This text follows on from the previous article "The CAD-CAM Process Chain: The Status Quo".
An important prerequisite for the optimization of the CAD-CAM process chain is the avoidance of media breaks and the associated loss of information.
Media breaks occur when different, unconnected, often even incompatible systems are used for different tasks in the product development process.
This does not mean that only systems "from one casting" are suitable for modern production. Coordinated systems and, above all, the underlying processes also offer many possibilities.
First we consider the problem of information loss and its origin.
The example from the previous article, in which the classic production drawing is interposed between design and production, is taken up here. The most obvious loss of information here is the loss of 3D representation.
The three-dimensional CAD model of a workpiece is displayed on a two-dimensional drawing and then recreated as a three-dimensional product.
A contradiction that is almost too obvious to be recognized.
In practice, an NC program is often created online on the basis of a drawing on the machine tool. However, this program is displayed three-dimensionally by the machine within the scope of the existing possibilities in order to simplify orientation on the workpiece for the machine operator. From this point of view, the workpiece is modelled twice.
If you stick to this example, production-relevant information is often lost on the way through the drawing. Of course, this should no longer happen after many decades of standardization of technical drawings including production drawings.
The process of creating the drawing is, as a rule, still partly manual and therefore potentially faulty.
If surface information, tolerances and similar are not explicitly stated in the drawing, they are no longer available in the production process, with the result that parts are initially unusable.
But even without the involvement of the classic paper drawing, information is often lost.
This can be seen very clearly in the example of two independent systems involved: System A is the CAD system, system B the CAM system. The exchange between the two systems takes place via classic generic exchange formats. These can be for example step, Iges, 3dXML or JT.
Between the systems there is usually a purely geometric exchange of information. This means that bodies and surfaces of the workpiece are transferable. In addition, general attributes or separately imported PMIs (Product Manufacturing Information) can often be transferred. Usually, these attributes have to be entered manually by the designer.
In addition to some mathematical inaccuracies, which are initially neglected here, that occur when using these formats, much information is nevertheless lost. This type of information transfer turns a bore into a cylindrical area in the part where no material may later be present.
For the NC programmer in the CAM system, this does not initially determine whether it is a functional bore, a circular pocket or even a burnt-out section.
In the sense of manufacturing tolerances, for example, this is a crucial piece of information.
These problems are solved in daily practice by, as a rule, individual agreements and standards. For example, Colour codes on surfaces, supplementary descriptions in enclosed documents or additional production drawings are used.
A further disadvantage is that features such as feature-based programming on the CAM side are usually not possible with neutral data formats. This is due to the above mentioned lack of information about the concrete design of the geometry. This is associated with the fact that efficiency in the CAM area is lost, which naturally also makes the overall process less efficient. It is therefore clear that a pure data exchange does not mean a full exchange of information.
Important for the correct planning and implementation of a CAx landscape, especially in the CAD-CAM area, as already mentioned, is the avoidance of information loss. Roughly speaking, every piece of information that has been entered into a model should also be available for the rest of the process without being converted, moved, or duplicated again by external intervention. On the one hand, these processes mean additional effort, on the other hand they increase the error potential and therefore also the necessary rework and corrections.
This ideal state is certainly not achievable everywhere, but it should be aimed at when designing CAx systems, especially for new acquisitions or major changes.
It is therefore imperative to first record the necessary level of information for each individual process step, such as design, NC programming, work preparation or fixture construction, and to define the information to be transmitted at the interfaces. A suitable system landscape can then be created.
If, for individual reasons, this landscape is to consist of several incompatible individual systems, one of the most important tasks is to find suitable possibilities for the transfer of information. For the reasons already discussed, these cannot consist of the known possibilities for the exchange of CAx data in neutral formats, but should in any case also have an automated transfer of further relevant information.
Large manufacturers of CAx systems have already recognized this necessity for a long time and rely on a seamless cooperation of the modules in their respective systems.
Here it can often be trusted that almost all available information is also available for all process steps. The question of information exchange should always play a role when weighing up such holistic systems - which certainly also have their weaknesses - against a landscape of selectively ideal systems for the respective process step.