A factory acceptance test (FAT) is commonly treated as the final milestone before equipment is released for shipment, but in practical terms it is better understood as a structured observation period. The FAT gives the buyer, the controls engineer, and the maintenance team a rare opportunity to watch a system run under controlled conditions, question its behavior, and identify issues while changes are still relatively inexpensive. For warehouse operators, the value of a well-executed FAT lies less in a formal pass/fail signature and more in the quality of the evidence gathered. Early warning signs seen at this stage, if recognized and documented, can prevent costly disruptions during commissioning and the first months of production. This article outlines the key inspection points at which those signs appear, the methods for collecting reliable evidence, and the boundaries that determine when a finding should stop a shipment.
Purpose and Scope of Factory Acceptance Testing #
The stated purpose of a FAT is to verify that a system meets the agreed functional and performance criteria before it leaves the supplier’s facility. In practice, that means three separate activities are being carried out at once. First, the system is compared against its design specification: physical dimensions, component types, wiring standards, and software revisions. Second, the system is run through normal operating cycles to demonstrate that it can move product, communicate with controls, and respond to commands as intended. Third, the system is stressed to a limited degree to expose weaknesses in build quality, alignment, or programming logic.
It is important to understand what a FAT is not intended to do. It is not a reliability test that proves the system will run flawlessly for years. It is not a substitute for commissioning, and it is not a certification that every conceivable fault condition has been handled correctly. The FAT is a snapshot of a system at a particular moment, under a limited set of conditions, with a specific configuration. The interpretation of that snapshot requires a clear understanding of the boundary between what was tested and what was only simulated.
That boundary becomes the basis for the early warning signs discussed later in this article. An anomaly observed at FAT may be a minor adjustment issue, a design flaw, or a completely benign behavior that simply looks different under artificial conditions. The observer’s challenge is to determine which category the anomaly falls into before the machine is packed into a shipping container.
Operating Context: FAT vs. the Live Warehouse #
Before evaluating any observed behavior, it is essential to define the ways in which the FAT environment differs from the eventual operating environment. These differences shape both the severity of findings and the interpretation of test results.
- Physical layout: At the FAT, conveyors are often shorter, merges may be fewer, and accumulation zones may be abbreviated. A system that handles 100 cartons per minute on a 20-meter test track may behave differently on a 200-meter layout with multiple infeed points.
- Load characteristics: Test product is frequently clean, uniform in size, and within the nominal specification. Real product includes damaged cases, shrink-wrap tails, overfilled boxes, and unusual center-of-gravity positions.
- Control interaction: At the FAT, the system may be running standalone or connected to a simplified controller. The live warehouse will impose a higher-level warehouse control system (WCS) or warehouse execution system (WES) with real-time decisions, order batching, and fault recovery sequences.
- Operational rhythm: The FAT typically runs in repeat cycles. The live environment introduces random arrivals, operator intervention, shift changes, and maintenance pauses that expose different failure modes.
Against this background, an early warning sign at FAT is not necessarily a deviation from nominal performance. More often, it is a deviation from expected behavior under the test conditions themselves. For example, a conveyor that overheats on a 30-minute continuous run at one-tenth of its rated load is not showing a normal phenomenon that will disappear at full load. It is showing a characteristic that will likely worsen when load and continuous runtime increase.
Mechanical Inspection Points #
Mechanical inspection is the starting point of any meaningful FAT, and it should be done before the system is powered up. A component that is poorly aligned or inadequately mounted at the factory will rarely fix itself in the field. The following inspection points merit particular attention.
Belt and Chain Tension #
Check belt tension with the drive stopped. A belt that appears taut but has a slight bow in the slack side may be within tolerance for a short test run but may slip under sustained load. Chain tension should be checked along the longest unsupported span. Also listen for a slapping sound during acceleration and deceleration, as this can indicate excessive play in the chain or a misaligned sprocket.
Roller and Idler Alignment #
Misaligned rollers create a lateral force on the belt or product. Track the belt through a full rotation and observe whether it drifts toward one side. A small drift that returns to center at the end of the cycle is common. A drift that increases with speed, or that causes the belt edge to contact the frame, is an early warning sign of frame twist or roller crowning issues.
Gearbox and Motor Temperature #
Infrared temperature readings on gearboxes and motor housings are useful, but they must be interpreted with knowledge of the duty cycle. A motor that reaches 60°C after a 10-minute idle run may be normal if the thermal mass is small. A gearbox that rises several degrees per minute under light load is a different matter. Take temperature readings at regular intervals and note the rate of rise, not just the final value.
Bearing Noise and Vibration #
Bearing noise is best detected with the drive disengaged and the conveyor manually rotated, if possible. A grinding or intermittent click that appears only under load should be documented and investigated before shipment. Vibration measurement is valuable, but for a single conveyor it may be overkill. A hand-held vibration pen can provide comparative readings across identical conveyor sections, which is often more useful than absolute values.
