×

Manufacturing

Extend CAD to CAM

30 Jun, 2003 By: Mark Huxley

---

Capable, affordable, and not just for architects anymore.

More Information  • Vendor Guide [PDF] Additional Reading In previous issues we've discussed the use of the digital model for such things as reducing time to market, obtaining and tweaking mass properties calculations before any physical parts are produced, virtually analyzing and improving your product early in the design phase, and obtaining physical models as fast as possible through rapid prototyping. "Today's Solid Modelers?Key to Art-to-Part Success," January 2003, Mark Huxley and Steven Weisberg "Desktop Rapid Prototyping," August 2002, Mark Huxley and Steven Weisberg,

Among the benefits of a 3D digital model is the ability to leverage it to create molds or actual parts in the portion of the design process called CAM (computer-aided manufacturing). Advances in CAD technology help countless users, with varying degrees of mastery, create increasingly complex geometry faster than ever before. Likewise, the tools used to create tool paths that match this topography have also progressed. Even if you ignore all the other benefits (see additional reading), the ability to fabricate parts based solely on a 3D model file is a compelling reason to venture into solid/surface modeling. Of course, there are always limitations. One major issue is the exchange of 3D model files. 3D modeling applications use various mathematical engines called geometry kernels to define their 3D file formats. The ideal setup for exchanging files between CAD and CAM occurs when the applications use the same data file format, thereby eliminating translation issues. Unfortunately, that isn't always practical in our heterogeneous CAD/CAM software environment where others in the workflow process use applications that differ from yours.

All of the CAD data in the world is created by about two handfuls of proprietary and licensed geometry kernels. Three mainstream avenues help you get your data from the CAD to the CAM system:

• Native file formats
• IGES (an industry standard)
• STEP (a newer industry standard)

Figure 1. The yellow edge shows that two surface edges in the domed area are not knitted together properly. Zooming in shows the problem area more clearly. This problem occurred because of mismatched tolerances between the CAD and CAM software packages.

Whether done explicitly or via a background process, each can pose problems. Failures are akin to a stitched quilt that's been ripped apart. The quilt in this case is a "watertight" group of surfaces that represents the exterior shell of your model (figure 1). Disparate tolerances between packages can be tuned, but this takes time and can be hit or miss. Some systems require you to make these settings prior to the creation of geometry-that is, changing the setting won't rebuild the existing model geometry to the higher/lower tolerance. This can be a problem if you don't know who is going to be making your parts.

Regardless of whether the geometry was created at an "inferior" tolerance or the translation process has corrupted it, the CAM software you are considering should definitely provide some sort of geometry healing tools to repair the data. How extensive, useful, and automated they are may or may not be important to your company. If you deal with data from several different CAD packages and many different suppliers, this will be of vital interest to you. If you create all of your geometry in-house, with a single CAD program, translations may not be an issue at all (figure 2).

Integration can offer one seamless window that addresses both your CAD and CAM needs. This eliminates problems that occur when changes are made in one program and not passed on to the other. With integrated associativity, the same model can be used directly by both

Figure 3. Unigraphics, PTC, and Delcam each offer capable software suites with impressive CAM packages. Unigraphics (below) performs machining operations on a gearbox casting. PTC (top right) shows its range of process-oriented applications that span PLM (product lifecycle management). PowerMill from Delcam (bottom right) offers the choice of CAD, CAM, or both from the same vendor.

programs. The ability to update changed geometry rather than reprogram it is a huge benefit. You can obtain this type of associativity by using a software suite developed by a single vendor or by purchasing compatible products from separate sources.

DO YOUR HOMEWORK
What should managers consider when evaluating CAM software? Many of today's programs have a similar look and feel, yet are worlds apart in strengths and weaknesses. Ease of use may get you up and running quickly, but is usually associated with a lack of options that may hinder future growth. As capabilities increase, generally so does training time.

Many qualified vendors offer CAM applications. The vendor table [PDF] shows a cross-section of offerings-large to small, bare-bones to world-class-with a staggering price range of $600-$60,000. Note that the prices don't reflect an apples-to-apples comparison. Some include the bundled price of full-fledged CAD systems along with the CAM software. Some applications have been in development for 35 years, while other companies have just introduced their codes.

