Engineers developing heavy equipment such as tractors, cranes, wood chippers, and construction equipment need to include incline sensors and indicators—both to ensure that workers do not operate the equipment in unsafe locations and to prevent their equipment from tipping over during use. They also have to be assured that the sensors will work despite vibrations and shocks (common conditions for off-road equipment).
These sensors, which are based on accelerometers, must also function reliably despite high shocks and vibrations. But there are generally two types of accelerometers currently available—capacitor-based and thermal versions—and both are microelectromechanical devices (MEMS). So what’s the difference between the two?
To select the right accelerometer for the application, engineers must consider several design variables. These include structure, resonance, reliability, stability, bandwidth, power consumption, and cost. Designers also need to understand the key principles of each device and how it measures inclination.
Accelerometers and Tilt
Accelerometers on heavy equipment measure tilt and roll, just as they do in aircraft. The devices can use two (2D or dual-axis) or three (3D or three-axis) orthogonal axes. They measure the acceleration due to gravity, then calculate inclination from the amount of acceleration due to gravity measured on the axes.
A 2D accelerometer can measure both pitch and roll, and the acceleration measured on the inclined axis is a function of sine of the angle. This limits its theoretical usable range to 0 ± < 90°, but in actual use, the limit shrinks to 0 ± 70° because the sine function starts to flatten out as it approaches 90°. However, a 2D accelerometer can measure pitch or roll through a range of 0 ± 180°.
It takes a 3D or two 2D accelerometers to measure pitch and roll over the full range of an object’s orientation with respect to gravity—i.e., 0 ±90° of pitch and 0 ±180° of roll. In most cases, a pair of 2D accelerometers are a better option than a 3D version because many 3D devices have degraded performance on the Z-axis.
Capacitive and Thermal
A 3D capacitive accelerometer contains a cantilevered beam, and calculates acceleration by measuring the force gravity exerts on it. Acceleration due to gravity causes bends the beam and makes it change position relative to two fixed electrodes. This changes the capacitance between the electrodes; the change is proportional to the acceleration.
A 2D thermal accelerometer uses a monolithic approach that combines the sensor and electronics onto an IC, which is hermetically sealed. The IC (above image) includes a heating element and a pair of thermopiles all suspended over a cavity etched into the chip’s surface. The thermopiles measure the movement gas molecules warmed by the heating element to detect acceleration. When there is acceleration, heated molecules move in the direction of acceleration, and with zero acceleration the heated gas is symmetrical dispersed around heater.
Capacitive sensors use a cantilevered beam with moving parts. They are inherently wideband transducers (>5 kHz), with a mechanical resonant frequency near 2 kHz for low-g devices used to detect and measure inclination. When vibration energy is greater than the capability of the sensor or near its resonant frequency, it can experience clipping or sensor resonance. In some cases, clipping or resonance may cause a large DC offset shift—particularly on the Z-axis, making it impossible for the sensor to recover the signal in high-vibration environments. This is an inherent disadvantage of capacitive accelerometers in high-vibration environments. To compensate for it, engineers can try a variety of mitigation techniques to isolate vibrations from the sensor. However, in some environments, the vibration can be too large to eliminate.
One mitigation approach to isolate the accelerometer from the vibration is to suspend the accelerometer on rubber bushings, or springs and dampers. Another approach is to use a less-sensitive device with stiffer cantilever beams, which will give the sensor a higher resonant frequency and the ability to withstand larger mechanical shocks and vibration. But these techniques add cost, sacrifice performance, and require a longer design times, which could affect time to market and product launch deadlines. It also makes the sensor less reliable.
Even if vibrations can be lowered to a level that lets the capacitive sensor function properly, there is still the issue of aliasing to address. The sensor’s wide bandwidth will let higher frequency vibrations alias (downconvert) into measurements and degrade them. To get around this, engineers can employ heavy oversampling and more microprocessor horsepower—i.e., a more expensive processor to apply heavy DSP filtering to remove the out-of- band energy and prevent the aliasing. Mechanical shocks and impacts also affect aliasing and resonance in capacitive sensors. A mechanical shock—a high-magnitude, short-duration event—contains a wide range of frequencies. If the shock packs enough energy at or near the sensor’s resonance frequency, it will send the sensor into resonance, rendering it unable to make accurate measurements.
Mechanical shock, if it is strong enough, causes stiction in capacitive sensors. Stiction makes cantilevered beams stick to each other if they touch; it is a property of extremely small structures. If the beams get stuck together, the sensor’s output remains constant. Extreme shock can move beams past their mechanical stops and damage them. Mechanical shocks can also cause capacitive sensors to become uncalibrated by changing the device’s zero g bias or its sensitivity.
In comparison, thermal sensors use the movement of heat to measure acceleration and this acts like a front-end low-pass filter, mitigating vibrations and shocks. For example, MEMSIC’s MXC6244AU features excellent low-pass filter response (10 Hz/-3 db) and integrated 2nd order filtering to further attenuate out-of-band vibration, thanks to the thermal sensor. This results in a more accurate inclination measurement that can be determined without consuming computing power from the processor.
Capacitance-based inclination sensors for high-vibration environments require a two-die manufacturing approach to build a MEMS device—one for the sensing element and another for the ASIC. In most cases, the sensing-element dies is much larger than the ASIC die.
Tale of the Tape
A thermal sensor combines the sensor and electronics on a single IC, resulting in a smaller and more reliable device. The integration of sensor and electronics also translates into less costly manufacturing and simpler assembly.
Together, the monolithic design and lack of cantilevers of thermal sensors translates into high shock survivability and best-in-class reliability (repeatability). With no moving parts, the thermal MEMS device exhibits no variance due to shock and vibration that could affect any stored calibration.
Thermal MEMS devices also exhibit no measureable resonance, delivering immunity to vibration; no temperature hysteresis, excellent zero-g offset stability; and 50,000-g shock tolerance, all of which contribute to the devices’ reliability.
Capacitive inclination sensors, on the other hand, use less power—a major advantage for power critical applications such as battery-operated devices. In fact, they consume an order of magnitude less power than thermal sensors that typically use 3 milliwatts of power. That’s because it takes current to generate the heat in the thermal sensors.
Capacitive MEMS accelerometers also have higher bandwidths—an advantage for higher frequency applications, typically above 100 Hz—for high-g sensing applications such as crash detection for airbag deployment. In these cases, the sensor needs the wide bandwidth and high-g capability offered by the capacitive transducer.
James Fennelly, Product Manager
Looking for parts? Go to SourceESB.