Troubleshooting

Costly Foundry, Casting and Forging Mistakes and How to Catch Them

The specific errors that throw off casting yield, pour weight, melt loss, and forging tonnage, each with a symptom, a root cause, and a fix tied to a number.

Symptom: your Casting Yield reads 78 percent on paper but the floor reports 62 percent. Root cause is almost always counting only the runner and riser in the numerator while ignoring gates, overflows, and the sprue well. A gray iron plate might carry 12 percent in gates and 4 percent in overflows that never make the tally. Fix: weigh a full cluster including everything cut off at shakeout, then divide casting weight by poured weight. If cast part is 6.4 kg and the full tree pours at 10.3 kg, yield is 62 percent, not 78. Reconcile monthly against actual metal purchased per good casting shipped.

Symptom: melt loss looks like 1.5 percent but the Furnace Charge Calculator keeps under-delivering liquid metal. Root cause is treating melt loss as a single flat number across alloys and furnace types. Coreless induction on clean charge runs 1 to 3 percent, but rusty steel scrap or high-turbulence pouring pushes oxidation losses to 4 to 6 percent, and aluminum with poor cover flux can hit 5 to 8 percent. Fix: measure loss per charge by weighing input minus tapped metal minus slag, then feed that alloy-specific figure into the Melt Loss Calculator. Using 2 percent when reality is 5 leaves you 30 kg short on a 1000 kg heat.

Symptom: parts short-fill even though the Pour Weight Calculator said the ladle held enough. Root cause is mixing shrinkage and pour weight in the wrong direction, or forgetting that liquid volume must exceed solid casting volume by the solidification shrinkage. Steel shrinks about 3 percent by volume during freezing, aluminum near 6 percent, so a 5.0 kg finished aluminum part needs the feed metal plus roughly 6 percent extra fed from risers, not just the nominal part mass. Fix: size risers to hold at least 1.2 times the casting shrinkage volume, and confirm pour weight covers casting plus gating plus that feed demand before you tap.

Symptom: your Gating Ratio Calculator says 1:2:2 but you still see inclusions and misruns. Root cause is running a pressurized ratio where a non-pressurized one belongs. Iron often uses 1:2:2 or 1:4:4, aluminum needs non-pressurized systems like 1:2:4 or 1:4:8 to keep the choke at the sprue base and slow the metal below the 0.5 m per second critical velocity. Fix: put the smallest area at the sprue exit, not the ingates, and recompute. If the sprue choke is 200 square mm, ingate total area should be 400 to 1600 square mm depending on alloy, never smaller than the choke.

Symptom: die life runs half of what the Die Life Estimator projected. Root cause is entering nominal preheat and cycle temperature instead of the real thermal load. H13 die steel fatigue-cracks fast when surface temperature swings past 550 to 600 C each shot, and skipping die preheat to 200 to 300 C at startup causes early heat checking. Fix: log actual die surface temperature and shot count, then recalibrate. A die rated for 100,000 aluminum shots at controlled temperature may deliver 45,000 if spray cooling is uneven, so tie the estimate to measured cooling-line flow, not the catalog number.

Symptom: the press stalls even though Forging Tonnage said it would clear. Root cause is using room-temperature flow stress or ignoring the friction and shape factors. Forge load equals projected area times flow stress at forging temperature times a constraint factor of 3 to 8 for closed dies. A 200 square cm steel part at a flow stress of 100 MPa needs 200 tonnes at a factor of 1, but at a realistic constraint factor of 5 it needs closer to 1000 tonnes. Fix: use hot flow stress at your billet temperature and apply the shape factor before sizing the press, then add 20 percent margin for die chilling.

Symptom: sand costs creep and molds vary in strength. Root cause is a wrong sand-to-metal ratio in the Sand Usage Calculator, or not tracking new sand addition against return sand. Green sand systems typically run a 4:1 to 10:1 sand-to-metal ratio by weight, and a chemically bonded system needs 1 to 2 percent binder on sand weight. Fix: weigh flask sand per mold against casting weight and hold the ratio steady. If a 20 kg casting suddenly pulls 220 kg of sand instead of 140 kg, your flask sizing or ratio drifted, and burn-on and inclusion defects follow.

Symptom: shakeout backs up and cools castings unevenly. Root cause is loading the line above the Shakeout Capacity in throughput, measured in molds or tonnes per hour, not in count alone. A grid rated for 30 tonnes per hour chokes at 40 even if the mold count looks fine, because heavier flasks dwell longer. Fix: match pour rate to shakeout tonnage capacity with a 15 percent buffer, and stage hot molds for the correct cooling time. Pulling steel castings before they drop below 500 C invites distortion and cracked thin sections at the gate junctions.

Symptom: quoted scrap looks fine but the Foundry Scrap Cost line bleeds margin. Root cause is booking scrap only as returned metal value and ignoring the lost value added, the labor, molding, energy, and machining already spent on a rejected casting. A part that scraps after machining costs far more than its metal weight times return price. Fix: cost scrap at full accumulated value, not remelt value. If internal scrap runs 8 percent and each rejected part carried 30 minutes of machining, the true loss can be 3 to 5 times the metal-only figure you were tracking.

Published 2026-07-01.