For many years, sensors used to measure quantities such as pressure, temperature, and acceleration have been the mainstays of automotive electronics. However, a large number of systems requiring such functionality, such as fuel injection control, fuel economy, and safety systems (including "smart" airbags and tire pressure monitoring), have moved beyond simply placing a sensor where needed and sending the signal back to the control unit. The emergence of various digital bus systems has facilitated centralized processing and simplified wiring harness connections, making it possible to fully utilize the processing intelligence integrated within the sensors responsible for data acquisition. However, this architecture also raises reliability concerns; furthermore, typical automotive applications face significant challenges in continuously reducing costs. High-speed buses such as Controller Area Networks (CAN) and the more powerful FlexRay bus have traditionally been used in applications requiring intensive computing and rapid processing, such as engine and chassis control. In contrast, low-cost, single-wire Local Interconnect Networks (LIN) were developed for less speed-critical, but more simplicity and low-cost body electronics applications, such as seat positioning and temperature control. Single-wire LIN also means lighter weight, resulting in better fuel economy. Matthias Poppel, Global Advanced Embedded Control Marketing Manager at Texas Instruments (TI), stated that placing the control IC and sensor on the monitored mechanical part can save space and simplify the processing workload of the system's central processing unit. However, he added that the reliability of such mechatronic sensors that integrate mechanics and electronics is an issue. "Their flexibility is also insufficient (for design engineers) because such integrated sensors may only have one supplier, while components assembled from discrete parts generally have several suppliers," he added. According to Poppel, "The transition to 32-bit MCUs and the satellite processor/sensor architecture attached to the mechanical part still need to be tested." He cited the example that a PCB mounting a mechatronic sensor attached to a measuring device must be tested and verified to ensure stability and reliability under any foreseeable operating, load, temperature, and vibration conditions. Steve Henry, Senior Marketing Manager for Programmable Controller Power at Freescale Semiconductor's Sensor Business Unit, also mentioned concerns about sensor reliability and the impact of packaging. “For example, the challenge with sensor-accelerometers is that users want the integrated sensor and controller ICs in a smaller package,” Henry said. Freescale offers a 6mm x 6mm QFN surface-mount microelectromechanical systems (MEMS) device, but the dies must be stacked to meet smaller overall package requirements so they can be mounted in increasingly compact spaces. When stacking an accelerometer (which monitors mass and resonance) on a processor chip, the effects of external environmental factors such as load and vibration on stress sensitivity need to be considered, Henry said. “Wrapping the accelerometer in silicone, RTV, or other materials will insulate it from the package,” he noted. But then a different die bonding technology needs to be developed, because “it might try to use wire bonding to attach it to a pillow,” Henry explained. This also affects reliability. “Because you can’t integrate all the functionality onto a single silicon die, you need to stack the dies to optimize the process,” said Mark Shaw, Marketing, Applications, and Systems Manager for Freescale’s Sensors Products Group. "You can take advantage of high chip logic density and high voltage withstand capability (for sensor processors) without being limited by MEMS processes," he said. Henry, however, believes sensors should be viewed from a packaging perspective. This leads to the conclusion that instead of placing them on a single silicon wafer, they should be separated into two chips: the processor and the sensor. Shaw points out that larger MCU chips also have larger die sizes. "MEMS has a higher failure rate and lower yield," he notes, adding that putting the processor and MEMS sensor on a single chip would be far more costly. If the MEMS in the sensor area fails, even the best processor will be rendered unusable. While agreeing on the need to improve sensor yield, Frank Cooper, president of mixed-signal chip supplier ZMD USA, believes that moving to what he calls a "single-wafer solution" offers advantages in terms of simplified packaging. This approach is superior to wire-bonding ASICs to sensors and connectors before packaging. "A true single-chip solution puts a G sensor (accelerometer), temperature sensor, or flow sensor on a single chip with signal processing components, thus achieving minimal wire bonding," Cooper says. This reduces the likelihood of sensor bursts, short circuits, fatigue, and contamination during installation. Safety and engine efficiency will remain key concerns for automotive sensor applications now and in the future, especially when tires are in contact with the road . Reliability is also a concern because sensors face harsher environments they've never encountered before. A representative application is tire pressure sensing (TPS). The National Highway Traffic Safety Administration mandated that 20% of new cars in the 2006 model year be equipped with TPS sensors. John McGowan, head of sensing and control at Infineon Technologies, stated that TPS sensors are used in "crowded, hot environments" and must be robust, durable, and cost-effective. Infineon engineers developed such a sensor by placing a