Label the flow lines for the major type of stress occurring at the location: Remember that force is transmitted at interfaces primarily by compression.

262 CHAPTER 9 Product Generation Figure 9.21 Evolution of a battery contact.

9.3 Form Generation

263 Design perfection is achieved not when there is nothing more to add, but rather when there is nothing more to take away. —Antoine de Saint-Exupéry beginning of the design process with it and considering new requirements and functions. ■ MagnifyingMinifying : Make a component or some feature of it biggersmaller relative to adjacent items. Exaggerating the size or number of a feature will often increase one’s understanding of it. Make one dimension very short or very long. Think about what will happen if it goes to zero or infinity. Try this with multiple dimensions. Sometimes eliminating, streamlining, or condens- ing a feature will improve the design. ■ Rearranging : Reconfigure the components or their features. This often leads to new ideas, because the reconfigured shapes force rethinking of how the component fulfills the functions. It may be helpful to rearrange the order of the functions in the functional flow. Take the current order of things and switch them around. Put what is on top, on the bottom; or what is first, last. ■ Reversing : Transposing or changing the view of the component or feature; it is a subset of rearranging. Try taking what is the inside of something and making it the outside or vice versa. ■ Substituting : Identify other concepts, components, or features that will work in place of the current idea. Care must be taken because new ideas sometimes carry with them new functions. Sometimes the best approach here is to revert to conceptual design techniques in order to aid in the development of new ideas. ■ Stiffening : Make something that is rigid, flexible or something that is flexible, rigid. ■ Reshaping : Make something that is first thought of as straight, curved. Think of it as cooked spaghetti that can be in any form it wants to be and then hardened in that position. Do this with planar objects or surfaces. A more complete list of ideas for patching can be found in TRIZ’s 40 Inventive Principles, discussed in Section 7.7. These principles suggest many ideas for patching products. The primary goal of patching is to make things work and to make them sim- pler. The most elegant designs are those that provide the needed functionality but look simple. The quote in the aphorism above states this thought best. Ex- cessive patching implies trouble. If design progress is stuck on one function or component and patching does not seem to be resolving the difficulty, it may be a waste of time to continue the effort. To relieve the problem, apply these three suggestions. 264 CHAPTER 9 Product Generation ■ Return to the techniques in conceptual design; try to develop new concepts based on the functional breakdown and the resources for ideas given in Chap. 7. ■ Consider that certain design decisions have altered or added unknowingly to the functions of the component. As products evolve, many design decisions are made; it is easy to unintentionally change the function of a component in the process. It is always worthwhile, when stuck on finding a quality solution, to investigate what functions the component is fulfilling. ■ If investigating the changes in functionality does not aid in resolving the problem, the requirements on the design may be too tight. It is possible that the targets based on engineering requirements were unrealistic; the rationale behind them should be reviewed. The results of efforts to refine or patch any aspect of the product can lead in either of two directions. First, and most often, the refinement or patching is part of the generateevaluate loop in product design. After each patch or refinement, it is good practice to revisit the decisions that have been made in developing the product to this point before reevaluating. As the product becomes more refined, evaluation usually requires more time and resources; therefore, double-checking can lead to savings. Second, if no satisfactory solution can be found, the result of the refining or patching effort requires a return to an earlier phase of the design process. 9.4 MATERIALS AND PROCESS SELECTION At the same time form is being developed, it is important to identify materials and production techniques and to be aware of their specific engineering requirements. An experienced designer has a short list of materials and processes in mind even with the earliest concepts. In developing an understanding of the product, we may have Form Function Material Production Assembly Manufacture Connections Components Configuration Constraints set requirements on materials, manufacturing, and assembly. At a minimum we did competitive benchmarking on similar devices, studying them for conceptual ideas and for what they were made of and how they were made. All this information influences the embodiment of the product in several ways: First, the quantity of the product to be manufactured greatly influences the selection of the manufacturing processes to be used. For a product that will be built only once, it is difficult to justify the use of a process that requires high tooling costs. Such is the case with injection molding, in which the mold cost almost exclusively determines the component cost for low-volume production see Section 11.2.4. In general, injection-molded plastic components are only cost-effective if the production run is at least 15,000. A second major influence on the selection of a material and a manufacturing process is prior-use knowledge for similar applications. This knowledge can be

