Controller Path Fidelity with Acceleration and Override Control

In the traditional CNC method, acceleration control is applied after interpolation with a causal (not using future information) low-pass filter. In the earliest CNC controller, this was implemented by applying a large time constant (often several hundred milliseconds) in the analog velocity loop of the servo drives. Now it is more commonly implemented with a dynamically equivalent digital low-pass filter, usually first-order, acting on the data stream out of the interpolator. In many path-based positioning applications, such as CNC machining, the controller’s commanded pathgeneration algorithms need to satisfy many user desires simultaneously. First, they must follow the desired path within a certain required accuracy. Second, where possible, they must follow this path at a programmed vector tool speed, or “feedrate” (the rate at which the material is fed into the tool). Third, they must keep the path traversal within the physical capabilities of the machine, usually expressed as limits on the higher derivatives of motor position, particularly acceleration. Finally, in many systems, the operator must be given an interactive “feedrate override” capability to be able to adjust the speed manually and instantaneously without changing the part program. Typically this is done with a knob that sets the override as a percentage of the programmed feedrate. Figure 1 shows the resulting position and velocity profiles for a simple constant-speed move passed through a digital filter with a 100-msec time constant.