A factory acceptance test is the last point in the project where a problem belongs to the supplier. After the line ships, every defect becomes a discussion about freight, site conditions and who pays for the engineer. That asymmetry is the reason a FAT deserves more preparation than most buyers give it.
This is a practical checklist, in the order a test should run.
Before the test: agree the criteria in writing
The single most useful thing you can do happens before you travel. Put the acceptance criteria in the contract, expressed as measurable outcomes rather than descriptions. "No.4 finish to sample" is not a criterion; "No.4 finish, Ra within the agreed range, no visible striping under the agreed lighting at the agreed distance, matching the retained sample" is.
Also agree what happens if a criterion fails: rework, retest, delay, or price adjustment. A FAT without a failure clause is a demonstration, not a test.
1. Run your material, not theirs
Suppliers test on material they have on hand, which is usually clean, flat and well within tolerance. Your material may be a different grade, a different hardness, a different coil condition, or have surface defects that the machine must remove. Send your material ahead of time and require that the test is run on it.
If the line cannot produce the finish on your material, everything else in the test is irrelevant.
2. Run the full sequence, not a single pass
A common shortcut is to demonstrate the final polishing head on pre-ground material. That proves the last head works; it does not prove the line can take raw material to finished product in one run. Require a full run from the incoming condition to the finished condition, with the abrasive sequence that will actually be used in production.
3. Measure the finish, do not judge it
- Ra at several points across the width and along the length, measured in the agreed direction
- Gloss if the specification includes it, at the agreed angle
- Visual comparison against the retained sample, under the agreed light and viewing distance
- Consistency between the first metre and the last metre of a long run — this is where belt wear and pressure drift show up
Ask to see the measurements recorded, not just performed. A supplier who measures but does not record is not giving you a baseline for later comparison.
4. Check the things that are hard to fix later
- Flatness of the finished product, especially on wide material. A finish can be correct on a panel that is not flat.
- Edge quality — burr, edge marking and whether the heads can reach the edge without marking it.
- Tracking on coil lines — does the strip stay centred through the full speed range, including during acceleration?
- Extraction performance — dust or mist at the head, measured or at least visibly demonstrated, not described.
- Noise and vibration at operating speed. These are difficult and expensive to correct after installation.
5. Run the speed range, not just the nominal speed
A line is usually demonstrated at the speed that gives the best result. Production runs at the speed that gives the required output. Test at the upper end of the intended production speed and confirm the finish still meets specification. If it does not, the real capacity is lower than the quoted capacity, and that changes the payback calculation.
6. Test the changeovers
Time an abrasive change and, if the line is intended for multiple finishes, a finish change. Changeover time is a capacity assumption in every sizing calculation, and it is rarely measured during a demonstration. Ask the operator to do it at normal pace, not at demonstration pace.
7. Verify the controls and the recipes
- Recipe recall actually restores pressure, speed and head positions
- Alarms and interlocks function — ask to see one triggered deliberately
- Data logging works and can be exported, if it was specified
- Language of the interface and documentation matches the contract
- Spare parts list is complete and matches what is installed
8. Retain samples and records
Keep signed finished samples from the FAT — several, from different points in the run — and keep the measurement records with them. These become the reference for site acceptance, for the first production batch, and for any dispute six months later. A sample in a photograph is worth very little; a physical sample with a signed record is worth a great deal.
Photograph the line as built, including nameplates, motor labels and the extraction connections. When a part fails in year three, those photographs are how you identify it.
9. Site acceptance is a separate test
A line that passed the FAT can still fail on site, because the foundation, power supply, compressed air and extraction ducting are different. Plan a site acceptance test with the same criteria, run after commissioning and before final payment. The most common site-specific problems are vibration from an inadequate foundation, unstable power affecting drive performance, and extraction that was sized for the factory rather than for the installed duct run.
10. Payment structure is part of acceptance
Whatever the commercial terms, retain a meaningful final payment until site acceptance is signed. That is not distrust; it is the mechanism that makes the acceptance criteria matter. A line with 100 percent paid before shipment has no remaining leverage if the last five percent of performance is missing.
A one-page FAT checklist
- Criteria agreed in writing, with a failure clause
- Test run on your material, full sequence
- Ra and gloss measured across width and length, recorded
- First metre versus last metre compared
- Flatness and edge quality checked
- Tracking verified through the speed range
- Extraction demonstrated, not described
- Changeover times measured
- Recipes, alarms and documentation verified
- Samples and records retained and signed
- Site acceptance criteria and payment retention agreed