Nearly every defect on a ferrous sand casting traces to one of five causes: gas the mold could not vent, contraction the risers could not feed, sand that washed or expanded into the cavity, metal that froze before the mold filled, or mold parts that moved after the pattern was drawn. The table identifies each defect by what it looks like. The sections after it say which ones are the foundry’s process to fix, which start on the drawing, and which are cosmetic on a face that gets machined anyway.
| Defect | What It Looks Like | Common Causes | Prevention |
|---|---|---|---|
| Gas porosity / blowholes | Smooth, rounded cavities inside or on the casting | Excessive moisture, poor venting, dissolved gas, low sand permeability | Control moisture, improve venting, degas metal, improve sand permeability |
| Shrinkage cavity | Irregular internal or surface voids | Inadequate feeding during solidification | Improve riser design, use chills, promote directional solidification |
| Sand inclusion | Sand particles embedded in the casting | Mold erosion, weak sand, turbulent metal flow | Improve mold strength and gating design, reduce turbulence |
| Cold shut | Thin seam or line where two metal streams meet | Low pouring temperature, slow filling, poor gating | Increase pouring temperature and filling speed, redesign gates |
| Misrun | Casting is incomplete or some areas are missing | Metal solidifies before filling the mold | Increase fluidity and pouring temperature, enlarge gates, improve venting |
| Hot tear / hot crack | Irregular cracks formed during solidification | Restricted contraction, poor casting geometry | Improve mold collapsibility, add fillets, optimize section transitions |
| Metal penetration | Rough surface with metal penetrating between sand grains | Coarse sand, high pouring temperature, weak mold surface | Use finer sand or coatings and optimize pouring temperature |
| Scab | Raised irregular metal area on the casting surface | Sand surface lifts or cracks during pouring | Improve sand strength and moisture control |
| Cuts and washes | Grooves or eroded areas on the casting | High-velocity molten metal erodes the mold | Reduce metal velocity and improve runner/gate design |
| Rat tails / buckles | Long, shallow surface lines | Expansion of mold sand caused by heat | Use proper additives and sand properties |
| Swell | Bulged casting surface | Mold wall deforms under metallostatic pressure | Increase mold hardness and strength |
| Mismatch / mold shift | Upper and lower halves of casting do not align | Cope and drag become misaligned | Improve flask alignment and locating pins |
| Core shift | Internal cavity is incorrectly positioned | Core movement during pouring | Improve core prints and core supports |
| Run-out | Molten metal leaks from the mold | Poor mold closure or damaged mold | Secure mold halves and inspect molds before pouring |
| Flash / fins | Thin metal projections along parting lines | Gaps between mold sections or excessive pressure | Improve mold closing and clamping |
| Drop | Irregular lump caused by sand falling from the cope | Weak mold sand or insufficient ramming | Increase sand strength and improve mold preparation |
1. Gas Porosity and Blowholes
Gas porosity is gas that the mold or the melt could not shed before the skin froze: smooth, rounded cavities, from pinholes just under the surface to blowholes big enough to take a fingertip. The rounded wall is the tell. A shrinkage cavity is ragged.
In green sand the gas is mostly steam. The sand carries roughly 2–5% water, and the metal front turns that to steam in the first instant of contact; if permeability is low or the cope is not vented, the steam pushes back into the metal instead of out through the sand.
Every lever here is the foundry’s — moisture and permeability, vents in cope and cores, deoxidation and degassing practice at the furnace. The one thing the drawing can contribute is a blind core somewhere to vent to. Subsurface pinholes usually declare themselves on the first machined face, so a pressure-tight body should be radiographed before it goes to the machine shop, not after.
2. Shrinkage Defects
Shrinkage is contraction that nothing fed: irregular, rough-walled voids at the last place to freeze — the center of a heavy section, the root of a boss, the junction where two walls meet. It hides inside and shows up when the machinist breaks into it.
