Porosity in Die Casting: What Causes It, Why It Matters After Machining, and How Suppliers Control It

01Machining only uncovered it

A die-cast housing can look acceptable after trimming and visual inspection, then show pinholes across a machined sealing face. A threaded boss can pass an as-cast check and leak after drilling. A milled surface can reveal a cluster of cavities that was never visible on the original part. These failures are often blamed on the CNC operation. The more useful diagnosis is that machining exposed a condition already present inside the casting. CNC machining removed the material that was hiding it. That distinction matters. If the problem is treated as a cutting-parameter issue, the team may change tools, feeds, or fixtures without addressing metal flow, solidification, venting, local wall thickness, or the location of the machined surface. The result is repeated scrap and an argument between the caster and the machine shop. This article focuses on the production question buyers and engineers actually face: why does porosity become visible after machining, and what should a capable supplier do before that happens?

02Porosity is a casting condition, not a cosmetic defect

Porosity means voids inside or connected to the surface of a casting. In high-pressure die casting, the two most important forms are gas porosity and shrinkage porosity. Gas can be trapped when fast metal flow creates swirls, backfills, or insufficiently vented pockets. Shrinkage porosity forms when a hot, heavy section continues to solidify but cannot receive enough liquid metal to compensate for volume contraction. According to NADCA, porosity is one of the most common internal discontinuities in die castings, and it is rarely practical or economical to eliminate every discontinuity. The correct acceptance level depends on what the part must do. A non-pressure-bearing internal zone can tolerate a condition that would be unacceptable in a sealing land, a threaded boss, or a fatigue-critical section. The Gating Manual makes the process link explicit: shrinkage porosity is associated with the last areas to solidify, while gas porosity is strongly influenced by flow pattern, venting, and vacuum. If the gating system freezes too early or delivers metal to the wrong location, the defect can remain inside the part even when the external surface looks sound.

03Why the defect is hidden in the as-cast condition

Die casting solidifies from the die wall inward. The outer layer cools rapidly and forms a relatively dense, fine-grained skin. According to NADCA, this skin is typically about 0.015–0.020 in (0.38–0.50 mm) thick. Porosity is more likely in the less-dense center of the wall. The skin acts like a cover over the defect. A visual inspection sees a continuous surface. A dimensional inspection may also pass because the pore is below the surface and has not changed the external geometry. The part is not necessarily sound; it is simply unsectioned. When a cutter removes the skin, the internal condition becomes a surface condition. A single pore may appear as a pinhole. A group of pores may look like a rough patch or cratered area. An interconnected shrinkage network may become a leak path through a drilled hole or across a sealing face.

04From our experience: three kinds of void, and sinks versus shrinkage

When we investigate a defect we sort voids into three kinds, because their origins and their remedies are entirely different. Entrapped-air pores form when turbulent filling wraps air into the metal; they are fairly round with smooth walls, and tend to sit in the last-to-fill regions or where the stream strikes. Precipitated gas pores come from hydrogen dissolved in the melt coming out of solution as it freezes; they are small and widely dispersed, and trace back to melting and degassing. Shrinkage cavities form where solidification shrinkage is not fed; they are irregular with dendritic edges and sit in the same hot spot every time. Surface appearance alone rarely settles it — position, distribution and a section or X-ray are needed — and if this call is wrong, the parameter changes that follow are usually wasted. Sink marks on the surface and shrinkage porosity inside are, as we see it, two expressions of one cause: local shrinkage that is not fed. If the skin has frozen while the inside is still contracting, and the skin is not yet strong enough to resist the pull, the surface is drawn in and a sink forms; if the skin is already strong enough, the contraction opens a void inside instead. The same hot spot can therefore show up as a sink or as shrinkage porosity depending on wall thickness and cooling — or as both. That is why a part that looks perfect can show a cluster of voids under X-ray.

Three kinds of void look different — and call for different fixes.

05Machining reveals porosity at the exact place where function is measured

Machining is usually applied to the surfaces with the highest functional requirements: gasket faces, O-ring grooves, bearing bores, threaded holes, locating datums, and mating planes. That is why the defect often appears late and has a disproportionate impact. The casting may pass incoming inspection, trimming, and coating, yet fail when the final CNC operation creates the actual pressure boundary or assembly interface. The supplier has already paid for the casting, transport, setup, tooling, and machining. The customer discovers the defect only after the most expensive value has been added.

