2. Crash test sensor suite architecture

A crash test is only as trustworthy as the measurement chain that feeds it. Before any pulse is analyzed or any injury criterion is computed, the data acquisition system must capture the physical quantities that define the event: rigid-body deceleration, structural deformation, restraint forces, occupant kinematics, and contact loads. This chapter maps the sensor suite that performs that capture — accelerometers, load cells, strain gauges, and displacement sensors — and explains where each transducer is mounted on the vehicle and on the anthropomorphic test device (ATD, or "dummy"), and why that placement matters for signal validity.

Transducer families and what each one measures

Each sensor type answers a different physical question, and mixing them up is a common source of invalid data. The table below summarizes the four families used in a typical frontal or side impact test.

The bandwidth column is not decorative. SAE J211 defines channel filter classes (CFC 60, 180, 600, 1000, 1650) that must match the physical phenomenon being measured. A chest accelerometer is normally filtered at CFC 180, while a head accelerometer uses CFC 1000 or 1650. Applying the wrong filter class during post-processing silently distorts the injury criteria — a topic developed in later chapters, but the sensor selection here already constrains it.

Vehicle-mounted instrumentation

Vehicle sensors establish the crash pulse and the structural response. The crash pulse is typically captured by a triaxial accelerometer rigidly bolted to the tunnel or the rear seat cross-member near the vehicle center of gravity, because that location best represents the rigid-body deceleration of the occupant compartment. Additional uniaxial accelerometers on the B-pillar base, the sill, and the engine block reveal how the pulse propagates and where the structure deforms non-rigidly.

Load cells on the seat belt system measure the restraint force path: one at the retractor or anchor, one at the buckle, and often one at the outboard anchor. These channels tell you whether the belt loaded as designed or whether it slipped, locked late, or routed over a hard point. Steering column load cells and displacement sensors quantify column collapse and driver airbag reaction. Strain gauges on the floor pan and rocker panel capture the bending and buckling that precede intrusion.

ATD instrumentation and placement

The dummy is a calibrated measurement platform in its own right. A Hybrid III 50th percentile male, the most common frontal ATD, carries accelerometers in the head (triaxial at the head CG), the upper and lower chest (uniaxial or triaxial), and the pelvis. Load cells measure neck axial force and moment, chest deflection via a chest potentiometer or IR-TRACC, femur axial force, and tibia force and moment. The head accelerometers feed HIC; the chest channels feed chest deflection and the chest acceleration criterion; the femur and tibia channels feed lower-extremity injury metrics.

Placement discipline on the ATD is stricter than on the vehicle. The head accelerometer must sit at the head center of gravity with its sensitive axis aligned to the SAE J211 sign convention: positive X forward, positive Y to the dummy's left, positive Z downward. A sensor rotated 90° in its mount will produce a plausible-looking but axis-swapped trace, and the error may only surface when the injury criterion is compared against the certification corridor. Every ATD channel should be verified against its mounting drawing before the test, not after.

How the suite fits together

The diagram below shows the signal chain from transducer to recorded channel, and how the four sensor families distribute across the vehicle and the ATD. Note that all channels converge on a single data acquisition system with a common time base — synchronization across channels is what makes cross-correlation between, say, belt force and chest deflection meaningful.

Placement errors that invalidate signals

Most bad channels in a crash test are not failed sensors — they are correctly functioning sensors in the wrong place or with the wrong orientation. Three failure patterns recur:

• Axis misalignment. A head accelerometer rotated in its mount swaps X and Z. The trace looks smooth and plausible, but HIC computed from it is wrong. Verify orientation against the mounting drawing and, where possible, a static tilt check before the test.
• Compliant mounting. A load cell bolted through a soft bushing, or an accelerometer on a thin bracket, records the mount's dynamics rather than the structure's. The symptom is a resonance peak in the 200–600 Hz range that does not appear on adjacent channels.
• Range saturation. A ±500 g accelerometer on the engine block during a 60 km/h frontal impact may clip. Clipping appears as a flat-topped trace; the peak value is a lower bound, not a measurement. Select range from the expected pulse plus margin, not from habit.

Pre-test verification checklist

The following checklist is a practical deliverable for the instrumentation phase. It is written for a frontal impact test with a Hybrid III 50th percentile driver, but the structure generalizes.

Running this checklist takes perhaps twenty minutes and prevents the most expensive failure mode in crash testing: a completed, costly test whose data cannot be used because one critical channel was misoriented or saturated. The sensor suite architecture described here — vehicle transducers for the pulse and structural response, ATD transducers for occupant loading, all converging on a synchronized acquisition system — is the foundation on which every subsequent validation and diagnostic step depends.