Software Design And Development On DSP-Based Multi-axes Motion Controller

Open NC technology (Open architecture numerical control) is a main developing direction of NC research, while motion controller is a key component of PC (personal computer) plus NC type open system. This thesis focuses on multi-axes motion controller, which runs on ISA bus and utilizes DSP and FPGA as its kernel processors. We develop motion controller software and some testing software. We also build and test a CNC interpolation platform with this multi-axes motion controller.We adopt two different levels of timer interrupts for the DSP, which is the kernel processor of the motion controller. We also design the four-axes serial D/A output module in FPGA.In this thesis, we design and develop the software frame of motion controller in DSP. For motion planning, we adopt trapezoidal, S-curve and flexible acceleration/deceleration control to meet the requirement of speed and stability in high speed CNC. For servo control, we adopt the PID plus velocity feed-forward plus acceleration feed-forward algorithm. We can achieve accurate and stable motion control performance by tuning control parameters. For NC interpolation, we adopt the look-ahead technique to calculate the corner's federate and implement of interpolation algorithms in the DSP.The testing software of the motion controller is developed in Visual C++ 6.0. The main functions include G-codes decoding, parameters setting, axes'status and sampling data monitoring. We test the motion controller in both functional module level and system level, and find that the motion controller works within our developing expectation.Finally, we build a CNC interpolation platform to test the open architecture numerical control system with our multi-axes motion controller. We do lots of testing experiences on this platform to validate its performance. The platform is flexible and effective. We can embed all kinds of interpolation algorithms and advanced motion control algorithms into the motion controller. It is useful for NC algorithm research and application.