Changeover Time Is Not the Cost — The Iterations Are

 A changeover is 20 minutes of spanner work and 70 minutes of everything else. How to price it, measure it, and cut the part that actually costs.

Primary keyword: thermoforming changeover time Secondary: quick mould change thermoforming machine, thermoforming setup time, high-mix thermoforming line

A buyer visits a factory, asks to see a mould change, and starts a stopwatch. The tooling comes out, the next set goes in, the clamps close, the operator taps a screen. Twenty-two minutes. He writes it down, compares it against the other two quotes on his desk, and buys the cheaper machine.

Three months later the same plant is running nine changeovers a week and the schedule has fallen apart. Nothing was mis-sold. He simply timed one of four things that happen during a changeover — and it was the only one a stopwatch can see.

A changeover on a four-station pressure former runs door-to-door from last good part of product A to first saleable part of product B. Inside that window there are four separate jobs:

Bucket 1 is what gets demonstrated. Bucket 4 is what the plant lives with.

The reason buyers never see buckets 2 and 4 is structural: they are not events, they are waiting. Nobody films a heater bank soaking. Nobody puts "operator walked around the machine three times" on a quotation sheet.

A trial loop is not free, and it is not one cycle:

1. Form one sheet
2. Walk to the stacker, take a part
3. Check wall thickness at the corners, check the cut, check the trim
4. Decide: raise this zone, shorten that delay, drop the cutting depth 0.2 mm
5. Wait for the change to take effect — on the thermal side, that wait is not instant

On a 650 × 750 mm four-station machine, one honest loop is 8 to 12 minutes. Three loops is half an hour of a machine that costs $45–80 per hour fully loaded. And the loops cluster: the first one is always wrong, because the sheet that comes off a heater bank that has just closed the doors is not the sheet that will come off it an hour later.

That last point is the one that separates a good operator from a fast one. The good ones stop trying to fix the first three sheets and simply let the machine soak, because they know the setting they "found" at minute four will be wrong at minute forty.

Take a plant running a mixed packaging catalogue — trays, blisters, clamshells — on one line.

Annual cost = 432 × (110 − 45) / 60 × $55
            = 432 × 1.083 h × $55
            ≈ $25,700 per year

Twenty-five thousand dollars a year is not a rounding error — it is the size of the price difference between two machines of this class. But note what the number is not: it is not the 22 minutes on the stopwatch. If the buyer had priced the changeover at 45 minutes, he would have written down roughly $17,800 and understated the gap by a third.

There is a second number worth calculating, and almost nobody does:

Minimum economic batch = changeover cost per event ÷ contribution per part

At 101 per changeover (110 minutes at $55/h) and $0.12 contribution per tray, the first 840 parts of every run exist only to pay for the changeover. Below that batch size, the line is running to make scrap look productive.

Speed comes from removing decisions, not from moving faster. Six things are worth paying for, and they cost the machine builder very little:

1. A recipe that stores heater set-points, not just times. If the operator has to dial in eight temperatures by hand for every product, bucket 3 never gets below ten minutes and bucket 4 starts from scratch every time.

2. Zone count. Four heater zones top and four bottom gives you the ability to trim one cold edge without touching the whole bank. Three and three forces a compromise that shows up as a thin corner.

3. A common base plate. If every mould shares the same bolt pattern and the same mounting height, bucket 1 collapses to a lift and four clamps. If every mould is a bespoke frame, it does not.

4. Quick clamps and a chain-rail width scale. Rail width is the adjustment most often "eyeballed" — and then re-done after the first sheet comes out with a wandering trim line. A scale turns a two-attempt job into a one-attempt job.

5. Reference marks for cutting depth and force. Removes one full iteration from every changeover, forever.

6. A written first-off checklist — five items, signed. Not for quality theatre. It exists so that the operator stops re-litigating the same four checks in a different order every time.

Before you compare two machines on changeover time, define the measurement so that it cannot flatter anyone.

1. Define door-to-door, in writing. From the last good part of product A to the first part of product B that the QC bench accepts. Not "mould in place".
2. Time five changeovers, not one. The first one after a maintenance day is always slow. The one before a deadline is always suspiciously fast.
3. Record iterations separately from minutes. "34 minutes, 2 loops" tells you where to attack. "34 minutes" does not.

If a supplier cannot give you those three numbers from their own shop floor, they are quoting you a demonstration, not a capability.

1. Which parameters does the recipe store — times only, or heater set-points as well?
2. Do the moulds share a master base plate? What is the bolt pattern and the mounting height?
3. How is the chain-rail width set — scale, gauge, or by eye?
4. What is the heater zone count, top and bottom?
5. Can you send me the changeover sheet for two products I actually run, filled in with your measured minutes?

The answers are checkable. The stopwatch number is not.

If you change over twice a week, none of the above is your problem — buy on mechanical quality and service, and ignore this article.

If you change over three times a day, changeover time is probably the largest line in your cost sheet that no one owns. It does not appear in the ERP, it is not in the quotation, and it is invisible to the salesman who is comparing prices with you.

That is the line worth arguing about.

Related reading

• From single-station to four-station: what Southeast Asian packers are specifying in 2026
• Cavity count is not a speed dial: how to calculate what your machine supports
• Compressed air is not a utility on a pressure former — it is the tool
• HX-6575 four-station pressure and vacuum thermoforming line

The four-bucket breakdown, the 8–12 minute loop and the 15–40 / 10–30 / 5–15 / 15–45 minute ranges are planning estimates from shop-floor practice on 650 × 750 mm class four-station machines. They are not measurements of a specific machine, and they are deliberately wide, because the spread between a well-tooled line and a poorly tooled one is that large.

The machine rate of 55/h is an illustrative fully loaded figure (labour, overhead and depreciation); substitute your own. The 432 changeovers per year follows from 9 per week over 48 weeks, and the $25,700 figure follows arithmetically from those inputs — change any of them and the result changes with it.

• Replace the illustrative $55/h rate with a figure your own market recognises, or leave it clearly labelled as an example (as written).
• If you can supply your measured changeover minutes for two real products, swap them into the section "Measure it in your own plant" — nothing builds more trust than a filled-in sheet.
• The article deliberately quotes no cycle times and no output figures, consistent with the other three pieces on the site.
• Both images are interim alicdn links. Move them to  on the site before publishing and update the alt text if the file names change.
• whatsapp:+86 18038149443