The quickest way to make an injection-molded part expensive is to leave one undercut in the geometry and let the moldmaker solve it for you. An undercut is any feature that hooks around the mold steel, so a plain two-plate straight-pull mold physically cannot open without tearing it off — and the standard fix is to add a moving mechanism (a side-action slide, a lifter, or an unscrewing core) that retracts before ejection. Each mechanism adds tooling hours, up-front cost, and long-term maintenance. The good news is that most undercuts on a typical product enclosure can be designed out so that a straight-pull tool forms the feature with no mechanism at all. This post walks through the six established ways to do it, the material limits on each, and how to catch every undercut before your file goes out for a quote.
None of these redesigns are exotic. They are the standard concessions that resin suppliers and moldmakers have documented for decades, and they all share one idea: change the geometry so that nothing in the part shadows another part of it along the direction the mold pulls. The hard part is almost never choosing the fix. The hard part is seeing the undercut in the first place, because the ones that cost you are usually on a face you cannot see from the outside of the model.
Why is an undercut worth redesigning before you quote?
Because the undercut you catch is a free file edit, and the one the moldmaker catches is a line item on the tool. When a straight-pull mold cannot open around a feature, the moldmaker does not send the part back — they quote it with a slide or a lifter built in, and that cost is baked into the tooling before you ever see it. A shop quoting the job has no structural reason to tell you the same feature could be re-drawn to drop the mechanism entirely; that is a redesign that costs them the sale of the more complex tool, not a dishonesty. (We wrote about how one detail like this inflates a quote in why one small design detail makes your injection molding quote so much more expensive.)
So the whole advantage is in the timing. While the part is still a STEP file, every fix below is a geometry change you make in your CAD tool for the cost of an afternoon. Once it is cut steel, the same fix is a mold modification or a re-cut. The whole point of finding undercuts early is to move the decision to the side of the line where it is cheap.
What actually counts as an undercut?
An undercut is any feature oriented so that it blocks the part from ejecting straight along the parting line — it "shadows" the mold steel. Side latches, external clips, snap-fit hooks, internal recesses, wire guides, threads, and windows in a side wall are all common undercuts. If you are unsure what separates an undercut from an insufficient-draft face, we cover both in draft angles and undercuts: what they are and why they break your part. The key property for this post is that an undercut is defined relative to the pull direction — which is exactly why several of the fixes below work by changing that relationship rather than removing the feature.
Below, the remedies are ordered roughly from the least invasive (a mold or parting-line concession that leaves your part looking the same) to the most invasive (splitting the part in two). In practice you work down the list and stop at the first one that fits.
Remedy 1: Pass-through shut-off (bypass steel)
Instead of a moving slide, extend the two mold halves so they meet and seal across the undercut, leaving a window in the wall where the overhang used to be. This is the most common way to lose a side-wall undercut. Rather than a horizontal slide cutting the latch hole or clip from the side, the core and cavity steel are designed to extend through the part and mate directly against one another — a "kiss-off" that keeps plastic out of the undercut zone and opens a hole or window directly beneath or beside the feature. When the tool opens, that steel simply withdraws straight out through the opening it created. It works well for side latch holes, external clips, wire guides, and shallow side-wall depressions.
The trade is that you accept an opening in the wall, and the sealing surfaces need care. To keep the mating faces from rubbing, galling, and flashing over time, the shut-off surfaces want a minimum 3 degrees of draft (5 degrees is preferred on the matching faces), and the transition needs about a 2.5 mm (0.1 in) step to absorb the tool's clamping tolerance. If your part can tolerate a window where the undercut was, this is usually the cheapest exit.
Remedy 2: Move or step the parting line
Because an undercut only exists relative to the pull direction, rotating the part in the tool or stepping the parting line can turn the same feature into a straight-pull one — no mechanism, no part change. Placing a feature directly on the parting plane lets the two halves split the geometry between them. External bosses, sloping wall steps, and external threads are the classic candidates: threads split across the parting line release cleanly on opening instead of demanding a hydraulic unscrewing rack. Stepping the parting line can also cut a deep feature's draw depth roughly in half, which relaxes the steep draft taper the full-depth feature would otherwise need.
This one is worth proposing to your moldmaker explicitly, because it is often a tooling decision they will make conservatively unless you have already shown that the feature can live on the split. Knowing where your undercuts are before the quote is what lets you have that conversation.
