T Joint TIG Welding: The Complete Guide to Perfect Fillet Welds

You clamped two pieces of steel at a perfect 90 degrees, struck the arc, and an hour later the joint looks like a row of lumpy caterpillars sitting on top of the metal instead of fused into it. That gap between a T joint that merely sticks and one that wets out smoothly into both plates is the single biggest skill test in TIG welding — and it is completely fixable once you understand the geometry, the heat and the technique.

What Is a T Joint and Why TIG Welding Makes It Harder Than It Looks

A T joint is easy to describe and genuinely hard to weld well. It forms when two members meet at roughly 90 degrees, one sitting perpendicular to the other like the letter T, joined by a fillet weld laid in the inside corner on one or both sides. There is no groove to fill, no butt seam to line up, and no backing bar to hide behind. The weld sits out in the open, visible from every angle, and it has to fuse into two surfaces that behave completely differently under the arc.

In practice, a T joint in TIG welding comes down to depositing a fillet weld in that corner. Whether it looks like a professional joint or a cold, spattery mess depends on four things working together: joint preparation, correct amperage, filler cadence, and torch angle. Get them right and the puddle wets out smoothly into both plates. Get them wrong and the bead sits on top of the metal instead of biting in.

Where You Run Into T Joints Every Day

You probably weld more T joints than you realize. Common examples include:

• Roll cage tube junctions, where a diagonal or horizontal tube meets the main hoop near a right angle and the weld carries real structural load.
• Box sections and equipment brackets, where a stiffener or mounting tab is welded to the face of a plate or tube.
• Sheet metal brackets and enclosures, including aluminum enclosure fabrication where heat control is everything.
• Stainless food-grade frames, where the weld must be smooth, cleanable, and free of crevices.
• Aircraft-style aluminum fabrication, where weight limits and alloy sensitivity leave almost no room for error.

That last category deserves a pause. Aluminum is where T joint technique separates comfortable welders from frustrated ones, and it connects to the wider world of Aerospace TIG Welding: Standards, Techniques, and Training. The same corner geometry that makes a steel T joint manageable can turn brutal once the base metal starts pulling heat out of the puddle.

Why the Corner Fights Back

TIG, or GTAW, is the most heat-controlled arc process available to a fabricator. The pedal gives you instant, analog amperage control, and you can hold a puddle nearly still without depositing much metal. That control is exactly what a fillet weld needs. The catch is that TIG delivers less preheat from the arc than MIG does, so an open corner puts the burden of heat distribution entirely on you. One plate heats fast while the other stays comparatively cold, and the puddle has to bridge that difference.

Three specific things make the T joint harder than it looks:

1. Gravity works against you. Unlike a groove weld, the fillet puddle is open on its upper face. Too much heat or too slow a travel speed and the molten metal sags or rolls out of the corner before it solidifies.
2. Heat splits unevenly. The arc concentrates more energy into whichever plate presents more surface area to it. The vertical member often stays cooler than the horizontal one, so fusion into that side fails first.
3. You manage a three-sided puddle with one arc. The root, the vertical toe, and the horizontal toe all need fusion at once, and the arc can only point in one direction.

The classic result is a bead that looks acceptable on the surface but has no fusion at the root, the single most common T joint failure mode. As Miller’s common TIG weld problems guide explains, poor fit-up, incorrect torch angle, excessive travel speed, and too little heat or filler all contribute to that hidden lack of fusion.

What the Code Says About the Geometry

You do not have to guess at what counts as a proper T joint. AWS D1.1 treats fillet welds on T joints made at angles between 80 and 100 degrees as prequalified, which covers the vast majority of shop fabrications without requiring a separate procedure qualification for the geometry itself. The code also defines the measurement that matters most: the effective throat is the shortest distance from the joint root to the weld surface. That number, not the visible leg length, is what determines whether a fillet weld is big enough.

Understanding that distinction early saves a lot of rework. Plenty of welders look at a bead, decide it reads thin, and pile on more filler when the throat already meets the requirement. Others weld a bead that looks right but has barely touched the root. Both mistakes come from judging the weld by appearance instead of by geometry.

