The U.S. Air Force had a tight budget for its F-117 fighter jet. Because they couldn't afford a clean-sheet-of-paper-design, Lockheed Martin engineers reused a number of major components from previous designs. By reusing the flight control computer and actuators, servomechanism, pilot seat, heads up display, and a number of other parts, Lockheed built two prototype planes for $30 million dollars. "At that time, a typical advanced technology prototype would normally cost three to four times as much," says Donald Reinertsen, product development consultant for Reinertsen and Associates (Redondo Beach, CA). "It was a brilliant example of reuse."
The concept of reuse has been around since the steam engine, Reinertsen adds. But people are taking it more seriously as the market place becomes increasingly competitive.
Planning for modularity at the beginning of a project means countless benefits for a company when manufacturing future products, says Paul Stasi, Senior Corporate Director for New Product Life Cycle Process at KLA-Tencor (San Jose, CA).
For the last two years, managers at KLA-Tencor, a semiconductor inspection tool manufacturer, have pursued the concept of modular design for software, and in the upcoming year will launch an initiative for hardware. "Two positions within the organization—a Chief Technology Officer for Systems and a Chief Technology Officer for Software—drive the philosophy as well as the process application of module design," says Stasi.
Writing modular code and using it for similar designs was initially quite challenging. KLA-Tencor is composed of 17 divisions, within which are several system groups. "We had System A in Division 1 using UNIX, while System B in Division 3 used DOS, and System C in Division 10 used NT," says Stasi. The trick was getting everyone on a common platform and writing in a common language.
Once this was accomplished, an engineer might write a thousand lines of code to direct a machine to inspect a certain pixel area on a wafer. Three or four divisions within KLA-Tencor could reuse such a program to look for a different defect in that same area.
Developing modular hardware was a different challenge. As the organization becomes centralized, KLA-Tencor is trying to develop a synergy for systems hardware. The future goal is for a systems engineering group to review, categorize, and departmentalize KLA's 70 different products on the assembly and sub-assembly level.
Implementing a program of this nature has associated costs, says Stasi, but the returns outweigh the risk. "If a typical product cycle is 14 to 18 months, and as a result of modularity we achieve market release two months earlier, that means another 3 to 5 bookings. From a revenue perspective, that can translate into another $3 to 8 million for the company," he adds. Even more important, modular design gives KLA an early presence in the segment.
To help facilitate reuse, KLA is presently evaluating enterprise applications. But in the short term, engineers use the CAD package Pro/Engineer from PTC. Pro/Engineer designs, for example, go right out to the manufacturing floor with all the associated documentation on the same system. If manufacturing needs to write an ECO to update a design, they work off of the paperwork that created the design.
"Modularity will not always make economic sense," says Reinertsen. "But a company must plan for it at the beginning in order to capture the benefits."
Eliminating chance. Most companies practice some form of reuse. And for good reason. There are enormous practical benefits, Reinertsen says. "While people are aware of the more obvious savings in time and cost, they are not always aware of some more subtle benefits. For example, reuse reduces the amount of unpredictability in the development process which greatly simplifies its management."
Another consideration is queuing. "A development process is a bit like traffic on a busy highway," says Reinertsen. "Remove one lane from a four-lane highway and you reduce the capacity by 25%. But you double or triple the cycle time."
No one would ever load a factory to more than 85 to 90% utilization because they know it would be impossible to get the product out the door. "But companies routinely load the engineering process to 99% and then managers scratch their head wondering why programs don't come out the door on time," says Reinertsen. Implementing a certain amount of reuse can help alleviate engineering overloads.
Richard Chilton, service-manufacturing engineer for Leatherman Tool Group (Portland, OR) agrees. Chilton designs the automatic feed systems for Leatherman PST or pocket survival tool punch presses. The PST key chain or microtool turns into a pair of scissors, pliers, a knife blade, or file. "Reuse takes away the chance of crashing the die and eliminates the unknown," he says. "Any time we reuse an existing design we save time because we don't have to reprove or rewrite tool programs."
To make the original PST handle, Chilton devised a flat condition stamping process with an automatic feeding system. When Leatherman wanted to release a similar tool whose handle had a slightly different width, length and shape, Chilton just tweaked the original process.
To help them in their reuse process, Chilton and associates use Inventor from AutoDesk. "With this program, we call up the original sketch, modify the necessary component, and the whole drawing changes accordingly," Chilton says. "Inventor also allows us to do a full 3D rendered design that shows the tools in development as they will appear as a product."
"I am a firm believer that in the mechanical sector, most true inventions have already been invented," says Chilton. "New ideas are just using existing designs in a different way."
For more information on Design for Profitability, Enter 548