Machined From Billet or Printed Then Machined: Choosing the Route for a Flight Part
Machine the part from certified billet, bar or plate unless the design needs something only additive manufacturing can deliver, such as internal passages or merging several welded pieces into one. Printing then finish machining can win on complex, low quantity parts, but it carries a heavier qualification, allowables and inspection burden that must be paid before the first flight part ships.
Updated
The short answer
Machine it from billet unless the design gains something only a printer can give: internal passages, several parts merged into one, or a shape that would waste most of the bar. For a typical prismatic bracket, housing or fitting, machining from certified wrought stock is cheaper to qualify, easier to inspect and backed by decades of material data.
The deciding factor is usually the paperwork, not the printer: a printed flight part brings a qualification and inspection burden that a machined part from a known alloy does not. For the process detail, see our additive manufacturing and CNC machining service pages.
Machined from billet vs printed then machined, side by side
| Factor | Machined from wrought billet, bar or plate | Metal AM, then finish machined |
|---|---|---|
| Qualification burden | Material spec plus mill cert; first article on the part | Qualified material process per machine, part production plan, witness testing, preproduction articles on flight work |
| Material allowables | Established handbook data for common alloys and tempers | Must be built for the specific qualified process; wrought data does not transfer |
| Property scatter | Controlled by the mill and the material spec | Sensitive to machine, parameters, powder lot, build position and orientation |
| Surface and fatigue | Machined surfaces with known fatigue behavior | As-built surfaces lower fatigue capability; critical surfaces need machining or polishing |
| Inspection | Dimensional inspection, material certs, surface NDT where specified | All of that, plus CT or other volumetric NDT and witness coupon testing |
| Main lead time drivers | Certified stock availability, programming and fixtures, outside processes | Qualification, build queue, stress relief, HIP and heat treat batches, machining, CT and coupon testing |
| Cost scales with | Material removed, setups, tolerances; setup cost spreads over quantity | Build height and volume, plate nesting, post processing and inspection per part |
| Design freedom | Limited by tool access | Internal channels, lattices, part consolidation |
| Best fit | Prismatic and turned parts, tight tolerances everywhere, steady volume | Complex internal flow paths, multi-piece welded assemblies, low quantity complex parts |
Qualification burden: where most of the difference lives
A machined flight part starts from a material that is already characterized. A printed flight part starts from a process that must be proven first, and that cost lands before the first part ships.
NASA-STD-6030, the NASA standard for additively manufactured spaceflight hardware, shows the scale of it. The standard classes parts A, B or C by consequence of failure, then splits Class A and B into subclasses 1 to 4 by structural demand and AM risk. Our additive hub covers what each class allows. What matters for this comparison is what the standard asks for before production:
- A Qualified Material Process for each machine. The standard requires the process capability to be documented in a configuration controlled QMP record for each AM machine. A second printer is a second qualification, not a spare.
- A Material Property Suite. Test data for that material and condition, used to set allowables, design values and the control limits witness specimens are judged against.
- Witness testing on every independent build. For metals, the standard's minimums include six tensile specimens per build for Class A and B parts and two for Class C, with high cycle fatigue specimens for Classes A1 through B1.
- Witness articles for the hardest cases. Class A1 parts, those with high structural demand and high AM risk, need one witness article evaluated for every six flight parts produced.
- Preproduction articles and a review. A part production plan, preproduction articles and a readiness review come before production, with NASA materials and processes attendance required for Class A1 and A2 parts.
Your program may not invoke NASA-STD-6030, but many AM specifications follow a similar structure. Compare that list with a machined part, where the material cert, the drawing and an AS9102 first article usually carry the qualification.
Material allowables and property scatter
Wrought alloys come with published allowables; printed alloys come with the allowables your qualified process earns. That difference decides whether a stress engineer can close margins on day one.
For common wrought alloys, designers draw on established handbook data such as MMPDS, and mill certs tie each bar or plate to a specification (see material certs and traceability).
Printed metal does not inherit that data. NASA-STD-6030 states that non-AM product forms such as forging, bar and plate are not intended to be used to infer AM performance, even for the same alloy. It also defines a lot-mature property suite as one with data from at least five unique feedstock lots and ten build cycles and heat treat lots. Until a process reaches that point, design values carry more uncertainty.
