Development and experimental teaching design of a distributed synchronous data acquisition platform for the industrial internet of things

This study aims to develop a distributed data acquisition platform with both high-precision time–frequency–phase synchronization and engineering practicality for the Industrial Internet of Things (IIoT). It also addresses the persistent teaching bottleneck where experiments in electronic information courses are disconnected from real engineering scenarios, failing to cultivate students' abilities to solve complex engineering problems. Furthermore, it intends to systematically transform the platform's core technologies and research methods into experimental teaching resources, thereby enhancing students' capabilities in solving complex engineering problems and hardware–software codesign.

The platform takes the Fudan Micro Qing Long series Programmable Logic (PL) + Processing System (PS) heterogeneous System-on-Chip as the core and constructs a distributed synchronous data acquisition architecture based on the Precision Time Protocol (PTP). To address the key problems of substantial jitter in soft timestamps, poor scalability of hard timestamps, and lack of sampling synchronization in traditional PTP implementations, this study proposes a three-level clock synchronization mechanism, double-edge triggering technology, and a sampling clock phase alignment method. Through hardware–software codesign, high-precision synchronization in three dimensions (time, frequency, and phase) is achieved. Specifically, the PL section undertakes real-time logical tasks such as hardware timestamping, clock synthesis, and sampling triggering. In contrast, the PS section hosts the protocol stack and upper-layer control functions. A closed-loop regulation system composed of a digital-to-analog converter and a voltage-controlled crystal oscillator is adopted for fine frequency synchronization, and a mixed-mode clock manager is used to dynamically adjust the phase of the sampling clock, effectively solving core issues such as soft timestamp jitter, insufficient scalability of hard timestamps, and inconsistent sampling edges in traditional schemes. Meanwhile, a four-stage progressive experimental teaching system (basic verification, advanced testing, comprehensive evaluation, and open innovation) is constructed that centers on the developed platform, forming a 20-class-hour project-based teaching chain that covers the entire process of hardware deployment, core algorithm optimization, and overall system performance tuning.

Test results show that the average time synchronization error between distributed nodes of the developed platform is 3.3 ns with a standard deviation of 4.9 ns and a maximum time deviation of 16.6 ns. Additionally, the frequency synchronization deviation converges to ±15 ppb, and the time deviation converted from the sampling phase deviation ranges from-12.2 ns to 16.4 ns. The comprehensive synchronization performance is superior to typical baseline methods and meets the requirements of industrial-grade high-precision data acquisition. Teaching practice verifies that hierarchical progressive project-based experiments can effectively guide students to complete the full-process training from system construction to algorithm optimization and performance evaluation, greatly improving their abilities in system integration, algorithm optimization, and engineering expression.

This study successfully develops a high-precision distributed synchronous data acquisition platform that meets IIoT requirements, and the proposed multiple synchronization methods effectively ensure the consistency of cross-node data in time, frequency, and phase. By systematically transforming scientific research achievements into a stepped experimental teaching chain, a virtuous cycle of technology development, educational empowerment, and competency enhancement is formed, providing a referenceable and scalable practical paradigm for industry–education integration in the IIoT field.