How to Weld Thin Metal With an Arc Welder: Heat Control That Actually Works

A customer once brought me a hand-fabricated steel toolbox lid, two sheets of 20 gauge mild steel rolled and welded along a seam, and asked why it looked like a potato chip. The welds themselves were not terrible — decent penetration, no holes, not much spatter. But the panel had oiled into a gentle curve that no amount of hammering was going to talk out of. “I turned the amps down,” he told me, genuinely confused. “Everyone says turn the amps down on thin metal.”

He had followed half of the right advice. Turning the amperage down is real advice — and it is also where most welders stop looking, which is a problem, because amperage is not the number that decides whether a thin panel survives. Heat input per inch of travel does, and that number is a calculation you can do before you ever strike an arc.

This article is about that calculation and everything that hangs off it: how thin is actually thin (with a gauge table that is correct, because most are not), where stick welding genuinely stops being sensible, the techniques that keep a panel flat, and when to stop fighting the machine and switch processes instead.

Short answer: Thin metal is a heat input problem, not an amperage problem. Heat input = (volts × amps × 60) ÷ (travel speed in ipm × 1000). Travel speed is your biggest lever — double it and you halve the heat. Use the smallest rod that holds an arc, never weave, put a copper backing bar behind the joint, tack before you weld, and back-step. Below about 16 gauge, put the stick welder down and use MIG or TIG.

What Thickness Is Actually Thin?

Welding content invents a minimum thickness for arc welding. There is no such limit in any code, AWS specification, or manufacturer document — we checked. What exists instead is gauge-to-inch conversion, and a lot of it is quoted wrong.

Steel sheet uses the Mild Steel Gauge (MSG) system, which is a weight table, not a mathematical curve: thickness in inches equals weight in pounds per square foot divided by 41.82. Aluminum, brass and copper use Birmingham Wire Gauge or AWG, where thickness equals 0.005 × 92^{(36 − gauge)} ÷ 39. The two systems give different answers for the same gauge number, which is why you always name the metal alongside the gauge.

Common errors worth knowing so you can spot them: 22 gauge is sometimes quoted as 0.0275 inches, which matches no standard table, and 26 gauge is quoted as 0.0201 inches — which is actually the value for 24 gauge aluminum. And “16 gauge” spans 0.0508 to 0.0635 inches depending on the material. For a machine reference point, Miller’s Maxstar 161 spec sheet lists a minimum material thickness of 0.020 inch for steel — a machine rating rather than a process limit, but the only published hard number of its kind.

Where Stick Welding Stops Being Sensible

Lincoln Electric’s Stick Electrode Welding Guide devotes its sheet metal section to 18 through 12 gauge. Not 20, not 22, not 24. It also states that overhead welding of 18 gauge and thinner is not recommended. That is a manufacturer’s ceiling, and it excludes most of the material people actually mean when they say thin.

Why the arc fails at low amperage

At low current the electrode cannot maintain a coherent molten tip. Arc length becomes uncontrolled, the arc wanders or extinguishes, and the electrode sticks to the work. Miller’s own SMAW guidance lists excessive heat input as a cause of burn-through and amperage too low as a cause of sticking — the same problem arriving from both directions, which is the trap: too hot and you burn through, too cold and you stick and gouge.

The minimum output problem, in numbers

The typical floor is 20 to 25 amps, with 10 amps at the low end. Welding 24 gauge steel needs roughly 20 to 30 amps — which puts you at the absolute bottom of the machine’s range, exactly where arc stability is worst. A GTAW machine of that same class reaches 5 amps. That gap, not opinion or skill, is the entire reason TIG wins on thin metal.

So the honest framing of this topic is not “here is how to make SMAW work on thin metal.” It is: here is where SMAW genuinely reaches, here is the heat control that gets you there, and here is when to switch. If your thin-metal job happens to be an exhaust system, our stick welding exhaust pipe guide walks through the same physics at 16 through 20 gauge.

Heat Input Is the Whole Subject

Distortion and burn-through are not caused by amperage. They are caused by heat input per inch of travel — a number you can calculate before the arc is lit:

H = (V × A × 60) ÷ (travel speed in ipm × 1000) kilojoules per inch.

