Downtown Muncie, Indiana, has, for some reason, a remarkable array of jewelry stores, many—if not most—of which are abandoned. No more metalworking inside of these cobwebbed facilities. But just a couple blocks away from these former shops, there is a facility wherein metal is being worked with the exquisiteness of a jeweler, but in this case, the material is not gold but hardened 4140 and H13 steel that's being crafted, and the pieces being turned out of Delaware Machinery and Tool Company don't weigh fractions of an ounce, but 80,000 pounds and more. Sometimes the 75-ton overhead cranes in the 150,000-ft2 plant site are strained toward their limits.
At Delaware Machinery the 200+ people are involved in making molds. Mainly, these molds are for large automotive applications, such as those for producing aluminum engine blocks or transmission cases.
Delaware has been in business, and privately held, since 1938. In the early `80s, the present CEO, Robert A. Haas, Jr., decided the company would position itself as being the best in the business by providing molds that were different than those that were then generally available. These molds would provide interchangeability.
How It Was
Historically, a given mold, James McDonald, Delaware assistant superintendent, explains, really was one-of-a-kind. Slides, cores and details were precisely machined and then hand-finished for a particular mold. When there was a problem at the user's facility with an element of a mold, it was necessary to go back to the mold maker so that a tailored replacement piece could be produced. Haas's concept was that this would be unnecessary because every mold would be made the same way. This meant tightening up on tolerances so that there would be dimensional accuracy throughout the mold assembly. Slide components and details were made so that when a problem occurred, the user could take out the problem piece and insert a spare.
To do this with reliability and consistency, Delaware began making an investment in computer-aided manufacturing (CAM) software and hardware. "People were using CAD [computer-aided design]," McDonald recalls, "but they were creating models and then using duplicating machines." These models and machines couldn't provide the repeatability that Haas was looking for. Now the machine shop is full of an array of EDM machines (Elox . . . Makino . . . Mitsubishi . . .), graphite machining machinery (including an Okada GMC 1088, a vertical-spindle machining center for electrodes, the first of its type delivered to the U.S.—and Delaware received it in May, 1997), machining centers (such as the Makino 1513, which is fitted with a six-pallet changer; it arrived in March, 1997), coordinate measuring machines (three DEAs and a Brown & Sharpe), what may be one of the largest collections of Cincinnati Gilbert boring mills to be found anywhere, a Helisys rapid prototyping machine (McDonald says they'd been unfamiliar with the laminated object manufacturing [LOM] process, but saw the machine at IMTS `96 and bought it)...and this is just the tip of the iceberg.
McDonald suggests that the necessary investments in CAD/CAM technology had the effects of (a) putting some tool and die shops out of the big mold market or (b) making those that remained that didn't make the investment not nearly as competitive as those companies that did. In his estimation, there are three other competitors in North America that Delaware tends to encounter when it is making bids.
Although the issue of mold consistency has been addressed by tighter tolerances on components and surfaces, the mold is just one aspect of the casting process. There are a number of variables involved in diecasting. Take the machine's condition, for example. A new machine is likely to be tighter all around than an old one, so, for example, there can be a big difference between the two machines when it comes to clamping pressure. There is the temperature of the molten material. There's injection timing. And so on. McDonald says that if you put the same mold in three different machines, you'll get three different parts.
The procedure had been that a customer would send prints (or a CAD file: Delaware has 20 workstations—Sun SPARCstations, mainly—and CATIA for Chrysler, Computervision PDGS for Ford [it hasn't moved to SDRC I-DEAS yet], and EDS Unigraphics for GM). Delaware (or any of its main competitors) would make the mold as specified and send it to the customer. The mold would go into a tryout machine at the customer's facility. Then, typically, it would be back to Delaware for modifications because it was discovered that the part as-cast wasn't what was expected. Remember: this mold is one that has been made with CNC equipment to tolerances within 0.005 in. Sometimes it was discovered by Delaware personnel that the mold was dead-on to the requirements as submitted by the customer. The issue was with the machine's variability, not the mold. It could have been just a matter of making an adjustment to a particular setting.
One thing that occurred at Delaware in 1987 was the creation of a new business, Matrix Technologies, which developed an air gaging system that can be used to monitor the behavior of machine slides during casting. (Why air gaging in this era of electronics? Simple. Think of the heat and pressures involved during diecasting and the likelihood of a silicon-based sensor lasting under those conditions.) This capability allowed Delaware people to zero-in on the problem areas that occurred with the mold that was made right but not performing as expected. They use the Matrix equipment at their shop; it is also made available to customers who want to have better process control at their own facilities.
One thing that has happened at engine and transmission plants of late is that the people there have realized that it isn't particularly efficient in terms of time and space for this back-and-forth of multi-ton molds. Not only is it time-consuming, having a tryout machine on site wasn't as useful as having an operating diecasting machine. So they asked the mold-makers to do the first-piece sampling.
One way that mold makers could provide this sampling is to contract with a third party that happens to have a diecasting machine that's of the appropriate size and which has open time when it is needed for the sampling. Realize that mold making shops tend to be filled with EDM machines and machining centers, not diecasting machines. This had been the case at Delaware. And going to a third party was what they did in order to meet this sampling requirement.
But Bob Haas had been thinking about doing something else at Delaware, and the customers' requests for sampling turned the concept into a reality. An on-site sampling center was constructed. Some $12-million were invested in the facility, which includes three Prince Machine (Holland, MI) diecasting machines (800-, 2,000- and 3,500-ton units) that were installed in 1996. The machines were built by Prince per Delaware's requirements, to accommodate the needs for different locking tonnage's, shot speeds, etc. McDonald says that the people at Delaware had originally figured that it would be possible to have just one diecasting machine that could handle 75 to 80% of the molds, but then they realized that more equipment was necessary.
As Chrysler Kokomo, a transmission manufacturing facility, had a major mold order with Delaware when the center was being established, one of the machines was setup to duplicate a cell in the Kokomo plant: the same Fanuc robots (S-420i's), degating press, and controls. This way, they can have full assurance that the molds will perform as expected under Chrysler-like circumstances.
"We would like to do more than sampling for our customers," McDonald says. "We would like to do R&D, as well." The sampling center is the place where they can do this. What could happen is that not only would Delaware supply a mold to a customer but a "recipe" along with the mold, instructions that would include all of the necessary settings for good parts.
Will it happen? Well, as diecasters look for reductions in scrap rates, it is entirely likely.—GSV
How Long Does It Take?
According to Jim McDonald, it can require from 10,000 to 12,000 hours to produce a mold for diecasting a large aluminum automotive component. A similar amount of time can be involved in the pre-manufacturing programming, too. Consequently, multiple dies per order can help spread some of the costs.