CNC Parts Processing: How Metal Stock Becomes a Finished Part

• 1
  What CNC parts processing actually is
• 2
  Stock choice and how it sets your limits
• 3
  Fixturing: where most tolerance is lost
• 4
  Cutting parameters and surface finish
• 5
  Inspection and what the report tells you
• 6
  When CNC parts processing is the wrong answer
• 7
  CNC parts processing vs other processes

CNC parts processing is subtractive manufacturing: a rotating or stationary cutting tool removes material from a solid block until the remaining geometry matches a CAD model. The machine never sees the drawing. It follows G-code, a list of coordinates and feed rates generated by CAM software from your 3D file.

That sounds simple, and for a flat bracket it is. The complexity appears when tolerances tighten, walls get thin, or features sit on five faces. Then the question stops being how do we cut this and becomes how do we hold it while we cut it.

The loop has four stages: stock preparation, fixturing, cutting, and inspection. Each stage can add error. A perfect toolpath cannot rescue a part that moved 0.02 mm in the vise during a heavy roughing pass.

We run this loop across 127 high-precision CNC machines, including 16 simultaneous 5-axis centers and 16 mill-turn centers. Most jobs start with a free DFM analysis inside 12 hours, before any metal is cut.

Stock is not a neutral input. Aluminum 6061 machines fast and holds tight tolerances, but it galls if you push a dull tap into a deep hole. Stainless 316 work-hardens the moment the tool rubs instead of cuts, so feed per tooth has to stay above a floor. Titanium Ti-6Al-4V moves heat into the tool rather than the chip, which is why cutter life drops and cost rises.

Stock form matters too. Plate is cheaper than bar but can carry internal stress that releases after the first heavy cut. A long thin part machined from plate may bow 0.1 mm or more even when the machine is accurate. Normalizing or stress-relieving before finish passes is the usual answer.

For parts above 500 mm, we often start from near-net stock to cut cycle time and tool wear. Our largest travel is 4,000 × 400 × 150 mm, so oversized one-piece frames are possible, but the fixturing plan has to be designed before the quote, not after.

• 1
  Aluminum 6061 / 7075Fast, stable, good for tight-tolerance housings and fixtures.
• 2
  Stainless 303 / 316 / 17-4PHCorrosion resistance, but work-hardening punishes light feeds.
• 3
  Steel 1045 / 4140Strong and cheap, yet needs stress relief on slender parts.
• 4
  Titanium and InconelHigh heat and tool wear; budget more cycle time and cost.

A three-axis vise is fine for a block with features on one face. The moment features sit on the back, you either flip the part or move to a 4-axis or 5-axis setup. Every flip introduces a new datum error, typically 0.01–0.03 mm unless you build a dedicated fixture.

Five-axis machining removes flips. The tool reaches the underside while the part stays clamped, so the datum never changes. That is why we hold ±0.005 mm on complex parts that would drift out of tolerance on a flip-heavy three-axis plan.

Thin walls are the other trap. Below about 1.5 mm in aluminum, cutting forces start to deflect the wall away from the tool. The cutter leaves a tapered surface and the wall springs back. Light finish passes, sharp tooling, and sometimes a wax or low-melt fixturing compound keep the part rigid.

If a part needs a soft jaw, a vacuum plate, or a custom nest, say so early. Fixture design changes the quote more than the toolpath does.

Surface finish is a function of tool geometry, feed per tooth, spindle speed, and rigidity. A sharp carbide end mill at the right chipload can leave Ra 0.8–1.6 μm as machined. Push the feed too high and the finish turns scalloped. Drop it too low and the tool rubs, which is worse.

For cosmetic or sealing surfaces, we plan a separate finish pass with a smaller stepover. That costs cycle time but avoids hand polishing, which is hard to control on a curved face.

Fine finishes reach Ra 0.2–0.8 μm and usually mean slower passes and more inspection. As-machined at Ra 1.6–3.2 μm is enough for most brackets, covers, and internal structural parts.

Anodizing, bead blasting, and electroless nickel all change the final dimension slightly. Hardcoat anodizing can add 0.02–0.05 mm per surface, so if a bore has a tight tolerance, mask it or plan the pre-plate size.

Inspection is not a final gate. We check incoming stock, monitor in-process dimensions on critical features, and run a final inspection before shipment. Every part is inspected; reports are available on request.

For a ±0.005 mm feature, a caliper is not enough. That tolerance needs a micrometer, a bore gauge, or a CMM depending on geometry. If your drawing calls for ±0.005 mm, expect the quote to include measurement time.

Our historical qualification rate is 99.99%. That number comes from catching drift during the run, not from hoping the last part is good.

Ask for the inspection method, not just the result. A reported dimension without the gauge that produced it is hard to audit.

CNC is not always the right process. If you need 50,000 identical plastic housings, injection molding beats milling on unit cost once tooling is amortized. If the part is a thin sheet bracket, sheet metal fabrication is faster and cheaper. If it is a lattice or internal channel that no tool can reach, 3D printing wins.

CNC wins when geometry is complex, tolerance is tight, material is hard, or quantity is low to medium. It also wins when the part must be metal and functional, not just a visual model.

The crossover is not fixed. A 200-piece aluminum run can still be cheaper on a mill than on a die-cast tool. A 5,000-piece run usually flips. The deciding variables are tooling cost, wall thickness, and how much post-processing the casting needs.

We quote no minimum order quantity, from one prototype to 10,000+ part runs, so you can test the decision with real parts before committing to tooling.