For a German Mittelstand shop holding tolerances below ±10 µm on batches of 50 or more, thermal growth is a bigger scrap risk than machine age or country.
For a German Mittelstand shop holding tolerances below ±10 µm on batches of 50 or more, thermal growth is a bigger scrap risk than machine age or country of origin. A new Chinese machining centre with active thermal compensation is usually the lower-risk choice over a used European machine without it, unless your parts are thermally forgiving, your runs are very short, or your shop already controls temperature tightly enough to make the simpler option work.
This article is written for shop owners and purchasing engineers comparing a new Chinese-built machine against a used European one. The comparison is not about brand pride. It is about the physical fact that metal changes size when it gets warm, and that this change happens slowly, without an alarm, while the spindle is cutting.
Thermal growth is the dimensional change of a material when its temperature changes. Every material has a coefficient of thermal expansion (CTE), which tells you how many micrometres one metre of that material grows per degree Celsius. For cast iron and steel, the CTE is typically 10–12 µm per metre per degree Celsius. That is an industry-wide physical property, not a machine specification.
A machining centre column that is one metre tall will grow roughly 50–60 µm if its temperature rises from 20 °C to 25 °C. That is 0.05–0.06 mm. If you are boring a hole with a ±10 µm tolerance, that growth alone can push you out of spec before the cutting tool even touches the part.
The main heat sources in a machine tool are:
The problem is not that these parts get warm. The problem is that they do not all get warm at the same rate or by the same amount. That uneven heating bends the machine structure slightly and changes the relative position between the tool tip and the workpiece.
| Heat source | Typical temperature rise | Typical dimensional drift | Impact on a tight-tolerance part | Common countermeasure |
|---|---|---|---|---|
| Spindle bearings and motor | 10–25 °C after 30–60 min | 20–50 µm in Z axis | Bore depth and face position drift | Warm-up cycle, active spindle cooling |
| Ballscrew friction | 5–15 °C after 30 min | 10–30 µm per axis | Position drift in X, Y, or Z | Linear scales, ballscrew cooling, thermal compensation |
| Machine column and bed | 2–8 °C over several hours | 20–80 µm depending on size and casting | Overall geometry distortion | Symmetrical casting design, temperature sensors, compensation |
| Ambient shop temperature | ±2–5 °C typical unclimatised | 10–25 µm per metre of structure | Random variation batch to batch | Climate control ±1 °C or better |
| Coolant temperature | 5–20 °C rise during long runs | 10–40 µm transmitted into table and fixtures | Part and fixture growth | Coolant temperature control unit |
The values in this table are typical industry ranges found in machine tool thermal research and manufacturer application notes. They are not specifications for any single machine. Any machine you evaluate must be tested or confirmed against the official catalogue.
In one-off work, thermal drift is annoying but manageable. You can measure the first part, adjust offsets, cut, measure again. The machine drift between those steps is small, and you correct it manually.
In batch work, the risk multiplies silently. A typical 50-part batch takes two to four hours. During that time the spindle warms up, the ballscrews heat up, and the machine frame slowly changes shape. The first 10 parts may be within tolerance. Parts 30 to 50 may drift out. No alarm sounds. The operator sees good parts and keeps loading blanks.
Here is a realistic Mittelstand example. You are machining 60 aluminium valve bodies with a 20 mm H7 bore. The tolerance on bore position is ±8 µm. Your machine starts cold at 18 °C. After 40 minutes of cutting, the spindle nose has grown 25 µm in Z and the ballscrew has shifted X position by 12 µm. You do not re-zero because the first parts measured well. The last 20 parts may have bores shifted by 20–30 µm relative to the datum. That is scrap or expensive rework, and you only find out when the customer rejects the batch or when you measure at final inspection.
A warm-up cycle helps, but it is not a complete cure. Running the spindle at medium speed for 20–30 minutes before production stabilises the spindle bearings only. It does not stabilise the ballscrews, the machine column, or the coolant if they have not been running under load. In a shop without strict climate control, ambient temperature drift during the batch can still move the machine frame by 10–25 µm over two hours.
A used European machining centre from 10–20 years ago often has excellent mechanical rigidity and a well-documented design. But most machines from that era do not have active thermal compensation. Some have linear scales, which are glass or magnetic rulers that measure the actual position of the axis slide rather than counting ballscrew rotations. Linear scales remove ballscrew thermal error, but they do not compensate spindle growth or machine frame distortion. In a vertical machining centre, spindle growth in Z is usually the largest single thermal error. So even a used machine with linear scales can still drift 20–50 µm in Z during a warm-up period.
