A 4x8 CNC Plasma Cutting Machine: When It Fits and When It Does Not

The 4x8 refers to the cutting bed: 4 ft x 8 ft, or roughly 1,220 x 2,440 mm. That is the sheet size the machine was built around, and it matches the standard plate format most steel service centers stock. A gantry carries a plasma torch over the plate on X and Y axes, and a CNC control follows a 2D toolpath generated from a DXF or nested layout file.

The cut itself is thermal. A constricted arc ionizes gas into plasma at roughly 25,000 °C, melting the metal along the kerf. The torch gas then blows the molten material out of the cut and down through the plate. There is no mechanical contact, so plate hardness matters less than plate thickness and thermal conductivity.

Most 4x8 tables run from a 45 A to 200 A plasma power supply. Low amperage handles 0.5-6 mm sheet with a fine kerf. High amperage pushes through 25-50 mm depending on material. Steel, stainless, aluminum, and copper alloys all cut, but they do not cut the same way.

• 1
  Carbon steelCleanest plasma results. 0.5-50 mm depending on amperage.
• 2
  Stainless steelNeeds nitrogen or H35 gas mix; dross control is harder.
• 3
  AluminumCuts fast but the kerf edge hardens and oxide forms quickly.
• 4
  Copper and brassPossible but slow; the heat spreads into the plate instead of the kerf.

A well-tuned 4x8 CNC plasma cutting machine holds roughly ±0.5 mm on profile position for thin plate, and it drifts wider as the plate thickens. On 20 mm steel, ±1.0 mm or worse is normal because the arc lags the corner slightly and the kerf widens with heat. That is not a machine defect. It is the process.

The cut face tells you more than the dimension. Plasma leaves a bevel on one side of the kerf, usually 2-8° depending on speed and amperage. It also leaves a recast layer and a heat-affected zone a few tenths of a millimeter deep. On mild steel this usually machines off without trouble. On 17-4PH or a hardened tool steel, that HAZ can be harder than the base metal and it will fight a carbide end mill.

Surface finish off the torch lands around Ra 12-25 μm on the cut face. That is a saw-cut level of roughness. If a drawing calls for Ra 0.8-1.6 μm or a sealing face, plasma is a blanking step only. The finished geometry has to come from a machining operation afterward.

Choose plasma when the part is flat, the profile is 2D, and the drawing controls a shape rather than a fit. Mounting plates, machine guards, weldment gussets, HVAC transitions, trailer frame rails, and base plates are all good candidates. If a plasma-cut edge gets welded, ground, or covered by another part, the bevel and the roughness stop mattering. Cutting speed is high and cost per part is low.

Do not choose plasma when the cut edge is a functional surface. A bearing bore, a sealing groove, a dowel hole, a thread, or a mating face cannot come off a torch. Nor should plasma be used where the HAZ will affect fatigue life on a cycled part, or where the material is a precipitation-hardening or hardened alloy that will crack or work-harden at the cut edge.

Thin sheet is another mismatch. Below about 1 mm, plasma heat distorts the plate and the kerf is a large fraction of the feature size. Fiber laser owns that range. Very thick sections, above roughly 50 mm, are usually better on a waterjet or an oxy-fuel torch.

• 1
  Good fit2D profile, edge gets welded or covered, no tight fit.
• 2
  Poor fitFunctional edge, tight bore, thin sheet, cycled load.
• 3
  Wrong toolAny feature needing Ra under 3.2 μm or ±0.05 mm.

The most common real-world pattern is not plasma or machining. It is plasma first, machining second. The torch does the rough profile cut on the plate at low cost and high speed, then the part goes onto a mill for the features that actually control function. This keeps the expensive spindle time focused where it earns tolerance.

At GreatLight, a plasma-cut blank typically goes onto a 3-axis machine for face milling and hole patterns, or onto a 5-axis center when the part needs compound angles or features on more than one face. The 4,000 x 400 x 150 mm travel on our large mills handles long plasma-cut profiles without re-fixturing. We run 16 simultaneous 5-axis machining centers, 12 four-axis mills, 27 three-axis machines, and 16 mill-turn centers across 127 machines.

The tolerance split matters. Plasma holds the profile at ±0.5 mm. Milling holds the finished feature at ±0.005 mm with a surface finish of Ra 0.8-1.6 μm on a standard cut, or Ra 0.2-0.8 μm when the drawing calls for it. If you need that split on one part, say so on the RFQ. We quote a blank-and-machine route against a full-machined route so you can see the cost difference.

Mild steel is forgiving. The HAZ is shallow, the recast layer machines off with a normal carbide tool, and post-weld distortion is predictable. A36, 1018, 1045, and 4140 blanks all behave this way, though 4140 should be cut in the annealed condition if a later hardening step is planned.

Stainless steels are trickier. The oxide that forms on a plasma-cut 304 or 316 edge will blunt tooling and can cause weld porosity if it is not ground back. On 17-4PH (SUS630), the cut edge can reach a hardness that damages HSS tooling, so we usually specify a rough grind before milling. Aluminum cuts cleanly but the kerf oxide layer has to be removed before anodizing or the coating will not take evenly.

Titanium and Inconel are the cases where plasma is usually the wrong process. The HAZ on Ti-6Al-4V and Inconel is deep and metallurgically significant, and both materials need a controlled atmosphere or an abrasive process instead. For those, we route the blank through waterjet or start from solid stock.