3D Printed Aerospike Rocket Engine: How the Spike Works

• 1
  How a 3D printed aerospike rocket engine adjusts to altitude
• 2
  Why additive manufacturing unlocked the spike
• 3
  Material choice for a spike that runs hot
• 4
  Where CNC machining still fits the aerospike program
• 5
  How to qualify a printed spike before it sees fire
• 6
  Aerospike vs bell nozzle: when each one wins

A bell nozzle is a fixed tube. Designers size the exit area for one altitude, usually a compromise between sea level and vacuum. Below that altitude the flow over-expands and separates from the wall. Above it the flow under-expands and you leave thrust behind.

An aerospike flips that arrangement. Instead of a closed skirt, the exhaust expands against an open spike and against the surrounding atmosphere. Ambient pressure becomes part of the nozzle. At sea level the atmosphere squeezes the plume inward. As the vehicle climbs, the plume spreads outward on its own.

That self-adjusting behavior is the whole point. The spike does not move. The pressure field around it changes. Engineers call this altitude compensation, and it matters most on a first stage that has to work from the pad to separation.

The trade-off is thermal. A bell carries its heat in a wide, continuous wall that is easy to cool with channels. A spike is a thin, exposed body sitting directly in the hot gas path. Every square millimeter of its surface sees the full flame.

• 1
  Bell nozzleOne design altitude, simple cooling, well understood
• 2
  AerospikeAltitude compensation, harder cooling, harder to make
• 3
  Ambient pressureActs as the outer nozzle wall on a spike

A regeneratively cooled spike needs internal coolant channels that follow a curved, tapering body. Machining those channels into a solid cone means drilling from the outside and plugging the holes, or splitting the part and brazing it back together. Both routes add joints.

Powder bed fusion builds the spike and its channels in one piece. The laser melts each layer, so a channel can turn, branch and narrow without a tool ever reaching inside. That is the real reason a 3D printed aerospike rocket engine became practical rather than theoretical.

The geometry still has limits. Overhanging surfaces below roughly 45 degrees from the build plate need support, and support inside a coolant channel is nearly impossible to remove. Designers rotate the part, self-support the channel roofs, or accept a rougher wall.

Surface finish inside those channels is the second constraint. As-built walls land around Ra 8–15 μm. That roughness raises pressure drop and can trigger local hot spots. Honing or abrasive flow finishing brings the wall toward Ra 1.6–3.2 μm.

• 1
  One-piece channelsNo drilled plugs, no brazed joints in the hot path
• 2
  Self-supportKeep overhangs above about 45 degrees
• 3
  Internal finishAs-built Ra 8–15 μm, finished near Ra 1.6–3.2 μm

Copper alloys such as GRCop-42 and CuCrZr conduct heat far better than nickel alloys, which is why they appear in combustion chamber liners. The catch is that copper is soft and creeps. A thin copper spike wall can deform under its own thermal load.

Nickel alloys like Inconel 718 hold strength to roughly 700 °C and print cleanly. They conduct heat poorly, so the coolant has to work harder and the wall has to be thinner. Thin Inconel walls are where print defects show up first.

Some designs split the job. A copper liner carries the heat away, and a nickel or steel jacket carries the structural load. That means two materials and a joint, so the manufacturing plan gets longer and the inspection plan gets stricter.

For a subscale demonstrator, many teams start with a single nickel alloy. It is more forgiving to print and easier to inspect. Move to a copper liner only when the chamber pressure and run time demand it.

• 1
  Copper alloysBest conductivity, lower creep strength
• 2
  Inconel 718High strength, poor conductivity, thin walls
• 3
  BimetallicBetter performance, more joining and inspection

Additive does not replace machining on a rocket engine. It feeds it. The injector face, the flange bolt circles, the seal grooves and the mounting interfaces all need tight tolerances that no printer holds as-built.

A printed spike typically arrives with 0.5–1 mm of stock on critical faces. The first operation establishes a datum from the chamber axis, then the flange faces and bolt holes come in on a 4-axis or 5-axis mill. Concentricity between the throat and the mounting flange is the number that decides whether the nozzle sits straight.

On our 16 simultaneous 5-axis centers we hold ±0.005 mm on turned and milled features, with finishes to Ra 0.2–0.8 μm on seal faces. That is the range where a metal O-ring or a copper gasket actually seals.

Small demonstrator parts also run well as fully machined hardware. A regeneratively cooled spike with simple straight channels can be milled in two halves from 6061 or 316L and joined, which skips the print queue entirely.

• 1
  Printed near-net0.5–1 mm stock on sealing and mounting faces
• 2
  Datum firstEstablish chamber axis before any flange work
• 3
  Critical numberThroat-to-flange concentricity

You cannot cut a printed spike open and still fly it, so the inspection plan has to be non-destructive. Computed tomography catches internal porosity, un-melted powder and channel blockages that a visual check misses entirely.

Flow testing is the second gate. Push water or air through the cooling channels and record pressure drop against flow rate. A channel that is half blocked shows up as a higher pressure drop at the same flow, long before it becomes a burn-through.

Wall thickness on the spike tip deserves its own measurement. The tip runs hottest and has the least material. Ultrasonic gauging or a CT slice through the tip tells you whether the print held the nominal wall.

Finally, pressure test the jacket. A hydrostatic test at 1.5 times the design coolant pressure finds leaks at joints and fittings while the part is still on the bench. Once it is on the test stand, a leak becomes a fire.

• 1
  CT scanPorosity, trapped powder, channel blockage
• 2
  Flow testPressure drop vs flow rate per channel
• 3
  Tip wallUltrasonic or CT slice at the hottest point
• 4
  Hydrostatic1.5× design coolant pressure before hot fire