Inertial measurement units (IMUs), once cool and shiny in their new MEMS editions, are now familiar old friends. We’ve become accustomed to motion sensors in our phones, so this particular integration of linear and rotational acceleration feels rather established in comparison with some of the new sensors being considered for consumer use.
But there’s still stuff going on in the motion world, and a few of the recent announcements seemed worthy of note. Oddly enough, they all came out within a week of each other; we’ll simply take them in chronological order.
Small 6-axis
First, ST Microelectronics announced what they say is the world’s smallest six-axis motion sensor, the ASM330LXH. The package is a 3x3x1.1-mm3 land grid array; the size comes largely from integrating all six individual sensors on a single chip. Ranges are 2/4/8/16 g for the accelerometer and 125/245/500/1000/2000 dps for the gyroscope.
While targeted for consumer applications in general, there are lots of hints in their release that they’d love to see some of these in cars. They’ve been qualified for “non-critical” automotive apps, which, surprisingly, includes navigation. I suspect that will change when driverless cars arrive…
Bridging bulk and surface
A few days later, ST made another announcement – this not so much about a specific sensor as much as a fundamental process they’re using for their motion sensors. In fact, this process is used for the 6-axis device we just discussed, which they announced the week before.
The deal here is about what they call their THELMA process, which distinguishes itself by having a thicker epi layer than normal – 60 µm. Perhaps a little background here will help provide some context.
Motion sensors have “proof masses” – chunks of mass that react to changes in motion. The sensors detect the effects of motion on the proof mass (exactly how varies by sensor) to provide the readings. In general, the heavier the proof mass, the better the signal.
Bulk-machined proof masses are carved out of the bulk silicon wafer. They’re beefy and perform well, but they’re also more expensive to make. Surface-machined proof masses, by contrast, aren’t built from the wafer itself; they’re made by growing a layer of epitaxial silicon on the wafer. This is much easier to fashion into a proof mass, so it’s how inexpensive consumer-grade sensors are typically made. You lose something in performance; it’s a “you get what you pay for” thing.
According to ST, the typical epitaxial layer for surface machining is 25 µm. By making that layer thicker, they’ve added more bulk to the proof mass; the idea, then, is that their motion sensors should perform better than your typical surface-machined device while still keeping much of the cost advantage.
Compass and Gyro
The next day, mCube announced the smallest eCompass and “iGyro.” I suspect not coicidentally, these are both accelerometer/magnetometer combo sensors. In both cases, the magnetometer takes the lead role, with the accelerometer providing corrections.
In the case of the eCompass, the accelerometer provides “tilt compensation” since most of us can’t hold a compass exactly flat. The accelerometer can detect the acceleration of gravity and therefore knows which direction “down” is, and sensor fusion software can then provide a corrected compass reading.
The iGyro is a “soft” or “emulated” gyroscope. Here the magnetometer provides the rotational information, but the accelerometer is used to help reject “magnetic anomalies” – big metallic items that can distort magnetometer readings.
So, in reality, the difference between the two devices is the sensor fusion software used to turn the raw sensor signals into either compass direction or angular rate outputs.
These aren’t new to mCube, but they announced their smallest versions, both in 2x2x0.95-mm3 packaging.
Analog Vibration
Finally, Silicon Designs announced a new line of accelerometers suitable for vibration sensing, the SDI Model 1510 Series. These appear to be rather different from the highly-integrated-and-digitized sensors going into consumer devices: this provides analog outputs. Then again, vibration sensing isn’t something people have been asking for in their phones. Unless, perhaps, to help distinguish between a real silent incoming call from that phantom vibration feeling.
It can be used to measure either DC or AC acceleration, with a single-ended or differential output. Different family members provide an acceleration range from 5 to 100 g.
That’s all the IMU news for now. Well, actually not: there’s one other spin on motion sensing that we’ll look at next week. It’s worth a separate discussion.
(Images courtesy mCube, Silicon Designs)