Hold onto the core idea: a T joint asks you to feed one arc into two unequal heat sinks while gravity pulls the puddle down, and TIG gives you the finesse to do it but not the free preheat that easier processes provide. Everything that follows, from amperage settings to filler selection to torch angle, exists to solve that one problem.

Fillet Weld Geometry: Leg Size, Throat Thickness and What Codes Require

A fillet weld is easy to draw and easy to misjudge. In a T joint the weld sits in the inside corner where the web meets the flange, and its size gets described two different ways: the leg length you can see, and the effective throat you actually have to trust. Codes care about the throat. Confusing the two is why welders look at a bead that measures perfectly and declare it undersized, or stack extra passes on a joint that was already strong enough.

Leg Length vs Effective Throat: The 0.707 Rule

Leg length runs from the root of the joint out along one plate to the toe of the weld. Effective throat is the shortest distance from the root to the face of the weld, which is the narrowest cross-section the load has to pull through. On an equal-leg fillet laid at 45 degrees, that shortest path runs diagonally across the corner, and the trigonometry gives a fixed ratio, as covered in standard fillet weld geometry references:

Leg size	Effective throat (0.707 × leg)
1/8 in (3 mm)	0.088 in (2.1 mm)
3/16 in (5 mm)	0.133 in (3.3 mm)
1/4 in (6 mm)	0.177 in (4.5 mm)
3/8 in (10 mm)	0.265 in (6.7 mm)

So a 1/4 inch leg on each plate gives an effective throat of roughly 0.177 inch, not 1/4 inch. The bead carries about 70 percent of the leg you laid down. With unequal legs the relationship gets even less flattering: the throat equals (a × b) divided by the square root of (a² + b²), and the smaller leg drags the result down fast.

Minimum and Maximum Fillet Sizes Under AWS D1.1

Structural codes set floors and ceilings on fillet size, and both matter in a T joint. The AWS D1.1 Structural Welding Code sets a minimum fillet weld size against the thickness of the thicker part joined:

• Material 1/4 in (6 mm) or less: minimum 1/8 in (3 mm) leg
• Over 1/4 in to 1/2 in: minimum 3/16 in (5 mm)
• Over 1/2 in to 3/4 in: minimum 1/4 in (6 mm)
• Over 3/4 in: minimum 5/16 in (8 mm)

These minimums exist because a tiny weld on thick plate cools too fast, hardens the heat-affected zone, and fails by brittle fracture rather than by tearing the weld itself. Thin sheet is the documented exception, where codes allow smaller legs on thinner stock, which is exactly the situation you meet when repairing thin aluminum components such as the wheel work described in our Alloy Wheel Welding: A Comprehensive Guide to Safe and Effective Aluminum Wheel Repair.

The ceiling works the other way. AWS D1.1 restricts the maximum fillet weld size along the edge of material less than 1/4 inch thick to the full thickness of that material. Weld past that and you are no longer joining two plates, you are building a lip that hangs over the edge, fusing through thin material, warping the corner and creating a notch where the bead rolls over. For material 1/4 inch and thicker, the code keeps the edge fillet slightly under the thinner member’s thickness unless the detail specifically calls for welding to full penetration.

Convex, Concave and Overlap: Reading the Bead Profile

Profile is where geometry and aesthetics collide:

• Convex beads bulge outward. A modest crown is fine and even desirable, but heavy convexity means slow travel, too much filler, or both, and the abrupt toe angle becomes a stress raiser.
• Concave beads dish inward. They look tidy and delicate, and they are the classic throat-killer: the leg measures out to size while the actual throat falls short of it.
• Overlap is weld metal spilling over the toe without fusing into the plate. It masquerades as a big leg on visual inspection while contributing nothing structurally, and it hides a cold, unfused edge underneath.

