| Coping with RoHS -- the case for fast-turn design |
by Lawrence Ricci (Jun. 5, 2006)
 Foreword: This guest perspective explores the options available to board and device makers that are faced with premature chip obsolescence due to the EU's new Reduction of Hazardous Substances (RoHS) guidelines. The author contends that product manufacturers basically have two alternatives: buy a ton of the EOL (end-of-life) chips, and risk becoming "sitting ducks" knocked off by more agile competitors; or, use the opportunity to redesign and upgrade the impacted product, thereby rising "Phoenix-like" from the ashes of obsolescence.
Both options carry risks, writes Lawrence Ricci, Business Development Manager at board-maker Applied Data Systems (ADS). However, the risks of the Phoenix approach can be minimized -- and its advantages can be maximized -- by utilizing fast-turn designs based on "integrated platforms" that combine off-the-shelf single-board computers (SBCs) with industry-standard operating systems such as Windows CE.
Coping with RoHS -- the case for fast-turn design
by Lawrence Ricci
Introduction
A few years ago, when the RoHS (Reduction of Hazardous Substances) initative was being developed, it seemed it would have a soft impact on high-tech devices. This has proved not to be true. The tiny amount of lead-based solder on PCBs is of issue, and the even tinier, microscopic, quantity of lead inside semiconductor chips is of issue as well. Semiconductor suppliers, always ready to move to next-generation technology, were quick to issue EOL (end-of-life) notices for non-RoHS parts. This earlier-than-expected component obsolescence, from the supplier side, impacts all OEMs, including those like medical device manufacturers who are shielded from direct application of the RoHS statute. These OEMs (original equipment manufacturers) are now seeking alternatives.
Lifetime buy or redesign -- be a Sitting Duck or a Phoenix?
The tradeoff is simple -- make a "lifetime buy" of the non-RoHS chips and fight approvals on a device-by-device basis, or design a new device with compliant parts and, probably, enhanced features. The two approaches differ in cost, risk, and potential benefit. Ultimately, the measurements of cost, risk, and benefit are made in a competitive area, where the wisdom of each OEM's choice is compared, by the invisible hand of the market, to the choices of the OEM's competitors. The real impact of the choices made today may not be known for years, or even a decade, but it is likely to be a dramatic and definitive impact, well understood in retrospect.
The powerful dynamics associated with the impact of RoHS became evident to me when two OEMs contacted ADS in the same week. For several years, these companies (in completely different industries) have been evaluating the move to a next generation product, but have not yet made the commitment to embark on a new design. Suddenly, the same week, I got a call indicating that both programs were now on the front burner. Why? Because both companies got the same EOL notice, and needed to make plans to either buy over a million dollars' worth of inventory, or accelerate a redesign/replacement program. In short, they were faced with I call the "Sitting Duck" vs. "Phoenix" choice.
Both companies are major enterprises, and have been wrestling with powerful but sometimes ill-defined constraints and objectives from their markets, their technologies, and their internal "politics." The RoHS-linked EOL notices, however, injected a laser clear demarcation into their planning process. A line was drawn in the sands of time. They realized they had to put a million dollar level investment into either (a) obsolete inventory, or (b) a new product design. Failure to do either would be tantamount to withdrawal from their business segments. These are creditable, defendable Phoenix and Sitting Duck options, but the "dead duck" option is not really a business proposition you can take to shareholders.
Sitting Duck vs. Phoenix tradeoffs
Let's consider three key tradeoffs in the Sitting Duck vs. Phoenix decision -- costs, features, and risks.
Costs
The initial lifetime buy is just the first of many costs the Sitting Duck must pay. During the period of EOL, other components will disappear from the supply chain, most likely at an increasing rate. Many of these will have no drop-in replacement, so more lifetime buys will be required.
From a business point of view, inventory is not "money in the bank." In business, money in the bank pays interest. Aging inventory does the opposite -- you must reserve income each quarter against eventual write off. In other words, you buy it with cash, and then you get a constant drain on your income. Depending on the GAPP standards invoked, a million dollars of inventory could demand tens of thousands of dollars a quarter from income (e.g. margin) just to sit there.
The initial cash demand for a lifetime buy is typically more than the complete cost to develop a new product, and the implied commitment to subsequent lifetime buys multiplies the eventual cost to the Sitting Duck.
Features
The lifetime buy does allow the continued supply of the "same old product." The problem is, in today's dynamic and international markets, continual feature enhancement is demanded. Devices that were once based on stand-alone 8-bit microcontrollers are now expected to have network interfaces, Web accessibility, graphics, and other high-end functions that the user community gets and expects, in devices as modest as cell phones and thermostats. Likewise, users have come to expect that even as functions go up, costs come down.
Still, there are markets where features remain static for decades, like certain power meters used in electrical utilities. Each OEM that considers adopting the Sitting Duck strategy must decide whether the impacted product's feature set and price point can remain static for the duration of the lifetime buy.
Risks
OEMs always want new products, but the development process is not without risks. Many projects, once started, never finish. When the continued supply of the older product is not assured, the risk is serious. Risk is the reason why both companies that inspired my investigation into RoHS EOL issues did not start their programs years ago. Now, however, it is risk-avoidance that may compel them to move forward.
