FPGA-based High Performance Diagnostics For Fusion

Abstract

High performance diagnostics are an important aspect of fusion research. Increasing shot-lengths paired with the requirement for higher accuracy and speed make it mandatory to employ new technology to cope with the increasing demands on digitization and data handling. Field programmable gate arrays (FPGAs) are well known in high performance applications. Their ability to handle multiple fast data streams whilst remaining programmable make them an ideal tool for diagnostic development. Both the improvement of old and the design of new diagnostics can benefit from FPGA-technology and increase the amount of accessible physics significantly. In this work the developments on two FPGA-based diagnostics are presented.

In the first part a new open-hardware low-cost FPGA-based digitizer is presented for the MAST-Upgrade (MAST-U) integral electron density interferometer. The system is shown to have an optically limited phase accuracy and a detection bandwidth of over 3.5 MHz. Data is acquired continuously at 20 MS/s and streamed to an acquisition PC via optical fiber. By employing a dual-FPGA approach real-time processing of the density signal can be achieved despite severly limited resources, thus providing a control signal for the MAST-U plasma control system system with less than 8 μs latency. Due to MAST-U being still inoperable, in-situ testing has been conducted on the ASDEX Upgrade, where fast wave physics up to 3.5 MHz could first be observed.

The second part presents developments to the Synthetic Aperture Microwave Imaging (SAMI) diagnostic. In addition to improving the utilization of long shot-lengths and enabling dual-polarized acquisition the system has been enhanced to continuously acquire active probing profiles for 2D Doppler back-scattering (DBS), a technique recently developed using SAMI. The aim is to measure pitch angle profiles to derive the edge current density. SAMI has been transferred to the NSTX-Upgrade and integrated into the experiment’s infrastructure, where it has been acquiring data since May 2016. As part of this move an investigation into near-field effects on SAMI’s image reconstruction algorithms was conducted.