11. 12, 2026
A pre-shipment acceptance test is the structured verification a machine tool undergoes at the builder's plant before it is dismantled, packed and shipped, and it is the buyer's last practical opportunity to prove that the machine meets the contract before it is thousands of kilometres away and surrounded by installation costs. Compared with a final site acceptance test, the factory test can be run on a stable foundation, with the builder's metrology equipment, technicians and spare parts on hand, and with the option of correcting a fault before the machine leaves. I evaluate the process by what is written into the contract, which measurements are recorded, whether a representative workpiece is actually cut, how the buyer can witness the test, and what happens when a result falls outside the agreed limit.
A factory acceptance test is only enforceable if the test method, instrument, acceptance limit and remedy are written into the purchase contract before the order is placed.
Static geometry and positioning records are necessary but not sufficient; a cutting test on a representative workpiece reveals thermal drift, chatter, tooling behaviour and control setup that no-load tests cannot.
Witnessing can be done remotely, but only with an agreed protocol that fixes camera angles, instrument displays, part serial numbers, measurement method and a recording that the buyer keeps.
Vibration, spindle runout, circular interpolation and tool change repeatability should be recorded as numbers, not described as "normal".
Preservation, packing, lifting points and container loading are part of the acceptance scope, because damage in transit is easy to prevent and expensive to resolve.
By the time a machine reaches the buyer's floor, three things have usually happened. The foundation has been built to suit whatever arrived. The rigging and installation costs have been paid. And the machine is in a building where the builder's metrology equipment and specialists are not present.
A fault found at the factory is a scheduling problem. The same fault found after installation is a commercial dispute involving freight, technicians, travel visas, downtime and often a temporary repair that becomes permanent.
This asymmetry is why I treat the factory test as the primary acceptance event and the site test as confirmation that nothing changed in transit and that the foundation is adequate. The contract should reflect that priority, including who pays for a repeat test if the first one fails.
The most common failure in acceptance testing is not a machine that fails; it is a machine that passes an undefined test. If the contract says "accuracy according to standard" without naming the standard, the measurement method and the limit, then almost any result can be argued as compliant.
Before signing, I want the following attached to the order as technical annexes.
A configuration list naming the exact model, controller and software version, spindle motor, spindle taper, axis drive and motor, guideway type, ballscrew specification, bearing arrangement, coolant system, chip conveyor, tool magazine capacity, and every option with its part number where available.
An accuracy specification sheet stating positioning accuracy and repeatability for each linear axis and, on machines with rotary axes, for each rotary axis, with the test method named.
A test procedure describing the sequence of checks, the instruments used and their calibration status, the environmental conditions, and the warm-up routine.
A cutting test specification defining the test part drawing, material, tooling, cutting parameters, the features to be measured, and the acceptance limits for dimensions, roundness, flatness, parallelism, squareness and surface finish.
A remedy clause stating what happens if a result is outside the limit: rework at the factory, replacement of a component, a price adjustment, or cancellation, and who bears the cost of a repeat test.
A documentation package list covering the final inspection report, the accuracy records, the hydraulic and pneumatic diagrams, the electrical schematic, the ladder and parameter backup, the operation and maintenance manuals, the spare parts list and the warranty certificate.
Without these annexes, the acceptance meeting becomes a negotiation about what "good" means.
The order of the checks matters, because some tests disturb the machine and others need a settled machine.
Unpacking and visual inspection. Before power is applied, check for transit damage, oil leaks, loose components, corroded surfaces, and whether the accessories listed in the packing list are all present.
Level and foundation check. Confirm that the machine is level within the agreed tolerance on its own foundation or on temporary supports. A machine that is out of level will fail geometry tests for reasons unrelated to manufacturing quality.
Warm-up. Run the spindle and axes through the specified warm-up cycle. Thermal stability is a precondition for meaningful accuracy measurement, not an optional step.
Geometric accuracy. Measure straightness, squareness, parallelism, spindle axis to table relationship, and axis-to-axis perpendicularity using the method named in the contract, typically a recognised procedure such as ISO 230-1 for machine tool geometric tests.
Positioning accuracy and repeatability. Measure each linear axis at defined intervals in both directions, and each rotary axis at defined angular positions, using a laser interferometer or an equivalent calibrated instrument, following a procedure such as ISO 230-2.
Circular interpolation and servo behaviour. A circular test, for example with a ball bar following ISO 230-4, reveals servo mismatch, backlash, stick-slip and geometry errors that linear tests do not show. This is one of the most informative tests in the whole sequence and it is often skipped in low-cost purchases.
Spindle checks. Record spindle runout at the taper and at a test mandrel at a defined distance, spindle speed accuracy across the range, spindle orientation repeatability if the machine performs tapping or tool changing, and spindle thermal growth over a defined period.