Obviously, the abilities of the different products vary greatly. You don't expect a rusty old Datsun pickup truck to perform like a new Mercedes. Though each has its place in the world, it would be foolish to expect a $600 package to perform all the functions of software that costs 10 times or 100 times more. One vendor says that you can get "80% of the functions for 20% the cost" of its competition. This is fantastic-if you won't be needing that minority of functions.

When you shop for CAM software, an important point to remember is the software must perform all the tasks you need it to. This sounds simple enough but if one key function is not present, productivity can suffer dramatically. It is wise to have some capabilities that you just may need at a later date. Erring slightly on the safe side ensures that you can meet the majority of your future needs.

Figure 4. Complicated geometry that requires 5-axis capability can be created and milled with equal ease or difficulty, depending on the tools you have. Here, Mastercam V9's automated multiaxis flowline machining hones a precise finish.

KICKING THE TIRES

• Does the system include the capabilities you need now and in the future?
  You want the right tool for the task at hand. Why pay for 5-axis software if your machine tools can't or won't be using it (figure 4)?
• What kind of stability and market strength does the vendor possess? Will they be around in 5 years? 10 years? Will the product be supported then? Seat count can provide a valuable reference point in this area. Many vendors cringe when asked for this data, fearing that just this one number will be used to judge the "king of the hill." Though not the best indicator of the suitability of the package for your company's particular needs, we've included it in the table to provide insight about maturity, prevalence, number of potential users in the job market, and so forth. A product that is widely used and accepted will likely be productive.
• Yet another translation? Are postprocessors included to translate the tool paths into G-code for your machine tools? If not, how much are they, and how are they made available? Is there a post builder in the CAM system? Will it accommodate customization, and to what degree?
• Are the postprocessors efficient? This becomes important on large jobs or where tool paths are more complex.
• Surface finish and accuracy capabilities. If necessary, are Class A surfaces supported? What kind of file sizes will you be dealing with? Will it take an hour to transfer the file to the machine tool? Will your employees have to waste time chopping the file into several parts to make the transfer?
• Is the system user-friendly, intuitive, and easy to learn? How much training will your users need to use the program efficiently? There is a very wide range in estimates from the listed vendors. Some say that no training is necessary because their code is so easy to use. Online help and tutorials may offer sufficient guidance. Others advise that you may need as many as four weeks to fully adapt the program to your methods.
• Investigate vendor and/or reseller support and training capabilities. Even a widely used application may not have a strong presence in your vicinity, so personalized technical support may be limited or pricey. Will your personnel be relegated to an 800 number for answers? Users with pressing questions who are forced to wait on hold or spend time bringing green technical support people up to speed can get frustrated quickly.
• If you have an existing CAD application, which CAM packages work well with it? If you make changes to your CAD data, will the tool paths adapt?

  Figure 5. Elapsed time estimates based on CAM simulations make it easy to schedule and plan your machine tool use. STEP-NC technology lets you program standard features such as those shown in green, blue, and magenta.

• To what extent will you be customizing the routines?
• How much time will the code take to run on the machining floor (figure 5)?

Once you narrow down your choices, ask the vendor sales team(s) to perform a real world test with one of your company's more intricate parts. You don't need to be General Motors to get this type of service. Be an active participant in this test. See where the time is invested. What goes smoothly and what doesn't? If your company performs a task differently from how the application engineer approaches it, see what it takes or if it is even possible to wrestle the code into adherence.

Once you get a good feel for some of these issues, you'll have a better idea of how much time and money you'll need to invest in training and setting up the software.

THE FUTURE
STEP-NC, or AP-238, is a proposed extension of the STEP (Standard for Exchange of Product model data) standard to support NC programming information. Basically, STEP-NC adds manufacturing data such as tooling, stock, and process sequence to the STEP design information, which includes geometry, topology, tolerances, relationships, attributes, and assemblies.

According to its developers, the goal of STEP-NC is to eliminate bottlenecks to manufacturing productivity such as cumbersome postprocessors, antiquated G- and M-codes, data redundancies, multiple CAD files, and more.

