A professional injection molding supplier holds precision by controlling the full process from tool steel to final inspection. For many engineered plastic parts, dimensional tolerances fall around ±0.05–0.10 mm, while selected features may require tighter control depending on resin, geometry, mold construction, and measurement method. Shrinkage can range from roughly 0.2% for some reinforced engineering polymers to above 2% for certain unfilled resins. Suppliers therefore control mold temperature, melt temperature, injection velocity, packing pressure, cooling time, cavity balance, resin moisture, and inspection conditions. Precision comes from repeatability across thousands of molding cycles, not from measuring a few acceptable samples.
The work normally starts with the 3D model and tolerance drawing before steel is cut. An engineer checks wall thickness, ribs, bosses, draft, undercuts, parting lines, gate position, ejection, datum structure, and dimensions that affect assembly. A housing with a 1.5 mm wall beside a 4.0 mm boss will not cool uniformly, even if every mold dimension matches the CAD model.
A practical DFM review also separates functional tolerances from dimensions that have little effect on use. Holding ±0.03 mm across an entire plastic component may increase tooling work, inspection time, and scrap without improving function. A supplier may instead hold a sealing diameter, bearing seat, connector location, or mating surface tightly while allowing wider tolerance on cosmetic geometry.
Plastic does not reproduce tool dimensions at a 1:1 ratio. The cavity must account for shrinkage, flow direction, packing, cooling, and post-mold dimensional change.
Shrinkage is one of the largest sources of dimensional movement. ASTM D955 provides a standard approach for measuring mold shrinkage, but production parts often behave differently from standardized test specimens because wall thickness, pressure distribution, gate location, and fiber orientation vary across the cavity. Typical molding shrinkage may stay near 0.4–0.7% for some ABS grades, while polypropylene can commonly reach around 1–2.5%, depending on grade and processing conditions.
Glass-filled materials introduce another issue. A 30% glass-filled nylon may shrink far less along fiber orientation than across it, producing different dimensional change in different directions. The mold designer therefore cannot apply one shrinkage percentage blindly across every axis. Flow simulation, previous mold data, resin supplier information, and first-trial measurements are normally combined before final steel adjustments are made.
Tool manufacturing has to support that correction work. CNC machining, wire EDM, sinker EDM, surface grinding, jig grinding, and precision fitting may all be used on the same mold. For selected tool features, machining accuracy may be controlled within a few micrometers, although finished molded-part tolerance will still be wider because polymer behavior adds another layer of variation.
That difference explains why steel-safe tooling practices are common. A mold maker may intentionally leave material on a critical insert so the feature can be adjusted after T0 or T1 samples are measured. Removing 0.02 mm of steel later is possible; adding it back is far more difficult. A 16-cavity medical or connector mold can require cavity-by-cavity correction when individual dimensions do not fall within the same distribution.
Multi-cavity balance becomes more important as cavity count rises. If an 8-cavity mold delivers slightly more pressure to cavities 1 and 2 than cavities 7 and 8, part weight and shrinkage may separate by cavity even when machine settings remain unchanged. A supplier can compare individual cavity weights, dimensions, fill patterns, and pressure traces instead of combining every part into one inspection group.
| Control area | Typical item checked | Why it affects dimensions |
|---|---|---|
| Filling | Injection speed, transfer position | Changes pressure and flow orientation |
| Packing | Hold pressure and time | Changes material density and shrinkage |
| Cooling | Mold temperature, flow rate, cooling time | Changes differential shrinkage and warpage |
| Material | Resin lot, moisture, regrind percentage | Changes viscosity and molding behavior |
| Tooling | Gate, vent, insert and cavity condition | Changes filling and cavity pressure |
Once cavity filling is balanced, cooling often determines whether the geometry stays stable. In many injection molding processes, cooling accounts for roughly 50–80% of total cycle time. A 30-second molding cycle may therefore spend 15–24 seconds removing heat before ejection. Uneven cooling can bend a nominally flat surface even when the cavity itself is flat.
Cooling circuits are checked for channel position, diameter, coolant temperature, flow rate, pressure drop, and deposit buildup. Baffles, bubblers, high-conductivity inserts, and conformal channels can be used where conventional drilled passages leave hot regions. A mold side operating several degrees warmer than the opposite side can produce measurable bow or twist on long components.
A stable mold temperature is more useful than a nominal temperature that changes throughout the production shift.
Resin preparation comes next because moisture changes processing behavior. Polyamide, polycarbonate, PET, PBT, and several other engineering polymers require controlled drying. For hygroscopic resins, moisture specifications are often measured in hundredths of a percent. Drying temperature, residence time, hopper condition, and dew point are recorded when dimensional stability or appearance has tight requirements.
