MIG Welding Faults: Causes, Fixes, and Prevention

The bead came out shiny and flat, and the welder was already reaching for the grinder. Then the inspector ran a dye penetrant on it and a bright orange streak crawled straight out of the middle of the pass. Nobody had checked anything — the fault had simply been hiding in plain sight. Almost every MIG welding fault looks fine until something forces you to actually look for it.

What MIG Welding Faults Are, and Why the Terminology Matters

Most people use fault, defect, and discontinuity interchangeably. In a hobby garage that rarely matters. In a fabrication shop working to a code, the same words can mean the difference between a signed-off weld and a rejected one.

Here is the short version. A MIG welding fault is anything in the finished weld that was not there when you started. Officially, MIG is GMAW, or Gas Metal Arc Welding: a shielded metal arc process where a continuously fed consumable electrode and a shielding gas combine to make the joint. That process produces certain imperfections more often than others, and the classification system lets you name them and fix them.

Discontinuity, Defect, Indication: The Three Words That Matter

Michigan Tech training material draws a clean line between these terms. A discontinuity is any deviation from the ideal weld geometry: an irregular surface, an unexpected bulge, a void in the metal. A defect is a discontinuity that exceeds the acceptance criteria of the governing code or specification. An indication is what a nondestructive examination method flags during inspection, and is not necessarily a defect at all.

Read that middle definition twice: a fault does not become a defect until a rule says so. The same visible pore can be acceptable in one application and a rejection in another.

Consider a decorative gate, statically loaded, where the owner cares about how it looks from the driveway. Now consider a bridge component under cyclic loading, where fatigue cracks initiate at stress concentrations and a small pore becomes a crack starter. Identical geometry, opposite verdicts.

The Standard Fault Catalogue

Michigan Tech’s training materials lay out the classic catalogue of GMAW imperfections, and it is essentially the map of the rest of this article:

• Porosity — gas pockets trapped in the weld metal
• Spatter — droplets thrown off the arc, stuck to the bead or base metal
• Undercut — a groove melted into the base metal at the weld toe
• Lack of fusion — weld metal that did not bond with the base metal
• Cold roll — incomplete fusion at the toe, leaving the bead loose on the surface
• Arc strikes — arcs struck outside the joint, generally on the work
• Inclusions — oxide or flux trapped in the weld
• Cracking — the one you always feel in your stomach when you see it

Beyond that list sit incomplete penetration, burn-through, wire whiskers, bead shape problems, arc blow, and bird-nesting in the wire feeder.

The Four Root-Cause Families

Every fault in that catalogue traces back to one of four families, a framework that turns troubleshooting from guesswork into a process.

Root-Cause Family	What It Covers	Typical Faults
Gas and shielding	Gas coverage, flow rate, leaks, wind, nozzle condition	Porosity, oxidation, poor bead color
Parameters and heat input	Voltage, amperage, wire feed speed, travel speed, arc length	Undercut, burn-through, lack of fusion, spatter
Technique and workmanship	Gun angle, travel speed, arc placement, joint prep	Cold lap, incomplete penetration, arc strikes
Material and equipment	Base metal and wire condition, machine and gun wear	Cracking, bird-nesting, erratic arc

The families overlap. A bead that looks oxidized and full of holes might be a gas problem, but it could also be dirty base metal or a nozzle set too far back, which is a technique problem. That is why diagnostic order matters.

How Do You Actually Fix a MIG Welding Fault?

This is the answer most people are searching for, and it is worth stating plainly: the fix is rarely exotic.

1. Identify the symptom. Look at the weld, then run dye penetrant or a close visual read on a coupon. Name what you see: porosity, undercut, lack of fusion, spatter.
2. Work backward to the cause: is the shielding gas right and covering the arc, are your parameters delivering the right heat input, is your technique sound, is your material and equipment in condition?
3. Change one thing at a time, re-test, and write down what changed and what happened. That log is what makes you faster next month.
4. Verify under realistic conditions. A fault gone from a ten-minute coupon can still show up in a long production weld that heats up differently.

Why the Terminology Matters in Real Work

Getting this language right is not pedantry. When a shop claims a job has “no defects,” that means nothing; when it claims the welds meet the acceptance criteria of a specific code, that means something a third party can verify. Codes such as AWS D1.1 for structural steel and ASME BPVC Section IX for procedure and welder qualification define what counts as acceptable, and inspection exists to measure against those definitions. A Certified Welding Inspector is effectively translating this vocabulary into a pass or fail decision.

That vocabulary also speeds up your own troubleshooting. Once you can name the symptom, you can search for the specific fault, compare notes with other welders, and read the right troubleshooting charts from Miller’s MIG troubleshooting guide or ESAB’s defect resolution guide instead of guessing.

One last note before the deep dives. If you are new to GMAW or moving to a new material, the fundamentals matter more than any troubleshooting chart. Getting comfortable with gas flow, arc length, and gun angle on a simple coupon is what makes the rest of this article click. Aluminum, for instance, behaves nothing like mild steel, and our complete guide to aluminum MIG welding covers the setup that prevents an entirely separate set of faults.

Porosity: Diagnosing and Eliminating the Most Common MIG Fault

Porosity shows up in more than one way, and each version tells you something different. Sometimes it is obvious: a cluster of tiny craters across the bead crown, like orange peel. Sometimes it hides. You get a bead that passes a glance, a grinder, and dye penetrant check, then a pressure test finds gas trapped mid-pass. That is why welding troubleshooting starts with where the bubbles are, not just that they exist.

What is the MIG fault called when gas gets trapped in the bead? Porosity. It is a discontinuity: a gas bubble that formed in the molten pool and failed to escape before the metal solidified. ESAB lists porosity among the most common MIG welding defects, blaming atmospheric air getting into the weld zone or contamination in the joint such as rust, paint, grease, or dirt. Nearly every case traces to four causes: not enough shielding gas, gas not reaching the joint, dirty metal, or hydrogen in the pool.

Surface porosity vs. subsurface porosity

Surface porosity is visible: open holes or craters on the crown, pinholes near the toes, or a rough, cratered texture. It forms when gas is still leaving the pool as it freezes, meaning the gas arrived too late, too thin, or too dirty.

