TIG Welding Defects: Causes, Signs, and How to Fix Them

A fabricator once spent a full shift laying what looked like a beautiful stack-of-dimes bead on stainless tube, only for the radiographer to flag bright specks and a hairline crater crack at the very end of the run. That is the cruel irony of TIG welding: the process that produces the prettiest welds also punishes the smallest mistakes hardest.

What Counts as a TIG Welding Defect (and Why TIG Is So Unforgiving)

Every TIG welder has a story like this: hours spent laying what looked to be a perfect stack-of-dimes bead, only for it to be scrapped after inspection. For one fabricator working with thin-walled stainless tube, the rejection came not from looks, but from radiography. Tiny specks and a hairline crater crack at the very end of the run were enough to condemn the entire joint. It’s a cruel reminder that with TIG welding, aesthetics are never a guarantee of soundness.

Before diving into any of those flaws, it’s important to draw a line between two terms that are often used interchangeably. According to the American Welding Society, a discontinuity is an interruption in the typical structure of a weldment. It only becomes a defect when that interruption renders the item unserviceable, or when it violates an applicable acceptance criterion. In short, the weld code makes the call, not the naked eye.

The Precision That Comes At A Cost

TIG welding, more formally known as Gas Tungsten Arc Welding (GTAW), uses a non-consumable tungsten electrode and an inert shielding gas to protect the molten puddle. Unlike MIG or stick welding, the weld metal comes solely from a separate filler rod that the welder feeds by hand. This manual feed, paired with an incredibly focused arc, is exactly what allows GTAW to produce some of the cleanest, highest-integrity welds possible. However, it’s also exactly what makes it so unforgiving.

There are a few technical reasons for this strictness. GTAW must use a drooping constant-current power source to avoid excessively high current if the electrode were to accidentally touch the workpiece. Additionally, because arc heat is distributed roughly one-third at the cathode and two-thirds at the anode, the tungsten is run negative on DCEN to prevent it from overheating. On top of this, with such low heat input and such a narrow shield, even the smallest lapse in cleanliness or gas coverage can ruin a joint. For this reason, proper surface preparation, such as using the methods outlined within Aluminum Cleaner for Welding: How to Prep Aluminum Right, becomes just as important as the arc itself.

Who Needs To Know This, And What To Ask

This knowledge isn’t reserved for one group. Hobbyists chasing their first clean root pass, welding students learning fundamentals, certified welders working to code, and inspectors holding the final say all benefit from learning to spot GTAW flaws early. With all of these perspectives in mind, the principal question that this article aims to answer is simple: what do TIG defects look like, why do they happen, and how can they be prevented and repaired?

To answer that, the flaws ahead will be broken down by their families. Some are atmospheric, such as porosity caused by a compromised shield. Others are electrode-based, such as the signature tungsten inclusions. Some are fusion-based, such as lack of fusion or incomplete penetration. A select few are crack-based, and any of these are considered unacceptable regardless of size. Finally, there are geometry-based flaws, material-specific issues for aluminum and stainless, and most importantly, how each of these are judged through inspection against an acceptance criteria.

With that groundwork laid, every one of these defects can be traced back to the same four culprits: shielding, cleanliness, heat input, and termination technique. Each of the sections to come will focus on controlling those four, as they are the difference between a rejected tube, and a sound one.

How the GTAW Process Itself Creates Defects: Arc, Gas, and Electrode

Before cataloguing individual defects, it helps to know where they come from. Almost every TIG (GTAW) defect traces back to one of four variables: heat balance, shielding, cleanliness, or technique. Heat balance is amperage, arc length and travel speed. Shielding is gas delivery, flow rate and post-flow. Cleanliness is the plate, the filler and the tungsten. Technique is arc starting, torch angle, joint tracking and arc termination. When a weld goes wrong, name the bucket first and the fix usually follows.

This mental model matters more in GTAW than in MIG or stick, because those processes carry their own shielding with the flux or the core wire. As TWI’s overview of the TIG process explains, GTAW protects the arc with an externally supplied inert gas and relies on a non-consumable tungsten electrode, so the arc’s quality is only as good as the settings and hands behind it.

How you strike the arc decides how the weld starts

High-frequency (HF) start is the preferred method for high-quality work because the arc jumps the gap without the electrode ever touching the plate. Scratch starting, by contrast, drags the tungsten across the work like a matchstick, and the risk is obvious: the electrode can stick, short out, and drop a piece of tungsten straight into the molten pool. Lift-arc starting sits in between, with the tungsten briefly contacting the plate before the current ramps up. On coded or appearance-critical work, from aerospace components to food-grade tube, HF start removes an entire category of human error.

Shielding gas delivery: laminar beats turbulent

Gas only protects if it arrives at the arc in a calm, steady column. A standard baffle or collet body gives turbulent flow at the nozzle exit, which leaves the weld pool partly exposed, especially on vertical joints, inside corners and edges where the shielding curtain is easily disturbed. Fitting a gas lens in the torch restructures the flow into a laminar jet that reaches further and stays stable, allowing a lower cylinder flow rate with better coverage.

Tungsten preparation: grind direction and tip shape

Ground tungsten is a precision tool. Grind longitudinally, along the length of the electrode, so the scratches run in the direction of current flow; a circumferential or faceted grind creates erratic arc starts and wandering. The included angle should match how you weld: a shorter, blunter taper for overhead and out-of-position work, a longer point for flat butt welds where you want a focused, penetrating arc. Regrind on a dedicated, uncontaminated wheel, never the one used for steel, and remove any contaminated tip completely before restarting. Consistent tungsten prep is one of the habits drilled into aerospace TIG welding programs, where arc stability is non-negotiable.

