A professional injection molding supplier controls more than mold production. It reviews part geometry, resin behavior, shrinkage, cooling, tooling steel, process settings, inspection methods, and production capacity before volume manufacturing starts. Injection molding tolerances can reach about ±0.076 mm at the tool level in some commercial processes, while resin shrinkage adds further variation. Cooling may consume around 80% of a molding cycle, so mold design has a measurable effect on output and unit cost. A supplier that manages tooling, molding, measurement, maintenance, and secondary operations under one production system can reduce repeated sampling, scrap, dimensional variation, and schedule changes across programs running from thousands to more than 1 million parts.
Custom plastic parts rarely fail because the injection molding machine cannot fill a cavity. Problems usually begin earlier, when wall thickness, draft, gate position, ribs, bosses, tolerances, material shrinkage, or assembly requirements are not reviewed together. ISO 20457:2026 specifically addresses dimensional and geometrical tolerances for molded plastic parts because plastics behave differently from machined metals during cooling and shrinkage.
That difference becomes important before mold steel is cut. A housing with a 2.0 mm wall beside a 5.0 mm boss cools unevenly; the heavier section remains hot longer and can pull surrounding surfaces inward. Adding a radius, coring the boss, or changing wall transitions during DFM costs far less than welding and remachining a finished cavity after T1 samples reveal sink or distortion.
The same review should cover draft. Even a small draft change can affect ejection force, surface scuffing, texture release, and cycle consistency. A supplier should check which faces move with the cavity, which stay on the core, where ejector pins can contact the part, and whether cosmetic areas remain free from pin marks and parting-line mismatch.
The drawing tells the supplier what the part should measure. The mold design determines whether thousands of parts can repeatedly reach those measurements.
Tolerance planning follows the geometry review because molded dimensions depend on both tool machining and polymer behavior. Protolabs publishes a mold machining capability of about ±0.003 in., or ±0.076 mm, while noting that finished-part resin tolerance is separate and depends on material and design. Its published resin-related variation starts around 0.002 in./in. for dimensionally stable materials and can reach about 0.025 in./in. for less stable materials.
A professional supplier therefore does not treat every ±0.05 mm callout the same way. A sealing edge, bearing location, snap feature, PCB locator, or connector interface may justify a tight tolerance, while an internal rib or non-mating surface can often use a wider range. In 2026, ISO 20457 remains a useful reference because it separates general molded-part tolerance requirements from tighter directly specified dimensions.
Measurement must then match the drawing. A 30-part capability study provides far more production information than measuring one visually acceptable sample. Protolabs, for example, lists a 30-part dimensional capability study for selected production programs and a three-part dimensional first-article option for smaller checks.
Common inspection methods include:
-
CMM measurement for datum-based dimensions, position, profile, and complex geometry.
-
Optical systems for small edges, slots, profiles, and features that are difficult to contact.
-
Pin, plug, thread, and go/no-go gauges for repetitive production checks.
-
Calipers and micrometers for suitable linear dimensions.
-
Dedicated fixtures where an assembly relationship matters more than a standalone dimension.
Inspection cannot compensate for unstable molding, so process development comes next. Melt temperature, mold temperature, fill speed, transfer position, packing pressure, packing time, screw recovery, back pressure, cooling time, and resin moisture can all change part dimensions or appearance. Recording an approved processing window gives technicians a defined operating range instead of relying on repeated machine adjustment.
Cooling receives particular attention because it often occupies most of the cycle. RJG reports that cooling can account for roughly 80% of total molding time; one documented process example showed about 85% of the cycle associated with cooling before the part could withstand ejection.
That percentage explains why cooling-channel layout deserves the same engineering attention as cavity geometry. Channels placed too far from a thick section may leave one region significantly hotter than another at ejection. The result can be longer cycles, post-mold distortion, uneven shrinkage, or a need to hold the part in fixtures after molding.
A 30-second cycle produces 120 cycles per machine hour before normal stoppages. Reducing the same process to 25 seconds raises the theoretical rate to 144 cycles per hour, a 20% increase. With a four-cavity mold, theoretical hourly output rises from 480 to 576 parts without adding another molding machine.
Cycle speed only helps when quality remains stable, which leads to mold construction. Cavity steel, inserts, gates, runners, vents, ejector pins, slides, lifters, wear plates, hot-runner components, and water circuits need to suit expected production volume and resin selection. Glass-filled engineering polymers, for example, are more abrasive than many unfilled commodity plastics and may require greater attention to wear surfaces.
