Follow the yellow brick road...map

The longest road map begins with a single real part.

By Bill Schweber, Technical Editor -- EDN, 1/6/2000

I recently received a call from the public-relations agency of a budding semiconductor vendor. Would I be interested, the company asked, in hearing about a client's wireless-component road map?

Normally, because I'm a nice guy, I'd consider giving them some of my time, but this was the third time in six months that a representative for this vendor had approached me with potential road-map stories. The first time, I patiently listened to the representative talk about the company's impending killer RF IC and associated plans, but now all I wanted to hear was that the promised part was real. The representative's reiteration of and elaboration on that road map was no longer of interest to me; I was looking for the company's ability to deliver, not to plan and promise. To paraphrase the words of the sages, "the longest road map begins with a single real part."

Increasingly, I'm seeing both start-up companies and established vendors saying they want to get into this or that market because they believe they have superior design skills, process technology, or some other attribute that will let them succeed. Sometimes it's true, but often it's simple opportunism as they reach for the next hot market. The road map is part of a company's plan to prove it's not a one-shot wonder. But what if there's only a lot of bluster but little reality behind the wizard's curtain, just as Dorothy and friends found when they reached the Emerald City?

The real problem comes when you bet the completion of your next product on the availability of a company's component. It's really very simple: When it doesn't deliver, you don't deliver. You undoubtedly have a narrow window of market opportunity, and you need the part at least several months before your shipping date so you can properly evaluate it, design it in, and make sure you can properly integrate it and its subtleties with the rest of your system.

Unfortunately, it is hard to tell in advance which vendors have what it takes and which don't. I remember a representative from one vendor who started speaking about the company's IC—which was in its early design phase—using the sentence "It will do this" but ended up using the sentence "It does this." The representative wasn't trying to deceive me. It's just that, because the company lives with the design day and night as it tries to design and fabricate the IC, the goal and the reality start to blur. I don't think the vendor could separate its desire from the actual tangible product.

But you have to do the reality check. When you're considering going with a new vendor, or one with little history in a given application, you know you're taking a chance, and that's fine. You have to carefully look at the risk-versus-return balance. Even experienced designers and vendors stumble, and their chances of unexpected problems are higher if they are new to the application; using a new process; using a new foundry; or haven't yet been through the entire design, layout, fab, and test cycle. Even luck plays a role in the outcome, because there's a lot of magic in designing leading-edge ICs. But remember what athletes say: "The more I practice, the luckier I get."

One issue I always raise with vendors, especially those that are using an independent foundry for a new design, is this: "Tell me about your final test." If the company says the foundry will test the parts, I'm a little skeptical. If the company says it has purchased or leased the appropriate automatic test equipment, I feel a little better, because the vendor is taking direct responsibility for the final pass/fail hurdle. You all know that it's the second- and third-order subtleties in a design and subsequent test that make the difference in breakthrough parts. But uncovering these subtleties is a challenge, and one that your proposed vendor has to face directly. Road maps are of little use to potential users unless the vendor can put his feet firmly on the road and actually begin the journey, not just talk about the planned trip.