Moline, IL-A farmer brings in the harvest before the weather turns ugly. A heavy equipment operator pushes around a few dozen tons of earth every day. And a homeowner cuts the grass, hopefully finishing in time to watch a ball game. The size of their equipment may differ, but all three of these people have something in common: They all care about quality. And they care about it in a way that all too often has little to do with the way engineers think about quality. Forget about conformance to manufacturing specifications. Equipment users want machines that do what they should, run when they should, feel right, and look good. Taken together, these real-world quality attributes produce what Bob Moulds, vice president of engineering for Deere & Company, calls "fitness for use."
 |
As much as the nitty-gritty of manufacturing specifications do matter, "fitness for use" shapes the end-user's perceptions of quality. Led by Moulds, Deere's design engineers understand this fact and have made it the cornerstone of their efforts to deliver quality in products ranging from huge harvesters to tiny grass trimmers. "There's often a tendency for engineers to focus on what they know, rather than what the customer wants," says Moulds. "We do both."
And what do Deere's customers want? Performance tops the list. "If we don't deliver enough functionality, everything else is for naught," says Moulds, who worked as a quality engineer early in his career. Then there's reliability. It makes an obvious difference to farmers up against a shrinking harvest window and a bank loan. "Farmers tell us, 'Whatever you do, don't take me out of the field,'" says Don Yarbrough, Deere's manager of worldwide combine development. "Anything that interferes with uptime, we jump on like crazy." Reliability also matters to heavy-equipment and commercial landscapers, who view their machines as moneymakers. And even homeowners with lawns in the suburbs won't tolerate downtime-in part, because they probably don't know how to make even simple repairs.
Users also place stock in "softer" attributes, such as ergonomic design and even aesthetics. "Farmers can be very critical of appearance and design," Moulds notes. "Some even wash and wax their combines."
Finally the users don't want to overpay. Think "frugal farmer," and you wouldn't be too far off the mark. But corporate buyers of off-highway equipment and homeowners can be every bit as sensitive to sticker price. Says Moulds, "We have to deliver quality but do so at a low cost."
Deere engineers sow the seeds of quality through a formal product development methodology called the "Product Delivery Process," or PDP. "It is the heart and soul of our quality efforts," Moulds says. Equal parts common sense and high technology, the process combines classic customer research with sophisticated tools that verify designs both virtually and in the lab. And the process addresses quality concerns long before designs have been set in steel. "PDP recognizes that most quality problems start in the design world," Yarbrough says.
Bits and pieces of the PDP have been in place for about a decade now. But the process as a whole has started to pay off over the past two years as large-scale products emerge from it.
Customer is king. PDP begins with so much customer input-from focus groups and field research-that farmers, heavy-equipment operators, and landscapers just about qualify as bona fide members of the product development teams. "We try to get all the customer testers together in a room," says Yarbrough, "And we don't want any yes-men." Even as a design progresses into the prototype stage, users evaluate proposed design improvements.
According to Yarbrough, customers played an extensive role on Deere's 50 Series combines, which debuted early last year and represent the first fruit of the modern PDP process. "Customers influenced nearly every aspect of the 50 Series," Yarbrough says, before reeling off a list that includes how to clean the grain tank, how the chopper works, power, capacity, and servicing intervals. For the sake of aesthetics, one customer suggested that Deere raise a back wall six inches to hide the top of a black air cleaner. "He was right," Yarbrough says.
User input can get right down to the smallest design details. Brian Rauch, manager of advanced r&d in Deere's construction equipment division, recalls that one customer came up with a "really simple" solution to cleaning the tracks on the H-Series crawlers, another recent product of the PDP. "He suggested we recess the bolts on the crawler-track frame so that cleaning tools wouldn't bump up against them," he says.
Or consider visibility, a key quality attribute for those who run heavy equipment. Customers wanting better visibility at night drove Deere engineers to add sidelights to the recently developed G-Series Backhoe Loader. "That was a first in our industry, and we ended up filing a patent on them," Rauch says.
Perhaps the greatest design influence from users has been over ergonomics. Customers can highlight simple things that even the best engineer could miss. Yarbrough cites the location of the monitor within the cab. "Is it easy to see? Can you read it with bifocals? These are things a young engineer wouldn't necessarily think of," he says. On the 50 Series, for example, customer testers literally helped shape the control handle for the machine's hydraulic system. "We started with a text-book approach to the handle design," Yarbrough remembers. "But some testers said their arm ached after an hour using it." That feedback set off a trial-and-error design cycle that modified the handle's tilt, geometry, and button location until it felt right to users.
Only after Deere engineers fully understand what the customer wants, do they begin work on solid engineering specs, which can take two years to develop on large projects. Yarbrough observes that PDP guides Deere toward firm specs earlier in the design process than in years past. "We don't go through any big swings in specs like back in the old days," he says, adding that even major combine elements like engine or grain tank sizing might not be fixed until a prototype had been built. "Now customer input drives these sorts of major decisions early in the process," he says. And holding the line on specifications benefits quality. "Any late spec change is a reliability killer," Yarbrough remarks.
