4. Instrumentation for linear impact testing

Linear impact testing produces valid, repeatable data only when the measurement chain is designed as a system. A sled, a pneumatic or servo-hydraulic accelerator, and a rigid reaction surface define the mechanical input; the instrumentation defines what you can actually observe. This section covers the five core transducer families used in linear impact work — accelerometers, load cells, pressure sensors, high-speed cameras, and the data acquisition (DAQ) system that ties them together — and explains how to select, mount, and configure each one so the recorded signals represent the physical event rather than the measurement artifacts.

Accelerometers

Piezoresistive (MEMS) accelerometers are the default choice for sled and component impact testing because they respond to DC, which lets you capture the full velocity change during a pulse. Piezoelectric (charge-mode) accelerometers offer higher sensitivity and wider bandwidth but cannot measure sustained acceleration, so they are used mainly for high-frequency shock and structural response measurements.

Selection starts with the expected peak level. A frontal sled pulse might reach 40–60 g; a component-level impact can exceed 500 g. Choose a full-scale range at least 1.5× the expected peak to avoid clipping, but not so high that resolution suffers. Bandwidth must cover the highest frequency of interest — for a 30 ms half-sine pulse, the meaningful content is below roughly 1 kHz, but mounting resonance and structural ringing can extend to 10 kHz, so a 10–20 kHz sensor is common.

Mounting is where most errors originate. A stud-mounted accelerometer on a machined flat surface has a resonance above 30 kHz. Adhesive mounting drops that to 5–10 kHz. A magnetic base can fall below 2 kHz and will ring within the measurement band. For linear impact work, always stud-mount or use a rigid adapter block, and torque to the manufacturer's specification. Apply a thin film of silicone grease or couplant to improve contact stiffness.

Orientation matters. Define the vehicle or sled coordinate system before mounting: typically +X forward, +Y lateral, +Z vertical. Record the sign convention in the test log. A sensor mounted backwards will produce a pulse with inverted polarity, which is easy to miss when reviewing data later.

Load Cells

Load cells measure the force transmitted through a structure. In linear impact testing they appear in two roles: as the reaction load cell in a crash wall or barrier, and as inline load cells in seat belt, airbag module, or steering column fixtures.

Strain-gauge load cells are standard. Their full-scale capacity should be at least 2× the expected peak force to survive occasional overloads without permanent zero shift. For a barrier load cell measuring a full-vehicle frontal impact, capacities of 500 kN to 2 MN are typical. For an airbag module deployment fixture, 5–50 kN is more appropriate.

Mounting stiffness is critical. A load cell installed with soft bushings or long bolts will measure a filtered version of the true force because the compliance acts as a mechanical low-pass filter. Use rigid, preloaded fixtures and verify that the cell's first structural mode is well above the frequency content of the impact.

Calibration should be performed in the installed orientation and, where possible, with the actual fixture attached. A load cell calibrated on a universal testing machine and then bolted into a compliant bracket will not read correctly.

Pressure Sensors

Pressure measurement in linear impact testing serves two purposes: monitoring gas generator output in airbag modules, and measuring pressure inside inflatable structures or closed volumes during deployment.

Piezoresistive pressure transducers with a range of 0–10 MPa (0–100 bar) cover most airbag inflator applications. For lower-pressure work such as seat cushion or pedestrian protection airbag testing, 0–1 MPa sensors are more suitable. Response time should be under 1 ms to capture the rapid pressure rise during deployment.

Mounting the sensor close to the measurement point minimizes acoustic delay and tubing resonance. If a remote mount is unavoidable, keep the connecting tube short and rigid, and account for the propagation delay in your data analysis. For inflator can pressure, a flush-mounted sensor in a specially machined boss is preferred.

High-Speed Cameras

High-speed cameras provide the spatial and kinematic context that point sensors cannot. They capture deployment sequence, fabric unfolding, tether engagement, and occupant kinematics. Typical frame rates range from 1,000 fps for full-vehicle events to 10,000–20,000 fps for airbag deployment details.

Resolution and frame rate trade off directly. At 1,000 fps, a 1280×1024 sensor is common. At 10,000 fps, resolution may drop to 512×512 or lower depending on the camera's internal memory and readout architecture. Calculate the spatial resolution you need: if the field of view is 1 m wide and you need 1 mm resolution, you need at least 1000 pixels across the frame.

Lighting must be synchronized with the camera. High-speed capture requires intense, continuous light — typically 2–10 kW of tungsten or LED illumination — because exposure times are on the order of microseconds. Strobe or pulsed LEDs can reduce heat but must be synchronized to the frame clock.

Camera synchronization with the DAQ is essential. Use a common trigger signal, usually a contact closure or TTL pulse from the sled release mechanism, to start both the DAQ and the cameras. Record the trigger time in both systems so you can align the data during post-processing.

Data Acquisition Setup

The DAQ system must sample all channels simultaneously at a rate high enough to avoid aliasing. For impact testing, a minimum of 10× the highest frequency of interest is a practical rule. If your accelerometer bandwidth is 10 kHz, sample at 100 kHz per channel. Many labs use 200 kHz for critical channels.

Anti-alias filters must be enabled and set to less than half the sample rate. A 10 kHz anti-alias filter with a 100 kHz sample rate is a common configuration. Do not rely on software filtering alone; the hardware filter prevents aliasing before digitization.

Use a common time base for all channels. If the DAQ has multiple cards, ensure they share a clock or are synchronized via a backplane trigger. Channel-to-channel skew should be less than one sample period.

Bridge completion and excitation for strain-gauge load cells and pressure sensors must match the transducer specification. Quarter-bridge, half-bridge, and full-bridge configurations require different completion modules. Verify excitation voltage and check for offset before each test.

Sample rate:        200 kHz per channel
Anti-alias filter:  20 kHz (hardware)
Resolution:         16-bit minimum, 24-bit preferred
Trigger:            TTL rising edge from sled release
Pre-trigger:        10 ms (to capture baseline)
Record length:      200 ms (sufficient for most sled pulses)
Channels:           8 accelerometers, 2 load cells, 2 pressure, 1 trigger

System Integration and Verification

Before a live test, perform an end-to-end check. Tap each accelerometer gently and confirm the correct channel responds. Apply a known shunt calibration to each strain-gauge channel and verify the expected voltage change. Fire a test flash on the high-speed cameras and confirm the frame count and trigger alignment. Check that all channels share the same time base by injecting a common pulse and verifying zero skew.

During the test, monitor the data in real time if possible. Look for clipping, excessive noise, or dropped channels. After the test, review the raw data before any filtering. A clean accelerometer trace should show a smooth pulse with a clear rise and decay; ringing or spikes suggest mounting problems or electrical interference.

Key Takeaways

• Choose accelerometer range at 1.5× expected peak and bandwidth at 10× the highest frequency of interest; stud-mount for linear impact work.
• Load cells must be calibrated in their installed orientation and fixture; soft mounting acts as an unwanted mechanical filter.
• Pressure sensors should be flush-mounted or connected with short rigid tubing to preserve rise time.
• High-speed camera frame rate and resolution trade off; calculate spatial resolution from field of view and required detail.
• DAQ sample rate should be at least 10× the highest frequency of interest, with hardware anti-alias filtering and a common time base across all channels.
• Always verify the full chain before a live test: tap test, shunt calibration, camera sync, and time-base skew check.