06Removing more material can make the result worse

More machining stock does not guarantee a better surface. According to NADCA, removing the dense skin beyond approximately 0.020 in (0.50 mm) increases the chance of exposing core porosity. The normal minimum machining allowance is about 0.010 in (0.25 mm), but the final value must account for casting variation, flatness, parting-line effects, and the operation itself. The practical rule is to remove enough material to achieve the required geometry, but not so much that the process routinely cuts into the porous core. If the casting varies too much for that rule to work, the solution is better die and process control, not an automatically larger machining allowance.

07Gas porosity and shrinkage porosity create different machining risks

Small, rounded gas pores may have limited structural effect when they are dispersed and outside a pressure boundary. Shrinkage porosity is more irregular and can be interconnected. When machining opens it, the result may be a leak, a weak thread, a fatigue initiation site, or a surface that cannot be sealed by coating. This is why a supplier should not report only “porosity present” or “porosity absent.” The location, type, size, spacing, and relationship to the machined surface matter. A radiograph that is acceptable for a non-functional rib may be unacceptable for a drilled hydraulic passage.

08A visual check cannot prove a machined surface is sound

If the failure mechanism is created by stock removal, the inspection plan must reproduce that condition. Depending on the application, this can include sectioning through the planned machining zone, simulated machining, radiography, ultrasonic testing, dye penetrant inspection for surface-connected defects, or pressure testing after machining. According to NADCA, porosity limits should be agreed by critical zone and application, with reference radiographs or sectioned samples where appropriate. Pressure-tight specifications require special test methods and should be defined before production rather than added after the first leak.

09From our experience: why more intensification pressure rarely cures shrinkage

The usual response on the shop floor is to raise intensification pressure and hold it longer. In our experience that does little for shrinkage porosity. Pressure has to travel through a path that is still liquid — from the plunger through the biscuit, the runner and the gate to the hot spot. The gate is the narrowest point on that path and the first to freeze; once it has frozen, no amount of pressure at the shot end reaches the hot spot. What usually works is building intensification earlier, thickening the gate so it freezes later, or adding a squeeze pin to press the hot spot locally. For every defect type in the table below, our first remedy is a change to the part or the die rather than to a process parameter: a thick solid boss is a hot spot however the machine is set. Zinc alloys freeze over a much narrower range than aluminum alloys and leave a very short feeding window, so on our zinc parts the trouble lies more in filling — cold shuts, flow lines, entrapped air — and part design carries even more weight; aluminum freezes over a wide range, and shrinkage is the bigger problem.

From our troubleshooting: defect, main cause and remedies in order of priority
Defect	Main cause	Remedies, in order
表面缩水 / Sink marks	筋或凸台过厚、壁厚突变 / Over-thick ribs and bosses, section steps	① 改结构：减薄筋位、凸台掏空 ② 加强该区冷却 ③ 增压提前 / ① Redesign: thinner ribs, cored bosses ② More local cooling ③ Earlier intensification
内部缩松 / Shrinkage porosity	热节补缩路径被内浇口凝固切断 / Feeding path cut off when the gate freezes	① 挤压销局部补缩 ② 加厚内浇口 ③ 随形冷却缩短凝固时间 ④ 结构减料 / ① Squeeze pin ② Thicker gate ③ Conformal cooling ④ Remove mass
卷气孔 / Entrapped air	紊流充填、排气不足 / Turbulent fill, insufficient venting	① 重算内浇口面积与充填时间 ② 加大排气 / ① Recalculate gate area and fill time ② More venting
析出气孔 / Precipitated gas	熔体含气量高 / High hydrogen in the melt	① 精炼除气 ② 控制回炉料比例与清洁度 ③ 控制熔炼温度与保温时间 / ① Degas ② Control returns ③ Control melt temperature and holding time

← Swipe sideways for all columns →

10Common mistakes

Inspecting only the as-cast exterior. An intact skin can hide a defect that will be opened by the first drilling or facing operation. Checking the wrong surface creates confidence without testing the actual risk. Treating every pore as equally serious. A dispersed pore in a non-functional interior region is not equivalent to interconnected shrinkage below a sealing face. A blanket zero-porosity requirement can increase cost without improving product performance, while an undefined acceptance rule leaves the supplier guessing. Adding machining stock to compensate for unstable castings. Excess stock increases cycle time, tool wear, chip volume, and the probability of cutting through the dense skin. It may also increase clamping loads on a weak wall. First control the casting variation and identify the real datum. Locating the gate or overflow without considering the machined zone. Gate and overflow locations affect metal flow, local heat, trimming, and internal quality. If the final machined surface is not identified during die design, the process may place a high-risk flow or solidification region directly under the cutter. Blaming the machine shop for a repeating defect pattern. One isolated tool mark may be a machining issue. A repeatable cluster of pits at the same boss, face, or hole location is evidence that the casting process needs investigation. Review shot conditions, venting, vacuum, wall thickness, gating, and batch history before rewriting the CNC program. Our own order of investigation is the die first (cavity, ejection balance, gates and vents), then the alloy (composition, returns, melting), and only then the shot parameters — die and material problems persist, while parameters can only find the best point the die allows.