For steel the fix is feeding, and it is not optional. A carbon steel such as A216 WCB gives up about 3.6% of its volume between liquidus and solidus, and that metal has to come from a riser that stays liquid longer than the section it feeds, through a neck that does not freeze first. Chills at the heavy end and a taper toward the riser make the freezing front travel one way, so the last liquid is in the riser and not in the casting.
Gray iron does not follow that rule. A Bureau of Standards (now NIST) measurement series found the volume change on freezing ran from a 1.65% expansion in gray iron to a 5.85% contraction in a low-carbon, low-silicon iron: graphite precipitating during solidification pushes outward. A high-carbon-equivalent gray iron counterweight can pour sound with no riser at all.
When gray iron does shrink, the cause is usually that expansion shoving a soft mold wall outward, so there is not enough metal left to fill the enlarged cavity. Mold rigidity does for iron what the riser does for steel. Ductile iron expands the same way and leans on the mold at least as hard, which is why a ductile housing that shrinks in green sand will often pour sound in a rigid no-bake mold off the same pattern.
Shrinkage is also the defect the drawing causes most. An isolated heavy section, a boss a riser cannot reach, a wall that steps from thin to thick with no transition — each creates a hot spot that freezes last and starves. When the foundry asks to add a riser pad, taper a section, or put a chill against a boss, that is the request to say yes to.
3. Sand Inclusions
Sand inclusions are loose sand the metal picked up and carried into the casting: grains or clumps embedded in the surface, or sitting just under a machined face as a dark, gritty pocket. Sand floats in iron and steel, so inclusions collect against cope surfaces and under cores, and their position tells you where the sand came from.
Wherever it came from, the fix is the foundry’s — gating that slows the metal at the cavity, sand strength and cure, and mold handling.
The drawing contributes only through geometry: a sharp inside corner or a thin fin on the pattern makes a friable sand edge, and a fillet there costs nothing.
4. Cold Shut
A cold shut is a seam where two metal fronts met and did not fuse: a thin, sharp-edged line on the surface, often with slightly rounded lips, that runs into the section as a crack-like discontinuity. Each front carries an oxide skin, and if neither is hot enough to remelt the other’s, they lie against each other and freeze. It happens wherever streams converge cold — the far side of a core, the top of a tall thin wall fed from two gates, the end of a long flow path.
The reflex is to pour hotter, and that does help, but for a narrower reason than most people assume. The Steel Founders’ Society of America puts it plainly in its guidance for casting designers: for a given alloy, fluid life does not increase with superheat. A hotter pour does not make the metal flow better; it only delays the moment the skin freezes.
For steel there is also little room to go hotter. The pouring windows below are what the four alloys we pour actually run at, and carbon steel already sits against the refractory limit of the sand itself, 3000–3330 °F (1650–1820 °C).
| Alloy | Pouring temperature |
|---|---|
| Gray iron | 2450–2700 °F (1340–1480 °C) |
| Ductile iron | 2450–2700 °F (1340–1480 °C) |
| Stainless and high-alloy steel | 2700–2900 °F (1480–1600 °C) |
| Carbon and low-alloy steel | 2850–3100 °F (1565–1700 °C) |
The durable fixes are in the gating and the section: more ingates so the fronts meet sooner and hotter, shorter flow paths, a faster fill, and on the drawing a wall thick enough for the alloy to run the distance. Pouring at the top of the window is a stopgap, and it is paid for elsewhere — hotter metal means more shrinkage porosity and a higher risk of hot tears.
5. Misrun
A misrun is a mold that never filled: the casting is complete up to a smooth, rounded edge and then stops — a missing corner, a wall that ends short, a lug that is not there. The metal in the thin section froze before the rest arrived.
Thin walls far from the gate are where it happens, and it gets worse with a slow pour — a choked sprue, a ladle too small for the mold — and with a cavity that cannot vent the air ahead of the metal. Steel misruns more readily than iron: it pours near the top of the windows in the table above and has a short fluid life, while gray iron runs thin sections with ease.