The translation between a customer's cries of "I don't want to break down," to the engineer's language of "mean time between failure," is where Deere's collective engineering know-how enters the picture. Every product-and by extension, every component and subsystem-has specific mean- time-between-failure targets that spring from years of experience. "If we meet those targets, we will end up with satisfied or very satisfied customers," says Klaus Hoehn, Deere's director of worldwide tractor and component engineering.
Making models. Though rooted in basic customer research, the PDP does have a high-tech side. Over the past decade, Deere has dramatically increased its reliance on design automation tools. The first step in this direction began when the entire organization-with engineers spread around the globe-moved into solids-based design and standardized on Pro/ENGINEER from PTC (Needham, MA). "It's the best thing the company ever did for us," says Yarbrough, echoing a view expressed by Deere's other engineering managers. More than just providing a common platform for design and the communication of design concepts-though this role can't be underestimated-solid modeling has also had a positive effect on engineering quality.
Early in the design process, Deere uses Pro/ENGINEER to see how its assemblies will come together and to analyze manufacturing tolerances. "We model everything, right down to the fasteners, to see how it all fits together," reports Moulds. These electronic "fit-up" procedures have made a big difference in complex combine or tractor assemblies, which could potentially contain dozens of moving-part conflicts. And it has likewise improved the tolerance stack-up that can interfere with the alignment of boltholes and other sheet metal features. Quality improvements result because electronic fit-up takes the place of a far less refined tool for overcoming poor alignment-the mallet. "There is always tendency in a manufacturing plant for people to make things fit," Yarbrough says. "With electronic tools, we know that things already do fit."
Solid models also serve as the framework for Deere's push to analytically validate designs. "We verify more and more designs virtually and only then go to hardware," says Hoehn. Indeed, there doesn't seem to be a class of analysis tool that Deere engineers haven't tried. They use computational fluid dynamics to model air flows-both in engine manifolds for performance and in the cab for operator comfort. They use hydraulics simulation software, mostly Easy 5, to size components and model oil flows. NVH analysis takes place too, particularly in the consumer division that produces lawn mowers. The list could go on. "Analysis has become a real job requirement for engineers on every product development team," says Yarbrough, who notes that Deere also launched a central technology group to help each division's design engineers with the most advanced simulations.
Of all the analysis work, fatigue simulation has had the biggest impact. "Deere is a leader in fatigue analysis-on or off the highway," says Tom Cordes, a former aerospace engineer who manages the Engineering Mechanics Department at Deere's technical center. The company's engineers have tried just about every commercial finite element analysis and dynamic structural analysis program on the market as they simulate the fatigue performance of equipment ranging from lawn mowers to combines.
All the structural analysis work doesn't necessarily boost quality directly. "We've always delivered a structurally durable product," Cordes notes. But virtual fatigue analysis does help Deere deliver that durability faster. As Cordes explains, developing a new piece of equipment a few years ago meant building a series of prototypes and running them for thousands of hours in the field-and more recently, in the lab. Computerized structural analysis provides some of the same data up front, allowing only relatively mature designs to make it into steel. "Analysis gives us an up-front understanding of load paths where we had to guess or find out in the prototype in the past," says Yarbrough.
And by flagging a given design's understressed areas without guesswork, structural analysis helps Deere engineers remove unnecessary material. "In the old days, we'd sometimes put on extra steel," says Yarbrough. This brute-force approach, while in keeping with Deere's focus on durability, pitted structural integrity against the low weight and cost that also define quality for users. Structural analysis eliminates that tradeoff. Or as Cordes puts it, "It allows us to remove material and cost without any risk."
In a prime example of how advanced engineering tools can address simple ergonomic concerns, Deere for the past five years has also used its solid modeling capabilities to drive virtual reality tools that put people in equipment that doesn't yet exist in steel. "The customer can interact with the machine before we build it," Rauch says.
Deere currently uses virtual reality tools to optimize the location of control levers and flag any impediments to operator visibility. And it also uses it to get an early indication of a design's serviceability. "You can reach underneath a transmission with your own arm to find the drain plug," says Hoehn. "Or we can check the location of the dipstick or air filter." Even better simulations are in the works. Hoehn reports that the company will soon have the ability to run virtual simulations of a tractor's ride characteristics over different types of terrain.
Fieldwork. Another ingredient in Deere's approach to engineering quality is knowing when to turn off the workstations and put genuine prototypes through their paces. Despite a growing confidence in computer models-which Cordes describes as "very, very good and getting better all the time"-some design qualities still elude virtual validation.
