MEMS accelerometers: a fast-track to design success?

LONDON — Motion sensing, largely enabled by Micro Electro-Mechanical Systems (MEMS) accelerometers, has taken more than its usual share of the headlines recently. Whilst not a rags to riches tale – accelerometers have been used in automotive applications, among others, over many years – the acceptance of accelerometer-based features enjoyed by Nintendo Wii's motion-sensing remote controls and Apple iPhone's self-orienting display, has brought accelerometer manufacturers attention they could but dream of in year's gone by.

According to Douglas McEuen, a senior analyst at ABI Research, a handful of manufacturers share the market: “A 'big three' – Analog Devices (ADI), Freescale, and ST Microelectronics (ST) – are joined by just a few others, including OKI Semiconductor and Hitachi Metals America.” Other smaller players include Bosch Sensortec, Honeywell, Kionix, MEMSIC and VTI Technologies.
Mathieu Potin, MEMS analyst at Yole Developpement, believes the MEMS accelerometer market was worth $876million in 2008 – a revised figure, following the financial downturn.

With the total MEMS market worth an estimated $7.6bn in 2008, the technology is an important contributor to MEMS sales as a whole. Though MEMS accelerometer sales are likely to decrease to $846million in 2009 due to a slump in automotive demand, he estimates that the market will recover to be worth $1.6billion by 2012, fuelled by a boom in consumer applications.
For comparison, Yole projects that the total MEMS market in 2009 will be worth $8billion, and by 2012, the entire market will be on the same projection path as accelerometers, to be worth $16.7billion.

Within the MEMS accelerometer market, there are a plethora of uses. Yole cites 25 in its latest report, although Potin acknowledges that the opportunities are endless. The best known are now in the consumer electronics domain, such as the provision of horizontal and vertical sensing to determine portrait and landscape requirements in mobile phones or digital still cameras, enabling pedometer features, supplementing gps navigation, power management, or recognising freefall conditions to provide hard disk drive protection for notebooks or portable audio and video players. The latter was the main driver of accelerometer integration in consumer devices until 2008.

Accelerometers also feature in home appliances such as washing machines, where they are used to adapt the spin cycle to reduce noise, water and energy consumption, as well as to extend the life of the machine. Similarly, vibration monitoring is an important application in heavy-duty equipment and shipping-container management.

What may be surprising to many is that the first MEMS devices were created in the late 80s, with volume driven initially by air bag sensors. This has been followed by integration for tyre pressure measurement and stability control purposes in the automotive sector. Indeed, the automotive market accounted for the most sales with a 45 to 50 percent share of the MEMS accelerometer market today, estimates Potin. However, he believes that the consumer electronics sector could grow its share from approximately 30 percent today to 45 percent or more by 2012.

Explaining the opportunities, Benedetto Vigna, Group Vice President and General Manager of ST's MEMS and Healthcare, RF Transceivers and Sensors Division says: “We are strong believers of MEMS 'consumerisation'. We still see a lot of growth in existing consumer market such as laptops and mobile phones. But we see opportunities in other markets such as remote controllers, digital cameras and portable audio and video players.”

Much of this growth is likely to be contributed by human machine interface applications – something that has surprised some industry watchers. As early as 1998, Analog Devices' accelerometers were designed into game pad controllers by Microsoft, Logitech and Pellican. However, it was the iPhone and Wii game controller that revolutionised their usage. Acknowledges Christoph Wagner, a field application engineer for ADI, which supplied accelerometers for Nintendo's GameBoy in 1999 and the Wii in 2006: “People bought Nintendo's Game Boy, but not the extra motion sensing module. The difference with the Wii is that motion sensing is at the heart of the product.”

Potin says that the success of these products is also a reflection of how the games have been created: “Take Nintendo; most games that make use of motion sensors are ones that Nintendo has developed itself. In contrast, most games for the Sony PS3 have been developed by third parties. From my understanding, it has been complex for designers – especially those outside an organisation ” to adapt to these new capabilities.”

It is the incontestable popularity of accelerometer based features in the Wii and iPhone, plus comments such as McEuen's view that 'the sky's the limit' for accelerometers in the consumer space, that suggest that motion sensing is not a passing fad but becoming almost a 'must-have' feature in some end user markets.

From a technical perspective, Wagner points out that motion sensing is simply the next step for the analog front end: “Audio was the first real-world signal to be processed by signal processors, and the next was video. We see motion sensing as the third big signal processing trend.”