Polarity, amperage, arc length, travel speed and post-flow

• Polarity: DC electrode negative (DCEN) gives deep, narrow penetration with little cleaning action, ideal for steel, stainless and titanium. AC is used for aluminum, where the electrode-positive half of the cycle blasts away the oxide film and the negative half drives penetration. Balance control sets that ratio: too much cleaning and you etch the surface, too little and the oxide contaminates the pool.
• Amperage versus electrode size: match the tungsten to the current. Undersized tungsten overheats, blunts, balls unpredictably or spits particles into the pool; oversized tungsten wastes a focused arc. A common shop guideline is roughly 100 amps per 1/16 in (1.6 mm) of tungsten for DCEN work, with 3/32 in (2.4 mm) the safer choice for heavier currents.
• Arc length: keep it short, about the diameter of the electrode. A long arc spreads heat, weakens the gas shield and invites both porosity and tungsten contamination.
• Travel speed: fast freezes the pool before gas can escape; slow cooks the joint, opens the arc up and risks burn-through or overheated tungsten.
• Post-flow: the shield does not stop when the arc does. Post-flow keeps gas over the hot crater and the still-glowing electrode until both drop below oxidation temperature, preventing crater oxidation and tungsten degradation on the next start.

Get those five dials right and you will have already prevented a surprising share of the defects in the sections ahead. Cleanliness comes next, and that starts before the arc is ever struck, as covered in aluminum cleaning for welding.

Porosity in TIG Welds: Types, Causes, and What the Pores Tell You

Porosity is the defect most TIG welders run into first, and the one that stings hardest, because the bead often looks flawless from the outside. A weld can stack a perfect row of ripples, wet in beautifully at the toes, and still hide a cluster of gas pockets underneath. It is also the defect you are most likely to get caught with, since distributed porosity is exactly what radiography and ultrasonic testing are designed to find.

The Gas Metallurgy, in Plain English

Molten metal can hold a lot of dissolved gas, much like a cold soda holds carbon dioxide. Steel, stainless and aluminum pools absorb nitrogen, oxygen and hydrogen from the surrounding air while they are liquid. The moment the pool solidifies, that solubility collapses and the gas has nowhere to go except into bubbles. If the bubbles float out before the puddle crusts over, you never know they existed. If the surface freezes first, they are trapped, and you have porosity.

This is why TIG is both blessed and cursed. There is no flux coating generating its own shielding blanket, so the whole job rests on the inert gas envelope around the arc, as TWI’s technical guidance on weld porosity explains. Anything that lets air reach the molten pool can produce pores.

The Five Shapes Porosity Takes in a GTAW Weld

Not all porosity looks the same, and the shape is a clue:

• Distributed porosity: fine, evenly scattered bubbles through the weld volume, usually from a shielding or cleanliness problem that lasted most of the run.
• Surface-breaking pores: pinholes open at the weld face, often the visible tip of a larger subsurface cluster.
• Wormholes: elongated, tube-like pores stretched along the direction of travel, formed when gas is still rising through a pool that is freezing around it. On a radiograph they show a distinctive herringbone pattern, which is why inspectors call them out so quickly.
• Worm tracks: elongated surface depressions left where gas worked its way out of a mushy puddle, often seen on aluminum.
• Crater pipes: a shrinkage void at the arc termination, formed during final solidification of the pool and frequently accompanied by gas porosity. Autogenous TIG stops, or simply stopping the filler feed while the arc is still running, are classic invitations.

Shielding Failures: Where the Gas Envelope Breaks Down

Start with the gas path. A leaking hose, a loose nut at the flowmeter or solenoid, a cracked torch head or a perished o-ring will all pull air into the line under the slight vacuum created by flowing gas. The cheapest diagnostic in the shop is a soapy-water leak check on every fitting. Then look at the environment: a fan blowing across the bench, an open roller door, or a fume extraction arm parked too close will strip the shielding envelope right off the arc. Excessive travel speed causes porosity too, because the pool freezes before bubbles can float out, and an arc that is too long wanders, leaving the shield gas a thin, easily disturbed envelope to work in.

Contamination: The Other Half of the Problem

Gas does not only come in from the air; it is often already on the metal. Paint, primer, cutting fluids, oil, grease and rust all decompose in the arc and release gas straight into the pool. Zinc coatings are notorious, which is why alloy wheel welding guides insist on stripping the weld zone back to bare aluminum. Moisture is the hydrogen source that catches people out: damp filler wire, a workpiece left out overnight, or a box of aluminum rod stored in a humid shop. The same moisture rules apply if you switch processes, as covered in our aluminum MIG welding guide. For aluminum specifically, the hydrated oxide film is the enemy, and proper mechanical and chemical preparation per aluminum cleaning practice prevents much of the porosity people blame on their machine.

Some porosity sits in the base metal itself. Sulfur, phosphorus and especially lead in the alloy promote gas-related problems, and free-machining or leaded grades are generally considered unweldable by GTAW.

Reading the Pores to Find the Cause

What you see	Most likely cause	First check
Fine porosity through the whole weld	Shielding loss or dirty filler	Gas leaks, drafts, filler storage
Porosity clustered at one spot	Local contamination or arc wander	Paint, oil, gap in fit-up at that point
Wormholes on radiograph	Gas trapped by a fast-freezing pool	Travel speed and amperage balance
Crater pipe at the stop	Bad termination technique	Downslope, crater fill, filler added at stop

The pattern is worth more than any single pore. Fix the pattern, and the porosity goes with it.

