by Debra Vogler, senior technical editor, Small Times
November 9, 2009 – HP recently announced an inertial sensing technology that enables the development of digital microelectromechanical systems (MEMS) accelerometers that are up to 1,000× more sensitive than high-volume products currently available. According to David Erickson, engineering manager in HP’s technology development organization, Imaging and Printing Group, the new sensors based on this technology can achieve noise density performance in the sub-100 nano-g/square root Hz range to enable dramatic improvements in data quality. The MEMS device can be customized with single or multiple axes per chip to meet individual system requirements; its dynamic range is >130dB with a bandwidth of 0-250Hz that is extendible to 10kHz.
HP views the current inertial sensing landscape as comprising consumer devices (i.e., low-performance/low-cost) and aircraft/navigation applications (i.e., high-performance, high-resolution, expensive, and power-hungry), according to Erickson. "What we are disclosing bridges the gap, bringing together the very high performance with the cost, size, and power you could expect from a MEMS device," he said.
One key to the inertial MEMS sensor technology is a design that uses a 3-wafer (single crystal silicon) construction as opposed to silicon-on-insulator (SOI) (see figure) — this enables temperature stability of the device, which plays directly into its low-noise sensitivity, Erickson told Small Times. Additionally, a large proof mass and the electrode design, which features constant gap sensing surface electrodes, are significant contributors to the performance. Specifically, device performance is enabled by increasing the area of the electrodes and decreasing the gap distance.
Most inertial sensors have a proof mass on springs that moves when the device moves (sensing the position of the mass with capacitive electrodes is the generally accepted approach), Erickson explained, but HP extended that principle by using a proof mass that is 1000× more massive than traditional MEMS devices. That translates into a three orders-of-magnitude improvement in noise floor performance, he said, and combined with HP’s electrode design, results in a device possessing unique features: "a much lower noise floor and a much broader dynamic range and much more thermally stable," he said. The proof mass is suspended by very high-aspect-ratio silicon springs that provide the required stiffness; very deep etches are required to make them. Though HP primarily uses standard etchers, the company has developed proprietary processes, Erickson noted.
Developing the manufacturing processes to release such a large proof mass was also a difficult challenge, according to Erickson: "It’s thick and big, and most MEMS processes don’t deal well with removing a lot of material." He said the innovation came about because the company was looking for ways to construct a micromover — i.e., a way to very precisely move a piece of silicon in a MEMS device. Because a MEMS accelerometer is akin to a motor in reverse, when the researchers built a motor with parallel electrodes to move the proof mass to a precise location, "that insight from the motor problem was applied to accelerometers […] We realized you could sense very precisely a proof mass motion with a similar electrode structure, and then we went to work on a large proof mass. So we came at it from a different angle than most sensor designers are using."
Other manufacturing processes also play a major role. "We’ve been working on manufacturing processes and process capabilities that allowed us to get to much smaller gaps," Erickson told Small Times. Smaller gaps mean larger capacitance, and therefore greater the signal-to-noise ratio (S:N). The way in which the electrodes are structured, the very small gaps and the parallelism of the electrodes — all contribute to an orders of magnitude change when the proof mass moves (i.e. the S:N ratio is higher).
The company believes that its test data suggest the new devices can enable new classes of applications — for example, bridge monitoring and seismic monitoring. "We’ve done quite a bit of work looking into bridge monitoring," said Erickson. "Sensors available today are insufficient to detect vibrations at frequencies that are needed for structural analysis."
"HP envisions that sensor networks utilizing HP’s new inertial sensing technology will create a new paradigm in bridge maintenance and safety," Grant Pease, business development manager, told Small Times. "The technology will provide higher resolution to measure vibration modes in a bridge, which in turn provides a better understanding of the structural health and usage. The inherent low power usage of the technology enables wireless operation over an extended period of time allowing for cost-effective implementation and use." (For additional discussion of bridge and infrastructure monitoring, see Small Times‘ interview with Michael O’Halloran of CH2M Hill: CH2M Hill, HP eye progress in infrastructure monitoring.)
The sensing technology is a key enabler of HP’s vision for a new information ecosystem, the Central Nervous System for the Earth (CeNSE). Integrating the devices within a complete system that encompasses numerous sensor types, networks, storage, computation, and software solutions enables a new level of awareness that facilitates communication between objects and people.