Failing electrical equipment crackles at 40 kHz weeks before it burns. OhmWatch puts always-listening sensor nodes on your motors, pumps and panels — and explains every risk score in language an electrician can verify with their own meter.
Every electrical failure announces itself — in order, on a schedule. Most monitoring joins the story at the last chapter.
Partial discharge, arcing and tracking chirp at ~40 kHz — inaudible, invisible, and unambiguous.
The loose lug or failing bearing starts to heat. IR arrays see it glow long before smoke.
Draw creeps up, RPM sags, vibration signatures shift as damage compounds.
Where calendar-based maintenance and breaker trips finally find out.
One ESP32-class gateway and five sensing channels cover an asset: ultrasound (the earliest warning), IR thermal array, split-core current clamp (no rewiring — clips over the feeder), vibration, and contact temperature.
No machine-controller integration, no customer Wi-Fi dependency — nodes ship with cellular eSIM and speak MQTT to the platform.
Works on 40-year-old equipment. Especially on 40-year-old equipment.
Risk climbs through four bands — HEALTHY, WATCH, WARNING, CRITICAL — with hysteresis so statuses don't flap at the boundary. And every assessment says why, in sentences a technician can walk out and verify.
No black-box scores. If OhmWatch can't explain it, it doesn't alert on it.
See it live →One prevented panel fire, seized pump, or unplanned line-down pays for instrumenting the entire building. That's the wedge: continuous monitoring at a price the plants everyone else ignores can actually adopt.