Printed Lattices, Machined Housings: Both Halves of a Medical Device

The first electron-beam-melted hip cup was implanted in Italy in 2007, and it looked neither machined, cast nor forged. Its bone-facing side was a tangle of struts — deliberately rough, deliberately porous — because its designers wanted bone to grow into the implant, not merely sit against it. Nineteen years on, a lattice face is simply how a modern press-fit acetabular cup is designed.

Why bone wants a lattice, not a polish

Osseointegration is mechanical before it is biological. Bone grows into interconnected pores most reliably in the 500–700 µm range, and it keeps that bond only if the implant lets it carry load instead of shielding it. A machined titanium surface, however clean, offers nothing to grip. Plasma-sprayed porous coatings help, but a coating is an added layer with an interface that can degrade. EBM reframed the question: instead of coating a solid cup, print the cup so its outer millimeters are the porous structure.

Stiffness matters as much as texture. Solid Ti-6Al-4V sits around 110 GPa in elastic modulus while cortical bone lives near 10–20 GPa, and bone that stops carrying load resorbs. An engineered lattice lets the designer tune effective stiffness strut by strut, easing the mismatch at exactly the interface where bone is supposed to keep working.

The 2007 cup that made additive orthopedic-grade

Lima's Delta-TT acetabular cup earned its CE mark and first implantation in Italy in 2007, built on Arcam electron-beam-melting machines from Ti-6Al-4V ELI powder. EBM suits implants unusually well: the build chamber runs at several hundred degrees Celsius, so parts emerge with minimal residual stress, and the vacuum environment keeps reactive titanium clean. By the late 2010s, more than 100,000 patients were carrying implants printed on those machines, tens of thousands of them acetabular cups with printed trabecular structures, and follow-up series past seven years show the fixation holding.

The design move that matters: the cup is one part with two jobs. The load-bearing shell is printed to near-full density and finished by machining wherever geometry demands precision — the rim, the taper, the locking features. The bone-facing surface is an engineered lattice, printed in the same build, with no bonded interface anywhere. Coating delamination, a classic failure mode, is designed out because there is no coating.

The unglamorous other half: housings, handles and the ISO 13485 mindset

Most medical hardware is not an implant. For every printed cup there is a cart-mounted diagnostic instrument, a benchtop analyzer, a surgical handle, a fluidics manifold — parts that never touch a patient's bloodstream but still live inside a quality system. This is overwhelmingly CNC territory, and what distinguishes medical machining is not tighter tolerances; it is discipline: material certificates traced to the melt, first-article inspection against the drawing, and no silent process changes. A supplier can hold ±0.01 mm and still be unusable for medical work if they swap a resin or an alloy without telling you.

Machining PEEK without ruining it

PEEK earns its place in spinal cages, trauma trials and instrument components because it is radiolucent and its stiffness — around 3.6 GPa unfilled — sits far closer to bone than any metal. It machines like a polymer with a grudge: internal stresses in the extruded stock release as you cut, so precision parts get rough-machined, stress-relief annealed, then finished. Keep tools sharp, keep heat down, and for implant-grade work control the coolant, because contamination you machine in does not wash out.

Device-grade parts without a device giant's supply chain

If you are a diagnostics startup in Hsinchu rather than an orthopedics multinational, the practical question is what to demand from suppliers. You do not need your machine shop to hold your regulatory clearance — you need it to behave like part of your quality system: certs with every lot, declared subcontractors, first-article reports, and tolerances quoted only where a dimension is actually critical. Blanket ±0.05 mm across a housing drawing can easily add a third to the machining cost and buys you nothing; call out the six dimensions that matter.

The split to internalize: the implant itself will be printed and finished by a certified orthopedic manufacturer, but everything around it — surgical trials, instrument handles, cases and trays, the analyzer enclosure — is conventional manufacturing under good paperwork, and Taiwan's machining base is exceptionally strong at exactly that. Reserve the exotic-process budget for the part that touches bone; spend discipline, not novelty, on the rest.

The real lesson of the 2007 cup is that additive won by doing what machining could not — growing a functional lattice as part of the implant — while everything else in the device stayed machined for the reasons it always was: precision, traceability and cost. When the unglamorous half of your device needs quoting, CNC machining with full material traceability is work we handle every week.