← All Articles

CNC Machining Tolerances Explained: What ±0.01mm Really Costs

Published 27 October 2026 · 8 min read

Every CNC quote you receive is built on tolerance. Two drawings of the same part, one at ±0.1mm and one at ±0.01mm, can differ by 40% in price and by a full week in lead time — and most buyers never find out why.

This guide explains what tolerance ranges standard CNC machining actually holds, when tightening to ±0.01mm is genuinely necessary, and how to read a tolerance block without over-specifying your way into a higher price.

The tolerance bands that matter

CNC machining tolerances are not a single number — they are a ladder. Each rung requires different equipment, different cutting strategy, and different inspection.

Band Typical achievable How it is produced Relative cost
General ±0.2 mm Roughing + single finishing pass, standard vise fixturing 1.0× (baseline)
Commercial ±0.1 mm Finishing pass with sharp tooling, basic inspection 1.1–1.3×
Precision ±0.05 mm Semi-finish + finish, controlled coolant, first-article check 1.3–1.6×
High precision ±0.01 mm Multiple light finishing passes, temperature-stable shop, CMM verification 1.6–2.5×
Ultra precision ±0.005 mm and below Grinding, lapping, jig boring, climate-controlled room 2.5×+

The important thing to understand: the jump from ±0.1mm to ±0.05mm is cheap. The jump from ±0.05mm to ±0.01mm is where cost accelerates.

That is because at ±0.01mm you are no longer fighting the cutting tool — you are fighting everything around it.

What actually limits you at ±0.01mm

At this level, the machine itself is rarely the bottleneck. A modern 3-axis VMC with a good spindle holds position to within a few microns all day. The problems come from the environment:

Thermal growth. Aluminium expands roughly 23 µm per metre per °C. A 200mm aluminium part that warms by 5°C during machining grows by about 23 µm — more than double your entire tolerance band. This is why high-precision shops measure at 20°C and let parts stabilise before inspection.

Tool wear. A carbide end mill wears progressively. On a run of 500 parts, the last part can be 10–15 µm smaller than the first unless the operator compensates. Tight-tolerance work means shorter tool life and more frequent offset checks.

Fixturing stress. Clamping force deforms thin walls. The part springs back when released, and the measured dimension no longer matches the cut dimension.

Material internal stress. Rolled aluminium plate carries residual stress from the mill. Machine one side and the part bows. The standard answer is stress-relieved stock, or rough-machine, stress-relieve, then finish — which adds a process step and days to the schedule.

Measurement uncertainty. You cannot hold ±0.01mm if your inspection method has ±0.008mm of uncertainty. This is why a ZEISS CMM is not a marketing prop — at this tolerance level it is a prerequisite.

When do you actually need ±0.01mm?

Genuinely, fewer parts need it than the drawings suggest. High precision earns its cost when the tolerance is functional, not decorative:

Conversely, these usually do not need ±0.01mm:

A useful rule: put the tight tolerance only on the dimensions that touch something. Everything else can sit at ±0.1mm on the same drawing. Mixed-tolerance drawings are normal and expected — a good supplier will quote them naturally.

How to read a tolerance block

A typical drawing carries a general tolerance note plus specific callouts. Something like:

Unless otherwise specified: linear ±0.1mm, angular ±0.5°, surface finish Ra 3.2.

Everything inside that block is priced at commercial rates. Only the explicitly callout dimensions — the ones with ±0.01 or a fit designation like H7 written directly on them — are priced at high-precision rates.

Where buyers lose money:

  1. Tightening everything to be “safe.” A drawing stamped ±0.01mm globally can double the quote, and half of those dimensions will never be measured by anyone.
  2. Forgetting geometric tolerance. A dimension can be ±0.01mm and still be unusable if flatness or perpendicularity is uncontrolled. GD&T callouts (position, profile, runout) matter more than size tolerance on functional parts.
  3. No datum scheme. Without clear datums, the inspector and the machinist interpret the part differently, and you get a reject that was actually a drawing problem.

The tolerance versus cost trade-off, in practice

Take a 120 × 80 × 20mm aluminium housing, quantity 100:

The cost driver is never the number in the title block. It is how many extra operations the tolerance forces.

What to ask your supplier

Before accepting a quote, ask these four questions:

  1. What is your standard general tolerance? A serious shop answers immediately — ±0.1mm, ISO 2768-m, or similar. A vague answer means inspection is informal.
  2. How do you verify ±0.01mm features? The answer should name specific equipment — CMM, air gauge, micrometer with known calibration. “We check with calipers” is a disqualifier at this tolerance.
  3. Do you charge for first article inspection? A real FAI report with actual measured values, not a signed checklist, is what protects you at mass production.
  4. What happens on a tolerance reject? Who pays for the rework, and what is the rework plan? Get this in writing before the PO.

Summary

If you have a drawing where the tolerance stack-up is unclear, send it over — we will mark which dimensions genuinely need tightening and which can relax. Free DFM feedback, quote back within 24 hours.

Request a Quote →