What accelerometers are and how their sensors work
The article explains the theory of accelerometers and its sensors.
Have you ever wondered how your phone knows which way you’re holding it? You can thank a gadget called an accelerometer for that. It pulls off its little magic trick by sensing a change in gravity and recalculating its orientation, much like your inner ears.
As you can tell by the name, the accelerometer device measures acceleration forces, both static and dynamic. Static usually refers to a continuous force, like gravity. And dynamic means movements like tilting or turning your phone.
How does an accelerometer’s sensor work?
With MEMS, of course! MEMS stands for Micro-Electro-Mechanical Systems. You’ll find that the most common accelerometers have piezoelectric sensors. Many industries use them for monitoring, control, and testing processes. And of course, the phone industry uses them so that your phone knows whether it’s upside down or not!
You start with a semiconductor layer and stick it to a wafer – not the vanilla kind, the electronic kind. Stick the layer to the wafer with a thick oxide, then pattern it to the accelerometer’s geometry. The semiconductor layer has holes, so the stuff under it needs matching holes.
Next, you carve a cavity in the oxide to contain a small mass. This mass has a couple of arms holding it up. Directly below the geometry, another cavity lets the mass in the first cavity move at right angles to the surface.
Okay, now we have a couple of transducers stuck on either end of a hollow tube. You polarize these transducers and set them to a specific capacitance. Then you partially fill the tube with a heavy liquid and excite the accelerometer.
While you have the device excited, you need to measure the total output voltage. You may also need to tweak the volume of the liquid until you get the voltage you want. Finally, you check the outputs of each transducer, find the voltage difference, and note the dominant transducer. There you have it, an accelerometer with a piezoelectric sensor.
Temperature transducer with thermocouple
Because we live in a three-dimensional world, accelerometers must use two-axis or three-axis models. Cars often stick with two, as they rarely go up or down unless they’re hovercars. They just need to know whether something is coming from the left, right, back, or front.
Most phones, on the other hand, use three, as people wave them around and drop them at some point in their lives. These sensors are now so accurate that you can even control some games by tilting your phone.
If you want to know more about accelerometers, please ask our engineers!