- SPECIAL REPORTS
The best time to start controlling the factors that drive assembly costs is early in the design process. Every manufacturer has its own approach to controlling costs, and some have been successful in using the computer-based Design for Manufacturing and Assembly (DFMA) methodology for this purpose. DFMA, a combination of two complementary software tools—Design for Manufacture (DFM) and Design for Assembly (DFA)—provides a strategy for identifying early cost drivers and developing tactics to reduce costs throughout the manufacturing process.
Four experts from diverse manufacturing companies shared their collective insights and experiences in tackling the challenge of getting cost drivers under control early in the product cycle. The occasion was a panel discussion held at The 2001 International Forum on DFMA, hosted by Boothroyd Dewhurst Inc.
The panel included Paul Zimmermann, principal mechanical engineer with honors, Raytheon Systems Co. (Dallas); Dantar Oosterwal, director of cost management, Harley-Davidson Inc. (Wauwatosa, WI); Robert A. Williams, customer experience strategy manager, Modular Network Storage Div., Hewlett-Packard Co. (Loveland, CO); and Ken Swift, Lucas Professor of Manufacturing Systems Engineering, University of Hull, and director of the British Aerospace Systems Research Center in stress concentration and fatigue analysis (Hull, England). Following are excerpts of their comments on just three of the questions the panel addressed.
Identifying and Managing CostA critical component of the overall DFMA process is predictive cost estimating while early design decisions are being made. So moderator Winston Knight, senior vice president, Boothroyd Dewhurst Inc., challenged each panelist in turn to discuss the importance of early cost assessment and predictive cost development, and how these tools are actually used in managing manufacturing costs.
Williams: In my view there are three progressive steps in the decision-making that is required to design good products, and make a profit for our corporations and our stockholders.
The first step is the acquisition of knowledge. Because we live in the information age, that’s not a problem in my mind. There’s plenty of knowledge out there, but there are also plenty of errors being made every day, both by corporations and by individuals. The second step is gaining understanding. And once we have the understanding that comes from the knowledge, the final step is the application of wisdom. We, as first- or second-level engineering managers, are usually advising people in high positions of leadership and decision-making, so that they can make wise choices.
Lord Kelvin once said, "If you cannot measure it, you cannot fully understand it." So our task then is to bring this understanding to using tools that measure. The general concept of product development is to come up with an optimum concept and then move into a prototype. But, so often, you want to rush right to the prototype because you want to get the product to market as quickly as you can.
What we need are tools that help us get that understanding. One such tool is Design for Assembly (DFA) analysis, and out of that come suggestions for redesign. Then we need to make some initial materials and process decisions. By running those parts that came from the DFA analysis through the cost analysis tool, we are able to get to our final goal and that is to plan our design for manufacture and give it some detail and then lead into the prototype stage. This, admittedly, is an ideal situation. But, it is one that we should constantly be moving toward.
Zimmerman: The overall objective is to reduce the cost of designs. Motorola’s Six Sigma principles focus on reducing defects, and by reducing defects we automatically reduce the costs. The first, and perhaps most important, principle for accomplishing this is to minimize the number of parts.
I like this quote from Albert Einstein: "The best design is the simplest one that works." And this is true in any industry, regardless of the product. The simplest design that works is probably the best design. Not only that, but it will also have the fewest parts and is likely to be the most reliable. DFMA simplifies the design, and it reduces parts, processes and assembly steps.
With defense-related products, we have to be concerned about maintainability. These products have to be easy to repair in the field, and modular subassemblies are ideal for this purpose. I particularly like to use modular subassemblies if I have something that is a defect driver. A good example of a defect driver is the focal plane array detector assembly in a forward looking infrared (FLIR) system, which you can think of as the eyeball for a night sight.
That FLIR eyeball has a very high failure rate, and the easier we can make it to take out and replace in the field, the easier it is for the soldier to repair it in the field. And, by the way, it also makes it easier to build it and fix it in the factory.
The design engineer selects the parts, materials and processes, and that dictates in large measure the cost of the product. So if you just minimize the number of parts, minimize the number of processes and minimize the number of assembly steps, you’re automatically going to reduce the cost, that’s a given.
Oosterwal: The importance of early cost assessment has been an issue for a very long time; it was even addressed in a very ancient text: "For which of you, intending to build a tower, sitteth not down first, and counteth the cost, whether he have sufficient to finish it? Lest haply, after he hath laid the foundation, and is not able to finish it, all that behold it begin to mock him, saying, this man began to build, and was not able to finish." This, of course, is from Luke 14:28-30, and serves to remind us that even in those Arial, cost overruns put one at risk of being ridiculed by one’s colleagues.