Subsurface porosity is the sneaky version: bubbles formed, got trapped under solid metal, and never surfaced. The bead can look smooth, slightly bulging, and even sound flat when you tap it. Finding it needs a method that reaches inside, such as dye penetrant testing, radiography, or in some cases ultrasonic. It is more dangerous, because it can cut the load-carrying cross-section of the joint without changing the outside appearance.

What you see	What it usually means	First place to look
Open craters, pinholes on crown	Gas enters pool late	Nozzle size, tip recess, wind
Pinholes at the bead toes	Poor coverage at pool edges	Nozzle, stickout, angle
Beefy, blistered surface	Gas trapped under the cap	Moisture, wet wire, dirty metal
Passes dye check, fails pressure	Internal porosity in the root	Hydrogen, wet cylinder, oil

The ranked diagnostic checklist

Work down this list in order. Gas-side problems are far more common than parameter problems, and they take seconds to check.

1. Shielding gas coverage. Gas must envelope the arc and hot pool. Miller calls inadequate coverage a leading cause of porosity.
2. Flow rate. ESAB gives 25–40 CFH for most MIG work. Too little lets air in; too much drags air in with turbulence.
3. Leaks and flow check. Inspect hose, regulator, and gun body for leaks, and confirm flow at the gun rather than at the regulator.
4. Drafts and wind. An open shop door spoils a bead; Miller advises blocking off the welding area with a curtain or windbreak.
5. Nozzle. Keep it clean and large enough to shield the whole pool; a spatter-caked nozzle starves the pool edges.
6. Contact tip recess and stickout. Too deep buries the wire inside the nozzle and shielding collapses; keep stickout within 1/2 inch.
7. Gun angle. Miller calls out excessive gun angle because it pulls the envelope off the pool. Keep the nozzle over the work.
8. Base metal and consumables. Clean back to bright metal; check for wet cylinders, oily wire, and damp storage.

Three different gases, three different problems

Identifying the gas mechanism tells you which fix will actually work.

Nitrogen and oxygen pickup happens when the envelope fails and air enters the pool: wrong flow, wrong nozzle size, a leak, a draft, or excessive stickout. It shows as widely scattered pinholes, often at the toes.

Hydrogen pickup is sneakier. Rust, scale, paint, oil, moisture, and skin oils introduce hydrogen or nitrogen into the pool. Hydrogen solubility drops sharply as the metal solidifies, so the gas is rejected and trapped as fine internal porosity along the root and center of the bead. This is the fault that passes visual inspection and then fails a pressure test. The fix is cleaning, not gas adjustment.

CO₂ issues behave differently again. Pure CO₂ is standard for short-circuit transfer MIG and GMAW and gives a deeper, more stable pool than mixed gas. But it is heavier than air, so drafts knock it aside easily and it needs more flow for a wide bead. Turbulent flow can entrain air rather than exclude it. In shops that rent gas, cylinder mix-ups are common: a 75/25 blend used where pure CO₂ is specified changes penetration and spatter, sending the welder after a settings problem that is really a gas problem.

Step-by-step fix sequence

1. Clean the joint. Bright steel, no paint, rust, scale, oil, or moisture.
2. Check the gas path. Leaks at the hose, gun, and regulator; flow confirmed at the gun.
3. Set flow and service the gun. 25–40 CFH, clean nozzle, correct tip recess.
4. Fix environment and geometry. Screen drafts, nozzle close, stickout under 1/2 inch, gun square, cylinder at full pressure.
5. Re-test on a coupon. A short bead on clean scrap of the same thickness before production.

When flux-cored wire helps

Self-shielded and gas-shielded flux-cored wires carry a flux core that decomposes in the arc, giving shielding plus a cleaning action that burns away oxide and contamination as the bead forms. ESAB notes that flux-cored wire can provide improved cleaning action and reduce the risk of porosity compared with solid wire. That makes it the sensible answer on dirty, rusty, or oily material, outdoors in wind, or when you want to avoid the shielding gas setup. It is not a universal fix: flux-cored produces more fume and more heat than solid wire and still will not tolerate thick rust, paint, or a coating. Aluminum has its own oxide and porosity problems, covered in our guide to cleaning aluminum for welding.

When you have the cause pinned down, document it. Porosity that returns on the next job or the next cylinder is usually a leak, a bad spool, or a gas delivery problem that was papered over. A settings log recording cylinder, gas mix, flow rate, and liner condition per job catches the pattern fastest.

Lack of Fusion, Cold Lap, and Incomplete Penetration

These three look alike on the surface: a bead that seems fine but is not bonded to the metal underneath. Lack of fusion means no bond to the base metal or previous bead. Cold lap is weld metal that overlapped without melting in, bulging over the toes. Incomplete joint penetration means the prepared root edge was never consumed.

Lack of Fusion: The Weld That Never Bonded

Each points to different causes, and lack of fusion is the most serious: a structural defect hiding inside a bead that can look convincing. Miller’s troubleshooting guide lists the usual suspects: geometry, heat, and cleanliness.

Start with geometry, the one welders rarely question. Miller advises a gun angle of 0 to 15 degrees from vertical with the arc on the leading edge of the pool, which drives arc force into the joint. If the arc trails the puddle, heat lands on molten metal instead of the base metal. Counterintuitively, Miller notes you sometimes need to increase travel speed to hold the arc in position, since a slow travel speed lets the puddle grow.

Next, heat. Insufficient heat is another cause, remedied by raising voltage settings or wire feed speeds, which is the diagnosis once geometry and cleanliness are correct. Higher voltage with matching wire feed speed widens the bead and spreads heat into the joint edges.

Then cleanliness. ESAB’s guidance is blunt about the first step: clean the joint surfaces to remove oxide layers and contaminants, then raise heat input by adjusting voltage, wire feed speed, and travel speed, and keep a consistent arc length. Rust, mill scale, paint, oil, and moisture all act as a thermal barrier between the puddle and the base metal.

Cold Lap: The Bead That Overfills

Cold lap is lack of fusion’s cousin. The weld metal flows over the base metal but never melts into it, so the bead sits proud, overlapping both toes with a profile that looks fat, wide, and lumpy at the edges.