Tungsten Inclusions and Tungsten Spitting: The Defect Only TIG Has

Every welding process has its signature flaw, and GTAW’s is unique: a fragment of the tungsten electrode itself ended up in the weld. A tungsten inclusion is exactly what it sounds like — a hard, high-melting-point particle of the non-consumable electrode embedded in the weld metal. Because the electrode is never supposed to melt, its presence in the bead means something in your setup or technique went wrong. MIG and stick welders never deal with this, since their electrodes are consumed on purpose, as TWI’s overview of the TIG process makes clear.

Three Ways Tungsten Gets Into the Weld

The routes in are predictable, and experienced welders have taken all three:

– Dipping the electrode into the puddle. The most common cause. Arc length collapses, the tungsten touches molten metal, and a bead of contaminated tungsten freezes into the joint the moment you pull away.
– Touching the filler rod to the electrode. Feeding the rod a little too eagerly, especially in a tight joint, transfers filler and tungsten into each other. The tungsten picks up a coating that then drops into the pool.
– Tungsten spitting. Excessive current for the electrode size, a badly ground or contaminated point, or a taper that is too sharp causes small particles to boil or shatter off the tip and spray into the weld. So does excessive electrode extension, which overheats the stick-out.

The last one is a parameter problem, not a hand-eye problem. A 2.4 mm tungsten carrying 200 amps will erode and spit no matter how steady your hands are. Correct electrode type, correct diameter for the current, and the right taper angle keep the tip stable.

Why Radiography Flags It So Clearly

Tungsten inclusions are dense, irregular, and show up during radiographic testing as bright indications on film — brighter than the surrounding weld because tungsten is denser than the lead penetrameter used to establish image quality. That brightness is why a welder who “got away with it” visually can still get flagged by the inspector.

Acceptance depends on the code. Some standards permit small, isolated inclusions below a stated size limit, on the reasoning that tungsten is inert and metallurgically harmless — it does not crack or corrode the way a lack-of-fusion void can. Cracks are different: no acceptance table anywhere lets a crack pass, regardless of size or location. Under visual criteria such as AWS D1.1, anything you can actually see sticking out of the bead surface is rejectable as exposed tungsten.

How to Fix an Inclusion

Once the flaw is found, the repair is mechanical and unforgiving:

1. Stop and correct the root cause first — dial back amperage, shorten arc length, regrind the tungsten, or switch to a larger electrode.
2. Grind out the contaminated weld metal to sound metal, extending far enough that the buried tungsten particle is gone, not just the visible speck.
3. Clean the groove and the back of the joint before rewelding; smeared tungsten dust left in the crater simply recreates the defect. Good joint prep discipline matters here, especially on aluminum, where residue hides easily.
4. Reweld using the qualified procedure and reinspect with the same method that caught the flaw.

Prevention Beats Grinding Out

Tungsten contamination is far cheaper to prevent than to excavate, and the controls are simple: start the arc with high-frequency rather than scratch starting, or lay the electrode on a copper striker plate (or on the filler rod itself) so the first strike never touches the workpiece. Keep the taper moderate — a long, needle-sharp point melts back instantly, while a slightly blunted tip survives. Match electrode diameter to current, use a gas lens to stabilize a longer arc, and never let the tip wander outside the gas envelope. In high-rigor work such as aerospace TIG welding, operators regrind and inspect tips between passes as a matter of routine. The habit pays off anywhere the weld will be x-rayed.

Lack of Fusion, Incomplete Penetration, and Cold Lap

Fusion defects are the quiet failures of TIG welding. Porosity shows on film and cracks announce themselves, but a weld can sit on the surface looking dense and tidy while never bonding to the joint underneath. That is what makes these TIG welding defects dangerous, and why inspectors treat them more seriously than their looks suggest.

Why GTAW Still Fails to Fuse

GTAW has a reputation for excellent penetration, and on paper it deserves it. The arc is focused, the heat adjusts in fine increments, and the welder watches the puddle the whole time. But that same manual control puts every variable in the operator’s hands. Fusion defects in TIG trace back to technique, work angle, or settings simply too cold for the thickness and joint geometry in front of you.

The common causes stack up quickly:

• Parameters that are too cold. Low amperage leaves a puddle sitting on the base metal instead of biting in.
• Torch not centered on the joint. Heat goes to one side, the other edge never reaches fusion temperature, and the root stays unfused.
• Wrong work and travel angles. Pushing too hard or dragging at an odd angle drives the arc away from the root and leaves the sidewall untouched.
• Travel speed off pace. Too fast and the puddle freezes before the edges wet out; too slow and heat wanders off the joint line.
• Oversized fillet in one pass. A single pass cannot put enough heat into a large joint, so the center builds while the sidewalls stay cold.
• No dwell at the edges. TIG needs a deliberate pause at each side of the puddle to melt the sidewall before adding filler.
• Poor fit-up and open gaps. Wide gaps pull heat away, distort the joint line, and leave the prepared root edge unconsumed.
• Vertical-down technique. The puddle runs ahead of the arc, solidifies, and the arc welds over frozen metal instead of into it. Vertical-up with a controlled weave is usually the fix.

The Four Flavors of Missing Fusion

On the shop floor these terms get used interchangeably, but they describe different failures. Telling them apart matters because the cause and the repair differ for each.