Remedy 3: Turn an internal undercut into a through-hole or window
An internal recess normally demands a lifter — an expensive, maintenance-heavy core that retracts inside the main core — but punching a hole through the adjacent outer wall lets a straight-pull core form the feature from outside instead. Internal side-wall latches, recesses, and mounting slots are the usual offenders. By opening a window in the outer wall, you convert the internal obstruction into a simple bypass shut-off: the core steel reaches in through the window, forms the underside of the feature, and pulls straight back out. Lifters are among the hardest mold components to build and keep running, so trading one for a window is often a large cost swing.
When you form slots this way, the core pins should seat about 3.18 mm (1/8 in) into the opposite cavity wall rather than just kissing it, so the resulting flash sits perpendicular to the surface and trims cleanly.
Remedy 4: Redesign the snap-fit hook with a core-access hole
Add a hole in the base wall directly beneath the hook, so a standing core pin from the other mold half can reach up, form the underside of the hook, and withdraw straight down through that hole. Cantilever snap hooks are the textbook undercut — the overhang locks onto the core steel. The standard fix is not to weaken the hook but to give the mold a straight path to it: the access hole under the hook lets the B-side pin form the catch surface and release on opening, with no sliding core. The hole must be at least as wide as the hook's overhang so the steel clears, and the mating faces that form the underside carry the same 3-degree shut-off draft as any bypass surface.
This is the single most common reason a snap-fit enclosure quotes with an avoidable slide. If your part uses snap hooks, this is the first geometry to check. Our teardown of a real open-source part shows how easily these hide — see the neutral DFM teardown of a real CAD part, where a finished-looking bracket carried 18 located undercuts.
Remedy 5: Strip it off (bumpoff) — within the material's limit
For a shallow, rounded undercut on a flexible enough resin, you can skip the mechanism entirely and let the part stretch over a fixed projection during ejection, then snap back. A bumpoff forces the part to deform over solid, non-retractable steel as it is pushed off, so the geometry has to be shallow, beveled or rounded, and — critically — sited away from ribs, gussets, and corners so the wall has room to flex. Get the resin or the depth wrong and you crack the part instead of releasing it.
The ceilings are resin-specific and small. Published resin-supplier design guides put the rough limits at:
| Material | Approximate strippable undercut | Notes |
|---|---|---|
| Acetal (Delrin) | up to ~5% of diameter | circular shapes only; bevel the lead edge |
| Unfilled nylon (Zytel) | ~6 to 10% | highly ductile |
| Flexible elastomers / TPE | ~5% (up to ~10% with air-assist) | needs room to deform freely |
| Polycarbonate, PC blends | ~1%, if attempted | stiff; stripping generally discouraged |
| Glass-reinforced grades | ~1% cool mold, ~2% hot mold | fibers restrict stretch |
Rectangular undercuts generally cannot be stripped at all, because the corners concentrate stress. Treat these as ceilings, not targets, and keep the undercut rounded.
Remedy 6: Split the part into two moldable halves
When an undercut is too deep or too rigid to strip and cannot be reached by any shut-off or parting-line trick, split the CAD model into two straight-pull-friendly pieces and join them in assembly. This is the fallback for closed-box geometries and complex enclosures, and it is often the right call when up-front tooling capital is tight, because two simple molds can be cheaper and faster than one mold with several actions. The pieces are joined after molding — commonly by ultrasonic welding (high-frequency vibration, typically around 20 kHz, using a tongue-and-groove or energy-director joint), by press- or snap-fitting, or by structural adhesive. If you ultrasonic weld, give the joint edges a generous radius — at least about 0.5 mm — so the vibration does not fracture a sharp corner.
Splitting adds an assembly step and a joint to design, so it is last on the list. But for a die-locked internal feature it is frequently cheaper over the life of the tool than the mechanism it replaces.
How do you decide which fix to use?
Work down the list and take the first remedy the feature allows, because the order runs cheapest-to-build first. In practice the decision looks like this:
- Is the undercut on a side wall you can open a window through? Use a shut-off (Remedy 1) or, for an internal recess, a through-hole (Remedy 3).
- Is it a snap hook? Add the core-access hole beneath it (Remedy 4).
- Can the feature live on the split, or would rotating the part fix it? Move the parting line (Remedy 2).
