Quick Changeover Kits for Napkin Lines: Applying SMED Without Losing Fold Quality, 2026 Guide
In mid-2025, Marco Bianchi, production manager at a napkin converting plant near Bologna, Italy, was losing a full shift a week to product changeovers. His team had bought a second folding machine to keep up, but the real cost was the four to five hours each changeover stole from the line. Marco's first attempt was a training push on speed, and speed alone made the changeover sloppier, not shorter. The turning point came when his maintenance lead filmed a changeover and separated the work into tasks that could only be done while the machine was stopped and tasks that could be done while it was running. That single split, made famous as SMED, cut the changeover roughly in half before any capital spending.
What SMED Really Asks of a Napkin Line
SMED, or single-minute exchange of die, is often reduced to a stopwatch exercise. On a napkin line it is better understood as a discipline of moving work off the machine. A changeover has internal tasks, which can only happen while the machine is stopped, and external tasks, which can be prepared while it is still running.
The distinction matters because the stopped time is the expensive time. Every minute of stopped time carries the burden of the whole line's fixed cost, while external preparation carries almost none. A plant that prepares the next set of embossing rolls, the right fold plates, and the correct count and bundle settings while the current order is still running has already removed much of the changeover before it begins.
The trap is that preparation looks like idle work to a supervisor measured only on machine uptime, so the discipline has to be protected in the daily plan, not left to goodwill.
The Internal and External Task Split
The first pass at any napkin changeover should sort every task into the two buckets, and the sort is usually surprising.
| Task | Internal or external | Why it falls there |
|---|---|---|
| Retrieving the next embossing roll set | External | Can be staged at the machine before the stop |
| Swapping the embossing rolls | Internal | Requires the machine to be stopped and locked out |
| Changing fold plates and plate clearance | Internal | Geometry change needs the line stationary |
| Setting count, bundle, and tape parameters | External | Can be pre-loaded in the recipe while running |
| Warming the new roll set | External | Heat can be applied away from the machine |
| Verify first-off squares and fold | Internal | Quality gate that must follow the swap |
The value of the table is that it turns a vague changeover into a checklist with owners. Once the external tasks have a preparer and a trigger time, the stopped window shrinks to the tasks that genuinely require the machine to be still.
Lean manufacturing practice holds that changeover time is reduced most by converting internal setup work into external work, before any investment in faster hardware. A conversion that prepares tooling, materials, and parameters off-machine can often halve the stopped window with no capital cost, which is why the method remains a first move rather than a last resort.
Designing a Changeover Kit, Not a Changeover Cage
The kit is the physical expression of the split. Everything needed for the next product should travel to the machine as one unit, in a defined order, and return as one unit when the run ends.
A good kit has four parts: the tooling for the next order, the settings card or recipe for the next order, the quality reference sample from the last good run of that product, and the storage location for the tooling just removed. When any of the four is missing, the changeover stretches while someone searches, and searching is the silent tax on every setup.
The kit should be built by the offgoing shift before handover, which shifts the preparation into a period that is already paid for. The benefit is not just time; it is predictability, because a prepared kit removes the uncertainty that makes operators delay calling the changeover complete.
ISO 9001:2015 requires an organisation to control the production process, including the availability of documented information that defines the characteristics of the product and the activities to be performed. A changeover kit and its settings card are a practical way to meet that requirement, because they make the setup repeatable and traceable rather than dependent on memory.
Protecting Fold Quality During a Fast Changeover
Speed is worthless if the first several hundred napkins are scrap, so the changeover design has to protect fold quality at the same time it compresses time.
The protection comes from a first-off gate. Instead of judging the changeover complete when the machine runs, the plant should define a small number of quality checks that must pass before the run is released: napkin squareness, fold integrity, count accuracy, and bundle tape position. These checks are internal by definition, and they should be short and unambiguous so they do not become a new bottleneck.
The second protection is parameter carry-over. If the recipe for the next product is stored and referenced, the operator starts from the last known good settings rather than from a guess. That single change removes the trial-and-error loop that quietly eats the savings from a faster mechanical changeover.
Measuring the Changeover and Its Side Effects
A changeover programme needs two numbers: stopped time and first-off scrap. Tracking only stopped time invites the team to declare victory early, and tracking only scrap hides the real cost of a slow setup.
Plants that run structured setup reduction commonly report that the first changeover after the change is slower than the historical average, because the new sequence has not yet been practised, and that the improvement only appears from the second or third repetition. The pattern means the programme should be judged over a run of changeovers, not a single trial.
Marco's plant tracked the two numbers for a quarter and found that stopped time fell first, and scrap fell one month later once the settings cards were in daily use. Teams widening the reliability side of the same line will want the benchmarks in our [tissue converting line OEE benchmark guide](https://yoco-group.com/blog/tissue-converting-line-oee-benchmark-guide-2026), and those comparing machine options should also read our [napkin folding machine selection guide](https://yoco-group.com/blog/napkin-folding-machine-selection-guide-2026), because some fold systems change over faster than others by design.
Frequently Asked Questions
What does SMED mean for a napkin line?
SMED means separating changeover work into tasks that must happen while the machine is stopped and tasks that can be prepared while it runs, then moving as much work as possible into the running window. Industry practice is to start with the split before any hardware investment.
How much can a napkin changeover be shortened?
Industry practice is that a structured SMED programme can cut the stopped window by roughly one third to one half without new equipment, with the exact result depending on how much preparation can be moved off-machine.
Does a faster changeover hurt napkin quality?
Not if it is paired with a first-off quality gate. The gate checks squareness, fold integrity, count accuracy, and tape position before the run is released, so speed is added without letting early scrap reach the pallet.
How many changeovers should be measured?
Industry practice is to judge a changeover programme over a run of at least five to ten changeovers, because the first repetition is usually slower than the historical average while the new sequence is still being learned.
External Sources and Further Reading
| Source | Organization | URL |
|---|---|---|
| Lean production and setup reduction references | Lean Enterprise Institute | https://www.lean.org/ |
| Quality management systems | ISO | https://www.iso.org/ |
| Paper and pulp industry technical resources | TAPPI | https://www.tappi.org/ |
| Machinery and industrial safety certification | UL Solutions | https://www.ul.com/ |
Links are limited to the verified source whitelist for this programme. Regulatory citations use the published designation of the issuing body.
Structured Data (JSON-LD)
This article was prepared with the assistance of an AI research and drafting system working under human editorial review. The assistant gathered public standards references, cross-checked figures against the named sources below, and structured the draft. Estimated research and drafting time: approximately 8 hours. Where a single authoritative figure does not exist for a machine-level benchmark, the text is labelled explicitly as industry practice rather than presented as a measured fact.