The bearing didn't fail without warning. Nobody was listening.

Rotating equipment almost always tells you it's dying — through vibration, heat and rising current — weeks before it stops the plant. addanode puts low-cost sensors on your critical motors, pumps and gearboxes and watches the trend 24/7, so maintenance acts at the start of a failure instead of the end. And we're honest about where condition monitoring pays back and where it doesn't.

Looking for the other kind of "asset"? This page answers is this machine healthy and how long will it last — vibration, temperature and current on rotating equipment. If your question is where is it, do we still have it, who has it — linen, stock, tools — that is RFID asset tracking.

What we watch

Three signals that predict most rotating-equipment failures.

Vibration

Imbalance, misalignment, bearing wear and looseness show up in the vibration signature long before failure.

Temperature

Bearing and winding temperature rise flags friction, overload and cooling problems.

Current

Motor current and power trends reveal load changes, mechanical drag and electrical faults.

The three signals above, placed in the order they change. Vibration moves first, then temperature, then current and power — by the time a fault is audible the decision about when to act has already been made for you. Order of appearance, not a schedule: how much warning each stage gives depends on the machine, its duty and how hard it is loaded.

The failures we actually find — and what each one asks you to do.

Most of what monitoring catches on rotating equipment comes from a short list, and each item on it points at a different job. The value is not the alert — it is knowing, before anyone opens a machine, whether this is an alignment job, a lubrication job, a foundation job or an electrical one.

Failure mode	What shows first	What it asks you to do
Misalignment	Radial vibration at 1× and 2× shaft speed, often with a coupling running warm	Laser-align cold, check for soft foot, then re-check at operating temperature — thermal growth moves the answer
Imbalance	A steady 1× peak that grows with speed, phase holding still	Look for build-up, erosion or a thrown balance weight first; balancing a fouled rotor only buys weeks
Bearing wear	High-frequency energy climbing well before overall vibration moves at all	Schedule the change at the next planned window — the trend gives you that choice, the overall level takes it away
Lubrication, too little or too much	Temperature rise with high-frequency noise, current unchanged	Re-grease on running hours rather than the calendar; over-greasing churns and heats as surely as starvation
Looseness	Several harmonics at once, and vibration much higher in one direction than the other	Torque the holding-down bolts and inspect the grout before anyone opens the machine
Belt or coupling wear	Peaks below shaft speed, or at belt-pass frequency	Re-tension and replace as a set — a new belt beside a worn one carries the load unevenly
Rotor or winding fault	Current sidebands and phase imbalance, while vibration still looks normal	Stop hunting mechanically. This one is found electrically, and it worsens under load rather than at idle
Cavitation or restricted suction	Broadband noise, a shift in current, and a head that no longer matches duty	Check level, strainer and suction line before the pump — the symptom is at the pump, the cause usually is not
Cooling loss	Winding temperature up while current and vibration stay where they were	Clean fins and filters, and confirm nothing has been stacked in the airflow path

Two of these are misdiagnosed far more often than the rest. Cavitation gets blamed on the pump, because that is where the noise is — but the fix is almost always upstream, in a level, a strainer or a suction line, and a rebuilt pump on an unchanged suction will cavitate again. Rotor and winding faults get chased mechanically, because the vibration still looks acceptable and nobody thinks to check current; that is exactly why the current trend earns its place beside the other two signals.

One South African detail worth adding to the list: several of these show up most clearly in the minutes after a restart, when a cold machine is loaded quickly. Alignment and looseness in particular reveal themselves in that window, so the readings around a return from an outage are worth keeping and comparing rather than filtering out as noise.

Done right, not just installed

Three, rather than one, because each answers a different question and together they turn an alarm into a diagnosis. Vibration says what is wrong — imbalance, misalignment and bearing wear each leave a different signature. Temperature says how fast: a rising trend is the machine converting a fault into heat, and the slope tells you whether you have weeks or days. Current says whether the machine changed or the work did — a pump drawing more power may be failing, or may simply be pushing against a blocked line, and only the current trend separates the two. One signal in isolation produces alerts that people learn to ignore; the three read together produce a short list of assets worth walking out to.