9.4 Materials and Process Selection

265 When in doubt, make it stout, out of things you know about. both a blessing and a curse. It can direct selection to reliable choices, yet it may also obscure new and better choices. In general it is best to be conservative, and heed the axiom below. When studying existing mechanical devices, get into the habit of determining what kind of materials were used for what types of functions. With practice, the identity of many different types of plastics and, to some degree, of the type of steel or aluminum can be determined simply by sight or feel. Appendix A provides an excellent reference for material selection. It includes two types of information: a compendium of the properties of the 25 materials most often used in mechanical devices and a list of the materials used in common mechanical devices. The 25 most commonly used materials include eight steels and irons, five aluminums, two other metals, five plastics, two ceramics, one wood, and two other composite materials. The properties listed include the standard mechanical properties, along with cost per unit volume and weight. This list is intended to serve as a starting place for material selection. Detailed information on the many thousands of different materials available can be found in the list of references given at the end of Appendix A. Additionally, the appendix contains a list of materials used in common products. Since many different materials can be used in the manufacture of most products, this list gives only those most commonly used. Knowledge and experience are the third influence on the choice of materials and manufacturing processes. Limited knowledge and experience limit choices. If only available resources can be utilized, then the materials and the processes are limited by these capabilities. However, knowledge can be extended by including on the design team vendors or consultants who have more knowledge of materials and manufacturing processes, so the number of choices can be increased. Probably the most compelling point in the selection of a material is its availability. A product that has a very small production run will probably use off-the-shelf materials. If the design requires structural shapes I-beams, chan- nels, or L shapes that must be light in weight, then extruded aluminum shapes could be used. This decision, however, limits the material choices. Aluminum ex- trusions are readily available in only a few alloytemper combinations 6061-T6, 6063-T6, and 6063-T52. Other alloys are available on special order. There is a setup charge to obtain these, and a minimum order of a few hundred pounds—a complete run of material—would also apply. If the available alloytempers have properties needed by the product, they can be used. If they do not, the product shape may need to be changed. During design, the material and production processes selected must evolve as the shape of the product evolves. As a product matures, its layout, details, ma- terials, and production techniques are refined become less abstract. At the same time a product is refined, changes are sometimes patched with no accompanying 266 CHAPTER 9 Product Generation refinement. Suppose the material initially chosen for a component was identified only as “aluminum”; this selection must now be refined and may be patched. For example, the refiningpatching history of the selection of material for one component is “Aluminum” → 2024 → 6061 → 6061-T6. That is, the selection of “aluminum” was refined to a specific alloy 2024, which was changed patched to a different alloy, 6061, which was then refined by identifying its specific heat treatment, T6. This evolution is typical of what occurs as a product is refined toward a final configuration. Sometimes during the design of a new product, the requirements cannot be met with existing materials or production techniques, no matter how much patch- ing and shape modification occurs. This situation gives rise to the development of new materials and manufacturing processes. Until recently, the thought of designing the materials and processes to meet the product design needs meant postponing the design project so that material or production technology could reach maturity Section 8.4. However, recent advancements in the knowledge of metal and plastic materials have, to a certain extent, allowed for material and process design on demand. 9.5 VENDOR DEVELOPMENT When specifying systems, assemblies, or components you either use what is available from vendors, or design new hardware. Mechanical designers seldom design basic mechanical components e.g., nuts, bolts, gears, or bearings for each new product, since these components are readily available from vendors. For example, few engineers outside of fastener manufacturing companies de- sign new types of fasteners. Similarly, few designers outside of gear companies design gears. When such basic components are needed in a product, they are usually specified by the designer and purchased from a vendor who specializes in manufacturing them. In general, finding an already existing product that meets the needs in the product is less expensive than designing and manufacturing it, since the companies that specialize in making a specific component have many advantages over an in-house design-and-build effort: ■ They have a history of designing and manufacturing the product, so they already have the expertise and machinery to produce a quality product. ■ They already know what can go wrong during design and production. A new design effort requires extensive time and experience before reaching the same level of expertise. ■ They specialize in the design and manufacture of the component, so they can make it in volumes high enough to keep the cost below what can be achieved through an in-house effort.

