A parcel logger that samples only occasionally can report a quiet journey even when a short handling event occurred between samples. Increasing the average sampling rate is one response, but it may spend energy recording hours of uneventful transport. A more useful design separates event detection, waveform capture, and the summary eventually uploaded to the user.
Define what an event record must answer
Decide whether the product needs to identify movement, estimate event severity, or retain a waveform for later analysis. These are different objectives. A threshold crossing can establish that the detector reacted; it does not by itself provide a calibrated shock measurement. Required bandwidth, sensor range, mounting stiffness, and sampling rate follow from the physical event you intend to characterize.
The ADXL362 documentation describes motion-related interrupt and FIFO features that can support an event-driven architecture. Treat that as an implementation example, not evidence that this particular sensor covers every parcel shock. Check the expected acceleration spectrum and clipping risk before choosing a part.
Preserve the beginning of the event
A processor that wakes after an interrupt cannot recover measurements that were never captured. A continuously operating sensor buffer can retain earlier samples while the host sleeps. Allocate a pretrigger interval deliberately, and specify what happens if a second event arrives while the first is being copied. The FIFO service-time budget is useful for checking whether the processor can drain data before overwrite.
As an illustrative storage calculation, three signed 16-bit axes require six payload bytes per sample. A 400-sample record therefore needs 2,400 bytes before timestamps, headers, checksums, and alignment. If the logger stores many events locally, use the complete record size and realistic event frequency rather than the sensor payload alone when selecting memory.
Avoid an endless wake cycle
Movement during normal vehicle travel may keep retriggering a detector. Define an activity episode, a maximum capture duration, and a rearming condition. Preserve a counter when events are suppressed or merged so the report does not imply that nothing happened. A low battery should change recording policy explicitly rather than silently shortening every waveform.
Review the accelerometer output-rate and filtering relationship before lowering the sample rate to save energy. A smaller data file does not prove that the remaining signal represents the same physical bandwidth. Freeze the sensor configuration with each stored record so later analysis knows which filtering and range were used.
Validate the complete package
Test the assembled logger in its intended mounting, including orientation changes and the enclosure attachment. Compare known scripted movements with captured timestamps and waveforms, then repeat with a nearly full buffer and a depleted but permitted battery. Check recovery after an interrupted memory write. The useful outcome is a defensible record of what was observed and what could not be retained. Related measurement and power checks are available in the technical resource library.
Move from reading to a build-ready project
Use the fabrication and assembly checklists to prepare drawings, production files and test requirements.
Open the DFM checklist