Die casting draft design refinement: optimizing taper angles for easy ejection while keeping wall thickness stable

Content Menu

● Introduction

● What Actually Drives Ejection Force

● Wall Uniformity and Its Effect on Quality

● Simulation-Based Optimization in Practice

● Material Differences You Can’t Ignore

● Tooling Tricks That Buy You Draft Reduction

● Rules of Thumb I Actually Use

● Conclusion

● Q&A – Questions I Get Every Week

Introduction

Draft angles in die casting are one of those details that separate tools that run 200,000 shots without a hiccup from tools that start flashing or galling after 20,000. I still remember the first time I had to explain to a program manager why we couldn’t just make the walls straight like the solid model showed. The part was an aluminum transmission case extension, 110 mm deep, and the OEM wanted zero draft for packaging reasons. We ended up compromising at 0.35° on the cover side and 1.4° on the moving-side cores, and it ran fine for 1.2 million pieces. That experience taught me that draft isn’t a fixed number from a handbook anymore – it’s a variable you tune once you understand how it fights with uniform wall thickness.

The basic conflict is simple. Larger draft reduces the force needed to push the casting off the tool, but it forces the wall to taper. In a 50 mm deep pocket a 1° draft on one side changes wall thickness by about 0.87 mm from top to bottom. If your target wall is 2.8 mm, the bottom section ends up 3.6 mm or thicker, which cools slower and almost always shows sink or porosity on the opposite face. Too little draft and the casting hangs up on the cores, you break pins, mark surfaces, or tear metal and create cold shuts on the next shot.

Over the last decade the acceptable range has moved dramatically downward because of better die coatings, more precise temperature control, and most importantly, simulation that can predict ejection loads before the first steel is cut.

What Actually Drives Ejection Force

When the aluminum solidifies it shrinks onto cores and away from cavity walls. The grip stress on a core is roughly σ = EαΔT × (shrinkage allowance), where E is the modulus at ejection temperature and α is the CTE. For A380 at 250 °C that grip can easily hit 30–50 MPa on a plain core with no draft. The ejection force is then that stress times the contact area times a friction coefficient that’s usually 0.4–0.5 with standard die spray and higher if the spray is lean.

Add 1° of draft and the normal force drops because the casting can slide instead of scrape flat against the steel. Most of the benefit happens in the first degree – going from 0° to 1° can cut ejection force 60–80 %, while going from 1° to 2° only gives another 15–20 %. That’s why chasing the last half-degree usually isn’t worth the extra wall variation it creates.

We saw this on a powertool housing in A380. The original tool had 1.8° uniform draft on 42 mm tall ribs. Ejection was fine, but the ribs were 2.2 mm at the tip and 3.4 mm at the base. Opposite those thick bases we had sink marks that failed cosmetic inspection. We rebuilt the cores with 0.6° draft for the top 25 mm and 1.6° for the bottom 17 mm. Wall thickness stayed 2.6–2.9 mm everywhere, sink disappeared, and peak ejector-pin load dropped from 11 kN to 7 kN per pin.

Wall Uniformity and Its Effect on Quality

Uniform wall thickness matters because solidification time goes with thickness squared. A section that’s 30 % thicker takes almost twice as long to solidify. That late-solidifying mass feeds shrinkage in the thinner areas nearby and leaves porosity bands exactly where you don’t want them – usually in load-bearing ribs or sealing flanges.

A classic example is a laptop bottom case in magnesium AZ91D. The designer started with 1.5° draft on all side walls. At 65 mm draw depth that gave a wall thickness change of 1.7 mm top to bottom. The bottom 15 mm was 2.4 mm instead of the target 1.5 mm and we had 8–12 % porosity in the corners. We changed to 0.4° on the cavity side and 1.8° on the core side only, plus a 4 mm straight land at the parting line. Effective wall became 1.52–1.61 mm everywhere. Porosity dropped below 1 %, and we shaved 9 grams off each part.

Another real part was a 5G base-station filter housing in AlSi10Mg. The cavities were 180 mm deep with 3.0 mm nominal walls. Uniform 1° draft would have made the walls 6.2 mm at the bottom – completely impossible for thermal performance. We used asymmetrical draft (0.25° cavity side, 2.2° core side) and the wall stayed 3.0 ± 0.08 mm from top to bottom. Leak testing went from 14 % reject to 0.3 %.

