BLOG

SMED without a consultant: cut your changeover time in 4 steps

Separate internal and external work, time each step, and test safer ways to prepare the next changeover.

Every changeover is stopped-machine time. If your press, injection moulder or packaging line changes over 3 times a day and each change takes 45 minutes, you're giving away more than 2 hours of production a day. Multiply by a year: that's whole weeks of capacity you already own and don't use.

SMED (Single-Minute Exchange of Die) is a method for reducing changeover time by separating work that requires a stopped machine from work that can happen while it runs. The good news: you don't need a consultant or a six-month project. You need to measure ONCE with a method and decide activity by activity. This is a useful approach for SMEs, and it's exactly what you'll be able to do yourself by the end of this article.

Why your changeover takes twice as long as it should

Almost every un-analysed changeover hides the same flaw: things get done with the machine stopped that could be done with the machine running. Hunting for fixtures, preparing material for the next format, adjusting parameters from the previous batch, walking off for tools. All of that is external work disguised as internal.

In SMED there are only two kinds of activity:

  • Internal: can only be done with the machine stopped (mounting the die, for example).
  • External: can be done while the machine is producing (fetching the die, preparing material, reading the work plan).

The opportunity varies by process. Time the steps before setting a target; some work may be movable, but safety and quality checks still come first.

Worked example: a press changeover

In this illustrative A→B press changeover, timed activity by activity: 43 minutes of downtime. Classifying the 14 activities showed 7 were external work done internally: preparing fixtures, locating the work plan, pre-heating the die, bringing in material for the next format.

The improvement plan fit on one page: every external activity moved out of the downtime (with an owner and a specific moment), and the internal ones were simplified (quick-release clamps, a changeover cart prepared the day before, a checklist stuck on the machine).

In the worked example, moving and simplifying steps takes the stop from 43 to 21 minutes. That is 22 minutes less per changeover; at four changes a day, it would be 88 minutes of potential machine availability. It is not a measured e2b client result.

The 4 steps

1. Time the current changeover, activity by activity

Pick the most frequent changeover in your plant. One person with a stopwatch (a phone works) writes down each activity and its duration - no opinions, no improving anything yet. One pass is usually enough to start; two if the changeover varies a lot.

2. Classify: internal or external

For each activity ask: can this be done with the machine running? If yes, it's external. This is where the surprise shows up: some steps may be in the wrong place. Measure yours rather than assuming a percentage.

3. Convert and simplify

First, move everything external outside: fixtures, material and paperwork ready BEFORE the stop. Then attack the internal work: quick-release clamps, position references, duplicate sets of fixtures, checklists. Every activity in the plan gets an owner and a new estimated duration.

4. Standardise and measure before/after

Time it again. If the first round has little effect, review your classification and constraints. A large cut is not guaranteed. When the new time holds for three changeovers in a row, document it: it's your new standard and the base for the next round.

The mistakes that kill a SMED

  • "Eyeballing" the times: without per-activity times there's no classification, only opinions.
  • Improving while measuring: first you watch, then you touch. Mix the two phases and you won't know what worked.
  • Leaving out the person who does the changeover: the operator knows where the time goes; a plan born in an office dies within a week.
  • Stopping after the first round: SMED is a cycle. A second analysis may uncover further opportunities, but measure the result.
  • Not standardising: without a checklist and a written standard, the changeover may drift back towards its old time.

The template I use for this

The full analysis in this article - the timing sheet, the internal/external classification, the improvement plan and the before/after calculated automatically - is the e2b SMED: Changeover Reduction template. It comes with an illustrative press example pre-filled (43 → 21 min, 51%) so you can copy the mechanics, plus a PDF guide in English and Spanish.

See the SMED template