Defect	What is actually wrong	Typical trigger
Lack of side-wall fusion	Weld metal never bonded to the prepared joint face	Off-center torch, no edge dwell, cold settings
Lack of interbead fusion	Adjacent beads or passes sit against each other without bonding	Too-fast travel, skipping the restart overlap, cold second pass
Incomplete joint penetration	The weld does not reach through; the prepared root edge is left un-melted	Amperage too low, gap too tight, arc not aimed at the root
Overlap / cold lap	Filler rolled over base metal that was never melted	Excessive filler, puddle too cold, wrong torch angle

TWI’s overview of the TIG process makes the key point plainly: the arc only delivers heat where it is pointed, so the joint root and sidewalls have to be deliberately included in the puddle every pass.

Why Fusion Defects Hide from the Eye

Lack of fusion is a volumetric defect living inside the joint. From the surface you see a continuous bead, correct reinforcement, clean toes. The void only reveals itself through non-destructive testing: radiography shows it as a dark linear indication along the joint line, ultrasonic testing maps it as a reflected echo, and penetrant or magnetic particle testing catches it only when the unfused area breaks the surface.

That invisibility is why the acceptance rules are strict. AWS D1.1 Table 8.1 requires complete fusion between adjacent weld layers and between the weld metal and the base metal. Visible lack of fusion carries no size threshold: if you can see it, it is rejectable, with no “small enough to pass” allowance like distributed porosity gets. MTU’s welding discontinuities training material groups these under the same zero-tolerance category as cracks for visual evaluation, precisely because they mean the joint never became a joint.

Repair: Remove to Sound Metal, Then Reweld

You cannot fix lack of fusion by running another bead over it. The affected area has to be ground, chipped, or air-arc gouged back to clean, sound metal, then re-inspected to confirm the unfused layer is gone. Then correct the cause before striking the arc again: raise the amperage or slow the travel, re-center the torch, restore proper work and travel angles, split an oversized fillet into multiple passes, and add a deliberate dwell at each sidewall. The rework needs the same inspection as the original, and the governing procedure has to be followed for requalification.

The lesson holds across every fusion defect: heat input, torch position, and travel timing decide whether metal joins metal. On critical TIG work such as aerospace fabrication, where radiography and ultrasonic testing are routine, those three variables are checked long before the first pass is made.

Cracking in TIG Welds: Crater Cracks, Hot Cracking, and Hydrogen Cracks

If porosity is the defect welders argue about, cracking is the one nobody argues with. In AWS D1.1 Table 8.1, cracking is the only absolute zero-tolerance criterion: any crack shall be unacceptable, regardless of size or location. A 0.02 inch pore in a quarter-inch plate can pass with limits. A hairline crack the width of a human hair at the very edge of a weld gets rejected, ground out, and rewelded. Codes treat cracks as unacceptable because a crack is a moving target — under load, vibration or thermal cycling, it propagates.

Crater Cracks: The Most Common TIG Crack

Most welders meet their first crack at the end of a bead. As the arc comes off the puddle, the last small volume of liquid metal shrinks as it solidifies, and there is no molten pool left behind it to feed that shrinkage. The result is a concave crater, sometimes with a pipe or a fine star-shaped crack running through the center. Terminating the arc right at the joint edge, where restraint is highest, makes it worse.

The fixes are mechanical and they work:

– Use the machine’s downslope or crater-fill function. Modern GTAW power sources ramp the current down over a set time so the puddle freezes gradually instead of being cut off. As TWI’s overview of the TIG process notes, current control at arc termination is a built-in advantage of the process — most welders simply never set it up.
– Back off the pedal slowly if you are running manual control, and keep the post-flow gas running over the crater.
– Add two or three extra drops of filler at the stop to build the crater back up to flush or slightly convex.
– Reverse direction and run a short overlap back into the previous solidified weld, then break the arc over metal that is already there rather than over an open edge.

Centerline Hot Cracking and Contamination Cracking

Hot cracking happens while the weld is still cooling, generally in single-pass bead-on-plate or narrow butt welds where the bead freezes from both toes inward and the centerline is the last liquid to solidify. Shrinkage pulls that final line apart. The usual suspects are a bead that is too deep and narrow for its width, high restraint from the joint, and low-melting constituents in the base metal such as sulfur, phosphorus or lead. In Michigan Tech’s discontinuities training, centerline cracking is presented as the classic solidification cracking pattern, driven by geometry and chemistry together.

Contamination drives cracking too. Copper from plating, zinc from galvanizing, oil, paint and heavy oxide films all contaminate the pool and create brittle, crack-prone regions. On aluminum this is doubly true, since the oxide film and any moisture it holds introduce hydrogen directly into the melt — proper aluminum prep is a crack-prevention step, not just a cosmetic one.

The geometry of the bead matters as much as the chemistry. A convex bead puts concentrated stress on the toes, while an over-concave bead lacks effective throat. The optimum fillet has a flat face with toes blended smoothly into the plate.

Post-Weld Cold Cracking

Cold cracking shows up hours after the weld, sometimes the next day. Three conditions combine to cause it: excessive joint restraint, rapid cooling, and hydrogen embrittlement. Hydrogen comes from moisture in coatings, humid electrode storage, oily fit-up surfaces, or dirty filler. The remedy is preheat where the procedure calls for it, larger multi-pass beads that cool more slowly, careful cleaning and drying of consumables, and joint design changes that reduce restraint — altering the sequence, tack pattern or fit-up so the joint is not fighting the weld as it shrinks.