Control System and Software Verification #
Software verification at FAT is frequently limited to watching the system complete a cycle and noting whether alarms appear. This is insufficient. The control system must be exercised in ways that expose its internal assumptions about the physical world.
Input/Output Mapping and Addressing #
Verify that every sensor used in the control logic corresponds to a physical position that matches the mechanical drawing. Common discrepancies include photoeyes placed on the wrong side of a stop gate, or signals split across multiple input cards in a way that creates race conditions. A practical method is to request a full I/O map from the supplier and then test each sensor by covering or interrupting it while watching the software tag change state.
Fault Recovery Logic #
Fault recovery is the most frequently under-tested part of a FAT. The system may start and run perfectly, but how does it behave when a product stalls at a transfer point, or when a sensor becomes dirty, or when a stop gate fails to confirm its position? Ask the supplier to force these faults and observe the recovery sequence. An early warning sign is a recovery sequence that requires a manual reset on a condition that should be automatically recoverable, such as a momentary sensor flicker.
Data and Statistics Generation #
If the system is expected to produce throughput statistics, cycle counts, or error rates, verify that this data is generated correctly and is timestamped. Many commissioning delays result from discovering only after installation that the SCADA or WCS interface is not producing usable data. At FAT, identify which reports are populated during the test run and check the values against manually counted cycles.
Electrical and Safety Circuit Inspection #
Electrical inspection at FAT is about more than confirming that connectors are tight. It is about verifying that the installation matches the design intent and that safety systems are functional. Observers should not attempt to override or bypass any safety device during the FAT. Site procedures, lockout requirements, OEM documentation and competent engineering judgment take priority for any interaction with safety circuits. This is not a suggestion; it is a boundary.
With that stated, there are legitimate inspection activities that do not involve defeating safety. Check cable routing for mechanical damage, verify that pneumatic lines are not in contact with sharp edges, and confirm that power supply ratings match the connected load. Also request the safety relay configuration and time-based logic. A safety circuit that requires a manual reset after a door switch opens is often correct, but a circuit that does so after a momentary voltage dip may be a design issue that will cause nuisance shutdowns in the warehouse.
Emergency Stop Functionality #
Every emergency stop (E-stop) should be tested from each location, and the test should confirm that the affected zone matches the design. Test one E-stop at a time first, then test multiple E-stops simultaneously. A system that does not shut down correctly, or that resets inconsistently when the E-stop is released, should be treated as a critical finding. The FAT is not the place to sign off on a safety system that has not been fully exercised.
Early Warning Signs in Performance Data #
Performance data collected during the FAT is the foundation for comparing factory behavior with site behavior later. The following table presents common observations, the likely underlying causes, and the appropriate action before shipment.
| Observation at FAT | Likely Underlying Cause | Action Before Shipment |
|---|---|---|
| Intermittent product jams at the same transfer point | Sensor height incorrect, timing window too tight, or transfer belt speed mismatch | Adjust sensor position or timing; document the change and re-run 100 cycles |
| Conveyor tracking drifts only after 30 minutes of continuous run | Thermal expansion of frame, belt stretch, or initial tension too high on one side | Measure frame temperature and belt tension; repeat the long-run test after adjustment |
| Gearbox temperature rises rapidly even under low load | Incorrect oil level, wrong oil grade, or over-tightened bearing | Check oil specification against build drawing; verify level; investigate before signing |
| Full-load test produces a higher motor current than nameplate rating | Conveyor misalignment, excessive belt tension, or undersized motor | Measure current on all phases; compare with motor curve; request engineering review |
| Fault recovery takes longer than the programmed alarm timeout | Software sequence delays, or a sensor that does not reset in time | Trace the PLC sequence; identify whether the delay is safe; document it as a known operational limitation |
| Same product position triggers a jam on every tenth cycle | Accumulation effect, or a speed variation that is not visible on the HMI | Record high-speed video; compare cycle times across multiple runs to identify the variable |
The key principle in using this table is that the observation is not the conclusion. A single jam at a transfer point may be a one-off caused by a slightly damaged test carton. Ten jams at the same point in twenty cycles is a repeatable event that requires root cause analysis before shipment.
Evidence Collection and Documentation #
Evidence collected at the FAT is only valuable if it can be referenced later. Without a clear, time-stamped record, a minor observation made by an engineer at the factory has no value at the commissioning site. The following documentation practices should be standard for every FAT.
- Use a single test log that records the test name, the system version, the operator, and the start and end times for each test.
- Record actual numeric values, not just pass/fail checkboxes. For example, instead of writing “conveyor speed OK,” write “conveyor speed measured at 1.22 m/s against target of 1.20 m/s.”
- Take short video clips of any unusual behavior, including flickering indicators, intermittent noise, or product contact with the frame.
