On most packaging lines, the biggest loss is not the breakdown everyone remembers. It is the steady run of 10, 20 and 40 second stops that nobody writes down: a film splice that trips registration, a photo-eye that sees a gap, a jam cleared in passing. Each one looks too small to matter, yet together they can cost hours a day. Packaging micro-stop analysis reads those stops straight from the PLC, groups them by cause and shows which few patterns to fix first. See your own line's patterns in a micro-stop demo.
Find the Short Stops Operators Never Log
iFactory Micro-Stop AI mines PLC events at second-level resolution, clusters every stop by its cause signature and opens improvement cards for the top patterns automatically. Your team sees what is really eating line time.
What Counts as a Micro-Stop
In OEE, short stops sit under Performance, not Availability. The classic six big losses call them idling and minor stops: brief halts, often a minute or two, that the operator clears without maintenance. Longer stops count as downtime and need a reason. Where the line falls is a site rule. Many plants pick a threshold of a few minutes, and anything shorter is counted as a Performance loss. Agree your own threshold with our OEE team.
Why Operators Don't Log Them, and Why It Matters
A 20 second stop is over before anyone could reach a terminal. Operators are busy clearing it, the cause looks obvious, and the same fault has happened so often that it feels normal. Stops like these are rarely tracked well, which is why they hide so easily. The arithmetic shows what they cost. The example below assumes 400 stops a day on one line, about one every three and a half minutes. Test the numbers on your own line in a line data review.
Not Every Stop Starts at the Machine That Stopped
A machine that is starved has no product coming in. One that is blocked can't push product out. Both stop without being faulty, and both are easy to blame on the wrong machine. The PackML state model helps separate them. A machine in Held wants to run but is stopped by a condition of its own, such as missing cartons. A machine in Suspended is waiting on a problem elsewhere on the line and restarts without operator action. Map your line states with our controls team.
In this illustration the wrapper looks like the problem, but the stop began as a gap at the infeed. Only events with exact timestamps from each machine show the order things happened in.
What Second-Level Event Data Reveals
Shift reports and minute-level counts blur short stops together. Event data from the PLC keeps each state change, fault code and sensor signal with its own timestamp, down to the second. Read in order, a few seconds of events tell a story no operator had time to record. The log below is an illustration from a flow wrapper. See how event mining connects to your PLCs in a PLC data walkthrough.
Stop Guessing Which Short Stops Matter
iFactory Micro-Stop AI reads PLC events at second-level resolution, groups stops by cause signature and opens an improvement card for each top pattern, with the evidence attached. Your team starts on the fix, not the investigation.
Clustering Stops by Cause Signature
A cause signature is the pattern of events around a stop: what happened just before, how long it lasted, how it was cleared and what else was going on. Stops with the same signature usually share a cause, even when the fault code differs. Grouping them, then laying the groups over time, shows when each cause bites. The AI does the grouping across thousands of events, and engineers decide what each group means. The heat map is illustrative. Ask our support team how clusters are formed on your data.
Typical Micro-Stop Causes and Their Signatures
Most packaging micro-stops come from a short list of causes: misfeeds, jams, blocked or misaligned sensors, wrong settings and material changes. What makes them findable is that each leaves a recognisable trace in the PLC data. Use the matrix to know what to look for, then confirm your own signatures in a pattern review.
Set Speeds So Short Stops Don't Spread
Many packaging lines use the V-curve. The key machine, often the slowest, runs at its rated speed. The machines either side are set to recover about 10% faster, and each step further away about another 10%, so buffers refill after a short stop. Micro-stop data shows whether buffers really refill in time or whether each stop ripples along the line. The method has a cost: machines other than the key one keep changing speed, and those changes can cause more short stops and wear. Check your own speed settings with our line engineers.
The Improvement Card Behind Each Top Pattern
Finding a pattern is only useful if someone owns the fix. Each top cluster becomes a card with the evidence, an owner and a check date, so the result can be measured against the same signature afterwards. Close the card only when the signature count has fallen and stayed down. Use these fields even if you run the review by hand. Build the card into your routine in a pattern review session.
Three Patterns and What They Pointed To
These are illustrations of how patterns lead to actions. They are not case studies, and real causes always need checking on the line. In each case the pattern came first, and the fix followed a check on the line by the people who run it. Discuss your own patterns with our support team.
Turn Recovered Seconds Into Packs
Size the value from your own line data, not from a vendor average. The worksheet carries on from the example above and uses assumed inputs to show the arithmetic only. Value the packs at your own margin, and count labour or overtime only where the extra output really replaces it. Run it with your figures in a value review.
Restarts often run below full speed for a while, so the real gain may differ. Check it by measuring the same signatures before and after each fix.
Event Mining on a Server Inside the Plant
iFactory delivers Micro-Stop AI as a turnkey bundle. A pre-configured NVIDIA AI server ships racked and ready with the event mining and clustering models pre-loaded, so second-level PLC data never has to leave the plant. You rack it, plug in power and Ethernet, and the AI is live. Scope it with our integration team.
- Pre-configured NVIDIA AI server, racked and ready
- Micro-Stop AI software pre-loaded
- PLC/SCADA links to each packaging machine
- Network and cabling
- State and fault code mapping per machine
- Training for operators and line engineers
- 24x7 remote monitoring
Server racked, PLCs linked, states and fault codes mapped.
Signatures built, clusters checked with line engineers.
Improvement cards on, teams trained, weekly review started.
Packaging Micro-Stop Analysis Software That Leads to a Fix
iFactory Micro-Stop AI turns raw PLC events into a short list of patterns with owners. See it on packaging data in a short demo.
Micro-Stop Questions From Plant Managers
See the Hours Hidden in Your Short Stops
Mine PLC events at second-level resolution, cluster stops by cause and get improvement cards for the patterns that matter. The full system is delivered turnkey and goes live in 6 to 12 weeks.
