Why MIM Parts Fail — and How to Tell a Good MIM Supplier from a Bad One

MIM's Trust Problem Is Real — and Mostly Deserved by the Bad Suppliers

Walk through any firearms, knife, or watch forum and you will find the same argument running for years: metal injection molding parts are "mud metal," they "always have voids," they fail where machined parts survive. One frequently quoted forum summary claims MIM parts are only "90-95% as dense as forged steel" and that "voids within an MIM part are invisible." The people writing those threads are not wrong about the parts they handled — they are wrong to generalize from them.

MIM (metal injection molding) is a mature aerospace and medical process: injection-mold a wax-polymer feedstock loaded with fine metal powder, remove the binder, and sinter the part to near-full density. Done well, it produces complex parts at 96-98% relative density with properties that satisfy structural and surgical applications. Done badly, it produces exactly the porous, void-ridden parts the forums describe. The gap between the two is not luck — it is process control, and process control is auditable. This page covers the three stages where MIM parts actually fail, the numbers that reveal quality, and the questions that expose a weak supplier before you find out with a field failure.

For the process-selection side of the question — when MIM beats press-and-sinter or machining for your part — read our PM vs MIM comparison; this page is about quality, not choosing the route.

Where MIM Parts Actually Fail: Three Stages, Three Defect Families

MIM shrinks the part roughly 18-22% during sintering (the feedstock is only about 40-68% metal powder by volume), and every defect family traces back to how that shrink was managed:

Two things follow from this table. First, the defects that kill parts in service are internal — a knit line or void does not show on a machined surface, which is exactly why forum posters distrust what they cannot see. Second, every one of these mechanisms is a control problem: fill profiling, debind ramp rates, sintering atmosphere and temperature uniformity. Good suppliers measure them every batch; bad suppliers discover them in returns.

Density: The Number That Separates Good MIM from Bad

Relative density is the single most diagnostic MIM metric, because it quietly aggregates every upstream defect:

• Well-run structural MIM sinters to roughly 96-98% of theoretical density for stainless and low-alloy steels. The residual 2-4% is fine, closed, evenly distributed porosity — it knocks some fatigue performance off wrought values but behaves predictably.
• The parts that earn the "mud metal" name sit meaningfully below that, with porosity that is coarser, connected, or clustered near knit lines. A part at 90-93% relative density can look identical, machine identically, and fail in a few thousand cycles.
• The measurement is cheap and standard: Archimedes (immersion) density on a per-lot basis is the industry baseline, and it catches sintering drift weeks before a customer notices. Internal voids need more — X-ray or CT scanning on a sampling plan is how serious MIM houses verify what density alone cannot see.

This is the audit hinge: a supplier who cannot show per-lot density data going back months is not running a controlled process, whatever their brochure says. When we assess a candidate supplier for our own supply chain — or when a customer audits us — density history is the first document requested, alongside the quality-system evidence described on our certifications page.

Because the fatal defects are internal, it also helps to know which incoming check can actually see which defect:

The pattern to notice: the defects that kill parts in service are exactly the ones a surface look misses. That is why every audit question in the next section asks for data rather than inspection — and why a supplier who treats X-ray sampling as optional is telling you where their parts sit in the table above.

Six Questions That Audit a MIM Supplier

Bring these to the audit or the RFQ. Each one has a right answer that costs the supplier nothing to produce if the data exists:

1. "Show me per-lot density results for a recent production part." Trended, in-spec, with a stated target — not a single certificate from three years ago.
2. "How do you detect internal voids?" X-ray or CT sampling on defined frequencies is the credible answer; "we haven't needed that" is a red flag for anything load-bearing.
3. "What is your debind process, and how do you know it completed?" Weight-loss tracking per debind cycle, with defined endpoints — the stage most correlated with cracking.
4. "What shrinkage value do you tool for, and how do you hold it?" A precise number (typically in that 18-22% band) with per-axis compensation and first-article dimensional data proving it.
5. "What happens when a lot drifts?" A real answer names the reaction plan — containment, dimensional check, density re-test — not "it doesn't drift."
6. "Can I have a PPAP-style first article with material certification?" The transition from prototype shop to production supplier is exactly here; a supplier who resists documentation will resist it forever.

Six yes answers do not make a supplier good, but any single no tells you where the failure modes live.

When MIM Is the Right Process — and When It Is Not

An honest quality page has to say this plainly: MIM is the correct route for small, complex, three-dimensional parts — external threads, undercuts, cross-drilled features, thin walls — at volumes from tens of thousands up, where the alternative is multi-axis machining of each feature. It is the wrong route for parts that are essentially flat (press-and-sinter does that far cheaper — see the PM vs MIM comparison), for a few hundred pieces (machining wins), and for applications where the buyer cannot afford any audit effort at all.

Cost-wise, the crossover between MIM and machining moves with part complexity, and the PM cost calculator shows the tooling-amortization arithmetic that frames that decision. Quality-wise, the choice of supplier matters more than the choice of process: a controlled MIM line outperforms an uncontrolled machining cell in consistency, and vice versa. If you want to know who you are dealing with before sending a drawing, our about page says who we are, and we will happily answer the six questions above about our own shop floor.