Die Casting draft angle optimization: balancing ejection ease with wall uniformity
Content Menu
● Draft Basics – What Actually Happens in the Die
● Factors That Move the Minimum Draft Number
● Practical Ways to Cheat Physics
● Real Programs That Pushed the Limits
● Simulation That Actually Predicts Ejection Force
● Surface Treatments Worth the Money
● Q&A – Questions We Get on the Floor Every Week
Introduction
Draft angles in die casting decide whether a part pops out clean on every shot or fights the die until something breaks. Most of us have been there: the toolmaker cuts 1.5° because “that’s what we always do,” then the first sampling run shows ejector-pin marks deep enough to fail cosmetics, or the walls vary so much the structural simulation no longer matches reality.
The real fight is rarely about having draft or not having draft. It’s about how little you can get away with while still keeping ejection forces low, die wear acceptable, and wall thickness consistent enough to pass the customer’s CTQ dimensions. In high-volume aluminum structural castings and thin-wall magnesium enclosures the difference between 0.7° and 1.3° can easily be tens of thousands of dollars per year in scrap, machining, or tooling maintenance.
This article pulls together lessons from actual programs, recent papers, and tricks that have worked on the floor when the designer refuses to add draft and the program manager refuses to add cycle time.
Draft Basics – What Actually Happens in the Die
When the metal solidifies it shrinks onto every core and away from every cavity surface. A 100 mm tall aluminum core sees the casting grab it with roughly 0.5–0.7 mm of radial shrinkage, depending on alloy and cooling rate. Without draft that shrinkage turns into a mechanical lock. The deeper the feature, the worse it gets – ejection force rises roughly with the square of depth if draft stays constant.
External walls are the opposite: the part shrinks away, so 0.1–0.3° is often enough unless the surface is heavily textured.
Factors That Move the Minimum Draft Number
Alloy matters more than most guidelines admit. AlSi9Cu3(Fe) – the workhorse 380/383 family – sticks harder than A356 or Silafont-36 because of the higher iron and copper. Zinc is forgiving down to 0.2–0.4°. Magnesium AZ91D sits between aluminum and zinc but releases beautifully once you break the initial vacuum.
Depth-to-draft ratio still rules. A good starting point that survives most production floors: internal draft in degrees ≈ (depth in mm)/70 + 0.3°. So 35 mm pocket → ~0.8°, 100 mm core → ~1.7°. Adjust down for zinc or coated dies, up for polished deep pockets in 380.
Texture and coating change everything. A standard EDM finish needs the full calculated draft. Polish to 0.2 Ra and you can usually shave 0.3–0.5°. CrN or DLC coating plus a good graphite spray can buy another 0.2–0.4°.
Practical Ways to Cheat Physics
Variable Draft Profiles
Instead of a constant taper, run low draft near the top where the casting has already shrunk away and increase it toward the root where grip is strongest. A linear ramp from 0.4° at the top to 1.8° at 60 mm depth kept wall-thickness variation under 0.22 mm on a battery tray side wall that originally varied 0.95 mm at constant 1.2°.
Strategic Core-Outs and Thickness Recovery
Add small core-outs or lightening pockets near the base of deep ribs so the average wall stays on nominal even with higher draft at the bottom. Looks ugly in CAD but disappears after machining or gets hidden by assembly.
Air Ejection and Vacuum Valves
0.5–0.6° becomes realistic on 70–90 mm cores if you add air poppets and proper venting. The air breaks the seal before the pins even move. Cycle penalty is 2–4 seconds, but on a $2 M tool running 800 000 shots the payback is fast.
Slide-on-Die vs Fixed Cores
Whenever possible put deep features on hydraulic slides or lifters. The slide retracts first, leaving almost no draft required on that face. Multi-slide zinc tooling principles applied to aluminum have saved several laptop and phone programs that demanded flush exterior walls.
Real Programs That Pushed the Limits
An EV drive-unit housing had 112 mm deep cooling jackets. Original 1.4° gave clean ejection but walls varied 1.1 mm top-to-bottom, failing fatigue correlation. Final solution: 0.55° average using variable draft (0.3° top 40 mm ramping to 1.1° bottom), CrN coating, and four air valves per core. Thickness variation dropped to 0.26 mm, ejection force 720 N per pin, zero marks after 1.4 million shots.
A telecom 5G base-station heat sink needed fins exactly 2.00 ± 0.15 mm for thermal performance. Fixed cores with 0.9° would have exceeded tolerance. Switched the tall fins to hydraulic slides: effective draft 0.15° on the critical faces, uniform thickness achieved, still running at 2.8 million shots.
Magnesium laptop bottom case – visible A-surface could not show any draft step. Used four-slide tooling (expensive but one-time cost) and ran 0.25–0.35° effective draft. Cosmetic approval on first tool trial.
Simulation That Actually Predicts Ejection Force
Modern flow + structural coupled simulation (Magma 5.5+, Flow-3D CAST, Ansys Mechanical) can predict ejection force within 15 % if you feed it the right friction curve versus temperature. Run the solidification, map the contact pressure, then simulate pin push with a temperature-dependent μ (0.3 at 180 °C down to 0.12 at 120 °C with good spray). Iterate draft until peak pin load stays below 65 % of buckling limit.
Surface Treatments Worth the Money
- CrN or WDLC on cores: 0.3–0.5° reduction, lasts 150 000–300 000 shots.
- Laser-textured micro-pockets that hold release agent: another 0.2–0.3° on deep cores.
- High-temperature nano-ceramic sprays: cut friction 20–25 % over standard graphite.
Conclusion
There is no magic minimum draft number that works for every part. The winning approach treats draft as one more design variable in a system that also includes coating, spray, air, slides, and selective machining. Start with the depth/70 rule, simulate ejection force and thickness variation, trial the tool, measure actual pin loads, then adjust.
The programs that achieve 0.6–0.8° on deep aluminum features while keeping walls uniform are the ones that refuse to accept the old 1–2° blanket rule and instead optimize every millimeter of the feature. Do the work up front – variable draft profiles, coatings, air, and clever parting choices – and you’ll spend less time fighting stuck parts and more time shipping on time.
Q&A – Questions We Get on the Floor Every Week
- Q: Brand-new 800-ton Buhler with vacuum – how low can I push draft on 85 mm aluminum 380 cores?
A: 0.65–0.75° with CrN and good vacuum/valve layout. Below 0.6° you’ll get occasional sticking once the coating starts to wear. - Q: Customer rejected 1.2° because walls are no longer uniform. Machining afterward is too expensive. What now?
A: Variable draft + small core-outs at the base to recover nominal thickness, or move the deep feature to a slide. - Q: Does die temperature affect required draft?
A: Absolutely. Running cores 20 °C hotter usually buys you 0.2–0.3° because shrinkage attachment is weaker. - Q: Any success with zero draft on internal walls?
A: Only on very shallow features (<15 mm) or with collapsible cores / lost cores. Otherwise forget it in high-pressure die casting. - Q: How do I measure whether my draft is marginal without waiting for 100k shots?
A: Install load cells on two ejector pins during sampling. If force rises more than 8–10 % over 200 shots, you’re too close to the edge.