Why Tolerances Define Whether Your Part Succeeds or Fails
Every engineered component exists on paper as a perfect number — a shaft diameter of 12.000 mm, a housing depth of 8.500 mm. In the real world, those perfect numbers are never achievable. Every cutting tool flexes slightly under load, every workpiece expands a fraction as it heats up, and every fixture introduces its own microscopic positional variation. The gap between the ideal dimension and what the machine actually produces is inevitable.
What separates a well-engineered part from one destined for the scrap bin is not the elimination of this variation — it is the intelligent management of it. That management happens through CNC machining tolerances: the defined boundaries within which a dimension must fall for a part to be considered acceptable.
Getting tolerances right has a direct impact on assembly success, functional reliability, production cost, and lead time. Call them too tight without engineering justification, and you escalate costs, slow throughput, and put strain on supplier relationships. Leave them too loose, and parts won’t assemble cleanly, wear faster, or fail under load. This guide explains how tolerances work, what achievable precision looks like across different machining processes, and how to make decisions that align your design requirements with practical manufacturing realities.
What CNC Machining Tolerances Actually Mean
A tolerance is the total permissible variation in a dimension. For a nominal (target) size of 25.00 mm with a tolerance of ±0.05 mm, the acceptable range runs from 24.95 mm to 25.05 mm. Any measured part dimension that falls within this band passes inspection. Anything outside it is rejected.
Several core terms appear consistently in tolerance specifications:
- Nominal dimension: The intended target size stated on the drawing
- Upper limit: The maximum acceptable measurement
- Lower limit: The minimum acceptable measurement
- Tolerance band: The arithmetic difference between upper and lower limits
- Deviation: The difference between the actual measured size and the nominal dimension
- Datum: A reference point, line, or plane used as the origin for measurement
Three ways of expressing dimensional tolerances appear on engineering drawings:
Bilateral tolerance — the variation is split symmetrically around the nominal: 25.00 ±0.05 mm. This is the most common format and gives the machinist equal room on both sides of the target.
Unilateral tolerance — all allowable variation sits on one side of the nominal. A pin might be specified as 12.00 +0.00 / −0.02 mm, meaning it can only be at or below the nominal, not above it. Useful when interference or clearance in one direction is structurally critical.
Limit dimensions — the drawing states only the upper and lower limits directly, for example 24.95 / 25.05 mm, leaving the machinist to meet any dimension within that range.
The Two Main Tolerance Families: Dimensional and Geometric
Dimensional Tolerances
Dimensional tolerances govern the size of features: lengths, diameters, depths, widths, and thread dimensions. They are the starting point for most drawings and the first tolerances a machinist consults.
Geometric Dimensioning and Tolerancing (GD&T)
Size alone is often insufficient to fully describe how a part must perform. A shaft might have a diameter within tolerance but still be bent. A flat surface might be the right height but badly warped. GD&T addresses this by adding control over form, orientation, location, and runout through a standardized symbol language.
Common GD&T controls include:
- Flatness: How much a surface may deviate from a perfect plane
- Circularity (roundness): How closely a cross-section approaches a perfect circle
- Cylindricity: Combined control of diameter and straightness across a cylinder’s full length
- Perpendicularity: How precisely a feature sits at 90° to a reference datum
- True position: The allowed deviation from where a feature’s center must ideally be located
- Runout and total runout: How much a surface deviates as the part rotates around a datum axis — critical for rotating components such as shafts and bearing seats
GD&T is especially important in assemblies where multiple parts must align. Our CNC-Bearbeitungsdienstleistungen team reviews GD&T callouts as part of the drawing review process to flag any specifications that may create unnecessary cost or production risk before machining begins.
Standard Tolerance Grades: ISO 2768 and IT Grades
ISO 2768
ISO 2768 is the most commonly referenced general tolerance standard for CNC machined parts. It defines default tolerance grades for linear dimensions, angular dimensions, and geometric characteristics when the drawing does not carry individual feature tolerances.
