When the motor rotates, the control unit will compare the command (Command) composed of the speed set by the driver and the acceleration/deceleration rate with the speed of the hall-sensor signal change (or by software calculation) and then determine the next group ( AH, BL or AH, CL or BH, CL or...) The switch is turned on, and the length of the turn-on time. If the speed is not enough, the opening will be longer, and the speed will be reduced if the speed is too high.
This part of the work is done by PWM. PWM is the way to determine whether the motor speed is fast or slow. How to generate such PWM is the core to achieve more precise speed control. High-speed speed control must consider whether the clock resolution of the system is sufficient to master the time to process software instructions. In addition, the data access method for hall-sensor signal changes also affects the processor's performance and judgment accuracy and real-time performance. As for low-speed speed control, especially low-speed starting, the change of the returned hall-sensor signal becomes slower. How to acquire the signal method, processing timing, and properly configure the control parameter values according to the characteristics of the motor becomes very important. Or the speed return change is based on the encoder change to increase the signal resolution for better control. The motor can run smoothly and respond well, and the proper P.I.D. control cannot be ignored. As mentioned earlier, the brushless DC motor is a closed-loop control, so the feedback signal is equivalent to telling the control department how far the motor speed is from the target speed. This is an error. If you know the error, you have to compensate it naturally, with traditional engineering control such as P.I.D. control. However, the control status and environment are actually complex and changeable. If you want to control the robustness and durability, the factors to be considered may not be fully mastered by traditional engineering control, so fuzzy control, expert systems and neural networks will also be incorporated into intelligent types. The important theory of PID control.