What It Means When the 3D Printing Sector Grew 2% on November 4

• 1
  Why a 2% Day Does Not Change Your Process Plan
• 2
  How Additive Builds a Part, Layer by Layer
• 3
  Where Additive Holds Tolerance and Where It Drifts
• 4
  Material Choice Sets the Real Boundary
• 5
  What Volume Does to the Economics
• 6
  How to Inspect a Printed Part Without Fooling Yourself
• 7
  Additive vs Subtractive: Matching Process to Feature

Publicly traded companies in the 3D printing sector grew 2% on November 4. That number is an equity move. It reflects what buyers of stock expect about future revenue, not what happened on any shop floor that afternoon. If you are sourcing parts, the index is not a signal. Your geometry, volume and tolerance are.

Sector indices bundle hardware makers, service bureaus and software vendors into one ticker. A 2% move can come from one company's earnings call, a rate decision, or a supplier announcement that never touches the parts you buy. The underlying demand for printed parts moves much slower than the quote does.

What does change your plan is where the process boundary sits. Additive wins on internal channels, lattice, and low counts. Subtractive wins on tight fits, hard materials, and anything that has to survive a million cycles. A green day on the index tells you nothing about which side of that line your part falls on.

So treat the headline as context, not instruction. The rest of this page is about the mechanics: how a printed layer forms, why that limits tolerance, and how to decide when to stop printing and start cutting.

Every additive process does the same thing in different physics. A slicer cuts the CAD solid into horizontal layers, typically 0.05–0.2 mm thick for polymer extrusion, 0.03–0.06 mm for resin, and 0.02–0.06 mm for metal powder bed. The machine then deposits or fuses material inside each layer's contour.

That layered construction is the whole story of the process. Bond strength between layers is lower than strength within a layer, because the joint depends on partial re-melting or adhesion rather than a continuous crystal structure. Printed parts are anisotropic. A bracket loaded across the layer plane can fail well below the bulk material's datasheet strength.

Shrinkage is the second constraint. Metal powder bed parts shrink 3–20% depending on alloy and geometry as the melt pool solidifies. Resin shrinks during cure. FDM polymer shrinks as it cools. Machine software compensates with a global scale factor, but thin walls and thick bosses cool at different rates, so the compensation is never exact everywhere.

Layer thickness also sets your surface finish. A 0.2 mm layer leaves visible stair steps on any sloped or curved face. You can sand, tumble, or bead blast them down, but each of those operations costs time and rounds corners. A machined face at Ra 0.8–1.6 μm needs no such follow-up.

Dimensional accuracy on a well-tuned polymer printer runs roughly ±0.2 mm on small parts, and worse as the part grows. Industrial resin and metal systems can reach ±0.05–0.1 mm on a good day and a controlled build chamber. That is fine for a housing that bolts to a bracket with slotted holes. It is not fine for a bearing bore.

The drift is not random. It follows geometry. Long thin features warp. Large flat faces curl at the edges. Holes print undersize because the melt pool or resin creep closes the top of the circle. Vertical walls that should be 3.00 mm often land at 3.05–3.10 mm.

Post-processing changes dimensions too. Support removal leaves witness marks. Heat treat for metal parts relieves stress and moves the part again. If a printed metal part is HIPed or sintered, plan for a second measurement pass, not a first-article sign-off.

The practical rule: use additive where the feature has clearance built in, and use subtractive where two surfaces have to mate to a number. If your drawing calls out ±0.005 mm, no printing process will sign that print.

Polymers cover most prototyping. ABS, PC, PMMA, POM, PA, PEEK, PP and HDPE all print in some form, and PA with carbon fibre filler gives a stiff, light part for jigs and covers. These are the same families we machine every day, so a printed prototype and a machined version can share a material name even when the mechanical properties differ.

Metals narrow the field. Titanium Ti-6Al-4V, Inconel, 316L stainless and aluminium alloys print well in powder bed. Printed Ti-6Al-4V reaches high strength after heat treat, but fatigue life stays below wrought bar stock unless you HIP and machine the critical surfaces. For anything rotating or pressure-holding, that gap matters.

There is a hybrid path that gets used more than people expect. Print the near-net shape, then machine the datums, bores and sealing faces. You skip most of the roughing stock and still hold ±0.005 mm where the drawing demands it.

If the part needs anodizing, hardcoat, electroless nickel or laser marking, check that the printed substrate takes the finish. Porosity in printed metal can trap plating solution and bleed out later. Machined 6061-T6 or 17-4PH does not have that problem.

Printing has almost no fixed cost. No fixture, no program proving, no first-article loop beyond a dimensional check. That makes one part cheap and ten parts cheaper per unit than ten machined parts, because the machine time scales linearly.

CNC inverts that curve. Setup and programming are fixed, so part one costs the most and part one thousand costs the least. The crossover for a simple bracket often sits somewhere between 30 and 200 pieces. For a complex part with five setups, it can sit higher. For a part with one setup and loose tolerance, it can sit lower.

The other variable is machine time. A printed part that takes 14 hours in the chamber occupies that chamber for 14 hours no matter how many you order. A machined part that runs 4 minutes on a 5-axis center can ship thousands per week. When the order grows, cycle time becomes the constraint, not setup cost.

Our rule of thumb for customers: print to prove the fit, machine to prove the function. Once the design is frozen and the volume is real, move it to the mill.

Calipers are a poor tool for printed parts. They touch the high points of a rough surface and read optimistically. Use a micrometer on flat faces or a optical comparator on profiles, and always measure after support removal and any heat treat, not before.

Check the features that carry load. Layer boundaries matter most on parts in tension or bending. A simple visual check with a low-power loupe will show delamination, incomplete fusion, or resin pockets that will crack under vibration.

For anything headed into a regulated program, keep the powder lot, build orientation, and post-processing record. Build orientation alone changes tensile strength by 20–30% in some polymers, so a part that passed in one orientation may fail in another.

If the part will be machined after printing, do the first-article inspection on the finished geometry. The printed blank's dimensions are irrelevant once the datums are cut.