Three-axis accelerometer chip being developed for MicroNavigator
Mike Daily of Integrated Microsystems Dept. 1738 and a team of more than 40 in Sandia’s Microelectronics Development Laboratory (MDL) are moving full steam ahead to bring a prototype three-axis accelerometer chip and the processes required to make it to product status.
"Moving research results into products is the mission of the Integrated Microsystems Department," Mike says. "And that’s our task with the MicroNavigator."
With sensors no bigger than a grain of pollen, the MicroNavigator is an integrated microsystem that includes minuscule gyroscopes that sense rotation, a Global Positioning System (GPS) receiver (GPS is a satellite system used to locate an object to within a few meters), a navigation computer, and accelerometers. Compared to commercially available navigators in the same performance class, the MicroNavigator will have less than 1/50th the size and weight, as well as reduced power consumption.
First product versions due in FY99
Mike says the first product accelerometers, being built for Kent Meeks of Advanced Weapon Technologies Dept. 2168, will be out by the end of FY99, with the full MicroNavigator units planned for FY2002. Variations of the accelerometer and gyros are also planned for additional applications such as flight dynamics measurements on reentry vehicles and other national security applications.
In fact, demand for micronavigation systems, microaccelerometers, and microgyros is strong.
"I get calls every week from system developers — both internal and external — who want something now, but they want a product, something they can count on," Mike says. "We’re working hard to satisfy that demand."
So why hasn’t someone, somewhere already done it?
"No one else has the technology we have, and that makes the difference," Mike says.
The technology, Integrated MicroElectroMechanical Systems (IMEMS), was developed by a team lead by Jim Smith and Steve Montague of Intelligent Micromachines Dept. 1725. It combines the conventional complementary metal oxide semiconductor (CMOS) manufacturing process of integrated circuits with micromachined mechanical structures on a single silicon chip. By putting both on one chip, the device becomes more reliable, smaller, and can be batch- processed, making it less expensive.
The project, started by Michael Callahan, Manager of National Security Initiatives Dept. 5133, grew out of two developments.
The IMEMS technology developers, jointly funded by Michael and Tom Hitchcock, Manager of Joint DoD/DOE Munitions Technology, arranged a partnership with the University of California, Berkeley, from which a prototype three-axis accelerometer with on-chip control electronics was designed and built. Evaluations by Ragon Kinney of Guidance Subsystems Dept. 2334 demonstrated that the accelerometer worked better than expected, despite some problems that needed to be engineered out for the part to be useful.
"Feasibility had been demonstrated and the results were very encouraging," Mike says.
At the same time John Ellis of Advanced Sensors and Navigation Dept. 2525 was studying ways to apply this new technology to navigation applications, coming up with a variety of potential national security uses and establishing their performance requirements.
Support for the product development effort came from Michael Callahan and Harry Weaver, Director of Microelectronics Technologies Unit 1720, who asked Mike Daily and his team to come up with a "product a systems engineer could really use."
The first step was to determine what it would take to go from a research prototype to a product. Mike put together a group of experts from a variety of departments, including wafer processing, integrated circuit design, modeling, testing, reliability, failure analysis, and packaging, to look into the matter.
They found two pieces missing — a product infrastructure and the need for a better understanding of the product they were to develop.
Missing infrastructure
"Since this is the first time we’ve actually tried to develop a product line with IMEMS technology, we invested heavily in creating parts of the infrastructure that we needed. This involved creating a product development methodology, developing a model-based design capability for the product designers, and building the ability to test and package the parts in a clean environment, basically a clean room," Mike says.
Cleanliness is necessary until final packaging since moisture, particles, and other contaminants can ruin the exposed mechanical structures once they have been "released" by etching away the sacrificial oxide used to protect them during the wafer fabrication process.
Separate clean rooms for packaging and testing were constructed, a move that will benefit other microsystem projects under way at the Microelectronics Development Laboratory.
Necessary knowledge
While the product infrastructure was being established, the IMEMS designers, led by Dahlon Chu of Analog Microelectronics Dept. 1736, were acquiring the knowledge needed to construct accurate models.
"The first thing we actually built was a wafer design with test structures on it," Dahlon says. "These were put together by our experts in IC [integrated circuit] processing, reliability, failure analysis, packaging, test, mechanical design — you name it. Along with a few functional structures, it has hundreds of test structures, which are characterized to extract the information we need to construct detailed models suitable for product design."
As part of the initial infrastructure planning, the MicroNavigator team defined a process flow for how the IMEMS wafers would be handled and processed from wafer fabrication through the delivery of packaged parts. Yields had to be improved for product manufacturing.
To do this, "yield hit" information was needed to identify where and what the problems were.
"We needed to know if a failure was due to bad CMOS, damage during the mechanical release process, or damage during handling and packaging," Mike says. "We probed wafers before mechanical release to establish which die had good electronics, then sent the wafers back into the Class I fab for mechanical release — a potential contamination threat for the fab."
They solved the contamination problem by developing a way to move the wafers out of the fabrication area, probe them, and return them still clean. They also improved the release process, increasing overall yields by 35 percent.
Mike and his team are currently working their way through other problems — such as figuring out how best to handle and package the parts and eliminating long-term drift in the accelerometer.
Mike says the idea behind the MicroNavigator is to build very small, low-power inertial instruments and navigation systems that can be used in testing stockpiled weapons or be put in small munitions to guide them to a target.
Telemetry systems will be tiny
Currently when nuclear weapons are tested, the testing instrument or telemetry system is large. Because of its size, before it can be put in the weapon, the nuclear pit (actually a mockup of the pit with the same weight characteristics) — the heaviest part of the weapon — must be removed to make room. As a result, the measurements have less "fidelity" or realism because a key piece of the weapon is missing.
By reducing the size of the telemetry package and making it tiny, the testing instrument can be put into the weapon without removing the pit, making measurements more realistic. These Enhanced Fidelity Instrumentation (EFI) telemetry units, under development in Telemetry Systems Engineering Dept. 8416, need accelerometers and gyros for flight dynamics measurements, as well as for navigation
measurements.
The possibility also exists that these Micro-Navigators could be made small and cheap enough to be part of an artillery shell or other small munitions, serving the function of guiding the munition to the target.
"The potential use for the MicroNavigator is endless," Mike says. "If they are batch processed, the cost and size could be reduced by orders of magnitude. You could put an inertial navigator in almost anything."