Because the puddle is not wetting into the base metal, the extra heat goes into the puddle itself, and cold lap becomes self-reinforcing: the bead gets bigger, the pool gets heavier, and the arc gets pushed further off the leading edge. Miller notes that increasing travel speed helps prevent cold lap, since the puddle gets less time to sit and spread. Cold lap also appears at slow travel speed on hot or dirty surfaces, or when the operator tries to fill a gap in one pass and pours metal on top.

Incomplete Joint Penetration: The Bevel That Was Never Consumed

A bevel or V-groove creates a root face, a root opening, and angled bevel faces. Penetration means the weld metal consumed that prepared root edge. Michigan Technological University’s welding defects material defines incomplete joint penetration as a condition in which “the prepared root edge of the bevel has not been consumed by the weld,” and defines lack of interbead fusion separately: the weld failing to fuse to previous beads.

Incomplete penetration usually means the heat went somewhere else: the arc was too long, so the puddle formed on the surface instead of reaching the root; travel speed was too fast for the thickness; the root face was too wide; or the wire was too heavy for the current to melt fast enough.

A quick field test is a bend test coupon on a single V-groove with a root face. A clean, ductile fracture through the weld metal means full penetration, while a flat, dull, unfused face at the bottom means the root was never consumed. For production work, that verification belongs to qualification and inspection, where ASME codes and standards come in.

How to Raise Heat Input Without Causing Burn-Through

Once you know the joint is heat-starved, the instinct is to turn up the heat. On material under about 1/8 inch, that puts you straight into burn-through, a different fault with a different repair strategy. Heat input is a budget across three dials: voltage, wire feed speed, and travel speed.

• Raise wire feed speed and voltage together in small, matched steps to widen the bead and add penetration, but keep travel speed honest.
• Hold the arc on the leading edge at a 0 to 15 degree gun angle and shorten the stickout so more heat lands on the joint.
• Use a smaller wire on thin material; 0.035 inch with more passes delivers heat without melting through the base.
• Watch the root. Add a backing bar or copper chill so you can push penetration without going through.

And clean the joint: remove the oxide, oil, paint, and mill scale before you touch a parameter. On aluminum that step is non-negotiable, and our guide on how to prep aluminum for welding covers the oxide film in detail.

Quick Diagnostic Table

What you see	Likely fault	Check first
Dark unfused line along one toe, or weld metal peels off the base	Lack of fusion	Gun angle, arc position, joint cleanliness
Bead overlaps both toes and looks fat and rounded	Cold lap	Travel speed, cleanliness
Flat unfused face on a break test	Incomplete penetration	Arc length, root face, wire

One last habit worth building: cut a section before you re-run the job. A cross section of a single V-groove tells you immediately whether you had penetration, fusion, or both, and it takes five minutes instead of an afternoon of guessing at settings.

Undercut and Burn-Through: Too Much Heat, Too Little Control

When it comes to MIG welding faults caused by excess heat, undercut and burn-through are two of the easiest to create, and two of the most costly to ignore. Both stem from putting too much energy into the joint, but they look entirely different. One erodes the base metal, the other blows straight through it. Knowing how to tell the two apart is the key to correcting them before they become a rejection on the job.

Undercut: The Notch That Eats Fatigue Life

Undercut is a groove melted into the base metal at the weld toe, the exact spot where the weld metal meets the untouched parent metal. Rather than blending in smoothly, the excessive heat has washed away a notch along that edge. This is far more serious than it looks. Undercut is a stress concentrator at the weld toe — it reduces the fatigue life of the joint far more than its size suggests, which is why codes limit it tightly.

The causes for this fault are usually technique or heat related. Some of the most common include excessive heat, too long an arc, wrong gun angle, too much travel speed, poor joint design, and arc blow. Arc blow, caused by magnetic forces deflecting the arc, produces uneven penetration and shallow welds and can also trigger porosity or incomplete fusion. It’s more typical with DC welding and at high currents on thick sections, so if your arc is suddenly wandering with no change to your hand placement, it’s worth checking your work lead routing.

For correction, follow ESAB’s guidance for undercutting: adjust welding parameters, primarily voltage, to achieve proper weld size; maintain a steady arc length and proper gun angles to prevent excessive melting; and ensure proper joint preparation and consistency. A good starting point is to drop your voltage slightly, shorten your arc length, and focus on directing your leading edge of the arc towards the joint to properly tie in the toe without gouging it.

Burn-Through: When There Isn’t Enough Metal Left

Burn-through is the polar opposite. Instead of eroding a groove, the weld puddle melts completely through the base metal, leaving an actual hole. Miller confirms burn-through is especially common on thin material under 1/8 inch (about 12 gauge) and that excessive heat is the main cause; the fixes are reducing voltage or wire feed speed and increasing travel speed, particularly on heat-prone materials such as thin aluminum.

This matches ESAB’s guidance for burn-through: use a lower voltage or reduce the wire feed speed to decrease the amperage, and increase travel speed to reduce total heat input. These small parameter changes are your first defense. If you’re working with applications such as aluminum boat welding, this becomes even more crucial, as thin aluminum dissipates heat unpredictably and can go from a stable puddle to a hole in seconds.

Reading The Bead To Know Which One You Have

The fastest way to diagnose these two faults is to look at the profile of your bead. Use this quick breakdown to tell them apart on the floor:

• Undercut: The weld crown may look acceptable, but you’ll see a sharp, washed out groove running along either toe of the weld. The base metal is missing, not penetrated through.
• Burn-Through: You’ll see a hole, a blown out crater, or a pass that has sagged through the backside of the metal. Oftentimes, you’ll see a ropey drop of weld metal hanging underneath your joint on thin sheet.

As a concrete starting fix, try this order of operations: shorten your arc, correct your gun angle to a more neutral or slight drag depending on your joint, reduce voltage or wire feed speed, and increase your travel speed. Verify with a quick test coupon before returning to your part. These two faults may both be products of too much heat, but with a trained eye and controlled parameters, they’re entirely preventable.

Spatter, Wire Whiskers, and Bad Bead Shape

These three faults sit on the outside of the joint, where you can see them without a dye penetrant or an X-ray, and they are often dismissed as cosmetic. They are not blemishes. They are visible evidence of something happening inside the arc, and a finished bead reads out your gas, your parameters, and your technique.