11Our typical approach

Experienced suppliers treat porosity control as a design and process responsibility shared by casting, machining, and quality teams. They do not wait for the machine shop to discover the internal structure. They identify where material will be removed, predict where porosity is likely to form, and validate the finished condition with the customer's actual functional test. First, we mark every machined and pressure-critical zone on the casting model before tooling is released. This includes sealing faces, bores, tapped bosses, drilled passages, and datums. We define the maximum depth and area of material removal so the die team knows where porosity must be minimized. Second, we review geometry for heavy sections, abrupt wall changes, deep bosses, and flow splits behind cores or openings. We use metal-saver cores, ribs, transitions, and revised feature locations where they reduce hot spots and shrinkage risk. The objective is not simply a visually attractive casting; it is a casting that remains functional after the planned machining operation. Third, we develop the gating, overflow, venting, and thermal-control strategy around the critical zones. The Gating Manual recommends controlling flow direction, avoiding unstable backfill, and providing a path for air and late-solidifying metal. Flow simulation is used where the cost of a hidden defect justifies it. Fourth, we set machining stock from the actual casting capability. We check whether the allowance is enough to clean the surface but small enough to protect the dense skin. We also align machining datums and clamping pads with the die structure so that casting variation does not become an avoidable position error. Finally, we validate the complete sequence. We inspect as-cast samples, perform simulated or actual machining through the critical zones, and then run the relevant leak, thread, surface, or dimensional tests. For high-risk parts, we retain a reference section or radiograph and define the acceptance limit by zone. Process changes involving gates, vacuum, die temperature, alloy source, or machining depth trigger a review rather than being treated as routine adjustments.

12From our projects: an engine thermostat housing

An engine thermostat housing (coolant outlet) we make carries cooling-system pressure and is leak-tested at 0.25 MPa for 30 s with no leak allowed. It faces exactly the risk this article describes: once an internal pore connects to the outside, coolant gets in, corrosion accelerates, and over time a leak path opens. We reduce internal porosity through optimised gating and venting, check internal quality by X-ray sampling and leak-test every finished part; every alloy batch is checked by spectrometer to control iron and other impurities, and blanks are stress-relieved before machining so the sealing face does not move after thermal cycling. In stable production, first-pass yield on the leak test is above 99%, and in-service leak complaints are below 0.1%.

13Questions buyers should ask

1. Which machined surfaces are porosity-critical, and were they identified before die design? 2. What is the expected dense-skin depth in the alloy and section being supplied? 3. How much material will CNC machining remove from each critical face, bore, or boss? 4. What evidence shows that the machined depth will not routinely expose core porosity? 5. Is the expected defect gas porosity, shrinkage porosity, or both? 6. How do the gate, overflow, vent, vacuum, and cooling layouts protect the machined zones? 7. Will the supplier provide sectioned samples, simulated-machining results, X-ray images, or pressure-test data? 8. Are porosity limits defined by location, pore size, spacing, and allowable count rather than by a vague zero-defect statement? 9. Is leak testing performed before machining, after machining, or both? What are the exact test conditions? 10. What process changes require revalidation, and how will the customer be notified? The strongest supplier answer connects a defect limit to a feature, a machining depth, a test method, and a production control. That is the level of definition required to make porosity manageable.

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. Porosity Control and Secondary Machining Preplanning, pp. 2-13 to 2-15; Machining Stock Allowance S/P-4A-13-24, p. 4A-34; Quality Assurance, 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 (NADCA). Product Design for Die Casting: In Recyclable Aluminum, Magnesium, Zinc and ZA Alloys, 7th edition, 2015. Skin Effect and Porosity, pp. 16-17; Machining and Porosity, pp. 63-65; Conventional Die Casting Process and Skin Effect, pp. 93-100.
[3]North American Die Casting Association (NADCA). Gating Manual, Publication #512, 2006. Casting Quality Requirements and Porosity Considerations, pp. 12-13; Flow Pattern and Ingate/Outgate Location, pp. 22-25.
[4]North American Die Casting Association. Basic Operator Training Program, 2015 revision. Shrinkage and internal void formation, pp. 47-49.

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