Technical article

CNC Machining vs 3D Printing: Which Process Actually Fits Your Part? A Scenario Guide

Why There's No Single "Best" Manufacturing Process

In our Q1 2025 quality audit, I rejected 14% of first deliveries. Not because the parts were defective — because they were made with the wrong process for the job. One supplier shipped resin-printed fixtures when the spec called for load-bearing aluminum brackets. Another quoted us 5-axis machining time for a part a 3-axis mill could've handled at half the cost.

Both mistakes had the same root cause: someone assumed there's one "right" way to make a part. There isn't. What you choose depends on what the part actually needs to do. That's the whole game.

Four scenarios. Different answers for each. (And yes, I know "it depends" is the least satisfying answer in manufacturing. It's also the honest one — so let me be specific instead of vague.)

  • Scenario 1: Production metal parts with tight tolerances
  • Scenario 2: Medical or highly custom small-batch parts
  • Scenario 3: Full-color visual or presentation models
  • Scenario 4: Early-stage FDM prototypes in plastic

Scenario 1: You Need Production Metal Parts With Real Tolerances

This is where CNC earns its place. If the part bears load, sees heat, or has to mate with another machined component, you're looking at subtractive manufacturing — full stop. No print is replacing a structural bracket in a drivetrain.

The decision inside this scenario is 3-axis vs. 5-axis. Plenty of shops will quote you a dmg mori 5 axis cnc machine for everything because that's what's on their floor and that's what they want to keep busy. Don't let that happen without asking why. Five-axis earns its cost when:

  • The part has features on multiple faces — complex brackets, impellers, medical implant housings
  • You need to minimize setups, because every setup is another chance for tolerance to stack
  • The geometry genuinely can't be reached with a 3-axis approach

For everything else, a well-run 3-axis job is faster and cheaper. Repeat work on a cnc dmg mori turning center or a simple vertical mill, done right, beats an over-specced 5-axis job every time. The machine doesn't make the part — the person running it and the process around it does.

What I actually check in a dmg-mori delivery: setup documentation, in-process inspection records, and whether the first-article inspection matches the drawing — not just whether everything landed inside the tolerance band. A part that's "within tolerance" but drifts across the full band from unit 1 to unit 40 isn't in control. It's just lucky. Lucky doesn't repeat on the next order.

If a supplier can't tell you which setups they used and where they measured, the tolerance claim is a guess with a number attached.

One more spec review tip: ask for the measurement method, not just the result. CMM, hand gauge, optical — each has different uncertainty. A tight tolerance verified by the wrong method is theater.

Scenario 2: You Need Medical Parts in Small Batches

Here's where people get excited about medical 3d printing service providers and start comparing printer resolutions. I get it. Resolution sounds like a spec you can compare. But that's the wrong filter.

The first thing I check with any additive medical supplier isn't the machine — it's traceability. Material certificates. Logged process parameters. Documented post-processing steps. If they can't show you a lot number tied to your part, walk. That's the difference between a medical service and a hobbyist with a resin tank and a website.

Never expected it, but the cheapest per-part quote is usually the most expensive decision here. Turns out the cost of a failed validation run dwarfs whatever you saved on the print itself. I had one vendor quote us 40% below the next bid; they couldn't produce a material cert for the biocompatible resin. We passed. That "saving" would've cost us a re-validation cycle measured in weeks, not dollars.

Small batches matter here too, and this is where I get stubborn. A 5-part trial run for a new surgical guide still deserves a proper spec review — not a "we'll just print it and see" shrug. The vendors who treat your 5-part order like a 5,000-part order are the ones you'll still be using when you scale. (Note to self: I should formalize a vendor scorecard for this — the spreadsheet's getting unwieldy.)

Scenario 3: You Need Full-Color Visual or Presentation Models

For client-facing models, trade show pieces, or design review props, a full color 3d resin printer is the right tool. The color fidelity is genuinely impressive now. Photorealistic, if the file's set up well and the operator knows what they're doing.

Here's the part nobody warns you about: those beautiful parts are fragile. The surprise wasn't the color quality — it was how badly a "professional-grade" print handled a basic drop test. Resin, by nature, is brittle. Full-color resin even more so, because the dye load reduces the toughness of the cured polymer. If your model is going to be handled, passed around, or shipped, budget for protective cases or accept that a replacement is one clumsy handshake away.

Cost note: the jump from standard resin to a tougher engineering resin is real, but it's usually worth it if the part has to survive the show floor. On a 20-piece run, the upcharge per piece is modest against the cost of a broken demo in front of a prospective client. Presentation value collapses fast when the presentation breaks.

Scenario 4: You're Choosing FDM Material for Early Prototypes

The petg vs pla 3d printing question comes up constantly, and the default answer — "just use PLA, it's easier" — is often wrong.

PLA is easy to print, dimensionally stable, and cheap. It's also brittle and softens at relatively low temperatures. Fine for a form check. Not fine for a prototype that'll be handled, dropped, or left in a hot car during a field test.

PETG is tougher, handles heat better, and its layers bond more durably. It's also more finicky — stringing, slower print speeds, and it likes to stick to the bed a little too well. Not a big deal once you've dialed it in. A real nuisance if you haven't.

My rule: if the prototype is going to be touched by someone who isn't you, lean PETG. If it's a fit-and-form check you'll toss in a drawer, PLA is fine. Pick based on how the part will be treated, not which spool is already on the shelf. That's not a preference — it's a process decision that shows up later as either "the test went fine" or "why is the housing cracked?"

How to Tell Which Scenario You're Actually In

Stop and answer four questions before you send an RFQ anywhere:

  1. Does the part bear load or see heat? Yes to either, and you're in Scenario 1 (metal) or Scenario 4 (functional FDM, not just visual).
  2. Does it touch a patient? Scenario 2. Documentation first, machine second.
  3. Does color or visual realism matter more than strength? Scenario 3.
  4. What's the batch size, really? Five parts or 5,000 — the answer changes the process, but not the standard of care.

A quick word on that last point, because I've seen it go wrong too often: small orders get treated as an afterthought. A 5-part trial gets pushed to the back of the queue, specced loosely, inspected casually. That's a bad trade. Today's 5-part trial is often next year's production order — and the supplier who gave it a real spec review is the one you keep. Small doesn't mean unimportant. It means potential.

This guidance reflects practices we audited through Q1 2025. Materials, machine capabilities, and service offerings change fast in both additive and subtractive manufacturing — verify current specs and certifications with your supplier before specifying.

Tomas Eriksen

Tomas Eriksen is an independent cutting tool and machine tool analyst covering end mills, drills, taps, reamers, inserts, boring bars, tool holders, lathes, and machining centers. He uses ISO 13399 tool-data concepts while comparing substrate, coating, geometry, runout, rigidity, spindle interface, cutting speed, feed, chip evacuation, and cost per finished part. His procurement guides help shops match tools and equipment to materials, operations, capacity, and measurable productivity goals.