Effective heat input at the workpiece is arc energy multiplied by an efficiency factor — and this is where three different numbers get conflated into one wrong answer.

Efficiency, and the trap

Arc (thermal) efficiency applies to the heat input calculation: TWI and ISO 1011-1 give 0.8 for SMAW, GMAW and FCAW, and 0.6 for GTAW. Measured arc efficiency from published studies runs roughly 0.66–0.85 for SMAW, 0.84 for GMAW and 0.67 for GTAW. Deposition efficiency is a different animal entirely — the share of purchased filler that ends up as weld metal: about 65% for SMAW, 92–98% for GMAW, 87% for FCAW and 99% for GTAW.

Those last figures are the ones that get misused. “TIG is 99 percent efficient” is true for deposition and irrelevant to your heat problem. GTAW actually has the lowest arc efficiency of the four — tungsten and argon carry heat away. It wins on thin metal because the foot pedal meters heat in real time, not because it wastes less energy.

What actually moves the number

• Travel speed — linear in the denominator. Double it and heat input halves. This is the largest lever you have, and the one welders use least consciously.
• Amperage — linear in the numerator, but constrained by the machine minimum and arc stability.
• Arc length — sets voltage. On a constant-current machine you do not set voltage; you set arc length, and voltage follows. Miller’s formulation: voltage is governed by arc length and influenced by electrode diameter.

A worked example

Take 20 gauge mild steel (0.0359 in) with a 3/32 inch E6013 on a constant-current stick machine, running 75 amps and a 20 volt arc at 15 inches per minute:

H = (20 × 75 × 60) ÷ (15 × 1000) = 6.0 kJ/in arc energy, about 4.8 kJ/in effective at 0.8 efficiency.

Now travel at 25 inches per minute with everything else unchanged: H = (20 × 75 × 60) ÷ (25 × 1000) = 3.6 kJ/in, about 2.9 kJ/in effective. That is a 40 percent reduction in heat, bought with technique rather than a new machine. It is the single most useful thing to understand about welding thin metal.

Why Thin Metal Specifically Fails

The mechanism is not melting point — it is thermal mass per unit area. Heat travels into the plate at a rate set by conductivity. On a thick section it sinks away and the plate stays cool behind the puddle. On a thin panel the total thermal mass is small enough that the whole panel rises toward melting behind the arc, and the puddle falls through. Welders say it in plain language: on a thin panel there is nothing carrying the heat away.

Distortion follows the same physics. The weld zone expands, is restrained by cold surrounding metal, then contracts on cooling, leaving angular and longitudinal distortion. Because thin material has almost no stiffness, it oil-cans instead of holding shape. TWI notes that distortion is worst with low thermal conductivity, high thermal expansion and low elastic modulus — which describes thin steel precisely.

Techniques That Work on Thin Metal

Do not weave

The clearest rule in the subject, and often the opposite of what newcomers are told. Miller’s SMAW guidance: do not whip or weave, because the more time the arc spends in an area the hotter it becomes — travel in a straight line. Lincoln’s stick guide says it plainly: do not weave. Weaving means dwelling at the edges, where the pool already was, and on thin material that dwell is how you burn through. The technique people often confuse with weaving is whip and pause (stitch welding): advance past the puddle, pause to let it freeze, resume. That one is real, documented, and reduces heat in a way weaving never does.

Tacking before welding

TWI treats tack welding as the primary distortion control on thin material — tacks let you verify fit-up and let the part settle before you commit to the full joint. The sequence matters too: welding through from one end, starting at one end and back-stepping, or starting at the centre and back-stepping.

Copper backing bars

One of the most effective single interventions available. Copper’s conductivity is around 391 W/m·K against roughly 50 for steel, so it pulls heat out of the joint several times faster than the plate can. Finite element studies of chill bars report heat-affected zone width cut from 15 mm to 9.2 mm, distortion reduced, and residual stress down around 62 percent. This is the technique that turns an impossible thin weld into a routine one.

Tabs, backing, and the smallest rod that holds an arc

Welding tabs — sacrificial extensions on the joint ends — solve two problems at once: the crater crack, and letting the ends reach temperature without being in the weld zone. They are a documented remedy for crater cracking on aluminum. Run-on and run-off tabs do the same job differently. On stainless, TWI recommends dedicated backing, ideally chromium-plated copper, which releases more cleanly than bare copper. For the electrode itself: use the smallest rod that maintains a stable arc — Miller’s arc length guidance gives 1/16 inch for 1/16 and 3/32 inch electrodes — with a drag angle of roughly 10 to 20 degrees, and never dwell.

