Additive manufacturing in production: when a printed part belongs on a real bill of materials

Prototype economics do not carry across

In a prototype shop the machine sits idle between jobs and turnaround is the only cost anyone tracks. In production the platform has to be filled, builds have to be planned, and cost per part depends on how well parts nest and how long the build runs. Unlike machining or moulding, unit cost barely falls as volume rises, so the crossover against a tooled process arrives sooner than people expect. Model it honestly using build time, material, gas or resin consumption, an allowance for failed builds, and the post-processing hours that follow, rather than the price a bureau charged for a single sample.

Repeatability is a machine, feedstock and parameter question

A printed part is the output of a machine state, a feedstock lot and a parameter set, and all three drift. Powder reused across builds shifts in size distribution and chemistry. Parameters that suit one machine are not automatically valid on its sibling. Orientation on the platform changes surface finish, support requirement and mechanical properties, so orientation itself belongs in the specification and has to be frozen. Production use therefore means parameters locked under change control, witness coupons within builds to evidence material behaviour, and a record of machine, lot and platform position against every serialised part that leaves.

Post-processing is most of the labour

The build is the beginning. Parts come off a platform needing depowdering or resin removal, stress relief for many metal processes, separation from the plate, support removal that is often manual and skilled, and machining of any feature carrying a real tolerance or a sealing duty. Internal channels are the awkward case, since proving a powder-filled passage is clear can be harder than printing it. Surface finish rarely meets a drawing without secondary work. Cost the post-processing route before committing to additive, because it drives labour content, floor space and frequently whether the part is economic at all.

Feedstock handling is a regime, not a task

Feedstock brings a handling regime with it. Fine metal powders can be reactive and present inhalation risk and, for some alloys, fire and explosion risk, which drives inert handling, controlled vacuuming, earthing and restricted areas. Photopolymer resins bring skin sensitisation and waste disposal duties. Powder needs storage conditions, moisture control and lot traceability, and reused powder needs a documented rule for how far it may be recycled and how it is tested. Whoever writes the safety case for the room must understand the specific material, since controls for a polymer machine and a reactive metal machine are not interchangeable.

Where a printed part genuinely wins

Additive wins where geometry rather than volume is the constraint. Consolidating a fabricated assembly into one printed component removes joints, fasteners and inspection points. Internal cooling passages that cannot be drilled make tooling inserts perform better. Low-volume production, where tooling would never amortise, and obsolete spares whose original tooling disappeared long ago, are the two commercial cases that recur across industries. It loses where a part is simple, demand is steady and a mould or fixture already exists. Judging case by case, with post-processing costed, avoids both fashionable adoption and blanket refusal.

Frequently asked questions

How do we qualify a printed part for a production assembly?

Treat it as a special process. Fix the machine, the material specification, the parameter set and the build orientation, then demonstrate that parts made under those conditions meet the drawing and the functional requirement, including any property that depends on build direction. Include witness coupons in each build to evidence material behaviour, define what a failed coupon does to that build, and place the parameter set under change control so a service visit or software update cannot silently invalidate the qualification.

Is additive cheaper than machining for small batches?

Sometimes, and the comparison has to run to the end of the route. Additive avoids fixtures, tooling and setup, which is where machining loses money at small quantities. It then adds depowdering, stress relief, support removal, finishing and machining of any toleranced feature, which is where the advantage flows back. Complex parts with few critical surfaces favour printing. Simple parts carrying several tight features usually do not, since they end up on a machine tool regardless.

Can we replace an obsolete spare by scanning and printing it?

Scanning gives you geometry, not specification. You still need the material, the heat treatment, the surface requirement and an understanding of what the part does in service, and a worn part scanned as found reproduces its own wear. For non-critical items carrying no load the route is practical and widely used. For anything bearing load, sealing pressure or affecting safety, treat it as a new part needing engineering approval, and expect that approval to take longer than the build.