4. Sensor and instrumentation fundamentals

Thoracic injury metrics are only as trustworthy as the instrumentation that produces them. In the THOR Hybrid III and WorldSID anthropomorphic test devices (ATDs), three sensor families dominate the thoracic measurement chain: chest deflection potentiometers, rib accelerometers, and load cells integrated into the spine and rib structure. Each family answers a different physical question, and each imposes its own mounting, signal-conditioning, and coordinate-system conventions. Understanding how these sensors are physically arranged, how their outputs are scaled, and how they are combined into injury criteria is the prerequisite for any credible sled, crash, or component test.

What each sensor family actually measures

The chest deflection potentiometer measures the relative displacement between the sternum and the spine. In the THOR Hybrid III, a rotary or linear potentiometer is anchored to the spine box and connected to the sternum via a rod-and-linkage arrangement; as the sternum moves posteriorly under belt loading, the linkage rotates or translates the potentiometer wiper, producing a voltage proportional to deflection. The WorldSID uses a comparable but geometrically different arrangement, and in both dummies the measurement is a relative deflection, not an absolute chest compression in space. This distinction matters: the reported value depends on the reference frame defined by the spine box mounting, so any re-rigging of the linkage changes the calibration constant.

Rib accelerometers are uniaxial or triaxial piezoelectric or piezoresistive devices bonded or bolted to individual ribs, typically at the lateral or anterior rib surface. They capture the local acceleration history of the rib cage during impact, which is then integrated to velocity and displacement for rib fracture and viscous-criterion analysis. Because ribs deform and rotate, accelerometer orientation must be documented precisely; a 10° mounting error on a lateral rib accelerometer can rotate a significant fraction of the measured signal into an unintended axis.

Load cells in the thoracic region are most commonly integrated at the spine box and at the rib-to-spine attachment points. They measure the reaction forces transmitted through the thoracic structure, and in some configurations the shoulder and lap belt loads are captured by separate load cells in the belt path. The spine load cell provides the axial and shear force components needed for chest compression force and for cross-checking deflection-derived metrics.

Signal chain and integration architecture

The three families do not operate independently. A typical thoracic instrumentation chain runs from sensor, through an in-dummy amplifier or pass-through, to the data acquisition system (DAS), where channels are sampled synchronously and filtered according to the applicable standard (commonly SAE J211 for crash instrumentation). The figure below shows the integration architecture and the physical reference points that define each measurement.

Scaling, calibration, and the numbers that matter

Every channel must be traceable to a physical unit before it enters an injury criterion. For the deflection potentiometer, the calibration constant is established on a bench fixture by displacing the sternum linkage through a known travel—commonly 0 to 75 mm for THOR—and recording the output at each step. A representative calibration yields a slope of about 0.025 V/mm with an offset near 2.5 V at zero deflection; the DAS then applies deflection_mm = (V_measured − V_zero) / 0.025. If the zero reference drifts between tests, the entire deflection trace shifts, and the peak value used for the criterion becomes biased.

Rib accelerometers are calibrated against a reference shaker at a known acceleration, typically 10 g at 100 Hz, and the sensitivity is recorded in mV/g. A nominal 10 mV/g device with a ±5 V input range saturates at 500 g, which is adequate for most thoracic impacts but marginal in high-severity sled tests; checking the expected peak against the channel range before the test prevents clipped traces that silently invalidate the viscous criterion.

Load cells are calibrated with dead weights or a proving ring, and the sensitivity is expressed in mV/V at rated load. Because the spine load cell sits in the load path, its mounting preload must be recorded: an over-torqued bolt introduces an offset that appears as a constant force bias in every subsequent test.

From channels to thoracic injury metrics

The integrated measurement set feeds three principal thoracic criteria. Chest deflection comes directly from the potentiometer and is compared against injury assessment reference values (IARVs) that differ between THOR and WorldSID and between belt and airbag loading. Viscous criterion is computed as the product of chest compression velocity and compression, normalized by chest depth: VC = (dD/dt) × (D/C), where D is deflection, C is the initial chest depth, and dD/dt is obtained by differentiating the filtered deflection signal. Because differentiation amplifies high-frequency noise, the deflection channel must be filtered at the appropriate channel frequency class (typically CFC 180 for deflection) before VC is calculated. Chest compression force is taken from the spine load cell and, in some protocols, combined with belt load cells to estimate the force actually delivered to the thorax.

A practical integration check is to compare the deflection-derived compression against the force-derived compression at the same instant. If the two diverge sharply—for example, deflection continues to rise while spine force plateaus—the likely cause is a slipping potentiometer linkage or a load cell mounted outside the primary load path. This cross-check catches rigging errors that a single-channel review would miss.

Common integration pitfalls

• Reference-frame mismatch: rib accelerometers mounted with the sensitive axis reversed produce inverted traces that still look plausible until integrated. Document the axis convention on the mounting drawing and verify with a tap test in the expected direction.
• Time-zero misalignment: deflection and acceleration channels triggered on different events shift the VC peak. Use a common trigger and verify alignment with a pre-test tap on a shared reference.
• Filter mismatch: applying CFC 1000 to acceleration but CFC 180 to deflection and then combining them without resampling introduces phase error. Filter and resample all channels to a common rate before computing composite metrics.
• Linkage preload: an over-tensioned deflection linkage reads a nonzero offset at rest and under-reports peak travel. Set the linkage to the specified free-length and confirm zero output before closing the chest.

Treat the thoracic instrumentation as a single calibrated system rather than three independent channels. When the deflection, acceleration, and force traces are mutually consistent—peak timing aligned, signs matching the documented reference frame, and magnitudes within the expected envelope for the test severity—the resulting injury metrics can be defended. When they are not, the discrepancy itself is the most useful diagnostic signal available.