Powder Metallurgy Parts: Design Rules and When Sintered Components Make Sense

I still remember the first time a purchasing manager asked me to convert a machined steel bracket to powder metallurgy “to save cost,” as if it were a button I could press in CAD. Twenty years into a career spent mostly as a contract mechanical designer embedded inside large Japanese manufacturers, I’ve learned that PM parts are one of the most misunderstood processes on the shop floor. Everyone knows the word “sintering,” far fewer people know what it actually does to tolerances, strength, and design freedom. This article walks through the practical rules I use when deciding whether a part should go to a powder metallurgy supplier, and the mistakes I’ve made along the way so you don’t have to repeat them.

What Sintering Actually Does to a Design

Powder metallurgy starts with metal powder — usually iron-based, sometimes bronze, stainless, or more exotic alloys — compacted in a die under enormous pressure, then heated below the melting point so the particles fuse together. That “below melting point” detail is the single most important thing to internalize as a designer, because it means the part never becomes fully dense the way a forging or casting does. There is always some residual porosity, typically 5-15% depending on the compaction pressure and whether the part gets a secondary densification step.

On one project, I was asked to redesign a small planetary gear carrier that had been machined from bar stock. My first instinct, trained on solid-metal thinking, was to treat the PM version as a drop-in replacement with the same fillets and wall thicknesses. It wasn’t. The porosity changed the fatigue behavior enough that I had to add material in the highest-stress transition radius, something a CNC machinist would never have flagged because in solid steel that radius was already comfortably oversized. Sintering isn’t just “a cheaper way to make the same shape” — it’s a different material system wearing the same geometry, and it needs to be designed as such.

Density Is the Variable That Controls Everything

If there is one number I ask a PM supplier for before I even start detailing a part, it’s target density, usually expressed as a percentage of theoretical (fully dense) density or in g/cm³. Standard single-press, single-sinter parts land around 6.6-6.8 g/cm³ for iron-based powders, versus about 7.85 g/cm³ for wrought steel. That gap directly scales strength, and it scales it nonlinearly — going from 90% to 95% density might buy you a much bigger jump in fatigue strength than the percentage suggests, because porosity acts as a network of stress risers rather than uniformly distributed voids.

I’ve seen engineers spec a PM part using yield strength numbers pulled from a generic material data sheet without checking what density that number assumed. When the actual part comes in at a lower density because of its geometry (thin sections compact less uniformly), the real strength can be 20-30% below what the datasheet implied. Now I always ask the supplier to confirm expected density by section, not just as a single figure for the whole part, especially on parts with a mix of thick bosses and thin webs.

Wall Thickness, Draft, and the Rules That Keep Tooling Sane

PM tooling is unforgiving compared to injection molding or die casting in one specific way: the powder has to flow into the die cavity and compact uniformly along the pressing axis, which is almost always a single straight-line direction (top-and-bottom punches). That means every feature with any undercut relative to that pressing direction is off the table unless you’re willing to add secondary machining. I keep a simple mental checklist: minimum wall thickness around 1.5 mm for most iron-based powders, avoid sharp corners that create stress concentrations during ejection, and never spec a length-to-diameter ratio on a thin wall much beyond 6:1 without discussing it with the toolmaker first.

Draft angles are a different story than in casting — PM parts generally don’t need draft on the pressing faces themselves because the compaction and ejection happens along the same axis, but any feature perpendicular to that axis (a hole, a slot) still needs to be considered relative to how the tooling core pins will retract. I once specified a cross-hole too close to a flange edge, not realizing the core pin needed clearance to withdraw without shearing the still-green (unsintered) compact. The supplier caught it, but it cost us two weeks of re-tooling discussion that a five-minute conversation up front would have avoided.

Where PM Beats Machining and Where It Doesn’t

The economic case for PM is almost entirely about volume and shape complexity. Below a few thousand units a year, the tooling cost rarely pays back versus machining from bar stock. Above tens of thousands of units, especially for parts with features like flats, small gear teeth, or non-round profiles that would require multiple machining operations, PM often wins by a wide margin because the shape comes out of the press essentially net or near-net.

Where PM struggles is in applications demanding high toughness or impact resistance. The residual porosity that helps keep cost down also reduces elongation and impact strength compared to wrought material of the same nominal alloy. On a case I worked years ago involving a small sprocket for a conveyor drive, the client wanted to switch from forged steel to PM for cost reasons. I ran the numbers and pushed back — the application saw frequent shock loading from jammed product, and PM’s lower fracture toughness made it a poor fit even though the static load calculations looked fine on paper. We ended up compromising on a PM body with a heat-treated, higher-density rim near the teeth, which is a good example of how PM lets you tailor properties locally in a way solid machining cannot.

Secondary Operations: Sizing, Infiltration, and Heat Treatment

Most functional PM parts don’t leave the process as-sintered. Sizing — essentially re-pressing the sintered part in a precision die — tightens dimensional tolerances that as-sintered parts can’t hold reliably, especially on hole diameters and flatness. If your design calls for a tolerance tighter than roughly ±0.05-0.1 mm on a critical dimension, plan for a sizing operation from the start rather than hoping the as-sintered process delivers it.

Infiltration, where copper is drawn into the pore network during a second heat cycle, is something I use when I need better machinability or improved strength without going all the way to a wrought equivalent. Steam treatment, which forms a thin oxide layer in the surface pores, is a low-cost way to improve wear and corrosion resistance on parts that don’t need full plating. And conventional heat treatment — carburizing, induction hardening — works on PM parts too, but the porosity means case depth and hardness profiles behave differently than in wrought steel, so I always ask for sectioned hardness traverse data on the first article rather than trusting a generic heat-treat spec.

A Gear Project That Taught Me About Porosity

The clearest lesson I’ve had on PM came from a small helical gear used in an office equipment drive train. The part had run fine in prototype using machined steel, and the switch to PM for cost reduction seemed routine. In production testing, we started seeing intermittent tooth pitting failures that hadn’t shown up in the machined prototypes. After a few frustrating weeks, metallurgical sectioning showed the pitting initiated at surface-connected pores right at the pitch line, where contact stress was highest. The porosity that was perfectly acceptable in the tooth root, where stresses were lower, became the failure origin exactly where load was concentrated.

The fix wasn’t exotic — we specified a higher compaction density for that gear along with a light surface densification pass on the tooth flanks — but the underlying lesson stuck with me. Porosity is not uniformly “bad,” it’s a variable you have to manage regionally, matching density to the local stress field rather than treating the whole part as a single spec. That project changed how I read PM drawings from anyone else on my team; I now specifically look for whether density requirements are called out by feature, not just as a blanket callout in the title block.

Working With a Powder Metallurgy Supplier

Because PM tooling is expensive and design changes after tooling cut are painful, I’ve found the projects that go smoothly are the ones where the supplier is involved during concept, not after the drawing is “final.” Good PM suppliers can tell you within a day whether a feature is pressable as drawn, and they often suggest small geometry changes — a slightly different fillet, a chamfer instead of a sharp edge — that make a meaningful difference in tool life and part consistency without affecting function.

I also make it a habit to request first-article density measurements and, for anything structural, a short fatigue or burst test if the application justifies it, rather than relying purely on the supplier’s standard material certification. It costs a little time up front, but I have never regretted verifying rather than assuming with PM parts. The process rewards designers who treat it as its own discipline instead of a cheaper twin of machining, and it punishes those who don’t, often quietly, in a failure mode that only shows up months into production.