Simulation-Based Optimization in Practice

These days we rarely guess. The workflow that works best is:

Build the CAD with zero draft on non-critical surfaces and minimal draft where you suspect problems.
Run a full filling + solidification + stress analysis (MAGMA, FLOW-3D Cf or AnyCasting).
Look at the “core grip stress” or “ejection force” results contour plot. Anything over 40 MPa on aluminum cores is a red flag.
Parameterize the draft angles (often 5–10 separate zones) and let the optimizer minimize peak ejection stress while constraining wall thickness deviation to ±0.1 mm or whatever your spec allows.
Verify with a couple of manual runs because optimizers sometimes find weird local minima.

On an EV inverter housing we did exactly that. The optimizer settled on 0.18° for the shallow areas, ramping to 1.65° only in the last 40 mm of the 120 mm deep pockets. Total ejection force dropped 78 % compared with the original 1.5° uniform design, and wall thickness variation was 0.07 mm – better than the uniform-draft version by far.

Material Differences You Can’t Ignore

Zinc lets you get away with murder – 0.2–0.4° external is routine because shrinkage is low and the alloy is soft on the tool. We run some connector shells at 0.15° total draft included and they eject at 25–30 bar.

Aluminum is pickier. A380/A383 need 0.5–1.0° external minimum on a new tool, but with CrN or DLC coatings and good spray control we’re down to 0.3–0.4° on several current programs.

Magnesium AZ91D and AM60 shrink more and stick harder – plan on 0.5–0.7° more draft than aluminum unless you love rebuilding cores.

Tooling Tricks That Buy You Draft Reduction

PVD coatings (DLC, CrN, TiAlN) drop friction coefficient to 0.15–0.20 and let you cut draft 30–50 %.
Core temperature 20–30 °C higher than cavity reduces grip dramatically.
Polished cores (SPI A1 or better) plus frequent nitriding keep the surface slick.
Stepped draft or small straight lands at the top give you a running start for ejection.
Texture adds 1° draft per 0.03 mm of depth – don’t forget it.

We once took a cosmetic cover that needed leather grain and was failing ejection with 2° draft. Added DLC to the cores, raised core temp 25 °C, and dropped draft to 1.4°. It ran 850,000 shots with only one polish.

Rules of Thumb I Actually Use

External walls (cavity): start at 0.3–0.5° for aluminum, add 0.2° if textured.
Internal cores: 0.8–1.5° minimum, higher for magnesium or depths >80 mm.
Never let wall thickness vary more than ±12 % unless you have feeding designed for it.
Anything deeper than 40 mm should have stepped or variable draft.
Always simulate ejection before you sign off on the model.

Conclusion

Draft angle optimization has gone from “look it up in the NADCA table” to a proper engineering trade-off you solve with simulation and a few clever tooling choices. The parts that win cost-down exercises and quality awards are the ones where the team spent the time to make walls uniform first, then figured out the absolute minimum draft that still lets the casting come out cleanly every shot.

Next time you’re reviewing a new die cast model, don’t accept blanket 1° or 1.5°. Ask how deep the features are, what the wall tolerance is, and whether anyone has run an ejection analysis. The difference between a good tool and a great one is usually a few tenths of a degree applied in exactly the right places.

Q&A – Questions I Get Every Week

Q: Can we run 0° draft on external walls?
A: Yes, up to about 15 mm depth on polished, coated cavity surfaces. Beyond that you start dragging and marking.
Q: How do I convince the designer to accept more draft on cores?
A: Show them the ejection force plot and the broken-pin repair bill from the last program that tried zero draft.
Q: Is variable draft expensive to machine?
A: Five-axis machining and EDM make it almost free compared to the material and quality savings.
Q: Does die temperature affect draft requirements?
A: Yes – hotter cores = less grip = less draft needed. 20–30 °C difference is worth 0.3–0.5°.
Q: When should I just add ribs instead of more draft?
A: When the wall variation from draft would exceed ±10 % and you can’t use asymmetrical or stepped draft.