PC-based solid modeling CAD systems are making waves today. When properly implemented, these new products offer enormous speed and capability at a fraction of the price of workstation-based solid modelers. This new breed of solid modelers is having its greatest impact on small- to medium-sized machine shops that can now afford to own the software necessary to directly import their customers' solid models for machining.
Once a solid model of a product has been completed, there still remains the challenge of machining the components, molds or dies. Machinists understand the process of NC programming and machining and generally work most efficiently when they can use a common CAM system from job to job, regardless of which CAD system generated the initial file. Therefore, using an integrated CAD/CAM product to machine the computer model may require an unnecessary amount of learning time to gain proficiency when a universal, best-in-class, PC-based CAM system is often the more efficient, lower cost solution.
There are many approaches to machining a complex part, each of which is ideally suited to a particular range of applications. Shops typically must deal with a broad assortment of machining challenges, however, and it only makes sense that they have an equally broad understanding of the CAD models they import as well as of the CAM tools at their disposal. The following is an overview of the critical aspects of both issues, as they relate to today's powerful PC-based technology.
A good way to begin is to understand how CAD systems define 3D geometry. There are three basic ways:
* Wire frame models are simple stick model representations of a solid. Wire frame modelers that only save the stick model do not work with the true mathematical surface data, but may be able to produce shaded images, though not reliably. However, they cannot produce very accurate 3D tool paths because not enough information is stored in the data file.
* Surface modelers create true mathematical models of the surface of an object. The object may be represented graphically as a wire frame structure, but in this case, the entire continuous surface is defined with mathematics. Even though it was constructed of many individual entities joined or linked together, it is now defined as one complex surface. The important difference here (between a surface modeler as opposed to a wire frame model) is that the surface model mathematically describes the entire surface, not just the points that make up the wire grid.
* Solid modelers also have the ability to represent surfaces. However, a solid modeler also saves the information about how the various surfaces meet, which allows for Boolean operations such as addition or subtraction of volumes. A solid model is created by using cubes, cylinders, cones, spheres or extrusions of shapes. It is a fully enclosed model and has volume.
Solids can be added, subtracted or intersected with one another to create the desired finished part. Entire edges of a part can be easily rounded with one command. In the same way, intersecting surfaces can be filleted. All edges are automatically trimmed as the model is constructed. Solid modeling, however, is not without limitations as CAD/CAM Publishing pointed out in the October 1995 issue of CAD Report: "...no solid modeling system is yet capable of doing complex free-form surfaces well. If your products look like a car dashboard or fender or even Microsoft's curvy mouse, you'll need to stay with surface modeling for a while longer." The technology necessary to machine surfaces directly from solids is still lacking because of the inherent problems of doing the tens of thousands of Boolean operations necessary to create tool paths from solid geometry.
Some solid modeling systems are based on a concept known as boundary representations (B-reps). It is important to remember that a B-rep solid is simply a collection of trimmed surfaces that allows for operations that span more than one surface. The fundamental mathematical definition of the surfaces is the same in solid models as it is in surface models.
Regardless of how solids are created, operator visualization of models can be greatly enhanced if the system has a fast, interactive graphical display. Wire frame representations of surface models can be quickly rotated to visualize the entire model, then stopped at any point, and rendered in seconds. In addition, display and manipulation of a shaded surface model is now possible on a PC with CAM software that has been written to support Open GL, the 3D Graphics Library defined by Silicon Graphics. This option works best with a high-end GL graphics card.
Importing Surfaces And Solids
Because customers are supplying both surface and solid models to machine shops, a truly universal CAM system must be able to import both types of geometry. And this capability will only become more important as more companies move to PC-based solid modelers. Terry Wohlers, an independent consultant with Wohlers Associates, had this to say about CAM's role in the onslaught of Windows-based solid modelers: "As this market develops, so will the need for CAM products that can machine the model data produced by these solid modelers. CAM products that support SAT, STL and other popular file formats will be the favorites. Those that don't will be left in the dust."
