Factory acceptance test (FAT) on Chinese machinery — how to pay for evidence, not promises

There is a precise moment in every Chinese machinery purchase when your negotiating power is at its maximum: the machine is built, it is standing on the factory floor, it has consumed months of the factory's production capacity — and a large part of your money has not yet left your account. That moment is the factory acceptance test. Everything before it is promises; everything after it is logistics. Whatever leverage you have not written into the FAT, you do not have at all.

We sell Chinese aluminium-extrusion machinery to European buyers on commission, and a FAT before shipment is a standard part of every transaction we run — not because it is fashionable, but because it is the only point in the process where a problem costs days instead of months. This article explains what a proper FAT protocol contains, when you should fly to China and when a live video test is genuinely enough, how to tie your payment milestones to FAT acceptance, and what to do when the test finds deviations — which, on a machine with thousands of parts, it usually will.

What a FAT is — and what it is not

A factory acceptance test is a contractual acceptance procedure performed at the manufacturer's plant before shipment, in which the machine is run against written, measurable criteria agreed when the contract was signed. Three words in that sentence carry all the weight: contractual, measurable, before.

It is worth being equally clear about what a FAT is not:

• It is not the marketing video. Every factory will happily send you a clip of a machine running. A machine running is not a machine meeting its specification — a press cycles beautifully on camera whether its dead cycle is 18 seconds or 28.
• It is not the SAT. The site acceptance test happens after installation in your plant and confirms the machine performs in your conditions. The SAT protects you too — but by then the machine has been paid for, shipped for 6–8 weeks and craned onto your foundations. The cost of discovering a problem at SAT is a multiple of discovering it at FAT.
• It is not something you can add later. A FAT bolted onto a contract that says only "buyer may inspect the goods" is theatre. The factory's obligation to pass defined tests, your right to withhold shipment and payment until it does, and the retest procedure must all be in the purchase contract from day one.

The practical consequence: the FAT is negotiated at the same time as the price. A supplier who resists a written FAT protocol during contract negotiation is telling you something important about how the acceptance will go.

What a proper FAT protocol contains

A serious FAT protocol for an extrusion press or a complete line runs to a dozen or more pages. The length is not bureaucracy — every section exists because someone, somewhere, got burned without it. The core sections:

1. Reference documents. The contract technical specification, approved drawings and the agreed component list. The FAT tests the machine against these — not against the factory's current catalogue.
2. Identity verification. Nameplate data, serial numbers of the machine and of the key bought-in components, photographed and matched against the contract. This is where you confirm that the hydraulic pumps and the PLC are the genuine branded units you paid for, not "equivalent" substitutes — component swapping is the single most common quality failure in this trade, and it takes ten minutes with a camera to rule out.
3. Dimensional and build checks. Key dimensions against drawings, weld quality, guarding, cable routing, paint. Boring — and where sloppy workmanship shows first.
4. Dry run. The machine cycles without material: all movements, limit positions, interlocks, emergency stops, alarm handling. Every safety function is triggered deliberately and its response recorded. If you intend to write the EU safety package into the contract — and you should, as we explain in our CE guide — this is where it gets accepted.
5. Load test with material. The machine does its actual job, ideally with your material: a press extrudes through a real die, a furnace runs a full thermal cycle, a packing machine packs your profiles. Duration matters — an uninterrupted run of several hours reveals what a five-minute demonstration hides: oil temperature drift, instability after warm-up, recurring minor stops.
6. Performance measurements against numbers. Each criterion has a contract value, a tolerance, a measurement method and a pass/fail box. No number, no criterion.
7. Documentation handover checklist. Manuals in the agreed language, electrical and hydraulic schematics, spare parts list, test certificates. Checked at FAT, because chasing documents after final payment is a hobby with no end.
8. Punch list and sign-off. A structured list of findings with categories and deadlines (more below), and a signature page that makes the result contractual.

What do measurable criteria look like in practice? Examples of the type of line items we write into protocols for typical machines:

Machine	Example criterion	Contract value	Method
1000 t extrusion press	Dead cycle time	≤ 18 s	Stopwatch over 10 consecutive cycles
1000 t extrusion press	Main ram pressure at rated force	Per spec, held stable	Calibrated gauge in frame, sustained test
Aging oven	Temperature uniformity at setpoint	±3–5°C across chamber	Multi-point thermocouple survey, logged
Billet furnace	Heat-up from cold; billet temperature spread	2–4 h; per spec	Furnace log + contact pyrometer on billets
Packing line	Line speed with your profiles	60 m/min	Timed run with buyer's material
Any machine	Safety functions	All operative	Each interlock and e-stop triggered and logged

The exact values come from your contract specification — the point of the table is the structure: criterion, number, method. A protocol whose test lines read "machine runs normally — OK" is a rubber stamp, not a test.

One more argument for buying a complete extrusion line from a single manufacturer: one FAT covers the whole train, including the interfaces between machines — the billet furnace synchronised to the press cycle, the puller to the extrusion speed. With three suppliers you get three FATs and nobody accountable for the seams.

In person, live video, or a hired inspector?

Honest answer: it depends on what is at stake, and the industry's dirty secret is that most buyers choose by inertia rather than by risk.

