MIM parts contract about 15–20% linearly during sintering — the dominant dimensional change in the whole process — so the mold cavity is deliberately cut oversized by a scale factor of roughly 1.18–1.25× so that the sintered part lands on nominal. The shrinkage itself is predictable and designed around; what actually pushes a part out of tolerance is variation in shrinkage. Holding the as-sintered band of ±0.3–0.5% of nominal dimension therefore depends on uniform green density and furnace control, not on the scale factor alone.
This page covers the mechanics: where the shrinkage comes from, how it converts into a tool scaling factor, why it is near-isotropic but never perfectly so, and which process levers keep lot-to-lot variation inside the tolerance band. For what the process can hold on the finished part — including secondary operations that reach approximately ±0.01 mm — see the companion MIM tolerances guide; this page stays on the mechanism and the compensation method.
Where MIM Shrinkage Comes From
A molded MIM green part is a mixture of metal powder and binder, full of porosity. Debinding removes most of the binder and changes dimensions only slightly. The real contraction happens in sintering: as the furnace temperature rises, the powder particles weld together at contact points, pores close, and the compact densifies toward 95–99% of theoretical density. The volume loss has to come out of the linear dimensions, and that is the shrinkage the tool must pre-compensate.
Published values for the linear contraction span roughly 12–20% depending on alloy, feedstock solids loading, and process window, with about 15–20% being the most common design assumption for first-pass tooling. Because the driving mechanism is densification, the shrinkage of a given geometry is repeatable once the feedstock and furnace recipe are fixed — that repeatability is exactly what makes compensation practical. The MIM process steps overview shows where debinding and sintering sit in the chain.
Converting Shrinkage Into a Tool Scale Factor
The cavity is not cut at the shrinkage percentage — it is cut at the inverse of the remaining fraction. A part expected to shrink 15% linearly needs a cavity about 1.18× nominal, because the sintered part retains only 1/1.18 of the molded length. The same logic gives about 1.25× for 20% shrinkage. Expressed as a cavity expansion ratio rather than a percentage, published factors for common MIM alloys cluster around roughly 1.18–1.23×.
| Expected linear shrinkage | Tool scale factor | Where it typically applies |
|---|---|---|
| About 12–15% | ≈1.14–1.18× | Lower-shrinkage feedstocks and tighter process windows |
| About 15–20% (common design assumption) | ≈1.18–1.25× | Most common MIM alloys, first-pass tool design |
| About 20% or more | ≈1.25× and above | Selected alloys or conditions; confirm by trial shots |
The factor is empirical, not a material constant. Two parts in the same alloy can need different factors when their geometry, packing behavior, or sintering support differs, so a serious supplier validates the scale with trial shots: mold, measure the sintered part, adjust the cavity or the process, then lock the tool. This is why tooling lead time matters — Emitech cuts MIM tooling in 15–20 days from approved drawing, and that window includes first-article verification rather than blind scaling from a generic table.
Why Shrinkage Is Never Perfectly Isotropic
First-pass design treats shrinkage as near-isotropic, and for compact, uniform parts that assumption holds well. But several mechanisms bias contraction along one axis more than another: wall-thickness gradients that densify unevenly, fill patterns that leave density gradients in the green part, gravity and support conditions on the sintering setter, and friction where the part rests during firing. Long slender parts and sections with abrupt thickness changes show the effect most.
The practical consequences are two. First, a single scale factor may need directional refinement after the first sintered measurements — a few tenths of a percent of extra contraction in one axis is enough to eat a large share of the tolerance band. Second, anisotropy is best designed out before it is measured out: uniform walls, gradual section transitions, and self-supporting sintering postures reduce directional bias at the source. The DFM guidelines for MIM cover those geometry rules in detail.