A supplier should also match cavity count to annual demand rather than automatically proposing more cavities. Moving from one cavity to four cavities can multiply output per cycle, but it also increases mold size, runner balance requirements, cooling complexity, upfront tooling cost, and the number of cavities that must remain dimensionally matched.
| Production item | What should be reviewed | Example measurement |
|---|---|---|
| Tool accuracy | Cavity, core, inserts, shutoffs | Around ±0.076 mm machining capability in some commercial services |
| Part consistency | CTQ dimensions across samples | 3-part FAI or 30-part study |
| Cycle performance | Fill, pack, cool, eject | Cooling may approach 80% of total cycle |
| Volume planning | Cavities × cycles × available hours | 4 cavities at 25 sec = 576 theoretical parts/hour |
| Scrap control | Rejects by defect category | 1% scrap = 10,000 parts per 1 million produced |
The last row becomes expensive at production scale. A 1% rejection rate may appear small during a 500-piece trial because it represents only five parts. At one million pieces, the same percentage becomes 10,000 rejected parts, plus consumed resin, machine time, inspection labor, handling, and replacement production.
Reducing rejection from 4% to 1% removes 30,000 rejected pieces from every one million produced. For a 20 g component, those 30,000 pieces represent 600 kg of molded material before runners or purge waste are counted. Cost comparison therefore needs scrap, cycle time, inspection, packaging, secondary processing, freight, and tooling service in addition to quoted piece price.
Material selection has the same scale effect. ABS, PC, PP, POM, PA, PC/ABS, TPE, PPS, and glass-filled grades have different shrinkage, moisture, stiffness, temperature performance, chemical resistance, and molding conditions. Selecting a resin from one strength number on a technical data sheet ignores processing behavior and the environment the finished part will see.
A supplier should ask whether the part will contact oil, fuel, detergents, UV exposure, outdoor moisture, elevated temperature, electrical current, or repeated mechanical loading. It should also know whether the molded surface will later be painted, plated, laser marked, ultrasonically welded, heat staked, bonded, or assembled with inserts.
Those downstream operations affect early mold choices. An ultrasonic-weld joint requires suitable energy-director geometry; a threaded insert needs enough surrounding plastic to handle installation and service loads; overmolding needs compatible substrate and overmold materials. Treating secondary work only after the mold is complete can force unnecessary tooling revisions.
For buyers comparing suppliers such as Qlution China Mold Supplier, the useful questions are technical: who performs DFM, who builds and maintains the mold, which molding machines are available, how resin lots are controlled, which dimensions are measured during trials, how tool changes are documented, and what production records remain available after shipment.
Documentation becomes more important as production continues. ISO 9001-based quality systems commonly use controlled drawings, revision records, inspection procedures, calibration records, nonconforming-material controls, and corrective-action records. ISO 20457:2026 adds a current international reference for molded-part tolerance and acceptance requirements.
A drawing revision from Rev. B to Rev. C should therefore reach mold engineering, production, inspection, assembly, and packaging before the next lot starts. Manufacturing from an obsolete dimension for even 8 hours can create thousands of unusable parts when a multi-cavity mold is running a 20- to 30-second cycle.
Tool maintenance continues the same control after production approval. Vents can become contaminated, ejector pins wear, slides lose lubrication, water channels collect deposits, gates erode, and textured surfaces can be damaged through poor handling. Maintenance intervals should reflect resin type, cavity count, mold construction, and accumulated cycles rather than an arbitrary calendar date.
A mold running 500,000 cycles at four cavities can theoretically produce two million molded pieces before scrap and downtime. At that scale, records for cleaning, wear inspection, replacement components, dimensional checks, and mold repairs make it easier to see gradual changes before they affect a large shipment.
Capacity should be checked with the same numbers. A four-cavity tool operating on a 20-second cycle completes 180 cycles per theoretical machine hour and produces 720 parts. At 85% usable production efficiency, output falls to about 612 parts per hour, so a requirement for 1 million parts needs roughly 1,634 production hours before additional allowance for maintenance, material changes, sampling, or unexpected stoppages.
Those calculations also help evaluate backup capacity. If one machine is allocated for 1,600 hours and the supplier has no compatible press available during maintenance, delivery depends on a single production path. A professional molding program should already define suitable machine tonnage, mold interfaces, auxiliary equipment, material-drying requirements, inspection frequency, packaging method, and the records required for each production lot.