Crack Prevention Checklist

1. Set downslope and crater fill before you strike the arc, not after.
2. Never stop a bead on an open edge; overlap back into solid metal.
3. Keep bead width-to-depth healthy — avoid narrow, deep single passes.
4. Control bead profile: flat face, smooth toes, no sharp notches.
5. Clean, dry and degrease everything that touches the puddle.
6. Preheat restrained or hydrogen-sensitive joints per the procedure.

Cracks do not get better with grinding unless you remove them entirely to sound metal and requalify the repair. In critical work, from pressure components to aerospace TIG welding, the acceptance rule is the same one every code shares: find it, remove it, prove it is gone.

Undercut, Weld Profile Problems, Worm Tracks, and Burn-Through

Plenty of TIG rejections have nothing to do with gas or cracks. A whole family of defects is purely geometric: the metal is sound and fused, but the weld is the wrong shape. Inspectors call these profile discontinuities — undercut, excessive convexity, concavity, undersized fillets, overlap, worm tracks, and burn-through. Geometry controls stress. A notch at the toe concentrates load the way a nick concentrates force in a rope, and a thin concave throat carries less force than the joint was designed for. On the plus side, these are the most visible defects here, quickest to catch by eye and gauge.

Undercut: The Notch Nobody Wants

Undercut is a groove melted into the base metal along the toe or root of a weld and never filled with deposited metal. It reads as a sharp crescent on one or both edges. In TIG it almost always means too much heat at the edge relative to the filler being added: excessive current, an arc long enough to spread heat past the puddle, travel too fast for the filler to wash into the groove, or, in fillets, the arc aimed squarely at one plate instead of centered in the joint. Aborting a stop without enough filler carves undercut right at the crater.

The danger is structural, not cosmetic. Undercut is a stress raiser, and on high-strength steel or cyclically loaded parts it becomes a crack initiation site. AWS D1.1 draws the line in its visual acceptance criteria: for statically loaded connections, undercut is limited to 1/32 in [1 mm] where the base metal is under 1 in thick, and 1/16 in [2 mm] where it is 1 in or thicker. For cyclically loaded primary tension members, undercut transverse to the direction of tensile stress is capped at 0.01 in [0.25 mm], which is effectively zero (AWS D1.1/D1.1M, Table 8.1). Codes that run tighter still, such as aerospace work, treat any visible undercut on primary structure as grounds for rejection outright (Aerospace TIG Welding: Standards, Techniques, and Training).

Fix it at the torch: drop the amperage or shorten the arc, slow travel slightly, center the electrode in the joint, and feed filler into the edge rather than running bare. Add metal at every stop. Shallow undercut on non-critical, statically loaded material can sometimes be ground smooth and blended, but only where the governing code allows it and only after confirming the remaining wall thickness still meets design minimums.

Profile Problems: Convexity, Concavity, and Undersized Fillets

Each profile flaw has its own signature:

– Excessive convexity (a humped bead) comes from too much filler for the heat input, or travel so slow the bead piles up. The sharp transition at the toe becomes a second stress raiser, similar in effect to undercut.
– Concavity is the opposite — a sagging, dish-shaped face, most common out of position, where gravity pulls the puddle down before it freezes. The result is a reduced effective throat: the weld is thinner than the size stamped on the drawing.
– Undersized fillets happen when the operator attempts too large a leg in a single pass or never dwells at the toes to wet them in. D1.1 tolerates this only narrowly: undersized portions of a fillet weld must not exceed 10% of the total weld length.
– Overlap (cold lap) is filler rolled over the base metal without fusing it, usually from too much work angle, too little current, or a puddle too viscous to wet the edge. It looks acceptable to beginners because there is plenty of metal — but no bond.

Most profile problems are corrected by rebalancing current and travel speed and slowing down at the toes, not by adding more metal.

Worm Tracks and Wagon Tracks

Those long, shallow grooves running lengthwise along an otherwise healthy-looking bead are surface evidence of elongated subsurface porosity. The gas forms as the pool solidifies and gets stretched in the direction of travel. They signal one of two things: shielding turbulence or travel speed outrunning the gas envelope. High flow rates that stir the shielding gas, drafts hitting the arc zone, or simply moving faster than the trailing gas can protect all produce them. Treat a worm track as a warning, not a standalone defect — where you can see one, more porosity usually hides underneath.

Burn-Through and Melt-Through on Thin Material

On thin sheet, TIG’s precision becomes its liability. Burn-through (an actual hole) and melt-through (a sagging back side) are usually caused by excessive current for the application, low travel speed, excessive gaps, or filler wire too small to sink heat away from the puddle. Thin aluminum is the classic victim, since it conducts heat so aggressively that the area around the hole stays hot long after the arc passes; proper surface prep with the right chemistry helps keep heat input predictable (Aluminum Cleaner for Welding: How to Prep Aluminum Right). Countermeasures: reduce amperage, use pulsed current, increase travel speed, backstep, step up to a larger filler rod, and stitch across wide gaps in short segments rather than bridging them in one pass.

Defect, Typical Cause, First Fix

Defect	Typical cause	First fix
Undercut	High current, long arc, fast travel, off-center arc	Reduce amperage, shorten arc, add filler at toes
Excessive convexity	Too much filler, slow travel	Feed less rod, move a bit faster
Concavity	Puddle sag, out-of-position welding	Lower heat, tighten puddle control
Undersized fillet	Too large a leg in one pass, no toe dwell	Make multiple passes, dwell at edges
Overlap	Low current, wrong work angle	Raise amperage, correct torch angle
Worm tracks	Gas turbulence or travel too fast	Screen drafts, check flow, slow down
Burn-through	Excess current, slow travel, wide gaps, thin filler	Reduce amps, pulse, larger rod, backstep

Grind It Out or Weld It Over?

