Someone once said, "Come up with a great idea and the world will beat a path to your door." Sorry, but I have to chuckle. The easiest part of product development is coming up with the idea. To be frank, lots of people come up with inventions everyday. I would speculate that literally millions of new product ideas are spawned each week. The hard part is all the stuff that follows. It takes real persistence, tremendous fortitude, and a whole lot of time and money to bring a new product to life, and not everybody knows how to do it.
Products take on a life of their own. They get conceived, they're nurtured, they mature, and the healthy ones spawn new ideas. In their development stage, it is important that their direction be clear. What they look like and what they do is also important. The developer needs to get quality feedback to ensure good product acceptance and development. The drive to "make your mistakes quicker" is paramount to quickening the product development schedule. As a result, make a plan and stick to it. That way, you can plow through the program, make your mistakes, and reduce the number of surprises at the end.
Plenty of techniques and products can help you in the process of bringing a product to market. I'll talk a little bit about windows of opportunity and the importance of time to market, scheduling techniques, and of course, CAD's role in rapid prototyping and gaining customer buy-in.
This article covers design from concept to production -- we'll call this the Product Development Cycle. At Walter Dorwin Teague Associates, we often break the cycle into four logical, yet chronological, phases.
Hand sketches, CAD sketches, Concept Breadboarding, and Concept Review. Phase 1 deals with the vision and proofing of concepts. An important part of Phase 1 is the evaluation of a product's market potential and acceptance. For the designer, it's important to have a good specification to work from. If you don't have one by now, create one even if it is wrong. Just get something down on paper. Know where you're going before you get there. How much is the product supposed to cost? What's it supposed to do? How many are going to be made per year? What's the environment like for the product? What's the projected life of the product? Are there any special regulatory requirements like FDA, FCC, UL, or CSA?
In this phase, "blue sky" ideas are acceptable and encouraged. Cardboard cutouts, whiteboard brainstorming, soft foam models, freeform hand sketches, and quick-and-dirty calculations are all part of Phase 1. Anything that can help get an idea across quickly is acceptable. By the way, there's a big market here for the smart CAD vendor or developer that can make CAD simple and intuitive enough to use at this stage. Get the ideas down quickly and unfettered. End up with hundreds of ideas, but be sure to evaluate them before moving on to the next phase. You don't want to continue with anything that's obviously undoable. Therefore, near the end of Phase 1, conduct a thorough design review. Pick the best ideas for continuation, and chuck the rest.
Styling, Human Factors, and Generation of Styling Mockups. Industrial designers are great! They apply artistic and ergonomic sense to complex, technical, and often inanimate objects. They give them life. In a highly competitive commercial arena, the industrial design of a product may be the only thing that distinguishes one product from another. The product's look and feel is critical. The user interface, ergonomics, and visual cues need to make sense.
In addition to looking right, the product must function properly. Phase 2 is a good time to bring in the manufacturers and engineers to ensure the validity of the design. Get them involved. Get their buy-in. "Make your mistakes quicker." It's important to know if the direction that's being taken is doable.
The net result of this phase is often a full-size, 3D, hard-foam mockup, a CAD rendering, and an industrial design control drawing.
An actual size, properly weighted mockup provides the developer an opportunity to get a tactile and spatial sense for the size and volume of the resulting design. No matter how good or expensive the system is, currently, CAD cannot take the place of holding something in your hands. However, a good photorealistic CAD rendering can provide excellent visualization with the ability to change colors, textures, and lighting in ways difficult to achieve otherwise. In fact, some photorealistic renderings are so good that it is possible to cut and paste them into real life scenes using scanners and common graphics software. This technique is commonly used very convincingly in actual sales brochures prior to the existence of the real product.
Industrial design control drawings are used to define product exterior and feature intent. The process forces the industrial designer to completely define all exterior and functional features that the engineer is going to have to deal with. The resulting definition may call out unusual draft considerations, "blends" of surfaces, and "swoopy" cutouts that are not always easy to generate with today's "low cost" 3D modeling software like AutoCAD. However, the integrity of the design must be maintained. Very often, the engineer fudges the result to get something close because the chosen CAD software cannot deal with the actual design. To the industrial designer, this method is usually unacceptable. The slightest deviation in geometry can have a dramatic impact on the look and feel of the resulting product. This problem has driven many of my colleagues to more expensive software costing tens of thousands of dollars more than AutoCAD. Fortunately, relief from this dilemma may be on the horizon.
Engineering and Functional Prototypes. Phase 3 is where some of the latitude enjoyed earlier gives way to a number of unyielding and very specific constraints. We stuff the box, implement the functions, and make it as low cost and easy to assemble as possible. Phase 1 was quick and dirty, but Phase 3 deals with precision and accuracy.
