Historically, there hasn't been much of a debate on where to create 3D tool paths for mold machining because there hasn't been much of a choice. It took a complicated CAD/CAM system and a lot of concentration to surface and program these difficult jobs, a task best left to the guys with the clean shirts in a darkened room well away from the shop floor.
But that's all changing quickly with the advent of NC programming technology designed expressly for use in the manufacturing environment. This is not just old CAM technology wrapped up in a shop-hardened computer, but a very new approach to tool path preparation that represents a fundamental rethinking of the whole mold manufacturing process. Now it is practical, if not to move all NC programming out of the CAD/CAM department, then certainly to share that responsibility with the shop floor. Indeed, shops can once again place control of 3D tool paths back in the hands of an able machinist.
Bringing that control to the shop floor does much more than just making the machinist's job interesting again, though that benefit alone is larger than some may think. It also can take a huge load off of the CAD/CAM department, get jobs through the shop faster, and improve the productivity of both CAD/CAM and shop-floor personnel. Those are big claims for sure. But they are very real possibilities if shops are willing to invest a little and change the way they think about how NC programming has to be.
The Programming Tools
Contrary to how it may sound so far, this argument is not aimed at eliminating the CAD/CAM department's role in creating tool paths. Rather, the idea is to provide the shop floor with solid, easy-to-use programming tools, and then to allow the work to be done when and where it makes the most sense. Mold design and surfacing tasks are clearly best suited to the CAD/CAM department, for instance, which is not going to change anytime soon. And there is still a pretty good argument for creating a lot of tool path there too, though there are proponents on both sides of that issue.
But there are many cases where the machinist is simply in a better position to judge the effectiveness of a cutting strategy. At these times it makes an awful lot of sense to give the machinist the capability to create his own program, or at least to modify the one he already has that's not getting the job done correctly. It is here that new programming technology is providing an effective tool that shops have not had before.
The core of such technology is knowledge. That is, a system is constructed with a specific set of manufacturing processes in mind and imbued with a great deal of methods information appropriate to those processes. This frees the user from much of the mundane and repetitive decision-making typical of conventional interactive programming and allows him instead to focus on selecting the best general machining strategy for the task at hand. In die and mold work, rather than forcing the machinist to think in terms of mathematical descriptions of form, he is presented with graphic images that look like the metal he is about to cut. And he is presented with process options much more akin to the kinds of decisions he wants to make in his very physical world: Is a climb cut or conventional cut appropriate here? Should this tapered wall be finished with an X-Z planar cut, or is constant Z the way to go? Spiral or zigzag? A true knowledge-based system will not only recognize which decisions are important, it will suggest appropriate processes within context and automate much of the routine decision-making for the machinist that is really just a function of the specific circumstances at hand.
A good example of such technology is the Prospector system from Cimlinc (now Softech, Inc., Troy, Michigan). Designed for making molds, Cimlinc refers to the product as "total process optimization software" meaning both that many best-practice mold machining methods are captured in the database and that they are readily accessible to the operator on the shop floor who is creating or modifying a program.
In a typical shopfloor programming scenario, such a system may function first to graphically verify a program that has been sent down from the CAD/CAM department. Then the machinist uses it to create new tool paths where he sees problems that must be corrected or opportunities to otherwise improve the process. As Cimlinc's Brian Lambert puts it, "With the aid of a knowledge-based tool path creation facility, operators can quickly modify the program to reflect real-world conditions that might not have been apparent to the original CAM system. In effect, we're moving process planning right down to the shop floor where the adjustments can be made in the form of real-time process optimization."
Much of the programming process is presented in the form of "Wizards" that lead the operator though a rational decision-making process and suggest appropriate alternatives along the way. In cavity work, for instance, the best tool is usually the shortest tool capable of doing the job, and the Wizard will guide the operator into choosing the shortest combination of tool, toolholder and adapter.
This programming methodology could be applied to an entire core or cavity. More likely, though, it is going to be applied to certain features within the workpiece. In this case, the shopfloor programmer calls up the workpiece model, finds the area in question, and draws a "window" around the specific area to be reprogrammed. A window is simply a 2D or 3D boundary that constrains the cuts to be made. Thereafter, the system automatically figures all the tool paths based on the operator's cutting strategy.
