Why Die-Cast Parts Fail During CNC Machining: The Hidden Connection Between Casting Quality and Machining Performance

01Machining is often only where the problem shows

When a die-cast part fails on a CNC machine, the immediate suspect is often the machining supplier: the wrong tool, excessive feed, poor fixturing, or an operator who removed too much material. Sometimes that diagnosis is correct. In many recurring cases, however, machining is only where a casting problem becomes visible. A hole that breaks through, a milled face full of pinholes, a thread that will not hold torque, or a housing that leaks after machining can all originate earlier — in metal flow, solidification, venting, die temperature, or part design. CNC machining exposes the casting's internal structure. It does not automatically create that structure. This distinction matters to both engineering and procurement. A supplier can hold every machined dimension and still deliver a failed component if the casting contains interconnected shrinkage porosity beneath the machined surface. Conversely, a casting with acceptable, non-critical internal porosity may machine economically if the process, stock allowance, and inspection plan are designed around it. The practical question is whether the supplier can control the casting so that the specified machining operation remains reliable.

02Die casting is near-net shape, not defect-free material

In the die-casting industry, high-pressure die casting fills a steel die rapidly under high pressure. This enables thin walls, complex geometry, and high production rates, but it also increases the risk of trapped air and localized solidification defects. In our production experience, optimized gating, venting, and flow simulation can reduce porosity, but they rarely eliminate it completely. According to NADCA, it is rarely practical or economical to eliminate every casting discontinuity. Acceptance should be related to the function of the component, not to an abstract “zero defect” expectation. Alloy choice is part of the same decision. Aluminum alloys differ in pressure tightness, density, strength, and machining ease; the alloy that is easiest to cut is not automatically the best choice for a pressure boundary.

03The dense skin is part of the machining equation

During solidification, the outer surface freezes first. According to NADCA, a typical dense, fine-grained skin is approximately 0.015–0.020 in (0.38–0.50 mm) thick. Porosity is more likely in the less-dense center of the wall. Removing more than about 0.020 in (0.50 mm) through secondary operations increases the chance of exposing that core porosity. This is the hidden connection between casting quality and machining performance. A machining operation is not removing uniform, homogeneous stock. It may be cutting through a sound skin into a region containing gas pores or shrinkage cavities. The result can be: · pinholes and open pores on a milled sealing face; · torn edges or tool chatter as the cutter crosses voids; · leaking drilled or tapped bosses; · unstable clamping because a datum surface is not continuous; · reduced thread pull-out strength; · cosmetic defects after anodizing, painting, or plating. According to the 2024 NADCA standard, the normal minimum machining allowance is 0.010 in (0.25 mm), intended to avoid excessive tool wear and minimize exposure of porosity. The actual allowance must also account for casting and machining variation, parting-line tolerance, flatness, and datum strategy. The guide also stresses that datums should exist throughout the process and that the best results come from locating the casting from datum points in the same die half as the feature being machined.

04Gas porosity and shrinkage porosity behave differently

Gas porosity is often relatively round and dispersed. According to NADCA, small, smooth pores — commonly less than 1 mm, with individual pores generally below 2 mm — may have limited effect on strength and may not leak. Shrinkage porosity is more irregular, can be interconnected, and is more likely to create a leak path when a drilled, bored, or milled surface opens it. That difference should change the inspection plan. A visual check of the as-cast surface cannot prove that a machined pressure boundary is sound. Depending on risk, appropriate methods include sectioning or simulated machining, X-ray/radiography, ultrasonic testing, dye penetrant inspection for surface-connected defects, and pressure testing. Agree on the detection method, acceptance level, and reference radiographs or sections in advance.

05The failure often appears several process steps late

The casting may pass visual, trimming, and as-cast dimensional checks, then fail when CNC machining removes the dense skin. By then, the lot has already incurred casting, transport, setup, tooling, and machining cost. In our manufacturing experience, buyers should treat this as a supply-chain problem: the relevant cost includes scrap, rework, leak-test failures, sorting, line stoppages, and expedited replacement parts. Post-casting machining details should be planned at the earliest design stages because they affect die options and final machining cost.

