A CNC machine can create a misleading impression.
You load a drawing, press a button and the machine makes the part.
In reality, several hours of work, testing and occasional backtracking may lie between an idea and the first chip. The machine carries out exactly what it is told to do. The difficulty is telling it the right thing.
Learning CNC machining therefore developed along two parallel paths for me: understanding how wood is machined and learning how to turn an idea into instructions the machine can actually use.
Today, that work mainly runs through two programs, each of which has found a different place in my method: Fusion 360 and MillMage.
Before the CNC starts, the project has to be built
Some projects begin with a very simple drawing.
An outline, a pocket, a series of holes or an engraved motif can be prepared quickly. But as soon as a form becomes more complex, the way of thinking has to change.
A CNC machine does not see a box, a pattern or a jig. It receives coordinates, depths and toolpaths.
The idea must therefore be converted into usable geometry.
This is where Fusion 360 became important in my work. I use it for projects that require genuine design work: three-dimensional forms, mechanical wooden parts, prototypes, complex jigs or components that must fit together.
The task is not simply to draw a silhouette.
Real dimensions, the radii left by the cutter, reachable depths, workholding and the order of operations all have to be considered.
A shape that looks perfectly logical on screen may be impossible to machine if the cutter that must physically enter it was never considered.
That was probably one of the first major changes CNC machining brought to the way I design: drawing while already thinking about manufacture.
Fusion 360: from modelling to machining
Fusion 360 allows two closely connected parts of the project to be developed.
The first is design.
I can draw a part to precise dimensions, change a parameter, create a hole pattern, build several components or work on a three-dimensional form.
The second is machining preparation.
Once the model is finished, the way in which material will actually be removed still has to be defined.
That is a different logic.
Tools must be selected, toolpaths created, depths determined and operations organised. Roughing does not have the same purpose as finishing. A pocket, a contour and relief machining also require different strategies.
I have spent a great deal of time working on the transition between flat drawings and genuine three-dimensional forms.
Importing an SVG and extruding it is relatively simple. Producing a convincing relief, with gradual transitions, parts that pass over or under other forms and surfaces genuinely suitable for 3D machining, demands much more work.
The volume then has to be constructed, sometimes with several profiles, lofts, fillets or freer forms, before deciding how it can be machined.
At that point it becomes clear that modelling and manufacturing are almost two different trades, even when they are brought together in the same software.
MillMage: working directly with the machine
MillMage occupies a different place in my use of the CNC.
For many projects, I do not need to build a complete 3D model in Fusion 360.
A vector drawing may be enough.
I can prepare a form, import it into MillMage and directly create the necessary operations: contours, pockets, drilling or engraving.
This approach is particularly well suited to 2D and 2.5D work, where the forms are defined by different depth levels.
The advantage is a shorter workflow.
The file is prepared and positioned on the work area, operations are defined, and then come the settings linked to the real part: origin, position, height and preparation before machining.
MillMage is also where I recently began working more seriously with V-carving.
The principle is particularly interesting for engraving and future inlay tests. Unlike a pocket cut at a constant depth, the depth of a V-bit varies according to the width of the shape being machined.
The first results behaved as I expected. The next stage is to push the tests further, particularly for joints and inlays where positive and negative parts must genuinely correspond.
This kind of test sums up my approach to CNC machining quite well: understand a function on a simple example before entrusting it with an important piece.
There is no single workflow between Fusion 360 and MillMage
For a long time, I tried to define one universal workflow.
In reality, that is probably not the right approach.
Not every project requires the same level of preparation.
For a jig, a cut-out, a pocket or a hole pattern, producing a complete model in Fusion 360 can be unnecessarily heavy.
Conversely, for a complex form or a true relief, a simple vector file quickly reaches its limits.
My method has therefore become more flexible.
For the most direct work, MillMage makes it possible to start quickly from 2D geometry and build the machining operations.
For projects requiring genuine volume design, Fusion 360 becomes the starting point. The model and machining strategies can be prepared more precisely before moving to the machine.
Separating these roles took a great deal of testing. I had to understand file formats, the difference between a simple extrusion and a true relief, toolpath generation and the transition from the digital project to the workpiece actually fixed on the CNC.
On screen, the part is always perfectly straight.
On the machine, it still has to be positioned correctly.
Origin and positioning: details that can ruin a good program
A machining program can be perfectly prepared and still produce a poor result.
All it takes is for the material not to be positioned as planned.
Choosing the origin, positioning the stock and setting zero are therefore integral parts of the work.
These questions appear secondary when first learning CAD and CAM. In front of the machine, however, they immediately become concrete.
Where is the program zero?
Is it set on the top of the material or somewhere else?
How can a precise position be found again for a later operation?
How can a part or jig be installed several times with sufficient consistency?
These problems led me to work on positioning methods and dedicated fixtures.
The CNC then becomes capable of making tools that simplify its own future work.
That is an aspect I particularly enjoy.
Making jigs in order to return to handwork
I do not use the CNC only to make finished components.
It also produces workshop jigs.
In my view, this is one of its most useful roles in cabinetmaking.
A well-designed jig can then be used with a router, to position holes, reproduce a form or reinstall a part several times in the same position.
In that situation, digital manufacturing does not replace traditional work at all.
It prepares a precise, repeatable reference that will later be used with other tools.
I have also worked on shape matrices and drilling grids. They look like simple operations, but require thought about spacing, offsets between rows and how the design can be changed without starting again from scratch.
Once again, the point is not merely to succeed with the first part.
The real value lies in building a method that can be reused.
The cutter does not know that it is cutting wood
A computer simulation works in perfectly predictable material.
Wood is not predictable.
Two areas of the same board can react differently. Grain direction, density and internal stress can all change the behaviour of the cut.
Pass depth, feed rate and cutter choice therefore have to be adapted.
A small cutter can produce finer detail but is also more fragile. A ball-nose cutter suits certain relief and finishing operations, while a straight or spiral cutter will be used for others.
Theoretical parameters are necessary, but testing in the real material is just as important.
I learned not to treat a successful software simulation as a finished project.
The simulation is one stage.
The first real test is another.
Relief: the next frontier of CNC work
Relief machining is one of the areas I continue to explore.
The difficulty is not simply to create a form that rises and falls.
For a relief to be convincing, its volumes, transitions and the way light will reveal the machined and finished surface all need to be considered.
It requires a different way of drawing.
Work in Fusion 360 has allowed me to explore several methods of constructing these volumes. At the same time, I follow the development of MillMage’s tools and progressively test new possibilities related to toolpaths and relief work.
I prefer to move forward in stages.
A simple motif helps to understand a toolpath strategy. A V-carve test shows how a V-bit behaves. A first inlay can then confront the digital drawing with a real assembly.
CNC machining rewards shortcuts rather poorly.
What learning CNC changed in my work
At first, I thought the main challenge would be learning the machine.
In reality, most of the work happened elsewhere.
I had to learn how to construct clean geometry, understand formats, think about toolpaths, organise operations and, above all, connect the digital model with the reality of the workshop.
Fusion 360 allowed me to take design further and consider projects in three dimensions before making them.
MillMage simplified another part of the process by creating a much more direct link between files, common operations and the CNC machine.
The two approaches are not opposed. They answer different needs.
Today, the CNC is as much a development tool as a manufacturing tool.
It can test an idea, make a prototype, prepare a jig, reproduce geometry precisely or approach forms that I would not previously have handled in the same way.
But it also taught me something less spectacular: the more precise and automated a machine appears, the more important the work done before it starts becomes.
The chip is simply the visible part of the project.