Guidelines for plasma CNC cutting

• 1
  How plasma CNC cutting removes metal
• 2
  Which metals suit plasma CNC cutting
• 3
  Tolerance, kerf and edge quality you can expect
• 4
  Parameters that decide cut quality
• 5
  Where plasma fits before precision machining
• 6
  How to plan a plasma cut job
• 7
  Plasma cutting versus milling for plate profiles

Plasma CNC cutting is a thermal process. A constricted arc forms between a tungsten or hafnium electrode inside the torch and the workpiece itself. Compressed gas flows through a narrow nozzle, gets heated by the arc, and turns into plasma — an ionized gas that conducts current. The jet leaves the nozzle at high velocity and at a temperature that reaches roughly 16,600 °C at the core.

That heat does two things at once. It melts a narrow band of metal, and the same gas stream blows the molten material out of the kerf before it can resolidify and bridge the cut. The motion is controlled by the CNC table, so the torch path, travel speed and standoff height are all programmed. This is why the process is called plasma CNC cutting rather than simple hand cutting: the arc does the cutting, the machine decides where.

The cut is a melting process, not an ablation process. There is no mechanical contact between tool and workpiece, so the plate does not need to be clamped against cutting forces the way a milled part does. That makes it fast on flat stock. It also means the heat-affected zone and the dross on the bottom edge are built into the result, not optional side effects.

• 1
  Arc temperatureAround 16,600 °C at the core, far above the melting point of steel or aluminum.
• 2
  Gas does the clearingThe same jet that melts the metal pushes the molten pool out of the kerf.

The arc must complete a circuit through the workpiece, so the material has to conduct electricity. Mild steel, stainless steel, aluminum, copper, brass and most conductive alloys all cut. Non-conductive materials such as wood, acrylic, glass or ceramic do not, no matter how the parameters are set.

Thickness is the second filter. Low-carbon steel is the friendliest material for the process and is usually cut from 1 mm up to roughly 25 mm on a conventional table. Stainless and aluminum behave differently because they conduct heat away faster and form refractory oxides. The practical ceiling drops, often to around 12–20 mm depending on the power source and gas.

Copper and brass cut cleanly but pull heat away from the arc very quickly. On thin sections this is fine. On thick plate the cut can lose its bottom edge or require a larger nozzle and a slower travel speed. Titanium cuts but demands argon shielding to avoid contamination of the cut face.

One more condition: the plate has to be flat and reasonably clean. Mill scale, rust and heavy oil change the arc start and the kerf width. Light surface rust is tolerable. Loose scale is not.

Plasma is a roughing process. Across a typical table, a well-tuned cut on 6 mm mild steel holds about ±0.5 mm on profile position, and the kerf is usually 1.0–3.0 mm wide depending on nozzle size and current. Holes under about 1.5× the plate thickness start to lose roundness and taper, so small holes are often drilled or milled after cutting.

The cut face is never a finished surface. It carries a slight bevel, a heat-affected zone a few tenths of a millimeter deep, and often a hard dross bead on the underside. As-cut roughness sits in the range of roughly Ra 6.3–25 μm, far coarser than a machined face. If a drawing calls for Ra 0.8–1.6 μm or a sealing surface, plan a secondary machining pass.

Edge squareness is where thin and thick plate diverge most. On 3 mm steel the edge can look nearly square. On 20 mm plate the top edge stays close to the programmed line while the bottom edge lags, producing a visible bevel. If the part has to sit flush against another face, that bevel matters.

Corner behavior also differs. The machine has to slow down for a sharp corner, which dumps more heat into that spot and can round it. Adding a small corner radius, or programming a lead-out, keeps the corner crisp.

• 1
  Kerf compensationThe CAM path must be offset by half the kerf, or every part comes out undersize.
• 2
  Heat-affected zoneUsually a few tenths of a millimeter; hard for taps and reamers if not removed.

Travel speed is the single biggest lever. Run too fast and the arc cannot melt through, so the jet leaves a heavy dross bead and the bottom edge lags. Run too slow and the kerf widens, the heat-affected zone deepens, and the top edge washes out into a rounded shoulder. On 6 mm mild steel, a 100 A nozzle typically runs somewhere in the 2,500–3,500 mm/min band; the exact number comes from the cut chart and a test coupon.

Standoff height, the gap between nozzle and plate, controls arc shape. Too much gap spreads the arc and widens the kerf. Too little risks a double arc that destroys the nozzle. Most cut charts specify 1.0–2.0 mm, and a torch height controller holds that value as the plate warps.

Gas choice changes the cut face. Compressed air or oxygen is common on mild steel. Nitrogen suits stainless and aluminum because it limits oxidation on the cut face. Argon-hydrogen mixes are used where a cleaner edge matters more than cost.

Consumable condition is not a detail you can ignore. A worn nozzle or a pitted electrode changes the arc and shifts the kerf within a single nest. Track pierce counts and swap consumables on a schedule rather than when a cut visibly fails.

Most plate parts do not need a milled profile from a solid blank. A plasma table turns a 4,000 × 400 × 150 mm plate envelope into near-net profiles quickly, and the machining center then handles only the features that need tolerance: bores, slots, tapped holes, faces and datums.

That split changes cost and lead time. The thermal cut removes the bulk of the material with almost no tool wear. The mill then works on a part that is already close to shape, so cycle time drops and the ±0.005 mm tolerance is spent on the geometry that actually needs it.

There is a limit to this. If the part is a thin-walled housing where the heat-affected zone would distort the wall, plasma is the wrong first step. If the part is a bracket with generous tolerances and no sealing faces, plasma may be the only step needed.

In our Dongguan and Singapore plants, plasma-cut blanks feed the 5-axis and mill-turn cells when the geometry calls for it. The decision is made on the drawing, before the plate is nested.