FAST DSP USING FPGAs AND DSOs FOR MACHINE DIAGNOSTICS

Digital signal processing using digital signal processors is now a mature field for machine diagnostics, giving significant benefits, in particular when used to analyse BPM signals for tune measurement and fast feedback systems. We discuss here digital signal processing using Field Programmable Gate arrays (FPGAs) with large gate counts and intelligent oscilloscopes. These offer great potential for the analysis of very fast signals to maximize the information extracted from high bandwidth sensors. i) FPGAs allow data to be filtered numerically and treated at the speed of data collection of A/D converters in the 100MHz range. Parallel, fast and continuous treatment of BPM and FCT signals is possible. Examples are given of injection efficiency, turn-by-turn injection efficiency, turn-by-turn beam position, amplitude and phase calculation with averaging over each turn or many turns. ii) Modern oscilloscopes include much computational power. In-built DSPs can perform correlations on the traces allowing the application of FIR filters. Some oscilloscopes incorporate a PC and allow on-board manipulation of the data using MATLAB. An example is given using an FIR applied to a 5GHz oscilloscope to extend its time response to measure electron bunch lengths less than 100ps with 1ps resolution