Buyer guide for engineers

Precision CNC CMM Manufacturer Guide

This precision CNC CMM manufacturer guide is written for design engineers and sourcing teams who have to accept or reject a machined lot on paper. Read it and you will know which CMM numbers matter, which ones are decoration, and when to ask for a different measurement method.

Get Instant Quote Talk to an Engineer

±0.005 mm machining tolerance12-hour quote + DFM100% inspection before shipmentISO 9001 / IATF 16949

Precision CNC CMM manufacturer guide: bridge CMM measuring a machined part

On this page

1
Key takeaways
2
Which measurement method fits which part
3
Reading a precision CNC CMM report like an engineer
4
What a precision CNC CMM manufacturer should be able to hold
5
When a CMM is the right tool, and when it is not
6
Sampling plans, FAI and what to put in the PO
7
Seven steps to qualify a precision CNC CMM supplier

Key takeaways

CMM data is a starting point, not a verdictA printed deviation list tells you nothing until you know probe uncertainty and how the part was held.

Ask for the uncertainty, not the resolutionA machine that displays 0.1 μm does not measure to 0.1 μm. Probe uncertainty limits you.

Match the method to the featureContact probing suits holes, bores and planes. Thin walls and soft plastics often need optical or low-force probing.

Certificates without scope are weak evidenceISO 9001:2015 and IATF 16949:2016 only count if the certificate scope covers your part family.

One CMM report is a sample, not a trendAsk how often the critical dimensions are re-measured across the run, not just on the first article.

Decision table

Which measurement method fits which part

Use this table before you write the inspection requirement into the PO.

Part feature	Recommended method	Typical uncertainty	Watch out for
Bores, holes, slots, planes	Contact CMM, ruby stylus	±2 to ±3 μm	Stylus deflection on deep small holes
Thin walls under 0.8 mm	Optical or laser scanning	±5 to ±10 μm	Surface finish changes the reading
Soft plastics and elastomers	Low-force probing or optical	±5 to ±10 μm	Contact force deforms the feature
Freeform and 5-axis surfaces	Scanning CMM, dense point cloud	±3 to ±8 μm	Filter settings hide real waviness
Large frames over 1,000 mm	Bridge CMM, temperature controlled	±5 to ±15 μm	Thermal growth over the day
First article on a new process	CMM plus hand tools cross-check	±2 to ±10 μm	Single method gives false confidence
High-volume production parts	CMM sampling plus gauges	±2 to ±5 μm	Gauge wear drifts without checks

What the data means

Reading a precision CNC CMM report like an engineer

A CMM report is a list of measured points and the deviation from nominal. That is all it is. The value comes from the conditions behind those points: which probe was used, how the part was clamped, what the ambient temperature was, and whether the datum scheme matches your drawing. Ask for those four things and most reports become readable in a few minutes.

Start with the datum. If the drawing calls for datum A as a primary face and the inspection report aligns the part on a vise jaw instead, every position number downstream is suspect. Datum mismatch is the single most common reason a good part looks out of tolerance on paper.

Then look at how many points define each feature. A hole measured with four points gives you a diameter and very little else. Roundness and cylindricity need far more points, usually 12 or more around the circumference, and often two or three levels through the depth. If the report shows a cylindricity value from a four-point circle, treat it as a rough indicator, not a controlled result.

Finally, compare the reported uncertainty against the tolerance band. If your tolerance is ±0.010 mm and the measurement uncertainty is ±0.003 mm, the measurement eats 30 percent of the band. That is workable but tight. If uncertainty is ±0.008 mm against the same tolerance, you cannot reliably separate good parts from bad ones, and the inspection method needs to change before the parts do.

1
Check the datum scheme firstIt should match the drawing callouts, not the shop fixture.
2
Count the points per featureRoundness needs 12+ points around the circle, not four.
3
Compare uncertainty to toleranceKeep measurement uncertainty under about one third of the band.