Grinding and blending are legitimate repairs for shallow profile defects — undercut, a proud hump, a rough stop — as long as you blend the toe smoothly, verify thickness, and the code you are working to allows it. Anything that removes a fusion boundary (overlap, deep undercut at the root, suspected lack of fusion) should be gouged or ground back to sound metal and rewelded, because grinding merely hides what lies beneath. For a plain-language map of how these discontinuities are classified and judged, Michigan Technological University’s welding defects webinar is a useful reference (MTU Welding Discontinuities and Defects). The cheapest repair is always the one you prevent: correct heat, correct angle, correct speed, and one extra second of filler at every stop.

Aluminum and Stainless TIG Defects: Oxide Inclusion, Sugaring, and Discoloration

Porosity and cracks can happen in any metal, but aluminum and stainless steel each have signature failures of their own, and they catch out experienced welders constantly. The surface chemistry is different, so the way the weld goes wrong is different too.

Why Aluminum Is a Separate Game

Aluminum carries a tenacious oxide film that melts around 2,000°C while the base metal softens near 660°C. If that oxide is not stripped away as the arc passes, it simply gets buried in the pool. This is why GTAW on aluminium is run on AC: the half-cycle where the workpiece is cathodic provides a cleaning action that blasts the oxide off just ahead of the puddle, while the other half delivers penetration.

That split is adjustable on modern machines as AC balance. Push toward more cleaning and you get a wide etched zone but shallower penetration; push toward penetration and the oxide may not lift cleanly. Find the balance where the cleaning etch sits about 1–2 mm either side of the arc without the bead going cloudy.

Aluminum-Specific Defects

• Oxide inclusions. Insufficient cleaning, a dirty or worn wire brush, or an untuned AC balance leaves brown or black streaks of oxide trapped in the weld. Heavy surface oxides that were not removed are a specifically listed cause of tungsten and oxide inclusions, with AC cleaning action, balance adjustment and wire brushing before welding cited as the remedies. Brush with a dedicated stainless brush used only on aluminum, always in one direction, never with contaminated tools.
• Burn-through on thin sheet. Aluminum conducts heat so well that the operator chases a cold-looking surface while the area behind the arc has already collapsed. Slow down your reaction to the puddle, drop the amperage, and use a smaller filler diameter.
• Sagging and insufficient throat on out-of-position fillets, where the pool stays fluid too long and gravity pulls the face convex at the top and thin at the root. Stitch the filler, control heat with a foot pedal, and reconsider position before you reconsider settings.
• Porosity from filler and surface condition. Aluminum sponge stores moisture, and oily or hydrated filler releases hydrogen straight into the pool. Etched, chalky surfaces left by chemical cleaners do the same. Proper prep matters more here than almost anywhere else; our Aluminum Cleaner for Welding: How to Prep Aluminum Right guide walks through the sequence. Paint, primer and zinc coatings add their own fume, which can be trapped as porosity, with typical maximum primer thickness around 20 micrometres at joint edges.

Stainless Steel: Heat Tint, Sugaring, and Warping

Stainless fails from heat rather than dirt. Too much heat input or too slow a travel speed produces heat tint, the straw-to-blue discoloration that signals chromium has oxidized locally. On a decorative piece it is a nuisance; on a corrosive-service weld it means sensitization, where chromium precipitates as carbides at the grain boundaries and the area beside the weld loses its corrosion resistance. The cure is simple: lower current, faster travel, stringer beads, and pickling or passivating after the fact.

The dramatic one is sugaring: thick, grey, cinders-of-bread chromium oxide forming on the back side of a root or autogenous weld when the underside is not protected by inert backing gas. It is not just cosmetic; the affected root must be ground out and rewelded. Backing gas, a purged chamber, or a trailing backing arrangement is non-negotiable for root passes. Conversely, too little heat with an over-fast travel speed causes warping: the weld shrinks against cold, rigid surroundings and pulls the assembly out of shape.

Contamination cracking follows any residue that survives into the pool. Cutting fluids, marking inks, and sulfur-bearing free-machining grades are common offenders. Clean with a dedicated stainless wire brush and a solvent that leaves no film.

Controlling Coverage on Both Alloys

Three controls do most of the work:

1. Balance control on AC, tuned so cleaning and penetration are both adequate without over-etching.
2. Backing and trailing gas — purge the root on stainless, and extend the shield behind the arc on fast aluminum runs.
3. Surface preparation — mechanically clean immediately before welding, keep primers thin at joint edges, and store filler dry.

Handled this way, aluminum and stainless stop being the metals that punish you and become what TIG is best at: clean, controlled, X-ray-quality joins.

Inspection and Acceptance: When a Discontinuity Becomes a Rejectable Defect

Here is a truth that surprises a lot of new TIG welders: the welder almost never has the final say on whether a weld passes. The inspector does, and inspectors do not judge a weld by how pretty it looks. They judge it by whether the flaws exceed the limits written into a code. That gap between “looks good to me” and “meets the acceptance criteria” is where most shop-floor arguments begin.

The vocabulary matters. A discontinuity is any interruption in the normal structure of a weld — a pore, a slag line, an undercut toe, an inclusion. A defect is a discontinuity severe enough to be rejectable under the applicable code or specification. The same pore can be acceptable in one job and rejectable in another. Nothing about the pore changed. The standard did. So the real question is never “is there a flaw?” There always is. It is “how is the flaw found, and does it exceed the limit?”

