Real-Time Trigger System for BULLKID-DM
by
IPE Bldg. 242, Room 413 + Zoom
Final presentation of Monika Prakash's master thesis
Abstract:
The BULLKID-DM experiment targets the detection of sub-GeV WIMP-like dark matter candidates using kinetic inductance detectors. The DAQ system, based on an RFSoC platform, generates the excitation tones and performs demultiplexing of the modulated frequency comb.
This thesis presents the design and implementation of a real-time trigger system for BULLKID-DM. Although two trigger algorithms were already available in the system—a simple differentiator and a moving average filter—both struggled to reliably detect low-energy events, which are especially important in dark matter searches. This limitation motivated the design of a more robust and sensitive filtering approach. Using Python simulations, several candidate algorithms were evaluated . Based on performance and implementation feasibility, an algorithm based on IIR filtering was subsequently integrated into the FPGA. Notably, the IIR filter stage, typically used to smooth high-frequency noise, is emulated using two cascaded FIR filters, enabling efficient implementation within the FPGA's resource constraints while preserving filter characteristics. Performance benchmarks demonstrate substantial improvements in signal-to-noise discrimination, detection reliability, and algorithmic flexibility compared to existing trigger methods.
To increase flexibility of the DAQ system, an additional controller module has been developed which enables dynamic selection between multiple trigger algorithms during runtime without requiring FPGA reconfiguration. This modular approach allows rapid prototyping and comparative evaluation of different event detection strategies under realistic conditions. The full system is designed with portability and adaptability in mind, ensuring compatibility not only with BULLKID-DM but also with other cryogenic detector and low-background experiments.
Timo Muscheid