The system models both the target surface (that is, the final workpiece model) and the stock so that with the generation of each cutter pass, it knows what stock remains. With this information it can then go on to automatically generate "airless" tool paths to clean more material with each new cutting tool or Z level. Moreover, the picture of the process that is presented to the operator on his PC monitor looks just like the material on his machine. The remaining stock can be highlighted in a color solid model representation, and at any Z level of the operator's choosing, a capability that can be used to quickly ascertain the right cutting depth to pick a surface feature or combination of features to be cut.
On The Floor
But how does shopfloor programming play in a real shop environment? After all, we're talking here about shifting responsibilities, and work, from one department to another, seldom an easy proposition under the best of circumstances. And we're asking machine operators to master a skill that has ostensibly been far removed from the shop floor since the onslaught of CNC.
At Summit Services (Sterling Heights, Michigan), shopfloor programming seems to be working out just fine, both for some reasons they anticipated and for some they did not. Summit Services is a 25-person mold shop that builds tools exclusively for the two plants of Summit Plastic molding, a tier-two auto supplier. According to shop manager Gary Verrier, they moved to shopfloor programming out of basic necessity because they were having trouble finding skilled programmers for the CAD/CAM department and were looking for some way to lighten the load.
And it worked, both from the CAD/CAM and the shopfloor perspective. According to CNC manager Rene Bossio, shopfloor programming helps the CAD/CAM department become more efficient in several ways. For one thing, it's cut time out of the planning process that used to be necessary for the machinist to provide his input on cutting strategy and for the NC programmer to familiarize the machinist with the details of the job. It's still important for the CAM room and the machinist to be on the same page as to general cutting strategies, but it's no longer necessary to work out every last detail on the front end. Now the CAM room can more quickly create a program and send it down to the shop where the machinist can decide for himself where he wants modifications, and go ahead and execute the change.
A critical component of this capability is the machinist's newfound ability to graphically preview the program on the shop floor. "It helps them visualize where they are going to cut since they now can actually see the path," says Mr. Bossio. The capability is also enthusiastically welcomed by the machinists who, before, sometimes had little idea what was going to happen once the cycle start button was pressed.
But Summit's machinists do far more than look at moving pictures; they use the programming capability to optimize programs on the fly, which can result in reduced cycle time and in a better machined product. Says machinist Dave Planitz, "As you're cutting the job, if it's not doing what you want, you can rewrite the program right there." Mr. Planitz asserts that the machinist has a particular advantage in this regard, since he is not imagining the cutting process as the off-line programmer must do; he is seeing and hearing it play out in real time. Mr. Planitz also suggests that an attentive machinist will be extremely familiar with how a given tool cuts best in a material, and will have a better appreciation for many other process subtleties, particularly as increasingly capable cutting tools and machines are brought into the mix over time.
And so Summit's machinists now play a larger role, not just as process executors, but as process "replanners" and improvers. If they see a path cutting air, they quickly edit it out. If they see that they can cut faster, they do it. When they see an opportunity to more thoroughly or smoothly clean an area, they open a window and generate the tool path right on the spot. And sometimes they find that they can completely machine features that were originally planned to be EDMed, and eliminate a time consuming operation.
Out The Door
The experience has been similar at J.B. Rath Co. (Sterling Heights, Michigan), a larger mold shop (140 people) just down the street. Also a captive shop, Rath is a division of Becker Group International, which develops and produces automotive interior parts and component modules. Rath makes more than 180 molds a year, all of which go to Becker plastic molding plants in the United States and Germany.
According to engineering manager Bill Hoover, Rath's initial interest in shopfloor programming had to do with increasing output and eliminating the backlog of their nine-person programming department. It was not that the programmers were inefficient, he thought, but that they were spending too much time in the final detailing of programs and in executing changes once a program was supposedly complete.
