For its 26 machining centers alone, PHD's Huntington, Indiana manufacturing facility has 4,875 tools in its active library. And "active" is the operative word. The rate at which tools are swapped in and out of machining centers is increasing. Last year, the plant did 63,717 tool setups. This year, it will do more than 79,000. No matter how you look at it, this plant uses a lot of tools.

Yet just one tool presetter measures all of the relevant data for all of the tool setups for these 26 machines. The software associated with the presetter manages lathe tooling as well. Two employees, one for machining centers and one for lathes, serve as the gatekeepers who maintain the integrity of this information. In short, while the plant uses a lot of tooling, it has a tightly controlled and centralized system for keeping that tooling in order.

PHD started building this system about a decade ago. At that time, it wasn't clear just how important the system would become. The company's business is changing. This maker of automation components--including cylinders, grippers, slides and rotary actuators--is seeing lot sizes and leadtimes shrink, while the number of product designs proliferates. In greater numbers, customers are asking for just-in-time service at the same time that they ask for custom products in place of catalog items. These changes are good, because PHD feels particularly capable of meeting these demands. However, the response to the demands is effectively transforming the Huntington production plant, along with a sister plant in Fort Wayne, into something more like a job shop.

However, the difficulty is that PHD lacks many of a job shop's options. In a job shop, a smaller number of machining centers might have substantial tool capacity in each machine. The shop might equip these machines with a standard complement of general-purpose tools that could be applied to almost any job coming in the door. In other words, a job shop wouldn't have to swap out tools so much.

PHD can't afford these kinds of concessions. It can't afford to devote that much floorspace to tool magazines, and it can't afford to hold that much tool inventory in every machine. Nor can it afford the cycle-time compromises that come from using general-purpose tooling instead of tools specifically suited to specific details of the part. What this plant needs is a system controlled and responsive enough to handle a large volume and variety of tooling. The plant had the foresight to begin putting such a system in place in 1994.

Over the years, the system has reduced human error, reduced the plant's overall scrap rate and improved the change-over time between jobs. Today, this system is facing a challenge, but it's not a challenge related to effectiveness. The challenge has more to do with physical limits. Part of the system's elegance lies in the fact that one presetter can serve so many machines, but the plant is now running this presetter around the clock. At 79,000 tool setups, the plant is pushing the upper limit of how many tools per year a presetter can measure.

The first presetter that the plant installed, like the plant's current model, came from Zoller, Inc. (Ann Arbor, Michigan). Even though the model PHD was using in 1994 was quite possibly the most sophisticated presetter installed in the United States at the time, the technology has improved significantly since then. The plant still has this first model sitting in a corner, because the plant can't find a buyer for it. The current model, purchased 4 years ago, beats it handily in terms of both precision and ease of use.

At least a year went by before presetting was integrated into the plant's process in something like the way it is today. The presetter itself is only part of a package that also includes tool management software--a vital element for using the presetter well. Tooling technicians at this plant used that first year to populate this software with the shop's preferred tools. They assigned tool names and ID numbers, associated toolholders with the tools, and input nominal dimensions and cutting parameters for the plant's various workpiece materials. All of this information had to be entered one tool at a time, in spare moments as PHD's production continued. Only after a year was there enough information in the system that a sizeable proportion of the plant's tools could be called up from memory instead of being entered for the first time. The tool crib personnel called up tools in this way, but just as importantly, so did the programmers. Their ability to select from a common reserve of tooling saved them time and guesswork, and it made the process more consistent by ensuring that standard tools were used in standard ways. At about this same time, the presetter itself was connected to the shopfloor network.

There was resistance from the shop floor then, and understandably so. Operators had long been accustomed to keying in their own tool offsets, and in many cases, even measuring their own tools. Now they were being asked to hit "cycle start" on programs using tool data they had never even touched.