Medical Device & Diagnostic Industry Magazine
MDDI Article Index

An MD&DI September 1997 Column

PRODUCT DEVELOPMENT

Accelerating the Product Development Cycle

Bill Evans

For managers in the highly regulated and competitive medical environment, accelerating product through the development cycle is critical. A faster cycle gives manufacturers a better chance of getting to market quickly and hitting a home run. A six-month delay can reduce a product's life cycle profits by 33%, according to a well-recognized McKinsey & Co. study.

This handle for a laparoscopic instrument was made by direct data machining of Ultratem (G.E. Plastics, Pittsfield, MA). Photo by Mark Hundley, courtesy of Bridge Design

It is unlikely that a single initiative, such as raising salaries or creative bonus structures, will accelerate the product development process. Instead, a combination of actions, including changes in management style and in the tools and techniques used in the development process, will be required.

Where does one start, and how can one determine which actions will have the greatest impact? Select an area that can result in the most improvement with the least effort. Be patient and persistent: 20% of the ideas will probably yield 80% of the time savings. Consistently reaping the advantages of this evaluative approach will require diligence. To avoid missteps and resets, managers should check and validate their product development process at each phase. During some intermediate steps, it may be necessary to spend more money to achieve a faster result, but the big overall savings for a project will occur if a polished product--one that is done right the first time--hits the market ahead of its competition.

This article offers tips and strategies for engineering project managers, vice presidents, and CEOs on efficient and practical ways to speed up the product development cycle.

PRODUCT DEFINITION

The first step is perhaps the most important one. Before scheduling a new project, analyze a recently completed job. Compare original schedules with the actual dates achieved to identify the stages in which time was lost and why.

For example, if failure of early production samples caused delays, more product validation tools, such as finite element analysis, may be needed, or an outside specialist could be consulted. Alternatively, a delay may have been caused by the product's marketing definition being changed late in the development process. Make sure simple prototypes--or even breadboards--are given to users as early as possible.

A handmade foam model like this enclosure can be made in a day (Bridge Design, San Francisco).

The best stage in which to invest time and money to correct problems is before a project is under way. Waiting until a new project is started puts the team under too much pressure to learn new techniques, qualify new vendors, and purchase new design tools at the same time that they must handle critical path design.

Encourage communication between front room marketing teams and the engineers in the back. Some of the best ideas are developed by interdisciplinary teams rather than by individuals. There is a great potential for improvement when representatives from ergonomics, electronics, manufacturing, management, engineering, service, regulatory, and industrial design all participate at the product development stage. There is some truth to the Dilbert cartoon stereotypes of marketers and engineers. However, opening genuine channels of communication will allow marketing and sales personnel to temper the nerdy engineers, while engineers can resist product ideas from marketing that are complex, inefficient, and impractical. Given the chance to understand the market, engineers grasp nuances that make products more marketable and are further motivated to give their greatest effort.

Engineers should regularly go to trade shows, hospitals, doctors' offices, and academies where they can meet the medical professionals who will use their products. They should also obtain competitors' products so they can dismantle them for analysis.

Adding outside members to the interdisciplinary product development team can make or break the outcome of the product development process. While consultants can be subject to conflicting demands from many clients, they also tend to be exposed to a wider variety of product development techniques, allowing them to keep clients up-to-date and give advice on the most appropriate rapid prototyping (RP) or tooling processes.

Innovative ideas come from experiments, and experiments are not always 100% successful. It's important for product definition teams to know that some avenues will turn into blind alleys. Willingly taking these risks should help the team stay one step ahead of the competition.

Using this risk-and-repair approach can make a product more innovative and get it to preproduction more quickly. While a mistake may cost money in the short term, the holders of a project's purse strings must understand that there is a net savings in the long run. The product development process is something like a freight train consisting of many linked parts, and everyone has to be on board for the duration of the ride.

EARLY PROTOTYPING

Three rules of product development are feedback, feedback, and feedback. This mantra is just as important to product definition. Special prototyping should take place in the product definition stage, not just in the traditional prototyping stage. A brainstorming session that quickly leads to an early-stage hands-on model or visual representation increases the project's chance for success down the road.

Of course, this prototype isn't a full-blown costly model but a mock-up. For example, if portability of a product is a concern, buy an appropriately sized suitcase and put weights in it to test the potential problem before production is under way. Be sure to have real users, as opposed to in-house staff, give feedback on the prototype.

At this point, prototypes should be made with a minimum of effort, dollars, and time. Don't be obsessed with computer-aided design (CAD) systems. Use a glue gun and cardboard, or hack up blocks of inexpensive foam. The physical models will reveal much more than the CAD system. Make sure that someone at each brainstorming meeting is sketching ideas. Industrial designers offer a lot at this early stage because of their strong rapid visualization skills.

