Plasma CNC Cutting Proficiency: How the Arc Cuts and Where It Stops
Plasma CNC cutting proficiency is not about owning a bigger table. It is about reading kerf, dross, and bevel on the part in front of you. This page explains the arc physics, the variables you can actually control, and the tolerance band where a milled edge is the better call.
What the Arc Actually Does
Plasma CNC cutting starts with a pilot arc between the electrode and the nozzle. Gas passes through that arc, heats into a plasma state, and leaves the nozzle at roughly 20,000 °C. The workpiece closes the circuit, the arc transfers to the plate, and the cut begins.
The arc does two jobs at once. It melts a narrow band of metal, and the same gas flow pushes that molten metal down and out of the kerf. Too little flow and dross stays welded to the bottom edge. Too much and the arc wobbles, so the kerf widens and the cut face roughens.
Cut quality is mostly a story about heat balance. Travel too slowly and the plate absorbs extra heat, so the top edge rounds over and the cut face shows heavy drag lines. Travel too fast and the arc lags behind the torch, leaving a beveled edge with a ragged bottom.
This is why plasma CNC cutting proficiency is a skill of reading the cut, not of memorizing a chart. Steel, stainless, and aluminum each want different speeds, gases, and standoff. The numbers below are starting points, not laws.
Kerf, Bevel, and Dross: The Three Signals
Kerf is the width of metal the arc removes. On a 6 mm carbon steel plate a typical kerf runs 1.0–1.8 mm. Move to 25 mm plate and the same torch cuts 2.5–4.0 mm wide. That gap is not waste you can ignore when you are nesting parts.
Bevel is the angle of the cut face. A square edge is the goal, but the arc is a cone, so the top of the kerf is always a little wider than the bottom. On thin plate this shows as a few hundredths of a millimeter. On 25 mm plate a 3–5° bevel is normal.
Dross is the resolidified metal hanging on the bottom edge. Low-speed dross is a heavy blob that needs a chisel. High-speed dross is a fine, brittle bead that brushes off. Both mean the speed is off. Clean cuts have almost none.
Read the three together. A wide kerf with light dross usually means too much current for the thickness. A narrow kerf with a heavy bevel means the torch is running too fast or the standoff is too high.
Variables You Can Actually Control
Amperage sets how much metal the arc can melt. A 45 A nozzle is happy on 6–10 mm carbon steel. A 130 A nozzle handles 25–32 mm. Running a small nozzle at high amperage burns the tip in minutes and widens the kerf for no gain.
Cutting speed is the single biggest lever on edge quality. Most manufacturer charts are conservative by 10–15% for hand-fed tables. On a CNC table with a good height control, you can run closer to the top of the range and get a squarer edge.
Standoff is the distance between the nozzle and the plate. It is usually 1.0–2.5 mm. Pierce height is higher, often 3–6 mm, so the molten splash does not plug the nozzle. A pierce that is too low kills consumables fast.
Gas choice changes the chemistry of the cut. Oxygen on carbon steel gives a hot, fast cut with a slightly oxidized edge. Air or nitrogen on stainless and aluminum avoids oxidation but needs more current. Shield gas is optional and mainly protects the nozzle.
Consumables and the Cost of a Bad Tip
The electrode and nozzle wear together. A worn electrode loses its hafnium insert, the arc starts wandering, and the kerf widens by 0.2–0.5 mm before you notice anything else. If your parts suddenly need more grinding, check the consumables first.
Track pierces, not hours. A 45 A electrode often lasts 600–1,000 pierces on 6 mm steel. A 130 A electrode may last 200–400 pierces on 25 mm plate. Writing the count on the torch body is a simple habit that pays for itself.
Piercing thick plate is the harshest moment of the whole cycle. The molten splash hits the nozzle face and erodes it. Raise the pierce height, shorten the pierce delay, and drill a pilot hole when the plate is over 20 mm.
Air quality matters more than most shops admit. Moisture and oil in the line cause a sputtering arc, which shows up as a rough cut face and short tip life. A refrigerated dryer and a coalescing filter are cheap compared to a week of scrapped parts.
Where Plasma Reaches Its Limit
Plasma holds a comfortable band up to about 50 mm on carbon steel, less on stainless and aluminum. Past that, the bevel grows and the heat-affected zone widens until the edge needs heavy cleanup. Oxy-fuel or waterjet becomes the better tool.
The heat-affected zone is the hidden cost. On 6 mm steel it may be 0.2–0.5 mm deep. On 25 mm plate it can reach 1.5 mm. If the part is a bearing seat or a sealing face, that softened layer has to be machined away.
Hole quality is the weakest point. Plasma rarely holds a hole under 1.5× the plate thickness without taper. A 10 mm hole in 12 mm plate will show noticeable taper. If the hole is a bolt clearance, fine. If it is a dowel bore, drill or mill it after.