Witnessing in person is the strongest option. You see the factory itself — order, tooling, other machines in build, the things a camera is never pointed at. You can ask for any test to be repeated, put your own instruments on the machine, and have the punch-list conversation face to face, where it is hardest to stonewall. Budget realistically €1,500–2,500 per person for a 3–4 day trip to a Chinese industrial region (flights, hotel, local transport); check current visa rules for your passport before booking, as China has run visa-free schemes for many European nationals in recent years. Against a press in the €55–85k class — let alone a €210–540k line — that is between 1% and 3% of the purchase, protecting the other 97%. For a first transaction with a given factory, for any press, or for any complete line, we consider in-person witnessing non-negotiable.

Live video FAT is a legitimate tool at the right scale — for a €8–25k anodising rectifier, a packing machine or a repeat purchase of a machine type you already run, flying may genuinely not pay. But a video FAT is only worth anything under rules agreed in advance:

• Live, not pre-recorded — you name the date, the stream is continuous, and you direct the camera: "show the nameplate again, closer".
• Serial numbers and instrument readings on camera, not read out loud by the operator.
• Uninterrupted runs — a stream that conveniently drops during the load test is a failed test, and the protocol should say so.
• The full recording is handed over and referenced in the signed protocol.

A hired third-party inspector is the middle path: an inspection engineer from an international or specialist agency witnesses the FAT against your protocol and reports with photos and measurements. Indicative cost is a few hundred euros per inspector-day plus travel — typically well under €2,000 for a two-day machinery FAT. The limitation is that a generalist inspector verifies what is written down but rarely knows extrusion machinery well enough to catch what the protocol missed. The protocol has to be good, because the inspector will not improve it.

In our own transactions the default is simple: we or our representative attend the FAT, the client is welcome to join, and the client receives the signed protocol and the complete footage either way. The machine goes to port only after FAT acceptance — that sequencing, described on our how-we-work page, is the whole point.

Payment milestones: the FAT is only as strong as the money behind it

A FAT with no payment consequences is a suggestion. The factory's incentive to fix findings quickly is exactly proportional to the money that moves on acceptance — so the payment schedule and the FAT clause are one mechanism, not two.

The standard opening position of most Chinese machinery factories is 30% deposit with order, 70% before shipment. Note what that structure does: by the time you learn anything about the real machine, 100% of the price is committed. The structure worth negotiating ties the largest tranche to the signed FAT protocol:

Milestone	Typical share	What it buys you
Deposit at contract signing	30%	Production slot; factory buys materials and major components
After FAT acceptance, before shipment	55–65%	The machine has demonstrably met its criteria before the bulk of your money moves
After delivery / commissioning	5–15% retention	Leverage for punch-list items, documentation gaps and commissioning support

Three honest observations from practice. First, factories accept FAT-linked payment far more readily than they accept retention — a 10–15% holdback after delivery is a hard negotiation, and 5–10% is the realistic landing zone; anyone promising you 30% retention from a Chinese factory has not bought many machines there. Second, for larger transactions a letter of credit with the FAT certificate among the negotiation documents is a clean solution, but it carries bank costs and many mid-size factories dislike it; weigh it for line-scale purchases, not for a single auxiliary machine. Third, and most important: never let the pre-FAT payments exceed the amount you could afford to walk away from. That is the true function of the 30/60-plus-retention structure — it keeps the decision to reject a failed machine economically possible, which is precisely what makes rejection unnecessary.

Write the mechanism into the contract explicitly: shipment occurs only after the FAT protocol is signed by both parties, and the pre-shipment payment falls due only on that signature. Not "after successful testing" — after the signed protocol.

When the FAT finds deviations — and it will

A FAT on a machine with thousands of parts that finds nothing at all is not a good sign; it is a sign nobody looked. The purpose of the FAT is not a perfect scorecard — it is to find problems while they are still cheap. What matters is having a deviation procedure agreed before anyone is standing on the factory floor arguing. We use three categories:

• Category A — blockers. Missed performance criteria, failed safety functions, wrong or substituted major components. The machine does not ship. The factory fixes at its own cost, and the affected tests are repeated — with the retest, including your repeat attendance costs if you have negotiated it, on the factory's account. No exceptions, and above all no "we will fix it at your site": a promise that travels 9,000 km by sea arrives lighter than it left.
• Category B — must fix before shipment, no retest visit needed. Minor defects that do not touch performance or safety: a damaged guard, mislabelled cabling, a missing manual chapter. Fixed before packing, verified by photo and video, listed in the protocol.
• Category C — accepted deviations. Cosmetic issues or agreed substitutions you choose to live with — in writing, with a stated remedy: a price adjustment, additional spare parts, or a fix during commissioning with a named deadline. Undocumented tolerance today is a warranty argument next year.

Time is the honest cost here. A clean FAT plus shipment preparation takes 1–2 weeks; a category A finding with rework and retest can add several more. Plan that buffer into your project schedule from the start — on top of the 18–28 weeks of production and 6–8 weeks of sea freight typical for press-class machinery — rather than letting schedule pressure talk you into shipping an unaccepted machine. Every bad machinery import story we know contains the sentence "we were in a hurry, so we shipped it anyway".

Our FAT protocol template — yours for the asking

We maintain FAT protocol templates for the machine types we sell — presses, furnaces and ovens, finishing and packing equipment — built from the structure above and adapted to each contract's specification. If you are buying through us, the protocol, the FAT attendance and the footage are simply part of the process. If you are buying directly from a factory you found yourself, we will still send you the template on request — write to us and mention the machine type. We would rather the market learns to test properly than watch another buyer wire 70% against a marketing video.

And if you are still upstream of all this — comparing suppliers and budgets — start with our guide to what a complete extrusion line really costs and how to start an extrusion plant. The FAT is the moment your leverage peaks; the contract you sign months earlier decides how much leverage there is to use.