The standard divides linear tolerances into four grades:
| Grade | Symbol | Application |
|---|---|---|
| Fine | f | High-precision parts; tight-fit assemblies |
| Mittel | m | General engineering; most machined components |
| Coarse | c | Fabricated structures; less critical dimensions |
| Very Coarse | v | Heavy structural work; rough machining |
Most machining shops, including Dimud, default to ISO 2768-m (medium) for metal parts and ISO 2768-c (coarse) for plastic components unless the drawing specifies otherwise. When tighter dimensional control is needed, individual feature tolerances are added directly to the drawing.
ISO 286 IT Grades
For shaft and bore fits — the precision pairing of cylindrical features — ISO 286 defines a system of tolerance grades called International Tolerance (IT) grades, ranging from IT01 (most precise) through IT18 (least precise).
In general machining practice, IT6 to IT8 covers precision fits such as those used in bearing journals and locating pins. IT9 to IT11 addresses standard clearance fits. IT12 to IT14 handles loose fits and general dimensions where assembly tolerance is generous.
Achievable Tolerances by Machining Process
Different machining operations inherently achieve different precision levels. Choosing the right process for the required tolerance is a fundamental part of cost-effective part design.
CNC Milling Tolerances
CNC-Fräsen removes material with rotating multi-point cutters across three or more axes. Dimensional tolerances achievable with modern 3-axis and 5-axis milling centers typically range as follows:
- Standard tolerance: ±0.1 mm — suitable for non-critical features, clearance fits, and cosmetic surfaces
- Medium precision: ±0.05 mm — the practical default for most structural and functional parts
- High precision: ±0.01 mm — requires controlled environment, rigid fixturing, and sharp tooling
- Near-precision: ±0.005 mm — achievable on specific features with careful process setup
Surface finish from milling typically falls between Ra 0.8 μm and Ra 3.2 μm depending on toolpath strategy and cutter geometry.
CNC Turning Tolerances
CNC-Drehen rotates the workpiece against a stationary cutting tool, producing cylindrical features with exceptional diameter consistency. For rotational parts, turning consistently outperforms milling in diameter control:
- Standard: ±0.05 mm — adequate for shafts, bushings, and stepped cylinders used in general assemblies
- Precision: ±0.01 mm — appropriate for bearing journal fits and connectors requiring controlled clearance
- High precision: ±0.005 mm — achievable on diameters requiring interference or precision transition fits
Live tooling on multi-axis turning centers allows secondary milled features to be added without re-fixturing, preserving the positional relationship between turned and milled features.
CNC Grinding Tolerances
CNC grinding uses abrasive wheels operating at high speed to remove minimal amounts of material with exceptional surface integrity. It is the process of choice when dimensional precision must reach the extreme end of what CNC machining can achieve:
- Standard grinding: ±0,005 mm
- High-precision grinding: ±0.002 mm
- Ultra-precision surface and cylindrical grinding: ±0.001 mm
Surface roughness from grinding can reach Ra 0.1 μm to Ra 0.4 μm — significantly finer than milling or turning. Components destined for grinding are typically rough-machined first, then heat-treated or hardened if required, and finish-ground to final dimension. Dimud’s CNC grinding capability supports precision mold inserts, bearing-class components, and high-tolerance functional parts in our Präzisionsformenbau workflow.
How Material Choice Affects Achievable Tolerances
Material behavior during cutting fundamentally limits how tight a tolerance can be held without driving up cost. Some materials machine predictably and hold dimension well; others introduce variability that requires additional process steps to manage.
Aluminum alloys are among the easiest materials for tight-tolerance work. They cut cleanly, generate lower cutting forces, dissipate heat efficiently, and exhibit stable dimensions once machined. Aluminum parts can routinely achieve ±0.01 mm and, with care, ±0.005 mm on critical features.
Stainless steel work-hardens during machining, generating heat that accelerates tool wear and causes thermal expansion of the workpiece. Achieving ±0.01 mm on stainless is feasible but requires sharper tooling, slower feeds, and more frequent inspection.
Brass and copper machine very freely with excellent surface finish but can smear at cut edges rather than shearing cleanly, which can affect thread and bore surface quality. Tight tolerances on these materials are achievable with the right toolpath strategy.
Titan combines high strength, low thermal conductivity, and chemical reactivity with cutting tools — a challenging combination. Holding ±0.02 mm on titanium parts requires careful tool selection, rigid fixturing, and controlled feed rates.