- Is it shallow, rounded, and on a flexible resin within the table above? Strip it (Remedy 5).
- Is it deep, rigid, and unreachable? Split the part (Remedy 6) or, if none of these fit, accept the mechanism knowingly.
The only prerequisite that runs through every branch is the same: you need to know exactly where each undercut is, on which face, and which direction the mold would pull. That is the information a redesign starts from, and it is the part a quick look at the model does not give you.
How do you find every undercut before you start redesigning?
By checking the geometry face by face against the mold-pull direction — the check a general once-over of the model reliably misses. Undercuts hide on the inner wall of a standoff, the back face of a cutout, and the underside of a snap hook: faces that are not visible from outside the part. A tool that walks every face against the draw direction does not skip them, and it hands you the exact face to open in your CAD program.
Fabdose reads your STEP or STP file on your own computer and locates, face by face, which features create undercuts and which walls fall below the draft threshold — before the file reaches anyone who profits from the answer. It covers the deterministic, geometry-only problems for injection molding: undercuts and draft. It does not simulate the flow-dependent defects — weld lines, air traps, sink marks — that depend on gate and cooling and need a manufacturer's DFM review or a flow simulation; a geometry check can flag those as a risk but cannot pin them to a face. Your CAD geometry is processed locally on your own computer; the defect findings and their descriptions are processed by AI. The free tier covers 5 analyses and 30 chats with no credit card; Pro is $35/month for unlimited analyses and a multi-process comparison report.
Once you have the list of undercuts and the face IDs, the six remedies above are the menu. The part that used to be a surprise on the quote becomes a set of edits you make first — while every one of them is still just a file.
Check your design with Fabdose
FAQ
Can you always eliminate an undercut, or do some parts really need a slide?
Most undercuts on a typical product enclosure can be designed out, but not all. Shut-off and through-hole tricks work when the undercut sits on a wall you can open a window through, or when you can move the parting line to reach it on a straight pull. When the feature is deep, buried inside a closed volume, and unreachable from any straight-pull direction, a side-action, lifter, unscrewing core, or splitting the part is the honest answer. The goal is not to avoid mechanisms at any cost — it is to avoid paying for one you did not need because nobody caught the undercut while the part was still a file.
How deep an undercut can you strip off the mold without any mechanism?
It depends almost entirely on the resin, and the limits are small. Published resin-supplier design guides put the rough ceilings at about 5% of the feature diameter for acetal, 6 to 10% for unfilled nylon, around 5% for flexible elastomers and TPEs (up to 10% with air-assisted ejection), and roughly 1% for rigid resins such as polycarbonate, where stripping is generally discouraged. Glass-reinforced grades sit near 1 to 2%. Stripped undercuts also need to be rounded or beveled, and for acetal roughly circular; sharp corners and rectangles concentrate stress and crack.
What is the difference between a shut-off and a side-action slide?
A side-action slide is a moving piece of steel that retracts sideways before ejection to form a side-wall undercut. A shut-off is not a mechanism at all: the two mold halves are extended so they seal against each other and close off the undercut zone, leaving a window or hole in the part instead. The mold opens straight with nothing to actuate, which is why a shut-off is cheaper to build and maintain; the trade is the opening in the wall.
Does moving the parting line change the part?
Often not. An undercut is defined only relative to the pull direction, so rotating the part in the tool or stepping the parting line can turn an undercut into a straight-pull feature with no functional change. External threads split across the parting plane are the classic example — they release without an unscrewing mechanism.
How do I find every undercut before I redesign?
Check the geometry face by face against the mold-pull direction, which is what a general look at the model misses. Undercuts hide on inner standoff walls, the backs of cutouts, and snap hooks. Fabdose locates each undercut and each zero-draft face by face ID on your own computer, before the file goes to a shop. The free tier covers 5 analyses with no card, and your CAD geometry stays on your computer.
Fabdose is a desktop tool for checking STEP and STP files against injection molding design rules. It locates draft violations and undercuts face by face, before the file goes out for a quote. It does not quote jobs or sell manufacturing. Your CAD geometry is processed on your own computer; defect findings and descriptions are processed by AI. The redesign remedies and material limits above are drawn from established plastics and mold-design references; treat the strip-undercut percentages as approximate ceilings that depend on your resin, wall, and geometry, not guarantees.