If there are multiple competitors in a business sector, some will choose the lifetime buy Sitting Duck option and some will choose the "accelerated redesign" Phoenix option. These opening moves may set the pattern of competition for a decade or more. The companies that select the stasis of a lifetime buy strategy will risk becoming Sitting Duck targets, with fixed features and cost. Sooner or later, one of the companies that adopts the more flexible Phoenix fast-turn development option will get the Duck in its sights and knock it off.
OEMs need to realize that their "ponds" are open to the sky -- new birds can fly in from anywhere at any time. A few fat and happy Sitting Ducks are how these more competitive birds can best find, and eventually own, the pond.
In fact, the situation for the Sitting Duck is even riskier, going beyond competitive forces. Other non-competitive forces like additional regulation could intervene during the period of their lifetime buy and render the product, and its millions of dollars of stockpiled inventory, worthless.
The Phoenix rises from the ashes of EOL with a newer, more up-to-date product |
Minimizing risk for fast-turn development
So, faced with a supply-side EOL decision, the path forward is best understood as a risk vs. risk decision. The risk of a Sitting Duck strategy, while not immediate, is there to estimate and live with. The risk for Phoenix-like fast-turn development, however, can be reduced.
The technology that an OEM's customers value is usually based on the OEM's unique interface hardware and application software. No matter how complex, elaborate, and difficult to develop, platform-level technology such as networking, graphics, security, and so forth are typically commodities designed against industry standard compatibility specs. Vintners make wine, they don't make the bottles. OEMs, especially those pressured by EOL choices, should focus on their core value.
By outsourcing the system's platform-level hardware and software components, an OEM can begin application development on day one of the project, thereby dramatically reducing risk. Indeed, it may be possible to have a proof of concept device running in a few weeks or months, before the final decisions have to be made on EOL lifetime buy investments.
Reducing risks with integrated platform solutions
The use of a complete and operating system such as Windows CE, integrated with a complete board-level single-board computer (SBC), may not yield the lowest product cost, but it can often yield the lowest overall product lifecycle cost (encompassing the "DIME" of Development, Introduction to factory, Maintenance, and EOL considerations).
The ADS Bitsy SBC is fully supported by Windows CE
(Click image for details) |
What is most important, when faced with a hard EOL deadline, is for a platform solution like this to be a low risk program. Both the hardware and software must have already been tested and proven in dozens of applications. A custom build of the OS can be performed in a few days, providing the OEM with an exact match to their target application's environments. Later, once critical deadlines have been met, and there are opportunities for cost reduction and other optimizations.
Windows CE is very complete system, distributed on seven DVDs. It includes network, graphics, CODECs, database, and other functions required to implement almost every application. It offers all this at a cost, but in a modular fashion. For example, the Windows CE license can drop from $20+ to below $5, for OEMs willing to develop their own graphics interface -- something they'd need to if they chose to use a so-called "free as beer" OS such as Linux.
Adding other third-party components
There are many other third parties in the embedded software community as well. These companies can provide drop-in packages for database, security, device management, and other vertical applications representing hundreds of years of design equity. For example, one very successful such company, OSI Software Inc., has an informal marketing slogan: "spec us in, design us out." There is some wisdom here -- OSI, with its ECHO product, can get a customer to market quickly with a terrific variable- and event-vs.-time database (like an event recorder). Later, if the OEM feels it can do better and wants to capture the license cost as margin, it can design ECHO out of the product.
OSI is only one company with products for the OEM developer. The market of third-party embedded software products is rich with innovation and opportunity for OEMs who want to fly like a Phoenix and move fast to next generation product functions.
Hardware cost reduction of a fast-turn product development is also possible. For example, many SBC suppliers are able to move from "standard" embedded SBCs, which are designed to be paired up with an OEM's custom application baseboard, to low-cost all-in-one designs that combine the SBC along with OEM's baseboard design and made just for the OEM under an exclusive supply agreement. At higher volumes, even the production of the all-in-one board can be transferred to the OEM's factory or designated contract manufacturer. And, of course, once the Phoenix-like OEM has completed its first post-EOL rebirth, it can potentially do its own design for the subsequent generation.
As the product is managed across Phoenix generations for low lifecycle cost, each component of the lifecycle cost DIME can be optimized, in turn, and interactively with the others. But facing a hard EOL deadline, it's essential for design cost and factory introduction time to be limited and maintained as risk-free as possible.
Conclusion
We live in dynamic times -- change is forced on us from all directions. In the evaluation of internally engineered vs. purchased platform based products, there is always a tradeoff between margin and speed of development.
When entire industries are faced with simultaneous EOL notices, is is currently happening due to the EU's new RoHS requirements, this may translate into a choice between being a Sitting Duck, knocked off by more agile competitors, or becoming a Phoenix and rising to new heights from the ashes of obsolescence.
 About the author: Lawrence Ricci is Business Development Manager at Applied Data Systems in Columbia, Maryland. A Microsoft Windows CE MVP, Ricci is active in the embedded community and publishes extensively. His particular interests include development of smart, communicating devices based on open standard Web services technologies. He can be reached via email.
Related stories:
(Click here for further information)
|
|
|
news feed