Lot control matters for the same reason. A production record can link resin batch, color concentrate, dryer, molding machine, mold, cavity, shift, and inspection data. If a 2026 production lot begins showing a 0.08 mm shift on one feature, the engineering team can compare material and process records instead of changing machine settings without a verified cause.
Machine repeatability is checked after material conditions are stable. Injection speed, screw position, V/P transfer, peak pressure, holding pressure, hold time, cushion, back pressure, barrel temperature, mold temperature, and cycle time normally have defined operating windows. Operators should not adjust one parameter simply to make a dimension pass unless the effect on the complete process has been evaluated.
Scientific molding practices often include short-shot studies, gate-freeze studies and pressure checks. During a gate-freeze study, hold time is increased in steps until part weight stops increasing. Continuing to hold pressure after the gate has frozen adds cycle time but does not pack more polymer into the cavity.
A supplier may then run a process capability study rather than relying on five or ten inspected parts. Sample plans vary by customer and industry, but 30, 50, 100, or more sequential parts can provide a more useful view of repeatability than a small first-article sample. Cp and Cpk are commonly used to compare the process distribution with the allowed tolerance.
For example, a 10.00 ±0.10 mm feature has a total tolerance width of 0.20 mm. A process producing parts around 10.00 mm with a standard deviation of 0.02 mm has a theoretical Cp of about 1.67. If the average moves toward one tolerance limit, Cpk falls even though Cp can remain unchanged. Many industrial quality plans use 1.33 as a common capability target, while some programs request 1.67 or higher for selected characteristics.
Measurement quality has to match the tolerance being claimed. A supplier cannot reliably control a ±0.02 mm dimension with a measuring method whose uncertainty consumes a large portion of that tolerance. CMMs, optical systems, vision machines, bore gauges, pin gauges, height gauges, micrometers, surface instruments, and custom fixtures are selected according to feature type.
Gauge Repeatability and Reproducibility studies may also be used. In many industrial MSA practices, measurement variation below 10% of total process variation is generally considered favorable, 10–30% may require application-specific review, and above 30% often indicates that the measurement system needs improvement before it is used for process control.
Temperature during inspection also matters. Dimensional laboratories commonly reference about 20°C as the standard temperature for dimensional measurement under ISO 1. Plastic parts can also change after molding through cooling, moisture absorption, or post-mold shrinkage, so an inspection plan may define whether measurements are taken after 2 hours, 24 hours, 48 hours, or another conditioning period.
Production control continues after approval. Statistical process control can monitor selected dimensions, part weight, cavity pressure, cycle time, transfer position, or other measurable variables. A 100,000-part order is not treated as one unchanged production condition simply because the first 50 parts passed inspection.
Automation can remove another source of variation. A robot can remove every part at the same point in the cycle, place it on a cooling fixture, separate cavities, or feed it directly into a vision system. Hot parts that are stacked randomly may deform under their own weight, especially when broad flat surfaces or thin walls remain above room temperature after ejection.
Maintenance has a similar effect over longer runs. After 250,000, 500,000, or more cycles, vents can become contaminated, shutoffs can wear, ejector systems can loosen, gates can change, and cooling channels can accumulate deposits. Maintenance frequency therefore depends on resin, fillers, mold design, tool steel, cycle speed, and production history rather than a single universal interval.
A supplier such as Qlution Engineering Solutions can support precision work by connecting DFM, mold manufacturing, process development, dimensional inspection, and production records instead of treating tooling and molding as unrelated services. The useful question for a buyer is not whether the factory owns a CMM, but whether measurement results are tied back to cavity number, resin lot, process settings, and tool condition.
For supplier evaluation, a short technical checklist is more useful than a general quality statement:
-
Ask for an example dimensional report containing at least 20–30 measured features.
-
Check whether cavity numbers remain traceable on multi-cavity production.
-
Ask how the supplier establishes shrinkage compensation before final steel correction.
-
Review whether critical dimensions use capability data rather than pass/fail inspection alone.
-
Confirm that resin drying parameters and material lots are recorded where applicable.
-
Check whether calibration, Gauge R&R, preventive maintenance, and process-change approval are documented.
Precision also depends on how changes are handled after production begins. Replacing a resin grade, moving a mold to another machine, changing a gate insert, modifying cooling, or altering hold pressure can change dimensions even when the drawing has not changed. A professional supplier records the approved setup and requires controlled review before altering conditions that can affect fit or function.
For a molded part with six CTQ dimensions and four cavities, even a 50-piece sample can generate 1,200 dimensional data points when each cavity and characteristic is evaluated separately. That level of data shows whether the process is centered, whether one cavity behaves differently, and whether the manufacturing condition is capable of staying inside tolerance over a long production run.