Speaker
Description
Bunch-by-bunch beam-size monitoring at SuperKEKB requires acquisition and processing of signals from bunches separated by approximately 4 ns. A silicon-based X-ray beam size monitor (SiXRM), installed in the high-energy ring, has demonstrated measurements of the vertical beam size with a precision better than 6.4 µm and agreement with the existing CMOS-based monitor. The present acquisition system, adapted from the Belle II time-of-propagation detector electronics, does not support continuous streaming. We are developing a data acquisition system based on the AMD RFSoC ZCU216 platform to enable continuous digitization synchronized to the accelerator RF and online beam-size reconstruction. The proposed architecture combines waveform acquisition, bunch identification, and FPGA processing to produce a stream of bunch-resolved beam-size estimates. A compact neural network implemented in the FPGA is planned to perform the reconstruction and reduce the volume of data transferred to the host system. This approach is intended to support studies of rapid beam-size changes, including possible growth preceding sudden beam loss events. We present the beam measurements that motivate the upgrade, the streaming acquisition and processing architecture, and the validation strategy using recorded beam signals and DAC-to-ADC loopback tests.