The lasting fixes are a faster fill, an ingate placed into the thin section rather than upstream of it, and a vent at the far end of the cavity. If a wall still misruns with the metal at the top of its pouring window and the gating already reworked, the wall is thinner than the alloy will run over that distance, and the drawing is where the fix is.
6. Hot Tears
A hot tear is a crack that opened while the casting was still partly liquid: ragged, dark with oxide, following the grain, and almost always at a junction where a heavy section held a thin one back as it contracted. A crack that formed cold is straight and bright by comparison.
The mechanism is restraint. The casting wants to contract as it cools, the mold and cores resist — a strong core inside a ring, sand packed between two flanges — and the stress concentrates at the junction that is still mushy, which is the weakest point in the casting.
Steel is the most tear-prone alloy of the four we pour because it contracts on freezing; gray iron, which barely contracts at all, almost never tears. Pouring hotter raises the thermal stress and makes tearing more likely, which is the second cost of pouring at the top of the window.
Hot tears are a design defect more often than a process one. Sharp inside corners, abrupt steps from thin to thick, and long thin webs between heavy flanges each create a restrained junction. Fillets and gradual transitions remove the stress raiser; where the geometry cannot change, the foundry adds temporary tie bars and machines them off, and softens the core mix so it collapses instead of resisting.
7. Metal Penetration
Metal penetration is metal that soaked into the sand instead of stopping at it: a rough, sand-crusted surface where the metal ran between the grains and locked them in, so blasting will not clean it. The casting is to size underneath; the crust is the problem.
There is a trade here: fine sand closes the pores between the grains but lowers permeability and invites the gas defects in section 1, so the usual answer is not finer sand across the whole mold but a refractory wash on the faces that need it and on the cores.
On a face that gets machined, penetration costs nothing. On an as-cast sealing face or a visible face, it is a coating or a sand change at the foundry, and it is worth marking those faces on the drawing so the foundry knows which ones to coat.
8. Scabs
A scab is a thin, rough plate of metal standing proud of the surface with a layer of sand trapped under it: the mold face lifted or cracked during pouring and metal ran in behind it. It sits on cope faces and on broad, flat surfaces that the metal took a long time to cover.
The lifting is silica expansion. Radiant heat from the rising metal reaches the cope face before the metal does; as the sand skin passes 573 °C the quartz inverts and expands abruptly while the sand behind it is still cool, so the skin buckles away from the mold and metal fills the gap. Excess moisture and a slow fill both lengthen the exposure and make it worse.
Scabs are the foundry’s to fix — additives that cushion the expansion, moisture and compaction control, and a faster fill so the cope face is covered before it heats through. The drawing can help by avoiding a large, flat, unbroken cope face where the part orientation allows it.
9. Cuts and Washes
Cuts and washes are grooves and eroded channels where moving metal cut sand out of the mold: a rough, streaked surface downstream of an ingate, with the casting oversize where the sand used to be and the sand itself turning up as inclusions somewhere else. Metal penetration leaves the casting to size with a crust on it; a wash leaves it oversize and rough.
The metal was moving too fast where it met the mold, and the sand was too weak to take it — an ingate aimed at a mold wall or a core, a runner that lets the stream accelerate, sand under-compacted or under-cured. The fix is in the gating: slow the metal at the entry, do not gate onto a core or a thin mold wall, and coat cores that sit in the stream. What the drawing contributes is information — tell the foundry which faces are critical so the ingates go somewhere else.
10. Rat Tails and Buckles
Rat tails and buckles are the same sand-expansion failure as a scab, caught before the metal got behind the lifted skin: shallow, irregular lines (rat tails) or a broad, low ridge (buckle) on an otherwise flat surface. They are cosmetic, and they vanish on any face that gets machined.
Rat tails prevail on the drag face, the floor of the mold that hot metal covers first and heats through fastest; scabs sit on the cope face, which radiation heats before the metal arrives. Where the defect is tells you which one you are looking at.