06Pressure-tight parts require a different conversation

According to NADCA, pressure-tightness depends on product design, die design, alloy, and process, and may require special testing or impregnation. Pressure-tight specifications are deviations from normal production practice and should be agreed upon before production starts. Do not write “leak-proof” as a general note and leave the supplier to interpret it. Define the medium, test pressure, test duration, temperature, fixture method, allowable leak rate, sampling plan, and whether testing occurs before or after machining. A casting can pass before machining and fail afterward because the machined surface intersects an interconnected pore network.

07The best machining process starts in the die design review

Discuss pressure tightness, secondary machining, and surface finishing before production tooling is designed. According to NADCA, machined areas should be identified so that gates, overflows, vents, chills, and process parameters can be arranged to minimize porosity in those locations. This is also why a small change in machining datum or stock can have a large production effect. A supplier that understands both casting and machining may move a gate, add a core, change a local section, or place a locator on the same die half as the machined feature. The change may cost more in tooling but save much more in recurring machining and scrap.

08Not only porosity: blank distortion and clamping spring-back

The failures listed above include unstable clamping on a datum surface that is not continuous. In our experience there is a further class of machining problem that also starts in the casting: the blank arrives at the machine already carrying residual stress and distortion. Uneven walls, unbalanced ejection and uneven cooling can all distort a blank before it is ever cut; machining only reveals it. The check is simple: measure the form error of the blank first. If it is already more than the machining stock can absorb, the problem is in the casting, and the fix belongs there too — more ejector pins, balanced ejection, adjusted cooling — rather than more parameter changes at the machine. The other is clamping distortion on thin-walled parts: the part measures good in the fixture and springs out of tolerance when released. On our projects this is usually not an operator problem but a mismatch between part stiffness and the way it is held; it calls for dedicated soft jaws, vacuum fixtures or multi-point supports that spread the clamping load onto the stiff sections of the part.

09Common mistakes

Treating the casting drawing and machining drawing as separate documents. The casting supplier needs to know which faces will be milled, how deep holes will be drilled, and which bosses will be tapped. If the machining plan is hidden until after the die is built, the die may place porosity exactly where the cutter must go. Specifying “zero porosity” without a functional limit. Zero porosity is usually not a measurable or economical requirement. Define critical zones, maximum pore size, spacing, density, or leak performance. A practical specification can state the maximum pore size, minimum spacing, and maximum number of pores per defined area. Removing excessive stock to compensate for casting variation. More stock is not automatically safer. It can remove the dense skin and expose the porous core. Solve excessive variation with better die control, datums, process capability, and casting tolerances rather than simply increasing the machining allowance. Inspecting only the as-cast part. If the risk is created by machining, inspection should reproduce the risk. Section a sample through the planned machined zone, use simulated machining, or perform NDT at the relevant location. Inspecting an unrelated surface provides weak evidence. Blaming cutting parameters for every broken tool or bad hole. Incorrect speeds and feeds matter, but recurring chip-out, sudden tool wear, or leaks concentrated at one feature can indicate casting variation. Review wall thickness, porosity, hardness, and batch history before repeatedly changing the CNC program.

10Our typical approach

Experienced suppliers treat the casting and CNC process as one manufacturing system rather than two separate operations. They connect part design, die design, casting controls, machining fixtures, inspection, and change management before production starts. This is how recurring porosity, datum, access, and cost problems are prevented instead of being discovered after machining. First, we mark the critical machined surfaces, holes, threads, sealing lands, and datums on a shared drawing. We then review wall thickness, bosses, ribs, core geometry, die-opening direction, and machining access. Cores and stepped cores can reduce finish stock, heavy sections, and porosity. Second, we use flow and thermal analysis where the risk justifies it. We select gating, overflows, venting, die cooling, and shot parameters to keep residual porosity away from critical machined areas. In our experience, early design review and mold-flow analysis are especially valuable for porosity-sensitive parts. Third, we validate the process with a representative pilot run. Validation should include as-cast dimensions, simulated or actual machining, surface inspection, thread checks, leak testing where applicable, and capability data for critical features. According to NADCA, castings may require separate gages before and after machining, while variable gages can document process data and Cpk. Finally, we control change. Alloy source, remelt practice, die temperature, vacuum performance, gate repair, fixtures, tool geometry, and machining depth can all change results. Our control plan identifies which changes require revalidation.

