This article is for production, manufacturing engineering and continuous improvement managers preparing a SMED project, or who ran one without seeing the expected OEE gain.
- A changeover runs from the last good part of batch A to the first good part of batch B, at nominal speed.
- The stopwatch only sees the stop, not the slowdown or the ramp-up.
- How often a changeover happens matters as much as how long it lasts.
- The changeover matrix tells you what to tackle with SMED and what to standardize.
A SMED project always starts with a measurement. If that measurement only covers part of the changeover, the project optimizes the visible part and leaves the rest untouched. Before reducing changeover time, you need to know where it starts and where it ends.
Where a changeover starts and ends
In What is the SMED method and how do you apply it? (in French), Julien Charles of 7-Shapes gives the right definition from step one. The changeover runs from the last part of batch A to the first good part of batch B. It breaks down into three phases: the slowdown to a full stop, the no-production phase, then the ramp-up to nominal speed.
The article also recommends filming the changeover to analyze it afterwards, then walks through the classic steps: separate internal and external operations, convert internal to external, reduce what remains, standardize.
This three-phase definition is the key. In many plants, only the second phase is measured.
What the stopwatch doesn’t see
- The slowdown and the ramp-up. The stopwatch starts when the machine stops and ends when it restarts. The minutes spent emptying the line and getting back to nominal speed are not counted, even though output during that time is below plan.
- The first nonconforming parts. After restart, adjustments are often made on the first parts. Until the part is good, the changeover is not over.
- Variability. The same changeover does not take the same time depending on the crew, the shift or the starting product. A one-off timing gives you a value, never a spread.
- Frequency. A 20-minute changeover repeated three times a day weighs more than a 45-minute changeover done once a week. Without counting, you target the longest changeover instead of the most expensive one.
Three ways to measure, compared
| Method | What it brings | Limitations | When to use it |
|---|---|---|---|
| Stopwatch timing on the floor | Detail of each task, discussion with the crew | A single changeover observed, stop phase only, observer effect | To break down a changeover you have already chosen |
| Video | Review every operation, classify internal and external with a clear head | Heavy to analyze, small sample | During the SMED workshop itself |
| Automatic measurement of stops and speed | Every changeover, every crew, ramp-up included | Does not tell you which tasks are performed | To choose what to tackle, then verify the gain |
The three complement each other. Automatic measurement tells you which changeovers cost the most and how much they vary. Stopwatch timing and video tell you why.
Reading the data: the changeover matrix
Once every changeover is measured, one simple view is enough to choose the target: a matrix with the starting product in rows and the next product in columns. Each cell shows the median changeover time, the gap between the fastest and the slowest, and its frequency over the period.
Three readings come out of it:
- The most expensive pairs in cumulative time, to tackle first with SMED.
- The most variable pairs, where the gain comes first from standardization: the best crew already knows how to do it.
- The avoidable pairs, which a different schedule could sequence better or eliminate.
The second reading is often the biggest surprise. When the same changeover varies widely from one crew to another, the problem is not technical: it is a standard that does not exist yet.
After measurement, SMED
With complete measurement, the SMED approach described by 7-Shapes delivers its full value. The article cites changeover time reductions of 30 to 40%, and notes that simply moving internal operations to external brings part of the gain with little effort.
The same measurement is then used to check that the gain holds over time, crew by crew. Without it, a SMED standard erodes without anyone noticing.
TEEPTRAK measures every stop and the line speed continuously, with no PLC connection. The first figures are available within 48 hours, after an installation of about one hour per machine.
Further reading
- What is the SMED method and how do you apply it?, Julien Charles, 7-Shapes: the six steps of SMED, illustrated step by step.
- SMED: how to reduce changeover time and improve OEE with data
FAQ
How do you calculate changeover time? Measure the time between the last good part of the previous batch and the first good part produced at nominal speed on the next batch. Slowdown, stop and ramp-up are all included.
Why film a changeover? Video lets you review every operation and decide, with a clear head, what must be done with the machine stopped and what can be prepared beforehand. It complements continuous measurement, which tells you which changeovers to film first.
Which changeover should you tackle first? The one that costs the most in cumulative time, meaning its duration multiplied by its frequency, which is not necessarily the longest one.
0 Comments