Toolpath selection based on the minimum deflection cutting forces in the programming of complex surfaces milling

In this paper, a new methodology for the selection of the milling toolpaths on complex surfaces that minimize dimensional errors due to tool defection is presented. In this way, an improvement on the accuracy of milled surfaces is achieved. The methodology can be applied to both three and five axes milling. In the three axes case, it is based on the calculation of the minimum cutting force component that is related with the tool deflection. This component has been previously defined as that perpendicular to the tool axis and contained on the plane defined by the tool axis and the normal vector to the workpiece surface. Cutting forces are calculated for each 15 sense on the tangent plane to the milled surface, in a grid of control points defined by the user, both for dowmilling and upmilling. With this information there are two possibilities. First, select a general toolpath direction that minimizes the mean value of the tool deflection force on the surface, and bearing this in mind, the CAM operator can produce a CNC program which leads to an accuracy improvement. The second option is the selection of different milling directions at each control point. Joining these points, the minimum force lines are defined on the workpiece surface. These can be used as the master guides for the toolpath programming of a complete surface. In the case of five axes milling, the approach is different, because in this case the tool-axis orientation with respect to the workpiece surface may be changed using the two rotary axes. Therefore, for each workpiece area both tool-axis orientation and machining direction can be selected to keep tool deflection force below a threshold value. Some case studies of both techniques and in-deep discussion of results are presented. Applying this approach, in three axes milling dimensional errors fall down from 30 mu m to below 4 mu m. In five axes milling errors can be kept below 15 mu m in most of the cases. (c) 2006 Elsevier Ltd. All rights reserved.