Get a Quote
← Back to blog

Injection molding · DFM

5 CAD Mistakes That Delay Injection Molding Tooling

27 Jul 2026 · Bertrand Mezatio · 6 min read

Almost every injection-molded part I review for the first time has at least one of these five issues. None of them are exotic. They're not modeling errors in the traditional sense — the CAD is usually clean, dimensioned correctly, and looks finished. The problem is that "finished" and "moldable" are two different things, and the gap between them is where tooling delays come from.

Here's what shows up most often, in roughly the order it costs you money.

1. Draft angle left at zero

This is the single most common issue. A part is modeled with perfectly vertical walls because that's how it was sketched, and draft gets treated as a "we'll add it later" detail. The problem is that draft affects the geometry everywhere a wall meets a parting line, so adding it late often means reworking every connected face, fillet, and rib in the model — not just tilting a wall.

As a rule of thumb, most general-purpose parts need at least 1–2° of draft per side, and textured surfaces need more (often 3–5°, depending on the texture depth). If a face genuinely can't take draft, that's a signal the part needs a design change, not a tooling workaround.

2. Wall thickness that varies without a reason

Thick bosses dropped straight onto thin walls, ribs that are thicker than the wall they're reinforcing, sudden jumps in section thickness — these all create the same downstream problems: sink marks on the visible face, warping, and longer cooling times because thick sections take longer to solidify than thin ones.

The fix is almost always the same: core out material so local thickness stays close to a consistent percentage of the nominal wall (commonly 60–80% for ribs and bosses), and use a gradual transition instead of an abrupt step.

3. Undercuts that weren't necessary

An undercut isn't automatically a problem — sometimes the part genuinely needs one. But a lot of undercuts show up because a feature was placed without thinking about the mold-opening direction, not because the design required it. Every undercut that isn't essential adds a slide, lifter, or collapsing core to the tool, which adds cost and adds a maintenance point that can fail over the tool's life.

Before accepting an undercut, it's worth asking whether the feature could be reoriented, split into two parts, or redesigned to release in the primary pull direction instead.

4. Sharp internal corners

Zero-radius internal corners are easy to model and easy to miss, because they don't look wrong in a CAD viewer. But sharp internal corners concentrate stress right where a part is most likely to see load, and they can cause the melt to hesitate briefly during filling, which shows up as a visible line or weak spot in the finished part.

A small fillet — often 0.5–1.0mm depending on wall thickness — fixes this at effectively no cost, which makes it one of the easiest wins in a DFM review.

5. No draft or clearance plan for gate location

Gate location gets decided by the toolmaker if the designer doesn't specify one, and that's usually the point where "the part I designed" and "the part that gets built" start to diverge. Gate placement affects fill pattern, weld lines, cosmetic surfaces, and where any witness mark ends up — all things that are much easier to plan for in CAD than to renegotiate after a quote has already gone out.

The common thread: every one of these is inexpensive to fix while the part is still just a CAD file, and expensive to fix once a tool is being cut. That gap is basically the entire economic case for a DFM review before quoting.

Want these caught before your part goes to a toolmaker?

I review CAD models for exactly these issues — plus wall thickness, parting line feasibility, tolerance stack-up, and process-specific rules — and send back a priority-ranked report within 72 hours.

Send your CAD →