Capability

What a precision CNC CMM manufacturer should be able to hold

Machining capability and measurement capability have to move together. A shop that can hold ±0.005 mm on a mill but cannot measure below ±0.010 mm is guessing on the tight features. At GreatLight, machining tolerance is stated at ±0.005 mm, and surface finish ranges from Ra 0.2–0.8 μm on fine finishes to Ra 1.6–3.2 μm as machined. Those are the numbers the CMM has to confirm.

Probe measurement uncertainty on our bridge CMMs is normally below ±3 μm, which keeps it inside one third of a ±0.010 mm band. For parts with ±0.005 mm tolerances, that margin disappears, so the inspection plan changes: more points per feature, temperature soak before measurement, and often a second method to cross-check the critical dimensions.

Temperature is not a detail. Aluminum 6061 grows roughly 23 μm per meter per degree Celsius. A 300 mm part measured 5 °C warmer than the calibration reference shifts about 34 μm. That is larger than most tolerances on the drawing. If your supplier measures in a shop that swings with the weather, the numbers move with it.

We measure after the part has soaked, with the CMM and the part in the same room. For long parts we note the temperature in the report. It is a small line item, but it explains a lot when a dimension drifts between the first article and the production run.

1
Machining tolerance: ±0.005 mmStated capability, not a marketing claim.
2
Probe uncertainty normally under ±3 μmRoughly one third of a ±0.010 mm band.
3
Thermal drift is realAluminum moves about 23 μm per meter per °C.

Which parts fit

When a CMM is the right tool, and when it is not

Contact CMM probing is strong on prismatic parts: housings, brackets, manifolds, engine and transmission components, fixture plates. These have well-defined datums, rigid geometry and features that a ruby stylus can reach. If your part has bores, counterbores, slots and flat faces with position callouts, a CMM report is the right document to ask for.

It is weaker on thin-walled and compliant parts. A 0.5 mm aluminum wall deflects under a few millinewtons of contact force, and the reading becomes a record of the deflection rather than the part. The same applies to plastics such as POM, PA and PEEK, and to anything with a soft coating. Low-force probing or optical scanning is the better choice there.

Deep small holes are another limit. A Ø1 mm hole 20 mm deep needs a stylus long enough to reach the bottom, and long styli bend. Shaft stiffness falls with the cube of length, so a stylus that works at 10 mm depth may be unusable at 30 mm. Suppliers who show you the stylus selection, not just the result, are usually the ones who understand this.

Very large parts bring a different problem. Our maximum processing size is 4,000 mm, with travels of 4,000 × 400 × 150 mm on the large machines. Measuring a part that long needs a bridge CMM with matching volume and thermal stability, or a portable arm with an accuracy penalty. Ask which one is being used.

1
Good fit: prismatic and rigidHousings, brackets, plates, engine and transmission parts.
2
Poor fit: thin walls and soft plasticsContact force deforms the feature being measured.
3
Poor fit: deep small holesStylus stiffness drops sharply with length.

Process control

Sampling plans, FAI and what to put in the PO

A single inspection report proves one part at one moment. Process control is what keeps the run inside tolerance. The standard structure is a first article inspection on the first good part, then periodic verification of critical dimensions on production parts according to a statistically based sampling plan, with 100 percent inspection reserved for critical components or low-volume high-precision work.

Write the sampling frequency into the purchase order. If the drawing has a position tolerance on four holes, say how often those holes are re-measured: every 20th part, every lot, or at each tool change. Vague wording like suitable inspection at the supplier's discretion gives you nothing to hold anyone to when a bad lot arrives.

Ask what happens when a dimension goes out. The useful answer describes containment, re-measurement with a second method, and a root cause on the process side, whether that is tool wear, thermal drift or fixture wear. A supplier who can only re-inspect the same part twice is not controlling the process.

