Products · Predictive Maintenance

Fix it before it breaks

The sensors are off-the-shelf and the PLC is already yours. The difference is the algorithm in between — the piece Gangdolf built — that learns each machine's own baseline and failure modes, then loads a finished monitor straight onto the PLC.

The core

One algorithm, from raw data to a running PLC monitor

Everything around it is commodity — the sensors, the wiring, the PLC. The algorithm in the middle is the part Gangdolf built, and it's what turns raw readings no one can read into thresholds a machine can act on.

Raw · IO-Link + REST API

2f a1 07 5c e3 9b 40 12
11 c8 4d 20 f6 3a 88 d1
9e 03 b7 6a 2c 55 0f e9
74 dd 19 8b 4f a0 63 27
c5 32 ea 96 71 0d bc 48
6f 90 23 fd 5e a7 14 db

Raw readings off the machine. No pattern a person could read off the wire.

the algorithm

Baseline + failure modes

Impacting Rising vibration Overheating

Learned for this exact machine — not a factory-default preset.

Deployed · your PLC

Watching around the clock, on-machine. No cloud, no PC, no subscription.

One repeatable algorithm — raw data in, a running PLC monitor out. No black box, no cloud subscription, and no hand-tuning machine by machine.

Works across every drive VFD · variable speedDOL · hard startSoft starter · ramped

Live view

See the algorithm catch a fault

Overall vibration tells you how hard a machine is shaking, not what's wrong with it. Here the level sits inside its learned baseline — until a sharp impact starts repeating on every axis at once. The algorithm reads the shape, not just the size: repeating peaks that dwarf the steady level flag it as impacting — an early bearing-or-gear signature — and point you at the spectrum check that confirms the root cause.

How it works

From raw data to a machine that watches itself

Four steps, the same every time. Sensors you can buy from any distributor, an algorithm that does the thinking, and the monitoring left running where your team already looks.

01

Collect

Smart IO-Link sensors read vibration and temperature straight off your machine. A short run of normal operation is pulled in over a REST API — no rewiring, no shutdown, nothing new bolted onto the plant.

02

Learn

The algorithm turns that raw block of data into a baseline — what healthy looks like for this exact machine — and a set of distinct failure modes, each with its own warning, alarm and trip level.

03

Deploy

The algorithm loads the finished monitor straight onto your PLC. Automatically, and the same repeatable way on every machine — no hand-tuning.

04

Run

From then on your PLC does the watching: on-machine, no cloud, no PC left running. It trends each failure mode, calls the service window before the breakdown, and can stop the machine before real damage.

What we monitor

Start where a breakdown hurts most

A pilot is a handful of IO-Link sensors on the one machine that keeps your maintenance manager up at night — and the failure modes the algorithm learns to catch on it.

Motors & gearboxes

Imbalance, misalignment and bearing wear each leave a fingerprint in vibration weeks before failure — and a bearing on its way out runs hot long before it seizes.

ImbalanceMisalignmentBearing wearBearing temp

Pumps & fans

Cavitation, a blocked line and a damaged impeller each have their own vibration signature. Caught early, it costs a seal kit instead of a whole pump.

CavitationImbalanceLooseness

Conveyors & bearings

Failing rollers, tension problems and dry bearings show up as rising vibration and heat on the lines that never get looked at — until they stop.

Roller wearBearing tempLooseness

Works with the AI assistant

Condition data feeds straight into the AI Maintenance Assistant's Connected tier — so your team can ask "why is P-201 trending up?" and get an answer grounded in the actual readings.

See the assistant
Gangdolf Automation · Adelaide SA

Start with one critical machine

Pick the machine whose failure costs you a shift. We'll instrument it, let the algorithm learn it, and load the monitor onto your PLC.

Plan a pilot