This fledgling standard, though still in the experimental phase, shows a lot of promise and is slated to become a draft standard this year. This is more significant when you consider the fact that about 1.5 million seats of software with standard STEP translators are currently deployed worldwide.

However, "though STEP-NC holds significant promise for the manufacturing industry, it is most likely five to ten years away from being a widely available, production-ready technology," says John Callen of CAM software developer Gibbs and Associates.

As it develops, STEP-NC is likely to face many of the same obstacles to adoption that beset the original STEP standard, described in detail in "STEP vs. IGES," Cadalyst, October 2001, www.cadalyst. com/features/1001 vendors/stepiges.htm.

Figure 6. A part similar to this was milled during a STEP-NC demonstration in January at NASA's Jet Propulsion Laboratory.

Still, recent demonstrations and the participation of manufacturers such as JPL (Jet Propulsion Laboratories) give STEP-NC significant credence. Dr. Martin Hardwick, president and CEO of STEP Tools, a company that develops STEP software tools (www.steptools.com), says, "At JPL, we illustrated what a typical manufacturing scenario would be like using STEP-NC (figure 6). Machinists will receive a STEP-NC part via the Web or traditional means, program the job with a logical step-by-step 'wizardlike' approach, and in return, get feedback on production schedules, machining strategy, tool selection, tolerances, and areas of complexity."

In June, NIST (National Institute of Science and Technology) co-hosted a STEP-NC meeting where a demonstration focused on a complex surface model provided by Boeing Aircraft.

STEP Tools also showed how to control a probing application using its new STEP Index Library. The free, downloadable library gives developers of open architecture controls and other applications access to the NC data defined by STEP-NC.

Figure 7. FeatureCAM's AFR (automatic feature recognition) function automatically recognized and programmed the yellow holes in this part.

AFR (automatic feature recognition) is a technology also in its infancy. Some vendors currently advertise feature recognition capabilities, in which you run a wizard or other tool to identify holes or pockets that can be automatically programmed (figure 7).

Current deployments implement somewhat limited AFR, such as prismatic shapes. In other words, feature recognition works only for simpler shapes. One user I spoke with says that his software recognizes about one-third of the eligible features-holes in his example-but not the remaining two-thirds. The AFR tool is not considered efficient enough to be used at his facility.

Continued development will assure that AFR works reliably and quickly. This is quite a chore considering the myriad of variations in a "simple" hole (blind, through, counterbored, countersunk, tapped, reamed, and so on). Add issues such as tolerancing and surface finish to the other variants, and things become complicated quickly.

Long used to supply materials to CNC machines and move parts from one station to another, robots are now poised to take over certain machining tasks. Delcam has been working with robot maker KUKA to develop machining by robot as a lower-cost alternative for large-scale manufacturing operations on softer materials. "The cost of installing a robot is far less than the price of a large machine tool with a similar working envelope," says Brett Green, general sales manager for KUKA. Robots are best suited to machining softer materials in applications such as pattern making and trimming composite components. Machine tools also have the edge in maintaining tight tolerances. Robots can be accurate to tenths of a millimeter.

Knowedge-based machining refers to the capture of machining information and reusing it for subsequent jobs and also taking it into account during the design process.

Consolidation. There has been a significant consolidation of CAD vendors in the marketplace during the past decade. CAD customers often find it impossible to switch programs quickly, and the surviving vendors have made good progress getting previously competing products to work together well. Interoperability between programs from the surviving vendors has improved considerably as well, despite some vendor's encrypting their file formats (something that serves them in the short term, perhaps, but not their customers). The CAM market will likely see a similar consolidation take place, but it would be hard to imagine the pace matching what was seen in CAD.

Price/performance ratio will certainly continue to fall as lower-end packages become more capable. You can achieve more today with $5,000 worth of software than you could with $150,000 worth of software ten years ago. The functionality provided by software suites that comprise applications from independent vendors is impressive. However, all-encompassing suites from a single company can be more stable and offer some key advantages. Bargains can be had if you understand your entire process. Buy what you need and can justify.