3. Accelerometer and pressure sensor principles for crash detection

Crash detection begins with physical sensing. Before any deployment decision can be made, the restraint control module (RCM) must convert mechanical motion and cabin pressure changes into trustworthy electrical signals. This chapter examines the four sensing elements that dominate modern airbag systems — MEMS accelerometers, remote (satellite) accelerometers, pressure satellites, and the signal-conditioning chain that connects them to the microcontroller.

3.1 MEMS Accelerometers: The Primary Crash Sensor

A MEMS accelerometer is a micromachined capacitive structure. A proof mass is suspended by silicon springs above a set of fixed electrodes. Under acceleration, the proof mass displaces, changing the differential capacitance between the moving and fixed plates. On-chip circuitry converts this capacitance change into an analog voltage or a digital word proportional to acceleration.

Two parameters dominate sensor selection. The full-scale range must be large enough to avoid saturation during a severe crash — typical frontal sensors use ±50 g to ±250 g, while side-impact sensors may reach ±500 g because side events are shorter and more violent. The sensitivity (for example, 20 mV/g or 4 LSB/g) determines the smallest resolvable acceleration, which in turn sets the detection threshold floor.

Bandwidth matters as much as range. A frontal crash pulse contains energy up to roughly 300–500 Hz; a side impact can excite content above 1 kHz. The sensor's mechanical resonance must sit well above the band of interest, and the anti-alias filter must attenuate everything above the Nyquist frequency before sampling. A common design samples at 2 kHz with an analog bandwidth of about 400 Hz, giving comfortable margin.

Orientation is a frequent source of field error. The sensing axis must be defined explicitly. In this course we adopt the convention that positive X points forward along the vehicle longitudinal axis, positive Y points to the driver's right, and positive Z points downward. A frontal impact therefore produces a large negative X acceleration (the vehicle decelerates while the proof mass continues forward). Every threshold, sign, and diagram in this course follows this convention; if a sensor is mounted rotated, the sign inversion must be corrected in software or by mounting keying.

3.2 Satellite Sensors and Pressure Satellites

Placing all sensing inside the RCM is inadequate for side impacts, where the crush zone is far from the center tunnel. Satellite sensors are small remote accelerometers mounted in the B-pillar, door, or sill. They communicate with the RCM over a two-wire current-loop or PSI5 bus, typically sending acceleration data at 1–2 kHz. Because they sit close to the impact, they see the crash earlier and with less structural attenuation than a central sensor.

Pressure satellites exploit a different physical principle. A side impact compresses the door cavity, producing a rapid pressure rise inside the sealed door volume. A piezoresistive or capacitive pressure element measures this rise. The key advantage is speed: pressure propagates through the cavity at the speed of sound, so the sensor responds within a few milliseconds — often faster than the door structure deforms enough to register on an accelerometer. Pressure satellites are therefore common in door-mounted side airbag systems.

Pressure sensors are also vulnerable to false triggers. Slamming a door, a minor parking impact, or a pressure change from altitude and temperature can all produce signals that resemble a crash. This is why pressure satellites are almost never used alone; their output is fused with satellite accelerometer data and validated against a plausibility window.

3.3 Signal Conditioning

Raw sensor output is not directly usable by the deployment algorithm. The conditioning chain performs four functions:

• Amplification and level shifting — the small differential signal is amplified and biased to mid-supply so that both positive and negative accelerations fit within the ADC input range.
• Anti-alias filtering — a low-pass filter removes energy above the Nyquist frequency before sampling. Without it, high-frequency crash noise folds back into the baseband and corrupts the pulse shape.
• Analog-to-digital conversion — typically 10 to 12 bits at 1–2 kHz per channel, with the RCM sampling all channels synchronously so that cross-axis timing is preserved.
• Self-test and plausibility — a built-in self-test (BST) applies an electrostatic force to the proof mass and verifies the output responds within an expected window. A failed BST at key-on typically sets a diagnostic trouble code and may inhibit deployment.

The following figure shows how the sensing elements, conditioning stages, and the RCM's decision logic connect, including the direction of signal flow and the point at which each channel is validated.

3.4 Practical Example: Verifying a Satellite Channel

Consider a bench check of a B-pillar satellite accelerometer before vehicle integration. The goal is to confirm that the sensor, its wiring, and the RCM's conditioning path produce a correctly scaled and correctly signed reading.

Setup. The satellite is mounted on a shaker fixture with its sensing axis aligned to the shaker's motion axis. The RCM is powered at 12 V nominal, and the PSI5 bus is monitored with an oscilloscope. The RCM is connected to a CAN tool that reports the decoded acceleration channel.

Procedure. Apply a 10 g peak, 100 Hz sinusoidal excitation along the positive sensing axis. Observe the decoded channel on the CAN tool and the raw PSI5 frame on the oscilloscope.

Expected result. The decoded channel should read approximately +10 g peak, with the PSI5 frame period matching the configured 1 kHz update rate. The waveform should be a clean sinusoid with no clipping and no visible aliasing artifacts.

Typical failure symptoms and diagnosis.

3.5 Key Takeaways

MEMS accelerometers provide the primary, high-bandwidth acceleration measurement inside the RCM. Satellite accelerometers extend coverage into the crush zone, while pressure satellites exploit door-cavity pressure rise for the fastest possible side-impact detection. All three depend on a conditioning chain that amplifies, filters, digitizes, and validates the signal before the deployment algorithm ever sees it. Understanding the axis convention, the sensitivity constant, and the self-test behavior is essential for both algorithm development and field diagnosis.