implicit benefit of the vertical zoom feature. In the narrow zoom configuration (top), fine details in the reconnection zone are visible, including the tendency for heat flux to be expressed at the edges of the particle flux surface puncture immediately before the crash (second and third frames, and ). One may also discern the transport of heat along the surface, creating an isolated cold region near the plasma core (fourth frame, ). At , the crash is complete and heat has exhausted from the core. The entire sequence of images encompasses . A wider zoom (bottom) reveals the character of larger features, such as the rotation of hot and cold islands in the poloidal plane. In the compound sawtooth crash that is imaged here, a smaller crash occurs at (second frame), followed by a larger release of energy from the core at (sixth frame). This sequence of images encompasses with poloidal symmetry restored in the final frame at .
The high field side of the surface is also accessible to the new ECEI system. The (a) time trace was obtained from a central channel imaging in the region defined by (b). Frames (c), (d), and (e) portray the transfer of heat from the core to a region outside the surface over a period of immediately following the final precursor oscillation, consistent with high field side sawtooth crashes observed in Ref. 9. Wide and narrow zoom views are possible at this radial location also. Here, a vertical coverage of greater than 30 cm ensures that the vertical extent of the surface is completely visible.
Time traces (top) and power spectra (bottom) obtained with the old (left) and new (right) ECEI diagnostics for comparable plasma conditions. Sawtoothing is present in both time traces. Frequencies above 10 kHz are taken to be noise, and the average relative value is compared. This method of comparison suggests better than a factor of 2 improvement in the system noise floor.
Article metrics loading...
Full text loading...