There is a specific kind of silence that happens when a buyer watches their new machine run for the first time in their own factory and it will not make a decent bag. It is usually followed by an email that begins, “we have a small problem”.
The factory acceptance test exists to prevent that email. It is the cheapest insurance in the whole purchase, and it is also the step buyers skip most often, usually because the flight is expensive and the salesperson says the machine will be tested “by our engineers anyway”.
First, the plain-English version
A factory acceptance test, or FAT, happens at the seller’s site before shipment. A site acceptance test, or SAT, happens at your plant after installation. Both matter, but only the FAT gives you leverage, because the balance payment has not left your account yet.
That leverage is the entire point. A FAT without a payment milestone attached is a factory tour with extra steps.
What to bring
Do not let anyone tell you the machine will be tested with “similar film”. Bring the real thing.
- At least 40 kg of your actual product, in the condition it will arrive on your line. Frozen means frozen. Sticky means sticky.
- Three rolls of the film you intend to buy, from your actual supplier, at your actual thickness and width.
- Calipers and a steel rule. Phone cameras lie about bag length.
- A spring scale or a small seal strength tester. Hand-pulling a seal until it opens tells you nothing repeatable.
- Your own specification sheet with tolerances written on it, printed, one page.
- A laptop and two empty USB drives.
- Comfortable shoes. You will stand for six hours.
Run it for hours, not minutes
The classic failure is the twenty-minute demonstration: cold machine, perfect film, operator standing by, everything beautiful. Then the machine ships, and in week three the film path starts drifting after four hours of running.
Ask for a continuous run of at least four hours, and count every reject bag in a notebook, by hand. Then ask for the machine to be stopped for thirty minutes and restarted cold, because startup and shutdown are where a large share of your real waste will happen. If the reject rate climbs in hour four, you have found something worth fixing now, while the engineer who built it is standing next to you.
Measure, do not watch
Watching is not testing. Write numbers down:
- Bag length: ten bags, measure each one. Ask what the tolerance is, and whether that tolerance holds at full speed.
- Fill weight: thirty consecutive bags on a check scale. Look at the spread, not the average. A stable average with a wide spread will fail your net content rules.
- Seal strength: five bags, peel test, record the values. Do the same test again after the machine has run for two hours, and again on a bag taken from the middle of the run.
- Leak or drop test: for liquids, squeeze test. For retail cartons or pouches, drop test one case from the height your customer will use.
- Speed: actual output over ten minutes, counted. Not the number on the brochure.
Force a changeover and time it
Ask them to change the machine from one bag size to another while you hold the stopwatch. This is the single most revealing test on the list. Some machines change format in twelve minutes with two tools. Others take ninety minutes, four people, and a manual that contradicts itself.
If you run many SKUs, repeat the changeover twice. The second time tells you what it will really be like after the engineer goes home.
Open the cabinet with the power off
Photograph everything. Every terminal strip, every drive, every label. Then check three things: are the wires numbered and ferruled, is there a schematic in the door, and is there dust or swarf sitting on the bottom plate from the build. The last one sounds petty. It is not; it means nobody cleaned the cabinet before closing it.
Safety and utilities, measured not assumed
Test every interlock and the emergency stop, and watch how long the machine takes to come to a full stop. Measure the noise level at the operator position with a phone app if you have nothing better, because 85 dB sounds abstract on a spec sheet and feels very different after eight hours.
Then ask for the real numbers on compressed air consumption and connected load. Air is where packaging lines quietly overspend: a machine rated at 0.6 MPa may still drink 400 litres a minute during sealing.
The documents to take home
Do not leave without these, and check them before you board the plane:
- Electrical schematic and pneumatic diagram, in PDF and one printed copy
- PLC program backup and HMI project backup, plus any passwords, in writing
- A parameter list showing every setpoint as it ran during the test
- Operation and maintenance manual in your language, not just English
- Spare parts list with part numbers
- Calibration certificates for temperature controllers and load cells
- CE declaration of conformity and the test report behind it
- A short video of the machine running, plus video of every settings screen you will need to touch
That last item is worth more than it sounds. Six months from now, when a new operator has changed a parameter by accident, a thirty-second video of the original screen will save you a service call.
If it fails
Machines fail FATs. That is normal and it is exactly why you are there. What matters is how the failure is handled.
Write the defect on paper, in front of the engineer, with a photo, a description, and a date for the fix. Do not accept “we will adjust it at your factory”. Do not pay the balance. Reschedule the test for the specific items that failed, not a full re-run of everything, and get the rescheduled date in writing before you leave the building.
The sign-off sheet
One page, signed by both sides, listing: machine model and serial number, film and product tested, run duration, measured output, bag dimensions and tolerance, seal strength values, reject rate, outstanding defects with dates, and the date the balance becomes payable. Photograph the signed page and email it to yourself before you leave the site.
That email is the document you will reach for if the machine arrives with a dented frame and a different servo brand.