If the terms SAT and STL are unfamiliar, it may because they are relatively new to the commercial world of CAD/CAM. Both provide a means of exchanging solids between systems, and one is particularly important as it is favored by a growing number of ACIS-based solid modeling systems.
ACIS is a B-rep type of solid modeling kernel developed by Spatial Technologies, (Boulder, Colorado), which many popular PC-based solid modelers use as their mathematical foundation for defining geometry. Any system using ACIS as its geometric engine can save its files in the SAT file format. For solid models created on ACIS-based CAD systems, the ACIS SAT file format can be used to bring a solid model into the CAM system as a surface model. Solid modeling systems that do not use the ACIS kernel use other file formats such as STL.
If the solid modeler is not ACIS-based, then files may be imported via IGES. An entire solid model can be represented in IGES as trimmed surfaces (IGES 144 entities). For machining, it is unnecessary to stitch together these surfaced faces because they have already been trimmed to one another. The volume data which the original solid model contained is lost when the data is converted into the IGES surfaces format. This is no cause for alarm, however, because volume data is not used in machining. For example, to machine a die cavity that produces an auto fender, only the skin of the fender is machined, not the volume that would be represented were it a solid object.
Some CAD vendors interpret entities in a slightly different way than is defined by the IGES standard. Some create entities that are not compliant with any of the standard IGES entities. These variations on IGES standards are sometimes called "flavors." In order to be of use for a wide variety of CAD systems, an IGES translator must be user-configurable to accommodate different flavors.
Machining Complex Models
There are two primary methods for cutting complex models:
1. Generating the tool path from the surfaces and the tool geometry while applying edge protection to the surfaces, or
2. Generating the tool path directly from the solid, using Boolean operators and accounting for the sweep volume of the tool. The direct, swept volume method of generating tool paths on a solid requires a substantial number of calculations based on the intersection of the solids involved because it must be continuously calculated as the tool moves along.
Because the mathematical solution used to directly cut solids is complex, the likelihood of encountering an unsolveable situation on a complex portion of the model is many times greater than if the tool paths are being generated from a surface model. When solid modelers are unable (for whatever reason) to perform an operation, they will return an error message to the operator. This can happen regularly on a simple model built with only fifty solid operations. Given this, consider tool path generation on a complex model where it is not uncommon to generate 10,000 to 100,000 individual tool path moves! The likelihood of one failure on this many moves is very high, therefore direct cutting is not yet practical. The ability to cut solids directly with the same reliability that we can currently cut surfaces will probably not be available for a few more years.
The most flexible method used for cutting complex models tessellates the part into a multi-faceted polyhedron. This method is flexible because it works on surfaces as well as solids. The surface of the polyhedron consists of a large number of triangles (facets). The number of facets is determined by the user's input of the accuracy requirement of the tool paths. The accuracy for a finished tool path is typically between 0.0005 and 0.0001 inch, but can be smaller if desired. The mathematical solution of the intersection of the tool geometry with a series of facets (which are actually a series of planes) is much quicker and more reliable than the method required to generate tool paths direct from solids, hence the tool path generation is many times faster. The way in which a CAM system approaches the mathematics required to generate tool paths is the single biggest factor determining the system's speed. No amount of software coding wizardry or lightning speed hardware can make a CAM system the fastest if it fundamentally approaches the problem of tool path generation using a slower technique.
But mathematics is only half of what a good CAM system is about. The other half is generating accurate tool paths that take good machining practice into account. And to this end, there are some important features that more advanced systems provide. For example:
* Gouge-Free Tool Support--It's important that any CAM system offer tool geometry support for squarenosed, bullnosed and ballnosed end mills, with full gouge avoidance on each. Each tool has advantages in machining various shapes. The better CAM systems do not limit the operator's choice of which tool is best for the job. Rough cutting with bullnosed tools is faster, and depending on the situation, bullnosed tools can also be effective for finish cutting.
