Why Composites Tolerances Are Different

If you are sourcing compression molded SMC parts or BMC components, you have probably noticed that tolerance discussions with suppliers do not follow the same rules as machined metal. A CNC-milled aluminum bracket holds whatever tolerance the machine can cut. A molded composite part, by contrast, has its dimensional behavior baked in at three stages: material formulation, tooling design, and process control. Industry experience suggests that mold design alone determines 60 to 70 percent of final part quality — which means by the time parts are coming off the press, most tolerance outcomes are already locked in.

The practical consequence for buyers: you cannot inspect tolerance capability into a composite part after the fact. You have to specify it correctly upfront, understand what the molding process can realistically hold, and build an acceptance plan around the right measurements.

Shrinkage: The Root Cause of Dimensional Variation

All molding compounds shrink during cure, and this shrinkage is the primary driver of dimensional variation in SMC and BMC parts. Two characteristics matter for procurement teams:

• Shrinkage is anisotropic. A typical polyester SMC shrinks 0.05 to 0.15 percent in-plane but 0.3 to 0.6 percent through the thickness. The part does not simply become uniformly smaller — it changes shape, which is why flatness and warpage callouts deserve as much attention as linear dimensions.
• Formulation dominates. The resin system determines the shrinkage band far more than press parameters do. Low-shrink and low-profile grades exist precisely to tighten dimensional control.

The table below summarizes typical shrinkage ranges by material system:

For context, a well-controlled BMC molding operation running at 150°C mold temperature with a 30-second hold can hold shrinkage within roughly 0.1 percent. Mold steel also expands measurably (about 0.012 to 0.014 percent over the working temperature range), and experienced toolmakers compensate for both effects when sizing the cavity: cavity dimension = part dimension × (1 + total shrinkage − steel expansion).

What Tolerances Can You Actually Specify?

Here is the honest answer experienced molders will give you, based on typical production outcomes:

• General dimensions: Approximately ±0.2 to 0.3 mm on parts up to 300 mm is a realistic band for well-maintained tooling. Do not blanket-apply ISO 2768 medium or tight machining grades to molded parts — that is a machined-metal standard, and quoting it will inflate your price or get silently ignored.
• Critical features: Bearing bores, sealing surfaces, and locating holes that must hold tighter than the molding band should be flagged for post-machining. CNC machining of key features after molding routinely holds assembly gaps within 0.05 mm — for example on precision electronics enclosures where fitted clearance matters.
• Large parts: Tolerance bands scale with size. On parts approaching one meter, expect ±0.5 mm or looser unless you accept machining costs and fixture-controlled measurement.
• Flatness and warpage: Often more important than linear tolerances on large panels. Because thickness shrinkage is three to five times higher than in-plane shrinkage, differential cooling across the part drives warp. Specify flatness on a stated support condition, not in free space.

The most common specification mistake buyers make is a uniform tight tolerance across every dimension on the drawing. Not only does this raise cost, it also drives the supplier to add post-processing everywhere — including on surfaces where machining marks create a new quality problem.

Defects That Show Up in Dimensional Inspection

Dimensional issues in molded composites are rarely random. They trace back to a handful of process causes, and recognizing the pattern helps you ask suppliers the right questions:

• Warpage and flatness deviation usually traces to uneven mold temperature. A variation of just ±5°C across the cavity produces visible gloss differences and 0.02 to 0.05 percent local shrinkage variation — enough to bow a large panel.
• Porosity and surface defects come from trapped volatiles. SMC generates 0.5 to 2.5 percent volatiles by weight, and vent design is the countermeasure: vent depths of 0.05 to 0.15 mm are typical. Vents deeper than about 0.2 mm produce flash; shallower than 0.03 mm cause gas porosity.
• Flash on parting lines indicates venting or clamp pressure problems, and it directly affects fitted dimensions at the split line.
• Dimensional drift over a production run often points to tooling wear (check whether the tool is P20 steel for under 50,000 cycles, or hardened H13 for higher volumes) or drifting cure parameters.
• Post-mold movement: Under-cured parts continue shrinking for days. Parts measured too early pass inspection, then drift out of spec in assembly. Cure verification is therefore part of dimensional control, not just a mechanical property question.

One countermeasure worth knowing about: for dynamically loaded parts, molders sometimes add a post-mold annealing step (holding at 150°C for around two hours) to relieve internal stress. The side benefit is dimensional stabilization — a useful question to ask if you see parts that were in-spec at incoming inspection but warped after heat exposure downstream.

Building a Practical Acceptance Plan

A workable dimensional acceptance plan for molded composite parts typically covers three layers:

1. First Article Inspection (FAI)

Full dimensional layout on the CMM, including GD&T callouts, before production release. Insist that critical-to-function dimensions and datums are explicitly listed, not just “per drawing.” The datum strategy should mirror how the part locates in your assembly, otherwise measured values do not predict real fit.

2. In-Process Checks

A short list of five to ten key dimensions checked each shift, with flatness on large surfaces. Beyond calipers and height gauges, a coordinate measuring machine catches the interaction effects (warp plus shrinkage) that single-point checks miss. Surface quality on visible parts is usually tracked as a roughness value — well-executed SMC and BMC surfaces routinely achieve Ra ≤ 0.8 μm, which matters not only for appearance but also in electrical housings where smooth surfaces reduce corona discharge risk.

3. Periodic Verification

Material-related dimensions deserve occasional root-cause level checks: thermal cycling stability (a well-formulated BMC battery box shows under 0.08 percent dimensional change across −40°C to +85°C cycling), warpage trend monitoring, and cure-degree verification by DSC if drift issues appear. A capable supplier will already run most of these as internal controls; a professional third-party test menu for BMC parts includes shrinkage rate, dimensional stability, warpage, CTE, water absorption, cure degree, and surface roughness.

Cost Consequences of Tightening Tolerances

Every step down in tolerance band adds real cost, and knowing where the money goes helps you decide where tolerance actually matters:

The cost-smart approach is almost always the same: classify dimensions into critical, functional, and non-critical; hold tight tolerances only where assembly or performance truly requires them; and let cosmetic and non-functional geometry run at the process’s natural band. Suppliers respond well to drawings that show this discipline, because it demonstrates you understand the process.

Questions to Ask Your Molding Supplier

• What shrinkage value was assumed when the cavity was cut, and what does end-of-run measurement show?
• How is mold temperature uniformity controlled and verified across the cavity?
• What is the venting strategy, and what flash standard applies at the parting line?
• Which dimensions are post-machined, and what is the measurement method and frequency for each?
• What dimensional stability data exists across the service temperature range?
• How long after molding are parts measured, and is cure verified before dimensional release?

A supplier who answers these concretely — with numbers, not assurances — is one whose tolerance capability you can build a program around. This is equally true whether you are molding SMC, BMC, GMT, or RTM components: the material changes, but the logic of shrinkage compensation, venting, and staged acceptance does not.

Need help defining a realistic tolerance and inspection plan for your composite parts? Contact our engineering team with your drawing and annual volumes — we will tell you what the molding process can hold as-molded, where post-machining earns its cost, and how to structure acceptance criteria your quality department can defend.