The first batch of MIM parts comes back from the furnace warped, and the room immediately blames the tool. That is usually the wrong first move. On the shop floor the diagnosis order is fixed, and it has four steps: measure the dimensions and decide whether this is a shrinkage offset or a shape that actually moved; cut a section open and look at the density distribution; check the ramp rate through the critical debind window; and look at how the part was supported during sintering.
Most warpage traces back to the tool anyway — but not in the way people mean. Uneven wall thickness and an off-centre gate leave a density gradient inside the green part. Debinding removes the binder that gave the part its strength, the gradient gets amplified, and in the high-temperature stage there is nothing left to hold the shape. Three things have to be settled on the drawing before the tool is cut: even wall thickness, a symmetric gate, and a usable support surface for sintering.
The Numbers First
Linear shrinkage for MIM commonly lands between 12% and 20%; 316L usually sits at 15–18%. The cavity is scaled up by one factor, and any anisotropy has to be handled on the process side — scaling the tool cannot fix it.
Green density runs about 60–65% of theoretical. Debinding has to carry away roughly half of the binder, and what remains is a small fraction of the pressed strength. The usual practice is to hold the ramp through the critical debind window at 0.5–1 °C/min.
Wall thickness belongs between 1 and 4 mm. Thickness variation within one part should stay inside ±10%, and ribs are usually taken at 0.5–0.6 times the main wall. Past that, the distortion caused by the density gradient exceeds what the form tolerance can absorb.
Shrinking Too Much and Losing Shape Are Two Different Failures
Spread the CMM report out and do not start with single-point deviations. If every critical dimension is off in the same direction and by the same proportion — if one scale factor brings length, width and height back onto the drawing — then this has nothing to do with warping. The scale factor is out of step with the actual powder fill. You change process parameters; the tool stays. It is common enough on the floor: a new powder lot, a different tap density, and the shrinkage drifts.
The second case is the expensive one. Individual dimensions run both plus and minus, or the part bows and twists as a whole. One recurring shape in our files is a long bracket: length is in tolerance, but the middle bows 0.2–0.4 mm and both ends lift to the same side. No scale factor rescues that. Go down into the density distribution.
Cut One Section and Half the Problem Is Settled
Take the worst part out of the batch and cut it along the direction of distortion. A uniform section has consistent granule distribution — the same colour and texture across the whole face. If the side near the gate is dense while the far end is loose, or a clear density boundary shows up across a wall transition, the green part was already uneven before it entered the furnace. Uneven green density means uneven shrinkage, and distortion follows necessarily.
Go back to the tool drawing and two problems usually match up. First, an abrupt wall change — a region stepping from 1.5 mm to 3.5 mm — where the thick side packs more powder, debinds more slowly and shrinks more, while the thin side does the opposite, and the two pull the part into a bow. Second, an asymmetric gate: a single side gate placed off the geometric centre drops the weld line in a thin-wall region, and that patch starts out with lower density.
| What you see after sintering | Where it is more likely to originate | Direction of the tool-side fix |
|---|---|---|
| Uniform scale up or down, shape intact | Scale factor, powder fill | Tool untouched; fix shrinkage in process |
| Bowing in one direction, worst at the largest wall difference | Green density gradient | Even wall thickness, symmetric gate |
| Corners or thin edges sagging, flat face dipping in the middle | No support in the high-temperature stage | Reserve a flat support face at design stage |
| Blisters or delamination cracks | Debind ramp too fast, gas cannot escape | Coordinate the process ramp, add venting in thick regions |
The tool-side answer is direct: level the wall thickness, and get stiffness from flanges and ribs instead of piling on material; move the gate to a symmetric layout — two-point or fan — so the weld line lands somewhere non-critical.
The Trap in Debinding Is the Default Ramp
Debind-stage distortion is the most frequently misdiagnosed as a tool problem. Binder removal is a process of escorting the binder out of the green part a little at a time; the part loses strength first and then holds itself up on the residual skeleton. Typical two-stage debinding pulls most of the low-molecular wax out by solvent extraction, then burns off the rest in a thermal debind furnace. If that furnace follows a default ramp of two or three degrees per minute, gas cannot diffuse out of the critical window, internal pressure pushes the softened part out of shape, and the severe cases blister and delaminate.
My test: if the distortion appears only after thermal debinding and before sintering — the green part measured fine — lock onto the debind stage. Do not touch the tool yet. Ask the factory what ramp rate that critical window actually ran at; in most cases holding it under one degree per minute contains the problem.
An uneven wall thickness makes debinding worse on its own: binder has to travel different distances in different regions, and that is an amplifier. Levelling the wall helps here as well.
In Those Few Minutes of Sintering, Support Decides the Shape
Late in the sinter cycle the particles weld and shrink, and the part is close to soft. Its own weight is enough to make it sag. Flat and long parts without a flat support face — resting on a few spacer points — will bow in the middle every time.
Support material is usually alumina or zirconia ceramic plate. The contact should be small and flat: line contact leaves marks, and full-surface contact blocks gas escape. The requirement this places on the tool design is specific. Work out the relationship between the parting plane and the visible faces in advance, put the support-allowed face on a non-A surface, and remove the marks with one more operation after sintering. Parts that are all curved surface with no flat anywhere need a process ear in the same material at the layout stage, cut off after sintering with the cut placed away from the primary visual face.
That adds one structure on the tool side, and it is far more stable than relying on straightening afterwards. Straightened parts carry internal stress, and they spring back after a few days on the shelf.
Run this sequence and the changes that land on the tool drawing usually concentrate in three places: walls made even at the design source, with stiffness carried by geometry; a symmetric gate layout so the density distribution stays as uniform as possible; and a clean support location reserved for sintering. Half of what happens in debinding and sintering is process parameters. The other half was written down before the tool was cut.
FAQ
How should the MIM cavity be sized, and what shrinkage value do I use?
Scale the cavity by a factor taken from measured shrinkage. Common alloys run 12–20% linear; 316L is usually taken at 15–18%. The value has to come from a sample of the same powder lot and process, not from an experience number. Shrinkage moves with particle size, fill density, sintering temperature and hold time, and even one supplier's lots drift. Run one or two trial moulds before finalising the cavity and you save a round of tool changes.
Thicker walls should be stiffer — why do they cause problems?
Past 4 mm the binder has a longer path out, gas cannot escape, and the debind stage blisters and delaminates. Thick and thin regions also shrink out of step, which pulls the part. MIM walls stay at 1–4 mm; get stiffness from flanges or ribs, with rib thickness at 0.5–0.6 times the main wall and a fillet at the root. Prefer a gradual ramp over a step at wall transitions, and keep total variation inside ±10%.
What kind of parts suit MIM, and when should it be avoided?
MIM fits small metal parts of 0.1–300 g, generally under 100 mm, with complex geometry and real volume. When the part is large and heavy, or the shape is simple, machining or die casting is usually more competitive and more stable. Parts with wildly uneven walls or extreme length-to-diameter ratios carry high distortion risk in debinding and sintering — check wall uniformity before choosing the process. And if the volume is too low to amortise the tool, that alone is a common reason to walk away from MIM.
Can slight distortion after sintering be fixed by straightening?
Straightening recovers some of it, but straightened parts carry internal stress, spring back over time and lose dimensional stability. When volume production relies on straightening as a fallback, the upstream process has a systematic problem, and the answer is back in density distribution and support. For one-off or small batches with loose tolerance, cold straightening is a reasonable supplement — but on visible faces, the contact marks usually need an extra finishing step to cover.
Top comments (0)