PM2.5 sensor drift after thermal soak
After a 4-hour 50 °C soak, unit #A3 shows +7–9% bias. Offset stable after cooldown. Next: repeat with airflow.
Lab Notebook
Selected entries. Everything is timestamped. Raw logs live in the repository.
uSD card endurance on SBC logging
F2FS with weekly trim doubled survival vs ext4 in identical duty cycles. Power-loss corruption remains the killer.
Restore drill: 3-2-1 check
Timed full-restore to bare metal: 2h14m. Bottleneck was decrypt-verify, not transfer. Notes updated.