Two settings arguments, settled

E6013 penetration. Miller classifies E6012 as mild penetration and E6013 as medium, with a soft arc for clean new sheet metal; E6010 and E6011 are the deep-penetration rods. Descriptions of E6013 as “mild penetration” are a terminology slip — light to moderate is correct.

The one amp per thousandth rule. It exists, but it originates as a machine capacity and GMAW/TIG heuristic, not an SMAW setting standard. On thin material it overstates the requirement — thin metal needs less than one amp per thousandth, not more. Follow the electrode data sheet.

Distortion Control Sequences

Two named techniques do real work. Back-stepping deposits short adjacent lengths opposite the direction of travel, so each new segment lands on metal that has already contracted — spreading distortion instead of concentrating it. IIETA measurements on GTAW aluminum quantify it: angular distortion reduction from 6.5 to 8.94 percent with 75 mm and 50 mm stitches, up to 24.06 percent with a 37 mm stitch. Skip welding deposits evenly spaced segments all in the same direction and works about as well.

One correction that gets stated backwards constantly: weld toward the free end of the joint. TWI’s guidance is explicit — not from both ends toward the centre. And on clamping: restraint reduces distortion but raises residual stress, with Lincoln’s figures on low-carbon plate around 45,000 psi, so clamp what is critical dimensionally and let the rest move.

When to Switch Process

GMAW — the best general answer

Short-circuit MIG is the production standard for thin sheet. Miller: 75 percent argon with 25 percent CO₂ will not promote burn-through on thinner metals, while pure CO₂ gives deeper penetration but more spatter. Wire: 0.023 and 0.024 inch for light gauge, 0.030 inch at 18 gauge and thicker. Electrode positive, push technique, pulse if the machine has it.

GTAW — the best answer for the thinnest material

Miller is direct: short-circuit MIG, TIG and pulsed TIG provide the most control over heat. On the thinnest material the foot pedal is the deciding factor, because it modulates heat during the weld instead of committing to a fixed setting. Sharp ceriated tungsten, filler matched to the material — typically 0.020 inch up to 1/16 inch for sheet. Pulsed GTAW and pulsed MIG both beat their unpulsed versions on thin sheet for the same reason.

FCAW — the wrong answer here

Recommended constantly, and a bad idea. Lincoln: self-shielded flux-cored welding is not recommended for very fine materials, specifically gauge less than 20. Miller is blunter — do not attempt thin sheet with flux-cored wire, which puts more heat into the base metal. That is the whole argument: flux-cored adds heat, and heat is the problem you are solving. (Our dual shield welding explainer covers where FCAW does belong.)

Brazing — a legitimate answer

For very thin or heat-sensitive material, silver braze solidus runs about 1,145 to 1,250 °F, well below steel’s melting point, so the base metal barely notices. Be honest about the trade: a silver braze joint is considerably weaker than a fusion weld. Structural joint — weld it. Seal, lap, or a repair where distortion would ruin the part — brazing may be the better engineering answer.

Material-Specific Notes

Mild steel

The easiest case. E6013 for clean material, E6011 for dirty. Nothing exotic.

Aluminum

GTAW strongly preferred — ESAB’s guidance is TIG for thin material, tight geometry or demanding cosmetics, with anything under 2 mm going to TIG. The reason is thermal: aluminum conducts at 205–237 W/m·K against 43–54 for carbon steel and 14.4–16 for stainless, so heat rushes away from the puddle three to four times faster than on steel. Do not MIG thin aluminum: burn-back (wire melting into the contact tip) is the failure mode — ESAB notes a too-short wire extension dramatically increases the risk, and Miller calls burn-back the most common aluminum MIG defect. Aluminum MIG also needs DC electrode negative or AC, 100 percent argon, U-groove drive rolls and usually a spool gun. One correction: claims that DCEN reduces heat input on aluminum are backwards — DCEN puts more heat into the workpiece. Gas choices are covered in our guide to welding aluminum gas.