* Surface Edge Protection--Individual surfaces that make up a surface or solid model may have small or sometimes large gaps. This is true even with a solid model if there are free form surfaces involved. How the gaps are handled differs among CAM systems. Some simply will add a straight line between the surfaces on either end of the gap, but this can be dangerous as it may gouge. The more reliable method is to "edge protect" each of the individual surfaces. With this method the tool path is guaranteed not to gouge the interior or edges of the surface.
* Tool Path Bounding--The ability to limit the extent of the tool path, so that only a portion of the model is cut, is an extremely useful CAM feature. Any arbitrary boundary should be available to the operator. For quick tool path generation, the bounding needs to be done at the time of tool path creation rather than having to wait until the full tool path is generated. That's particularly consequential in cases where a portion of the model is all you want to cut anyway.
* Fixed Distance Step-Overs And Scallop Height Control--CAM software should allow the user to both specify a fixed distance between each cut or to specify the maximum scallop height and maximum surface tolerance. The CAM system then generates the proper step-over to develop the shortest possible NC code and the fastest machining cycles possible.
* Feed Between Moves--Tool path generation usually follows rows as it cuts the parts. When the tool comes to the end of a row, the CAM system must decide how to position the tool at the start of the next row. Sometimes a simple "up and over" algorithm is used, but watch out, this can gouge the part if the system does not check for possible gouging. Better systems will have the option of following the part geometry in the transition to the start of the next row, in addition to full gouge avoidance.
* Gouge Checking Of Lead-In And Lead-Out Moves--The lead-in move refers to the positioning of the machine tool at the start of the first cutting row. It is important that the CAM system have a variety of options in generating this move and the corresponding lead-out move. The CAM system should ensure that both the lead-in and the lead-out move do not gouge.
The common methods of cutting complex models, whether solids or surfaces, are z-level roughing, z-level finishing, planar, single surface flow, multi-surface cutting (which also includes radial cutting), and project cutting. To give the machinist optimal control, the CAM system should be able to cut any specified portion of the entire model using any of these cutting methods.
Z-level roughing is the most efficient way to remove the majority of excess material when cutting a part. This method slices a model with constant z-level planes to generate paths which can then be spiral or zigzag cut. If there are overhanging areas on the part, it is important that the CAM system avoid gouging with the shank of the tool in addition to the tip.
Z-level finishing is the best way to finish areas of a part that are almost vertical. It is important that the operator can apply a boundary curve to limit the area of path generation.
Planar cutting is the oldest and most well-known method of cutting complex models. It allows the operator to generate parallel planar tool paths over any number of underlying surfaces. Current systems vary widely in the speed and quality of the tool paths generated. Even systems that claim to generate gouge-free tool paths may fail and gouge on complex shapes. If you are choosing a new CAM system, be sure to test the system on the most complex part you are likely to cut.
Single surface flow cutting may be the preferred method of cutting when a part's geometry is defined by only a few surfaces. This type of cutting follows only a single surface at a time, but the tool path can be gouge-checked against any number of surrounding surfaces. This method leaves the best surface finish.
Multi-surface flow cutting is a powerful feature and worth inquiring about. Flow cutting can be used to great advantage to optimally machine a part that is boomerang shaped, or has a bend, or many bends. The tool paths can be created so they flow along or across a specific flow surface. Flow cutting can be used to produce fine detail on a part, mold or die that is shaped like a wheel. For example, the detail on a wheel mold (or any round or oval shape) can be cut with the tool paths direction controlled to whatever will produce the best results. This also allows for radial cutting and cutting around the circumference of a part.
Project cutting is useful when machining an irregularly shaped part (or portion thereof) with a particular 2D tool path, as for engraving. It is important that the CAM system can create this tool path, project it onto the complex model, and cut it with full gouge avoidance.
But which of the above mentioned geometry modelers and cutting methods are most important? The truth of the matter is that they all are, or at least may be, depending on the job at hand. Smart shops will avail themselves of CAD/CAM technology that is both powerful enough and flexible enough to handle all new challenges as they come. The good news is precisely that technology is affordable today.
About The Author: Larry Diehl initially developed SURFCAM in the 1980s. He continues to program daily as Surfware's chief software developer and oversees all product development for Surfcam.