Additive Manufacturing of Mold Steel: How It Works and Where It Stops

• 1
  Key takeaways
• 2
  What additive manufacturing of mold steel actually does
• 3
  Powder chemistry controls density and hardness
• 4
  Conformal cooling: where the design gain comes from
• 5
  Post processing decides the final tolerance
• 6
  When printed mold steel is the wrong choice
• 7
  Printed insert vs CNC cut insert

Additive manufacturing of mold steel means building a tool insert layer by layer in a metal powder bed, usually by laser powder bed fusion (LPBF). A 200 W to 400 W fiber laser melts a 20 to 60 μm layer of gas atomized tool steel powder, then a recoater spreads the next layer. The part grows on a steel plate, supported by printed anchors that are cut off later.

The material is not exotic. Most shops run H13 (1.2343), 1.2709 maraging steel, or a 420 stainless variant. These are the same families used in die casting dies, injection molds, and stamping tools. The difference is thermal history. Each pass sees melt and rapid solidification, so the microstructure is fine and the residual stress is high.

That stress matters. A printed block will bend if you cut the supports and machine it without a stress relief cycle. Typical practice is a 500 to 600 °C furnace hold for 2 to 4 hours before any semi finish cut. Skip it and a 150 mm insert can move 0.3 mm or more once the supports are released.

So the process is not a replacement for milling. It is a way to place cooling channels, venting, and thin cores where a drill cannot reach. Everything else still goes through the same CNC and EDM route as a conventional insert.

• 1
  Layer thickness20–40 μm for fine detail, 40–60 μm for faster build on bulky inserts.
• 2
  Build plate1.2343 or 1.2709 plate, machined flat before the first layer.
• 3
  AtmosphereArgon or nitrogen, oxygen held below 100 ppm to limit oxide inclusions.

Powder is where most quality problems start. Good LPBF powder is gas atomized, spherical, and screened to 15–53 μm. Particles outside that band either fail to spread or fail to melt through. A batch with 0.05% oxygen behaves differently from one at 0.15%, even with identical laser power.

Density above 99.5% is the usual target for tooling. Below that, pores open up during polishing and show as pits on a mirror surface. Porosity also drops fatigue life in a core that sees 200,000 cycles. You cannot fix it after the fact with heat treatment.

Hardness after hardening depends on the alloy. H13 printed and hardened to 48–52 HRC is common for die casting. Maraging 1.2709 can reach 50–54 HRC with a simpler aging cycle, around 490 °C for 6 hours, and shrinks predictably, which helps when you need tight cavity dimensions.

One practical limit: recycled powder. Reuse is fine, but oxygen and satellite particles build up. Most shops cap reuse at 10 to 15 cycles and re-screen each time. Ask for the powder lot and the reuse count if the insert is for a high volume tool.

• 1
  Particle size15–53 μm, spherical, low satellite content.
• 2
  OxygenKeep under 0.10% for H13 and 1.2709 tooling powder.
• 3
  Density target≥99.5% for polished cavities, ≥99.0% for non cosmetic cores.
• 4
  ReuseCap at 10–15 cycles with re-screening between builds.

A drilled cooling line is straight. A printed one can follow the cavity contour at a constant 8 to 12 mm offset. In a deep core or a part with a long thin rib, that changes cycle time, because heat leaves through the steel instead of sitting in the hot spot.

Channel diameter matters. Below 3 mm, powder removal gets hard and the risk of a blocked line rises. Above 12 mm, the wall between channel and cavity gets thin, and the insert may not survive injection pressure. Most production tools sit in the 4 to 8 mm range.

Cross sections should stay circular or slightly oval. Sharp internal corners trap unmelted powder and create stress risers. If two channels cross, model a smooth blend rather than a T joint.

There is a trade. A conformal core often needs more support structure during the build, and those supports leave a rough surface that has to be machined or EDM'd away. Budget that work into the cost, not just the print time.

• 1
  Offset from cavity8–12 mm gives a good balance of heat transfer and wall strength.
• 2
  Channel size4–8 mm for production; 3 mm is the practical floor.
• 3
  Powder removalPlan at least two open ends per channel for evacuation and cleaning.

A printed insert comes off the plate at roughly ±0.2 to ±0.5 mm on external faces. That is not a mold tolerance. The parting line, the cavity, and any shutoff still need CNC milling, grinding, or wire EDM after the build.

The sequence matters. Cut the part from the plate, stress relieve, then semi finish. Harden and temper, then finish grind or EDM to size. If you harden before removing supports, the part can warp in the furnace and you lose the datum.

On a 5 axis machining center with a Ø400 mm rotary table, a printed insert up to 400 mm across can be finished in one setup, which keeps the cavity and the parting line aligned. Larger inserts, up to our 4,000 mm travel, need a different fixturing plan and usually more than one setup.

Surface finish after finishing is typically Ra 0.8–1.6 μm on machined faces and Ra 0.2–0.8 μm after polishing. A mirror cavity needs the higher density powder and a slower finish pass, since any pore at the surface will open during polish.

• 1
  As built tolerance±0.2 to ±0.5 mm before any machining.
• 2
  Machined toleranceDown to ±0.005 mm on critical shutoffs and fits.
• 3
  HardeningH13 at 48–52 HRC, 1.2709 aged to 50–54 HRC.

If the tool is a flat plate with straight cooling lines and a run under 50,000 shots, print it and you pay more for no gain. A CNC cut 1.2343 block is faster, cheaper, and easier to repair. We machine those every week.

Large simple cores also favor subtractive. A 600 mm block with no hot spots does not need a printed channel. The build time alone can run several days, and the risk of distortion grows with part height.

Repair is the other limit. Welding a printed H13 insert is possible, but the heat affected zone differs from the base metal. For a high volume die, a printed insert that cracks may not be repairable to the same life. Plan a spare.

Where printing wins is clear: deep ribs, tall thin cores, hot spots that drive cycle time, and small inserts with complex venting. If two or three of those apply, the business case usually holds.

• 1
  Choose printingHot spots, deep ribs, conformal channels, small complex cores.
• 2
  Choose CNCFlat plates, short runs, simple straight cooling, large plain cores.
• 3
  Watch repairPrinted inserts can be harder to weld back to original life.