This phase has no room for error. This effort requires the service of a precision 3D CAD system that can knock out a real-life model. The system needs to handle draft, variable fillets, constant wall sections, 3D sweeps, threaded holes, variable parting lines, sheetmetal design, casting considerations, surface variations, and painless creation of assemblies. The system needs to handle automatic, or semi-automatic, 2D drawing creation. The 2D drawing and the 3D model should reflect changes. The interface needs to be intuitive, and not require years of training to create and "stuff" common but complex geometries and assemblies.
The resulting dataset must be reliably transferable to CNC and SLA machines and competing CAD systems. It's important to communicate well with other CAD systems to build synergy with coworkers, clients, and vendors to conduct product development quicker.
With the 3D model designed, the dataset can then be rapidly prototyped and checked for form, fit, and function. Much to the dismay of the masses, mistakes continue to happen. I don't care if you've spent $50,000 on a state of the art CAD system or not. There are going to be screw-ups. Let's just face it. You're going to have to plan on at least one iteration before going to tooling. The fastest way to "make your mistakes quicker" is to take your best shot, make one or two prototypes, fix the errors, and then, move onto tooling. It's a big mistake to think you can go straight to tooling without building a prototype first. If you do, you'll just be prototyping your product in hardened-tool steel -- and that can get really expensive. Instead of fixing one thing, you'll be fixing everything that comes off the line.
Tooling and Manufacturing. This phase is where the rubber meets the road. If you've done your homework and included the toolers, casters, molders, stampers, industrial designers, and engineers every step of the way, this step should be a breeze, albeit a long breeze. Although there are ways to shorten time to market (such as rapid tooling, spray metal, and aluminum tooling), on the average, plastic injection mold tooling can take 16 to 20 weeks to build, test, and debug.
This philosophical phrase was coined by one of our aerospace clients, Bob Nelson. It means go quickly, but continually check each step of the way with numerous iterations. Use whatever means are at your disposal to get answers quickly. Don't lose focus by trying to be perfect. Perfection will come with the iterations. Use cardboard cutouts, Styrofoam, tape, hot glue, CAD, wood -- whatever it takes -- for modeling. Just keep making models and thinking of speed. It's okay to make mistakes. Just be thorough and make them as quickly as you can. It's a philosophy that works.
Use technology to cut steps out of the cycle. Use rapid prototyping like stereolithography, sintered laser technology (SLT), or skilled model makers to generate master models of your parts. Then, use the master model to generate poured urethane castings from silicon molds. Although it's expensive, urethane casting technology can provide an early product introduction while you wait for the actual injection molded tooling to be finished. Teague helped Virtual I/O introduce its virtual reality headgear this way. Rapid prototyping and urethane castings can really shorten the time to market.
One emerging technology worth tracking is rapid tooling. It's a process not unlike SLT. The difference is that you prototype the tooling instead of the part, and then mold the part from the tool. It's a pretty exciting technology that will let you generate parts from the actual materials instead of a substitute, like urethane.
Of course, CAD systems can help shorten the development cycle. At the same time, CAD puts more of the burden of product success onto the CAD operator. CAD system operators now need to be more experienced in tooling technologies, sheetmetal, manufacturing, assembly techniques, and engineering science, in addition to being an expert at the helm of their CAD system. It's a tough mix. A lot of product design engineers have the kind of expertise needed to make the product a success, but if they don't know the CAD system, they'll create a bottle neck in the design cycle. Therefore, to speed things up, provide your engineers with advanced CAD training. It'll pay for itself quickly by reducing your time to market.
Be sure to use a good, full featured 3D CAD system. Here's a revelation: people change their minds, companies change specs on identified components, and sometimes the designer goofs. CAD systems need to be able to handle these changes gracefully, to have an easily editable model, an UNDO command, and an easy way of adjusting parameters. That means being able to change the draft at any point in the design process, being able to move a hole from one place to another by changing a number, being able to change the type of thread from a #4-40 UNF to a Fast Lead Dbl Helix, and being able to tug on a surface to add a final tweak to the design. Also, the user interface should be so good that it doesn't take rocket science to figure out how to do it.
Believe it or not, I don't think any of these features are too much to ask for. Many of the more expensive systems, Pro/E, Alias, SDRC IDEAS, and HP SolidDesigner provide robust implementations of many of the features already. However, I think that with the advent and implementation of Spatial Technologies' ACIS open modeling system, we will see considerable competition in this arena real soon. It's already happening. Autodesk has started taking advantage of it in Release 13, along with HP's SolidDesigner, SDRC, and 3D/Eye's TriSpectives Professional. The ACIS standard provides a common, robust 3D database from which each vendor can build its own user interface. Because the core modeler, ACIS, has already been taken care of by Spatial Technology, the winner in this arena will be the CAD system with the better, easier-to-use interface. As John Walker, founder of Autodesk, once pointed out in his paper "The Last Days," cost and ease of use will become the main issues.
