Plasma Cutting Troubleshooting

Plasma Cut Problems Quick Answer

Use this when scrap of the same alloy/thickness was dialed to calculator IPM and PSI, and the production cut still failed. The guide lists fixes; this table splits which failure you have before you change the wrong variable.

1. Top Spatter vs Soft Bottom Dross vs Hard Fin — Split Before You Touch Speed

Top spatter: Fine metallic spray or paint-rough texture on the top face along the kerf. The arc is not fully clearing the thickness — travel is too fast for amps, or amps are low for thickness. Exit arc often looks vertical or leans forward.

Soft bottom dross: Rounded, globby beads hanging under the cut. Easy to chip on a slightly slow cut; welded-on if very slow. Exit arc trails more than ~15° or hangs lazy. This is the classic "too slow for amps" tell.

Hard fine bottom fin: Thin bonded blade along the bottom edge — harder to knock off than soft beads. Usually high speed + high amps pushing swirl residue to the bottom, often with PSI on the high side of the machine range.

Field differential: Top spray ≠ bottom beads. Slowing down for top spatter causes soft bottom dross. Speeding up for soft bottom dross causes top spray. Hard fin needs a small speed reduction and PSI check — not a big IPM jump.

First fix order: (1) Confirm tip/electrode not worn (oval orifice mimics every dross type). (2) Match the tell: top → −5–10 IPM or +amps; soft bottom → +5–10 IPM; hard fin → −a little speed and verify regulator PSI under load. (3) Re-run the Plasma Cutting Calculator only after consumables and air are ruled out — calculator numbers assume dry air and a sharp tip.

2. Mid-Cut Arc Dropout — CFM Sag vs Speed vs Surface Transfer

What it looks like: Arc extinguishes and restarts, leaving a notch or double-start scar. Common on long CNC nests or thick plate after a short scrap test looked fine.

Fork A — air volume / pressure sag: Tank pressure starts OK; at-machine regulator falls mid-cut; compressor never recovers before the next pierce. Short scrap works; long production cut fails. Fix: size CFM correctly (see compressor guide), shorten hose runs, fix leaks, raise tank cut-in, or lower duty (pause between pierces). Do not chase IPM until the gauge holds under a continuous 20–30 second cut.

Fork B — travel too fast for transfer: Gauge steady, but arc blows out at thick sections, corners, or rust bands. Fix: slow 5–10 IPM; ensure full transfer before moving after pierce; clean start points on painted/rusty plate.

Fork C — pilot / transfer timeout on dirty surface: Arc never fully transfers at pierce; machine times out. Fix: grind bright at start, increase pierce standoff slightly, use pilot-arc mode if equipped.

First measurement: Watch the machine regulator (not tank gauge) during a continuous cut. Falling PSI = Fork A. Steady PSI = Fork B or C. Pair with Air Compressor Sizing for Plasma and Air Compressor Troubleshooting when Fork A wins.

3. Excessive Bevel & Wide Kerf — Torch Angle vs Worn Tip vs Standoff

Normal: 1–3° bevel is typical air-plasma geometry (kerf slightly wider at top).

Rejectable lean (>3–5°): Torch not square to work; standoff too high; tip orifice oval/enlarged spreading the arc; CNC cut direction fighting swirl (one side of part always fat).

Wide / uneven kerf: Micrometer across scrap kerf exceeds tip rating; wanders left-right with electrode pit wear. Settings that looked right on the calculator cannot compensate for an oval orifice.

First fix: (1) Square the torch — use a known-square scrap edge as visual reference. (2) Drop standoff toward drag-shield or 1/8" handheld / THC set-point. (3) Inspect tip orifice with light; replace tip+electrode as a pair if oval. (4) Only then re-check calculator IPM. For CNC bevel that flips with cut direction, follow OEM preferred direction — do not invent a new amperage table.

4. Pierce Blowback vs Tip Life — Edge Start vs Tilt vs Height

What it looks like: Molten metal sprays up into the tip/shield on pierce; tips crater after a handful of starts; shield face peppered.

Cause fork: Center pierce on thick plate with low standoff; no pierce delay; moving before full penetration; worn tip already weak. Not the same as mid-cut dropout — blowback is a start event.

Fixes that actually save tips: Edge-start whenever the nest allows. Must pierce in-plate: tilt 30–45°, complete pierce, then bring torch vertical. On ≥1/4", raise pierce height to 3/8"–1/2" then return to cut height. Use machine pierce delay / pierce mode. Calculator pierce time is a starting estimate — watch slag stop blowing upward before travel.

Guide-level pierce notes stay short; this page is the decision tree when tip spend is the actual failure mode.

5. Calculator Settings Right, Cut Still Wild — Moisture & Filter Path

What it looks like: Fresh tip lasts minutes; arc sputters; cut face striates even at guide IPM; water in filter bowl or icing at regulator.

Why settings pages fail here: Amps/IPM/PSI tables assume dry, filtered air. Moisture destroys hafnium inserts and makes every dross/bevel symptom appear randomly. Changing speed cannot fix wet air.

First fix order: (1) Drain tank and dryer. (2) Replace coalescing/desiccant elements on schedule. (3) Confirm filter is between compressor and plasma — not after a long wet hose. (4) Run a dry-air scrap test with new tip+electrode before touching calculator values again. Shop compressor diagnosis: Air Compressor Troubleshooting.

6. Display Amps vs Cut Heat — Duty Cycle, Leads, Input Power

What it looks like: Machine set to calculator amps feels cold (won't sever) or hotter than scrap calibration. More common on multiprocess units and long extension cords.

Likely causes: Thermal foldback after long nests; undersized input circuit / generator sag; poor work clamp / dirty return; wrong consumable amp class for the tip installed (40 A tip on 60 A set).

Fixes: Verify tip amp rating matches setpoint. Clamp-meter the input circuit if breakers trip or lights dim mid-cut. Improve work clamp to clean metal. Let the unit cool if duty-cycle lamp is on. Re-derive from the Plasma Cutting Calculator and Settings Chart after hardware is confirmed — not while the machine is thermally limiting.