Measuring and Accepting a T Joint Fillet Weld

The tool for the job is a fillet weld gauge: a thin stamped or machined set that hooks into the corner and reads leg and throat in one pass, with notches for checking concavity and the smaller legs typical of Tee joints. Run the throat scale against the root, and you will know instantly whether a bead you think is small actually meets the required size. Acceptance then follows the governing standard, AWS D1.1 for structural steel or ISO 5817 for work specified to its quality levels B, C or D, each of which sets allowable deviations for leg size, concavity, convexity, overlap and incomplete fusion. The habit worth building is simple: decide the required throat first, back-calculate the leg (throat × 1.414), weld to that, and gauge it before the joint cools. Geometry does not care which process laid the bead.

Machine Setup for TIG T Joints: Amperage, Polarity, Gas and Tungsten

Before you ever strike an arc on a T joint, your machine setup will do more for fusion than any fancy hand movement. With the tight 90 degree corner, there’s very little room for error. Dialing in your polarity, gas, tungsten, and amperage ahead of time is the difference between wetting smoothly into both legs, or having a bead that simply sits on top of them.

Polarity: Get The Heat Where It Needs To Go

The first thing to lock in is polarity. For mild steel and stainless steel T joints, you’ll want DCEN (Direct Current Electrode Negative). This configuration drives the majority of the heat into the workpiece, allowing you to fuse both plates of the fillet without dumping unnecessary heat into your tungsten.

For aluminum and magnesium T joints, AC is non-negotiable. The electrode positive half of the AC cycle is what breaks up aluminum’s tenacious oxide layer, whilst the electrode negative half delivers the heat into the puddle. Without this cleaning action, fusion into the corner becomes nearly impossible. When prepping for this, it’s worth keeping in mind proper aluminum cleaning practices to get the most out of your AC settings.

Gas Coverage, Cups and Postflow

For all of the materials above, pure argon is your go-to shielding gas. In clean shop conditions, aim for an argon flow rate of 15-20 CFH (7-9 L/min). If you’re running a gas lens with a #7 or #8 cup, you can often sit closer to the lower end of that range whilst still obtaining adequate coverage. With a standard collet body and cup, bumping closer to 17-20 CFH is the safer route, especially if there’s any draft around your work station.

Don’t neglect postflow either. Setting a sufficient postflow time will protect both your red hot tungsten and your cooling fillet bead from oxidation. A reliable field starting point is around one second of postflow per 10 amps of welding current, fine tuning from there based off of your cup and gas lens setup.

Tungsten Size, Type and Grind Orientation

Arc focus is king inside of a T joint, so your tungsten needs to match both your process and amperage. For DCEN on steel and stainless, 2% lanthanated (EWLa-2) is incredibly consistent. For AC on aluminum, pure tungsten (EWP) or zirconiated (EWZr-1) are both solid picks as they form a stable, balled tip under AC current.

As per guidance from Miller Welds, 1/16″ (1.6 mm) tungsten is suitable for most thin sheet T joints, scaling up to 3/32″ (2.4 mm) as amperage and thickness increase. Always grind lengthwise, axial to the tungsten, as opposed to circumferentially. For DC, you’ll want a crisp, tapered tip, whilst for AC you’ll want to allow that controlled ball to form.

Amperage, Pedal Control and Downslope

The most widely used starting guideline amongst TIG welders is 1 amp per 0.001 inch (0.025 mm) of material thickness. As a more practical field rule of thumb for steel, many welders start around 40-50 A per 1/16 inch (1.6 mm). With that being said, a T joint will almost always require slightly more amperage than a flat butt weld on the same thickness, as you’re having to heat two plates at the same time.

For this exact reason, pedal control is invaluable. Rather than setting a fixed, high amperage and hoping for the best, use your foot to ease into the corner to establish a puddle, hold a stable heat whilst dabbing filler, and feather down as you near the end of your bead. Utilizing downslope on your machine paired with that pedal will drastically reduce the chance of crater cracking at your stop points.

Practical Starting Settings

Material	Polarity	Thickness	Starting Amps	Tungsten	Gas (CFH)
Mild Steel	DCEN	1/16″ (1.6 mm)	40 – 55 A	1/16″ 2% Lanthanated	15 – 18
304 Stainless	DCEN	1/16″ – 3/32″ (1.6 – 2.4 mm)	50 – 80 A	1/16″ – 3/32″ 2% Lanthanated	16 – 19
6061 Aluminum	AC	1/8″ (3.2 mm)	90 – 110 A	3/32″ Pure or Zirconiated (Balled)	17 – 20

These values are only starting points. Fit-up, joint tightness, and your machine’s characteristics will all shift them slightly. Focus less on hitting these numbers exactly, and more on achieving a tight, controlled puddle that wets out to both legs of your T joint. That control will carry directly into your puddle management for the rest of your weld.

