Plasma cutting machine CNC technology explained
A working explanation of how a CNC plasma torch forms and controls an arc, which conductive metals and thicknesses it suits, and where the process stops being the right choice. Written for engineers and buyers who need to judge a cut edge, not just a cut.
In this article
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
Key takeaways
How a plasma cutting machine CNC technology builds and holds an arc
A plasma cutter does not chip or shear metal. It melts a narrow band of it and then blows the molten material out of the kerf with a high-velocity gas jet. The heat source is an electric arc constricted through a small nozzle, which raises the arc temperature far above what an open arc reaches and concentrates it into a spot roughly the nozzle diameter.
The sequence starts with a pilot arc between the electrode and the nozzle. Once the torch is close enough to the workpiece, the arc transfers to the plate and the pilot circuit drops out. From that moment the workpiece is part of the circuit, so the cut only works on conductive material. Ceramics, glass and most plastics are out.
Current sets how much metal melts per second. A 30 A cut on 3 mm mild steel behaves very differently from a 200 A cut on 25 mm plate: the kerf widens, the heat-affected zone deepens, and the edge angle grows. Operators pick amperage from a cut chart, then trim travel speed to suit the actual plate.
The CNC side does three jobs at once. It moves the torch along the programmed path, it holds the standoff height through an arc-voltage feedback loop, and it schedules pierce delay and lead-in moves. Get any of the three wrong and the part is out of tolerance even though the arc itself was fine.
Kerf, taper and the heat-affected zone
Kerf is the width of material the arc destroys. On mild steel it runs about 1.5 mm at 40 A and 3–4 mm at 200 A. CAM software offsets the tool path by half the kerf, so the finished outline lands on the nominal dimension. If the operator uses the wrong kerf value, every part on the nest shifts by the same amount.
Taper is the difference between the top and bottom width of the cut face. A square cut has near-parallel walls; a tapered cut leans in or out by a few degrees. The lean comes from the arc's own shape, from worn consumables and from cutting too fast or too slow. On 6 mm plate a fresh nozzle can hold taper under about 3°, and a worn one will not.
The heat-affected zone sits just below the cut face. It is a thin layer where the microstructure changed but the metal did not melt. On low-carbon steel that layer machines easily. On 4130, 4140 or 4340 it hardens, and a subsequent milling pass may need reduced feed to avoid chipping the edge.
Dross is the resolidified metal that clings to the bottom edge. It usually means travel speed is off, standoff is too high, or the gas mixture is wrong for the alloy. Low-speed dross is a soft blob that scrapes off; high-speed dross is a hard bead that takes a grinder. Either way it signals a setup problem worth fixing before the next sheet.
Which metals suit it, and which do not
CNC plasma cutting works on conductive metals: mild steel, stainless steel, aluminium, copper, brass and most conductive alloys. Performance changes a lot with thickness and alloy. Aluminium conducts heat away quickly, so it needs more current and faster travel than the same thickness of steel. Copper and brass are worse still, because their thermal conductivity pulls heat out of the cut zone.
Stainless steel cuts cleanly but the oxide layer on the cut face is hard and chemically active. If the part will be welded or passivated later, that layer has to be removed. On thin stainless the edge can also discolour from heat, which matters on visible panels.
Thickness is the practical ceiling. Below roughly 1 mm, plasma is hard to control: the arc is wider than the plate is thick and the edge distorts. Above roughly 25–30 mm on mild steel, cut quality drops and the process competes badly with other methods. In between, plasma is fast and cheap per meter of cut.
Non-conductive material is simply out of scope. So is anything where the cut face is a functional surface: a bearing bore, a sealing face, a precision slot. Plasma will get you a blank close to shape, not a finished feature. That blank then goes to a mill or lathe.
Plasma versus laser, waterjet and milling
Laser cuts a narrower kerf and holds tighter tolerance, but capital cost and running cost rise steeply with thickness. On thin sheet, laser usually wins on edge quality. On 12 mm and up, plasma is the cheaper cut per meter, and it is far less fussy about surface condition and rust.
Waterjet cuts almost any material with no heat-affected zone, which matters on titanium, Inconel and heat-treated alloys. It is slower and the abrasive running cost is real. If your part cannot tolerate any thermal change at the edge, waterjet is the safer call even at a higher price.
Milling is a different question. Plasma gives you a profile; milling gives you dimensions, holes, pockets and surface finish. A bracket outline can be plasma cut and then drilled and faced on a CNC mill. For a part held to ±0.005 mm with a fine finish, plasma is only the first operation, never the last.
The judgment call is usually about what the edge must do. If it is a weld prep or a clearance edge, plasma is enough. If it is a mating face or a bearing seat, it is not. Engineers who decide this early save a second setup later.
What a plasma cut can and cannot hold
A well-set plasma cut on mild steel holds roughly ±0.5 mm on the profile and better on hole position, since position comes from the machine, not the arc. The arc affects size, not location. That split matters: a nest can be dimensionally consistent and still be off nominal if the kerf value is wrong.
Hole quality is the weak point. Small holes come out tapered and often out of round, because the arc has to pierce and then travel a very short path. A common shop practice is to plasma a pilot hole and drill or mill to final size. On holes under Ø6 mm, plasma alone rarely holds the tolerance.
