Precision Gold CNC Processing Guide

• 1
  Key takeaways
• 2
  Why gold behaves differently at the spindle
• 3
  How precision gold CNC setups control tolerance
• 4
  Feeds, speeds, and finish on soft metal
• 5
  Tool wear, smearing, and surface defects
• 6
  When this process is the wrong choice
• 7
  Three-axis vs five-axis for gold parts

Gold sits at the soft end of the machined metals. On the Mohs scale it is around 2.5, and pure gold is roughly as soft as a fingernail. That has one immediate consequence: the chip does not shear cleanly and break away. It stretches, curls, and tends to weld to the cutting edge. In practice we see built-up edge form within the first few passes if the surface speed is too low.

The second consequence is heat. Gold conducts heat well, so much of the cutting heat leaves with the chip and the workpiece rather than sitting in the tool. That sounds helpful, and it partly is. But the same property means the part itself warms up and expands. A small gold feature that measures correctly on a warm machine can shrink out of tolerance once it cools to 20 °C.

Alloys change the picture. Pure 24K gold is the most ductile and the most difficult to finish. Add copper, silver, or nickel and hardness climbs. Common jewelry and electronics alloys such as 18K yellow, 14K, and 18K white behave far more like a normal brass or a soft stainless. The cutting data we would use on C36000 brass is a reasonable starting point for many of them.

So the first question is never "can it be machined." It is which alloy, how much stock, and which faces actually carry the tolerance. A part where only a bore and a seating face are critical is a much easier job than one where every surface is called out.

• 1
  Ductile chipsPlan for chip control, not chip breaking.
• 2
  Thermal growthLet the part stabilize before final inspection.
• 3
  Alloy firstHardness range spans from near-pure to jewelry alloys.

Tolerance on a machined gold part is rarely lost at the cutting edge. It is lost in the transitions. Every time the part comes off the fixture and goes back on, you re-introduce workholding error, chip re-cutting risk, and datum shift. A part with five faces and three setups carries three chances to drift.

This is where five-axis work earns its place. With a trunnion table and a rotary C axis, the tool can reach five sides of a small gold component in one clamping. Datum stays fixed. We hold ±0.005 mm (±0.0002 in) on critical features routinely on this platform, but the number that matters more is repeatability across the batch, not the single best part.

The counter-argument is real. Five-axis programming and setup cost more per hour than a three-axis job. For a flat plate with holes on one face, a three-axis machine is the correct answer and it will be cheaper. The judgment call is geometry: if the part has compound angles, undercuts, or features on four or more faces, single-setup five-axis usually wins on total cost.

We run 16 simultaneous five-axis machining centers alongside 27 three-axis and 12 four-axis mills. That mix matters because it lets us route a job to the machine that fits it rather than pushing everything onto the most expensive platform.

• 1
  Count the setupsThree or more faces usually favors five-axis.
• 2
  Datum disciplineOne clamp, one datum, fewer drift sources.
• 3
  Batch repeatabilityBetter predictor of yield than a hero part.

For small gold parts, the working principle is light radial engagement at high spindle speed. A 3 mm or 4 mm carbide end mill at 12,000 to 20,000 rpm with a 0.03 to 0.08 mm tooth feed is a reasonable band for many jewelry alloys. The exact number depends on alloy hardness and tool coating, so this is a starting point, not a recipe.

Do not push depth of cut. Gold features are often thin walls, fine prongs, or small bosses. A heavy axial pass deflects the part and leaves a taper that shows up later at inspection. Two or three light axial passes cost more cycle time but hold the wall thickness.

Cooling matters in a specific way. Flood coolant keeps the part temperature stable and flushes soft chips out of the cut. Air blast alone often leaves chips in the pocket, and a re-cut chip will scratch a polished face in one pass. If the part will be polished afterward, that scratch becomes extra hand work.

Finish targets are achievable but need honest planning. With correct parameters we reach Ra 0.8–1.6 μm as-machined on gold alloys, and Ra 0.2–0.8 μm is reachable where the geometry allows a finishing pass with a fresh tool. Below that, you are buying polishing, not machining.

• 1
  High rpm, light passReduces deflection on thin gold features.
• 2
  Flood over airSoft chips left in the pocket cause scratches.
• 3
  Fresh tool for finishA worn edge smears instead of cutting.

The most common defect we see on machined gold is smearing, not chipping. It looks like a dull, dragged patch on an otherwise fine surface. The cause is almost always a worn cutting edge. Once the edge radius grows past roughly 5 to 10 μm, the tool stops shearing and starts pushing material along the surface.

The fix is a tool-change policy, not a parameter tweak. On small gold work we track cut time per tool and replace on a fixed interval rather than waiting for a visible problem. Uncoated carbide or a light diamond-like coating both work. Diamond tooling lasts far longer but the cost only makes sense on high-volume runs or on optical surfaces.

Burrs are the second issue. Gold burrs are soft and tend to fold over rather than stand up, which makes them easy to miss visually. A folded burr on a mating face will close a clearance gap after assembly, long after the part has passed incoming inspection. Deburring by hand on gold is risky because the same softness that formed the burr lets a file remove too much.

This is why we inspect 100% of parts before shipment, with raw material checks, in-process monitoring, and a final inspection stage. On gold, a visual pass under magnification is as important as a CMM report.

• 1
  Smearing = dull toolChange on time, not on failure.
• 2
  Folded burrsThey sit flat and escape a quick look.
• 3
  Report on requestDimensional data is available if you need it.

CNC machining is subtractive, so it starts from a solid billet. On a gold part, the raw material often costs more than the machining time. If your design removes 70% of the billet as chips, the scrap value may not cover the difference. In those cases, a casting route with a light machining pass on critical features is usually the better economic answer.

Very small batch sizes also shift the math. Programming and fixturing a complex five-axis gold part takes real hours that do not scale down. For a one-off, a simpler three-axis setup with one or two operations can be faster overall, even if the cycle time per part is longer.

There is also a geometry floor. Features narrower than roughly 0.5 mm, deep pockets with aspect ratios beyond about 5:1, and internal radii below the smallest available cutter radius will need a design change or a different process. We flag these in DFM review rather than quoting a job we cannot hold.

None of this means machining is a poor fit for gold. It means the decision should be driven by geometry and material yield, not by the material name. Send the model and we can say which route fits within the 12-hour quote and free DFM analysis window.

• 1
  High scrap ratioCasting plus finishing may cost less.
• 2
  One-off partsSetup time dominates a simple job.
• 3
  Thin featuresBelow ~0.5 mm needs a design review.