{"id":580,"date":"2026-06-24T01:53:45","date_gmt":"2026-06-24T01:53:45","guid":{"rendered":"https:\/\/kurtfoundry.com\/?p=580"},"modified":"2026-06-13T02:49:33","modified_gmt":"2026-06-13T02:49:33","slug":"ndt-methods-sand-castings-match-test-defect","status":"publish","type":"post","link":"https:\/\/kurtfoundry.com\/fr\/blog\/ndt-methods-sand-castings-match-test-defect\/","title":{"rendered":"M\u00e9thodes de CND pour les moulages en sable : Comment faire correspondre le test au d\u00e9faut"},"content":{"rendered":"\n

You specified RT on your last casting order \u2014 but did it actually catch the defect that mattered? Most NDT specifications I review treat testing like a checkbox: pick a method, call out a standard, move on. The problem is that each method has blind spots, and the wrong choice burns your inspection budget while the real defect slips through.<\/p>\n\n\n\n

About 7% of casting volume is lost during solidification. Internal voids are not a question of if but where and how big. The practical challenge is matching the right test to the right defect in the right location \u2014 and that is where most specifications fall short.<\/p>\n\n\n\n

Why No Single NDT Method Covers Every Casting Defect<\/h2>\n\n\n\n

Sand castings produce defects at every stage: porosity from trapped gas, shrinkage from inadequate feeding, inclusions from sand breakdown, and hot tears from restrained contraction. Each defect type has a different size, shape, orientation, and depth \u2014 and each NDT method responds differently to those variables.<\/p>\n\n\n\n

RT excels at detecting volumetric defects like porosity and shrinkage cavities but struggles with tight planar cracks oriented parallel to the beam. UT measures depth precisely but goes nearly blind on coarse-grained materials. MT catches surface and near-surface cracks on ferrous castings only. PT finds any surface-breaking discontinuity regardless of material but cannot see below the surface at all.<\/p>\n\n\n\n

No single method provides complete assessment. The real specification question is not which method to use, but which combination matches your casting defect profile<\/a> and acceptance requirements.<\/p>\n\n\n\n

Radiographic Testing for Internal Voids and Inclusions<\/h2>\n\n\n\n

RT remains the go-to volumetric inspection for sand castings, governed by ASTM E1030 for metallic castings with reference radiographs per ASTM E446. Its strength is the permanent film record \u2014 you get a visual map of internal discontinuities that buyers, engineers, and third-party inspectors can all review independently.<\/p>\n\n\n\n

For porosity, gas holes, and shrinkage cavities, RT is hard to beat. I have reviewed castings where 120 X-ray views across three pieces turned up only a single rejectable indication \u2014 that kind of result gives everyone confidence in the pour.<\/p>\n\n\n\n

RT also identifies inclusion type and morphology. You can distinguish between sand inclusions, oxide films, and metallic foreign material based on density contrast on the film. That morphology data drives root-cause analysis back at the melt deck.<\/p>\n\n\n\n

\"Radiographic<\/figure>\n\n\n\n

Where RT Falls Short<\/h3>\n\n\n\n

RT has real limitations that most specification writers overlook. Thick-section castings (above roughly 6 inches of steel equivalent) require high-energy sources and long exposure times, driving costs up fast. Complex geometries with overlapping wall sections create confusing images where defects hide behind sound metal.<\/p>\n\n\n\n

The bigger issue: RT cannot tell you how deep a defect sits. You see a shadow on film, but you do not know if that porosity cluster is 5 mm below the surface or 50 mm. When machining will remove material, that depth information is critical \u2014 and RT alone cannot provide it.<\/p>\n\n\n\n

For heavy-wall castings, the practical approach is to radiograph first, then ultrasonically examine only the flagged areas for depth measurement and precise excavation planning.<\/p>\n\n\n\n

\"Diagram<\/figure>\n\n\n\n

Ultrasonic Testing for Depth and Thickness<\/h2>\n\n\n\n

UT fills the gap RT leaves: it measures defect depth, size, and distance from the nearest surface. ASTM A609 covers UT examination of carbon, low-alloy, and martensitic stainless steel castings, providing standardized acceptance levels.<\/p>\n\n\n\n

Where UT shines is on thick sections and simple geometries \u2014 pump casings, valve bodies, flanges. A skilled operator can map the entire cross-section and locate discontinuities to within a few millimeters of their actual depth. That precision is what lets you decide whether a defect will be machined away or remain embedded in the finished part.<\/p>\n\n\n\n

The Cast Iron Problem<\/h3>\n\n\n\n

Standard UT frequencies fail on cast iron. At 5 MHz with a half-inch transducer, you lose the back-wall signal after only a few inches of gray iron. The coarse grain structure scatters the sound beam until the signal is useless.<\/p>\n\n\n\n

Before you pour a trial run of cast iron and specify UT, know the frequency requirements: 2.25 MHz works for 1-2 inch wall sections, and you need to drop to 1 MHz with a one-inch diameter transducer for sections around 4 inches thick. Lower frequencies are the only way through coarse-grained iron \u2014 and if your specification does not call out the frequency, the testing lab will default to standard settings and report the part as uninspectable.<\/p>\n\n\n\n