CMOS ASIC for data processing and signal modulation along with a piezoelectric pressure measuring element within a common lead frame. McGowan explained that the "three-layer silicon sandwich" structure, with the ASIC sandwiched between two layers of glass, is robust and durable. Freescale's Henry also discussed the issue of "media compatibility," noting that TPS sensors could be exposed to "interesting chemicals" and liquids could splash onto tires in a garage. These include: acid spilled from batteries, assembly lubricants, dust, chemical residues from manufacturing processes, and humid air inside inflated tires. Infineon's McGowan stated that integrating processing capabilities with the sensor ensures the accuracy of functions such as temperature compensation, self-calibration, and failure mode detection. Cost control can be achieved by integrating multiple functions and features onto a single chip (as opposed to the discrete passive devices used in the past) and by mass production. Finally, this intelligent auxiliary sensor allows for a smaller central processing unit to be freed from data processing, enabling faster decision-making. Current tire pressure monitoring sensors are either mounted as a raised piece on the outside of the tire or fixed inside the wheel rim. Because these devices are powered by button batteries, McGowan stated that Tier 1 suppliers are aiming for 10-year battery life. "To achieve this, we use vehicle information in the processing algorithm, reducing the sampling and transmission rate when the car is stationary," he added. Future pressure monitoring may be accomplished by sensors directly embedded within the tire's construction. These sensors must be powered by what McGowan calls "energy scavenging," a technique that utilizes tire flexure to drive a piezoelectric element that powers the sensor. This concept can be extended, for example, to use engine vibrations as the operating energy for a crash sensor. Another approach is to drive the embedded tire pressure sensor from outside the tire via induction. Considerations here include the impact of any metal antenna ring within the tire rim on tire physics. A team led by Magneti Marelli has been working on "smart tires," going beyond simple pressure detection. Their findings were presented at the SAE 2005 World Congress by Project Manager Andrea Neponte and Strategic Innovation Manager Piero De La Pierre (paper 2005-01-1481). Tire testing will go beyond pressure; a triaxial accelerometer on the inner liner will also provide tire force data along three axes, tire contact patch size, and road conditions (via vibration data). While batteries were used during testing to ensure reliable communication, the team believes the strain gauge method cannot provide sufficient power for the application in terms of the power level (300 milliwatts) required by the sensor. Such a tire data system could be used to inform automotive chassis control systems or determine if tire or suspension servicing is needed. Automotive Sensor Outlook In the next five years, other sensor applications will likely include more gyroscope-based devices, says Shaw of Freescale. These devices will provide angular rate data for rolling stability control and other shaft closed-loop control. These gyroscopes will be MEMS-based, and as production volumes increase, the manufacturing cost of MEMS will decrease. Peter Knittl, Marketing Manager for Pressure and Hall Effect Sensors at Infineon, believes that enhancing the performance of airbag-triggered impact sensors will require pressure-based devices instead of the currently used G-sensors. “This shift to ‘active’ sensors is driven by the new government legislation (FMVSS-201) to prevent side impacts,” he says. “When the structure deforms, the G sensor will trigger. But the pressure sensor inside the door will detect a sound wave shortly after (about 5 to 6 ms), while the G sensor's detection time is 10 ms.” Future airbag systems may use both types of sensors simultaneously to enhance redundancy. The development trend of automotive sensor systems not only indicates where sensors will be used, but also reflects “how various bus systems must work together and which application areas each bus is specifically designed for,” says Poppel of TI. Two-wire or three-wire? Cooper of ZMD says he is surprised that the single-wire LIN bus has not yet gained greater dominance in vehicles. Typical applications still use three-wire radiometric sensor interfaces. “Perhaps with the emergence of new digital protocols, no one in the automotive industry wants to risk a recall or a bloody accident.” Therefore, proven legacy devices are easy for developers to use, and they want to be able to easily use stock devices, he continues. “While the industry may be moving towards a digital interface, they are still looking at analog (three-wire) output signals.” Cooper anticipates smaller, lighter, and more powerful sensors within the next five years, but with only a few electronic control unit (ECU) modules processing them. He also believes that reduced wiring (resulting in lighter weight), improved fuel economy, lower radiation, and the elimination of an additional layer of data interpolation will help drive the shift towards digital interface standards infrastructure. “Some cars today have over 100 ECUs, and the goal is to reduce that number,” says Poppel of TI. “Code reusability will also contribute to cost reduction. Furthermore, while leading technologies are seen in handheld wireless applications today, these products don't have long lifecycles (due to their short design cycles). On the other hand, cars are ultimately long-term systems that require high-quality and reliable electronics.”