Changeover Video Analysis: How to Film a Changeover for SMED

Key takeaways

• Record a real changeover from the last good unit to the first good unit at rated speed.
• Watch it with the crew and write down every step with its minutes.
• Mark each step internal (machine stopped) or external (done while it runs).
• Move external work before the stop, then shorten the rest.
• On the sample line below, the changeover falls from 68 to 31 minutes, worth $222,000 a year if the line is sold out.

Changeover video analysis means you film one changeover, then list every step with its time. You sort the steps into work that needs the machine stopped and work that does not. Then you move the second kind out of the stop and measure again.

Fabrico is OEE (overall equipment effectiveness) and light MES (manufacturing execution system) software with AI cameras. It logs each changeover as a timed stop with its reason, so you can see whether a new method holds. The filming itself is done by your own team, with a phone or a fixed camera.

Why film a changeover instead of timing it with a stopwatch?

A stopwatch gives you one number. A video gives you every step, in order, with its time. You can pause it, replay a step and count every walk away from the machine.

Memory is the other problem. Ask a crew how long a format change takes and you get the good day. The video shows the real one, with the missing tool and the third try at a guide setting.

For the definition and its effect on OEE, see our guide to changeover time and how to cut it.

How do you set up the camera for a changeover video?

A phone on a tripod is enough for most lines. Some plants use a fixed camera over the machine instead.

• Start early: a few minutes before the last good unit of the old run.
• Stop late: only when the line makes good units at rated speed again.
• Frame the work: the guards, the change parts and the trolley, not faces.
• Keep one clock: the time stamp in the video, or a clock in view.

Tell the crew why you film, who will watch and when the file is deleted. In Germany, the works council has a say in any device that can monitor behaviour or performance. In France, the staff committee (CSE) is consulted before a camera goes in.

Talk to them before the first recording, and see our guide to cameras on the line without watching people.

How do you turn the video into a step list?

Watch the video with the crew who did the changeover, the same day or the next. They know why each step happened. Pause at every new task and write one row:

1. Step, in plant words, such as "fetch star wheels from the store".
2. Start, end and minutes, from the time stamps in the video.
3. Internal or external: could it have been done while the old run was still going?
4. Kind of time: work, walk, search, wait or adjust.

Do not argue about fixes yet. Get the list right first, then check that the minutes add to the full changeover.

Which steps can move outside the stop?

Internal work needs the machine stopped. External work can happen while it runs. That split is the core of SMED (single minute exchange of die), and the video shows where external work hides inside your stop.

• External work done inside the stop. Fetching change parts, finding the format sheet, bringing labels or caps. Move it before the stop, onto a trolley that waits at the machine.
• Internal work that can be shorter. Sort the parts in fitting order, keep the tools on the trolley, and work in parallel where it is safe.
• Trial and adjust. You cut it with marked positions and recorded settings for each format, not by working faster.

Waiting counts too. If the line waits for a quality sign off, see our guide to first off approval after a changeover. For the full before, during and after list, use our changeover checklist.

Worked example: what does one filmed changeover show?

Sample line (illustrative numbers, not a customer): a bottle line rated at 200 bottles a minute, with 2 format changes a day on 250 production days a year. The crew films one change from 0.5 litre to 1 litre bottles.

On the video, the change runs 68 minutes from last good bottle to first good bottle at rated speed.

• External work inside the stop = 9 + 5 + 4 = 18 minutes, or 26% of the changeover. Moving it out leaves 68 minus 18 = 50 minutes.
• Trial and adjust was the biggest step: 21 minutes, or 31% of the total. Marked positions cut it to 8 minutes and save 13.
• A shorter swap and a cleaner restart save 3 + 3 = 6 more minutes. New changeover: 31 minutes.

Now put money on it:

• Saved per changeover: 68 minus 31 = 37 minutes, or 37 x 200 = 7,400 bottles.
• Per year: 7,400 x 2 x 250 = 3,700,000 bottles.
• At a contribution of $0.06 a bottle (or €0.06 in a euro plant): $222,000 a year, or €222,000.

That is cash only if the line is sold out, or if the time saved cancels overtime. Agree that rule with your controller first.

Nobody worked faster. 18 minutes left the stop because the parts were ready, and 13 more because the settings were written down.

How do you measure the changeover after the workshop?

Measure every changeover the same way: from the last good unit of the old run to the first good unit of the new one at rated speed. If you stop the clock at the last fitted part, you hide the trial and adjust block.

1. Film the next changeover once more to check that the new steps hold.
2. Write the new standard: steps, settings per format and a standard time, here 31 minutes.
3. Code every minute over the standard with a reason. The standard part can sit in planned time.
4. Watch the trend for weeks, not one day. SMED gains slide back when settings are not kept up.

Here software helps and video does not. Fabrico logs each changeover as a timed stop, and the operator confirms the reason in one tap, such as "format change, 1 litre". Slow running after the restart shows in the Performance part of OEE, live by line and shift.

So you do not need to film every changeover, because every one is already timed. Film again when the time drifts or the formats change.

What mistakes ruin a changeover video study?

• Stopping the video at the last fitted part, so the restart never gets counted.
• Filming people instead of the work, then using the video to judge someone.
• Filming before the crew, and the works council where there is one, know why.
• Buying new change parts before moving the free external work out of the stop.

Where Fabrico fits

Fabrico is a computer vision enabled Manufacturing Performance Platform. It runs with OEE and light MES first. It does not run your SMED workshop or film the whole changeover. It gives you the measured before and after.

• Automatic downtime logging, so each changeover is a timed stop, not a number from memory.
• Every loss priced in money and losses ranked by money, so a long changeover sits next to the jams it competes with.
• OEE live by line and shift, so the trend after the workshop is on screen the next morning.

You start with a pilot on one line, with a baseline and a decision date agreed in advance. There is no free plan and there are no tiers: each plant gets a quote. More in our guide to light MES.

Frequently asked questions

How do you film a changeover for SMED?

Put a phone on a tripod with a wide view of the machine and the change parts. Start before the last good unit of the old run, and stop after the first good unit at rated speed. Aim at the work, not at faces, and tell the crew how the video will be used.

Who should watch the changeover video?

The crew who did the changeover, the line leader and whoever runs the SMED workshop. The crew explains why each step happened. Keep the group small, and delete the file when the step list is done if you promised that.

Can software turn a changeover video into a step list?

Do not count on it: people make the step list by watching the video and writing each task down. Fabrico does not film the whole changeover. It logs each changeover as a timed stop with its reason, so you can track the result after the workshop.

What is a good changeover time?

There is no general figure. SMED aims at single-digit minutes, under 10, but your target depends on the machine and the formats. Set a standard for each format pair from your own filmed changeover, then code every minute over it.

See your changeovers in a demo

Bring one line and its last changeover log. We will show each changeover as a timed stop, with its reason and what it cost.

Request a demo · Contact us