That is the honest boundary of plasma CNC cutting proficiency: it is a fast, low-cost way to make a near-net blank, not a finishing process. Plan the secondary operation before you nest the part, not after.
From Plasma Blank to Machined Part
Most plasma-cut parts in our shop are not finished parts. They are blanks that go straight to a 3-axis or 5-axis mill for the features that carry tolerance. The plasma step removes 80–90% of the material at a fraction of the milling cost.
Leave 1.0–2.0 mm of stock on any face that will be milled. On 6 mm plate, 1.0 mm is enough. On 25 mm plate, allow 2.0 mm because the bevel and heat-affected layer cut deeper than the visible edge suggests.
Watch the datums. Plasma edges are not reliable datums. Pick a milled face or a drilled hole as your primary datum, and let the plasma edge float. Clamping on a rough plasma edge can pull the part out of square by a few tenths.
Check the heat-affected zone before finishing. If the drawing calls for Ra 0.8–1.6 μm on a face that was plasma cut, take a roughing pass deep enough to clear the softened layer. Then finish. Skipping that step leaves a gummy surface that never polishes clean.
Plasma vs Milling vs Waterjet: Pick by Requirement
Typical values for steel plate. Actual numbers depend on thickness and machine.
| Requirement | Plasma | CNC milling | Waterjet |
|---|---|---|---|
| Edge tolerance | ±0.2–0.5 mm | ±0.005 mm | ±0.1 mm |
| Heat-affected zone | 0.2–1.5 mm | None | None |
| Cut thickness | 3–50 mm | Up to 4,000 mm travel | Up to 150 mm |
| Hole taper | Noticeable | None | Slight |
| Setup cost | Low | Medium | Medium |
| Best for | Blanks and brackets | Finished features | Heat-sensitive plate |
| Secondary op | Often needed | Usually none | Deburr only |
The Verdict
If the edge only has to fit, plasma is the cheapest and fastest route. If the edge has to seal, bear, or slide, plasma the blank and mill the feature. There is no version of plasma CNC cutting proficiency that skips the second operation.
Plasma Cutting Questions Engineers Ask
Can plasma cut aluminum and stainless as well as carbon steel?
Yes, but the gas changes. Carbon steel runs on oxygen for a fast, hot cut. Stainless and aluminum run on air or nitrogen to avoid oxidation, and they need more current for the same thickness. Aluminum also reflects heat, so the kerf tends to be wider.
Expect the edge on stainless to be a little rougher than on mild steel. If the part will be welded, grind or machine the edge first, because the plasma-cut oxide layer can cause porosity in the weld.
How thick can plasma cut before the edge becomes unusable?
On a well-maintained 130 A machine, carbon steel up to about 32 mm cuts cleanly, and 50 mm is the practical ceiling. Stainless and aluminum top out lower, often around 25–30 mm, before the bevel and dross make cleanup expensive.
Past those limits the bevel angle grows to 5° or more and the heat-affected zone deepens. At that point oxy-fuel or waterjet usually costs less per finished part.
Why does my kerf get wider over a shift?
Consumable wear is the usual answer. The nozzle orifice erodes as the arc runs, and a nozzle that started at 1.2 mm can open to 1.6 mm in a few hundred pierces. A wider orifice means a wider kerf and a softer arc.
Check the electrode too. A worn hafnium insert makes the arc wander, which shows as an uneven cut face. Replace both parts as a set; changing only the nozzle rarely fixes it.
Should I plasma cut holes or drill them?
For bolt clearance, plasma holes are fine if the diameter is at least 1.5× the plate thickness. Below that ratio, taper becomes obvious and the hole may not pass a gauge.
For dowel bores, bearing seats, or anything with a tolerance callout, plasma the hole undersize and bore it on the mill. That is faster than fighting the torch and it holds ±0.005 mm when you need it.
What causes dross on the bottom edge?
Speed is the first thing to check. Too slow and the molten metal pools and freezes into a heavy blob. Too fast and the arc does not fully penetrate, leaving a brittle bead that brushes off.
If speed is right, look at standoff and gas flow. A standoff above 3 mm softens the arc, and low gas pressure cannot push the molten metal clear. Contaminated air from a wet line causes the same symptom.
Do I need heat treatment after plasma cutting?
Not usually for mild steel brackets and plates. The heat-affected zone is thin and the part still meets its structural role. For 4130 or 4140 parts that see fatigue loads, the softened layer may matter.
If the drawing specifies a hardness or a fatigue life, machine off the heat-affected layer or plan a stress relief. Talk to the shop before the parts are cut, not after.
Send Us the Blank, We Will Finish It
Upload your drawing and we will tell you which faces should be plasma cut and which need milling. Quotation and free DFM analysis within 12 hours.
12-hour quote100% inspectionNo minimum order
Follow GreatLight
We publish setup notes, tooling trials and inspection data from the factory floor.
