Signal-analysis-driven Analog Conditioning for Weak Transient Signal Detection in Rapeseed Cleaning Loss Monitoring

In cleaning-loss monitoring during rapeseed combine harvesting, traditional experience-based circuit design struggles to achieve stable triggering because single-grain impact signals are weak and interference is complex. To address this problem, this study proposes a hardware design method for loss monitoring based on signal analysis and analog signal conditioning. By integrating the grain impact sensing mechanism, kinematic decomposition, and the impulse–charge conversion relationship, the study analyzes the differences between the target signal and interference signals in both the time and frequency domains. Furthermore, based on the ratio of the energy retention rate of the target signal to that of the interference signals, a signal separation margin metric is proposed to enhance target-signal identification capability for weak rapeseed grain impacts, and is then used as a quantitative basis for the coordinated design of filter cutoff frequency and order, operational-amplifier gain allocation, and the envelope time constant. A signal conditioning chain was then established, comprising piezoelectric sensing, active filtering, signal amplification and unidirectional rectification, envelope extraction, and hysteresis comparison. At the signal-chain level, signal-chain modeling, LTspice simulation, and experimental measurements at key PCB nodes were used to validate the rationality of the proposed design method and the effectiveness with which the signal conditioning chain fulfills its intended functions. At the system level, bench testing was further performed to verify the application performance of the system. The results showed that the designed circuit could effectively preserve rapeseed grain impact responses while suppressing interference, thereby enabling stable single-event triggering of grain impacts. The square-wave pulses generated by grain impacts had widths of no more than 5 ms, with an average width of 2.3 ms. The monitoring system achieved a bench-test accuracy of over 97.5% and was further validated for field applicability. This study proposes a quantitative parameter design method for optimizing the recognition of weak transient target signals, centered on signal separation margin, establishes an integrated research framework spanning mechanism analysis, signal separation, parameter design, and system validation, and on this basis completes the hardware design of a rapeseed cleaning-loss monitoring device, thereby providing a useful reference for the design of online weak transient signal sensing systems for agricultural equipment operating under complex conditions.