Flux-core burn through
Six causes, in the order to check them — and only two of them are settings on the machine.
Why does flux-core burn through thin metal?
Flux-core burns through because it is a high-heat-input process with an aggressive, digging penetration profile, and thin steel has nowhere to put the heat. In the order worth checking: wire diameter too large, travel speed too slow, voltage too high for the thickness, a gap in the fit-up, no heat sink behind the joint, and running a continuous bead where a stitch was needed.
Note what is not at the top of that list: the machine. People blame the welder, and then buy a different welder, and blow holes in things with that one too. Four of the six causes below are not machine settings at all.
This page is the same diagnostic shape as flux-core welding spatter applied to a different symptom, and the two causes lists barely overlap — spatter is mostly polarity and voltage, burn-through is mostly heat input and geometry.
1. Wire diameter — check this first
This is the highest-leverage change available to you and the one most often skipped, because a part-used spool of 0.035 inch feels like something you should finish.
Wire diameter sets how much current the arc carries at a given feed rate. A thicker wire needs more current to melt, and that current goes into your workpiece. Dropping from 0.035 to 0.030 inch lowers heat input across the board without touching a dial.
The welding wire size chart sets out which diameter belongs with which thickness. For anything you would call sheet, 0.030 inch is the starting point, and there is a floor: self-shielded flux-cored wire does not come thin enough for genuine 22-gauge panel work, which is why MIG versus flux-core for body work sends you to gas for panels.
2. Travel speed — too slow, almost always
Heat input is amperage multiplied by voltage, divided by travel speed. Speed is the one variable in that equation under continuous control of your hand, and it is the one beginners get wrong in a consistent direction: too slow.
Going slowly feels careful. What it actually does is pile heat into one spot while the puddle grows past the point where surface tension can hold it, and then the bottom drops out. The fix is counter-intuitive — move faster, and accept a flatter, less built-up bead.
You can diagnose this without a gauge: a bead that gets wider and then holes near the end is almost always a speed problem, because heat accumulated as you went.
3. Voltage set for the metal you wish you had
Now the machine. Voltage controls arc length and, with it, how wide and hot the puddle runs. Set for a quarter inch and used on sixteenth-inch sheet, it will hole the sheet no matter how well you move.
Use the chart inside your machine's wire-compartment door, matched to the actual thickness in front of you rather than the thickest thing you own. Our flux-core settings chart is for when that door chart is missing, not as a replacement for it.
Change one step at a time, and only after you have ruled out wire diameter and travel speed. Chasing voltage while the real problem is a 0.035 inch wire is how people end up with a machine set nowhere useful.
4. The gap you did not think counted
A gap in the fit-up is a burn-through waiting to happen, and it is a geometry problem rather than a heat one.
Where two pieces touch, the arc has metal to melt and heat to conduct into. Where there is a gap, the arc has air — nothing to fuse to, nothing to carry heat away, and the edges of the gap melt back and widen it. A one-millimetre gap on thin steel reliably becomes a hole.
Clamp tighter. If the pieces genuinely will not meet, tack at intervals and work towards the gap from both sides rather than running straight at it. On an edge or an outside corner, a gap you can see light through is not weldable at this thickness — it needs re-cutting.
5. Nothing behind the joint to take the heat
This is the cheapest fix on the list and the one almost nobody uses on a first attempt.
Thick steel is its own heat sink; thin steel is not. Clamping a copper or aluminum backing bar behind the joint gives the heat somewhere to go and gives the puddle something to bridge to. Copper does not fuse to steel, so it supports the molten pool without becoming part of the weld.
A piece of thick copper bar or plate is enough. If you are repairing something where you cannot get behind it, that constraint is itself a reason to use a thinner wire and shorter bursts, because you have lost one of your heat controls.
6. Running a bead where a stitch was needed
On thin steel, a continuous bead is the wrong technique no matter how well set up the machine is. The accumulated heat of a long run is what defeats you, not the heat at any single moment.
Stitch instead: short bursts of half an inch to an inch, then stop and let it cool. Skip around the joint rather than working along it, so no region accumulates heat. On a long seam, work from the middle outwards in alternating directions.
This also controls distortion, which arrives before burn-through does and ruins the part just as effectively. MIG welding sheet metal covers the stitch technique in detail; it is identical for flux-core.