Choosing the Right Filler Metal for a T Joint

In a T joint, the filler metal is not just a supplement to the base metal, it essentially is half the weld. A fillet weld is built almost entirely out of deposited rod, so the alloy you choose and the diameter you feed decide how easily the puddle wets into both plates, how fast undercut shows up at the toes, and whether the finished joint survives a load. Too many welders grab whatever rod is closest on the bench. In a corner joint that habit is exposed immediately, because a fillet puddle sits open on three sides and gives you nowhere to hide a bad filler choice.

Match the Filler to the Base Metal First

The first rule is simple: match the filler family to the material you are joining, then fine-tune within that family.

Base metal	Filler	Why it works
Mild steel, general fabrication	ER70S-2	Deoxidizers tolerate light mill scale and minor contamination, the most forgiving everyday choice
Mild steel, clean and critical	ER70S-6	Slightly higher deoxidizer content, smoother arc on clean fit-ups
304 stainless	ER308L	Low-carbon version that resists carbide precipitation in the heat-affected zone
304 to mild steel	ER309L	Higher alloy content buffers the dilution between dissimilar metals
Aluminum, general and cosmetic work	ER4043	About 5% silicon, lower melting temperature, smoother and far less crack-sensitive puddle
Aluminum, structural work	ER5356	Higher strength, better for joints that carry load or see elevated service heat
Aluminum castings and mixed jobs	ER4943	Silicon-modified 4043-type filler with improved strength over plain 4043

ER70S-2 deserves its status as the default mild steel TIG rod. The “-2” designation means a specific deoxidizer package rather than a strength difference, and that package is what lets it burn through light mill scale and a bit of oil residue without porosity exploding. On a structural stainless T joint, stick with the low-carbon “L” grades such as ER308L, especially on thin material where you cannot afford carbide precipitation. For aluminum, 4043 is the everyday fillet filler because the silicon lowers the melt temperature and produces a fluid puddle that flows into a corner without crawling. Move to 5356 when joint strength matters more than appearance, or when the part will be anodized, though the color match after anodizing is poor with 4043. The Hobart Brothers aluminum welding guide is a solid reference for this selection logic.

Match Filler Diameter to the Weld Throat, Not the Plate

A common beginner habit is to pick a rod diameter equal to the plate thickness. In a butt joint on thin sheet that heuristic works, but in a fillet weld it fails, because the rod feeds the throat of the weld rather than the sheet. A 3/16 inch plate welded with a 3/16 inch leg needs roughly a 1/16 inch or 3/32 inch rod, not a heavy 3/16 inch rod that will freeze in the corner before it wets out.

Useful working ranges:

• 1/16 inch (1.6 mm) rod: sheet and light fillets with a leg up to about 3/32 inch.
• 3/32 inch (2.4 mm) rod: the everyday workhorse for mild steel and stainless fillets in the 1/8 to 3/16 inch plate range.
• 1/8 inch (3.2 mm) rod: multi-pass fillet work and heavier structural joints where each dab needs to deposit real volume.

If you are welding out of position, a smaller rod helps you stay ahead of gravity, as covered in our All Position Welding Rod Guide.

Filler Is Your Cheapest Heat-Control Tool

Here is the part most beginners miss: filler metal is a heat sink. Every rod you dab into the puddle pulls heat out of it, and in a T joint that matters enormously. Because the fillet sits in an open corner, heat builds fast at the toes and undercut starts forming almost immediately. Rather than dropping amperage on the pedal every time the puddle gets wide, a skilled welder simply feeds more rod, and the puddle tightens up. This is why a cramped bead with an underfilled toe is often a filler-cadence problem, not an amperage problem, a pattern called out in Miller’s common TIG weld problems guide and in undercut diagnostics from CK Worldwide.