Motion sensing technologies have evolved considerably since the early 90s, when they were first used. According to Potin, Analog Devices gained a head start with capacitive technologies. After this, a number of Japanese companies began developing accelerometers and from 2003 onwards, captured much of the volume with piezo type sensors. Between 2003 and 2007, the market was shared between piezo resistive, capacitive and MEMSIC's CMOS-based thermal accelerometers. Since 2007, capacitive technologies have stolen a lead, largely due to the increased market share achieved by ADI, Kionix and ST, says Potin.

Whilst each accelerometer technology has its advantages, Vigna explains ST's preference for capacitive technology: “It allows us to make a smaller device, has good temperature stability, as well as lower power consumption. These are the driving forces in motion sensor development. We have just release the AIS326DQ which combines these features, and we are working on the next generation being smaller, even less power consuming, and offering yet more temperature stability.”

Another technology trend is the concept of embedding algorithms for ease of development purposes. For example, Kionix' recently released KXTE9 product is a low power, digital-output tri-axis accelerometer, that combines on-chip algorithms to perform orientation and activity monitoring. Similarly ADI's low power digital-output tri-axis ADXL345 accelerometer integrates a FIFO memory block that samples input data to determine if the system should be responding to changes in movement or acceleration. These features unburden the host processor, allowing it to remain in sleep mode and facilitating power usage reduction.

Nowadays, accelerometers are available in a wide variety of measuring ranges, reaching up to thousands of g's. There's normally a compromise to be made between sensitivity and the maximum acceleration that can be measured though. Designers can choose between single or multiple chip implementations, and analog or digital outputs. Notably, whilst many markets use analog output sensors, mobile phones typically use digital.

For cost reasons partly, accelerometers designed for mobile phone applications tend to consist of two chips in one package ” one chip with the MEMS accelerometer features, and the other with the sensing circuitry. Vigna predicts that this combination will dominate for the next 12 to 18 months. He adds: “We may go to a one chip solution, but only if it becomes cheaper. When an acceleration sensor decreases in size, then the cost of testing increases. Usually, the smaller something is, the cheaper it is, but this is one paradigm that is different from cmos, and it is something some suppliers forget.”

ADI's Wagner notes that whilst the base technology is fairly similar for automotive and consumer-grade accelerometers, consumer applications obviously require very low footprints, whereas in the automotive market the requirement is for more robust, reliable sensors that can withstand environmental stresses over a period of 10-15 years and several thousand operating hours.
Vigna says that the biggest challenge when integrating accelerometers for the system designer should be design for reliability: “You need to make sure that the specification of the sensor when it is standalone, continues to be the same as it is when soldered on the board. You will find that some companies' sensors will drift a lot in different temperature ranges after soldering. These are very difficult challenges that not all suppliers can manage, but we are one of few who can.”

There have been great strides in accelerometer development during the last few years, culminating in tiny 3-axis, relatively cheap MEMS accelerometers now widely available. Vigna suggests that temperature stability and combinations of MEMS sensors, such as acceleration and gyroscope features are on ST's development agenda.

Whilst ADI has already integrated signal conditioning functionality on the MEMS die itself, Wagner intimates that this will become an increasingly important: “Especially when dealing with very small signals, such as a gyroscope.”

Potin concludes that the goal for all manufacturers will be 'to make the accelerometer invisible to designers.' This means more embedded algorithms to better exploit the capabilities of the sensor itself. He also alludes to the integration trend, suggesting that we'll see more combinations of sensors, such as accelerometers, magnometers and gyroscopes. Some are well-ahead in imagining the possibilities. Tapani Ryhanen, head of Nokia Research in Cambridge, England presented his newest mobile phone concept design – called 'morph' to the MEMS Executive Congress recently. He envisages rf MEMS, silicon microphones, accelerometers, microbolometers and microfluidics converging to allow mobile phones to sense their environment and the health and temperament of the people in its vicinity. It is clear that despite the economic downturn, accelerometers and other MEMS devices, have much more value to bring integrators to their systems.

This story appeared in the February 2009 print edition of EE Times Europe European residents who wish to receive regular copies of EE Times Europe, subscribe here.

See other stories from this issue here.

You can download a digital edition of the latest EE Times Europe print edition here.

1 comment on “MEMS accelerometers: a fast-track to design success?