Injection Molding DFM Checklist: 12 Mistakes That Kill First-Time Tooling Projects

Why this checklist exists

About 60% of injection-mold cost overruns we see come from DFM problems caught after steel is cut. Each problem is solvable in CAD in minutes, but solving it in hardened P20 or S136 means either a re-machine cycle (1-2 weeks, $2K-15K) or living with the defect for the life of the tool.

This checklist walks through the 12 issues that account for the bulk of those overruns. None require advanced FEA or moldflow simulation — they’re spot-checks any mechanical engineer can do on their own model in 30 minutes.

1. Wall thickness uniformity

The single most under-thought decision in plastic part design. Plastic shrinks as it cools — if some sections are 3 mm thick and others 1 mm thick, the thick sections cool last, pull material away from the thin sections, and produce sink marks, warp and internal voids.

Rule: Hold wall thickness within ±25% across the entire part. Target 1.5-3 mm for typical engineering plastics; thicker is rarely better.

Quick check: Color-map your CAD model by wall thickness (most CAD tools have this). Anything red (thick) or blue (thin) outside the body norm is a redesign target.

2. Insufficient draft angles

Without draft, a part stays stuck in the mold. Operators force it out, scratch the surface, sometimes break the ejector pins. Customers complain about scuff marks; we add 2 weeks to revisit the tool.

Rule: Minimum 0.5° draft on smooth walls, 1° on textured walls, 3° on heavy textures (VDI 30+). Add 0.5° more for every 25 mm of wall depth.

Common mistake: Designing the part flat and letting the toolmaker “add draft later”. The CAD model is the contract — draft has to be in the model before the quote.

3. Sharp internal corners

Stress concentrates at sharp inside corners, both in the part (cracks under load) and the mold (early steel fatigue). For the mold, a sharp corner requires EDM rather than milling, adding ~$300-1000 per corner and weeks to lead time.

Rule: Add fillets equal to 0.5× wall thickness on all internal corners. External corners can be sharper but a 0.25× fillet helps mold release.

4. Ribs that warp the part

Ribs add stiffness without adding weight or material — when sized correctly. When wrong, they create localized thick sections that produce sink marks on the opposite face.

Rule: Rib base thickness ≤ 60% of the adjacent wall thickness. Rib height ≤ 3× wall thickness. Always add fillets at the base.

Trap: Designers add a 4 mm rib to a 3 mm wall thinking “thicker = stronger”. Result: sink mark visible right through the wall on the cosmetic side.

5. Undercuts that need slides

Anything that prevents the part from pulling straight out of the mold is an undercut. Solving them requires side-action mechanisms (slides, lifters) — each one adds $2K-8K to tool cost and a failure mode for the life of the mold.

Rule: Eliminate undercuts in the mold’s pull direction whenever you can. Common workarounds:

• Move external snap features to a parting line.
• Replace internal threads with heat-set inserts (post-mold operation).
• Use through-holes instead of pockets where possible.
• Split the part into two snap-fit halves if it has too many undercuts.

6. Threaded inserts vs molded threads

Plastic threads work — for non-load-bearing covers, single-use packaging — but for anything that takes a fastener under load and gets disassembled more than 5 times, you want a brass or steel threaded insert.

Rule: For M3 and larger fasteners that see torque, design for press-in or heat-set inserts. Add the cylindrical pocket geometry to the CAD model with the insert manufacturer’s recommended bore depth and ID. Don’t just call out “tapped 4-40” on the drawing and assume the molder figures it out.

7. Sink marks and how to predict them

Sink marks are visible dimples on a face directly opposite a thick boss, rib or wall thickness change. They’re cosmetic killers on consumer parts.

Rule: Any boss diameter ≥ adjacent wall thickness needs to be cored out (hollow boss) or separated by a gap from the wall. Add a 0.5× wall thickness fillet under the boss to spread the heat.

Test in CAD: Project the boss centerline onto the cosmetic surface. If the boss is more than 60% of wall thickness, plan for a sink mark there.

8. Weld lines and where to put them

When the molten plastic flows around a hole, an obstacle, or a thin section that splits the flow, the two streams meet on the far side. They cool slightly before re-fusing — that line is the weld line. Strength at the weld line is 30-50% of base material strength and it’s visually visible.

Rule: Predict weld lines using moldflow, or estimate by tracing the flow front from the gate around obstacles. Avoid placing weld lines at:

• Cosmetic A-surfaces
• Threaded boss centerlines (where they crack first)
• Snap-fit cantilevers (where load is highest)

If you can’t avoid them, move the gate so they shift to a non-critical area.