9.5 Vendor Development

267 Additionally, even if the exact product is not available, most vendors can help develop products or components that are similar to what they already man- ufacture. Sometimes “design” is specifying Commercial Off The Shelf COTS components. This is so common that the terms COTS and the government equiv- alent, GOTS, are commonly used. COTS and GOTS design is the placement and interfacing between available components. In past times, it was common for a company to send detailed drawings of components to a number of vendors and select the vendor that quoted the lowest cost. Over the last few years companies have been working with a small number of vendors in the design process from the beginning and including them in the decisions that affect what they will be supplying. In fact, large companies have reduced the number of vendors by an order of magnitude since the mid-1980s. Some companies financially invest in their vendors, and vice versa, to further improve the bond. These tight relationships lead to improved product quality. Whether to make or buy a component or to choose a component from what is available from vendors, there is need for decision making. For these types of decisions, a good set of criteria are given in the MakeBuy, Vendor Selection template shown in Fig. 9.22. Detailed descriptions of each of the criteria are ■ Low development cost —How much is it going to cost to develop the com- ponent. If it is truly COTS, then there are no development costs. However, if work is needed to change a COTS system or part, or one needs to be developed, then these costs may be significant. ■ Low product cost —Many decisions are based solely on this criterion. This cost is highly dependent on the volume the number purchased, delivery costs and many other factors. These will be addressed in Chap. 11 when we discuss DFC, Design For Cost Section 11.2. ■ High product life cost stability —Beyond the cost, it is important to consider how the cost may change over time. Cost can be controlled better when you make a component or can be locked in by contract. ■ Low development lead time —If this and the next criterion are important; they may dominate all the rest and force the purchase of a COTS component. COTS components need no development lead time. ■ Low order lead time —Even COTS components have an order lead time. Sometimes it can even be longer than the time needed to make the component in house. ■ High product quality —Sometimes quality must be traded off for cost or time. It is important to understand from the beginning, the level of quality needed to meet the engineering specifications. ■ Good product support —To address this criterion, two questions must be answered: Who will be responsible for failures and maintenance of the com- ponent or product? And, how much support will be needed? ■ Easy to change product —Sometimes it is necessary to change the product during its lifetime. If it is COTS then you have no control over changes. If 268 CHAPTER 9 Product Generation MakeBuy or Vendor Selection Decision to be made: Make or buy Date: 092310 Product: Part 234-4B in Espiral Rationale: Choose Barns as it is significantly better than the others in weighted total and has no great weakness. Team member: Bob Prepared by: lvin Team member: Alvin Checked by: Becky-Sue Team member: Becky-Sue Approved by: Fredrick Team member: The Mechanical Design Process Designed by Professor David G. Ullman Copyright 2008, McGraw-Hill Form 20.0 Criterion Wt. Vendor 1 Vendor 2 Vendor 3 Vendor 4 Make Allied Barns Crane Low development cost 5 2 3 2 4 Low product cost 22 4 2 3 4 High product life cost stability 2 5 3 4 4 Low development lead time 7 3 2 4 2 Low order lead time 11 3 2 5 1 High product quality 14 2 3 3 2 Good product support 6 1 4 2 3 Easy to change product 8 3 5 5 4 Strong IP control 18 4 2 4 2 Good control of order volumes 5 4 1 2 4 Good control of supply chain 2 4 4 2 2 Total 35 31 36 32 Weighted total 3.2 2.56 3.47 2.79 Figure 9.22 Makebuy or vendor selection example. 