- Photograph the motor nameplates, gearbox tags, and sensor model numbers so that maintenance teams can order spares with confidence.
- Capture a full set of HMI screens, including alarm history, after each test run. This provides a snapshot of system state that the PLC program download may not contain.
- Retain the exact software revision and site parameters used during the FAT. When the software is updated later, the FAT results can be compared against the new revision to determine whether behavior changed.
One of the most overlooked pieces of evidence is the set of operator observations during the manual mode. How easy is it to recover a jam? Does the stop gate release cleanly? Does the interface respond instantly to a command? These qualitative observations often reveal issues that will affect production floor staff more than any data point.
Common Interpretation Errors #
Many FAT discussions are undermined by the same types of interpretation errors. Being aware of these errors helps the observer distinguish between a real issue and a misreading of the situation.
The first error is treating the FAT pass as a guarantee of long-term reliability. A machine that runs flawlessly for four hours at the factory may still have a component that is near the end of its design life, or a software bug that only appears after 1,000 cycles. The FAT should be used to establish a baseline, not to conclude that future failures are impossible.
The second error is adjusting test parameters to make the system pass. For example, if the test protocol specifies a throughput of 30 cartons per minute, and the system consistently achieves only 27, the temptation is to reduce the carton size or increase the gap between cartons. This practice erases the value of the test. The correct response is to retain the original protocol, record the observed throughput, and escalate the gap to the engineering team for a formal decision.
The third error is ignoring alarms and warnings during the FAT. Some teams assume that an alarm that appears once during the test run is a one-off event that will not happen in production. In reality, an alarm that appears during a FAT is statistically more significant than one that does not appear, because the FAT runs for a much shorter period than the eventual production schedule. Every alarm should be logged, investigated, and either resolved or documented as a known condition.
The fourth error is failing to test fault recovery in the same way as normal operation. A system that performs perfectly in steady state may have a completely broken recovery sequence that only becomes apparent when a jam is deliberately induced. The operator’s ability to resolve a jam quickly and restart the line is part of the system’s operational performance, and it must be tested.
The fifth error is treating the FAT as a purely technical exercise when it is in fact a contractually significant milestone. The acceptance signature is often tied to payment milestones and shipment release. As such, the decision to accept or reject equipment must be made with reference to the original specification, not to the subjective impressions of the engineers present.
Maintenance Implications of FAT Findings #
FAT findings are not purely transactional events that end when the machine is shipped. They have direct consequences for the maintenance strategy that will be applied once the system is in the warehouse.
If a FAT reveals that a particular drive belt degrades quickly, the maintenance team should order additional spares and plan for more frequent inspection. If the FAT shows that a sensor requires constant cleaning due to its position, a change should be requested now, not after the system is installed. If the FAT demonstrates that the safety reset procedure is complex, the maintenance documentation and training materials must reflect that complexity, and the maintenance team should be involved in the FAT to learn those procedures firsthand.
Involvement of maintenance staff in the FAT is one of the most cost-effective ways to improve the eventual reliability of the system. When a maintainer observes that a bearing is difficult to reach, or that the lubrication points are awkwardly placed, that feedback can be transmitted to the factory while the machine is still in build. After installation, the same feedback would require downtime, special tools, or a permanent modification order.
Additionally, the FAT provides an opportunity to validate the spare parts list against the actual components on the machine. It is not uncommon for the supplier’s spare parts list to contain a different model of sensor or belt than the one actually installed. During the FAT, the mechanic should verify that each part number on the list corresponds to a visible component on the machine.
Decision Boundaries and Escalation #
When an anomaly is found at the FAT, a decision must be made about whether to ship, to delay, or to conditionally ship. This decision should be guided by the severity of the finding and the confidence that a fix can be completed without further testing.
Minor cosmetic issues, such as paint scratches or paint overspray, do not typically delay shipment. They are recorded and resolved through a punch list. Functional issues that affect safety or throughput are more serious. A photoeye mounted at the wrong height, a motor running at an excessive temperature, or a safety circuit that does not reset correctly should hold the machine at the factory until corrected and retested.
There is also the conditional shipment, in which the machine is shipped with a documented, agreed-upon list of items to be resolved at the site. This option should be used conservatively. A condition that can be fixed with a simple adjustment on site, such as a sensor bracket position, is a reasonable candidate for conditional shipment. A condition that affects the core control logic or the mechanical integrity of the machine is not.
At all times, the decision to accept, reject, or conditionally ship must be made by the authorized parties under the purchase contract, with the involvement of qualified engineering staff. No observer or inspector should make unilateral decisions on behalf of the operator. The FAT is a shared learning experience, but the power to release equipment is contractual.
Key Takeaways #
- A factory acceptance test is a controlled observation period, not a guarantee
Related Pearl Gateway Guides #