A new Chinese machining centre with active thermal compensation takes a different approach. Temperature sensors are placed on the spindle nose, spindle housing, ballscrew nuts, and sometimes the column. Software reads those sensors many times per second and adjusts the axis positions to cancel the predicted thermal drift. This is not a mechanical fix; it is a calculation that shifts the coordinate system. The quality of the result depends on how many sensors are used, where they are placed, and how well the compensation model was calibrated.
Typical industry data from machine tool thermal error research suggests that a machine without active compensation can show 40–80 µm of thermal drift during warm-up. A machine with well-implemented active compensation can often hold that drift below 10–15 µm. These are typical ranges, not guarantees. Some entry-level machines may have only one or two sensors and a simple linear model, which reduces drift but does not eliminate it.
Gree CNC publishes thermal compensation as a feature on several vertical machining centre series. The exact number of sensors, the compensation range, and the residual drift after warm-up must be read from the official catalogue for the exact model and confirmed by a test on your shop floor. Do not assume that “thermal compensation” means the same thing on every machine.
The honest limit is this: a new Chinese machine does not automatically have better thermal behaviour just because it has software. Build quality of castings, ribbing, and sensor calibration still matter. A used European machine that has been retrofitted with a temperature sensor kit and a good compensation model can also work, but the retrofit cost and validation effort often eat up the price advantage.
| Criteria | Used European machine, 10–20 years old, no active compensation | New Chinese machine with active thermal compensation | What to verify before purchase |
|---|---|---|---|
| Spindle thermal drift after 30–60 min warm-up | Typically 20–50 µm, can be higher on worn spindles | Often under 10–15 µm if compensation is well calibrated | Run a thermal drift test on your shop floor |
| Ballscrew thermal error | 10–30 µm unless linear scales are fitted | Typically reduced by temperature sensors and software | Ask for axis drift test data, not just brochure claims |
| Machine frame distortion from ambient temperature change | Full effect, 20–80 µm depending on structure | Partially compensated if frame sensors are installed | Check number and location of temperature sensors in the official catalogue |
| Long-term repeatability after years of use | Well documented in some cases, but dependent on maintenance and crash history | Less long-term field data in DACH climate for some newer Chinese models | Ask for reference customers with similar batch work |
| Purchase price | Lower upfront, but risk of hidden wear | Higher upfront, but includes warranty and newer control | Total cost of ownership over 5 years, not just invoice price |
| Local service and spare parts | Often strong in DACH and UK for established European brands | Improving, but check distance to service engineer and parts stock | Ask distributor for response time and spare part lead time in writing |
This table uses typical ranges from industry experience and published thermal research. It is not a comparison of any specific European brand against any specific Gree CNC model. Every machine must be evaluated individually.
This article is not saying you should always buy a new machine with thermal compensation. If your work does not challenge thermal stability, the simpler used European machine may be the lower-cost and lower-risk choice.
Here are the conditions where a used machine without active thermal compensation is usually good enough:
If all of these points are true for your shop, a used European machine can be a rational purchase. If any of them is false, the thermal risk grows quickly.
Before buying either machine, run a thermal drift test. The test does not need expensive equipment. You need a dial indicator or laser tool setter and a reference block.
The procedure is simple:
For tight-tolerance work below ±10 µm, a total drift of more than 15 µm is a warning sign. For moderate work at ±20–30 µm, a drift above 30 µm means you will be fighting the machine all day. These thresholds are typical industry practice, not a Gree CNC specification.
For a used machine, ask the seller if any thermal compensation retrofit is installed. If not, factor in the cost of adding temperature sensors and a compensation algorithm, or accept the manual correction burden.
For a new Chinese machine, ask the distributor for a factory thermal drift test report. The report should show the test procedure, the temperature sensor locations, and the measured drift. If the distributor cannot provide that report, treat the thermal compensation claim as unproven.
Gree CNC machines sold into the DACH and UK markets are supported by the official catalogue. The catalogue is the only source for exact sensor count, compensation range, and environmental operating limits. Do not rely on a sales conversation for these numbers.
Run the buyer self-check & book a machine selection call — contact the Gree CNC team.