**Before you release final payment, demand a written, pass/fail acceptance test that runs your part, on your floor, with measured capability data.
Before you release final payment, demand a written, pass/fail acceptance test that runs your part, on your floor, with measured capability data. A supplier demo in a temperature-controlled showroom is not enough; only a test under your own shop conditions proves the machine will make your parts at the agreed tolerance.
To: Shop owners and purchasing engineers, DACH / UK From: Gree CNC technical support, overseas distributor Re: Minimum acceptance testing before final payment
A CNC machine tool is a capital investment. The purchase price is only part of the cost. If the machine cannot hold your tolerances at your required cutting parameters, you lose time, scrap material, and customer trust.
Supplier catalogues list machine geometry and positioning data. Those numbers are measured under controlled conditions. Your workshop may differ: temperature swing, floor stiffness, power quality, and cutting loads all affect real performance.
That is why acceptance testing exists. It is not a formality. It is your last and strongest lever before you hand over the final payment.
In DACH and UK machine tool buying practice, a common structure is:
The exact split is contractual, not a Gree recommendation. The key point is simple: do not sign the final acceptance certificate until the machine has passed the agreed tests. After that, your leverage drops sharply.
A serious CNC acceptance test has four layers. Do not skip any of them if your parts require tolerances tighter than ±0.05 mm.
Each layer needs a pass/fail number. Vague wording such as “machine shall perform to manufacturer’s specification” is not enough. You need a measurable limit, a test method, and a defined measurement point.
The table below gives a starting point for a mid-size 3-axis vertical machining centre. All values are typical industry ranges for a standard-class machine, not promises. The exact figure for your selected Gree model must be confirmed against the current official catalogue and your agreed protocol.
| Test | Method | What it checks | Typical pass criterion (standard class VMC) | Confirm against |
|---|---|---|---|---|
| Linear positioning accuracy | Laser interferometer, ISO 230-2 | Axis positioning deviation and repeatability | Repeatability ≤0.005 mm; systematic deviation ≤0.010 mm over 300 mm | Current Gree catalogue for the model |
| Circular interpolation | Ballbar, 200 mm diameter circle at 1000 mm/min | Backlash, servo mismatch, squareness | Roundness deviation ≤0.015 mm | Machine class, not catalogue alone |
| Spindle runout | Dial indicator on precision test bar | Spindle taper and tool-holding quality | ≤0.005 mm at spindle nose; ≤0.015 mm at 300 mm from nose | Catalogue / ISO 230-7 |
| Cutting capability | Test part, minimum 5 parts measured | Process capability on your real tolerance | Cpk ≥1.33 for critical features | Agreed acceptance protocol |
| Thermal drift | Spindle warm-up cycle, drift at tool tip over 2 hours | Thermal growth and compensation effectiveness | ≤0.025 mm without compensation; ≤0.015 mm with compensation (typical ranges) | Buyer requirement |
Technical terms used above, explained briefly:
Below is a practical method for a medium-sized vertical machining centre, X/Y/Z travel 800/500/500 mm, cutting aluminium 7075 and C45 steel. Adapt the numbers to your part.
Write down the three most difficult features you need to machine. Example:
These three features become your acceptance part features. Do not accept a generic square block with loose tolerances. The machine must prove it can make your geometry.
The purchase order must state:
If the seller will not put this in writing, do not place the order.
Before the machine is shipped, the manufacturer should run static geometry tests and dynamic motion tests. You or a hired inspector should witness them, either in person or by live video with raw data files.
Worked example – laser positioning test:
The machine is commanded to move to five target positions along the X-axis: 0, 100, 200, 300, 400 mm. Each target is approached from both directions, three times. The laser interferometer records actual positions.
Suppose the results for the 300 mm point are:
Your pass criterion is systematic deviation ≤0.010 mm and repeatability ≤0.005 mm. The machine passes.
Worked example – ballbar test:
Programme a 200 mm diameter circle in the XY plane at 1000 mm/min. The ballbar records a roundness deviation of 0.012 mm. Your pass criterion is ≤0.015 mm. The machine passes.
If the roundness deviation is 0.025 mm, the machine fails. That often points to backlash or servo gain mismatch. Do not accept the machine until the seller corrects it and retests.
The machine is installed, levelled, and powered. The same tests from Step 3 are repeated on your floor. Do not skip this because the manufacturer already passed. Transport, rigging, and your foundation can change a machine.
For high-precision work, the workshop should be at 20 ± 2 °C during the test. If your shop swings more than ±5 °C day to night, then holding 0.005 mm on a standard machine is unrealistic. You either need a temperature-controlled room or you need to relax the tolerance.
This is the most important step. The machine cuts a test part with your three critical features. Use the same tooling, coolant, and speeds you plan to run in production.
Machine at least five parts. Ten is better for a meaningful Cpk calculation. Measure each part with calibrated equipment.
Worked example – bore diameter 40 H7:
Limits: lower specification limit LSL = 40.000 mm, upper specification limit USL = 40.025 mm.
You machine five parts and measure:
Mean = 40.013 mm Sample standard deviation s = 0.00158 mm (use n−1)
Cpk to upper limit = (40.025 − 40.013) / (3 × 0.00158) = 0.012 / 0.00474 = 2.53
Cpk to lower limit = (40.013 − 40.000) / (3 × 0.00158) = 0.013 / 0.00474 = 2.74
Cpk = 2.53. Pass if your criterion is ≥1.33.
Now suppose a different machine gives:
Mean = 40.021 mm s = 0.00224 mm
Cpk upper = (40.025 − 40.021) / (3 × 0.00224) = 0.004 / 0.00672 = 0.60
Cpk lower = (40.021 − 40.000) / 0.00672 = 3.13
Cpk = 0.60. That fails, even though every part is inside tolerance. The process is too close to the upper limit. The machine may be capable, but the tool offset or thermal drift is pushing the mean upward. Recentre the process and retest. Do not release payment until the corrected process reaches the agreed Cpk.
For high-volume automotive work, a Cpk of 1.67 is a common target. For small-batch job shops, 1.33 is a reasonable minimum. The number must be agreed before testing.
This is often forgotten. Machines change shape as they warm up. The spindle motor, axis drives, and coolant all add heat.
A basic thermal test:
Worked example – Z-axis drift:
At start, Z position = 0.000 mm. After 2 hours at 12,000 rpm, Z position = 0.018 mm. Your pass criterion is drift ≤0.025 mm without thermal compensation. The machine passes for standard work.
If you paid for thermal compensation, a typical pass criterion might be drift ≤0.015 mm. If the machine drifts 0.025 mm with compensation switched on, the compensation is not working. Fix it before final payment.
Write the results in a signed acceptance report. Include:
If the machine fails any agreed test, the seller must correct it and retest. You do not pay the final instalment until every pass/fail criterion is met. This is not unfair. It is exactly what the acceptance protocol was designed for.
This memo is a general due-diligence guide, not a final acceptance protocol for a specific order. Before you sign any purchase order, the following items are open and must be confirmed in writing:
Any acceptance protocol is only as strong as the written contract behind it. Do not rely on verbal promises.
This memo is written for general guidance. Specific machine specifications, options, and acceptance figures must be confirmed against the current official Gree catalogue and your purchase contract.
Run the buyer self-check & book a machine selection call — contact the Gree CNC team.