Injection Molding Design Specifications

• 1
  Why Thin and Thick Walls Behave Differently
• 2
  Core Rules in the Injection Molding Design Specifications
• 3
  What Tolerance Is Realistic on a Molded Part
• 4
  Signs the Design Belongs on a Mill Instead
• 5
  Material Choice Changes the Numbers
• 6
  What Actually Drives Tool and Part Cost
• 7
  Molding vs Machining: Which Fits the Part

Molten polymer enters the cavity through a gate and freezes from the outside in. The skin solidifies against the cold tool steel while the core stays fluid under packing pressure. That pressure has to travel through the already-frozen skin to reach the far end of the part. If the flow path is too long relative to the wall, the melt front stalls and the part comes out short.

Wall thickness sets two things at once: how far the melt can travel before it freezes, and how the part cools. Thick sections hold heat longer than thin ones. The difference in cooling rate between a 3 mm boss and a 1.5 mm wall is enough to pull the boss inward and leave a sink mark on the opposite face. This is why uniform wall is the first specification to check, before draft or radii.

Shrinkage is the second mechanism. Every polymer shrinks as it cools, from about 0.4% for glass-filled nylon up to roughly 2.0% for unfilled polypropylene. The tool is cut oversize to compensate, but only if the shrink rate is known and consistent. Thick walls and high packing pressure push shrink higher; ribs and bosses pull it lower in those local areas.

A useful number for early layout is flow length over wall thickness. Unfilled ABS commonly reaches a 150:1 ratio, glass-filled grades less, and high-flow PC grades more. Under 100:1 the part is comfortable. Past 200:1 you are adding gates, raising melt temperature, or changing material.

Wall thickness should sit between 1.0 mm and 4.0 mm for most commodity resins. Below 1.0 mm the cavity is hard to fill and the part is fragile during ejection. Above 4.0 mm cooling time grows roughly with the square of thickness, so a 4 mm wall takes about four times as long to cool as a 2 mm wall. If a design needs stiffness at 4 mm, ribs at 2 mm usually do the job with less mass.

Draft is not optional. The part has to slide off the core without scuffing. One degree per side is the usual minimum on smooth vertical faces, 1.5 to 3 degrees on textured surfaces, because texture depth eats into the clearance. Deep ribs and tall bosses need more, often 2 to 5 degrees, since they grip the steel as they shrink onto the core.

Radii do two jobs. They remove the stress concentration at an inside corner, and they let the melt flow around the corner instead of hesitating. An inside radius of half the wall thickness is a common floor; a full wall thickness is safer. Outside corners should match the inside radius plus the wall, so the wall stays even around the bend.

Ribs and bosses follow simple ratios. Rib thickness should be 50 to 60 percent of the adjoining wall. Boss wall thickness should be about 60 percent of the main wall, and the boss should be tied to a side wall with a gusset or rib rather than standing alone. These ratios exist to keep the local mass close to the surrounding wall, which keeps the sink marks small.

Injection molding is not a precision machining process. Molding tolerances published by resin suppliers typically start around ±0.10 mm for small features and widen with part size, often quoted as ±0.10 mm plus ±0.002 mm per mm of dimension. Warp, shrink variation, and gate location all move the part after the tool is closed.

Tolerance is also direction dependent. Dimensions across the flow direction hold tighter than dimensions along the flow, because shrink along the flow path is larger and less predictable. A part measured 100 mm across might hold ±0.25 mm while the same 100 mm along the flow drifts more. Designers should put critical fits across the flow wherever the layout allows.

If a feature needs better than ±0.05 mm, molding is the wrong process for that feature. Three options exist. Cut the feature as a secondary machining operation after molding. Mold a clearance and let a machined insert define the fit. Or make the whole part on a CNC machine, where ±0.005 mm is routine and no tooling is required.

Shrink data from the resin supplier is a starting point, not a guarantee. The same grade behaves differently at different wall thickness, gate size, and hold pressure. For new geometries, a short tooling trial with first-article inspection is the only way to confirm the real shrink before committing to volume.

The clearest signal is a tolerance callout tighter than molding can hold. If a bore needs to sit within ±0.02 mm of a mating shaft, or two faces need to be flat within 0.03 mm, the molded part will need a machining step anyway. At that point, compare the cost of tooling plus secondary operations against cutting the whole part from billet.

The second signal is quantity. Tooling is a fixed cost that only pays back over volume. Below a few hundred parts, the amortized tool cost per part is usually higher than machining the same part directly. A well-programmed 5-axis center holds ±0.005 mm on a 4,000 mm envelope, so part size is rarely the limit.

The third signal is geometry. Undercuts, internal threads, side holes, and zero-draft walls all need side actions or lifters in a mold, which adds cost and maintenance. A 5-axis machine reaches those features in one setup without any of that hardware. If the part has three or more undercuts, the mold gets expensive fast.

The fourth signal is change rate. Early-stage products change every few weeks. A mold locks the geometry in; a machining program does not. Many teams machine the first 50 to 200 units, freeze the design, then cut a tool once the geometry has stopped moving. We do both in the same plant, so the transition stays in one quality system.

Amorphous resins such as ABS, PC, and PMMA shrink less and more uniformly, roughly 0.4 to 0.8%. They fill well but need higher melt temperatures and dry before processing. Semi-crystalline resins such as PP, PE, and POM shrink more, often 1.5 to 2.5%, and the shrink depends heavily on cooling rate. That makes tight tolerances harder to hold.

Fillers change everything. Adding 30% glass fiber to a nylon cuts shrink roughly in half and drops the thermal expansion, which helps dimensional stability. It also raises melt viscosity, so flow length drops and wall thickness may need to grow. Glass-filled parts also wear the tool faster, especially at the gate.

Drying matters more than most designers expect. PC and PA absorb moisture from the air, and wet pellets hydrolyze during melting. The result is splay marks, weak weld lines, and inconsistent dimensions. Drying at the supplier's recommended temperature and time is a process specification, not an option.

Surface finish interacts with material too. Glossy ABS shows every weld line and sink mark. Matte textures and low-gloss grades hide them. If the part has cosmetic requirements, pick the texture before the tool is cut, because changing texture later means re-cutting the cavity.

Tool cost scales with complexity, not part size alone. A simple two-plate tool for a small cover is a different object from a three-plate tool with side actions, hot runners, and a collapsible core. Each added action means more steel, more machining hours, and more maintenance points over the tool life.

Cycle time drives part cost more than material in most high-volume programs. Cooling time dominates the cycle, and cooling time follows wall thickness. Taking a 3 mm wall to 2 mm can cut cycle time by more than a third. That single change often saves more money than switching to a cheaper resin.

Gate location and count drive both quality and cost. More gates fill faster and reduce warp, but they add weld lines and tool complexity. A single gate is cheapest but limits flow length and can leave high orientation near the gate. The right answer depends on the flow ratio and the visible surfaces.

Inspection is part of the specification, not an afterthought. First-article inspection confirms the tool, and ongoing checks catch drift. At GreatLight we run raw material verification, in-process monitoring, and 100% inspection before shipment on both molded and machined parts, with reports on request.