How Honda Uses Laser Powder Powder Fusion Technology for Transport Parts

• 1
  Key takeaways
• 2
  What laser powder powder fusion technology actually does
• 3
  The process window that keeps a build sound
• 4
  Powder chemistry and reuse limits
• 5
  Where the geometry fights back
• 6
  What printing changes for the whole assembly
• 7
  Cost, lead time, and when to skip printing
• 8
  Print, machine, or both

Laser powder powder fusion technology spreads a thin layer of metal powder across a build plate, then scans a laser over selected areas. The beam melts the powder into a small pool that fuses with the layer below. The plate drops by one layer thickness, a fresh coat of powder rolls across, and the cycle repeats. A transport part grows this way at 20 to 60 μm per layer.

The pool is small. On a typical 400 W system it stays near 100 to 200 μm wide and exists for less than a millisecond. Solidification happens fast, so the grain structure is fine and the alloy can differ from the same metal cast or forged. That is where the strength comes from, and also where residual stress comes from.

Honda used this method for Formula 1 pistons and turbine housing areas, plus wheelchair handlebars shaped by topological optimization. The common thread is geometry that is hard to cut: internal passages, organic ribs, or weight that must come out of a part without losing stiffness.

Think of it as micro-welding with a powder feed, not as printing. Every scan track is a weld bead. If you would not trust a weld in that location, do not expect the printed part to behave differently.

• 1
  Layer thickness20–60 μm is the usual range for metal transport parts.
• 2
  Melt pool widthRoughly 100–200 μm on a 400 W machine.
• 3
  Cooling rateFast enough to refine grains, fast enough to trap stress.

Laser power, scan speed, hatch spacing, and layer thickness set the energy density delivered to the powder. Too little energy and you get lack-of-fusion porosity: unmelted powder trapped inside the part. Too much and you get keyhole porosity plus spatter that lands back in the pool. Both defects survive to the finished component unless you catch them.

Honda's answer to that risk was in-situ monitoring. A high-speed camera images each layer, so the team sees smoke, spatter, and pool anomalies while the part is still being built. That feedback loop is what makes a repeatable process instead of a lucky one.

For transport parts, the practical window is narrow. A turbine housing area or a piston crown carries thermal load, so porosity is not cosmetic. If your part will see cyclic stress, budget for process qualification, not just a build.

We see the same logic on the machining side. A process that is not monitored is a process you cannot quote with confidence. Printing adds one more variable: the powder itself.

• 1
  Lack-of-fusionLow energy density leaves unmelted powder and weak bonds.
• 2
  Keyhole porosityExcess energy drills a deep vapor cavity that collapses.
• 3
  SpatterEjected droplets land in the pool and create inclusions.

Metal powder is not a commodity filler. Particle size distribution, sphericity, oxygen content, and flowability all shift the melt behavior. Gas-atomized powder in the 15 to 45 μm range is the common feed for transport work. A batch with too many satellites or too much oxygen will spatter more and build denser oxide inclusions.

Reuse is the quiet cost driver. Every build cycle exposes powder to heat, moisture, and handling. Oxygen picks up, fines accumulate, and flow changes. Most shops sieve after each run and cap the number of reuses, then blend virgin powder back in. That cap is a quality decision, not a purchasing one.

Cross-contamination matters when one machine runs aluminium, stainless, and titanium. Residual powder from a previous job can seed a defect in the next. Dedicated powder handling and clean-down procedures are part of the process, not housekeeping.

If a supplier cannot tell you the reuse count and the oxygen spec for your build, you do not have a controlled process. You have a print.

• 1
  Typical size range15–45 μm gas-atomized powder for metal AM.
• 2
  Reuse capSet per alloy; sieve every cycle and track oxygen.
• 3
  Contamination riskMixed alloys in one machine seed inclusions.

A printed part needs support where it overhangs, and that support has to be cut off later. Anything steeper than about 45° from the build plate usually needs it. Supports are not free: they add material, add build time, and leave witness marks on the surface.

Thick sections are the other problem. A 25 mm solid block cools slower in the middle than at the edge, so the part warps as it cools and can crack off the plate. Designers often hollow those regions or switch to a lattice so the heat has somewhere to go.

Thin walls have a floor too. Below roughly 0.4 mm, a single scan track may not form a stable wall, and the part becomes fragile during depowdering. Honda's wheelchair handlebars sit in the sweet spot: thin, organic, and stiff because of topology, not because of mass.

This is the same trade an engineer makes on a CNC part. You remove material where it does not carry load, and you keep it where it does. Printing just lets you do that in three dimensions instead of two.

• 1
  Overhang angleAbove about 45°, plan for supports.
• 2
  Thick sectionsHollow or lattice them to control cooling.
• 3
  Thin wallsKeep above roughly 0.4 mm for a stable track.

A printed part rarely ships as-built. Bearing bores, sealing faces, and thread features are cut after the build because the as-built surface is too rough and the tolerance is too loose. That hybrid route is normal for transport hardware: print the shape, machine the interfaces.

That is where a shop like ours fits. We run 16 simultaneous 5-axis machining centers, 12 four-axis mills, 27 three-axis machines, and 16 mill-turn centers across 127 high-precision CNC machines. We hold ±0.005 mm and finish to Ra 0.2–0.8 μm when a sealing face demands it.

The design freedom of laser powder powder fusion technology is real, but it shifts work downstream. You still need a datum, you still need a fixture, and you still need inspection. A printed part with a bad datum is harder to hold than a billet, not easier.

Treat the print as a near-net shape. Then the assembly tolerances close the way they always did, on the machine.

• 1
  Post-machiningBores, faces, and threads after the build.
• 2
  Datum strategyPick datums early; printing does not remove them.
• 3
  InspectionCT or sectioning for internal features, CMM for interfaces.

Printing wins when the geometry is the value. A part with internal channels, a lattice, or a shape that would take five setups on a mill is a strong candidate. Printing loses when the part is mostly prismatic, needs tight tolerances everywhere, or will be made in large numbers.

Build time is the cost driver, not material alone. A tall part with supports can run for days on one machine. That machine is not making anything else. For a transport program, that trade is usually settled by how many parts you need and how fast.

If you need one prototype next week, both routes work. If you need 10,000 brackets, machining and die casting will beat printing on unit cost. If you need 200 turbine housings with internal cooling, printing plus finish machining is the sensible path.

We quote both routes when a part is ambiguous. A DFM review within 12 hours usually settles it: the geometry tells you which process belongs on the job.

• 1
  Print whenInternal channels, lattices, low volume, hard geometry.
• 2
  Machine whenPrismatic shape, tight tolerances, high volume.
• 3
  Hybrid whenPrinted shape plus machined interfaces.