Think of it as a dial: slow, generous dabs cool and build the throat; a thin, hesitant feed lets the arc dig and undercut. Practicing a consistent dab rhythm on a corner tie is time better spent than chasing the perfect machine setting, and the next section walks through exactly how to run that bead.

Step-by-Step Technique: Welding a Mild Steel T Joint with TIG

Now we strike an arc. This walkthrough assumes a simple mild steel T joint: two pieces of hot-rolled plate, about 1/8 inch (3.2 mm) thick, cleaned to bright metal, tacked into a square corner and sitting flat on the bench. The sequence transfers to stainless directly, and to aluminum once you switch to AC.

Fit-Up and Tacking: The Weld Is Decided Before the Arc Starts

Set the vertical member (the web) dead center on the horizontal member (the flange) and check the corner with a small square. Gaps wider than about 0.5 mm (0.020 in) will eat filler and pull heat into the root, so clamp or shim until the joint is tight. Then place a tack at each end, roughly 10–15 mm long, using the same amperage and filler you plan to weld with. On joints over 300 mm, add a mid-tack and work outward with your sequencing, which we will come back to in the distortion section.

Useful habit: tack, re-check square, then tack again. Tacks shrink as they cool and will pull a light assembly out of true if you never re-measure.

Torch Angle: Bisect the Corner, Then Lean Slightly

The single most common cause of a lumpy, un-fused T joint is bad geometry. Hold the torch so the tungsten bisects the inside corner at a 45-degree work angle, pointed equally at both plates. Then add a 5 to 15 degree travel angle — a slight drag if you are right-handed, or a slight push. Most welders do best dragging.

Keep tungsten stickout short, around 1.5 to 2 times the cup diameter, and watch the arc cone. If one side of the corner washes before the other, fix the work angle, not the amperage. General technique guidance from mig-welding.co.uk’s TIG technique guide makes the same point: angle problems show up as asymmetry before lack of fusion.

Establishing the Puddle and Reading Its Width

Start the arc on one plate slightly off the corner, not directly in the root, and let the puddle form. Move it into the corner and hold until you see both plates wetting. When set correctly, the puddle should spread to roughly 1.5 times the filler rod diameter before you add anything. That width is your real amperage gauge. Too narrow, the arc is cold or you are moving too fast. Blooming wider almost immediately, it is too hot or you are dwelling too long.

Feather the pedal rather than starting at full amperage. Miller’s troubleshooting material on common TIG weld problems repeatedly ties lack of fusion in fillet joints to insufficient heat and excessive travel speed, so resist the urge to hurry.

Dab Cadence: Feed the Leading Edge, Not the Trailing Pool

Move forward in small increments and touch the filler into the leading edge of the puddle — the front of the pool, ahead of the arc. The rhythm is roughly one dab per half-torch-width of travel. If you stir the rod into the trailing pool behind the arc, you get a cold lump with no throat. The rod should disappear into the puddle the instant it touches; if it balls up, the puddle is cold.

On a 45-degree fillet you are also using filler as a heat sink. A hot, thin puddle with undercut developing at the top toe gets fixed by adding more rod, not by trimming the amperage.

Restarts, Craters and Finishing

Long T joints are welded in segments, and every stop is a risk. Crater cracks form because the arc leaves a shrinking, starved pool. Ease off the pedal over the last 10–15 mm, dropping current gradually, then dab two or three extra shots of filler into the crater before breaking the arc. If your machine has downslope, set 2–4 seconds. A Performance Racing’s TIG troubleshooting guide lists crater cracking among the most avoidable TIG defects.

When restarting, begin the arc about 10 mm behind the crater, run a short puddle back into it, then resume normal cadence. Never restart directly on the cold crater; you will get a cold-lap seam.

A Note on Work Position

Everything above describes a flat fillet — the horizontal plate flat, the web standing up, weld running along the corner. This is the easiest T joint position and where you should learn. Overhead fillets flip the physics: gravity pulls the puddle and filler out of the corner, so you need more filler per dab, less travel angle, and a cooler setting to prevent sagging. If you are still struggling with flat fillets, do not move to overhead yet. Position work becomes unavoidable in advanced work like aluminum boat welding.

