Accurate CNC machining service: where the tolerance actually comes from
A tolerance on a drawing is a promise. This page explains how an accurate CNC machining service keeps that promise: fewer setups on five-axis centers, in-process probing, thermal control, and final inspection. Written for design engineers and sourcing engineers who need to judge whether a shop can hold ±0.005 mm on their part.
Accuracy is a process result, not a machine spec
Every section below answers one question: what in the process decides whether your part lands inside tolerance?
Where accuracy is lost before the cutter touches metal
Most out-of-tolerance parts are not caused by a worn tool. Setup error, thermal drift, and fixture flex come first. A three-axis machine holds a prismatic block well when every feature is reachable from one direction. Add a side hole, an undercut, or a contoured face at an angle, and the part has to be re-fixtured. Each re-fixture adds a datum shift. Two setups of 0.010 mm each can stack into 0.020 mm of position error before any tool wear is counted.
Fixture design decides more than spindle speed. A part clamped on a thin wall will deflect under cutting force and spring back after the cut. The bore measures oversize only after the clamp is released. For thin-wall housings, we plan support and clamp points in the DFM review, and sometimes machine a soft jaw or a dedicated nest for the run.
Material behavior matters too. Aluminum 6061-T6 moves with heat; the same part can measure differently at 8:00 and 15:00 in a shop without temperature control. Titanium and Inconel work-harden and push the tool away. A shop that ignores these effects will still report a good first article and drift on part 200.
- 1Setup countEach additional setup adds a datum shift that cannot be measured away.
- 2Fixture stiffnessDeflection under cutting force shows up only after unclamping.
- 3Thermal driftSpindle, coolant, and ambient heat move the part during a long run.
- 4Tool push-offHard alloys deflect slender tools; the cut runs off nominal.
How five-axis machining removes setup error
A simultaneous five-axis center tilts the tool and rotates the part in the same setup. Undercuts, angled holes, deep pockets, and contoured surfaces come off one datum. That is the main accuracy gain: not a better spindle, but one less chance to lose position between operations.
GreatLight runs 16 simultaneous five-axis machining centers with a Ø400 mm rotary table, alongside 12 four-axis mills and 27 three-axis machines. The five-axis cells handle parts that would otherwise need three or four fixtures. Travel ranges cover compact work at 500 × 500 × 450 mm up to 4,000 × 400 × 150 mm for long parts.
Five-axis is not automatically the right choice. A flat bracket with six drilled holes runs faster and cheaper on a three-axis machine with a good fixture. Reach matters more than axis count. If the tool cannot reach the feature without a re-fixture, five-axis pays for itself. If it can, it usually does not.
- 1Good fitImpellers, angled ports, deep cavities, parts with features on five faces.
- 2Poor fitSimple plates and shafts that one three-axis setup already covers.
- 3Watch forLong slender tools still deflect; five-axis does not fix poor tool rigidity.
Tolerance and finish we hold by feature type
These are the working numbers we quote against, not the best result ever recorded on one part.
| Feature | Typical tolerance | Achievable finish | Notes |
|---|---|---|---|
| Milled pocket, 6061 | ±0.005 mm | Ra 0.8–1.6 μm | Rigid setup, sharp tool |
| Bored hole, 17-4PH | ±0.005 mm | Ra 0.2–0.8 μm | Boring pass after drilling |
| Turned shaft, 304 | ±0.005 mm | Ra 0.8–1.6 μm | Mill-turn, one chucking |
| Angled port, 5-axis | ±0.005 mm | Ra 1.6–3.2 μm | One setup, rotary table |
| Thin wall, 1.5 mm | ±0.005 mm | Ra 1.6–3.2 μm | Support needed, light passes |
| Long part, 4,000 mm | ±0.005 mm | Ra 1.6–3.2 μm | Thermal control critical |
Material choice changes what accuracy costs
Aluminum 6061, 7075, and 2024 cut freely and hold tolerance well. They also expand more than steel per degree of temperature change. On a 300 mm part, a 5 °C shop swing moves the length by roughly 0.035 mm in aluminum. That is larger than the tolerance we are trying to hold. Climate control and part-soak time are not optional on tight aluminum work.
Stainless 303 and 316 machine cleanly but work-harden if the tool rubs. 17-4PH in the H900 condition is stable and takes a fine bore. Titanium TC4 (Ti-6Al-4V) and Inconel push back hard; they need low cutting speeds, rigid tooling, and more in-process checks. We machine all of these, plus beryllium copper and magnesium AZ31B.