Spatter: the arc is throwing metal

Spatter is small droplets of molten metal that break off the pool and land on the bead or the plate. Miller’s troubleshooting guidance lists the usual suspects: insufficient shielding gas, dirty base material, contaminated or rusty wire, and voltage or travel speeds that are too high. Excessive wire stickout is another: a long arc wanders along the wire before it reaches the work.

The hardware counts too. Miller notes that a wrong size contact tip, a worn contact tip, or the wrong contact tip-to-nozzle recess can all cause excessive spatter. A worn tip has a ragged bore, and the wire arcs off-center the moment it leaves the tip.

Flux-cored wire follows different rules. Miller recommends welding self-shielded flux-cored wire using straight polarity (electrode negative) with a drag technique to minimize spatter, and notes that with flux-cored or metal-cored wire, low voltage can also produce excessive spatter, so raise the voltage as needed. For flux-cored filler on aluminum, our aluminum flux core welding guide covers the spool handling and feed behavior behind that spatter.

Wire whiskers: short lengths of wire left in the bead

Whiskers are short protruding lengths of welding wire stuck in the weld metal, characteristic of GMAW with continuously fed wire. The MTU training material on welding discontinuities and defects treats them as more than an annoyance: a wire fragment left in a joint is a hard inclusion that a pressure test or an inspector will find.

They come from four places: improper parameters, excessive current, technique issues, and contaminated wire. That means too much amperage for the wire diameter, a feed speed that does not match the setting, a gun that drifts as the pool advances, and wire carrying oil, rust, or drawing compound. The fix follows the same order: reduce current, confirm the feed speed, keep the gun steady and moving with the puddle, and inspect the spool before blaming the torch.

Shielding gas plays a role too: an erratic arc leaves wire ends standing proud rather than melting them into the bead. Our guide to MIG shielding gas covers flow rates and nozzle setup when spatter and whiskers appear together.

Bead shape: the profile that predicts other faults

Convex beads are high, rope-like welds. Miller notes they generally occur in flat and horizontal welding, but they can also show up in fillet welds when the parameters are too cold for the material. They normally come with poor fusion at the toes, since the cold bead has not wet out into the joint corners. Miller’s recommendation is to increase voltage.

Concave beads are the opposite: a hollowed-out, sunken profile. In flat or horizontal positions, a concave bead usually means voltage too high, wire feed speed too slow, or travel speed too fast. In vertical-down welding they are far more common, simply because the welder works against gravity, and the advice there is to lower the parameters so the pool is less fluid.

What you see	What it points to	Correction
Heavy spatter on the bead	Low gas flow, dirty metal, dirty wire, high voltage	Check gas and nozzle, clean the joint, inspect the spool
Ball of metal at the wire end	Stubbing from low voltage for the feed speed	Raise voltage, shorten stickout, check feed rollers
Wire stubs in the bead	Excess current, poor gun control, dirty wire	Lower amperage, steady the gun, replace spool
High rope-like convex bead	Cold parameters, lack of fusion at the toes	Increase voltage, re-check toe fusion
Hollow concave bead, flat or horizontal	Voltage too high, feed too slow, travel too fast	Lower voltage, raise feed speed, slow down
Concave bead running vertical-down	Gravity acting on a fluid pool	Lower parameters so the puddle stiffens

Bead shape leads. A convex bead usually arrives before lack of fusion does, and a concave bead usually arrives before undercut and burn-through do. Fix the profile and the root cause often goes with it.

Arc Blow, Arc Strikes, and Other Electrical Faults

Some MIG faults come from gas, parameters, or technique. Others come from the machine and the circuit around the work.

Arc Blow: When Magnetism Steals Your Arc

Arc blow happens when the magnetic field around the arc deflects the arc column off its path, giving poor penetration on one side, undercut or extra reinforcement on the other, and a wandering bead.

It is mainly a DC welding problem that worsens as thickness and current rise. Light gauge hides it; a 300 amp structural joint can push the arc off the bead.

The fix is work lead routing: run two work leads as one cable or in parallel rather than spreading them into separate paths.

• Move the work clamp opposite the joint or closer to the work.
• Route leads together and keep the loop tight; a big loop acts like a transformer.
• Fix continuity: dirty contacts, corroded lugs, and spatter under the clamp force current down an unwanted path.
• Consider AC on non-magnetic alloys; it cancels the deflection, and AC MIG is standard on aluminum and stainless.

Arc Strikes: The Fault That Is Always on the Welder

An arc strike is any arc made outside the joint. It is not a metallurgical defect and not a machine fault, but a process interruption. Michigan Tech’s welding discontinuity material lists arc strikes among the standard discontinuities and is blunt: they are 100% welder error.

A strike leaves a crater that must be ground out, can freeze the arc momentarily, and leaves spatter or slag. In code work, a strike outside the joint is rejectable under many specifications.

Prevent it by securing the work leads, cleaning the joint, and establishing the arc only on the joint, never on the plate edge or a tack weld. A poorly secured ground is a frequent companion: when it shifts mid-pass the welder restrikes elsewhere, leaving a cold joint too.

Erratic Arc, Stubbing, and Wire Feeding

An erratic arc shows as a bead that changes width and shape along its length, with irregular ripple and spatter. Causes: a worn drive roll, a kinked liner, a worn gun neck, a gas leak, or poor grounds.

Stubbing is the wire shorting against the contact tip or the work without a stable arc, heard as rapid popping. Causes: a worn contact tip, insufficient stickout, excess drive roll tension, the wrong tip size, and voltage too high for the feed speed.

• Replace the contact tip; check the gun neck for wear.
• Confirm the wire feeds without drag, drive roll tension at spec.
• Check the liner for kinks and debris; replace rather than clean.
• Verify gas flow and leaks at the regulator and hose.

Stubbing appears more often on aluminum because the wire is softer. The complete aluminum MIG guide covers the gun, liner, and spool gun differences.

Bird-nesting is the classic wire feed fault: the wire leaves the gun in a curly tangle instead of a straight feed. Causes are a worn drive roll, low drive roll pressure, a worn liner section, and a bent gun. Replace the rollers and liner and check tension; if nesting returns, suspect the gun body.

Liner wear is the quiet version of the same fault: rubbing wire and its debris cause drag, bird-nesting, erratic feeding, and unstable arc voltage. A worn liner sheds particles, and aluminum needs a dedicated liner because residue contaminates steel liners.