Prevention is in the sand, so it is the foundry’s: additives that cushion the expansion, a compaction that leaves the grains room to move, and, on a critical as-cast face that keeps failing, a non-silica facing sand that has no inversion to go through.
11. Swell
Swell is a casting that came out bigger than the pattern: a bulged wall, usually a large vertical face in the drag, where the metallostatic head pushed a soft mold wall outward before the skin was strong enough to hold. In gray and ductile iron the graphite expansion during freezing adds its own push.
The cost is not the bulge. Swell eats the machining allowance on that face, adds weight, and — because the metal that filled the bulge came from somewhere — often leaves a shrinkage cavity elsewhere in the casting. An overweight casting with unexplained shrinkage is a mold that moved.
Prevention is mold rigidity: harder ramming, compactability held low enough that the wall stays put, or a move to no-bake for heavy sections that green sand cannot hold. Nothing on the drawing causes it, and nothing on the drawing prevents it.
12. Mold Mismatch
Mismatch is a step at the parting line: the cope half of the casting is shifted relative to the drag half, so a boss is half on and half off, a wall is thick on one side of the line and thin on the other, and a cored hole is not concentric with the outside.
The two mold halves did not register — worn pins or bushings, a damaged flask, pattern halves not aligned on the plate, or a mold that shifted while it was closed and handled. All of that is the foundry’s maintenance.
The drawing decides how much a given mismatch costs: a parting line on a flat plane, away from machined datums and away from features that must line up across it, can tolerate a step that would scrap a part parted through a boss. Where the parting line falls is worth settling with the foundry before the pattern is cut, not after the first casting shows a step.
13. Core Shift
Core shift is a core that moved before the metal froze around it: the internal cavity is in the wrong place, so one wall comes out thick and the wall opposite thin, or a cored passage is off-center to the outside. It is often found only when a wall-thickness check or a machining cut shows the discrepancy.
Buoyancy does most of the damage. A sand core is far lighter than the iron or steel around it and lifts hard during the pour; a core print too small or too loose, or a long core supported at one end only, lets it float up or bend. Cores set in the wrong position or crushed on closing account for the rest.
The foundry’s fixes are longer or tighter core prints, chaplets that hold the core down against the cope, and a stiffer core. The drawing’s contribution is the biggest one: a through-cavity with a print at each end holds itself, where a blind cavity hanging off a single print depends on chaplets to stay put.
14. Run-Out
A run-out is metal that left the mold: the cavity drained through the parting line or through a crack in the mold, and what solidified is a short or thin casting with a sheet of metal running off it. You will not receive a run-out — it is scrapped at the foundry — but a supplier who reports them repeatedly has a mold-handling problem, and it reaches you as missed deliveries.
15. Flash and Fins
Flash is a thin sheet of metal along the parting line or around a core print: metal that got into a gap between mold parts and froze there. On the outside it is a finishing cost, ground off in cleaning. Inside a cored passage on a hydraulic manifold or a pump body, a fin is a loose particle waiting to break free in service, and it is worth specifying that cored passages are inspected for them.
The gap comes from parting faces that did not seat — sand on the joint, a worn pattern plate, a cope lifting slightly under pressure — or from clearance between a core and its print. Heavy flash on a batch is a warning: the same lift that let flash form also lets mismatch and, at the extreme, run-out happen, so check that batch for a step at the parting line. Prevention is mold closing and clamping, which is the foundry’s.
Conclusion
The surface defects — penetration, scabs, rat tails, buckles, cuts, swell, flash — are the foundry’s sand and gating to control; they cost finishing time, and they are gone from any face that gets machined. The ones that scrap castings are internal or structural: gas porosity, shrinkage, cold shuts, and hot tears. Two of those four, shrinkage and hot tears, start on the drawing as often as on the shop floor, and a fillet or a section taper agreed before the pattern is cut is cheaper than any riser.