Engineering plastics (ABS, PC, POM, and nylon) present a different set of challenges. Internal stresses in stock material can relax during machining, causing warping. Plastics absorb moisture from the environment, causing dimensional change over time. For this reason, ISO 2768-c (coarse) is often the practical standard tolerance for plastic CNC parts, with tighter values reserved for features where function demands it.
The Direct Relationship Between Tolerance and Cost
One of the most consequential decisions a designer makes is which features actually need tight tolerances and which can be left to standard default values. The cost impact of unnecessary precision is substantial and often invisible until a quote comes back.
Here is why tighter tolerances drive cost:
Slower cycle times. Holding ±0.005 mm requires reduced feed rates, lighter depth-of-cut passes, and more conservative machining parameters throughout. The same part that takes 20 minutes at ±0.1 mm may take 60 minutes at ±0.005 mm.
More frequent inspection. Tight features require in-process measurement with CMM (coordinate measuring machine), air gauges, or precision bore gauges — all of which add time. Parts at standard tolerance may be sampled; ultra-precision parts are often 100% inspected.
Grinding as a secondary operation. When tolerances are beyond what milling or turning can reliably hold, grinding is added as a secondary process — adding equipment time, setup, and handling.
Specialized tooling and fixturing. Holding 0.005 mm requires rigid, thermally stable fixtures and precisely balanced cutting tools. These add setup cost, especially for low-volume work.
Higher scrap rate. Tighter tolerance bands statistically catch more non-conforming parts, increasing rework and scrap costs.
A practical approach: apply tight tolerances only to features that are functionally critical — bearing bores, locating pins, mating faces, and thread-fit diameters. Leave every non-functional dimension at the standard ISO 2768-m default. Submitting your design for DFM-Analyse (Design for Manufacturability) before production begins is one of the most effective ways to catch over-toleranced features early and reduce costs without compromising part performance.
Industry-Specific Tolerance Requirements
Automobilkomponenten
Herstellung von Automobilteilen demands consistency above all else. Components like housings, brackets, and shaft interfaces typically require tolerances in the ±0.02 mm to ±0.05 mm range for functional fits, with some bearing-class features needing ±0.005 mm to ±0.01 mm. Geometric tolerances for flatness of mating surfaces and true position of mounting holes are equally important, as automotive assemblies involve numerous sub-assemblies that must align precisely.
Herstellung medizinischer Geräte
Herstellung medizinischer Geräte operates under some of the most stringent dimensional and process requirements of any industry. Surgical instrument components, implantable device housings, and diagnostic equipment parts commonly require tolerances of ±0.005 mm to ±0.01 mm, combined with surface finish specifications as fine as Ra 0.2 μm to ensure biocompatibility and prevent bacterial adhesion. Full traceability of material, process, and inspection records is expected for any medical-grade component.
Elektronik und Halbleiter
For electronics and semiconductor applications, tight-tolerance CNC-machined components appear in sensor housings, connector interfaces, heat sink structures, and IC package guides. Positional tolerances on connector pin holes and interface flanges often specify ±0.01 mm to ±0.02 mm, as misalignment at the PCB level creates assembly failures downstream in production. Flat surface tolerances on heat-dissipating components are also critical for thermal interface performance.
Robotik und Energiespeicherung
Structural frames and joint interfaces in robotics require both tight dimensional tolerances and geometric controls — particularly perpendicularity and true position — to ensure smooth motion under load. Energy storage components like battery housing inserts demand consistent press-fit or clearance diameters to maintain reliable electrical and mechanical contact across production batches.
How to Specify Tolerances Correctly on Technical Drawings
Poor tolerance specification is one of the most common sources of manufacturing delays and cost overruns. The following practices produce cleaner, more manufacturable drawings:
Use a title block tolerance. Every drawing should define a general tolerance — typically ISO 2768-m for metal parts — that applies to all dimensions without an individual callout. This avoids ambiguity and simplifies the drawing.
Only call out tolerances tighter than the default where the function genuinely demands it. If a feature does not participate in an assembly fit or a functional interface, the title block tolerance is sufficient.