Tool change and magazine performance. Record tool change time, repeatability of the tool change position, and behaviour with the heaviest and longest tools in the intended set. A magazine that works with an empty test tool and fails with a boring bar is not ready.
Functional and safety checks. Verify every interlock, door switch, emergency stop, overtravel limit, coolant and lubrication flow, chip conveyor direction, hydraulic pressure, pneumatic pressure, and alarm behaviour.
Cutting test. Machine the agreed test part and measure it.
Preservation and packing. After the test, confirm rust prevention, axis locking, spindle protection, desiccant, packing method, container loading plan and lifting points.
A machine can pass every no-load test and still fail in production. The cutting test is where thermal drift, structural rigidity, control tuning, coolant delivery and tooling behaviour appear together.
I specify the test part to include the operations the machine will actually perform: a face that must be flat, a bore that must be round and on position, a shoulder that must be square to the bore, a thread or a groove, and if applicable an interpolated contour. The material should match the intended production material, or at least be stated, because a test in aluminium proves very little about a machine destined for steel.
The record should include the drawing with numbered features, the measured values for each feature, the instrument used, the operator, the date, and the ambient temperature. Surface finish should be recorded as a value, not a comment.
Two additional details turn a cutting test into useful evidence. First, cut at least two parts with a deliberate gap between them, and record whether the second part differs from the first. That shows thermal behaviour under realistic conditions. Second, run a heavier roughing pass than the finishing pass and record spindle load, vibration and any audible or visible chatter. Buyers who only see a finishing cut on a light part learn almost nothing about the machine's capacity.
Travel is expensive and often impractical. Remote witnessing works, but only if the protocol is agreed in advance.
I ask for a fixed camera position that shows the machine display, the part and the instrument at the same time, or a sequence of shots that clearly links them. The machine serial number and the part serial number should be visible at the start of the recording. Instrument readings should be shown on screen rather than reported verbally. The full recording should be supplied to the buyer and kept as a contractual record.
For the laser interferometer and ball bar tests, the raw data file should be supplied along with the report, because a summary graph can hide an outlier. For the cutting test, photographs of the part at each stage, plus the measurement sheet, are more useful than a final photograph of a finished part.
Where a third-party inspection is used, the inspector's scope must be defined. An inspector who checks packing and quantity is not performing an accuracy acceptance test.
"The machine was accurate at the factory." Prevented by a recorded test with instruments, dates and limits, signed by both parties, before shipment.
"The foundation is the problem." Prevented by specifying foundation requirements in the contract, including concrete strength, thickness, anchor layout, isolation if required, and the level tolerance at installation, plus a site test procedure that separates foundation effects from machine effects.
"That feature is outside the scope." Prevented by numbering the features on the test part drawing and listing them in the acceptance schedule.
"The measurement method is different." Prevented by naming the standard, the instrument class, the number of measurement points and the environmental conditions in the annex.
"Transit damage is the buyer's risk." Prevented by photographing preservation and loading, specifying lifting points and container bracing, and agreeing who inspects on arrival and within what period.
"The technician will come next month." Prevented by warranty terms that state response time, who pays travel and accommodation, the spare parts policy, and the escalation path.
The site test should repeat a reduced set of the factory checks. Level and foundation first, then geometric checks for the items most likely to be disturbed in transit, then positioning accuracy on at least the main axes, then a cutting test on a simple but demanding part.
Any difference between site and factory results should be explained before the machine is accepted. If the site result is worse, the question is whether the foundation, the level, the power supply, the compressed air, the ambient temperature or the installation disturbed the machine. Answering that question requires the factory record, which is why the documentation package must arrive with the machine and not months later.
A pre-shipment acceptance test is not a formality and not a courtesy visit. It is the point at which a specification becomes evidence. The buyer who arrives with a defined procedure, a numbered test part, calibrated instruments and a remedy clause leaves with a machine that has been proven; the buyer who arrives to watch a spindle turn leaves with an opinion.
For any supplier, including GSK, the practical preparation is the same: put the configuration, the accuracy values, the test method, the test part, the record format and the remedy in writing before the order, then attend or witness the test and take the records away.
Published company information for GSK states a 100,000-square-metre manufacturing facility, output of approximately 2,500 sets per year, an 18-month warranty, in-house QC testing and multiple certification options. Those statements are a starting point for the acceptance discussion, not a substitute for the recorded results of the specific machine on order.
Inquiry CTA: To build an acceptance schedule around a specific machine, send GSK Equipment the intended workpiece drawings, the critical tolerances, the material, the controller preference and the destination port. The team can supply the model configuration, the accuracy specification, the proposed test part and the inspection record format for review before the order is placed, so the acceptance criteria are agreed while they can still be written into the contract.
Navigation
Tel.: +86 632 5083 879
Mob.: +86 151 6326 5369
Wechat: +86 151 6326 5369
E-mail: info@gskequipment.com
WhatsApp: +86 151 6326 5369