CNC Machine Volumetric Measurement Services: How to Choose a Supplier

• 1
  Key takeaways
• 2
  What to compare across suppliers
• 3
  What CNC machine volumetric measurement services actually test
• 4
  Matching the test method to your part tolerance
• 5
  Certifications, reports, and what they do not cover
• 6
  Lead time, order size, and the cost of getting it wrong
• 7
  Step by step: screening a volumetric measurement supplier

A volumetric test asks one question: if the machine commands a point anywhere inside its working volume, where does the tool tip actually go? Linear accuracy on each axis is not enough. Three axes can each be within 5 μm and still put a hole 25 μm off once squareness error and angular error stack up over a long travel.

The map covers 21 error terms in the classic model. Three positioning errors, six straightness errors, three squareness errors, and nine angular terms. On a 4,000 mm machine the angular terms dominate. A 10 μrad pitch error is 40 μm at the far end of travel. That is why a test near the vise tells you almost nothing about the part you cut at the other end.

Tools differ. A laser interferometer reads linear position and angular deviation along a line. A ballbar reads circular interpolation and picks up servo mismatch and squareness fast. A touch probe on a calibrated sphere set checks the whole chain as the machine would actually run. Good suppliers combine them rather than pick one.

The output should be a deviation map, not a single number. A map shows where the machine drifts and in which direction. That lets you place a tight-tolerance feature in the calm region of the envelope or decide the work needs a different machine.

Start from the tightest tolerance on the drawing and work backwards. If a bore is held at ±0.005 mm, the machine has to hold position well inside that, because the test itself carries uncertainty. A common rule is to keep machine error under one third of the part tolerance. That leaves room for tool wear, thermal growth, and clamping shift.

For ±0.05 mm work on a 500 mm part, a ballbar check and a probe sweep may be enough. For ±0.01 mm work on a 1,000 mm part, you want a full 21-term volumetric map with compensation loaded. For anything tighter, ask whether the feature can be finished on a smaller machine with shorter travel. Geometry favors the small machine.

Material matters less than people think. Aluminium 6061, 7075, 316L stainless, and Ti-6Al-4V all cut to the same commanded path; they differ in cutting force and heat. Heavy cuts in 4140 or Inconel push deflection and thermal drift, so a machine that passes cold may drift on the third hour of a roughing cycle.

Ask what happens after a crash. A spindle bump or a table collision shifts geometry. A supplier who re-tests after any crash will catch it. One who waits for the annual check may ship bad parts for months without knowing.

ISO 9001:2015, IATF 16949:2016, and ISO 13485:2016 tell you the quality system is audited. They do not certify the accuracy of any specific machine on any specific day. Volumetric data is a separate document, and it should be dated, signed, and tied to a machine serial number.

A useful report lists the instruments used with their calibration due dates. A laser interferometer with an expired calibration certificate proves nothing. The report should also state ambient temperature and whether the machine was warmed up. A cold machine reads differently from one that has run for two hours.

Watch for reports that only show final numbers. If the raw deviation file is missing, you cannot check whether the pass limit was set before the test or chosen after seeing the result. Suppliers who publish both are easier to trust.

Reports should travel with the parts. When GreatLight ships an order, inspection records are available on request. That includes raw material check, in-process monitoring, and final inspection, not just a cover page.

Metrology adds time. A full volumetric map on a large machine can take most of a day, including warm-up and thermal soak. Suppliers who quote volumetric measurement as a free same-day add-on are usually doing a probe sweep and calling it volumetric.

That said, the test is cheap compared to a scrapped run. If a 200-part batch of 7075 housings has to be reworked because the machine drifted, the loss is far larger than the measurement cost. Ask for the test as a line item so you can see what you are buying.

Order size changes the calculus. For a single prototype, a probe verification of the critical features may be enough. For 10,000 parts, the machine needs a stable, documented error map because the drift will accumulate over the run and show up in the last articles.

Confidentiality matters when you send CAD files for a metrology review. GreatLight keeps uploads secure and confidential, and an NDA is available on request.