Both are tasks that the programmers are happy to give up to the shop floor. Even under the best of circumstances, Mr. Hoover believes it is often an exercise in futility for a CAM room programmer to try to get every last cut exactly right in a complex mold. "With surface data on a computer screen, it's hard to know exactly what you're looking at," he says. So assumptions are made along the way that may not turn out to be the best way to cut a tool. Then when it comes time to make changes to the tool path, that task becomes extremely disruptive to the programmer who has already moved on to the next job. "The biggest part is the engineering changes, which the CAM guys don't like to do," he says. Whatever the cause, turning those tool path revisions over to the shop floor affords CAM room programmers the opportunity to concentrate on big jobs all the way through with fewer interruptions, a focus that itself is likely to reduce mistakes being made in surfacing and initial tool path creation.
J.B. Rath has made good on the initial plan to speed up the flow through the engineering department. Indeed, once the shopfloor programming capability was instituted, the programming backlog disappeared quickly. But increasing the flow of jobs through the shop has proven to be just as significant. Before, when modifications to a program were required, it sometimes was necessary to pull the workpiece from a machine, physically mark up the metal to document the necessary changes, generate the new tool paths, and then set up the job again. The only alternative was to let the machine sit idle while the new program was being generated, and both options wasted valuable machine time. But when the machinists have the capability to program, once a job goes to the floor, it stays on the floor. And it invariably gets out the door sooner.
It's a point on which Summit's Gary Verrier completely agrees. "If there was a problem with the job, the programmer had to pull it back to tweak it, and that caused a lot of downtime. Now when a machinist sees that an area didn't clean up the way it should, he can take care of it right there. It helps throughput tremendously. You don't ever have to take a tool off the machine until it's done."
Uncovering Hidden Resources
Another point on which both shops agree is that shopfloor programming creates the opportunity for skilled machinists to once again directly express their own considerable process knowledge, which is as personally satisfying to the machinist as it is economically satisfying to his employer. Mr. Hoover, himself a product of the shop floor, will hire no programmer without machining experience because he feels so strongly about the necessity of that first-hand knowledge. Mr. Verrier, who can only go so far with the notion of artificial intelligence, puts it succinctly: "If you don't know how to cut already, that computer isn't going to tell you how to do it."
Actually, in time intelligent programming software will probably do just that, but not quite yet. For now there is still no substitute for the flexible application of the process knowledge that resides in the mind of someone who cuts metal every day. The problem for many shops, though, is that they have no practical way to quickly get that knowledge codified in a workable part program. No way, that is, until now.
And there's another important benefit of shopfloor programming. Giving the machinist the opportunity to express his knowledge in a useful way makes his job more engaging and more fun. Maybe to a hard-nosed manager that sounds a little too squishy to be regarded as a real business asset. If so, consider instead that the shop holds a vast reserve of unused NC programming capacity, an asset that most mold shops these days find pretty hard to come by. But it's sitting right there—literally—with knowledgeable people who otherwise can do little more than watch the machines run.
Now the management question becomes, how do you go about tapping that reserve? If these two mold makers are any indication, just give the guys in the shop some better tools and a little training, and they'll take it from there.
It's Not Just About Programming
Is this story about more than how to create tool paths faster? Cimlinc vice president Jeanné Naysmith thinks so. The core issue, he believes, is not just to apply a single point solution for a narrow application. Rather, it is to strategically apply appropriate new tools that allow shops to best leverage their entire array of existing resources in order to create a more efficient mold making process overall.
It's about rationalizing all of the shop's programming technology and process expertise, and having the flexibility to quickly apply the right solution at the right time. It's having the ability to respond quickly to varying workloads in the CAD/CAM department or the shop floor. It's about process documentation. (To that end, the Prospector system also includes the capability to create electronic project files that include all the relevant information necessary to complete a job.) It's about the ability to do continuous process planning if necessary—planning the next cut within the context of how the last cut actually worked, and executing that plan immediately.
The short and long of it, says Mr. Naysmith, is that this new breed of programming technology "allows you to immediately enhance the process you already have in place, and to migrate to a new way of doing business over time." But change they must if shops hope to realize the larger potential that this technology puts within grasp.