Unplanned Downtime Is a Symptom: How to Find the Causes Your Logbook Hides

Every plant has a breakdown register. In most of them, the entries look like this: ‘Line 2 stopped — belt problem — 2 hours.’ It tells you the line stopped and roughly for how long. It does not tell you whether this is the fourth belt problem this month, whether it happens after a product changeover, or whether the spare was in stock.

Unplanned downtime is a symptom. The causes sit in maintenance practices, spare-part planning, operating conditions and sometimes in the design of the line. This article sets out a simple method to get from symptom to cause.

Three questions every stoppage record should answer

To be useful, each stoppage needs to answer: what stopped (machine and component), why it stopped (failure mode), and what allowed it to happen (root cause). Most registers stop at the first question.

Short stops: the downtime nobody writes down

A two-minute jam does not make it into the register. Twenty of them a shift do. Micro-stops are often the largest hidden loss on packing and assembly lines, and they only become visible when stoppage data comes directly from the machine — a PLC counter, a run signal or a simple sensor.

If you cannot connect machines yet, ask operators to tally short stops by reason for one week. It is crude, but it usually changes the priority list.

Waiting time is a separate problem

Many long breakdowns are short repairs with long waits: for the technician, for the spare part, for the purchase approval. Record repair time and waiting time separately. If waiting dominates, the fix is in spare-part planning and escalation, not in maintenance technique.

Building a Pareto that actually guides action

Once stoppages are categorised for four to six weeks, rank causes by total lost time and by frequency. The top three causes typically account for a large share of lost hours. Those are your projects.

Look at the Pareto by machine and by failure mode. A machine that appears at the top because of many different causes needs a condition review. A failure mode that appears across machines — for example lubrication-related wear — points to a practice problem.

MTBF and MTTR without the jargon

Mean time between failures (MTBF) tells you how often a machine fails. Mean time to repair (MTTR) tells you how long it takes to get it running again. Track both per critical machine. Rising MTBF means reliability is improving; falling MTTR means your response is getting faster. Both are only meaningful if every stoppage is recorded consistently.

A weekly downtime review that works

• 30 minutes, same day each week, with production, maintenance and stores.
• Look at the top three causes by lost production for the week.
• For each, agree one action, one owner and one date.
• Check last week’s actions first. If they are not done, find out why before adding more.

Classify stops into six buckets first

Before you go anywhere near root cause, every stoppage should land in one of a small number of buckets. Six is usually enough: breakdown, changeover, material shortage, utility failure (power, air, steam, water), quality hold and planned activity such as cleaning or trials. When buckets are fixed, two supervisors on different shifts classify the same event the same way, and a month later the data can be added up.

Resist the temptation to create forty categories on day one. Long lists get ignored and everything ends up under ‘others’. Start small, look at what lands in ‘others’ after two weeks, and only then add a category if it keeps appearing.

Running a five-why without blaming the operator

Five-why analysis works when the team asks ‘why did the process allow this?’ rather than ‘who did this?’. Take a recurring bearing failure. Why did it fail? It overheated. Why? It ran short of grease. Why? The weekly greasing was skipped. Why? The technician was pulled into a breakdown elsewhere. Why? There is one technician per shift and no rule about which job takes priority.

The root cause is not ‘technician forgot’. It is a staffing and priority rule. Fixing that prevents the next ten failures; blaming the technician prevents none.

Spares: the quiet multiplier of downtime

When repair takes forty minutes but the line is down for six hours, the gap is almost always a spare part. Critical spares — those whose absence stops the bottleneck — deserve their own list, minimum stock levels and a monthly check. Link each critical spare to the machines that use it, so that when a machine is flagged, stores already knows what to keep ready.

Plants that make this connection often find that a small investment in the right spares removes a large share of their longest stoppages.

When to add sensors and prediction

Condition monitoring — vibration, temperature, current — pays off on critical machines with known failure modes, such as bearings on expellers, pumps and fans. It is far more effective once you have a clean breakdown history, because the history tells you which signals matter and what ‘normal’ looks like.

This is the order we recommend: consistent stoppage records first, machine signals on the bottleneck second, predictive alerts third.