The inspection methods, and what each one catches

Different flaws live in different places, so different methods are needed to find them.

• Visual inspection (VT) — The baseline. Every production weld gets looked at, and most acceptance decisions start and often end here. VT catches surface-breaking porosity, undercut, surface-open cracks, incomplete fusion at the toes, wrong profile, and spatter. It cannot see inside.
• Penetrant testing (PT) — A dye is drawn into surface-breaking flaws and developed back out. PT finds tight cracks, laps, and surface pores the eye misses, and it works on non-magnetic material like austenitic stainless and aluminum.
• Magnetic particle testing (MT) — Ferromagnetic steel only. Flux leaks around surface and near-surface flaws and iron particles cluster at the leak. Fast and very sensitive to cracks.
• Radiographic testing (RT) — X-rays or gamma rays pass through the weld onto film or a digital detector. Dense flaws show dark, less dense spots show bright. RT excels at volumetric flaws: porosity, inclusions, internal voids.
• Ultrasonic testing (UT) — High-frequency sound finds internal planar flaws like lack of fusion and cracks, and it outperforms RT on thick sections. More skill-dependent, but highly capable.

Radiography is normally the stronger method for detecting and characterising porosity, but it has a real weakness: small pores become hard to resolve in thick sections, so a weld that would clearly fail in thin plate may look reasonably clean on film at 40 mm. That is one reason UT is often paired with RT on heavy work.

The acceptance criteria: AWS D1.1 as the worked example

Most structural steel TIG work in North America is judged against AWS D1.1, the Structural Welding Code — Steel. Its Table 8.1 lays out the visual acceptance limits, and two porosity rules come up constantly.

Condition	AWS D1.1 Table 8.1 limit
Piping porosity (visible surface pores)	Sum of pores 1/32 in or larger shall not exceed 3/8 in in any linear inch of weld, and 3/4 in in any 12 in of weld
Cyclically loaded fillet welds	No more than one pore per 4 in of weld length, maximum pore diameter 3/32 in
Cracks	Not permitted, regardless of size or location
Undercut (primary members, cyclically loaded)	Depth limited per the code; shall not exceed 0.01 in for certain transitions

That first rule is the one welders argue about most. It is not a count of pores — it is the sum of their diameters over a given length of weld. Ten tiny pores adding up past 3/8 in in a single inch fail, even though each one is individually under the size threshold. Inspectors measure and add; welders eyeball and shrug. That is why the argument happens.

The second rule applies to fillet welds in cyclically loaded structures — bridges, dynamically loaded machine frames. A single 3/32 in pore in a four-inch run is the absolute ceiling. It is strict, and it exists because cyclic loading turns small flaws into crack starters.

D1.1 is not the only code. ASME Section IX governs procedure and performance qualification and points toward the applicable construction code for acceptance, while API 1104 sets its own limits for pipeline girth welds — generally tighter on crack-like flaws, more permissive on some porosity depending on service. Always confirm which code your drawing actually invokes. Guessing is how good welds get rejected.

In aerospace and similarly regulated sectors, acceptance criteria often trace back to aerospace-specific standards rather than D1.1, with far more NDT coverage. Our guide to aerospace TIG welding standards and techniques covers how that changes day-to-day work.

Timing, repairs, and reinspection

When inspection happens matters, because some steels crack after they cool. Visual inspection may begin once welds reach ambient temperature — but there is an exception with real teeth. ASTM A514, A517, and A709 HPS 100W steels require waiting at least 48 hours before visual inspection. These high-strength quench-and-tempered grades are prone to delayed hydrogen cracking, and inspecting too early can pass a weld that cracks overnight. If you weld these grades, write the 48-hour hold into your inspection request, because plenty of inspectors will not catch it unless you flag it.

Repair clauses in most codes follow a similar shape. A rejected weld is not simply ground out and redone on a hunch. The flaw is located and removed to sound metal — verified by PT or MT or re-radiography — and the repair welded to a qualified procedure. The area is then reinspected using the same method that found the defect. If RT caught the porosity, RT verifies the repair. Partial repairs that leave a blind flaw behind are worse than the original defect because they hide it. Repeated repairs also heat-cycle the base metal, so many specifications cap the number of repair cycles. Knowing when to stop chasing and start again from a clean fit-up is a skill of its own.

What this means at the bench

Two takeaways carry forward. First, inspection is a sampling of evidence, not a guarantee. A weld that passes VT alone may still hide lack of fusion; a weld that passes RT may hide a tight crack only UT would find. Know which methods apply to your job and weld to that level of scrutiny. Second, the porosity rules come down to restraint control in practice — travel speed that freezes the pool too fast, a contaminated filler rod, a slightly off gas flow. None of these feel like a violation mid-run, but they all trap gas in a pool that cannot release it. The next section covers the setup checklist and repair steps that keep them from happening.

Prevention Playbook: Setup Checklist, Repair Steps, and Common Mistakes

Every defect covered in this article traces back to one of four controllable variables: shielding, cleanliness, heat input, and arc termination. The good news is that none of them require expensive equipment to control. They require a routine. The following playbook is the one worth taping inside your cabinet door.

