Choose an OEM injection mold supplier by comparing engineering review, mold specification, tool-life target, measurement capability, sampling records, production capacity, and after-sales support under the same project requirements. A high-production mold may be specified for 1,000,000+ cycles, while a four-cavity tool running a 30-second cycle can theoretically produce about 480 parts per hour before downtime. ISO 20457:2026 covers dimensional and geometrical tolerances for plastic molded parts, and ASTM D955-21 covers standardized thermoplastic shrinkage measurements at 24 and 48 hours. Price should be compared only after cavity count, steel grade, runner system, trials, inspection scope, spare parts, and warranty are aligned.
Start by sending every supplier the same technical package: 3D CAD, controlled 2D drawing, resin grade, annual quantity, color, surface specification, assembly interfaces, required tolerances, expected tool life, target cycle time, and molding location. A quotation that is 15% lower is not comparable when it excludes hardened inserts, two sampling rounds, spare wear parts, or dimensional inspection, so the shared specification becomes the basis for engineering review.
The engineering review should examine wall thickness, ribs, bosses, draft, undercuts, shutoffs, parting lines, gate position, venting, ejection, cooling access, weld lines, sink, and expected warpage. A nominal 2.0 mm wall beside a 4.0 mm boss creates a large local change in material volume, so the supplier should explain how geometry, coring, packing, or gate placement will be handled before machining begins.
Ask for comments on the actual CAD geometry rather than a generic DFM PDF. A useful note identifies a 0.5° draft, a 0.15 mm shutoff condition, a 3 mm rib, or a specific gate location and explains the proposed change.
That geometry review should lead into material behavior because shrinkage cannot be treated as one fixed number. ASTM D955-21 states that cooling time and applied pressure significantly affect molding shrinkage, covers measurements at 24 and 48 hours, and notes that standard specimens cannot predict absolute dimensions in real parts with different flow paths, wall thicknesses, temperature gradients, and pressure conditions.
A supplier should therefore work from the selected commercial resin grade, not a broad label such as “nylon” or “polycarbonate.” A 30% glass-fiber-filled grade can behave differently along and across the flow direction, and a gate moved 20 mm may alter local fiber orientation, filling pressure, weld-line position, and flatness; material experience then becomes more useful than a large portfolio of unrelated molds.
Compare previous work by technical similarity. A company may have built 500 polypropylene housings yet have limited experience with a precision glass-filled connector using multiple slides and tight datum relationships, while a smaller supplier with 40 comparable tools may understand the machining, sampling, wear, and dimensional issues more closely.
Useful reference projects should match several measurable conditions:
-
resin family, filler type, and filler percentage;
-
part size, weight, and nominal wall thickness;
-
tolerance range and inspection method;
-
cavity count and runner configuration;
-
annual output and expected tool cycles;
-
surface texture or polishing requirement;
-
slide, lifter, insert, or unscrewing mechanisms.
Once comparable experience has been confirmed, examine the mold specification rather than the outside appearance of the tool. PLASTICS mold-class guidance has historically defined Class 101 molds for 1,000,000 cycles or more, with requirements intended for high-production service, so expected cycle count should influence steel, hardness, wear plates, guided ejection, cooling, and replaceable components.
Tool life should still match the business case rather than automatically using the highest specification. A program needing 50,000 parts over two years does not require the same construction as one needing 750,000 parts per year for six years; the quotation should state proposed steel grades, hardness, cavity count, hot- or cold-runner design, surface treatment, and expected maintenance conditions.
| Quotation item | Information worth requiring |
|---|---|
| Core and cavity | Steel grade, hardness, heat treatment |
| Mold base | Material and standard |
| Cavities | Number and cavity identification |
| Runner | Cold runner or named hot-runner system |
| Cooling | Circuit layout and connection standard |
| Moving parts | Slides, lifters, cylinders, wear plates |
| Samples | Number of trials and sample quantities |
| Inspection | Dimensions, report type, sample count |
| Spares | Pins, inserts, seals, heaters, sensors |
| Service | Warranty scope and repair terms |
The specified construction should then be checked against actual manufacturing capability. Ask which work is done in-house across CNC milling, high-speed machining, EDM, wire EDM, grinding, polishing, fitting, assembly, mold sampling, and metrology; outside heat treatment or texturing is common, but the supplier should identify the process and inspection responsibility.