As an example, Dave Holm, director of advanced product engineering for consumer and commercial products, cites the design of a mower's cutting chamber, which remains a bit more art than science. "Electronic tools might get you 70% of the way there," Holm says. But electronic tools have trouble capturing the full variety of grass conditions-wet, dry, height, variety, time of year. "Testing is still a must," he says. The same goes for larger agricultural and construction implements, whose performance where steel meets the ground still can't be validated by electronic means alone. "There are some things you just can't see in Pro/ENGINEER, " Yarbrough remarks.
Deere relies on two types of hardware testing programs to put the finishing touches on its designs. One takes place in the lab, where Deere runs its Advanced Design Verification (ADV) program. ADV puts instrumented equipment prototypes on hydraulic test benches that can compress months of field testing into a few days. According to Yarbrough, Deere has built rigs that can accommodate equipment as big as a combine-and not just its individual subsystems. "We can simulate the worst field conditions, something like driving over a levee," he says. The other takes place in the field, where Deere's test engineers and selected customers put prototypes through their paces for months at a time in all different conditions. "We still go to the field," says Yarbrough, "because there are things only an operator can notice over the course of a season in the field."
Building quality. Poor quality may often start, as Yarbrough suggests, with poor design. But quality problems really come home to roost in the manufacturing plant. Deere takes its share of traditional quality control measures. All the company's manufacturing facilities use statistical quality control methods and have ISO 9001 certification, according to Moulds. Looking beyond these traditional approaches to manufacturing quality, PDP also plays a role. It addresses manufacturing concerns early in the design process by including manufacturing and quality engineers on all product design teams.
What's more, design-for-quality goals unearthed during the PDP's early research stages are pushing traditional manufacturing technology. Nowhere is this relationship between PDP and technology advances more apparent than with Deere's advances in casting. "We're pushing the envelope on casting," says Yarbrough. On the 50 Series combines, for example, Deere adopted a new cast iron feed housing, replacing the sheet metal design used on previous models. The new casting met goals that are good for customers and for Deere's bottom line. For one, the new housing resists wear. "Iron castings just don't wear down as fast as sheet metal," Yarbrough notes. And the complex geometries possible with casting added functionality too by enabling Deere engineers to optimize material flow through the housing. And finally, the cast part came in at a lower cost than sheet metal, which would have needed complex and expensive bends to achieve the same geometry. "Thin wall casting saved us a ton of money while improving wear life," Yarbrough says. He adds that the new housing, unlike its sheet metal predecessors, also comes out in sections for easier servicing. "This feature probably isn't as important since we've never had to replace one," Yarbrough quips.
Information gleaned from the PDP likewise drives Deere to increasingly replace sheet metal with composites (SMC) on the agricultural equipment exteriors. Yarbrough says these composites, such as a new composite shield for the rotor on the 50-Series, not only saves weight and cost but also stands up to more abuse than sheet metal prototypes, which tend to bend more easily. "Farmers throw these covers on the ground, " he says, adding that these shields play an important functional role by sealing against grain leaks.
With Deere becoming less vertically integrated with every passing year, the company's supply base has a growing effect on quality. By emphasizing early development of manufacturing specifications and by including purchasing representatives on design teams, the PDP steers Deere's engineers toward the best components. "The PDP includes a lot of work on supplier quality," says Moulds.
The development process also fosters a mentality that's best summed up as partner don't purchase. "We can't be experts on each and every one of 4000 parts that make up a tractor," Hoehn says. "So we've found great value in asking suppliers to do more than deliver parts. We're not after the cheapest components," he stresses. "We're after the most cost effective."
One world, one machine. Global companies may or may not have a business advantage over more localized competitors, but there's little doubt that a worldwide reach boosts the quality of the Deere's machines. More than just having international sales and manufacturing operations, Deere's approach to global business revolves around equipment that can work anywhere in the world. "We focus on a single global platform and make it the best it can be," says Yarbrough. And he argues that these global machines have an edge when they hit the fields. "We have to develop each machine for the toughest conditions it will encounter around the world, not just in North America," he says.
Being global also turns the world into Deere's proving ground. With field testing programs in North America, Germany, Brazil, and Australia, machine prototypes can be put through their paces under the actual conditions they'll see in the different regions. Global testing also shortens the product development cycle. Yarbrough explains that Deere once field-tested agricultural-equipment prototypes only in North America, which limited testing time from May to November. "On engineering changes involving long-lead items, we would have to wait a calendar year to validate the change," he says. "Now we just take a machine down to the Southern Hemisphere."
The 50 Series combines represent the first large-scale product to take advantage of Deere's global outlook and testing programs, according to Yarbrough. This single harvester platform tackles wheat in Europe, soybeans in Brazil, or corn in North America-three of the most grueling regions for these crops. To satisfy Australian farmers, it also supports faster harvest speeds than currently used in North America.
Balancing the needs of customers around the world may seem like a tall order. But Moulds, for one, isn't worried. "Our product delivery process has taught us what it really takes to exceed our customers quality expectations," he says. "No matter where they live."