Injection Molding: What Decides Whether a Molded Part Works

Tool design, part design and process window - the three things that determine cost, lead time and whether the part is still in tolerance at shot 200,000.

The mold is not the product

A common misunderstanding in sourcing: the mold is treated as the deliverable, and the parts as an afterthought. In reality the mold is a machine tool, and the part is what that tool produces under a specific process window. Get the tool right and the process wrong, and you get scrap. Get the process right and the part design wrong, and you get a mold that cannot be fixed without a re-cut.

Three things decide the outcome, in this order of leverage:

Part design - wall sections, draft, ribs, bosses, undercuts, gate location, tolerance callouts.

Tool design - gating, cooling layout, ejection, sliders and lifters, steel selection, cavitation.

Process window - temperatures, pressures, hold time, cooling time, and how much of that window the machine can hold repeatably.

The rest - machine brand, automation, inspection - determines how consistently you stay inside that window.

1. What is actually inside a mold

A mold is not a hollow block. Each element exists to solve a specific problem, and each one adds cost:

Core and cavity - the two halves that form the part's inner and outer surfaces.

Gating - where plastic enters. Gate type and location determine fill patterns, weld line positions and cosmetic marks.

Cooling - the channels that remove heat. Cooling typically dominates cycle time, and uneven cooling is the main cause of warp.

Ejection - pins, sleeves or stripper plates that push the part out without marking or distorting it.

Sliders and lifters - mechanical actions for undercuts that cannot be released along the parting line. Each one adds tooling cost, maintenance and a potential failure point.

Venting - where air escapes. Poor venting produces burn marks and short shots.

A quote that does not discuss gate location and cooling layout is not a tooling quote; it is a price for a block of steel with a cavity in it.

2. Part design rules that prevent tooling cost

Most expensive molds are the result of design decisions that looked harmless in CAD.

Uniform wall thickness. Thick sections cool slower, sink on the visible face, and create internal voids. Where a thick section is structurally required, core it out or redesign the geometry rather than accepting the sink.

Draft. Every vertical face needs draft to release from the tool. One or two degrees is usually enough, but zero draft means polishing marks, drag lines and, eventually, a damaged cavity.

Ribs and bosses. Ribs should be roughly half the nominal wall thickness to avoid sink on the opposite face. Bosses need support ribs or a nearby wall, and they should not sit in a thick section.

Undercuts. A snap fit that needs a slider adds cost to the tool and a maintenance item to production. If the undercut can be redesigned as a pass-through or a separate assembled part, do it.

Gate location. The gate decides where the weld lines land and where the part is strongest. Put it where the customer will not look, and where the flow front will not race through a thin section.

Tolerance. A molded part's dimensions come from a tool plus a process. Specifying ±0.01 mm on a dimension that spans a parting line is a request for a capability study, not a tolerance.

3. Material and shrinkage: the number everyone argues about

Shrinkage is not a single value. It varies with material grade, wall thickness, flow direction, gate location, packing pressure and cooling. Two families behave differently:

Amorphous materials (ABS, PC, PMMA) shrink relatively little and fairly uniformly, which makes them forgiving on dimensions but demanding on drying and cosmetic finish.

Semi-crystalline materials (PA, POM, PP) shrink more, and the shrinkage differs between flow and cross-flow directions - which is why glass-filled grades exist and why warp is more common.

The practical implication: shrinkage allowances are set in the tool, confirmed at T1, and adjusted by re-cutting. That is why design changes after steel is cut are expensive, and why a first-article inspection is not optional.

4. From T1 to stable production

The first trial is not a sample run; it is a diagnosis. A disciplined sequence:

T1 - fill and function. Does the tool fill, does the part eject, do the sliders move, are the critical dimensions close?

Cosmetic and dimensional review. Weld lines, sink, flash, short shots, warp - each mapped to a cause, not just noted.

Tool correction. Steel-safe changes where possible; welding or inserts where not.

Process window establishment. Find the window, not just a working point: the range of temperatures, pressures and times that produces good parts. A part made at one fragile setting will not survive production.

Capability check. Run a short production run and measure the critical dimensions across it. A dimension that is in tolerance but drifting is not capable.

Approval and handover. Signed samples, agreed inspection criteria, and a documented process sheet for production.

Skipping step 4 is the most common cause of a mold that "worked in the trial" and fails in the second week.

5. Consistency at volume is an equipment question

Once the tool and process are right, consistency comes from the press and the handling around it:

Tonnage matched to the part. Too small a press cannot hold the clamp force; too large a press wastes cycle time and can degrade material in the barrel. Our range runs 120T to 800T across 130 machines.

Repeatable process control. Shot-to-shot variation in melt temperature and hold pressure shows up as dimensional drift long before it shows up as a visible defect.

Automation. Robots for part removal and sprue handling reduce cycle time and eliminate the handling marks that come from manual extraction. We run 15 ABB robots on the molding floor.

Scheduling discipline. Production is scheduled in MES, with in-process inspection against sample drawings rather than a single end-of-run check.

6. Two-color and insert molding

Two-color molding produces parts with two materials or two colors in one cycle - hard-soft combinations for grips and seals, or color combinations for consumer products. It requires a two-color machine and a rotating or indexing tool, and it is worth it when the alternative is two parts plus an assembly step.

Insert molding places a metal insert - a threaded boss, a bushing, a contact - into the cavity before the shot. The result is a single part with a mechanically locked insert, which is stronger and more reliable than a post-installed one. Both capabilities change the design conversation: parts that would have been assembled can be molded as one.

7. What drives the price

The mistake buyers make most often is comparing unit prices without comparing cavity count, material grade and finishing scope. Two quotes that differ by 30% are usually not quoting the same part.

8. Common defects and what actually causes them

The pattern: most "molding problems" are tool or drying problems, not machine settings.

9. Lead time and what to send

Mold build typically runs 15–25 working days, followed by trial shots and approval; production is then scheduled in MES. Quotation with DFM feedback comes back within 8 working hours of receiving drawings.

To quote accurately, send:

3D file (STEP) and 2D drawing with tolerances and GD&T.

Material grade and color, with any certification requirement.

Cosmetic requirements - texture, gloss, class-A surfaces, and which faces are visible.

Annual volume and batch size, to set cavitation.

Assembly context - inserts, mating parts, secondary operations.

Target date, destination and packaging requirements.

The short version

Injection molding rewards early decisions. Design the wall sections and draft before the tool is quoted, decide the gate and cooling layout with the mold maker, and insist on a process window rather than a single working setting. The tool is expensive; the design decisions that determine whether it works are almost free.

About the manufacturer

Shenzhen Strongd Model Technology Co., Ltd. operates 130 injection molding machines from 120T to 800T - including Sumitomo and Haitian two-color machines - with 15 ABB robots on the floor, supported by an in-house mold division of 40 machines (high-speed CNC, mirror EDM, Sodick/GF slow wire-cut, grinders) and a full metrology lab. The 40,000 m² facility in Bao'an District, Shenzhen holds ISO 9001:2015 and IATF 16949:2016 certification, has exported for 14 years, and serves more than 10,000 enterprises.

Website: https://www.strongdcnc.com/