Ultra-fast 3D printed propeller launcher technology

• 1
  What an ultra-fast 3D printed propeller launcher has to survive
• 2
  How LPBF builds the part, layer by layer
• 3
  Where printing stops and CNC takes over
• 4
  Balance, fatigue, and inspection on a rotating part
• 5
  Additive or CNC: matching the process to the feature

A launcher is not a decorative cap. It is a loaded interface. The blades sit folded or stowed, a release mechanism lets go, and centrifugal force plus spring or gas pressure swings them out in milliseconds. Every gram on the hub adds to the inertia the motor has to spin up, and every millisecond of delay changes thrust response.

So the load case mixes three things at once: hoop stress from rotation, bending at the blade root, and impact when the blade hits its stop. The part also has to keep its balance. A few grams of asymmetry at Ø300 mm shows up as vibration long before it shows up as a crack.

That is the gap ultra-fast 3D printed propeller launcher designs try to close. Metal additive, usually laser powder bed fusion (LPBF), builds the hub as one piece with internal channels, hollow ribs, and blade-root pockets that would need five setups and a lot of tool reach on a mill.

The trade is real, though. Printing gives geometry freedom; it does not give you a bearing bore or a thread without follow-up machining. The rest of this page separates the two.

• 1
  RotationHoop stress rises with the square of rpm, so balance matters as much as wall thickness.
• 2
  ImpactThe blade stop sees a sharp load spike each cycle, not a smooth ramp.
• 3
  InertiaHollow ribs cut mass at the rim, which is where mass costs the most.

LPBF spreads a thin layer of metal powder, melts the cross-section with a laser, then drops the build plate and repeats. Layer thickness for aluminum and titanium usually sits between 30 μm and 60 μm. A 120 mm tall launcher hub is therefore 2,000 to 4,000 passes of the recoater.

That number explains the lead time and the surface finish. Down-facing surfaces and overhangs are supported by the powder itself or by printed supports. Anything steeper than roughly 45° from the build plate needs support, and support removal leaves witness marks that have to be machined or blasted off.

Material choice drives the rest. AlSi10Mg and 6061-class aluminum print well and keep mass low. Ti-6Al-4V (TC4) gives higher specific strength and better fatigue behavior, at a higher cost per kilogram and slower build rates. 17-4PH stainless is common where corrosion resistance matters more than weight.

Two details bite engineers new to the process. First, residual stress: the melt pool cools fast, so long thin sections can curl. Second, anisotropy: properties along the build direction differ from properties across it. Orient the hub so the highest tensile load runs in the plane of the layers, not across them.

• 1
  Layer thickness30–60 μm typical; thinner layers improve finish but slow the build.
• 2
  Overhang limitBelow about 45° from the plate, add support or redesign the rib.
• 3
  Build orientationAlign the main tensile axis with the layer plane.
• 4
  Stress reliefHeat treat after printing and before any finish machining.

As-printed LPBF parts hold roughly ±0.1 mm on a good day, and that tolerance moves with part size, orientation, and thermal history. If your launcher needs a shaft bore that runs true to the blade axis, printing alone will not get you there.

The usual route is hybrid: print near-net, heat treat, then machine the critical features. At GreatLight we hold ±0.005 mm (±0.0002 in) on the machined features, with finishes from Ra 1.6–3.2 μm as-machined down to Ra 0.2–0.8 μm where a bearing or seal seat needs it.

Which features go on the mill? Bearing bores, dowel holes, thread forms, sealing faces, and the blade pivot bore. Which stay as printed? Internal cooling or wiring channels, hollow ribs, lightening pockets, and non-critical mounting bosses.

One more boundary: wall thickness. Below about 0.4 mm, thin aluminum walls distort during support removal and are hard to inspect. If a rib needs to be that thin, redesign it as a truss instead of a plate.

• 1
  Machine itBearing bores, pivot bores, threads, seal faces, dowel holes.
• 2
  Leave it printedInternal channels, hollow ribs, pockets, non-critical bosses.
• 3
  Watch the wallsKeep aluminum walls at 0.4 mm or thicker where possible.

A launcher that is out of balance shakes the whole assembly. Printing lets you place material exactly where the moment calculation says it should go, but it also lets you place it wrong. Model the part with a nominal balance plane and keep the print symmetric about it.

Fatigue is the second issue. Rotating parts fail at stress concentrations, and printed surfaces have more of them than machined surfaces. A bead-blasted or polished root fillet raises fatigue life noticeably compared with an as-built one.

Inspection is the third. You cannot measure an internal channel with calipers. For launcher work we use raw material certificates, in-process monitoring during the build, and a final inspection before shipment, with reports available on request. That covers the dimensional side.

What nobody can promise is a fatigue number for a geometry that has never been tested. If the launcher will run at high rpm for thousands of cycles, build a test article first and spin it. Additive gives you the shape; it does not replace the test rig.

• 1
  Balance planeDefine it in the model and keep the print symmetric about it.
• 2
  Root filletsBlast or polish them; as-built surfaces start cracks earlier.
• 3
  Test articleSpin-test a printed prototype before committing to a run.