Embedded Bearing Fault Diagnosis System DSP Side Design

Rolling bearings are important machinery parts. Nowadays, there are many bearing fault diagnosis methods. One of them is judging the malfunction though listening to the sound of bearing vibration, which is not automatic. It also can be done though signal processing. But most of those systems are based on PC. We introduce an embedded rolling bearing fault diagnosis system. It's more stable and cheap than the system based on PC. The system composes of ARM and DSP. This thesis focuses on the design on DSP side of the system, which composes three modules: the front end data acquisition module, the DSP algorithm processing module and the back end HPI communication module. This thesis introduces mainly the hardware and software design of DSP side.First of all, the thesis gives the requirement analysis and the construction of the system, as well as the choice of the devices and the realization. And then it introduces the hardware and software design of data acquisition circuit. It then introduces on-chip resources of TMS320C6713 and the hardware design of the peripheral circuit in detail, including the memory extension, HPI circuit, clock and power circuit design. And also, it gives some corresponding codes. The software part mainly includes the usage of DSP/BIOS, task management, HPI communication protocol, and the method of code optimizing, and also gives an improved technique. The second-level bootloader design is also given. At last, some suggestions for further development and the conclusion are presented.Though real-time tests, we found that DSP could complete all the arithmetics in time and communicate with ARM successfully. The system runs stably.