Ground Clamps, Cable Condition, and the Rules

Everything on the work side of the circuit affects weld quality. OSHA 1910.254 requires that cables with splices within 10 feet (3 m) of the holder not be used, and that the work lead be firmly attached to the work. A loose ground causes an unstable arc, an erratic bead, and a shock hazard. The standard also directs operators to free magnetic clamps of adherent spatter and to check for leaks of cooling water, shielding gas, or engine fuel; spatter in the clamp contact lets current take an unwanted path.

• Replace cables with cracked or flattened insulation instead of taping them.
• Clamp the ground to the work, not to a scaffold frame.

For broader equipment and lead length requirements, see OSHA 1910.252.

When a fault resists the obvious fixes, ESAB’s guide to resolving common MIG welding defects covers the same causes in sequence.

Diagnose the circuit first, then the settings. These faults are not random and rarely respond to a voltage change.

Material-Specific Faults: Aluminum, Stainless Steel, and Flux-Cored Wire

Same bead size, same gun angle, same amperage, and the aluminum came out porous while the mild steel came out clean. That is not bad luck. It is chemistry.

Aluminum ships covered in a thin, tenacious oxide layer that reforms in minutes. The oxide melts at a much higher temperature than the base metal, so rather than fusing it sits on the joint face as a barrier. Weld metal flows around it, giving you porosity under a healthy-looking bead or lack of fusion along the toes.

Joint prep is not optional. Brush with a stainless brush dedicated to aluminum or solvent clean with a dedicated product, then weld immediately. Our guide to prepping aluminum for welding covers the solvents, the timing, and why a dedicated aluminum brush matters.

Wire feeding is the next battlefield. Aluminum wire is softer and feeds less predictably, and it needs a dedicated spool gun or push-pull setup with its own liner, not the steel one. Sharing a liner leaves oxide and steel fragments behind that shower the next aluminum job with porosity. If the gun feeds smoothly for ten minutes then stutters, suspect the liner before the machine.

Hot cracking is the other aluminum signature. Aluminum shrinks more as it solidifies than steel, so a wide hot zone and high restraint pull the weld metal apart as it cools, usually at the toe or in the root pass. On structural aluminum MIG work, watch the amperage-to-travel-speed ratio harder than you would on steel.

Burn-through on thin aluminum is a heat-management problem. Miller notes thin aluminum as prone to heat buildup, where increasing travel speed helps prevent burn-through, because the longer the arc dwells on one spot the more heat soaks in. Many shops switch to AC, where the alternating half-cycle acts like a scrubber on the oxide, though modern constant-current power sources with high-frequency start and good pulse control handle aluminum well on straight DC. Our guide to aluminum flux-core welding covers the self-shielded option for gas-awkward jobs.

Stainless Steel: Heat Tint, Carbides, and the Fume You Cannot See

Stainless fails in two ways carbon steel never does. The first is visible: heat tint, the rainbow oxide film that forms as the bead cools, ranging from straw through blue to dark gray. It is not a fault itself, but a warning label, because the color tells you how hot the metal got and for how long, which maps directly to chromium carbide precipitation, or sensitization, at grain boundaries in the heat-affected zone. AWS standards often limit heat tint on pressure and sanitary work, and passivation is frequently required afterward to restore corrosion resistance.

The second failure is fume, not metal. Stainless welding converts chromium in the base metal and consumables into the hexavalent state during the arc process, producing hexavalent chromium fume, a confirmed carcinogen. NIOSH calls it one of the most hazardous substances a welder can encounter, which is why stainless work demands closed-loop extraction rather than a helmet and a fan in the corner of the shop.

Heat input is the practical lever. Keeping travel speed up and arc length short limits the time the base metal spends between roughly 800°F and 1,500°F, the range where chromium carbides form. Stainless also needs backside purge on many grades, using argon or an argon-carbon dioxide mix, at flow rates low enough to avoid turbulence but high enough to displace air. Too little purge gives oxidation and a compromised root; too much disturbs front-side shielding and invites porosity.

Flux-Cored Wire: Same Gun, Different Rules

Flux-cored changes polarity, technique, and the spatter equation. Most self-shielded flux-cored wire runs DC electrode positive, the opposite of the DCEN used for solid wire MIG, and the manufacturer’s polarity chart is not optional reading. Get it wrong and the wire feeds stiffly, voltage climbs, and the bead looks like burnt brush. Gas-shielded flux-cored and metal-cored wires may run either polarity depending on the product, so read the wire and the machine.

Drag technique, dragging the electrode along the joint rather than pushing it, holds the arc at the leading edge of the puddle and is standard on flat and horizontal fillets. Voltage is the primary spatter control: Miller confirms that low voltage produces excessive spatter with flux-cored or metal-cored wire and that voltage should be increased as needed. Low voltage also risks stubbing, where the wire touches the puddle and shorts out, building a ball of filler metal on the tip until it hits the work and spits a slug. That spatter is a signal, not a dirty-work problem: the arc length is too short for the voltage you are running.

Flux-cored also runs hotter and smokier than solid wire. The flux core decomposes into shielding gas inside the arc, so you get a visible fume plume and a higher heat input than a solid wire MIG at the same amperage. That extra heat helps on heavy structural sections and hurts on thin material, and the fume control rules that apply to stainless apply in spades. Our gas and cylinder safety guide covers the handling side, including shielding gas choices for aluminum and stainless.

Material	Dominant fault signature	First thing to check
Aluminum	Porosity, lack of fusion, hot cracking	Joint cleanliness, oxide removal timing, dedicated spool liner
Stainless steel	Heat tint, sensitization, root oxidation	Heat input and travel speed, backside purge, fume extraction
Flux-cored	Spatter, stubbing, erratic bead	Polarity, arc voltage, drag versus push technique

Material is the one variable you cannot adjust on the fly. Learn how each filler family changes the arc, the puddle, the cooling rate, and the fume, and the fault catalogue in this guide starts to make sense.

When a Fault Becomes a Defect: Inspection, Codes, and Acceptance Criteria

Fixing a MIG welding fault is a shop-floor problem. Deciding whether that fault is acceptable is a code problem. A bead that looks slightly convex to a fabricator may be fine under a bridge spec, while a bead that looks gorgeous to a hobbyist may be a rejection under a building code. The difference is not taste, it is a written specification.

