Breakdown losses are the production lost when equipment stops unexpectedly and needs a repair to run again, the first of the Six Big Losses in OEE. They are unplanned, discrete, and long enough to need a maintenance response, which sets them apart from the brief minor stops an operator clears alone.
Breakdown losses and minor stops both drag OEE down, but they are different animals and they need different fixes. Treat a chronic string of minor stops like a breakdown and you throw a work order at a problem a design change would solve. Treat a breakdown like a minor stop and you clear it and move on while the real failure waits to happen again. This post pulls the two apart and shows the countermeasure each one actually needs.
What are breakdown losses in OEE?
Breakdown losses are unplanned equipment stops that require a repair, a work order, a technician, a part, before the machine can produce again. In the Six Big Losses framework they are loss number one, "equipment failure," and they hit the Availability side of the OEE calculation. Every minute the machine is down and every unit it should have made in that time is a breakdown loss.
The defining features are that the stop is unplanned and that clearing it takes more than an operator restart. A seal blows, a motor trips on overload, a gearbox seizes, the line is down until someone with tools and a part restores it. That is the loss OEE is built to expose, and per industry references like OEE.com's Six Big Losses it is usually the most visible and the most argued-over of the six.
How are breakdown losses different from minor stops?
Duration and response separate them. A breakdown is a long, unplanned stop that needs a repair; a minor stop is a brief pause, often under five minutes, that the operator clears without tools. Both cost output, but one is an Availability loss and the other is a Performance loss.
The common convention is a time threshold: stops shorter than about five minutes get logged as minor stops, longer ones as breakdowns. That line is a convention, not a law of physics, and each plant should set its own and hold to it. What matters is consistency, if the same jam is a minor stop on days and a breakdown on nights, your data cannot tell you where to spend.
| Breakdown loss | Minor stop | |
|---|---|---|
| OEE bucket | Availability (equipment failure) | Performance (idling and minor stops) |
| Typical duration | Minutes to hours | Seconds to a few minutes |
| Who clears it | Maintenance, with a part or tool | The operator, no tools |
| How it logs | Discrete event with a reason code | Often uncounted; hides as "slow running" |
| Countermeasure | Reliability: root cause, PM, redesign | Stability: cleaning, tolerances, poka-yoke |
Why do the two losses need different countermeasures?
A breakdown is a reliability problem; a run of minor stops is a stability problem. Reliability work asks why a part failed and keeps it from failing again. Stability work asks why the process keeps hiccuping and takes the wobble out. Aim the wrong tool and you spend money without moving the number.
Breakdowns respond to root cause analysis condition monitoring, and a real preventive-maintenance program, the core of total productive maintenance. You track them with MTBF (how often failures happen) and MTTR (how long each takes to fix). Minor stops respond to cleaning, tighter tolerances, better guarding, and mistake-proofing, the small-scale, operator-owned work of autonomous maintenance. One is a wrench; the other is a design tweak.
What causes most breakdown losses?
Most breakdowns trace back to a short list of preventable mechanisms, not bad luck. Lubrication problems, misalignment, contamination, and running parts past their wear life account for a large share of unplanned failures on rotating equipment. None of them are mysteries; all of them leave warning signs before the machine quits.
That is the uncomfortable truth behind breakdown loss: a lot of it is deferred basic care catching up with you. A bearing starved of grease, a belt left at the wrong tension, a filter never changed, each one shortens the life of the part and turns a scheduled five-minute task into an unscheduled four-hour repair. This is why the reliability world puts so much weight on the fundamentals, and why autonomous maintenance hands the daily cleaning, inspection, and lubrication to operators. Catching a failure at the "it's getting warm" stage costs a fraction of catching it at the "it's on the floor in pieces" stage. The failures that look random usually were not; they were quiet.
How do you reduce breakdown losses step by step?
Work breakdowns in a fixed order so you attack the failures that cost the most, not the ones that shout the loudest.
- Capture every breakdown with a reason code. No code, no data. Log the machine, the failure mode, the duration, and the repair. Reconstruct-from-memory logs undercount short breakdowns badly.
- Rank by total lost time, not frequency. A once-a-month seizure that costs four hours beats a weekly ten-minute trip. Sort your codes by minutes lost across the period and start at the top.
- Find the root cause of the top failure. Ask why until you reach a cause you can remove, a lubrication miss, a misalignment, a worn part run past its life. Stop at the mechanism, not the symptom.
- Convert the fix into a PM. If the cause was a bearing that fails at a predictable interval, schedule the change before it fails. Move the work from unplanned to planned and the loss disappears from Availability.
- Add condition monitoring where failure is random. When a failure has no clean interval, watch for its warning signs, vibration, temperature, current, so you catch it as a planned stop instead of a breakdown.
- Shrink MTTR for the ones you cannot prevent. Stage the common spare, write the repair steps, and pre-kit the tools. A four-hour repair that becomes forty minutes recovers most of the loss even if the failure still happens.
- Re-measure and move to the next code. Confirm MTBF rose and lost minutes fell, then start again on the new top offender. Reliability is a queue, not a project.
How do you measure breakdown losses?
Two rates tell the whole story: how often failures happen and how long they take to clear. Mean time between failures is total run time divided by the number of breakdowns, higher is better. Mean time to repair is total repair time divided by the number of breakdowns, lower is better. Availability rises when MTBF goes up, MTTR goes down, or both.
Feed both into the Availability term of your OEE calculation and watch the trend, not the single number. A month with one ugly breakdown and a month of steady small ones can show the same Availability while telling you completely different things about the equipment. Honest machine downtime data, coded at the moment of the stop, is what makes the two distinguishable.
Where do breakdown losses cost the most?
Breakdown losses cost the most at the constraint, where an hour down is an hour of plant output gone. A breakdown on a step with slack barely dents throughput because the buffer covers it; the same breakdown on the bottleneck stops the whole plant. That is why finding the constraint first, with bottleneck identification techniques changes how you rank your breakdowns. Two machines with identical MTBF are not equally important if one is the constraint and one is not.
The dollars back this up. Unplanned downtime is consistently among the largest controllable costs in a plant, and the losses concentrate on a handful of assets. Rank by lost constraint-hours and a short list of machines usually explains most of the pain, which is exactly the list your preventive maintenance schedule should be built around, and the list your capacity planning metrics should assume will occasionally go down.
How does tracking make breakdown losses visible?
You cannot separate breakdowns from minor stops if both land in one "downtime" bucket at the end of the shift. The split has to happen at the moment of the stop, with the reason coded by the person who saw it. That is the case for capturing downtime on the floor in real time rather than reconstructing it later, where short breakdowns quietly vanish and the reason codes turn into guesses.
Plants like CLS replaced paper downtime logs with real-time capture, so a breakdown and a swarm of minor stops show up as what they are the same shift, coded, and ready to rank. Once the two are visible on a live board, the countermeasures sort themselves, the reliability queue for breakdowns, the stability work for minor stops. If you want a first read on what your Availability losses are worth, run your line through a free OEE calculator.