Edge squareness on thick plate is limited by physics. The arc leans as it penetrates, so the bottom of a 20 mm cut is narrower than the top. If the part needs a square edge on thick plate, plan a machining pass. Budget for it in the routing rather than discovering it at assembly.
For parts that do need machined features, GreatLight runs 127 high-precision CNC machines across three wholly-owned plants, including 16 simultaneous 5-axis machining centers, with tolerances to ±0.005 mm and finishes from Ra 1.6–3.2 μm as machined down to Ra 0.2–0.8 μm. Plasma blanks feed that workflow; they do not replace it.
Consumables, gas and the cost of a bad setup
Consumables wear predictably. The electrode and nozzle erode with every pierce and every meter of cut, and the arc grows as they do. A nozzle that has cut a few hundred pierces will produce more taper and more dross than a fresh one. Shops that track pierce count get more consistent parts than shops that replace on failure.
Gas choice follows the metal. Oxygen gives the fastest, squarest cut on mild steel. Air is cheap and general purpose. Nitrogen suits stainless and aluminium where oxide and nitride formation matter, and argon-hydrogen mixes appear on thicker stainless. The wrong gas shows up as dross or a rough face, not as a machine fault.
Piercing is where most consumable damage happens. Thick plate needs a longer pierce delay and a higher standoff, and piercing over the finished part edge will blow a crater into it. CAM should place pierces on scrap whenever the geometry allows.
Water tables and downdraft tables handle fume differently. Water tables cut smoke and noise, which helps on thick plate, but they also splash and can affect the underside of the cut. Downdraft is cleaner for thin sheet. Neither changes the cut geometry.
Process choice by part requirement
Pick the row that matches the feature that matters most.
| Requirement | Plasma | Laser | Waterjet or mill |
|---|---|---|---|
| Sheet 1–6 mm, fast outline | Best fit | Better edge | Overkill |
| Plate 12–25 mm mild steel | Best fit | Costly | Slow |
| Tolerance below ±0.1 mm | Not suitable | Close | Mill |
| No heat-affected zone | Not suitable | Thin only | Waterjet |
| Holes under Ø6 mm | Poor | Good | Mill or drill |
| Weld prep edge | Good | Good | Good |
| Visible cosmetic edge | Needs finishing | Good | Good |
| One-off prototype blank | Fast and cheap | Costly | Slow |
Where plasma stops and milling starts
Cut a profile with plasma when the edge is a weld prep, a clearance edge or a blank to be finished later; switch to CNC milling the moment the feature carries a tolerance, a sealing face, a bearing bore or a cosmetic surface. One rule covers most cases: if the drawing calls out a dimension with a decimal point, plasma is the first operation, not the last.
Questions engineers ask next
Can plasma cut aluminium and stainless as well as mild steel?
Yes, all three are conductive, so the arc transfers and cuts. The setup changes: aluminium pulls heat away fast, so it needs more current and faster travel; stainless needs nitrogen or a mixed gas to keep the cut face clean.
Thickness limits are tighter on aluminium and stainless than on mild steel at the same amperage. Expect to drop to a lower thickness for the same edge quality.
Why does my cut have more taper on one side of the part?
Directional taper is normal. The arc swirls, so the leading edge and trailing edge of a contour lean differently. On a closed contour the two sides can lean in opposite directions.
Consumable wear makes it worse. Check nozzle condition first, then travel speed. If both are good and the part still needs square walls, plan a machining pass.
How much material should I allow for kerf in CAM?
Use the value from the cut chart for your amperage and thickness, typically 1.5 mm at 40 A and 3–4 mm at 200 A on mild steel. Offset the tool path by half that value.
Confirm it on the first part instead of trusting the chart. Measure the kerf on a scrap cut and update the tool library. A wrong kerf value shifts every part on the nest by the same amount.
Is a plasma cut edge good enough to weld without preparation?
Often yes. A clean plasma edge on mild steel takes a weld directly, and the slight surface roughness helps the weld wet out. Remove dross and any loose oxide first.
On stainless, the oxide layer has to come off before welding or passivation. On thick plate, check that the edge is square enough for your joint design; a strongly tapered edge changes the root gap.
What is the smallest hole plasma can cut reliably?
As a working rule, hole diameter should be at least equal to plate thickness, and ideally 1.5 times it. Below that, the arc cannot complete the contour cleanly and the hole comes out tapered and out of round.
For holes under Ø6 mm, plasma a pilot and drill or mill to size. That is faster than fighting the arc and gives a hole that holds tolerance.
Does plasma cutting change the material properties?
It changes a thin layer at the cut face. The heat-affected zone on low-carbon steel is soft and machines easily. On alloy steels such as 4130 or 4140 that layer hardens.
If the part sees fatigue loading or a subsequent machining pass, account for that hardened layer in the routing. On parts where no thermal change is acceptable at all, waterjet is the better process.
Send the drawing, get a manufacturability answer
Upload a STEP or DXF file and we will tell you whether the part should be plasma cut, milled, or both, with a quotation and free DFM analysis within 12 hours.
12-hour quote100% inspectionNo minimum order quantity
More process notes from the shop floor
We publish setup notes, tooling trials and inspection data from the factory floor.