Every code-backed job has a document that decides the argument: joint configurations, allowable defects, tolerance bands, test methods. The welder who made the joint does not get a vote, and neither does the foreman’s opinion of a good weld. If the weld misses the document, how it looks is irrelevant.

The Standards World

The American Welding Society has authored more than 350 standards on welding and joining practices, procedures, and inspection. For U.S. structural steel work, the governing documents include:

• AWS D1.1/D1.1M:2025-AMD1, Structural Welding Code, Steel. The default for buildings, towers, and general structural fabrication.
• AWS D1.2/D1.2M, Structural Welding Code, Aluminum. Aluminum welds carry different defect limits and often different inspection requirements because of heat sensitivity and crack susceptibility.
• AASHTO/AWS D1.5:2025, Bridge Welding Code, which sits alongside D1.1 for highway and bridge structures.
• ASME BPVC Section IX, Welding and Brazing Qualifications, which governs how procedures are written and how welders are qualified, not the defect limits themselves.

Keeping D1.1 and Section IX straight confuses people for years. Section IX answers “is this welder and this procedure qualified to do this work?” D1.1 answers “is this weld acceptable?” You need both. A perfectly executed weld made with an unqualified procedure still fails, and a fully qualified procedure producing a crater crack still fails.

For pressure vessels, boilers, and piping, the ASME Boiler and Pressure Vessel Code Section VIII sets the acceptance criteria that Section IX qualification supports.

Visual Acceptance Criteria in AWS D1.1

D1.1 does not use one master list of allowable defects. It sorts welds by type, then applies different limits to each category. That is why “what size undercut is allowed” has no single answer.

Joint Type	How D1.1 Treats It	Limit Focus
Fillet weld	Judged on leg dimensions, contour, toe condition	Undercut length and depth, convexity or concavity, throat size
Partial joint penetration (PJP) groove weld	Judged on required effective throat	Undercut, porosity, incomplete fusion in the effective throat
Complete joint penetration (CJP) groove weld	Must fully penetrate, then judged on soundness	Cracks, porosity, slag, lack of fusion, back-face condition

Cracks are treated differently from nearly everything else. A crack in any weld configuration is a rejectable discontinuity under D1.1, and where a code permits some porosity, it never permits a crack. That is why crack repair in structural welds is a mandatory procedure with its own limits on how many times a joint can be reworked, not a discretionary fix.

The Inspector

On a coded job there is usually a third party in the loop. The AWS Certified Welding Inspector (CWI) credential is valid for nine years, with renewal applications at years three and six after activation. CWIs are not the only route; many codes also accept a CWIP, a CAWI, or an inspector qualified directly to the applicable code.

The inspector’s job is not to like your weld. It is to read the spec, look at the weld the way the spec describes, and document what is found. A good inspector also tells you why something failed, which is the most useful part of the exercise.

The NDT Ladder

Visual inspection catches roughly 70 percent of weld problems, and the remaining 30 percent is where the expensive failures live. The American Society of Nondestructive Testing identifies the major methods as UT, RT, PT, MT, ET, and AE, each with a specific blind spot.

Method	Finds	Cannot Find
VT (visual)	Surface cracks, undercut, visible porosity, bad profile	Anything buried inside the weld
PT (liquid penetrant)	Fine surface cracks and porosity on clean, nonporous metal	Subsurface flaws, anything under a cap or slag
MT (magnetic particle)	Surface and near-surface cracks in ferromagnetic material	Aluminum, austenitic stainless, copper
UT (ultrasonic)	Internal flaws in CJP groove welds: lack of fusion, cracks, porosity	Thin or coarse-grained material; operator-dependent
RT (radiographic)	Internal porosity, inclusions, lack of fusion, cracks	Cracks parallel to the film; no surface condition
ET (eddy current)	Surface and near-surface flaws in conductive metals	Nonconductive materials, deep flaws

Each method has a range, a blind spot, and a calibration requirement. Codes often require a specific combination, and that combination is where the “but it looked fine” argument ends.

For structural CJP groove welds, UT is the workhorse because it finds buried lack of fusion nobody can see. A defect under a cap is invisible to the person who made the joint: the arc covers it, the slag covers it, and the grinder can even make it look better while leaving the flaw where it was. That is why the distinction between a discontinuity and a defect matters in practice.

Work is performed by qualified technicians, not by the welder who made the joint. ASNT is modernizing employer-based personnel qualification under SNT-TC-1A, tightening who can perform and interpret each method. A CWI may be qualified for VT and MT and still be the wrong person to interpret a radiograph or an ultrasonic readout.

When It Fails: Handling Repair

Rejected welds are not automatically scrapped. Codes allow repair, but they dictate how:

1. The defect is located, identified by method, and mapped against the acceptance criteria.
2. The cause is determined. A crack from a prequalified arc strike is treated differently from a crack from residual stress.
3. The repair uses a qualified procedure and qualified welders, often with a repair WPS that addresses the root cause instead of repeating the original parameters.
4. The repair is re-inspected by the same or a more stringent method than the original.
5. Limits apply. D1.1 caps how many times a weld can be repaired and requires the repair to restore full strength. Some cracks in critical joints cannot be repaired without engineering approval.

Grinding out a defect and rewelding is the classic repair, and it works until it does not. Every grinding pass removes material and changes joint geometry, which is why a weld rejected for lack of fusion frequently returns rejected for undercut the second time.

What This Means for Your Own Work

Even if you never pull a permit, the mindset transfers. Fixing a fault at the bench is about the weld. Judging it against a spec is about whether the job holds up when someone else examines it with different tools and expectations. For inspection culture at the strictest end of the spectrum, see our guide to aerospace welding processes, standards, and careers.

The takeaway is simple. Learn to read a weld the way an inspector reads it: profile, toe condition, sound, dimension, and what is happening underneath. Then remember the inspector’s answer comes from the document, not from the arc.

The Hidden Cost of Faulty Welds: Fumes, Safety, and Your Bottom Line

Chasing a fault is only half of a welder’s job. The other half happens the second the arc strikes, because the plume off the puddle is what your lungs inherit. Once a faulty weld leaves the shop, it becomes a balance sheet problem. Before you trim amperage for the tenth time to kill spatter, know what you are breathing.

