Two processes, two different questions
The comparison gets framed as a technology contest. It is really two different questions wearing one headline.
3D printing answers: can I hold this idea in my hand this week? A printer builds the shape layer by layer from a CAD file, with no tooling, no fixturing and no programming beyond a slice. Complexity is nearly free, and quantity one is its natural habitat.
CNC machining answers: can I make this exact part, in this exact material, to this exact number, repeatedly? The machine cuts from solid certified stock, so the part inherits the material's proven properties, holds tolerances printing cannot, and carries a surface finish measured in microns rather than layer lines.
Ask which question you are actually asking, and the choice mostly makes itself.
The comparison that decides it
| CNC machining | 3D printing | |
|---|---|---|
| Materials | Any machinable metal or plastic, from certified stock with EN 10204 3.1 traceability | Strong in polymers; metal printing exists but is specialist and expensive |
| Tolerance | ±0.127 mm standard, ±0.025 mm precision | Typically several tenths of a millimetre, process dependent |
| Strength | Properties of the parent material, uniform in all directions | Often weaker between layers; orientation matters |
| Surface | Machined finish, improvable to ground or polished | Layer texture; post-processing needed for smooth or sealing faces |
| Complexity | Costs money: every feature is cutting time and setups | Nearly free: internal channels and lattices print happily |
| Quantity one | Carries all setup cost alone | Its natural habitat |
| Repeat batches | Unit price falls steeply with quantity | Unit price barely moves; scales poorly past small numbers |
| Documentation | Material certs, FAIR and CMM inspection are routine | Traceability and certification are harder and process specific |
Decision rules that survive contact with real parts
- Metal part doing a structural, sealing or rotating job. Machine it. Proven properties, real tolerances, and the paperwork regulated sectors demand.
- Polymer part, complex geometry, quantity one to a handful. Print it. Nothing beats a printer's economics at quantity one with complexity thrown in.
- Form-and-fit prototype before committing to machining. Print first, deliberately. A £50 print that finds the interference problem saves a machined batch cut to the wrong revision.
- Tight tolerance on a printed shape. Hybrid: print or cast the near-net shape, then machine the critical faces. Common on complex housings.
- Repeat production at any volume that matters. Machine it, and use batch economics properly. Printing's flat cost curve becomes a penalty as numbers grow.
The mistake in both directions is loyalty. Teams that print everything ship weak, rough parts that should have been machined. Teams that machine everything pay setup money to answer questions a print would have answered by Thursday.
Where TrueNorth fits in this
TrueNorth Engineering is a machining-side supplier: we coordinate CNC machining, fabrication and finishing through a vetted UK network, with inspection before dispatch and traceability as standard. We do not run printers, and for a polymer prototype with no downstream dependency an online print service is the right call, as our platform comparison says plainly.
Where we earn our place is the moment the printed prototype has done its job and the part needs to become real: certified material, honest tolerances, a finish that seals, and a quote that covers the delivered, inspected part. Send the CAD you printed from, tell us what the part does, and the DFM review comes back within 48 hours, including which printed features want redesigning for the mill.