9.6 Generating a Suspension Design for the Marin 2008 Mount Vision Pro Bicycle 269 this is an important criterion, then it may be best to make the component or have a closely allied vendor make it. ■ Strong IP control —IP, or Intellectual Property, is a primary asset of a com- pany. IP includes patents, CAD files, drawings, and other documents that give details about the design or production of a product ■ Good control of order volumes —Sometimes the number of components ordered needs to be flexible. This is generally in response to market changes that can be controlled to some degree through inventory, but that is expensive. So, if order volumes are volatile, then this may be an important criterion. ■ Good control of supply chain —If you buy a component you can only control the supply chain through your contracts. If this is not sufficient, then this criterion may be important. These criteria are used in Fig. 9.22 to decide whether to make or buy a component from one of two vendors. This example is a combination of the com- mon makebuy decision and vendor selection decision. Here a simple decision matrix is used to find it. Vendor 3 is the best choice. An online, free robust decision maker is available. 9.6 GENERATING A SUSPENSION DESIGN FOR THE MARIN 2008 MOUNT VISION PRO BICYCLE The Marin Mount Vision Pro bike was designed for the cross-country mountain bike enthusiast. It is a quality and fairly expensive bicycle over 3000USD. The primary demographic for this bicycle is male, 25–50 years old. But, because of its modern look and marketing, it is also designed to attract females and riders of other age groups. It is intended for use on technical trails where there is a mix of uphill and downhill, where light weight and pedaling efficiency are of primary importance. In this section, we will explore how the rear suspension evolved. The story presented here has been tailored for this text, but it does not differ much from the reality of the Marin design process. 9.6.1 Understand the Spatial Constraints for the Mount Vision Bicycle Rear Suspension For the rear suspension of a mountain bicycle, the spatial constraints are shown in Fig. 9.23. Beyond the obvious need to connect the wheel to the frame, the Marin engineers also wanted to control the path the wheel made relative to the frame as the suspension deflected, the stiffness of the suspension, and the chain length. Ideally, the wheel of the bicycle should move “nearly” straight up and down as it deflects. If the suspension was designed as a simple bar with a single pivot 270 CHAPTER 9 Product Generation Figure 9.23 Physical constraints for the mount vision. see Fig. 9.24, then the wheel would make an arc with it moving closer to the front of the bike as it deflected. This would give the rider the feeling she was falling backward as the wheel deflected. The Marin engineers wanted to control the wheel path to manage the feel transmitted to the rider. As important as the wheel path, is the change in stiffness—flow of energy of the suspension system. The ideal suspension system for any vehicle is soft, has low stiffness, when it goes over small bumps and gets stiffer for large bumps. In other words, the larger the deflection, the stiffer the suspension system should become. This requirement may not seem “spatial” but it constrains how the shock is mounted between the frame and the moving parts as will be seen. To understand the desire to control the chain length, consider a suspension that was designed so that when the pedals were pressed, the resulting tension in the chain pulled the suspension up i.e., the frame down. The rider, when feeling the frame drop flow of information would then ease off the force and subsequently the frame would rise. Feeling the frame rise, the rider then reapplies the pedal force resulting in a “pogo” motion and a very uncomfortable ride. Thus, an additional constraint is that the motions and accelerations felt by the rider will not lead to poor suspension performance. Summarizing, the spatial constraints are