9. Gate position and type

The gate is where plastic enters the cavity. Gate position controls flow pattern, weld line location, sink-mark distribution, and final part appearance. Gate type (edge, sub, hot-runner, fan, tab) affects scar visibility and how much hand-finishing is needed.

Rule: For cosmetic parts, specify gate location away from A-surfaces. Sub-gates leave a tiny dot inside the wall (great for cosmetic parts). Edge gates are cheapest but leave a visible vestige. Hot-runners eliminate runner waste but add ~$8K-20K to tool cost — typically pay back above 50K shots/year.

Don’t leave gating up to the toolmaker silently — the wrong gate becomes a 2-week argument when first-shot samples show ugly marks on the photo-shoot face.

10. Ejector pin marks

Ejector pins push the part out of the cavity. They leave small circular marks (~3-8 mm diameter) on the side they push against. These are visible — typically a faint depression or witness ring.

Rule: Specify on the drawing which surfaces are A-surfaces (cosmetic, no ejector marks) and which are B-surfaces (acceptable to push against). The toolmaker locates ejector pins on B-surfaces only. If every face is A-surface, you’ll need a custom ejection scheme — sometimes 2-3× the cost.

11. Cooling channel layout (the molder’s job, but check it)

Uneven cooling causes warp and dimensional variation between cavities. The mold needs cooling channels close to the part surface (typically 1.5-3× the channel diameter from the cavity wall) and balanced across the cavity.

You don’t design the channels — the toolmaker does — but you should ask to see the cooling layout before steel is cut. If channels are far from the part, cycle time will be 30-60% longer than it should be, and tooling cost may be $5K higher.

12. Venting

Air trapped in the cavity gets compressed by incoming plastic, heats up, and burns the polymer (visible as black scorch marks). The mold needs vents — small grooves at the parting line that let air escape.

Toolmakers handle this, but the part design affects how much venting is needed. Ribs in dead-end pockets are notorious — they trap air. If you can eliminate the dead-end (run a small relief or vent pocket through the rib), you eliminate a venting problem the molder would otherwise have to solve with extra parting-line grooves.

DFM in practice — the 30-minute self-check

Before sending any part for an injection-molding quote, run this 30-minute checklist on your model:

• Wall thickness colored map shows ≤ ±25% variation
• Pull-direction analysis shows ≥ 0.5° draft on every face (1° if textured)
• All internal corners filleted at ≥ 0.5× wall thickness
• Ribs ≤ 60% of adjacent wall, ≤ 3× wall height, with base fillets
• No undercuts unless absolutely necessary (count them)
• Threaded fastener locations have insert pockets, not molded threads
• Boss diameters checked for sink-mark risk on opposite face
• Cosmetic A-surfaces marked on the drawing
• Gate location requested (or left up to molder with documented preferences)
• Ejector-pin-acceptable B-surfaces marked

Most of these are quick visual checks once you’ve trained your eye. The ones that aren’t — flow analysis, weld-line prediction, gate optimization — are part of every injection molding DFM review we run with quotes.

FAQ

What’s the most expensive DFM mistake to fix after the mold is cut?

Wall thickness errors. Adding material to the cavity is impossible — you’d have to weld up the steel and re-machine the cavity, which means a 1-2 week tool revision cycle and $5K-15K in tool repair cost. By comparison, fixing a draft issue is usually a half-day re-machining job.

Can DFM be skipped for a soft tool prototype run?

Most DFM checks still apply — soft tooling (P20 / aluminum) has the same draft, wall, and ejection requirements as hard tooling. The cooling and gate optimization are looser because you’re only running 100-1000 shots. But skipping DFM on the part itself just means you’ll re-cut a soft tool and re-cut the production tool — paying twice.

How long does a typical DFM review take?

For a simple part (housing, bracket), our process engineers complete a DFM review in 2-4 hours. Complex parts with multi-shot, undercuts, or living hinges take 1-2 days. The output is a marked-up STEP file with annotations for every flagged issue, plus a written summary of the change recommendations and their cost / quality tradeoffs.

What CAD tools are best for DFM self-check?

SolidWorks has a built-in “DraftXpert” and wall-thickness analyzer that catches 70% of issues. Fusion 360 has a similar but lighter analysis suite. For more advanced work — flow simulation, cooling channel optimization — SolidWorks Plastics, Moldex3D and Autodesk Moldflow are the industry tools. Free option: SimScale’s web-based moldflow demo for a quick first pass.

Should I send my part for DFM review before or after I finalize the design?

Send it when the geometry is ~80% done but BEFORE you’ve spent a week tightening tolerances and adding texture callouts. We’ll catch the structural issues early when they’re cheap to fix, and you avoid wasting late-stage detailing work on geometry that has to change anyway. A second pass at 95% completion catches the smaller issues.