Run this sequence on scrap coupons: tight fit-up, 45-degree work angle, slight drag, puddle width as your gauge, rod dabbed at the leading edge, and a filled crater to close out.

Welding Aluminum T Joints: The Real Difficulty Spike

If a mild steel T joint tests your heat control, an aluminum T joint tests everything at once: cleanliness, arc settings, timing and patience. Welders who run clean steel fillets often get porous, underfused or collapsed corners on aluminum the first time they try. Aluminum is not unforgiving by nature. Three physical facts simply work against you at the same time, and each one has a specific countermeasure.

Why Aluminum Oxide Changes the Whole Game

Pure aluminum melts at roughly 1,220°F (660°C). The oxide layer that reforms on its surface the moment you stop welding melts at about 3,700°F (2,050°C), more than three times higher. That mismatch is why aluminum welding behaves nothing like steel welding: your arc has to burn through a crust that will not melt at the temperature the base metal turns liquid. Left in place, the oxide gets trapped as a filmy inclusion in the puddle, blocks wetting at the toes and produces the classic root lack-of-fusion failure in a T joint.

This is why mechanical prep matters more on aluminum than on any other common material. Wipe the joint with acetone to remove oil, then scrub the fusion zone with a dedicated stainless steel wire brush you have never touched carbon steel with, until the surface is uniformly bright. Our Aluminum Cleaner for Welding: How to Prep Aluminum Right guide walks through solvents, mechanical cleaning and the mistakes that leave invisible contamination behind.

AC Balance: Cleaning Versus Penetration

Aluminum is welded on AC, and the balance setting decides how much of each half-cycle you get. During the electrode-positive (EP) portion the arc blasts oxide off the surface; during the electrode-negative (EN) portion it drives heat into the joint for penetration. Push EN percentage higher and you get deeper, narrower penetration but weaker cleaning. Push it toward more EP and the surface cleans beautifully, but the arc softens, the tungsten wears faster and more heat goes into the torch rather than the work. The Fabricator’s aluminum TIG settings article is a solid reference for finding a practical starting window and adjusting it by what the puddle tells you.

Frequency is the second dial. An output frequency around 150–250 Hz is recommended on thin aluminum because it focuses and tightens the arc, giving you precise placement in the corner and less heat spread into the surrounding plate, exactly what you want on a fillet where one plate is already fighting to stay cool. On thicker material you can drop back toward 100–120 Hz for a wider, softer arc. The Hobart Brothers aluminum welding guide covers balance, frequency and filler selection in more depth.

Thermal Conductivity: The Corner Betrays You

Aluminum conducts heat several times better than steel. In a T joint, that means heat poured into the corner is dragged away as fast as you add it, into the horizontal plate, into the vertical plate and out along both members. The result is a nasty asymmetry: the corner you are aiming at can slump and collapse while the vertical member is still below fusion temperature. And aluminum gives no color warning. Steel glows, dulls and shifts hue as it approaches its limit; aluminum sits there looking unchanged and then suddenly goes shiny and drops through.

The practical answers are:

• Dwell at the root. Hold the arc in the corner a beat longer than feels right so both members reach fusion temperature together.
• Feed more filler, more often. Each dip of rod removes heat and builds throat thickness, so a heavier cadence both cools a hot puddle and adds the throat you are trying to achieve.
• Keep travel speed honest. Speeding up to save a sagging bead is what creates the underfilled, unfused root.
• Preheat thick sections carefully. A modest preheat helps heavy plate (roughly 3/8 in and above) start closer to temperature, but keep it low and controlled: overheating before you strike an arc reduces the alloy’s properties and invites distortion.

When the corner heats too fast, start the weld on the thicker-feeling side and let the puddle carry heat into the thin one, then flood the transition with filler. On larger fabrication work such as Aluminum Boat Welding: The Complete Guide for Strong Hull Joints, where nearly every seam is a fillet in thin plate, this filler-first rhythm is the whole technique. And if TIG productivity becomes the bottleneck, it is worth comparing processes in our Aluminum MIG Welding: The Complete Practical Guide — though for a visible corner fillet with tight heat control, TIG remains the choice.

