Why Carbon Won the Airframe

The number that decides an airframe material is specific stiffness, not strength. A drone arm does not usually break; it flexes, and flex between the motor and the flight controller feeds vibration straight into the IMU, which the estimator reads as motion that never happened. Stiff arms make a stable aircraft.

Unidirectional carbon runs around 135 GPa (19.6 Msi) of modulus at a density of 0.056 lb/in3, against 6061 aluminum at 69 GPa and 0.098 lb/in3. Even a quasi-isotropic laminate, with fibers spread across four directions and an effective modulus near 50 GPa, still delivers roughly three times the stiffness per pound of aluminum. Carbon also damps vibration better than metal, which shows up directly in cleaner accelerometer data.

That is the whole case, and it is a strong one. The problems start when a team treats carbon as aluminum that weighs less.

Trap One: It Conducts Electricity

Carbon fiber is a conductor. Roughly a hundred to a thousand times more resistive than aluminum, but nowhere near an insulator, and this catches nearly every first-time team.

An antenna on or near a carbon plate is detuned and shadowed. GNSS reception under a carbon canopy degrades badly, video links go directional in ways nobody planned, and a receiver that worked on a foam test rig loses half its range on the production frame. The fix is planning, not shielding: keep antennas off the composite, put GNSS on a mast or a dedicated ground plane, and treat the frame as an RF obstacle from the first layout, using the discipline in antenna design for wireless products.

Conductivity also means bare carbon against bare aluminum in humid or salty air sets up a galvanic couple, with the aluminum corroding. Isolate with a glass-fiber ply, a coated interface, or a nylon shoulder washer. And a loose strand of carbon dust will short a board, which is why frames get cleaned and sealed before electronics go in.

Trap Two: It Does Not Bend, It Breaks

Aluminum yields. It deforms, warns you, and often keeps the aircraft flyable. Carbon is elastic right up to failure and then it is gone, usually splintering. Design consequences follow.

Impact and notch sensitivity are the practical issues. A hard landing that would dent an aluminum arm can leave a carbon arm with barely visible internal delamination that halves its strength. There is no external sign, and the arm fails three flights later in normal use. Treat arms as consumables with a replacement interval, not lifetime parts.

Holes are the other issue. Fasteners load a laminate in bearing, the weakest way to load it, and the drilled hole cuts fibers rather than pushing them aside. Bolted joints need generous edge distance, oversized washers, and controlled torque, and they are where cyclic loading eventually shows up, along the lines described in material fatigue in product design. Anisotropy also breaks intuition, so simulate before committing: an isotropic estimate of a laminate is simply wrong, and the case for running the analysis is made in FEA simulation in product design.

Trap Three: Cutting It Is a Process, Not a Task

Carbon dust is a conductive, abrasive respiratory irritant, and cutting or sanding it in a shared shop without extraction contaminates the space and everything electrical in it. Machining needs diamond-coated or carbide tooling, high speed with low feed, and dedicated dust collection with PPE.

Drilling is the specific hazard: exit-side delamination is standard unless you use a backing plate and a specialized bit. Abrasive waterjet is the cleanest way to cut plate, since it produces no heat-affected zone and no airborne dust, and the tradeoffs against laser and router are covered in waterjet cutting. For a prototype run, most teams simply buy cut plate from a supplier who already owns the extraction equipment.

How Frames Are Actually Made

Four methods cover essentially every drone frame, and they separate by volume.

  • Plate and tube assembly. CNC-cut flat plate for the center section, pultruded round or square tube for arms, aluminum clamps to join them. Tooling cost is zero, iteration is a day, and a full frame set runs $150 to $600. Nearly every prototype and most sub-100 unit products live here.
  • Wet layup. Hand-laid fabric and room-temperature resin over a mold. Cheap tooling, high labor, variable fiber volume fraction near 45 percent, and part-to-part variation that makes weight targets hard to hold. Good for one-off shells, poor for production.
  • Prepreg in a heated mold. Pre-impregnated fabric cured under vacuum or in an autoclave. Fiber volume near 60 percent, excellent repeatability, and the only route to a true monocoque shell with integrated features. Tooling runs $8,000 to $40,000 per part, so per-part cost stays high until volume justifies it.
  • Compression-molded chopped fiber. Short fibers in resin pressed in a matched steel tool. Fast cycles and complex geometry, at maybe half the stiffness of continuous fiber. Sensible for brackets and motor mounts at thousands of units, not for primary structure.

Weight-critical programs feel this most sharply at both extremes. A design squeezing under the threshold in sub-250g drone design may need molded parts to hit the number, while a heavy-lift cargo drone needs large tube sections whose buckling behavior has to be checked, not assumed.

Not Everything Should Be Carbon

Carbon everywhere is a beginner's signature: expensive, hard to repair, and wrong for several roles. Landing gear should absorb energy, so glass fiber or springy nylon does it better and cheaper. Motor mounts see concentrated bolt loads and heat, so machined aluminum wins. Covers and battery trays should be molded or printed. And anything under an antenna should be glass fiber or plastic on purpose.

A good airframe uses carbon for the load path and something else for everything else, which is exactly the mixed-material reasoning in how to choose materials for a new product.

Joining: Bonded, Bolted, or Both

Adhesive bonding beats bolting because it spreads load into the laminate instead of concentrating it at a hole. A structural epoxy or toughened methacrylate on a properly prepared surface, peel-ply removed or abraded and solvent-wiped immediately before bonding, with a controlled 0.005 to 0.015 in bond line, is stronger than the plies around it, and it seals the joint against moisture.

The cost is disassembly, which is why most frames are hybrids: bonded where structure is permanent, clamped where a crashed arm has to come off in the field with a hex key. Never run a self-tapping screw into a laminate or rely on threads cut into carbon. Use through-bolts with backing plates or bonded metal inserts.

Get the Structure Right Before You Cut Fiber

Projects House designs and sources composite airframes: layup schedule and ply orientation, joint and insert design, RF-aware layout, and the choice between plate assembly and molded structure at your volume. Send your payload and target weight through our contact form.