11From our projects: a twin throttle body

On a motorcycle twin throttle body we make, both throttle bores are held to ±0.02 mm on diameter, and the two bores on one part may differ by no more than 0.02 mm. Two typical machining problems came up in development. Machined in two separate setups, the bores differed by up to 0.05 mm: the datum-transfer error of the second setup, plus tool wear accumulating between the two operations, kept the difference from converging. And the 3 mm intake walls deflected elastically under normal clamping force — the bores measured good while clamped and sprang out by more than 0.03 mm when released. We now bore both throttle bores in one setup on a shared tool offset; we changed to dedicated soft jaws with controlled clamping force and measure each bore both clamped and released to compare spring-back; and the intake parting line was moved off the sealing band, with the flange face formed entirely in one die half. In stable production the difference between the two bores holds within 0.02 mm, first-pass yield on the leak test is above 98%, and key dimensions run at CPK ≥ 1.33.

12Questions buyers should ask

Before awarding a die-cast and CNC-machined program, ask: 1. Which machined areas are treated as porosity-critical, and how were they positioned in the die? 2. What maximum machining stock will be removed from each critical face? Is it consistent with the expected dense-skin depth? 3. What porosity type is being controlled: dispersed gas porosity, shrinkage porosity, or both? 4. What are the acceptance limits by zone — pore size, spacing, count, density, or leak rate? 5. Will the supplier provide a sectioned sample, simulated-machining result, or reference X-ray for approval? 6. Is improved venting, overflow design, local cooling, or a revised gate required? 7. What inspection occurs before machining and what inspection occurs after machining? 8. How are datums and fixtures designed to prevent tolerance accumulation and unstable clamping? 9. What evidence shows that the casting process is capable of meeting the CNC hole, flatness, thread, and leak requirements? 10. What process changes trigger customer notification and requalification? These questions separate a supplier that quotes a casting from one that owns the performance of the finished component.

SOURCES AND NOTE

[1]North American Die Casting Association (NADCA). NADCA Product Specification Standards for Die Castings: Aluminum, Aluminum-MMC, Copper, Magnesium, Zinc and ZA Alloys, 12th edition, revised for 2024. Key sections: Tooling for Die Casting and Porosity Control, pp. 2-13 to 2-15; Secondary Machining Preplanning, pp. 2-14 to 2-15; Machining Stock Allowance S/P-4A-13-24, p. 4A-34; Quality Assurance, including Porosity and Pressure-Tight Castings, pp. 7-12 to 7-16; Pressure Tightness G-6-1-24, p. 6-3.
[2]North American Die Casting Association. 2021 NADCA Aluminum Die Casting Alloys (2021). Used as a supporting reference for alloy selection and the relationship between alloy properties, pressure tightness, and machining performance.
[3]North American Die Casting Association (NADCA). Product Design for Die Casting: In Recyclable Aluminum, Magnesium, Zinc and ZA Alloys, 7th edition, 2015. Sections: Pressure Tightness, pp. 11-13; Skin Effect and Porosity, pp. 16-17; Machining, pp. 63-64.

This is a technical introduction and design guide, drawing on the sources listed above and on Jinxiong's many years of production practice. It is for reference only and is not a commitment for any specific project. For a specific project, the technical agreement between both parties governs.

No phone. No drawings. See your cost in 30 seconds.

Enter alloy, part weight and order quantity, and the tool splits the piece price into five lines — material, casting machine time, CNC machining, surface finishing and QC/packing — then prices it at 1,000 / 5,000 / 10,000 / 50,000 pcs. Upload a drawing only when you want the formal number: we reply within 24 hours and quote, tooling included, within 1–3 days.

Try the Cost Estimator ← More articles