Micro stops in manufacturing are brief interruptions in a production process that are easy to dismiss one at a time. In OEE language they are often called small stops or minor stops, but the threshold and KPI classification depend on the plant’s measurement policy. The practical test is simple: measure the repeated minutes, annualize them using the real operating schedule, and give the pattern an owner.
The 20-minute example in the video is an interview illustration, not a survey or an industry benchmark. At 260 operating days, 20 minutes per day equals 86.7 hours per year. The video rounds that scale to about 100 hours so the capacity question is hard to ignore.
What micro stops and minor stops are
A micro stop is a brief production pause that is not long enough to be treated as a formal equipment failure in the plant’s current system. The OEE glossary calls this loss family idling and minor stops, also known as small stops, and places it under performance loss. Siemens documentation makes the important qualification explicit: the definition can vary by use case, and a system may assign the event to performance or availability depending on its rule.
That means micro stop and minor stop are useful working terms, not a universal stopwatch rule. One plant may use two minutes, another five, and another a threshold that changes by line speed or capture method. Write the threshold into the measurement policy before comparing shifts or claiming improvement.
Why micro stops in manufacturing go unrecorded
An operator can clear a small jam, reset a sensor, move a misaligned part, or wait briefly for material before a formal downtime entry feels worthwhile. If the line resumes before anyone records the event, the loss stays inside the shift routine.
OEE.com’s implementation guide recommends deciding how stop time will be recorded and describes a tick sheet or start and end times as practical methods. It also treats the threshold as a decision: events at or above the threshold can be recorded as stop time, while shorter events may be treated as small stops and performance loss. The exact boundary belongs to the plant.
This is why a strong micro-stop measure is more than a count. It should let you answer:
- Which line and asset was affected?
- Which shift, product, or SKU was running?
- How long did the event last?
- What did the operator observe and do?
- Did the event affect the constraint?
If those answers are missing, the dashboard may show that performance is down without showing which condition can be changed.
Annualize it: minutes per day to hours per year
Use this calculation:
Annual hours lost = average micro-stop minutes per operating day x operating days / 60
The table below uses 260 operating days. It is an illustrative planning table, not a plant benchmark.
| Average minutes per day | Operating days | Annual hours lost | Reading |
|---|---|---|---|
| 5 | 260 | 21.7 | Easy to dismiss in one shift, visible over a year |
| 10 | 260 | 43.3 | A recurring half-hour each three operating days |
| 15 | 260 | 65.0 | A meaningful capacity question for a constrained line |
| 20 | 260 | 86.7 | The video example, often rounded to about 100 hours |
| 30 | 260 | 130.0 | A daily pattern that deserves an owner and countermeasure |
If your plant runs five days a week for 52 weeks, the 20-minute row uses the same arithmetic: 20 x 5 x 52 / 60 = 86.7 hours. Replace 260 with the schedule you actually run. Planned shutdowns, seasonal production, overtime, and maintenance windows change the result.
Do not turn the table into a claim that every line loses 100 hours. The useful output is your own number, with the schedule and measurement threshold written beside it.
What to log for every stop
Start with a simple record that a supervisor can review at the end of a shift. The goal is not to create a perfect software schema. It is to keep the event’s context attached to its duration.
| Field | Why it matters | Example |
|---|---|---|
| Date and shift | Lets you compare like with like | September 30, night shift |
| Line and asset | Identifies where the pattern occurs | Line 2, filler infeed |
| Product or SKU | Shows whether the event follows the recipe or package | 500 ml bottle |
| Start time and end time | Creates an auditable duration | 22:14:10 to 22:15:05 |
| Duration | Supports the threshold and annualization | 55 seconds |
| Event classification | Keeps the local rule visible | Micro stop, below 2 minutes |
| Reason code | Groups similar events | Sensor reset |
| Observed symptom | Preserves what happened, not just a label | Photo-eye false trigger |
| Immediate action | Shows how the line was recovered | Operator cleared and reset |
| Constraint impact | Separates lost line time from lost system capacity | Constraint affected |
| Units delayed or affected | Adds production context when it is available | 35 bottles |
| Source | Shows how the event was captured | Operator log or PLC |
| Owner and next action | Turns a record into follow-up | Maintenance, inspect bracket |
OEE.com’s measurement guidance supports starting with manual measurement when that is the practical way to learn the loss boundary. Automation can make short events easier to detect later, but it does not decide what the event means or who should change the condition.
Turning stop logs into a capacity number
Once the log has enough entries, separate the time calculation from the financial assumption:
Annual capacity opportunity = annual hours lost x agreed value per hour
For the 20-minute example at 260 operating days:
Annual capacity opportunity = 86.7 x V
Here, V is the value your plant agrees to use for one hour of the affected capacity. It may be a contribution-margin measure, a planning value, or another finance-approved basis. This article does not invent a dollar amount because revenue, labour cost, and contribution margin answer different questions.
The constraint matters. If a short stop affects the bottleneck that limits the whole line, the capacity opportunity may deserve immediate attention. If it affects a non-constraint step with available recovery time, the same minutes may not carry the same business consequence. Record the constraint field rather than assuming that every lost minute has an identical value.
If the short stop also creates rejects or rework, keep that loss visible separately. First-pass yield versus final yield explains why an output can be accepted after rework while the first-pass loss remains real.
Use the pattern to choose the first fix
Do not start by asking which AI tool detects micro stops in manufacturing. Start with the pattern in the log:
1. Sort events by total minutes, not just by event count. 2. Split the result by line, shift, product, and reason code. 3. Mark which events affect the constraint. 4. Review the symptom and immediate action with the operator, maintenance, and quality people closest to the condition. 5. Assign one owner and one next observation, then check whether the minutes change.
This keeps the page’s promise honest. A micro-stop log does not prove root cause, and an OEE score does not tell you which countermeasure will work. It gives the team enough shared evidence to choose the next question.
For the wider boundary between hidden waste, rework, and capacity, see how to measure the hidden cost of manufacturing waste and rework. To structure the first floor observation, use the Gemba Walk Checklist.
The practical rule is simple: define the local threshold, log the small events, annualize the minutes, and investigate the repeated pattern that affects the constraint.
