
If injection-molded part design had to be reduced to one rule, it would be this: plastic wants to be the same thickness everywhere. Most sink marks, voids, and warped parts trace back to a design that ignored that preference. Here is the physics behind the rule, and the rib design practices that let you add stiffness without breaking it.
Why thickness rules everything
Molten plastic shrinks as it cools — every resin, every time. The molding process compensates by packing extra material into the cavity while the part solidifies. That compensation only works while there is a molten path to the area that is shrinking.
Thin sections freeze first. Thick sections stay molten longest, keep shrinking after the gate has frozen shut, and have nowhere to pull material from. The result is one of three defects:
- Sink marks — the surface collapses inward over the thick region as the core shrinks.
- Voids — the skin is rigid enough to hold its shape, so the shrinkage pulls a vacuum bubble inside the wall instead.
- Warp — different areas of the part shrink by different amounts or at different times, and the part bends to relieve the stress.
None of these are process defects a good molder can simply dial out. Process can moderate them; only geometry can remove them.
Uniform walls: the practical rules
- Pick one nominal wall and hold it. Most thermoplastics mold well somewhere between roughly 0.040 and 0.140 in (1.0–3.5 mm); the right number depends on resin, flow length, and the part's job.
- Where thickness must change, change it gradually. A common guideline is to transition over a length of at least three times the difference in thickness — a taper, not a step. Abrupt steps concentrate stress and create cooling differentials.
- Core out thick sections. A solid block of plastic serves no one. Hollow it from the non-cosmetic side and, if stiffness was the point, put it back with ribs.
- Respect flow length. Long, thin flow paths need pressure. If the wall is thin and the part is large, the resin choice and gate layout have to cooperate — a moldflow question, not a guess.
Rib design: stiffness without sink
Ribs exist to add bending stiffness without adding wall thickness. Designed well, they are nearly free; designed thick, they are sink marks waiting to happen — because a rib meeting a wall creates a locally thick intersection. The proportions below are the widely used starting points our engineers apply on design reviews:
| Rib parameter | Rule of thumb | Why |
|---|---|---|
| Thickness at base | 50–60% of the adjoining wall | Keeps the intersection thin enough to avoid sink on the opposite face |
| Height | ≤ about 3× wall thickness | Taller ribs buckle, fill poorly, and stick in the tool |
| Draft | 0.5–1° per side | Lets the rib eject without drag marks or sticking |
| Base radius | ~25% of wall thickness | Kills the stress concentration without fattening the intersection |
| Spacing | ≥ 2× wall thickness apart | Closely packed ribs trap heat and distort the cooling pattern |
Need more stiffness than one rib allows? Use more ribs, not thicker ones. A grid of well-proportioned ribs outperforms a few heavy ones — in stiffness per gram and in cosmetics.
Bosses: the same physics, concentrated
Screw bosses are the most common sink generator we see, because a cylinder of plastic meeting a wall is a thick section by definition. The same medicine applies: core the boss, keep its wall at 50–60% of nominal, attach it to walls with thin gussets rather than solid webs, and radius the base. If the boss must take real load, gussets — not mass — carry it.
A note on warp and filled materials
Glass-filled resins complicate the picture: fibers align with flow, and the material shrinks differently along the fiber direction than across it. A perfectly uniform wall can still warp if the fill pattern is asymmetric. This is exactly the class of problem that moldflow simulation answers before tooling — fill pattern, fiber orientation, and predicted deflection, in software rather than in steel.
The takeaway
Wall sections and ribs are decided in CAD, but their consequences are paid for in tooling, cycle time, and scrap — for the life of the program. The cheapest moment to apply any of the rules above is before steel is cut, which is the entire argument of early supplier involvement. Send us your CAD and we'll review wall sections, ribs, and bosses against the resin you actually plan to run — that review is how most of our best programs started.
