PM Troubleshooting

Common Mistakes and Fixes in Powder Metallurgy and Sintered Parts

When a sintered part comes out undersized, cracks in the furnace, or blows the cost quote, the root cause is usually one of a dozen repeat mistakes. This guide gives each a symptom, a cause, and a numeric fix.

Symptom: green parts weigh light and heights vary shot to shot. Root cause is usually a fill weight set from theoretical density instead of measured apparent density. Iron powder apparent density runs 2.8 to 3.2 g/cm3, not the 7.87 g/cm3 of solid iron, so a fill computed on solid density overshoots mass by more than 2x and underfills the die on the real press stroke. Fix: pull apparent density from a Hall flow test and feed it into the Powder Fill Weight calculator. A 0.1 g/cm3 apparent density drift moves fill height about 3% to 4%, enough to push a part out of a 0.15 mm height tolerance.

Symptom: the press bottoms out or green density falls short of the 6.9 to 7.1 g/cm3 target. The usual mistake is estimating tonnage from part footprint alone and ignoring compaction pressure per unit area. Iron at 7.0 g/cm3 green needs roughly 400 to 500 MPa, and force equals pressure times projected area, so a 20 cm2 face at 450 MPa demands about 90 metric tons, not the 60 someone guessed. Fix: run the Compaction Tonnage calculator with the true projected area including all steps and bosses, then check it against the Green Density target before committing to a press class.

Symptom: sintered parts land 0.3 to 0.6 percent undersized versus print even though the die was cut correctly. Root cause is a shrinkage allowance applied in the wrong direction or built on a generic factor. FN-0205 typically shrinks 0.1 to 0.3 percent, while some copper-steel grows, and a bronze bearing mix can expand 0.5 percent. Fix: never reuse another alloy's factor. Measure sintered-to-green dimensional change on a trial lot, load it into the Shrinkage Allowance calculator, and confirm the die is oversized by exactly that ratio. A 0.2 percent sign error on a 40 mm bore is 0.16 mm, well outside a typical H7 fit.

Symptom: monthly powder purchases exceed part weight math by 8 to 15 percent and nobody can find the metal. The missed variable is cumulative powder loss across mixing, transfer, fill spillage, and green scrap. Real losses run 2 to 5 percent even in a tidy shop, and a single leaking hopper can double that. Fix: weigh incoming powder against good green weight for one week and put the ratio through the Powder Loss Rate calculator. At $2.50 per kg on 30 tonnes a month, cutting loss from 6 percent to 3 percent recovers about $2,250 monthly, which usually pays for better transfer containment in one quarter.

Symptom: furnace throughput is below the quoted rate and back orders build. The common error is loading planned belt speed instead of the speed that holds time at temperature. A mesh belt sinter run for iron needs 20 to 30 minutes above 1120 C, so a 3 meter hot zone caps belt speed near 100 to 150 mm per minute regardless of what the schedule claims. Fix: back-calculate belt speed from required sinter time and hot zone length, then size trays and run the Sinter Furnace Capacity calculator. Overrunning speed by even 20 percent drops density and hardness and shows up as failed transverse rupture tests, not as a clean throughput gain.

Symptom: the furnace energy line on the cost sheet is a flat guess and actual gas or power bills come in 25 to 40 percent higher. Root cause is ignoring standby and idle load. A continuous belt furnace pulls energy during weekend holds and empty running, so cost per part swings hard with utilization. Fix: split connected load from idle load and run the Furnace Energy Cost calculator against real loaded hours. At $0.12 per kWh a 150 kW furnace costs about $18 per hour loaded, but spread over a half-empty belt the per-part energy can easily double from $0.04 to $0.08.

Symptom: the quote looked profitable but the job loses money once scrap is counted. The bad assumption is quoting on gross parts pressed rather than good parts shipped. If green cracks, sinter distortion, and sizing rejects total 7 percent, a 100,000 part order ships 93,000, and fixed tooling and setup spread over fewer good parts. Fix: quote on net yield. Run the Part Yield calculator with real reject data, then flow that yield into Cost Per Sintered Part so amortized tooling and material carry the true multiplier. A yield drop from 95 to 90 percent raises per-part cost about 5.5 percent, which erases a thin margin.

Symptom: tooling cost per part looks fine on a 500,000 run but the same die kills margin on a 40,000 run. The mistake is amortizing a $45,000 compaction tool set over hoped-for lifetime volume instead of the actual purchase order. Fix: amortize over committed quantity only. Run the Tooling Amortization calculator against the real PO. That $45,000 spread over 500,000 parts is $0.09 each, but over 40,000 it is $1.13 each, a difference that must sit in the quote or in a documented tooling charge, not buried in the hope of a reorder that may never land.

Published 2026-07-01.