Each node clamps to the structure and picks up the bearing through the frame, not through the air. The same measurement is taken the same way every time, whoever is on shift.
Earliest mechanical evidence of distress.
Confirms a defect big enough to move the frame.
What the fitter will hear when they get there.
Late on the curve; counted only alongside another signal.
A distressed bearing produces ultrasonic stress waves before the defect is large enough to move the frame.
Belts, drives, loading and wind put most of their energy at lower frequencies, so a small bearing signal stands out.
Ultrasonic signal fades quickly through steel, so a node hears mainly the structure around it. A flag points to a zone, not a long stretch of structure.
Ultrasonic can also move for reasons that go nowhere, such as a short-lived change in lubrication, load or water ingress. That is why it is never read alone.
Nodes share one cable that carries both power and data, so there are no batteries to replace. A controller at the end of each section powers the nodes, runs the measurements and keeps every raw recording locally. Results go over the mobile network to a server in Australia set up for each client, or a machine on your premises. There is no connection to site IT or OT. Enclosures are designed to meet IP67.
Development and testing: what was built, what it found, and what it doesn’t show yet.
Bearing fault simulator: how bearing faults appear in the envelope spectrum.