Pre-Weld Setup Checklist

Run through this list before you strike an arc, every single time:

• Clean the joint. Remove mill scale, oil, paint, and oxide with the right solvent and a dedicated stainless brush. Aluminum prep deserves its own routine, covered in Aluminum Cleaner for Welding: How to Prep Aluminum Right.
• Dry the filler. Store rods clean and dry; wipe them down before use, and never feed rod that has been lying in a dusty shop drawer.
• Leak-check the gas path. Check hoses, connections, and the regulator for leaks with approved leak-detection solution. Use welding-grade shielding gas only. Moisture or aspirated air in the gas stream is the most common cause of contamination, and AllGas Welding Supply: Gases, Services & Safety Guide explains the grades worth buying. Never use Ar-O₂ or Ar-CO₂ GMAW shielding blends for TIG.
• Prepare the tungsten correctly. Grind longitudinally on a dedicated, contamination-free wheel, at the right included angle for the application, with a tip blunt enough to carry current.
• Set flow and fit a gas lens. A gas lens lets you run lower, calmer flow with better coverage.
• Screen the draft. Fans, open doors, and forklift traffic destroy shields invisibly.
• Program post-flow and crater fill. Post-flow of roughly one second per ten amps keeps the hot electrode from oxidizing as it cools. Details on why that matters are covered by TWI’s equipment overview.

Parameter Starting Points

Use manufacturer charts as the final word, but a reliable starting framework for DCEN on mild steel is roughly one amp per 0.001 inch (25 mm) of material thickness, adjusted upward for material that acts as a heat sink and downward for out-of-position work. Match electrode size to current so the tungsten stays cool enough not to spit:

Tungsten diameter	Typical DCEN range	Good starting use
1/16 in (1.6 mm)	30–80 A	Sheet metal, thin tube, autogenous work
3/32 in (2.4 mm)	60–150 A	The general-purpose shop size
1/8 in (3.2 mm)	100–250 A	Heavy plate, long arcs, AC on aluminum

If you are still building confidence with parameters and procedure qualification, formal training helps more than forum advice; Aerospace TIG Welding: Standards, Techniques, and Training shows how tightly controlled professional programs are run.

How to Repair a Rejectable Defect

1. Identify it precisely. Mark the indication and confirm its type and extent with VT, PT/MT, or RT/UT as appropriate.
2. Remove to sound metal. Grinding, gouging, or machining must eliminate the entire defect with no residual indication left behind.
3. Clean the repair area back to bright, dry, uncontaminated base metal and filler.
4. Requalify the procedure. Repair welding must follow a qualified welding procedure specification; AWS D1.1 sets the governing rules for structural work.
5. Reweld using correct parameters, shielding, and crater-fill technique.
6. Reinspect with the same method that found the original defect, at the same acceptance level.

Common Beginner and Professional Mistakes

• Beginners: dipping the tungsten, welding over paint or oxide, running arc lengths that wander, forgetting crater fill, and cranking amperage to compensate for poor fit-up.
• Experienced welders: skipping the gas leak check, reusing contaminated tungsten, accepting a “cosmetic only” repair without requalification, and trusting a beautiful bead without ever testing the root. Workshops reporting shop-floor failure patterns in Michigan Tech’s welding discontinuities webinar keep seeing the same avoidable causes.

Also remember the baseline safety obligations: OSHA 1910.252 requires ventilation and respiratory protection where welding fume exposure cannot be controlled by other means.

The bottom line: porosity comes from shielding failures, inclusions from electrode handling, fusion defects from heat balance, and cracks from termination. Shielding, cleanliness, heat input, and arc termination are all things you set before you strike the arc and check while you weld. Control those four, and the vast majority of TIG welding defects never have a chance to form.

Frequently Asked Questions

What is the most common TIG welding defect?

Porosity is by far the most common, caused by loss of shielding, contamination, or travel speed that freezes the pool before gas escapes. It shows as fine distributed pores, surface-breaking pores, wormholes, or crater pipes.

Why do I keep dipping the tungsten and getting tungsten inclusions?

Dipping usually comes from an arc length that is too long for your amperage, an over-sharpened taper, or fatigue. Regrind the electrode longitudinally, use the correct diameter for the current, shorten the arc, and cut out the contaminated area before rewelding.

How much shielding gas flow does TIG welding need?

Enough to exclude the atmosphere without creating turbulence—too little lets air in, and too much aspirates it. Check for line leaks first, use a gas lens for laminar flow, and shield the arc from drafts with screens.

Are TIG welding cracks ever acceptable?

No. Under AWS D1.1 any crack is unacceptable regardless of size or location. Cracked material must be removed to sound metal and rewelded under a qualified procedure.

Why does my weld end in a cracked or sunken crater?

The last pool solidifies with shrinkage and no filler to compensate, leaving a concave, highly strained crater that cracks easily. Use the downslope or crater-fill function and keep adding filler as the arc fades.

How do I prevent porosity when welding dirty or coated steel?

Clean the joint edges back to bright metal and degrease them, because paint, primer, zinc and oil release gas and hydrogen into the pool. Keep coatings within the manufacturer’s recommended thickness and weld to a clean, dry surface.

What causes lack of fusion in TIG welds?

Settings that are too cold, a torch that is not centered, wrong work or travel angle, and travel speed that lets the puddle run ahead of the arc. Increase heat input, hold correct angles, and give the arc time to melt both joint faces.

How do you inspect TIG welds for internal defects?

Visual inspection catches surface flaws, while penetrant or magnetic particle testing finds surface-breaking discontinuities. Subsurface porosity, tungsten inclusions and lack of fusion require radiography or ultrasonic testing.

Can a TIG weld with porosity be repaired by just welding over it?

No. The defective area must be ground or gouged out to sound metal, cleaned, and only then rewelded once the root cause is fixed. Welding over trapped pores hides the defect and can trap more gas.