What Are the Characteristics of CNC Machining?

• 1
  CNC machining versus casting, sheet metal and 3D printing
• 2
  Accuracy that comes from the machine, not the operator
• 3
  Material range: from 6061 aluminium to Inconel and PEEK
• 4
  Geometry freedom, and where it stops
• 5
  Tolerance, finish and how they trade against each other
• 6
  Setup, batch size and why quantity changes the answer
• 7
  Repeatability and traceability across a production run
• 8
  Where CNC is the wrong choice

The first thing people name when they list the characteristics of cnc machining is accuracy. It is worth being precise about where that accuracy comes from. The tool path is fixed in code before the first cut, so the same program produces the same geometry on Monday morning and Friday night. On our 5-axis centers we hold ±0.005 mm (±0.0002 in) on critical features, and the operator is not chasing that number by hand.

Manual turning and milling depend on the skill of whoever is at the spindle. Two machinists, two slightly different parts. That variation is hard to see on one sample and very visible in an assembly of fifty. CNC removes the hand from the geometry, so the drawing that passed first article keeps passing at part 500.

The limit is real, though. Accuracy applies to features the tool can reach. A deep pocket with a 4:1 depth-to-diameter ratio in a 3 mm end mill will deflect, and no controller fixes that. If a feature needs a long thin tool, expect to lose tolerance or pay for a second setup with a shorter tool.

We check 100% of parts before shipment, with raw material verification, in-process monitoring and a final inspection. Reports are available on request. That is how the ±0.005 mm figure stays a fact rather than a brochure line.

A second characteristic of cnc machining is how little the process cares about the material family. The same machine that cuts 6061-T6 will cut 17-4PH stainless, TC4 titanium or a block of PEEK. What changes is the cutting data, the tool grade and the coolant strategy, not the basic setup.

In practice we machine 6061, 6061-T6, 2024, 5052, 5083, 6063, 6082, 7075 and ADC12 in aluminium; 303, 304, 316, 316L, 420, 430, 431, 440C and 17-4PH (SUS630) in stainless; 1018, 1045, 4130, 4140, 4340 and A36 in steel; C101, C110, beryllium copper and C36000 brass; plus TA1, TA2, TC4, Inconel and magnesium AZ31B or AZ91D. Plastics run from ABS and POM to PEEK and carbon fibre.

Material choice does change cost, and it is not only the stock price. Titanium and Inconel cut slowly, wear tools fast and often need more light passes. A part designed in 7075 for stiffness may cost 30% more than the same part in 6061 without any gain the assembly can feel. Ask what the material is actually doing before specifying it.

One warning: some materials are not a good fit for machining at all. Very soft or gummy alloys can tear instead of cut, and thin carbon fibre plates can delaminate at the edge. If the part is mostly flat and thin, sheet metal or waterjet may be the better route.

CNC handles shapes that casting and stamping cannot produce without extra tooling. Undercuts, deep pockets, angled faces, blended radii and pockets with drafted walls all come off the same setup. With 16 simultaneous 5-axis machining centers we can reach features that would need three or four separate fixtures on a 3-axis machine.

That matters most for parts with compound angles, ports, or pockets that meet at odd angles. A 5-axis setup machines them in one operation, which also means one datum. Fewer setups usually means tighter true position between features, because the part is not being re-clamped and re-zeroed.

Geometry freedom is not unlimited. A square internal corner cannot be cut by a round tool. A pocket narrower than the smallest cutter we stock cannot be machined at all. Holes below roughly Ø1 mm in deep walls become a drill-breakage risk. Features on the far side of a tall wall may need a longer tool, and long tools chatter.

The practical rule: if a feature needs a tool that cannot physically reach it, no amount of 5-axis motion helps. Send the model early and we will flag those spots in the DFM report before the program is written.

Tolerance and surface finish are the two numbers that move price the most, and they move together. As-machined surfaces sit around Ra 1.6–3.2 μm. A high-quality finish is Ra 0.8–1.6 μm. Fine finishes reach Ra 0.2–0.8 μm, and those usually need a separate finishing pass with a smaller stepover, which adds machine time.

