Five parts you shouldn’t manufacture because of LPBF, even if they can be made

An LPBF machine can manufacture almost any geometry, and that's precisely the problem: being able doesn't mean you should. These five part families print without technical difficulty and almost never should. You'll recognize them because they all share one symptom: when it's over, nobody can explain what they've gained.

1. The solid block

What is it? A compact part, without cavities or channels, with thick walls and a large solid volume.

Why not? All that material has to be melted layer by layer, which translates into machine hours, and the accumulated mass is also accumulated stress: it's the kind of part that deforms the most and needs the most anchoring supports.

The sign. If when you look at the part you don't see any interior hollows, it's not a candidate.

What to do. Either redesign by removing what doesn't work —and then it can be interesting— or mechanize it.

2. The part copied exactly as from the milling machine drawing

What is it? A geometry designed for numerical control, with its tool radii, flat faces, and uniform thicknesses, which someone sends to be printed without touching anything.

Why not? It's the most expensive and most frequent mistake. That part doesn't take advantage of anything additive manufacturing can do, and instead suffers all its drawbacks: slower, more expensive, and with a worse finish than the original.

The signal. Nobody has modified the model before sending it.

What to do. Truly redesign it—integrate adjacent parts, remove material, add conduits where there were previously straight holes—or leave it on the milling machine. Manufacturing the same geometry using a different process isn't an additive manufacturing project; it's an expensive change of supplier.

3. The part with internal channels that cannot be emptied

What is it? Interior ducts with recesses, changes in section, or blind ends where unmelted powder remains inside.

Why not? After build is finished, that powder is still there and there's no way to remove it. In a hydraulic or cooling system, it ends up migrating into the system. In a part that's going to vibrate, it makes noise. And it's not a defect that can be fixed later.

The sign. If you can't explain where the powder is coming from, it's not coming from.

What to do. Design the pipe with a continuous slope towards an outlet, without dead ends, and include drain holes even if they need to be plugged later. This is a design decision, not a manufacturing one.

4. The large, simple part

What is it? A support, a frame, a half-meter flange. Simple geometry, large size.

Why not? Production volume is limited and expensive per liter. A single part that fills the build plate monopolizes the machine for days to produce something that a folding machine, a laser cutter, and two welds can accomplish in an afternoon.

The sign. It has to be broken into parts to fit.

What to do. Conventional manufacturing, or —if additive manufacturing is really necessary— processes designed for size, such as DED or WAAM, which deliver material much faster at the expense of detail.

5. The entire part to be machined

What is it? A component in which virtually all surfaces have tight tolerances or a fine finish.

Why not? It's going to go through the milling machine anyway, so you're paying for two complete processes to get what one would give you. And on top of that, you're starting from a more expensive preform.

The signal. The drawing has tolerance indications in almost all dimensions.

What to do. Machine bar stock. The legitimate exception is when the printed preform saves a lot of expensive material, and in that case the calculation must be done with numbers, not intuition.

A sixth case, just in case: the material that is not weldable

It's not a part shape but a material, but it kills projects just the same. The rule, with no useful exceptions: if the alloy cracks when welded, it will crack when printed.

Brass is a clear example: zinc boils well below the temperature of the melting bath, evaporates during the process, and what remains is no longer brass. The same is true for leaded or sulfur-based free-machining steels: the additives that facilitate machining are precisely what ruin the weld. And most alloys formulated for casting are designed to solidify slowly in a mold, not in milliseconds.

What this technology does seek

Having said all of the above in a positive light, the ideal part for LPBF has four characteristics:

• Internal geometry that cannot be machined —channels, lattices, cavities.
• Integrates what were previously several parts and their joints.
• Short or variable production runs, with no tooling to amortize.
• Only a few surfaces with tolerances, which are machined later.

If your part has three of the four, it's worth calculating. If it has one, probably not.

Why are we telling this story?

It might seem odd that a machine manufacturer would dedicate an article to explaining when not to use their machines. It's the opposite: projects that fail because the wrong part was chosen damage the entire technology sector, and we're the first to feel the pinch. We prefer to say no before manufacturing rather than argue about it afterward.

If you have a part in your head and you don't know which side it falls on, it's exactly the five-minute conversation we do best: bring it to us, or run it through the quote calculator and see what comes out.

—

The complete criteria for deciding between processes can be found in print or machine. The design rules that avoid most of these problems can be found in supports in 3D metal printing. And which alloys are viable and which are not can be found in what materials can be printed in metal.