What Is Actually in Welding Fume

Fume is not smoke, and it is not the slag you chip off the bead. It is an extremely fine aerosol of condensed metal oxides that changes with every job. NIOSH reports that welding fumes are composed of metals and that most contain a small percentage of manganese, with the amount generated depending on the wire, rods, flux, base metal, process, position and bay ventilation.

Manganese is why fume gets a reputation for being quietly nasty, and the limits look disarmingly close until you line them up.

Authority	Manganese limit
NIOSH REL	1 mg/m³ TWA, 3 mg/m³ STEL
OSHA PEL	5 mg/m³ ceiling
ACGIH TLV	0.02 mg/m³ respirable, 0.1 mg/m³ inhalable

That spread matters. The OSHA ceiling permits roughly fifty times more manganese than the current ACGIH inhalable limit, so working to the regulatory ceiling means working in an environment conservative guidance calls seriously over-exposed.

NIOSH notes prolonged exposure above 1 mg/m³ can lead to manganism, a Parkinsonian syndrome with tremor, gait and mood symptoms. Confined space welding is worse: the plume has nowhere to go but through your breathing zone. NIOSH also reports welders exposed under 0.2 mg/m³ may already perform more poorly on brain and motor skills tests, with no safe harbor below the limit. It also flags welder’s anthrax, a rare occupational disease caused by bacteria entering skin lesions on arms and hands.

Hexavalent Chromium and Stainless Steel

Stainless work adds a different hazard. OSHA 1910.1026 sets a limit of 5 micrograms per cubic meter of airborne chromium (VI) as an 8-hour TWA, with an action level of 2.5 µg/m³. Engineering and work practice controls must bring exposure to or below the PEL before respirators can be relied on, plus regulated areas, exposure monitoring, medical surveillance for workers at or above the action level 30 or more days a year, HEPA-filtered housekeeping, and sealed containers for contaminated clothing.

Note the trap: buying respirators does not close that requirement. A shop that issues masks and calls the problem solved is still out of compliance.

Then there is metal fume fever, the flu-like episode some welders get on the Monday after a weekend grinding galvanized or zinc-coated steel. Chills, fever, chest tightness, nausea. It clears in a day or two, which is why people dismiss it. The fever is a signal to change the job, not tough it out.

Use the Hierarchy of Controls, in Order

1. Eliminate. Do not weld the coating. Strip galvanized material back to clean metal.
2. Substitute. A different consumable or process may produce far less fume.
3. Engineering controls. Local exhaust ventilation with a hood or arm close to the work.
4. Administrative controls. Schedules, rotation away from high-exposure work, restricted areas, training, exposure monitoring, respiratory protection program.
5. PPE. Respirators, correct shade lenses, gloves, leather jackets, eye protection for grinding and slag chipping.

What a Fault Costs After It Ships

Faults are cheap to fix on the bench and expensive everywhere else. A porosity repair means chipping out a sound-looking bead, grinding back to clean metal, re-prepping, rewelding and re-inspecting, then adding wire, gas, labor and machine time. A rejected repair can consume more hours than the original weld did.

• Code rejections and failed NDT. The work stops, schedules slip, and someone explains it to the client.
• Downtime. A shop holding up a shutdown or a launch date over a disputed weld loses money every hour it waits.
• Repeat customers and reputation. The troubleshooting reference your team follows should be the one that prevents the fault; the customer remembers the outcome, not the arc. One failed pressure part travels further than a dozen good ones.

A Practical Control Checklist

• Local exhaust ventilation. A hood or arm within a foot or two of the arc, on the side where the plume goes, not behind you. Test it by watching the fume move.
• Respirators. Only after engineering controls, with a written program, fit testing and medical clearance. The NIOSH welding fume resources explain the constituents before you pick a filter.
• Cylinder and gas handling. Secure cylinders upright, cap the valve when not in use, and check hoses for leaks before every job. The AllGas welding supply safety guide and Airgas welding supplies buyer’s guide cover the details.
• Training. New welders should be able to name the fume constituents of their own process. OSHA 1910.254 covers equipment condition, lead splices and leak-free gas delivery.

None of this is glamorous, and it is where a surprising number of shop claims end up. A fault that never happened because the fume was captured at the source costs nothing at all.

A Step-by-Step Troubleshooting Routine (and What to Do Next)

Troubleshooting works best as a routine, not a reaction. The fault you chased last week is probably the fault you will chase again this week, and the fastest way out is a repeatable sequence that moves you from symptom to cause to verified fix.

The Miller welding troubleshooting guidance makes the logic simple: look for anything that changed during the course of welding, whether parameters or technique, and treat those as the primary suspects before anything else. That is the spine of the routine below.

The 10-Step Troubleshooting Routine

1. Stop before you grind. Grinding off a bad bead destroys your best evidence. Photograph the bead, the joint, and the wire spool. A cosmetic-looking pass can hide a serious subsurface problem, as the dye penetrant story in the opening showed.
2. Name the symptom precisely. Porosity, undercut, burn-through, lack of fusion, spatter, convex bead, concave bead, cracking. Vague observations produce vague fixes. If you cannot name it, ESAB’s guide to common MIG welding defects is a good vocabulary builder.
3. Locate the fault in the joint. Toe, root, mid-bead, or face only. Position usually points at the cause: toe problems point to heat and arc angle, root problems to heat input and fit-up, face-only problems to gas and surface chemistry.
4. Ask what changed. Parameters, technique, joint prep, material, gas, consumable, equipment. Work backward from the last known good weld. If nothing changed but the fault appeared, suspect the equipment.
5. Verify gas at the flow meter. Read the flow rate at the machine, not at the cylinder regulator. A kinked hose, leaking fitting, or blown diffuser in the gun starves the shield even when the cylinder gauge looks fine.
6. Inspect the gun and consumables. Nozzle condition and size, contact tip recess and wear, liner condition, bird-nesting at the drive rolls, and wire spool cleanliness.
7. Inspect the joint. Base metal cleaned of oil, rust, paint, and moisture, correct root gap and fit-up, bevel prepared as the procedure requires. A dirty joint sits behind a startling share of avoidable faults.
8. Recalculate heat input. Wire feed, amperage, travel speed, and arc voltage against the WPS. Many faults are a mismatch between wire feed and travel speed rather than an error in either value alone.
9. Check the work side. Ground clamp placement and condition, work lead routing, and stray magnetic influence near the joint. Damaged insulation is worth ruling out too, since OSHA 1910.254 requires equipment and leads to be inspected before use.
10. Change one variable and retest. Make a single change, run a short coupon or scout piece, judge it before returning to production, then log the result so the next occurrence starts from a known baseline.