Stainless

Thin stainless needs a copper backing bar and low heat input. The two risks are burn-through and sensitisation — chromium carbide precipitation at the grain boundaries when the material sits in roughly the 425–870 °C (800–1,600 °F) range, which destroys corrosion resistance along the weld. Go fast, keep peak temperatures down, and use ER308L for 304. Heat tint colour is your evidence: straw is fine, blue is a warning, grey or black means you lost the HAZ.

Safety: The Thin-Metal Edition

Galvanized steel is the serious one

Arc welding galvanized steel is a genuinely hazardous operation and more common than it should be. The zinc coating vaporises into zinc oxide fume. OSHA’s permissible exposure limit is 5 mg/m³ as an eight-hour time-weighted average; NIOSH recommends the same as a TWA and 10 mg/m³ short-term. The measurement that matters: NIOSH found arc welding galvanized steel produced zinc oxide concentrations from 31 to 185 mg/m³, averaging 109 mg/m³ — roughly twenty times the limit, routinely. Metal fume fever follows four to twelve hours later: chills, fever, muscle ache, nausea, vomiting. Either grind the coating back several inches from the joint or do not arc-weld it, with local exhaust at the source either way. More on this in whether welding fumes are toxic.

Thin-metal specific hazards

Close standoff means intense radiant heat on your face and hands, and a small hot pool spatters more readily. And the hazard people forget: with no backing, hot slag falls through the joint onto the floor — and onto your feet. Boots, not trainers, and a backing plate for that reason alone. Correct helmet shade, flame-resistant clothing with no polyester, screens for anyone nearby.

Measuring Thickness Without a Caliper

Gauge markings are unreliable and calipers are cheap, so buy calipers. Beyond that: a feeler gauge set works on non-ferrous material and anything thin enough for the blades to fit; a sheet metal gauge plate works by sight and is only accurate for the material it was cut for; the weight method (pounds per square foot ÷ 41.82) is fast and accurate on a large flat steel piece; and a spark test reads carbon content rather than thickness. One trap: a magnet tells you nothing about aluminum and nothing useful about stainless in the condition you care about.

Joint Prep, Fit-Up, and Tack Spacing

Butt joints: leave a land of one to two times the material thickness and a gap of about one eighth to one quarter of the thickness. Do not bevel very thin material — a bevel on 20 gauge gives you two thin edges to burn through and a root gap you cannot fill without adding heat. T-joints: where thin metal fails hardest, because the vertical member soaks heat away and the horizontal edge melts before the puddle is ready. A copper backing bar solves this too. Lap joints: the strongest option for thin material, because the sheets distribute load across the full thickness — a good lap beats a poorly penetrated butt every time. Clean everything back to bright metal: oxide and mill scale cause lack of fusion, and on thin material there is no second chance after the arc starts.

The tension is real: closer tacks hold alignment better but concentrate heat. If tacks are visibly distorting the panel, spread them and improve the fixturing. The best fixturing is a full backing plate clamped up before you tack — not tack-weld-clamp-tack-weld.

Setting Up GMAW for Thin Material

• Drive rolls: knurled V-groove for flux-cored and larger solid wire; for aluminum, a knurled-in V-groove, because soft aluminum wire gets chewed by a knurled-out roll.
• Liner: keep it short — 12 to 15 feet is the practical maximum for a push gun. A damaged liner shows up as erratic feed and a wandering arc that looks like a machine fault.
• Gun angle: 15 to 20 degree drag, pushing the bead.
• Wire feed speed: you can calculate deposition rather than guess — lb/hr ≈ WFS(ipm) × 60 × (πd² ÷ 4) × 0.283. At 250 ipm on 0.035 inch wire that is about 4 lb/hr, right for thin sheet.
• Stickout: wire tip 1/2 to 3/4 inch back from the contact tip. Shorter builds spatter on the tip; longer loses shielding.
• Burn-off: the arc melts wire faster than the feeder pushes it, and the wire should never run out onto the work. Tip running back inside the gun means feed too low for the amperage.
• Gas: 75/25 argon-CO₂ at 15–20 cfh. Note the direction of the trade: less flow means more of the workpiece’s heat reaches the puddle (gas preheats it less), but less shielding means more porosity — on thin metal you balance those two, which is why shielding matters more here, not less.
• Voltage: lower voltage means a shorter arc and less heat. Use the shortest arc that stays stable.