Records matter too. We run raw material checks, in-process monitoring and final inspection, with reports available on request. For regulated programs, the certificate scope is the part buyers forget: ISO 9001:2015, IATF 16949:2016, ISO 13485:2016 and ISO 27001:2022 each cover specific activities, and the scope line on the certificate tells you whether your part family is inside it.

1
FAI first, then samplingCritical dimensions re-measured on a defined schedule.
2
100% inspection for critical or low-volume partsReserved for the parts where a single escape is expensive.
3
Check the certificate scopeA certificate only covers what its scope line describes.

Buyer checklist

Seven steps to qualify a precision CNC CMM supplier

Run these in order. Each step can end the conversation, which saves time on both sides.

1
Send a part with a known hard featurePick a part with one tight position callout and send the drawing with the RFQ. Ask for the CMM report with the quote, not after. Typical turnaround on quote and DFM analysis is 12 hours.
2
Read the datum scheme against your drawingIf the primary datum is a vise jaw instead of datum A, stop and ask why before discussing any deviation number.
3
Demand the probe uncertainty figureAsk for it in writing. Below ±3 μm is normal for a bridge CMM. If the supplier quotes resolution instead, they are answering a different question.
4
Check points per critical feature12 or more points around a bore for roundness, two or three depth levels for cylindricity. Four points is a diameter check, nothing more.
5
Ask how the part is held and soakedRigid fixturing, temperature soak, CMM and part in the same room. A 300 mm aluminum part measured 5 °C off shifts about 34 μm.
6
Agree the sampling plan in writingState the frequency for each critical dimension: every 20th part, every lot, or at each tool change. Put it in the PO.
7
Confirm certificate scope and NDA coverageMatch ISO 9001:2015, IATF 16949:2016 or ISO 13485:2016 scope to your part family, and sign an NDA before drawings move. Uploads stay confidential.

FAQs

Frequently asked questions

Can a CMM measure soft or thin-walled parts without damaging them?

Yes, with the right setup. Low-force probing using ruby styli reduces contact force to a level that will not mark most surfaces, and optical or laser scanning avoids contact entirely.

Both methods trade some accuracy for safety, typically landing in the ±5 to ±10 μm range instead of ±2 to ±3 μm. If your tolerance on a thin wall is tighter than that, the feature usually needs a design change or a dedicated gauge.

What CMM accuracy should I expect from a machining supplier?

For most bridge CMMs used in job-shop inspection, probe measurement uncertainty sits around ±2 to ±3 μm under good thermal conditions. That is enough to support ±0.010 mm machining tolerances comfortably.

Below ±0.005 mm, the measurement and the process start competing for the same budget. At that point the inspection plan needs more points, temperature control and a second method.

Is 100 percent CMM inspection realistic on a production run?

It is realistic on critical components and low-volume high-precision work. On a 10,000-part run, full CMM inspection of every part is slow and expensive, and it does not improve the process.

The better structure is first article inspection, then a statistically based sampling plan on the critical dimensions, with gauges for the rest. Gauges need their own wear checks.

What should be in the CMM report I receive?

At minimum: the datum scheme, the probe and stylus used, the number of points per feature, the measured deviation against nominal, and the ambient temperature.

Also ask for the measurement uncertainty. A report with deviations but no uncertainty figure cannot tell you whether a part at the tolerance edge is good or bad.

Does an ISO 9001 certificate mean the CMM data is trustworthy?

Not by itself. ISO 9001:2015 covers the quality management system. It does not certify a specific measurement uncertainty or a specific part family.

Read the scope line on the certificate and ask how the CMM is calibrated and how that calibration is traced. The calibration record is the part that makes the numbers meaningful.

How early can inspection start on a new program?

Once the drawing and inspection plan are agreed. With 127 high-precision CNC machines and 16 simultaneous 5-axis machining centers on site, production can start within 24 hours, and parts typically ship in 3–5 days.

First article inspection runs on the first good part, so the report reaches you before the run is finished. That gives you a chance to stop the process early if something is off.