Pipeline-based real-time single event data sorter for coincidence detection

The single event data sorted in time order are desired in most coincidence detection methods. However, the singles transmitted using the serial data transmission method, unless FIFOs with different depths are implemented in front-end modules to compensate the differences of time delays between modules, are usually not in time order. Therefore, an event data sorter is usually necessary to arrange the singles. In this study, we investigate a category of real-time event data sorter based on the pipeline architecture. The possible ways to implement the pipeline-based sorter are studied. Their performances (in terms of maximum sorting span) and resource requirements (in terms of pipeline depth, number of comparators, number of data swappers, and complexities) are assessed and compared. The pipeline-based single event data sorter and coincidence detector circuit presented in this paper has been implemented in a FPGA of a wearable brain PET (named Mind-tracker) system to perform the real-time coincidence event detection on-line.