Equipment names alone do not establish accuracy, so the next check is measurement practice. ISO 20457:2026 was published in August 2026 and provides a plastics-specific framework for dimensional and geometrical tolerances because molded parts can show larger dimensional, form, and location deviations than metal parts due to shrinkage, geometry, processing conditions, warpage, and non-uniform cooling.
For a drawing containing ±0.05 mm dimensions, ask how each important feature will be measured, which datum setup is used, how parts are fixtured, and how long samples are conditioned before inspection. A 30-dimension report from one part only records one molded sample, so repeat-production approval should include an agreed sample count from every cavity and more than one molding period.
For an eight-cavity mold, measuring five parts from only cavity 1 gives little information about cavities 2–8. A 40-part study using five parts from each cavity provides a much better view of cavity-to-cavity differences.
Measurement results should connect to trial conditions, because a dimension can change when processing conditions change. T0 or T1 records should include resin grade, drying condition, melt and mold temperatures, injection speed, holding pressure, holding time, cooling time, total cycle time, part weight, cavity number, visible defects, and dimensional results; later T2 changes can then be compared against the first trial.
Cycle time deserves separate review once acceptable dimensions have been established. A four-cavity mold operating at 30 seconds completes 120 theoretical cycles per hour, or 480 parts; at 27 seconds it reaches about 133 cycles and 533 parts, an increase of roughly 11% before downtime and scrap are considered.
That 3-second difference should lead to a cooling-system review rather than simply demanding a faster press. Ask to see cooling circuits around deep cores, thick bosses, slides, and large surfaces, because circuit length, diameter, flow restriction, baffles, bubblers, inserts, and distance from the cavity surface affect mold-temperature uniformity during an 8-hour production shift.
Machine compatibility comes next because the approved mold must operate on the planned production press. Compare shot size, screw diameter, clamping force, injection pressure, tie-bar spacing, platen dimensions, minimum and maximum mold thickness, ejector stroke, nozzle interface, hot-runner controls, and robot clearance rather than selecting a press only by tonnage.
When evaluating an Injection molding manufacturer for OEM projects, request the proposed mold specification and molding-machine specification together. If qualification is performed on a 300-ton press but production later moves to a 500-ton machine, screw size, residence time, pressure response, platen behavior, and process settings may differ even when the same resin is used.
Quality-system review should follow machine capability. ISO 9001:2015 has been used across organizations in 189 countries, with ISO reporting more than one million certificates; ISO also lists a new 2026 edition as under publication in September 2026. Certification is useful, but project records should still show drawing revision control, calibration, resin identification, nonconformance handling, corrective work, and approved inspection methods.
Commercial comparison becomes meaningful after engineering and quality scope are aligned. If Supplier A quotes 12% less but includes one trial while Supplier B includes three trials, hardened slide inserts, a full dimensional report, and spare ejector components, the purchase prices describe different packages rather than the same mold.
A weighted review can prevent one quotation number from dominating the comparison:
| Evaluation area | Example weight |
|---|---|
| Engineering and DFM | 20% |
| Tool construction | 18% |
| Inspection capability | 15% |
| Similar project history | 12% |
| Sampling records | 10% |
| Price and payment terms | 10% |
| Capacity and schedule | 7% |
| Project reporting | 5% |
| Maintenance support | 3% |
The schedule should be checked immediately after scoring. An 8-week tooling promise should show design approval, steel purchase, rough machining, heat treatment where required, precision machining, fitting, assembly, T0/T1 sampling, dimensional inspection, corrections, repeat sampling, customer approval, and shipment; missing stages often explain unusually short quoted lead times.
Ask for weekly status using dates, completed operations, open technical items, and the next milestone. A photo of a mold base with “70% complete” provides less scheduling information than a record showing CNC finished on June 8, EDM due June 11, fitting due June 15, and T1 planned for June 19.
Ownership and maintenance terms should be written before the tooling deposit because production may continue for 5–10 years. The purchase order should state mold ownership, approved manufacturing location, transfer rights, CAD and mold-drawing availability, storage conditions, maintenance responsibility, warranty coverage, replacement-part pricing, and access to hot-runner or electrical documentation.
For a tool expected to exceed 1,000,000 cycles, request a maintenance schedule covering ejector pins, slides, lifters, shutoffs, gates, vents, leader components, seals, cooling channels, heaters, thermocouples, and replaceable inserts. Service intervals should be based on material abrasiveness, mold construction, cycle count, and inspection results, giving the buyer a practical basis for comparing suppliers beyond the initial tooling invoice.