To reduce the development cycle of MEMS inertial sensors destined for commercial applications, it is possible to reuse technologies developed for automotive products. This approach confers the added benefit of keeping the same high quality level as automotive products. As a result, commercial low-g accelerometers have spun off from the automotive realm to come very close to achieving shorter design cycles and higher quality as well as lower cost.
In particular, inertial sensors that have been developed using a two-chip (transducer and signal-conditioning) approach make it easy to reuse existing technologies. For example, IC technology that was developed for other products such as analog devices or microcontrollers can be reused for the signal-conditioning IC. Another advantage of the two-chip stack is shorter design cycle time because transducer and IC design can be done in parallel, also resulting in shorter wafer process time.
The automotive and commercial inertial-sensing markets overlap in an acceleration-sensing range in the low-g's realm, typically below about 8 g's, although automotive accelerometer sockets reflect needs that extend from about 1 g (for tilt, dynamic suspension and other functions) to 250 g's or more (for crash detection). In short, commercial applications cover the more benign acceleration environments commonly experienced by humans rather than the shock waves resulting from the crumpling of metal.
In addition, the commercial OEM sector is much more diversified and fragmented than the concentrated automotive arena. Recently, more and more consumer products, including cell phones, portable hard-disk-drive products, videogames, car navigation systems and digital cameras, have begun using inertial sensors to monitor tilt, motion, shock, vibration, position and so on. And with many cell phones now including games, a digital camera, GPS, a pedometer and anti-theft features that harness inertial sensors, consumer applications are becoming a huge market for the devices.
That said, in the commercial market, low price is the highest priority for selecting a supplier, sometimes to the point of slashing features until the price meets a customer's target cost. Therefore, reducing cost for consumer sensor products is critical to market entry. The conventional methods employed to reduce the cost are to reduce die size and test time, improve yield, use a low-cost package, produce in high volumes and reduce development cost and time.
The automotive and commercial markets also differ in how they rank the relative importance of performance, reliability and cost for low-g applications. Safety-critical automotive usage demands the highest in quality and reliability. The most important constraints for commercial markets are cost, time-to-market and low power consumption.
To demonstrate acceptable levels of quality and reliability, automotive inertial sensors must, at a minimum, pass a qualification battery designated by the Automotive Engineering Council. Often those parts must undergo additional and more rigorous customer-specific qualification tests.
Closely coupled customer reporting of each failing part further bolsters the quality and reliability process. The level of scrutiny demanded in the automotive market of all failures, in terms of an exhaustive failure analysis and regularly monitored customer failure rates, gives a realistic bite to a zero-defect initiative. Single-digit ppm failure rates are still considered key.
This whole quality and reliability backdrop isn't cheap. In some respects, automotive "grade" accelerometers are akin to the Class B parts of the mil-spec world. But unlike the military market, automotive customers still push pricing and time-to-delivery, keeping those attributes as part of their needs, albeit ranked lower in priority.
Since consumer inertial products do not require performance as high as that for automotive, the circuit can be simpler, which helps to reduce the die size and test time while improving yield. The package has to be low-cost (e.g., plastic package), but also small and thin for portable products. Also, high-volume production helps to reduce the per-unit cost.
The commercial market's seemingly endless quest for low cost and short product cycle times is satisfied using several approaches. Commercial parts spun off from automotive platforms carry the highest in quality and reliability as their legacy. This legacy's price tag represents sunk costs (that is, previously incurred capitalized and expensed costs). Additional reduced time-to-market and low power can be addressed for commercial products by tailoring or reusing automotive platform elements, building blocks and packaging.
Three of the biggest contributors to cost when developing and producing inertial sensors are qualification, test and burn-in. In line with the customarily accepted commercial practice, burn-in is removed from the production flow. Further cost is removed by the judicious choice of those parameters to be tested at certain operational corners. The remaining parameters at all other operational corners are guaranteed-by-design. The guaranteed-by-design label is justified by employing enhanced characterizations, relying on the already established very high parametric Cpk (capability of the process) values and extended qualification regimes.
The cost for this approach is slight to none. If the commercial part is specified well within the operational window of its automotive cousin, which many times is the case, the cost is zero. Otherwise the cost is minimal when additional test intervals are folded onto preexisting characterization and qualification activities. This approach also dramatically compresses new-product introduction cycle time, with commercial customers accepting the preexisting data from the "cousin" parts or part components and qualifying the part at their module level. This joint working relationship proves to be mutually beneficial.
Virgil P. LaBuda (Virgil.LaBuda@free scale.com) is a senior product engineer for low-g inertial sensors and Akihiro Ueda (Akihiro.Ueda@freescale.com) is a program manager for new products development at the Inertial Sensors Operation of Freescale Semiconductor Inc. (Austin, Texas).
See related chart