For a German Mittelstand shop deciding between a new Chinese CNC machine and a used European one, the foundation, levelling and the first 30 days.
For a German Mittelstand shop deciding between a new Chinese CNC machine and a used European one, the foundation, levelling and the first 30 days of accuracy matter more than the machine’s origin. A new machine with a proper foundation, repeated levelling and a planned 30-day verification process will usually hold tolerance better than a worn used machine that is simply placed on the existing floor and left alone.
This primer debunks three common myths. It is written for owners and purchasing engineers in the DACH region and the UK.
You are not comparing a nationality. You are comparing a measured installation result.
A used European machine may have 15,000 spindle hours, unknown crash history and worn linear guides. A new Chinese machine may have modern control systems and unworn ballscrews. But neither machine is accurate unless it sits on a foundation that resists movement, is levelled to the builder’s tolerance and is re-checked during the first 30 days.
The decision is therefore an engineering decision, not a badge decision.
This is the most expensive myth.
CNC machine tools are not rigid enough to correct a moving floor. A machine bed that twists by 0.02 mm over one metre changes the geometry of every axis that runs on it. A hard floor is not the same as a stable floor.
A foundation has three jobs:
The standard solution is a reinforced concrete inertia block. An inertia block is a separate, heavy concrete block that increases the machine’s mass and lowers its natural frequency, so it moves less.
For a medium vertical machining centre weighing 8–15 tonnes, foundation engineers often specify 300–500 mm of reinforced concrete. This is a typical industry range. It is not a Gree catalogue value. The official Gree installation manual for your specific model is the only authority. If the manual asks for a different thickness, use that number.
Do not trust a visual inspection.
A used European machine often arrives without its original foundation data. You may be buying a machine that was installed on a floor that was never surveyed. A new Gree CNC machine comes with a foundation drawing. You can build the foundation before delivery. That reduces commissioning time and risk.
If your workshop has a thin 100–150 mm warehouse slab, do not buy any 8-tonne machine and place it directly on that slab. It will settle, twist and lose accuracy. You need to cut the slab and pour a proper isolated foundation first. This applies equally to a new Chinese machine and a used European machine.
Levelling is not an event. It is a process with fixed checkpoints.
Levelling means adjusting the machine’s support bolts and pads so the machine bed is flat and true. The tool is a precision machinist’s level, usually with a resolution of 0.02 mm per metre. This means the bubble shows a height difference of 0.02 mm over one metre.
Many machine builders require final levelling within ±0.02 mm per metre. This is a common industry target, but you must confirm the exact tolerance in the official Gree installation manual. Some large gantry machines require tighter or looser values depending on bed length.
The process is not “turn four bolts until the bubble is centred”.
A practical sequence is:
Each check should be recorded with the ambient temperature, spindle temperature and level readings. Without records, you cannot separate foundation settlement from thermal growth.
Thermal growth is the change in machine dimensions as parts warm up. A cast iron bed grows roughly 0.012 mm per metre per 10 °C temperature rise. That is measurable, not imaginary.
Levelling removes twist. It does not remove wear.
If the guideways or ballscrews of a used machine are worn unevenly, the axis will not move in a straight line, even if the bed is perfectly level. You can level the machine, but the table will still climb and fall as it travels over a worn area.
A typical used European machine after 10–15 years may show 0.03–0.06 mm of ballscrew backlash. Backlash is lost motion when a screw reverses direction. A new machine is often specified at 0.005–0.015 mm backlash. These are typical industry ranges. Always ask for the measured acceptance report, not the original brochure.
If the used machine’s measured straightness error is 0.03 mm per 300 mm, no amount of levelling will bring it back to 0.01 mm. You would need re-machining or replacement of linear guides. That is a rebuild cost, not an installation cost.
The first 30 days are not a simple break-in period. They are a controlled accuracy verification window.
A new machine and a used machine both need this window, but for different reasons.
A practical first-30-day plan looks like this:
A ballbar test measures how accurately the machine cuts a circle. It reveals backlash, servo mismatch, squareness errors and thermal drift. A new machine should show a roundness error under 0.010 mm on a 300 mm circle. A used machine may show 0.020–0.040 mm or more if wear is present. These are typical ranges, not catalogue promises. The official Gree acceptance test for your model gives the exact pass value.
Thermal stability matters as much as geometry. ISO machining accuracy is usually referenced to 20 °C ± 2 °C. A Mittelstand shop in the DACH region or UK may heat to 20 °C in winter but reach 30 °C in summer. That 10 °C swing can change a 500 mm part length by 0.06 mm on a steel machine bed. If you need 0.01 mm tolerances, you need temperature control or active thermal compensation.
A new machine starts from a known baseline. The builder’s inspection report gives the as-built accuracy. The first 30 days confirm that transport, foundation and levelling have not degraded it.
A used machine has no known baseline unless the seller provides a fresh, independent measurement report. Most sellers provide the original 2006 report, not a current one. That is a risk. You may discover in the first 30 days that the spindle taper has 0.012 mm runout, the ballscrews have 0.05 mm backlash and the z-axis drops 0.02 mm when warm. None of these can be fixed by levelling.
The comparison is therefore not “new Chinese versus used European”. It is “known new condition with installation plan” versus “unknown used condition with hidden wear history”.
The table below gives typical industry ranges. None of these are promises. Always confirm exact specifications with the official Gree catalogue or the used machine’s current measurement report.
| Parameter | New machine (typical industry range; confirm official catalogue) | Used European machine 10+ years old (typical industry range) | What it tells you |
|---|---|---|---|
| Final levelling tolerance | 0.02 mm/m is common; some models 0.03 mm/m | Same target, but may be impossible if bed is worn or damaged | Ask for a levelling report, not a claim |
| Ballscrew backlash | 0.005–0.015 mm | 0.03–0.06 mm | Above 0.04 mm usually limits circular interpolation accuracy |
| Spindle taper runout | 0.002–0.005 mm | 0.005–0.015 mm | Directly affects surface finish and tool life |
| Straightness per 300 mm axis travel | 0.005–0.010 mm | 0.010–0.030 mm | Wear is uneven; levelling cannot correct it |
| Circular interpolation roundness on 300 mm circle | 0.005–0.010 mm | 0.020–0.040 mm | Ballbar test shows total machine health |
| Foundation settlement allowance in first 30 days | Specified in Gree foundation drawing; often 0.02–0.05 mm | Usually unknown because old foundation data are missing | Ask for a foundation survey before placing any machine |
All figures in this table are typical industry ranges. They are not a substitute for the official installation and acceptance documents.
Use this list before signing any purchase order, whether the machine is new or used.
The machine that holds tolerance in your shop is the one that is installed on the right foundation, levelled repeatedly and verified during the first 30 days. That is measured in millimetres, not in the country name on the nameplate.
Run the buyer self-check & book a machine selection call — contact the Gree CNC team.