For roughly a decade, MEMS accelerometers have been making automobiles safer by triggering air bags in the event of a crash. But manufacturers of the tiny sensors have always had larger ambitions: a world filled with gadgets that sense and respond to motion.
But despite occasional design wins and recurring buzz about the potential of motion-sensing devices, MEMS accelerometers have so far proven to be too expensive and too limited in function to be broadly adopted by the potentially lucrative computer and consumer electronics markets.
However, that seems likely to change soon. Within the past 12 months, several accelerometer manufacturers have announced and begun shipping sensors that can discern more-complex and subtle three-dimensional movements - that is, by sensing motion along three axes instead of only two. This smaller, cheaper, more capable generation of chips is already being used to protect hard drives in computers and other products from data loss due to a shock or a fall. A few CE manufacturers are actually shipping products that implement some of the motion-based interface fantasies that have been floating around for years.
Is 2005 the year that MEMS accelerometers conquer the consumer electronics promised land? Probably not. But the conditions are ripe for broader adoption in the near future.
Accelerometers that can detect motion along only one or two axes are well suited to environments where likely motion can be easily predicted. But laptops and consumer electronic devices are far more freewheeling. It’s possible to use a combination of two-axis accelerometers or MEMS gyroscopes to sense motion in three dimensions, and at least one company has done exactly that: IBM used two-axis sensors from Analog Devices to add drop detection to portions of its ThinkPad line in early 2004.
Third dimension adds viability
But building a system from two-axis chips adds cost and complexity. Marlene Bourne, vice president of research for EmTech Research, believes that the addition of a third axis to the single-chip packages will make the accelerometers viable on a much wider range of devices.
Industry observers see a clear opportunity for integrated three-axis MEMS accelerometers: protecting hard drives from data loss due to the shock of a fall. Over the past few years, hard drives have spread from laptop computers to audio and video recorders and players. As drive sizes and prices decrease, that trend will accelerate. In December 2004, Samsung introduced the first cell phone equipped with a hard drive, and other manufacturers are expected to follow in 2005 and 2006.
However, hard drives have an unpleasant Achilles’ heel: If they go crashing to the floor while reading or writing data, the shock can damage the delicate drive head or platter, essentially killing the drive and rendering the data practically irretrievable.
Adding an accelerometer to a device adds the ability to detect free-fall and notify the hard drive of a potential collision. The drive can then brace itself for impact, significantly increasing the likelihood that it and the data it contains will survive to play another day.
And those three-axis chips are approaching price levels that make them feasible in devices that cost a few hundred dollars. Freescale Semiconductor has announced prices of just over US$5 per chip in 10,000-piece quantities. Industry observers expect that increasing volumes and increased competition will drive down prices even further - in a recently published report, Bourne cites prices “approaching US$1 per axis”. By the end of 2005, at least 10 companies worldwide will offer the chips, including Analog Devices , Freescale (Avnet ), Kionix, Hitachi Metals, Oki Electric and STMicroelectronics (Arrow ), to name just a few of the high-profile players.
Determining the trend
However, many open questions remain about how the accelerometer trend will play out. For instance, should a consumer electronics OEM add the sensor directly to the logic board or request that the hard drive manufacturer build it directly into the drive?
Building the accelerometer directly onto the logic board permits an OEM to easily switch between hard drive vendors - a factor that reportedly played into Apple Computer’s decision to add a hard-drive-protecting accelerometer to the PowerBooks it launched in January 2005. But some manufacturers would prefer that the hard drive vendor do the heavy lifting. “If you’re a system designer, you might not want to mess with it,” says Andy Higgenbothem, senior director of consumer electronics for hard drive manufacturer Western Digital.
Demand is still light. Only a handful of MP3 players protected by accelerometers have come on the market. Not surprisingly, many OEMs are still balking at the price. “I’ve had MP3 player guys come to me and ask, ‘Can you do it?’” says Higgenbothem. “I say, ‘Sure, but it’s going to increase your costs,’ and they say, ‘Forget about it.’” Accelerometer prices are still high enough that they’re not simply a check-box item, and there appears to be little, if any, hard data about drive failures in computer notebooks or consumer electronic devices due to shock.
Beyond the pragmatism of data protection, proponents of motion sensors have been touting their capabilities in user interfaces for consumer electronics devices.
For example, user might tilt their PDA to scroll through a list of contacts. Previously, these interface ideas haven’t left the drawing board, but now a few consumer electronics companies are actually experimenting with motion-based interfaces in products that are being shipped.
Pantech, Sharp, LG Electronics and Samsung all have phones available in Asian markets that respond to motion.
The most ambitious may be the Samsung Electronics SCH-310 mobile, which responds to several complex movements: It recognises numbers that users dial in the air, shaking the phone twice deletes a message or ends a call, and users can make the embedded MP3 player skip forward or backward a track by giving the phone a sharp flick to the left or the right.
Motion-based interfaces may turn out to have practical applications. Bruce Finke, vice president of sales and marketing for Kionix, says that one of its customers is using the company’s tri-axis accelerometer to design devices intended for situations in which traditional input would fail. For instance, users might be wearing gloves that make it difficult to press small buttons, or they might not be able to use two hands.
Not surprisingly, consistent predictions of unit volumes or sales revenue are hard to come by at this stage. Even if accelerometers end up following hard drives into mobiles and other handheld consumer electronics, nobody can say for certain how many mobiles or how many hard drives.
Finke says that, based on what he’s seeing right now, he could easily envision Kionix selling several million accelerometers per month by the middle of 2007.
Koshi Okamoto, executive director of marketing for Hitachi Metals, says that his company has stated that it expects three-axis accelerometer sales to total US$26.6 billion in his company’s fiscal year 2007. At today’s prices, that works out to between five and eight million chips per year. Bourne, on the other hand, is far more conservative: “We’re talking a few million units,” she says. “Tens of millions of units in five years would be really strong growth.”
Uncertainty aside, MEMS accelerometers do seem to finally be poised for their big break into consumer electronics, if only to keep your hard drive safe.
Although the logic of data protection is easy to sell, the logic of shaking a device as an input mechanism is less clear. Either way, the market is bound to get shaken up soon.