A custom testing machine can look complete and still be unready for shipment. The frame may be painted, the actuator may move and the software may display a curve, yet the equipment can still fail at the customer’s laboratory because the approved specimen does not fit, the measurement range is wrong, the report omits required channels or a safety interlock has not been challenged.

The factory acceptance test, or FAT, is the point where commercial promises are converted into witnessed evidence. It should answer a practical question: if this exact machine is installed under the agreed site conditions, does it have the hardware, control logic, measurement performance, safety functions and documentation needed to run the intended test programme?

Freeze the Acceptance Baseline Before the FAT Date

The most common FAT failure begins before anyone enters the workshop: the buyer and supplier are testing against different expectations. One team refers to the quotation, another uses the latest email and the software engineer works from an older method file. The first FAT document should therefore be a controlled acceptance baseline.

That baseline should identify the purchase order, approved technical agreement, current machine layout, fixture drawings, electrical and pneumatic requirements, sensor list, software functions, report fields and applicable test clauses. Every item needs an acceptance method: visual inspection, dimensional check, document review, functional challenge or witnessed test run.

Classify each requirement as mandatory before shipment, acceptable for site completion or excluded from the current supply. This prevents minor cosmetic comments from blocking shipment while a critical measurement or safety gap remains hidden in a general punch list.

Inspect the Machine Build and Supplied Configuration

Begin with identity and completeness. Confirm machine model, serial identification, electrical rating, actuator type, controller, load cells, displacement sensors, fixtures, guards, accessories, tools and spare parts against the approved list. A photo record should capture nameplates, sensor serial numbers, fixture interfaces and the final cabinet layout.

Then check the physical envelope. Measure working width, actuator travel, fixture adjustment range, specimen access and service clearance. Install the largest and smallest agreed specimens when possible. A fixture that works only with the supplier’s demonstration sample is not evidence that the customer’s product range will fit.

Inspect workmanship without turning the FAT into a cosmetic audit. Focus on issues that affect function and serviceability: loose wiring, unlabelled terminals, unsupported hoses, inaccessible fasteners, sharp edges, leakage, cable interference, guard alignment and maintenance access.

Verify the Complete Measurement Chain

A calibration certificate for a load cell is useful, but it does not prove the installed machine measurement chain. The FAT should trace the signal from the sensor through conditioning, controller, software display, stored data and final report. Confirm sensor identity, range, units, resolution, zeroing behaviour, overload protection and calibration status.

Use suitable reference equipment to check representative points across the intended working range. The goal is not to invent a universal tolerance at the FAT; it is to demonstrate the tolerance and range already agreed for the project. Record the reference device, its status, environmental conditions, commanded value, indicated value and result.

ISO/IEC 17025 describes competence, impartiality and consistent operation for testing and calibration laboratories. ISO 10012:2026 addresses measurement management systems intended to support confidence in measurement validity and reliability. These principles support a practical FAT habit: preserve traceable evidence for the measurements that determine acceptance, instead of relying on a screen demonstration.

Run a Representative Test Method, Not an Empty Demonstration

The machine should run at least one representative method using the agreed specimen or a technically justified substitute. A meaningful run includes mounting, zeroing, preload, control-mode transition, waveform, target range, frequency, cycle counting, stop conditions and result export.

Choose a test that exercises the risky parts of the design. For a fatigue system, that may mean low-range force control, actuator reversal, fixture alignment and long-cycle stability. For a durability machine, it may mean synchronized channels and failure detection. For a dimensional or static system, it may mean repeatable positioning and report calculations.

Do not require the entire production test duration unless that was agreed. A shorter witnessed sequence can still prove method logic when it includes the critical stages. Long unattended endurance runs can be listed separately with their duration, monitoring method and acceptance evidence.

Challenge Software, Permissions and Data Integrity

Software acceptance should follow the operator’s real workflow. Create or load a method, enter specimen information, set limits, start and pause the test, trigger a stop condition, resume when permitted, complete the run and generate the report. Confirm that units, decimal places, channel names and calculations remain consistent.

Test user permissions and configuration protection. Operators should not accidentally overwrite engineering limits, while authorized users need a controlled way to edit methods. Confirm backup, restore and export procedures, along with the supplied software version and license arrangement.

Save a complete FAT data package rather than screenshots alone. Include the method file, raw data, processed result, final report and software version. This package becomes the reference when the same test is repeated during site acceptance.

Review Safety Functions, Documentation and Training

Challenge each agreed safety function: emergency stop, guard interlock, travel limit, overload response, loss of air or power, abnormal sensor signal and safe restart. IEC 61010-1 covers general safety requirements for electrical test, measurement, control and laboratory equipment within its scope. The actual machine may also require other standards or local requirements, so the FAT should follow the approved project safety plan.

Documentation should match the delivered configuration, not a generic machine family. Review the operation manual, maintenance schedule, electrical drawings, pneumatic or hydraulic diagrams, component list, calibration documents, software instructions, troubleshooting guide and spare-parts list.

Training should include normal operation and recovery from predictable faults. Record who attended, which functions were demonstrated and which topics remain for site training. A signed attendance sheet is not enough if the operator never mounted a specimen or recovered from an interlock.

Close Findings with Clear Shipment-Release Rules

Every FAT finding should have a unique number, description, evidence, severity, owner, target date and closure method. Separate shipment blockers from items that can be completed before installation. If a deviation is accepted, record who approved it and whether drawings, software or manuals must be updated.

Shipment release should require more than a signature at the end of the visit. Confirm that blocker items are closed, final backups are secured, transport locks and protective packaging are defined, loose fixtures are identified and the site-preparation data has been issued. Record which checks will be repeated during site acceptance.

A well-run FAT does not guarantee that installation will be effortless, but it removes avoidable uncertainty. It gives the buyer a defensible acceptance record and gives the supplier a clean technical baseline for commissioning and after-sales support.