The reason is scatter. The standard calls AM process sensitive and lists influence factors unique to it, including down-facing surfaces, local feature-driven anisotropy and part geometry. In practice, the same alloy from two machines, two powder lots or two build orientations can test differently, which is why the parameter set gets frozen and witness coupons ride on every build.
Surface finish and fatigue
If a flight part sees cyclic load, plan to machine or polish every fatigue-critical surface, whichever route you choose. On a printed part, that requirement shapes the design.
NASA-STD-6030 is direct about it. It states that AM materials are not considered to have an endurance limit. In its criteria for low structural demand on metal parts, high cycle fatigue stress range may reach 80 percent of the applicable fatigue limit on improved surfaces but only 60 percent on as-built surfaces. Surfaces left as-built must be evaluated with fatigue data from a representative as-built surface, not from polished specimens. In its AM risk scoring, a part whose as-built surface cannot be fully removed on all fatigue-critical surfaces picks up risk points that push it toward a more demanding subclass.
For buyers: leave machining stock wherever fatigue, sealing or fit matters, ask how unmachinable internal channels will be finished, and expect a printed part with mostly critical surfaces to need nearly as many setups as a billet part. Surface treatments are covered on our surface finishing page.
Inspection: what you can see decides what you can fly
A machined part from sound wrought stock is mostly a dimensional and surface inspection job. A printed part adds a volumetric one, because porosity, lack of fusion and trapped powder sit inside.
- Inspectability drives classification. In NASA-STD-6030's AM risk table, a part where not all surfaces and volumes can be reliably inspected, and that cannot be adequately proof tested, scores 5 points by itself, and a summed score of 5 or more puts the part in the high AM risk subclass.
- CT is the usual tool, with limits. ASTM E3166, the guide for NDT of metal AM aerospace parts after build, covers computed tomography, radiography, ultrasonic, eddy current, penetrant, thermography and resonance methods. It does not set accept or reject criteria, so your drawing or spec must state the flaw size and acceptance limits.
If an internal feature cannot be inspected or proof tested, redesign it. Methods and limits are compared on our nondestructive testing page.
Lead time and cost drivers
We do not publish prices or lead times; both depend on the part, the supplier and the qualification path. The drivers differ sharply.
Machined from billet
- Availability of certified bar, plate or forging in the right alloy, temper and size
- Material removed: a part that throws away most of the billet costs material and spindle time
- Number of setups, tight tolerances and fixtures; setup cost spreads as quantity rises
Printed then finish machined
- Whether a qualified process already exists for that alloy on that machine; if not, qualification comes first
- Build height and how many parts nest on one plate, since powder bed time scales with height
- Stress relief, HIP and heat treat, often batch processes at outside vendors
- Finish machining of critical surfaces, often awkward to fixture
- CT time and witness coupon testing on each build
Machining setup cost falls per part as volume rises; AM cost per part falls less because build time, post processing and inspection repeat. AM tends to win where geometry would otherwise force a welded or brazed assembly. NASA reported that 3D printing the pogo accumulator for the RS-25 engine eliminated more than 100 welds and cut cost by nearly 35 percent and production time by more than 80 percent for that part. That is a consolidation result, not a rule for single machined parts. For machining cost drivers in more depth, see aerospace machining cost.
Design freedom and part consolidation
AM pays off when the part is designed for it. Printing a part drawn for a mill rarely saves money.
The best known example is GE's fuel nozzle tip for the LEAP engine, which GE says went from about 20 pieces welded and brazed together to one printed piece, with weight cut by about 25 percent, and in production since 2015. The gains came from removing joints and their inspections and from internal geometry that could not be machined. Each weld or braze that disappears is one less joint to inspect and one less place for a leak or crack.
Choose machining when, choose AM when
Machine from billet when
- The part is prismatic or turned and a 5-axis mill can reach every feature
- Most surfaces are critical for fit, sealing or fatigue
- No qualified AM process exists for the alloy and schedule is tight
- Quantity is steady or rising and setup cost can spread
Print, then finish machine, when
- The part needs internal channels, cooling passages or flow paths that cannot be drilled
- It replaces a welded or brazed assembly of several details
- Machining would remove most of a large, expensive billet
- A qualified process for the alloy already exists at a supplier, or the program can fund one
| Typical candidate | Usual route |
|---|---|
| Brackets, clevises, fittings, bulkheads | Machined, unless topology optimized for mass |
| Valve bodies with drilled ports | Machined |
| Injectors, manifolds, heat exchangers | Printed, then finish machined |
| Multi-piece welded ducts and accumulators | Candidate for printing and consolidation |
| Fasteners and threaded flight parts | Machined; NASA-STD-6030 excludes fasteners from Class A and B |
| GSE brackets, fixtures, tooling | Either; AM is often faster for one-offs |
For alloy-specific notes, see the metal additive manufacturing library page and titanium alloys.