Tolerance bands work the same way. A general tolerance of ±0.1 mm on a milled aluminium part is routine. Tightening a few critical features to ±0.005 mm means slower feeds, temperature control and more inspection. Tightening every dimension on the print is expensive and usually unnecessary.

Good drawings mark only the features that matter. A bearing bore, a sealing face or a dowel hole deserves a tight callout. The clearance hole next to it does not. Over-tolerancing is one of the most common reasons a quote comes back higher than the buyer expected, and it is easy to fix before release.

Finish also depends on the alloy. Aluminium takes a fine finish easily. 316 stainless work-hardens and can smear if the feed is too light. Titanium tends to hold a burr. If the surface is cosmetic, say so on the drawing; if it is functional, give the Ra value and the gauge length.

CNC has no tooling cost. That single fact drives the quantity decision. A casting needs a steel die before part one, so it only pays off in the thousands. A CNC program needs a fixture and a proved-out tool path, and then the first part and the ten-thousandth part cost roughly the same per unit.

We run no minimum order quantity, from a single prototype to 10,000+ part runs. Prototype and low-volume work is where machining wins clearly. It is also the fastest way to get a real part in your hand: quotation and free DFM analysis within 12 hours, production can start within 24 hours, and parts ship in 3–5 days.

Above a few thousand pieces per year, the comparison changes. Die casting or forging plus a finishing cut on the critical faces can be cheaper per part. That route only works if the geometry suits a die, the wall thickness is above about 1.5 mm and the annual volume justifies the tool.

A mixed route is common and worth considering. Cast the body, machine the sealing face, the bore and the mounting holes. You get the low piece price of casting and the tight tolerance of machining exactly where the assembly needs it.

• 1
  Low volume, complex shapeCNC, no tooling, fastest to first part
• 2
  High volume, simple shapeCast or forge, then machine critical faces
• 3
  Flat and thinSheet metal usually beats milling on cost

Repeatability is what separates a good sample from a good supplier. Because the tool path lives in a file, part 500 matches part 1 as long as the machine is maintained and the tools are changed on schedule. Tool wear is the main drift source, and it is managed by measuring the tool and offsetting the wear value, not by hoping.

Traceability supports that. Material certificates follow the batch. In-process checks catch drift before a full lot is finished. Final inspection confirms the parts before they ship. If a customer needs dimensional reports, we can supply them with the shipment.

This matters in regulated work. We hold ISO 9001:2015, IATF 16949:2016, ISO 13485:2016 and ISO 27001:2022 across three wholly-owned plants totalling 7,600 m², with 150 technicians and 127 high-precision CNC machines. Aerospace, automotive and medical buyers usually ask for the certificate set before they ask about price.

One honest limit: repeatability assumes the same shop, the same machines and the same process sheet. Moving a mature part to a different supplier resets part of that history. Expect a fresh first-article run even when the drawing has not changed.

Machining is flexible, but it is not always the cheapest or the best. Knowing when to stop is as useful as knowing when to start. Four situations come up often in our quote queue, and in each one we will say so rather than take the order.

Thin, flat parts with large surface area are a poor fit. A 1 mm aluminium panel will bow and chatter, and you will pay for stress relief and multiple light passes. Laser cutting or stamping is faster and flatter.

Very high volume simple parts belong to casting or forging. If you need 50,000 identical brackets a year with no tight features, a die pays for itself. Machining them is technically possible and commercially wrong.

Parts with internal cavities that no tool can reach belong to 3D printing, investment casting or a split design. And hollow, thin-wall parts with large volume are often better printed or moulded, because removing material from a solid block wastes both time and stock.

The last case is cosmetic-only parts. If the geometry is simple and the surface is the whole point, vacuum casting or 3D printing plus finishing may reach the same look at a lower cost. Bring the function, not just the shape, and the route choice usually becomes obvious.