Plastics behave differently again. POM and PEEK move after machining as internal stress releases. A bore that is on size at the machine can shrink overnight. For tight plastic parts we rough, stress-relieve, then finish, and we tell you that the first article may need a second measurement the next day.
- 1AluminumFast to cut, high thermal expansion; needs temperature control.
- 2Stainless and 17-4PHStable and bore-friendly; work-hardens if the tool rubs.
- 3Titanium and InconelLow speeds, rigid tools, more checks; slower cycle.
- 4PlasticsStress release moves dimensions after machining; finish last.
How we verify a part is actually accurate
An accurate CNC machining service is only as good as its measurement. A CMM report proves nothing if the part was 4 °C warmer than the reference temperature when it was measured. We let parts soak to room temperature before final inspection, and we record the shop temperature on the report when the tolerance is tight.
Every order gets 100% inspection before shipment. That covers a raw material check on incoming stock, in-process monitoring during the run, and final inspection of the finished part. Reports are available on request, including first article inspection and dimensional layouts. Our historical qualification rate is 99.99%.
In-process probing is the other half. Touch probes on the five-axis centers check the datum and key features between passes, so a drift is caught on part 20 instead of discovered on part 200. On long runs we also check the first part, a mid-run part, and the last part against the drawing.
- 1IncomingRaw material check against the certificate before cutting.
- 2In-processProbe checks on datum and critical features during the run.
- 3Final100% dimensional inspection before shipment, reports on request.
Cases where a tight tolerance is the wrong request
A ±0.005 mm callout on every dimension raises cost without adding function. If a hole locates a cover, ±0.1 mm may do the job and cut inspection time. We flag over-toleranced features in the DFM review and ask what the dimension controls before quoting the tighter number.
Surface finish is the same story. Ra 0.2–0.8 μm needs a separate finishing pass and sometimes hand polishing. If the surface only needs to look clean, Ra 1.6–3.2 μm comes straight off the machine. Put the tight finish only on sealing faces and bearing bores.
Datums are the third common issue. A drawing with three datums that do not match the function will force us to hold dimensions that do not matter while the ones that do drift. If you can tell us how the part is assembled, we can usually suggest a cleaner datum scheme before the first chip is cut.
Questions engineers ask before they send a drawing
Can you really hold ±0.005 mm on a production run, not just a prototype?
Yes, on features we can reach in a stable setup and measure reliably. The number applies to the feature, not to the whole part. A 4,000 mm part will not hold ±0.005 mm over its full length.
We tell you which dimensions we can hold before quoting. If a callout is not realistic on your geometry, the DFM review says so in writing.
How do I know the five-axis machine is the right choice for my part?
Count the setups. If the part needs more than one orientation to reach all features, five-axis usually wins on both accuracy and cost.
If a single three-axis setup reaches every feature, the extra axes add cycle time without adding value. We quote the cheaper route when it holds tolerance.
What do I need to send for an accurate quote?
A 3D model plus a 2D drawing with datums, tolerances, material, finish, and quantity. The drawing matters more than the model for accuracy work.
Send both through the quote page. Uploads are secure and confidential, and an NDA is available on request. Quotation and free DFM analysis come back within 12 hours.
Which materials give you the most trouble on tight tolerances?
Thin-wall titanium and Inconel are the hardest. They work-harden, deflect the tool, and move with heat. They can be done, but expect slower cycles and more in-process checks.
Plastics like POM and PEEK are the opposite problem. They cut easily but continue to move after machining as stress releases.
Do you provide inspection reports?
Yes. Every order gets 100% inspection before shipment, and dimensional reports are available on request.
We run raw material checks, in-process monitoring, and final inspection. First article inspection reports are standard on new parts.
What is the smallest quantity you accept for an accurate part?
One piece. There is no minimum order quantity, so a single prototype and a 10,000-part run go through the same process.
Production can start within 24 hours of a confirmed order, and parts typically ship in 3–5 days.
Send a drawing and get a real tolerance answer
Upload your model and drawing. We review the datums, setups, and material, then quote the tolerance we can actually hold. Quotation and free DFM analysis within 12 hours, NDA on request.
12-hour quote100% inspectionNo minimum orderNDA available