The diagnostic table
| What you are seeing | Most likely cause | Change this |
|---|---|---|
| Hole appears almost immediately on striking | Voltage far too high for the thickness, or 0.035 in wire on sheet | Drop to 0.030 in wire, then reduce voltage one step |
| Bead widens along its length, then holes near the end | Travel speed too slow — heat accumulating | Move faster, and stitch rather than run |
| Holes only at the start of each weld | Dwelling before moving off | Start moving as the arc establishes, not after |
| Holes only where the pieces do not quite meet | Fit-up gap | Clamp tighter, or tack and work towards the gap |
| Holes on an outside corner or edge | No material behind the arc to absorb heat | Add a copper backing bar, or reduce heat and stitch |
| Fine on flat, holes on vertical | Gravity pulling an over-large puddle | Reduce heat, shorten bursts, use an all-position wire |
| Nothing works and the metal is very thin | Flux-core is the wrong process for this gauge | Switch to gas MIG with a thinner solid wire |
When the answer is to stop
There is a thickness below which self-shielded flux-core is simply the wrong process, and no setting rescues it. The thinnest commonly available wire diameters are still coarse for 20-gauge and thinner panel work. If you are repairing body panels, the correct answer is gas MIG with a thin solid wire — what gas for MIG welding sheet metal covers the mix to pair with it.
What to look for if you are buying to fix this
A 0.030 inch spool before anything else. It is the cheapest and most effective change available.
A wire with published procedure tables, so your starting settings are documented rather than inferred.
A machine with fine control at the low end — either infinite adjustment or several closely spaced taps. The lowest setting matters more than the highest on thin work.
Clamps and a copper backing bar. Neither is exciting and both fix more burn-through than a new welder does.
What we would buy
Arc & Bead is an Amazon Associate. If you buy through a link on this page we may earn a commission, at no extra cost to you. We have not physically tested the machines on this page and we say so on How We Pick.
1. The wire to switch to
Lincoln Electric Innershield NR-211-MP flux-cored wire
E71T-11 in 0.030 in, all-position, with published procedure tables — a documented starting point rather than a guess.
Lincoln publishes NR-211-MP as E71T-11, all-position, DCEN, in 0.030 and 0.035 inch, with a maximum plate thickness of 5/16 inch for 0.045 inch and smaller diameters.
Two reasons it belongs on a burn-through page. First, 0.030 inch is the diameter to be running on thin material, and switching down from 0.035 inch is the most effective single change on this entire list. Second, Lincoln publishes full procedure tables for it, so you have a documented voltage and wire-speed starting point for your thickness instead of working from a general chart.
All-position rating matters once you are repairing something in place, where you cannot lay the work flat and gravity is working against the puddle.
It costs noticeably more per pound than unbranded E71T-GS, and it still produces slag and spatter — that is the process, not the wire.
What works
- Lincoln publishes full procedure tables for it, so there is a documented starting setting for every thickness
- All-position rated — it will run a vertical fillet, not just a flat bead
- Available in 1 lb, 5 lb and 10 lb spools, so a small machine does not need a big spool
What does not
- Noticeably more expensive per pound than unbranded E71T-GS wire
- Like every self-shielded wire it produces slag and more spatter than a gas process
2. A machine with somewhere to go
Forney Easy Weld 261 140 FC-i
Infinite voltage and wire-feed adjustment on 120 V, so you can nudge heat rather than jump between taps.
Forney publishes this as a 120 V self-shielded flux-core machine with infinite voltage and wire-feed adjustment. On thin material, infinite adjustment is genuinely useful: tapped machines sometimes have no setting between too cold and too hot for 20 gauge, and a continuous dial gives you somewhere to sit.
The flip side, which is worth knowing while you are diagnosing: infinite dials are harder to return to. Mark the position that worked with a paint pen, because "about there" is not a setting you can reproduce next week.
What works
- 30% duty cycle at 90 A is unusually generous for a 120 V machine in this class
- Infinite voltage and wire-speed control rather than four fixed taps, so you can actually dial a setting in
- 19 lb — genuinely carryable to the job instead of the job coming to it
- No gas bottle, no regulator, no second purchase before your first weld
What does not
- Flux-core only. It will never run gas, so it cannot become your Level 2 machine
- Every weld leaves slag that has to be chipped and wire-brushed off
- Needs a 20 A circuit — a 15 A garage lighting circuit will trip at higher settings
3. The practice wire
Blue Demon E71TGS flux-cored wire
Cheap enough per pound that learning heat control on scrap stops being an expense.
Blue Demon publishes this as E71T-GS in 0.030 and 0.035 inch, DCEN. Burn-through is fixed by practice more than by purchases, and practice means spools. A cheap wire makes the twenty test beads this actually takes feel reasonable.
