Not sure which one you're looking at?
That's the conversation we have for free. Tell us what your line does today and what you need it to do — we'll tell you honestly what it takes, even when the honest answer is the smaller project.
Insights · Build · September 14, 2026 · 8 min read
Factory Acceptance Testing: The Thirty Minutes That Decide Whether Your Machine Works
FAT is where a machine either proves itself or gets fixed cheaply. Skip it, or run it badly, and you'll do the same work on your production floor at ten times the cost.
A machine that fails in the builder's shop costs a day and a bracket. The same machine failing on your floor costs a shift, a rigging crew, an engineer's flight, and the trust of the people who have to run it. Factory Acceptance Testing — running the machine at the builder's facility, on your product, against agreed criteria, before it ships — is the cheapest insurance in the project. It's also the step most often rushed, skipped, or run as a demo instead of a test.
What a FAT is, and what it isn't
A FAT is a test with pass/fail criteria that were written before the machine was designed. It's not a walkthrough, a demonstration, or a chance for the client to "see how it's coming along." Those are useful too — we invite clients to design reviews — but they're not acceptance.
The criteria are the ones from the proposal: rate, duration, quality tolerance, product(s), and the specific conditions. If the proposal didn't have them, the FAT can't have them either, and what you'll get is a machine that "seems to work" — which is a phrase that has preceded every commissioning disaster we've ever seen.
The structure of a FAT that means something
- Dry run — mechanical and controls. Every motion, every sensor, every interlock. E-stops actually stop things. Guards actually prevent access. Fault messages are readable by a human. This is boring and it catches most of the wiring mistakes.
- Product run — short. Your product, at rate, for long enough to see it behave. Jams, transfers, fill consistency, seal quality. Fix, adjust, repeat.
- Product run — sustained. The acceptance run itself. Thirty continuous minutes at the design rate is a common minimum for a single machine; longer for lines. Output counted from the PLC, not a stopwatch. Quality sampled per the agreed method.
- Changeover, if applicable. Full change to a second product, timed, by someone who isn't the builder's engineer.
- Punch list. Everything that failed or was marginal, with an owner and a fix-by date. Nothing ships against an open safety item.
Two rules make this real. First, the client runs the sustained test, or at least stands at the HMI while it runs. The builder's engineer knows the machine's moods and compensates without noticing. Your operator doesn't, and that's the point. Second, the numbers get written down as they happen — counts, times, sample results — in a form both parties sign. Memory is generous.
What to bring
Real product, in production quantity — not the perfect batch. Your packaging, your actual containers, your labels if they matter to the machine. Your operator, if you can spare them for a day; the training happens by accident and it's the best kind. And the proposal, with the acceptance criteria highlighted. You'd be surprised how often it stays in the folder.
What "pass with punch list" means
Machines rarely pass a FAT clean, and that's fine. The question is what kind of items are open. A cosmetic guard gap or an HMI label typo shouldn't hold a shipment. A rate 15% under criteria, a seal-quality failure, or any safety-circuit issue should — those are the ones that become "we'll fix it on site" and then don't get fixed. Our rule: nothing ships against an open performance or safety item. Everything else goes on the list with a name next to it.
Why builders skip it, and why you shouldn't let them
A real FAT costs the builder a day or two of engineering time, product handling, and the risk of finding something. A weak builder would rather ship and fix on site, where the client is invested and the pressure to accept is high. A good builder would rather find it in their own shop, where the fix is cheap and their reputation isn't on your floor. Which one you're working with is usually visible in how they respond when you ask to run the sustained test yourself.
FAT and SAT are different tests
Site Acceptance Testing — the same criteria, on your floor, integrated with your line — comes after install and is the real finish line. But SAT should confirm what FAT already proved. If the machine passed a real FAT and fails SAT, the problem is almost always an interface: air pressure, upstream feed, floor level, a product difference between the sample and the line. That's a much shorter list to debug than "the machine doesn't work." A skipped FAT collapses both lists into one, on your floor, on the clock.
Standing invitation. Every machine we build runs a FAT in our Arvada shop before it ships, against the criteria in your proposal, and you're invited to run the test. We photograph and document every one — partly for the record, partly because a machine running clean in the shop is the best evidence we have that we do what we say.