Pulse Welding

Pulse GMAW alternates two currents: a peak that melts the droplet and delivers penetration, and a background that cools the puddle without adding much heat. The two controls therefore do different jobs — peak current controls penetration, while the peak-to-background balance controls bead profile and melt pool size. That is exactly why pulse is the best answer for thin material: it lowers average heat into the puddle without lowering your peak, so you keep penetration while limiting how much metal is molten at once. Syncronic-class machines set this from material and thickness; older pulse functions take real tuning. Pulse is not a substitute for good fit-up — it manages heat on an already well-prepared joint.

Troubleshooting

Porosity in particular has a longer suspect list than most welders check — before blaming the gas, rule out the metal itself with our full guide to porosity in welding.

Projects That Are Forgiving

A useful ladder, easiest first: 16 gauge steel brackets — thick, backed, forgiving, a good first project. 20 gauge steel boxes — learn burn-through and heat control without much punishment. 20 gauge aluminum pans — same heat control, much less forgiving metal. 26 gauge stovepipe trim — visually obvious defects, great practice, very thin. 24 gauge stainless backsplash — cosmetic, and stainless shows every heat mistake; do this last, not first.

Two things to avoid as a first project: 26 gauge aluminum with no backing, and any joint where the weld has to look perfect. Thin metal has a learning curve, and both of those punish you for being new. If you are earlier in the journey than that, our welding tips for beginners covers the fundamentals these techniques build on.

Frequently Asked Questions

Can you arc weld thin metal?

Yes, down to roughly 18 to 16 gauge with stick welding on a machine with adequate low-end control. Below that the arc becomes unstable at the low amperage thin metal needs, and GMAW or GTAW is the right answer.

What is the minimum thickness for arc welding?

No such limit exists in any standard. It depends on your machine’s minimum output, the electrode, the joint design, and how much distortion you can tolerate. The only published figure of that kind is a machine rating: Miller’s Maxstar 161 lists 0.020 inch minimum for steel.

What is the best way to weld thin metal?

GMAW in short-circuit transfer with 75/25 argon-CO₂ for general work, GTAW with a foot pedal for the thinnest material or where appearance matters, and a copper backing bar on anything you care about. Very thin or heat-sensitive — consider brazing.

What is the best arc welding rod for thin metal?

E6013 for clean mild steel at the lower end of its published range — Miller rates it medium penetration with a soft arc for clean new sheet. E6011 only when the material is dirty. Never E7018, whose minimum amperage exceeds what thin sheet can tolerate.

Should I weave when welding thin metal?

No. Miller says do not whip or weave; Lincoln says do not weave. Stringer beads, or whip and pause if you need to manage heat.

Does welding thin metal cause distortion?

More than any other thickness, because thin material has almost no stiffness. Control it with a copper backing bar, tacks, back-stepping or skip welding, and welding toward the free end. Clamping reduces distortion but adds residual stress.

Is flux-cored welding good for thin metal?

No — both major manufacturers say so. Lincoln: FCAW-S is not recommended below 20 gauge. Miller: do not attempt thin sheet with flux-cored wire, because it puts more heat into the base metal.

Can I weld thin aluminum?

Yes, with GTAW — ESAB recommends TIG for anything under 2 mm. Avoid MIG on thin aluminum because of burn-back, where the wire melts back into the contact tip.

Wrapping Up

Thin metal welding is a heat input problem, and heat input is a calculation: volts times amps, divided by travel speed, times an efficiency factor. Once you can do that arithmetic in your head, you stop guessing with the amperage knob and start controlling the weld.

The practical sequence that works: smallest electrode that holds a stable arc, straight line with no weave, copper backing bar behind the joint if you can get one, tack first, and back-step or skip to manage distortion. Travel speed is your biggest lever — the customer’s warped toolbox lid from the start of this article would have been flat if the travel speed had gone up instead of the amperage coming down.

And know when to quit. Below about 16 gauge, a stick welder is fighting its own minimum output, and the answer is a MIG machine with pulse or a TIG machine with a foot pedal. Both are better tools for that job than any amount of technique on SMAW — the panel does not care how good you are.