Two things to hold in mind while you use it. It runs dirtier than NR-211-MP, so a messy bead may be the wire rather than your hands. And it is single-pass rated, so it is for learning and for single-run joints — flux-core versus MIG strength covers why that restriction matters.
What works
- Cheap enough per pound that practicing on scrap does not feel expensive
- Widely stocked, so a spool is rarely more than two days away
What does not
- E71T-GS is single-pass rated — it is not the wire for a multi-pass structural joint
- Runs dirtier than NR-211-MP, which matters when you are also learning to read your own bead
Who this is wrong for: Multi-pass or structural joints. Its own classification rules that out, however good the bead looks.
Questions people actually ask
Why does my flux-core welder keep burning through?
Because flux-core is a high-heat-input process with an aggressive penetration profile, and thin steel has nowhere to put the heat. In order of likelihood: wire diameter too large, travel speed too slow, voltage too high for the thickness, a fit-up gap, no backing behind the joint, and running a continuous bead instead of stitching.
Four of those six are not machine settings, which is why buying a different welder rarely solves it.
What wire size stops flux-core burning through?
Drop to 0.030 inch. Wire diameter sets how much current the arc carries at a given feed rate, so a thinner wire lowers heat input without touching a dial - it is the highest-leverage single change available.
There is a floor, though. Self-shielded flux-cored wire is not made thin enough for genuine 20-gauge panel work.
Should I weld faster or slower to stop burning through?
Faster. Heat input is amperage times voltage divided by travel speed, so slowing down raises heat input - and going slowly is the mistake beginners make most consistently because it feels careful.
Accept a flatter, less built-up bead. A bead that widens along its length and then holes near the end is a travel speed problem almost every time.
Does a copper backing bar really stop burn-through?
Yes, and it is the cheapest fix on the list. Copper does not fuse to steel, so a copper bar clamped behind the joint supports the molten pool and conducts heat away without becoming part of the weld.
Thin steel is not its own heat sink the way plate is, so giving it one addresses the cause rather than the symptom.
Can you flux-core weld sheet metal at all?
Light-gauge sheet, with 0.030 inch wire, a stitch technique and a backing bar - sometimes, and it is hard work.
For genuine body panel thicknesses the honest answer is no: gas MIG with a thin solid wire and an argon-rich mix is the right process, and flux-core's advantages do not apply to panel work.
Why do I only burn through at the start of each weld?
You are dwelling - holding the gun still while the arc establishes and then starting to move. All that heat goes into one spot before any of it gets spread along the joint.
Start moving as the arc establishes rather than after it. It feels rushed and it is correct.
Is burn-through a sign my welder is too powerful?
Almost never. Burn-through is a heat-input problem, and heat input is wire diameter, voltage, travel speed and technique - all of which you control on any machine.
What does matter is how low and finely the machine adjusts. A welder with no usable setting between too cold and too hot for your material is a genuine limitation, but a high maximum output is not.
Sources
- Lincoln Electric — Innershield NR-211-MP published data: E71T-11, all-position, DCEN, diameters and procedure tables
- Lincoln Electric — self-shielded flux-cored welding troubleshooting and heat-input guidance
- Forney Industries — Easy Weld 261 140 FC-i published specifications: 120 V, infinite voltage and wire-feed adjustment
- Blue Demon — E71TGS published data: single-pass rated, 0.030 and 0.035 in, DCEN
- Your machine's own manual and door chart, which govern its recommended settings
Read next
- Excessive flux-core spatter, diagnosed in order
Flux-core spatters by design. The six causes of excessive spatter, in order of likelihood.
- Flux-core welding settings chart
Starting settings by thickness, and a symptom-by-symptom guide to fixing a bad bead.
- Welding wire size chart
Which diameter for which thickness, and why the wrong wire looks like a machine fault.
- MIG welding sheet metal without burning through
Stitch welding, heat management and copper backing — the techniques thin sheet actually needs.
- Flux-core vs MIG: strength and penetration
Same 70 ksi class on both wires - the real divide is single-pass rating, plate limits and penetration profile.
- Level 1 · Flux-Core: one machine, one wall socket, no gas bottle
If you have never struck an arc, this is your page. One machine, a standard household outlet, and nothing else to buy before your first weld.
Back to Level 1 · Flux-Core, or start over at Find Your Level.
Ryder M.
Ryder M. is the hands-on owner behind Magoo Media Group. He writes Arc & Bead's buying guides and picks every machine on the site. What every recommendation is based on — and what it is not based on — is set out on How We Pick.