A cleaning machine that passes inspection at the factory and then underperforms on your line failed one thing first: nobody wrote down what “passing” meant. Define the factory acceptance test before the purchase order is signed, and the machine arrives with a record you can re-measure on your own floor.
CHUANGQI runs four acceptance stages on every machine before it ships: a spray-pressure and leak calibration, a temperature-uniformity check across the drying zone, a continuous run of 72 hours, and a cleanliness check on a test coupon. The list below is the one we send to buyers who ask what arrives with the machine, and the wording comes from that pre-shipment protocol.
Key takeaways
- A factory acceptance test is defined before the order, not improvised on shipping day.
- Measurable pass criteria belong in the purchase order, in writing.
- Four stages carry most machine-level risk: pressure, drying uniformity, a 72-hour run, cleanliness.
- Ask for raw records rather than a summary sheet.
- A factory test certifies the machine, not your process.
Factory Acceptance Test vs Site Acceptance Test: Two Different Risks
This article covers the machine-level acceptance test a buyer runs at the supplier’s factory before shipment, and the site test that follows installation. It is written for the procurement or process engineer who signs the acceptance page. The worked example is CHUANGQI’s own pre-shipment protocol for PCBA cleaning machines. No external test threshold is quoted here, because an acceptance number belongs to your product class and your customer requirement.
The factory test removes build risk. The site test removes installation risk. They answer different questions, and one does not cover the other.
| Question | Factory acceptance test | Site acceptance test |
|---|---|---|
| Where it runs | At the machine builder, before crating | On your floor, after installation |
| What it proves | The machine was assembled and wired to specification and the utilities behave | The machine performs on your boards, with your chemistry and your plant air and water |
| What it cannot prove | Performance on a board family the builder has never run | Whether the cause is a build fault or a process setting |
| Evidence you keep | Pressure and leak records, temperature uniformity log, long-run log, cleanliness test result | First-article results on production boards, plus the acceptance route agreed with your customer |
Source: CHUANGQI pre-shipment acceptance protocol for cleaning machines, 2026-10. The comparison describes which risk each test covers, not a pass value for any machine.
What to Write Into the Purchase Order Before the Machine Is Built
A specification that lists only dimensions and voltages is a purchase, not an acceptance plan. Each line below turns into a measurement your supplier has to record, and each one exists because we have seen a shipment become arguable without it.
- Board envelope — the largest and the smallest panel you intend to run, and the tallest component on the assembly. A machine sized for your average board fails at both ends of the range (the smallest panel and the largest panel each need their own check).
- Contamination type — rosin, water-soluble or no-clean flux, and whether solder paste residue or coating overspray also has to come off. Chemistry choice follows this line rather than the other way round.
- Throughput target — boards per hour at a stated panel size. Demand the figure at the panel size you actually build, not at the size that photographs well.
- Utility envelope — plant air pressure, DI water supply, drain, and the electrical configuration at the installation point. A machine that needs a utility you do not have is a change order later.
- Acceptance items and tolerances — the four stages listed further down, each with the tolerance you will accept and the record you expect to receive.
- Documentation package — English operation and maintenance manual, spare-parts list, and the test records in a form your quality team can file.
One limitation belongs on that list, stated plainly. A factory test cannot confirm how the machine will behave on a board family nobody has run through it yet. If your product mix changes every month, the factory test proves the machine was built correctly; it does not prove your process window, and it does not replace a sample run on your own assemblies.
The Four Stages CHUANGQI Runs Before a Machine Ships
These four stages are recorded on every machine we build, and each one closes a different failure path. The order matters: a leak test taken after a long run tells you something a leak test taken before it does not.
| Stage | What it removes | Record you receive |
|---|---|---|
| Spray pressure and leak calibration | Liquid that never arrives at the surfaces the process is meant to wet | Pressure reading per circuit, leak check result |
| Temperature uniformity in the drying zone | Wet pockets that push residue back across a board after rinsing | Temperature map taken across the drying tunnel |
| Continuous run of 72 hours | Faults that only show up after warm-up, not in the first hour of a demo | Run log with every stoppage marked |
| Cleanliness check on a test coupon | A rinse stage that underperforms before the machine leaves the workshop | Coupon extraction result taken before crating |
Source: CHUANGQI acceptance protocol for cleaning machines, 2026-10. Every unit leaves the workshop with four records: spray pressure and leak calibration, a temperature-uniformity log for the drying tunnel, a 72-hour run log, and an extraction result from a test coupon. Tolerance values are read from that machine’s own specification sheet.
Limits of applicability. This checklist covers machine-level acceptance for a cleaning machine bought as a discrete unit, and it is not suitable as the only acceptance route when your customer applies a cleanliness requirement of its own: those criteria stay with your customer, and a machine record cannot stand in for a first-article run on production boards. It also assumes the cleaning step is a separate purchase; where the washer is folded into a larger line, the acceptance milestones belong in that line’s schedule instead. A stock machine can complete the four stages in a shorter window, but the record still has to exist. Skipping it saves administration and returns later as an argument.
Two other pieces on this site sit beside this one. If you would rather prove a machine on your own boards before ordering, the sample cleaning test route sends back a record measured on your assemblies before and after the wash. If the order paperwork and shipping terms are the part that is new to you, importing SMT cleaning machines walks through that side. The platform we most often write acceptance records against is the inline PCBA water cleaning machine CQ-C9610.
Frequently asked questions
What is a factory acceptance test for a cleaning machine?
It is a set of measurements recorded at the builder’s plant before the machine is crated, against criteria you defined in the purchase order. Typical items are a pressure-and-leak check on the spray circuits, a temperature map across the drying zone, a long continuous run, and a cleanliness reading taken from a test coupon. The value sits in the record: a machine that arrives with measurements can be measured again, and a machine that arrives with a signature cannot.
Do we have to be present at the factory acceptance test?
Attendance gives you the logs first-hand, and remote works if the protocol is written down. If you attend, bring your pass criteria and ask for raw logs rather than a finished summary. If you cannot travel, agree in advance on which records will be produced, who signs them, and how a retest is triggered. Spoken agreements are the part that does not survive a change of engineer on either side.
Can we skip the factory acceptance test for a stock model?
You can shorten it, not delete it. A machine built to a design already in production still has to be wired, plumbed and calibrated, and those are the steps where a build fault hides. The four stages above still run on a stock machine; what changes is the schedule, not the need for a record.
What happens if the machine fails a stage?
Fix it and re-run that stage, and keep both results in the file. A recorded failure followed by a passing retest tells your quality team more than a clean sheet with no history, because it shows the supplier noticed and corrected something. If the same stage fails twice on one design, treat it as a design question rather than an assembly question, and raise it before crating rather than after installation.
Next step: send the board envelope, the flux class and the boards per hour you need to hold, and we will return the acceptance items that apply to that configuration with a quote within 24 hours — request a machine acceptance plan.