Some team members may not have a technical background; they may also be geographically dispersed. To get their input, use a variety of techniques, including CAD tools, sketches, and rapid prototypes--any alternative to traditional blueprints that are often too technical to understand. Try CAD drive-throughs on big color monitors when the whole team is present.

Some interesting Web tools allow use of a browser to examine 3-D CAD files without having the expensive CAD program resident at a remote location. Customers and team members can look at design ideas during a phone conference with the rest of the design team. Videoconferencing, which is relatively inexpensive today, is another option.

Small tubular components can be made in four to six weeks using the Keltool process (3D Systems, Valencia, CA), sintering steel powder around a stereolithography pattern.

Schedules are often drawn in a linear fashion. However, the shortest time to market is achieved with schedule sections that overlap and with data that easily move back and forth between sections. This kind of scheduling is commonly referred to as concurrent engineering. But not all concurrent schedules are created equal, and many design tools don't allow a bidirectional flow of information between industrial design, electrical engineering, and mechanical engineering.

Test run and think through your CAD data flow as early as possible in the product definition stage. Select CAD tools that allow a two-way process. For example, industrial design CAD packages look flashy, but it may be easier building the industrial design in the mechanical design CAD package. Solutions might also involve developing relationships with new vendors who have more suitable CAD, holding internal training sessions with existing vendors, or even investing in better translation software.

RAPID PROTOTYPING AND BREADBOARDING

The key words in rapid prototyping and breadboarding are early, often, and appropriately. These guidelines can provide major time savings. In fact, the savings potential is so great that it can pay to buy RP machinery to gain a competitive advantage.

Making accurate physical models of the design used to be the slow stage of the product development process. However, the new RP techniques have reduced both cost and time by about a third during the past five years. Three times as many prototypes can be made for each product in the old time frame, taking designs to a more sophisticated and complete level before committing to tooling.

Much has been written on the slice-and-dice RP techniques in which a special software package slices the CAD model into thin sections that are reconstructed on a platen. A laser or similar photonics-based system solidifies a photopolymerized resin or sinters a plastic powder to create the parts microlayer by microlayer.

In the product definition stage, use obvious tools such as cardboard and foam or laser cutting to build simple kits. Laser cutting is also useful to create easy-to-evaluate cross sections of the dynamic mechanisms. Use laser cutting of circuit board material to create overnight mechanical samples of printed circuit boards (PCBs) to check connector and key component placements rather than wait for turns of the real PCB, which can take several weeks. High-strength polymers such as acetal can be laser cut to evaluate dynamic strength issues early.

FINE-TUNING PROTOTYPES

During engineering refinement, consider all the standard techniques, including stereolithography, selected laser sintering (SLS), solid ground curing (SGC), and so on. Plan on doing several iterations rather than one final check model before tooling. Many RP techniques are acceptable for judging form and fit, but consider SLS when stronger polymers are needed. SLS is one of the few processes that can produce parts for a disposable that can be sterilized and used in vivo or made directly in an elastomer polymer.

Direct AIM uses a stereolithography resin and allows molding of limited quantities (5­50) of parts in the correct thermoplastics (3D Systems, Inc., Valencia, CA).

The cost and quality of straightforward computer numerical control (CNC) machining of plastics has improved dramatically to keep up with other rapid prototype techniques. Machining is the only option for serious strength analysis, though machined parts still need to be carefully evaluated if the final production process is injection molding. Even if the same plastic is used in both machining and molding processes, strength, fatigue, and sterilization performance may vary significantly.

Virtual prototyping is often overlooked, but it is a useful tool. Testers can interact with a product mock-up in simple multimedia software form before significant time and dollars are committed to a project. Virtual prototyping is especially useful for sophisticated user interfaces. A polished demo can be created in just two to four weeks and produces far superior feedback from users than traditional written explanations of product functions. As an added advantage, the user-debugged demo serves as an excellent specification document.

Indigo Medical (now owned by Johnson & Johnson) gained TUV approval for this urological laser only 18 months after conceptualization by using such techniques as multimedia software and silicone rubber tooling.

Work with vendors nationwide. RP techniques are fairly routine, and a local service connection is not as important as it used to be. Schedules often slip when vendors can't open files or the file transfer type is incorrect. Invest up front in data transfer tools and debug the system. Take advantage of the Internet. File transfer protocol (FTP) is a far better and faster way to transfer data than using modem-based bulletin board systems or mailing floppy disks.