Speaking of "The Last Days," there is a new player in this arena, 3D/Eye, a little known company in Ithaca, NY. It's bound to knock the socks off of a number of industry giants like Parametric Technologies, HP, SDRC, and Autodesk. 3D/Eye has financial backing from one of Microsoft's millionaire (billionaire?) founders, Paul Allen. 3D/Eye has been developing CAD software for other companies like HP's ME10 for around 14 years. As of this writing, the company is just polishing off the final beta release of a feature-based, true 3D, ACIS solid-modeling package that will let you work, very intuitively, in shaded perspective mode. The product is called TriSpectives. The user simply drags and drops intelligent shapes off of a side menu (similar to Visio) down onto the sides of other shapes to create complex part designs. The program exploits the concept of "properties" to alter the size, shape, draft, and wall thickness of the features. TriSpectives 1.0 does a pretty good job of covering the previous wish list and then some. It allows painless creation and re-creation of draft, shelling, feature generation, surface finishes, renderings, libraries of parts, and so on. Although it is very fast, it is currently a little weak when it comes to generating good ANSI and SI quality 2D drawings. On the other hand, it only costs $499 and is written for Windows 95 and NT. If Autodesk doesn't come through with a truly competitive 3D modeling package soon, it may make sense to simply use TriSpectives as a 3D front-end for AutoCAD until it does. The SAT files should be seamlessly interchangeable, at least in theory.
Given the variety of products that exist, it makes sense to ask if you are using the right tools for the job. Tools that work well with rectilinear shapes might be just fine for a company that focuses its business on a one-off type design that uses mostly CNC machined metal to develop its products. Complexity of shape and in-house expertise should govern the choice of a CAD system that will get you there quicker. Advantages of using systems, such as AutoCAD and AutoCAD Designer (which cost around $5,000 for software), come from the plethora of experts available to support your projects. The advantage of using systems, such as Pro/Engineer, SDRC, and Alias (around $20,000 for software, plus $2,000 annual maintenance agreement per seat) come from their ability to generate shapes that may be difficult to generate in AutoCAD and CadKey.
All of the previously mentioned programs can generate STL files for use with rapid prototyping systems and rendering files for visualization, but why use a 747 to cross the street? AutoCAD can handle rectilinear shapes and planer sweeps just beautifully. Editing in 3D is difficult, but the number of people who know how to operate the program greatly outnumber operators of the other programs.
Swoopy, organic shapes need the more expensive CAD systems, but definition may be difficult at any rate. However, out of all the systems I am aware of, Alias Studio seems to excel in its ability to provide organic shapes in short order. It's a great front-end tool. Alias Studio can also provide superior photorealistic renderings that fool you into thinking the product actually exists. It's quite impressive. I've included a number of renderings that were done at Teague to show you what I mean. Although Alias Studio is usually thought of as an industrial design and animation tool (it was used to generate the T-Rex in Jurassic Park) and is only a surface modeler (as opposed to a solids modeler), Alias can generate STL files to create actual prototypes for validation. Once validated, the file can go directly to tooling, which can really save a lot of time by cutting out the engineering phase altogether for projects that don't require engineering input (such as small, single-part electronic enclosures, bottles, combs, an so on).
Multipart products, such as super computer enclosures, information kiosks, virtual reality headgear, stereo headsets, and in-flight entertainment system seats, are complex, have a lot of interdependent relationships, and are ideally suited for more powerful software like Pro/Engineer, HP SolidDesigner, and SDRC IDEAS. It's important to have the right tool for the job. Don't get caught up in the idea that the design has to be done on this system or that. Each development program will be unique.
Time can be cut out of the cycle in other ways, too. Dovetail scheduled activity. If the program does require engineering, in addition to industrial design, get the teams busy in parallel, multitask the effort. Constantly, look for clever ways of doubling the effort and halving the time. "Make your mistakes quicker."
Michelle Pillers, P.E., is a licensed, practicing mechanical engineer and former Director of Product Development for Walter Dorwin Teague Associates Inc. As an avid user of AutoCAD's solids modeler, she can be found helping users on CompuServe's AutoCAD Forum as the acting Modeling ThreadMaster. She can be reached by email at mpillers@cheerful.com. Be sure to visit her new venture's - Evolution Design Engineering - world wide web site at http:\\www.edeinc.com.