Get both routes looked at before you commit
If you are not sure which route fits, send the model, the drawing, the part class or criticality and the quantity through the additive manufacturing quote form. We match the part to suppliers for the route that fits and will tell you if machining looks like the better answer. A person replies within 1 business day, and every order ships with a certificate of conformance plus the build and inspection records your request requires. Controlled models move only through our controlled programs process, never the form; see confidentiality.
Send the job. A person replies within 1 business day.
Part, quantity and date are enough to start. Drawings and the mutual NDA follow by reply, and controlled programs never send technical data through this site.
We have it.
Reference . A person replies within 1 business day. Watch for our email from [email protected].
Questions
Is a 3D printed metal part as strong as a machined part?
Static strength can come close when the process is qualified and the part is HIPed and heat treated, but you cannot assume it. Printed material has its own allowables, more scatter between machines and builds, and weaker fatigue performance on as-built surfaces. NASA-STD-6030 states that AM materials are not considered to have an endurance limit. Judge a printed part on witness coupon data from its own qualified process, not on wrought handbook values.
When is additive manufacturing cheaper than machining for flight parts?
Usually when it replaces an assembly or impossible geometry, not a single simple part. Printing wins when it removes welds, brazes and assembly steps, or when machining would waste most of a large billet. NASA reported that printing the RS-25 pogo accumulator removed more than 100 welds and cut its cost by nearly 35 percent. For a simple bracket in steady quantity, machining from certified stock is normally cheaper once qualification and inspection are counted.
Why do printed parts still need machining?
As-built metal surfaces are rough and dimensionally loose compared with machined ones, and they reduce fatigue capability. Sealing faces, bores, threads, datum features and fatigue-critical surfaces are almost always machined after printing. NASA-STD-6030 also bars printed threads on Class A and B parts. Plan machining stock into the design and expect finish machining setups on most flight parts, which is why the route is printed then machined.
Can I use MMPDS or wrought data to design a printed part?
Not for flight margins. NASA-STD-6030 states that non-AM product forms such as forgings, bar and plate are not intended to infer AM performance, even in the same alloy. Allowables and design values must come from a material property suite built on the qualified AM process. The standard treats a suite as lot-mature only with data from at least five feedstock lots and ten build and heat treat cycles.
Does a printed part need CT scanning?
Flight parts with internal features usually do, because porosity, lack of fusion and trapped powder are inside the part. ASTM E3166 guides NDT of metal AM aerospace parts, including CT, but it sets no accept or reject criteria, so your spec must define flaw size and limits. Under NASA-STD-6030, a part that cannot be reliably inspected or proof tested is automatically placed in the high AM risk subclass.
Which parts are the best candidates for metal AM?
Injectors, manifolds, heat exchangers, cooling channels, ducts and accumulators that are welded or brazed from several pieces today, plus topology optimized brackets where mass matters. GE's LEAP fuel nozzle tip, which GE says went from about 20 pieces to one, is the standard example. Simple prismatic parts, threaded parts and fasteners are usually better machined.
Related
Sources
- NASA-STD-6030 Additive Manufacturing Requirements for Spaceflight Systems (approved baseline PDF)
- NASA Technical Standards: NASA-STD-6030 status page
- ASTM E3166: Nondestructive Examination of Metal Additively Manufactured Aerospace Parts After Build
- NASA: NASA tests 3-D printed rocket part to reduce future SLS engine costs (RS-25 pogo accumulator)
- GE Aerospace: manufacturing milestone, 30,000 additive fuel nozzles
- ASTM F2924: Ti-6Al-4V with powder bed fusion
- ASTM F3301: thermal post processing of powder bed fusion metal parts
Ready to
source it?
A person replies within 1 business day. NDA first.
Request a quote