Quick Reference: Symptom to First Suspect

Symptom	Most Likely Cause	Check First
Porosity (pits or holes in the bead)	Gas coverage, contamination, leaks	Flow rate at the flow meter, hose and gun leaks, dirty joint
Undercut (groove at the toe)	Excessive heat, long arc, arc blow	Wire feed and travel speed, arc length, clamp location
Burn-through (hole through thin metal)	Excessive heat on thin material	Lower wire feed, slow the travel, add a backing bar
Lack of fusion (bead sits on the joint)	Low heat, wrong gun angle, dirty joint	Leading-edge arc placement, 0 to 15 degree gun angle, clean the joint
Spatter	High voltage or travel speed, dirty wire, worn contact tip	Spool condition, contact tip and liner, voltage balance
Convex or ropey bead	Too much wire feed for the travel speed	Slow down or reduce wire feed
Concave or flat bead	Too little wire feed, arc too long	Raise wire feed, shorten the arc
Cracking	Distortion, filler mismatch, hydrogen pickup	Filler selection, dry consumables, preheat per the WPS

Preventive Habits That Kill Most Faults Before They Start

• Follow the WPS. When a qualified procedure exists, weld to it. Deviations that are not documented are deviations that fail inspection.
• Prep the joint properly. Clean, dry, correct fit-up. On aluminum this step is non-negotiable, which is why proper aluminum cleaning before welding deserves its own routine.
• Match consumables and gas to the material. Wire, gas, and polarity combinations are not interchangeable. Our Airgas welding supplies guide covers how to choose shielding gases and cylinders for common shop work.
• Clean the nozzle and check contact tip recess before each job. A five-minute gun check prevents more rework than any parameter change.
• Keep a settings log. Wire feed, voltage, amperage, gas flow, wire, and gas per job. When the same fault reappears, you have a baseline instead of a guess.
• Practice tacks. Tack welds reveal fit-up and heat problems early, before the full pass locks them in.

Conclusion

MIG welding faults are not random, and they are not a matter of good luck or bad luck. They are the visible output of a system: gas coverage, heat input, technique, material condition, consumables, and equipment. Run the routine, change one thing at a time, and most faults resolve in minutes instead of hours.

One caveat carries real consequences. Once a weld is in a controlled structure, quality is judged against a written specification, not personal preference. Anything that fails the governing code’s acceptance criteria is a defect, and repair work must follow a qualified repair procedure rather than simply re-running the same settings. Current code text is published through AWS standards and ASME codes and standards, and the personnel side of it, the Certified Welding Inspector, is covered on the AWS professional certification page. For the non-destructive testing side, ASNT sets the personnel qualification standards used across the industry. If a fault reaches that stage, you are past troubleshooting and into inspection and repair procedures.

One more resource worth keeping on the bench: our complete aluminum MIG welding guide covers the parameter and consumable differences that make aluminum behave nothing like steel.

Frequently Asked Questions

What are the most common MIG welding faults?

Porosity, lack of fusion, undercut, burn-through and excessive spatter top the list, and they all trace back to a small number of root causes: gas coverage, heat input, technique, joint preparation and consumable condition. Almost every fault can be narrowed to one of those five categories before you start changing settings.

Is porosity in a MIG weld dangerous?

It depends on where it is and how much of it there is. A couple of surface pores on a non-critical bracket may be acceptable under the governing code, but porosity reduces the load-bearing cross-section, acts as a stress raiser, and can become a crack initiation site. Under AWS D1.1 and similar codes, porosity is limited by both individual size and the combined length of pores per inch of weld.

Why am I getting undercut in my MIG welds?

Undercut means too much heat is melting the base metal at the toe of the bead. The usual culprits are excessive voltage or amperage, too long an arc, a gun angle that pushes the arc toward the side of the joint, and travel speed that’s too fast for the amperage. Adjust voltage first, shorten the arc, and fix your gun angle before reaching for the wire feed speed.

Why won’t my MIG weld penetrate?

Nine times out of ten it’s one of four things: not enough heat, a dirty or oxidized joint, a root gap too wide for the wire to bridge, or the arc sitting on the wrong side of the pool. Increase voltage or wire feed speed, grind the joint back to bright metal, tighten fit-up, and keep the arc on the leading edge of the puddle with a 0–15 degree gun angle.

Can a faulty MIG weld be repaired, or does it have to be cut out?

It can often be repaired, but only with a qualified repair procedure that addresses the root cause. Cracking, lack of fusion and incomplete penetration usually get ground out and rewelded, with the area re-inspected to the same acceptance criteria as the original weld. Anything over a limited number of repair attempts needs an engineer’s approval.

How much shielding gas flow do I need for MIG welding?

ESAB recommends a typical flow rate of 25–40 CFH for MIG welding with the nozzle held close to the joint. Always set and verify the flow at the flow meter rather than at the cylinder, and remember that wind and a leaking torch or hose will destroy your shielding no matter what the gauge reads.

What is the difference between a weld discontinuity and a weld defect?

A discontinuity is any deviation from the ideal weld shape or internal structure — porosity, undercut, spatter, a small concavity. A defect is a discontinuity that exceeds the acceptance criteria of the code or specification governing the work. So the same visible pore might be perfectly acceptable on one job and a rejectable defect on a cyclically loaded structure.

Why does my MIG gun get so much spatter?

Usually it’s one of four things: voltage or travel speed set too high, a worn or wrong-size contact tip, excessive wire stickout, or dirty and contaminated base material and wire. With flux-cored and metal-cored wires, low voltage is a common cause of heavy spatter, so try raising the voltage a little before anything else.

Why do my stainless steel MIG welds crack?

Stainless welds crack most often from too much heat input and too little restraint control, which promotes hot cracking, or from picking up contamination — a copper-brass grinding wheel, a dirty wire brush, or iron particles from carbon steel work nearby. Reduce heat input, keep the bead smaller, and keep stainless completely separate from carbon steel tooling.