Knowing what feeds and speeds and related parameters (such as the degree
of overlap of parallel tooth paths and the depth of cut) to use for a particular type of wood is the most important issue. Once these parameters are known, other issues fall in place.The best way to determine these parameters is to take some samples of the wood to be machined and to try different kinds of milling operations on the samples such as rough cuts with large diameter tools and parallel finish cuts with different size of tools.
The next most critical issue is to find the minimum number of tool paths and the most appropriate ones that will achieve the desired shape and surface quality in minimum time. Less than optimum tool path strategies can run into extremely long milling times. For example, intuitively, one would think that parallel finishing with a very small diameter round ended tools (ball nose cutters) would lead to the most precise and smooth surface.While this may be true, the resulting milling times can be extremely long (literally days for a large model). Often, an equally smooth and precise surface can be achieved with a much larger diameter tool in much less time. The time that it will take to complete milling a scale model is, fortunately, easily obtained, because tool path generation software allows one to
simulate a milling operation and to estimate the time it will take. Determining the most appropriate tool paths is one area where creativity is involved.
Perhaps the most challenging aspect of CNC milling is cutting fully round shapes with 3-axis technology.This involves deciding how to divide the three-dimensional computer model of the object to be cut into two or more models depending on the number of sides to be cut. If an appropriate origin for each milling operation is chosen, then when the stock is turned to expose another side to be cut, the milled surfaces will align precisely. More
details on this operation follow in the case studies.
Further decisions regarding milling strategy are made after machining of the final model has already begun.We have, for example, occasionally decided to repeat all or part of a tool path with different parameters or after a slight readjustment of an improperly set origin.We have tried two different methods.We have actually learned to understand the basics of Gcodes, allowing us to remove the parts of the code that did not need to be
repeated and we have used a facility of the tool path software that allows us to repeat the tool path from a given percent of completion.