Specify critical fits using ISO 286 shaft and hole designations. A hole specified as Ø20H7 and its corresponding shaft as Ø20g6 immediately communicates the intended fit class (clearance, transition, or interference) to any competent machinist without requiring an additional note.
Accompany tight geometric tolerances with a clearly defined datum structure. A flatness or position callout is only meaningful if the inspection datum — the reference from which the measurement is made — is unambiguously identified on the drawing.
Avoid stacking tight tolerances across multiple features in an assembly. Tolerance stack-up occurs when the combined variation of several features in a chain produces an assembly that falls outside acceptable limits even though each individual part is in tolerance. Early-stage design review through CNC-Prototyping helps identify stack-up problems before they affect production parts.
Common Mistakes That Lead to Tolerance Problems in Production
Even experienced engineering teams encounter tolerance-related issues in production. The most frequent ones include:
Treating all features equally. Applying ±0.01 mm across an entire drawing when only two or three features actually need it inflates cost and extends lead time without engineering benefit.
Ignoring material behavior. Specifying the same tolerance for a POM plastic bushing and a stainless steel shaft without accounting for the different machinability and thermal expansion of each material leads to unpredictable fit conditions in service.
Omitting surface finish from critical features. A bore can be dimensionally in-tolerance but still fail functionally if its surface roughness is too high, causing accelerated wear or seal leakage. Tolerance and surface finish specifications work together.
Over-specifying angular tolerances. Angular features such as chamfers, taper angles, and mating flanges are often given tighter angular tolerances than their function requires, adding inspection complexity and cost.
How Dimud Maintains Tolerance Compliance Across Production
At Dimud, tolerance control is not a final inspection activity — it is embedded throughout the production workflow.
Pre-production drawing review identifies unclear specifications, over-tight tolerances relative to part geometry, and any GD&T callouts that require special fixturing or process planning before a single part is cut.
In-process dimensional verification uses calibrated measurement equipment — including CMM inspection for complex geometries and precision gauging for shaft and bore fits — to confirm that parts are running within the tolerance band before the full batch is completed.
First article inspection (FAI) on new parts validates that the production process consistently delivers dimensions within the specified tolerance range before volume production is approved.
Traceability through our ISO-aligned quality system links material certificates, process records, and inspection data to each production lot. Our full approach to quality management and certifications is detailed on our quality and certifications page.
This structured approach to tolerance management means that the precision your drawing specifies is the precision your delivered parts reflect — not just on the first sample, but consistently across every batch.
Häufig gestellte Fragen
For metal parts without individual tolerance callouts, the default at Dimud follows ISO 2768-m (medium grade), which for features in the 30–120 mm range corresponds to ±0.1 mm. More critical features are specified with tighter individual tolerances directly on the drawing.
Through CNC grinding in a thermally controlled environment, tolerances of ±0.001 mm to ±0.002 mm are achievable on specific features. Standard milling and turning operations reliably achieve ±0.01 mm to ±0.05 mm for most parts.
No. Tolerances should reflect functional requirements. An unnecessarily tight tolerance on a non-critical feature adds cost and lead time without improving part performance. The goal is to specify exactly the precision the function demands — no tighter, no looser.
Tight dimensional tolerances and fine surface finishes are often specified together because both influence how parts interact in assemblies. However, they are independently controlled. A part can be dimensionally accurate with a rough surface finish or dimensionally within tolerance with a very fine finish. Both parameters should be explicitly specified on the drawing when they matter.
The most useful submission includes a STEP or IGES 3D model accompanied by a 2D drawing with all critical tolerances called out, datum references for any GD&T controls, and material and surface finish specifications. The more complete the drawing, the faster and more accurate the quotation.
Work With a Team That Understands Tolerance from Design to Delivery
Tolerance management is where engineering intent meets manufacturing reality. At Dimud, our CNC machining team works with customers across the full tolerance spectrum — from standard general-engineering components to precision mold inserts requiring sub-10-micron dimensional control.
Whether your project is in the early design stage and needs DFM feedback on tolerance feasibility, or you have fully defined drawings ready for production, our engineering team is available to review your requirements, identify any specification risks, and provide a detailed quotation.
Contact Dimud for a CNC machining quote — upload your CAD files and drawings, and our team will respond with engineering feedback and pricing within 24 hours.