1. Wheel and chain must clear frame for all deflections.

2. Wheel should move straight up and down.

3. Low stiffness for small deflections, increasing with deflection. 4. Chain length should not change during deflection. 9.6 Generating a Suspension Design for the Marin 2008 Mount Vision Pro Bicycle 271 9.6.2 Configure Components for the Mount Vision Bicycle Rear Suspension The simplest type of suspension that can be put on a bicycle is a one with a single pivot as shown in Fig. 9.24. On the bike, the pivot is near the center of the crank and every point on the rear triangular structure called the rear “stay” rotates around this point. As the wheel deflects, it makes a circular arc and the chain gets shorter, violating two of the spatial constraints. As the wheel moves up, the shock gets shorter. Shocks on bicycles generally have an air or oil damper with a mechanical, coil spring wrapped around it. This spring has a stiffness that remains essentially constant as the wheel deflects. So the spring force increases as the wheel is deflected. Thus, it is clear that this type of suspension will not work for the Marin Mountain Vision Pro. In 2003, Marin introduced a more sophisticated suspension based on a four- bar linkage and referred to it as their “Quadlink” design. The Quadlink was not the first four-bar suspension used on a mountain bicycle, but it did bring this type of mechanism to a high level of refinement. To understand how Marin configured this suspension, a short refresher on four-bar linkages. Figure 9.25 shows two simple members, A and B connected by member C. Members A and B, the links, move about fixed points and member C, the “fol- lower,” connects the end points of A and B. Points 1 and 2 move in circular arcs about the fixed points as in Fig. 9.24. For this parallelogram four-bar link- age, member C effectively translates without rotating. This will be clarified in a moment. To better understand what link C is doing, consider a modification to this basic four-bar where the links are different lengths as shown in Fig. 9.26. The projection of the links intersects at a point called the instant center. The instant center is the point about which link C is rotating when the links are in the configuration shown. The reason for the term “instant” is that the same linkage all the member’s lengths held constant; with the members in a different position have a different instant center, as can be seen by comparing Figs. 9.26a and 9.26b. Figure 9.24 A simple, single pivot suspension. Reprinted with permission of Marin Bicycles. 272 CHAPTER 9 Product Generation B A C 2 1 Figure 9.25 A basic four-bar linkage. Instant center a B C A Instant center b B C A Figure 9.26 A linkage with two of its instant centers. Thus, as the linkage moves through different positions, the instant center traces a path describing the virtual pivot point for member C. The linkage in Fig. 9.25, the parallelogram has the instant center always at infinity, thus the link has an infinite radius of rotation—it translates. One further four-bar concept is needed to understand how the Marin Quadlink was designed. If link C, rather than being a straight member as shown in the figures so far, is a structure as in Fig. 9.27, then every point on this structure or stay is rotating about the instant center. Figure 9.27 is the same linkage as in Fig. 9.26 but with the addition of the stay, CDE. Point 3 at the left end of member CDE rotates about the instant center and makes a nearly straight line. This is the point where the wheel is mounted. In order to design the shape of the path followed point 3 the engineer specifies the lengths of A, B, C, D, and E and the relative positions for the two fixed pivot 9.6 Generating a Suspension Design for the Marin 2008 Mount Vision Pro Bicycle 273 B D C 2 A 3 E 1 4 5 Figure 9.27 A complete four-bar structure. Figure 9.28 Simulation of the Quad link suspension: a undeflected and b fully deflected. Marin Bicycles are designed on Autodesk Inventor TM . Reprinted with permission of Marin Bicycles. points the distance between them and angle of the line connecting them, for a total of seven variables. There is a lot of design freedom. The Marin engineers adjusted these variables to meet the spatial constraints. The final design is shown in Fig. 9.28. These solid models were developed in Autodesk Inventor TM . This program let the engineers see the motion as the sus- pension deflected. The block in the upper left corner controls the simulation so the designer can see the motion of the mechanism and instant center. The Marin designers used the solid modeling software and other computer simulations to determine the best values for the seven variables, one that gave a fairly straight wheel path with near constant chain length. Further, by controlling the location of the virtual center and the positioning of the shock described in the next section they were able to achieve low stiffness for small deflections, increasing with deflection. Specifically, when the virtual center is nearly under the crank Fig. 9.28a the moment arm of the rear stay is much shorter than when the suspension is deflected Fig. 9.28b. 274 CHAPTER 9 Product Generation 9.6.3 Develop Connections: Create and Refine Interfaces for Functions for the Mount Vision Bicycle Rear Suspension This section focuses on the connections between the components. On the Marin Mount Vision Pro, the connections are those between the links in the four-bar linkage, those connecting the shock to the bike and those that connect the fixed parts together. We will consider these in order. For the four-bar linkage, the con- nections are the four pivots. These must have one degree of freedom and thus can be either bearings or flexures. For most mountain bikes, either rolling ele- ment bearings or bushings are used, but some have used flexures. Considering Fig. 9.27, the shock can be mounted in many different ways. It can be mounted between any two elements that move closer together as the system deflects; for example, element C and the frame, elements A and B, and so on. The addition of the shock adds two more pivots to the assembly making a total of six pivoting connections. The Marin engineers reduced the number of pivots by mounting the shock between linkage pivots 2 and 4. As the suspension system deflects, pivot 2 moves toward pivot 4. In fact, the engineers, when determining the lengths of all the seven members, took the needed change in length of the shock as an additional constraint. The decision to mount the shock in this manner made the design of linkage more challenging and connections more complex, but the trade-off for fewer pivots made this worthwhile. Pivots 2 and 4 need to have the link and shock free to rotate about the axel shown as a centerline in Fig. 9.29. Note in Fig. 9.28, the amount of rotation of these elements is small, only a few degrees in some cases. Bearings that operate primarily in one position and only move a small amount from that position present Frame Shock Link Link Figure 9.29 The components in pivots 2 and 4.