TRANSITIONS TO MANUFACTURING

Some medium-run production processes, such as reaction injection molding (RIM) and various casting techniques, translate quickly from RP techniques. For example, RIM uses foamed thermoset polyurethane injected into low-pressure medium-cost tools. This option is useful for sophisticated or detailed components that will be manufactured in the 50- to 5000-piece range. For instance, an enclosure for a diagnostic or surgical product that has external panels may be tooled in RIM polyurethane panels.

Traditionally, this is how the process to transfer to RIM worked: an engineering check model of full-sized components was made, and a single or small number of prototypes were used to evaluate and tweak the design. Drawings were made for the tool, and the RIM vendor then took over the project. This vendor created a new pattern, about 0.05% oversize to allow for shrinkage, and cast the tool. The total elapsed time was about 12 weeks.

In the best concurrent process, multiple iterations are made of the part in a chosen RP technique during the development cycle. A final check model is made using RP, making it 0.05% oversize to allow for mold shrinkage. After confirmation of the final design, the check model is given as a pattern to the RIM vendor, who makes a tool from the supplied pattern instead of creating a new pattern. The total elapsed time is six to eight weeks.

Discuss the details of your prototyping and pilot production plans with vendors before committing to particular RP techniques. They may be able to make trade show models cast in solid polyurethane from silicone rubber tooling and bring up production tooling four to six weeks later. To fully exploit the various RP techniques, at least one person in the company should become an expert in this area. There are many subtleties to these RP development strategies that should be fully discussed with vendors and understood by the development team as early in the process as possible.

TOOLING

Techniques to shorten tooling time have evolved rapidly in the past few years. Many tooling and RP vendors realized that their customers were taking advantage of RP to create their initial models only to be met with a frustrating wait during hard tooling. To solve that problem, vendors are trying to move RP techniques into tooling. Referred to as bridge tooling, the resulting tooling is especially suitable for plastic injection molding.

The tooling can be produced in a quarter to half the time of conventional tooling. However, it may not be suitable for high-volume production. These processes are ideally suited for small, intricate parts. As part size increases, especially in injectionmolded plastic in which the pressure involved requires large tools, the methods become less useful. Keep checking on these processes at trade shows, since they change almost monthly.

For fast turnaround of small batches of tiny injection-molded parts, there are tooling variations that make standard mold-base inserts from a special material on existing stereolithography machines. One of these techniques is called Direct AIM (3D Systems, Valencia, CA), which is typically offered by toolers who are savvy about stereolithography. There are also numerous other techniques that take a stereolithography pattern and use either metal spray or aluminum-filled epoxy to create a low-volume tool. Some of these techniques have been refined for the automotive industry and are capable of making very large parts, such as a car dashboard.

For higher volume and more expensive tooling that can also be turned quickly, consider Keltool (3D Systems), which uses stereolithography patterns and a sintered metal process, or RapidTool (DTM Corp., Austin, TX). For larger parts and medium volumes of 5000 to 10,000, there is fast-turn aluminum or, for higher volumes, P20 steel tooling. Both types of tools can be made at a variety of quality and automation levels ranging from manual pickouts in the tool to sophisticated variations that mimic more conventional hard steel tools. Prices for this tooling vary widely depending on specific vendor or technique. Tooling methods from RP-driven vendors can be surprisingly cheaper than traditional approaches. Costs of fast-turn aluminum or P20 steel tooling can equal or exceed that of standard tooling, but it is possible to almost halve the build time. Getting the product to market that much faster can increase your overall profits and recoup the higher building costs.

Many molding and casting techniques, such as sand casting, shell casting, investment casting, and rubber plaster molding (RPM), can be accelerated with various rapid prototyping patterns or molds. Ask your rapid prototype vendors for suggestions and methods.

It's impossible in a short article to cover the many tooling advances. Their success is highly dependent on your vendor. Check with several vendors, since some are not well informed. Look for information at trade shows and seminars, and make sure you get multiple vendors involved.

PRODUCTION RAMP-UP

To save time in production ramp-up, it is essential that manufacturing engineers be involved from the beginning and have pilot evaluations and associated documentation that matures as the project progresses. Computers also are having a big impact in this area. Progressive medical manufacturers no longer make complex CAD-generated assembly drawings. One company, for example, invested $1000 in a digital camera and created an on-line digital assembly database that allows operators to pace through assembly procedures with photos and simple captions to guide the process. This documentation is still easy to control for ISO 9000 and GMP considerations.

CONCLUSION

Regardless of the project in question, the goal is the same: to get the best possible product to market ahead of the competition in order to extend the product's life cycle and increase profit. To be successful with both product quality and fast time to market, iterate, iterate, iterate at every stage of the design process. Build strong multidisciplinary teams, and use every rapid prototyping and tooling trick in the book to cut the time it takes to get these iterations into as many hands as possible.

Bill Evans is principal and founder of Bridge Design (San Francisco), a product development consulting company.