Whether the product is a simple machine, a complex device or anything in between, four elements need to be considered in designing and manufacturing the product: quality, cost, timing and the function. The challenge is to establish a balance among the demands of these four elements. And there is a multitude of potential solution sets for achieving a balance.
If we want a product to do a particular function, each one of us will come up with a different way of doing that particular function. If you consider the quality, cost and timing required to provide that particular function, you can come up with a number of solution sets. So the dilemma is selecting between the solution sets and, concurrently, of optimizing within any particular solution set.
Cost is obviously the element that we’re talking about here, and the impact of early and accurate cost information is specifically that it allows for earlier and better decision-making as you go through that dilemma. First and foremost, it serves to unify the development team around project objectives. If you get out of the emotion and get into the actuals, early cost information helps you figure out where you’re heading within the number of solution sets that you’re dealing with. The bottom line is that early and accurate cost information provides timely information in the evaluation of design alternatives among the solution sets, as well as optimization within the solution sets.
The early and accurate cost information developed through DFMA also forces concurrent development. If you’re looking at the costs early enough, before the decisions have been made, it allows an opportunity for the design team to work collaboratively together. And it also allows for an early understanding of project viability.
Swift: In the big companies that I work with, the directors all want to reduce time-to-market. They see that as the single most important thing that they can push through in their businesses. These businesses are all fighting to drastically reduce the number of prototypes that they create, because that is crucial to reducing time-to-market. And what this re-quires, then, is to change from the culture we have now of building and testing prototypes, to one of analysis.
Because cost is such an important driver, we need a computer-aided design (CAD)-based environment where product information and early design decisions, including costs, are accessible, so that engineers can understand the elements of the product development process. Within that environment, we need special tools, such as DFMA, which will help the engineers identify the costs of the various elements they are dealing with in a product design.
But it has to be kept in mind that for these techniques to be accepted they have to be well-founded, rigorous and capable of being validated correctly. We need to work very hard to make sure our cost estimating methods are on par with the methods that we have in the engineering science domain.
We someArial have masses of data but not very much knowledge of what to do with it and what it means. We need to get a better grip on the data so we can understand what it really means, and this goes right back to the need for rigor in our costing processes.
Creeping FeaturesEven though good cost estimates for manufacturing a product have been established in the early going, the inevitable addition of features leads to cost inflation. So the panel members were asked how they formally manage the challenge of feature creep.
Williams: At HP, we knew this creeping featurism was a threat, particularly in turning a project around in a 6 to 12 month timeframe. So we just had to make the conscious decision to pick a cutoff time, marked by a date on the calendar, for accepting changes and adding features. We knew what the consequences would be if we made changes after that date. Once that date passed, we didn’t stop collecting ideas or considering additional features and, if momentum built later to add specific features, we would revisit the question.
Oosterwal: There is always the temptation to just go ahead and add something under the table, but what we’ve had success with is getting the question out in the open for discussion, and if the feature is beneficial, if marketing believes it will help sell motorcycles, or if it can be added into the selling cost, then we’ll move forward. The key is to have a discussion with all concerned parties and address the question of whether the proposed feature is beneficial to the overall program. If it is, then do it, and if it isn’t, then drop the idea at that point and move on.
Zimmerman: We have a process within Six Sigma whereby we list what is in the scope of the project, and what is out of scope. And we focus on keeping everyone in scope when we’re doing designs. The marketing team is asked whether the proposed feature is, in fact, something that the customer wants, and the answer to that often makes the decision.
Who’s in Charge?Territorial issues often come into play concerning systems that link various functions within a company. So the panel members were asked their opinions on whether cost estimating packages should be in the domain of product design, manufacturing or some other specific function area such as cost management.
Zimmerman: At Raytheon, we seem to find that it works better if it’s part of design. At most companies, there is still a tendency for manufacturing to be almost a stepchild, and it seems to work better if they are part of the design team reporting to program management.
Oosterwal: From our experience, I don’t think there is any one right place. We have the system open to everybody. All of the various functional groups have different reasons for needing it and they all have different uses for it. In fact, we expect our suppliers to understand the connection to cost. And we expect our manufacturing engineers to understand what makes up the cost of the product.
Design is certainly critical, of course, and we’ve set up a cost management area to support the design initiatives. It won’t be there forever, but it’s necessary from the perspective of providing support and providing focus on what needs to be done, especially if you want to start changing the culture within the organization.
A panel of experts, engaged in both formal presentations and extensive Q&A, covers a great deal of ground, and this article only includes a few highlights. To learn more about DFMA and early cost drivers, to direct specific questions to any of the panel members, or to learn about the program and speakers at the 2002 International Forum on DFMA (June 10 to 12, 2002), please contact the sponsor of the Forum, Boothroyd Dewhurst Inc. at 401-783-5840 or www.dfma.com.