Holding the ±0.3–0.5% As-Sintered Band
Shrinkage and tolerance are different quantities, and confusing them causes most MIM dimensional disputes. Shrinkage is the average contraction — large, but compensated in the cavity. The tolerance band is the allowable part-to-part and feature-to-feature spread, and for well-controlled MIM work it is ±0.3–0.5% of nominal dimension, about ±0.025 mm on features under 3 mm. The band is filled not by average shrinkage but by its variation.
- Green density uniformity: a stable injection window and balanced fill keep density gradients — the root of both anisotropy and scatter — out of the green part.
- Feedstock consistency: powder particle size distribution and solids loading shift the shrinkage factor, so lot control on feedstock is lot control on dimensions.
- Furnace uniformity and recipe control: temperature spread across the load and drift across runs converts directly into dimensional spread; debinding and sintering equipment matter as much as the mold.
- Measurement-based tool lock: the scale factor is finalized from first-article data, then frozen — process discipline afterward keeps the centered part centered.
Equipment depth is the invisible variable behind a stable band. Emitech runs 22 sintering furnaces (4 continuous + 18 vacuum) with dedicated debinding upstream, which keeps families of parts on consistent furnace recipes instead of mixing them across mismatched equipment — full plant detail is on the MIM capabilities page.
When Compensation Alone Is Not Enough
Correct scaling centers the part on nominal; it does not make every feature capable of every print callout. Features that must hold closer than the stable as-sintered band — sealing lands, bearing bores, gauge datums — are routinely finished with secondary operations: sizing can press critical dimensions into place, and CNC machining or grinding takes treated features to approximately ±0.01 mm. Emitech maintains 24 sizing presses and 14 grinding machines for exactly this route.
The economical pattern is to let the mold handle the 90-something percent of dimensions that live comfortably inside ±0.3–0.5%, and reserve secondary work for the handful that do not. Deciding that split at DFM stage — rather than discovering it at first article — is what keeps both tolerance and unit cost on target.
Frequently Asked Questions
Q: What shrinkage value should be assumed for first-pass MIM tool design?
About 15–20% linear is the common first-pass assumption, giving a cavity scale factor near 1.18–1.25×. Published values span roughly 12–20% depending on alloy and process, so the starting value must be treated as provisional until first-article measurements confirm or refine it.
Q: How is expected shrinkage converted into a cavity scale factor?
Divide the nominal dimension by the fraction that remains after shrinkage: a 15% contraction needs about 1.18× nominal and a 20% contraction about 1.25×. The factor is then finalized from measured sintered parts, because geometry and furnace support shift it part by part.
Q: What causes anisotropic shrinkage in MIM parts?
Density gradients from unbalanced fill, wall-thickness transitions that densify unevenly, gravity and setter friction during sintering, and support conditions in the furnace. The effect is usually modest on compact, uniform parts and largest on long slender geometries and abrupt section changes.
Q: How does shrinkage compensation differ from tolerance control?
Compensation handles the average contraction by cutting the cavity oversized; tolerance control handles the variation around that average. The as-sintered band of ±0.3–0.5% of nominal is governed by green-density uniformity, feedstock consistency, and furnace control — scaling the CAD correctly does nothing for lot-to-lot scatter.
Q: When does a MIM part still need secondary operations after correct scaling?
When individual features must hold closer than the stable as-sintered band of ±0.3–0.5%, or when distortion on a specific feature cannot be designed out. Sizing, CNC machining, or grinding takes treated features to approximately ±0.01 mm; the economical split is decided at DFM, not at first article.
Shrinkage compensation is a methodology, not a number from a table: assume 15–20%, scale the cavity accordingly, verify on sintered first articles, and control green-density and furnace uniformity to keep the band tight. If you have a part whose dimensional callouts are pushing against what you expected from MIM, send the drawing through the contact page — Emitech returns a quotation with DFM feedback within 24 hours, including how each critical dimension would be held.
Precision Metal Parts from Emitech
Nanjing Emitech delivers MIM, CNC machining, and custom metal parts. MIM services · Request a quote