What makes any of them useful is the baseline, not the threshold. A motor is judged against its own normal, established over weeks of ordinary running, because a reading that would be alarming on one machine is unremarkable on the one beside it. That is also why the first month of a monitoring project produces no alarms and should not: it is buying the reference the rest of the programme is measured against.

We scope for payback — not for sensor count.

Most condition-monitoring projects fail not because the technology doesn't work, but because they're aimed at the wrong assets, drown people in alarms, or never tie a reading to a maintenance action. We start from your criticality list: which assets, if they stop, cost you the most — and only instrument where the maths works.

Why most vibration projects fail →

Precision industrial equipment under addanode condition monitoring — vibration, temperature and current

Frequently asked

The criticality list is easier to build than it sounds — three questions per asset. What does an hour of its downtime actually cost? Count lost production, scrapped work in progress and the overtime to catch up, not just the repair. How long until it runs again? An asset with a spare on the shelf and a fitter who has changed it before is a different risk from one with an imported lead time. Does its failure take anything else with it? A bearing that seizes and scores a shaft has converted a cheap job into a long one. Assets that score high on all three justify continuous monitoring; where the spare is on the shelf and production can route around the machine, a route-based check is the honest answer and costs less.

South African plants have a fourth variable that does not appear in most condition-monitoring literature: how often the machine is stopped and started. Scheduled outages put motors through inrush and thermal cycling that steady running never does, so on many sites the trend worth watching is not run hours but start counts and the temperature recovery after each restart. It is also the reason to ask any monitoring system what happens to its data during an outage — the readings either side of a restart are the ones you most want to compare, and a system that only records while the grid is up loses exactly those.

Condition-monitoring questions we answer first.

The right first question is never "how many sensors?" — it's "which assets actually justify monitoring?" Our guide to predictive vs preventive maintenance is the 20-minute way to answer it. For cold rooms, fridges and reefer moves, see the dedicated cold chain monitoring solution.

Start with critical, hard-to-replace rotating equipment whose failure stops production or safety — large motors, key pumps, fans, compressors and gearboxes. We rank by downtime cost and failure likelihood, then instrument the top of the list.

For trend-based early warning on most plant assets, yes. Wireless sensors are ideal for catching developing faults. For detailed diagnostics on the most critical machines, we combine them with higher-rate measurement where it's justified.

That's the most common failure mode. We baseline each asset, set thresholds to its normal behaviour, and route alerts to the right person — so an alert means "go look", not "ignore it".

Yes — condition data lives in addaNet alongside OEE, energy and process data, so maintenance and production see one picture.

Accurate enough that the trend can be trusted — one of the most common complaints about cheap monitoring hardware is temperature sensors reading many degrees off, which quietly poisons every alert built on them. We use industrial-grade, calibrated sensors, sanity-check readings against neighbouring points on the platform, and support recalibration in service — because a threshold alarm is only as good as the measurement underneath it.

Effectively, yes — most of our condition-monitoring projects run as a monthly subscription (monitoring-as-a-service) that covers the sensors, connectivity, platform and alerting, rather than a capital purchase of vibration equipment you then have to own, calibrate and staff. That keeps the entry cost close to what hiring instruments for a one-off survey would cost, but you get continuous trending instead of a snapshot.

The workhorse trio: a vibration sensor on the bearing housing (imbalance, misalignment, bearing wear), a temperature sensor at the same point (friction and cooling problems), and a current clamp on the supply (load changes, cavitation signatures, run-hours). Larger or critical machines add pressure and flow context. Start with the trio on your worst actors — it catches the majority of developing failures for the least money.

Stop the next unplanned breakdown.

Send us your critical-asset list. We'll tell you which ones are worth monitoring — and which aren't.

Book a Discovery Call WhatsApp us