\"Diagram<\/figure>\n\n\n\n

If the geometry or material makes UT impractical, switch to RT. Do not force UT where the physics will not cooperate.<\/p>\n\n\n\n

Material Grade Matters<\/h3>\n\n\n\n

Not all stainless steels respond to UT equally. Fine-grained martensitic grades like CA6NM produce clean, reliable ultrasonic signals. Coarse-grained austenitic grades \u2014 CF8M, CF3, and most nickel-based alloys \u2014 scatter the beam badly enough to make UT unreliable.<\/p>\n\n\n\n

Before specifying UT on a stainless steel casting, confirm the metallurgical grade. Martensitic stainless: UT works well. Austenitic stainless: specify RT instead. Writing “UT per ASTM A609” on an austenitic casting creates a specification that cannot be executed meaningfully.<\/p>\n\n\n\n

Surface Methods: MT and PT<\/h2>\n\n\n\n

Magnetic Particle Testing (MT) per ASTM E709 and Liquid Penetrant Testing (PT) per ASTM E165 detect what volumetric methods miss: surface-breaking and near-surface discontinuities.<\/p>\n\n\n\n

MT detects surface cracks and shallow subsurface indications to a depth of 0.003-0.007 inches on ferrous materials only. It is fast, relatively inexpensive, and catches hot tears, grinding cracks, and surface-connected shrinkage that RT and UT may not flag. The limitation is material: MT requires ferromagnetic response, so aluminum, bronze, and austenitic stainless castings are excluded.<\/p>\n\n\n\n

PT works on any material with a nonporous surface. Apply penetrant, allow dwell time, remove excess, apply developer, and inspect. PT catches any surface-breaking discontinuity regardless of material or geometry \u2014 but it sees nothing below the surface. Zero subsurface detection.<\/p>\n\n\n\n

The most common mistake I see in new specifications is treating MT and PT as interchangeable. They overlap only on surface cracks in ferrous castings. For subsurface indications on ferrous parts, MT has a slight edge. For non-ferrous castings, PT is your only surface option.<\/p>\n\n\n\n

\"Magnetic<\/figure>\n\n\n\n

I recommend specifying MT or PT as the final inspection step after machining, not just on the as-cast surface. Acceptance criteria that pass the casting before machining<\/a> may not hold once you cut into the part.<\/p>\n\n\n\n

Matching the Method to the Defect<\/h2>\n\n\n\n

Here is the selection framework I use when reviewing casting specifications:<\/p>\n\n\n\n

Defect Type<\/th>Best Primary Method<\/th>Backup \/ Confirmation<\/th>Key Limitation<\/th><\/tr><\/thead>
Gas porosity (internal)<\/td>RT<\/td>UT for depth<\/td>RT misses tight planar voids<\/td><\/tr>
Shrinkage cavity<\/td>RT<\/td>UT for depth mapping<\/td>Thick sections need high-energy RT<\/td><\/tr>
Hot tears \/ cracks<\/td>MT (ferrous) or PT<\/td>RT if subsurface extent suspected<\/td>MT\/PT surface only<\/td><\/tr>
Sand inclusions<\/td>RT<\/td>VT on machined surfaces<\/td>Low-density inclusions may lack contrast<\/td><\/tr>
Subsurface porosity<\/td>UT<\/td>RT for morphology<\/td>UT unreliable on coarse-grained iron<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
\"Machined<\/figure>\n\n\n\n

This table gets you started, but casting geometry and material override any generic recommendation. A thin-wall ductile iron valve body responds differently to UT than a thick-wall gray iron pump casing \u2014 same method, completely different reliability.<\/p>\n\n\n\n

The specification gap that causes the most rework is ignoring what happens after machining. I have pulled castings that passed Level 2 UT and Level 3 RT, only to see defects open to the surface the moment the machine tool cut into the part. The standard acceptance levels<\/a> assume the as-cast surface stays intact. When you machine 10 mm off a face, every subsurface defect within machining depth plus 5 mm becomes a potential surface defect.<\/p>\n\n\n\n

Specify tighter acceptance criteria in machining zones, or add a PT\/MT step after rough machining. That single line in your specification prevents the most frustrating rejection scenario in the business.<\/p>\n\n\n\n

How to Write NDT into Your Next RFQ<\/h2>\n\n\n\n

Three specification items prevent most inspection disputes:<\/p>\n\n\n\n

    \n
  • Name the method AND the governing standard. “UT per ASTM A609, Level 2” is enforceable. “Castings shall be free of injurious defects” is not \u2014 no inspector can work with that language.<\/li>\n<\/ul>\n\n\n\n
      \n
    • Call out the acceptance level per zone. Machining zones need tighter criteria than non-machined surfaces. Specify the acceptance level for each zone separately.<\/li>\n<\/ul>\n\n\n\n
        \n
      • Require surface inspection after machining on any critical casting. One line \u2014 “MT per ASTM E709 on all machined surfaces, post-machining” \u2014 closes the gap between volumetric NDT acceptance and actual fitness for service.<\/li>\n<\/ul>\n\n\n\n

        The cost difference between specifying NDT correctly upfront and sorting out rejected castings after machining is not close. Get the specification right before you issue the PO.<\/p>\n\n\n\n