Aluminum TIG setup details are also well covered in Miller’s beginner aluminum TIG walkthrough. Scrub the joint like your weld depends on it, set the balance to clean without washing the puddle flat, run frequency high enough to hold a tight arc, and give the corner more filler than you think you need.

Common T Joint TIG Welding Defects and How to Fix Them

Fillets don’t fail randomly. Almost every bad T joint you cut apart fails in a predictable way, and each defect leaves a signature you can read in seconds. T joint fillets are judged on size, profile, undercut, porosity and fusion, and once you know what each looks like at the arc, the fix is usually a small adjustment to fit-up, angle, heat or filler cadence rather than a new technique.

T Joint Defect Troubleshooting Table

Defect	What you see	Likely causes	The fix
Lack of fusion at root	Dark line or gap where the plates meet; coupon shows no root penetration	Poor fit-up, long arc, wrong angle, fast travel, not enough heat	Close the gap, shorten the arc, correct the angle, slow down, add amps
Undercut at the toes	Groove melted into the metal along a toe	Amps too high for travel speed, poor angle, filler skipped at the toes	Drop amps, aim the arc at the toe, dab filler there, slow down
Porosity	Pinholes or wormholes in the bead, often at a restart	Oily or oxidized plate, draft on the gas, long arc, contaminated filler	Clean and degrease, shorten the arc, shield the area, use fresh rod
Insufficient weld size	Leg or throat shorter than the drawing calls for	Fit-up gap ate the filler, travel too fast, rod too small	Fix fit-up, slow down, use a bigger rod, or add a pass
Overlap	Filler rolled over the toe without wetting in; cold edge	Travel too slow, too much filler, not enough heat	Speed up, add heat, wet the toe before adding rod
Hot cracking	Crack down the bead center, sometimes a ping on cooling	Wrong filler for the alloy, fast travel, open crater	Use the right filler, slow down, taper current, fill the crater
Spatter on AC	Frozen metal balls stuck to both plates around the weld	Cleaning balance too high, dirty surface, long arc	Reduce the cleaning balance, scrub the aluminum, shorten the arc

Lack of Fusion at the Root: The Classic T Joint Killer

This defect ruins more T joints than everything else combined, because it hides: the bead can look decent from the front while never tying into the corner. Miller’s guidance on common TIG weld problems names the same causes — improper fit-up, a torch held too far away, wrong angle, excessive travel speed, insufficient heat and not enough filler.

A T joint gives the arc two cold mass sinks and an open corner, so heat drains from the root faster than in a butt joint. Fix order: close fit-up until the plates touch, shorten the arc to 1/8 inch or less, aim into both members, then slow until the puddle wets both toes before adding rod. If it still will not tie in, you are short on amperage.

Undercut at the Toes

Undercut is a groove along either toe, usually from amperage too high for travel speed or an arc aimed at one plate instead of the corner. Each dab at the toe refills the crater the arc just carved. CK Worldwide’s breakdown of undercut in TIG welding covers the causes and the code limits, and the fix is consistent: trim amperage, correct the angle, slow down, feed the toe. On a hot fillet, more rod often fixes undercut faster than less heat.

Porosity and Contamination

Porosity in T joints is usually a cleanliness or shielding problem. Oil, mill scale, cutting fluid and moisture trapped in the corner boil out under the arc and gas the puddle. Aluminum is worse, because the oxide layer and residual cleaner release gas at welding temperature, so a scrub with a stainless brush and a solvent wipe is the cheapest insurance, covered in Aluminum Cleaner for Welding: How to Prep Aluminum Right. Drafts, a too-small cup and cold restarts compound the problem.

How to Confirm an Acceptable Weld

Do not trust your eye alone. Inspection of a T joint fillet covers size, profile, undercut, porosity and fusion, and a fillet weld gauge measures leg length and throat against the drawing. Check both legs; a weld can pass on one and be short on the other. Then compare against AWS D1.1 acceptance criteria and, for tighter work, ISO 5817 quality levels. Where root fusion is required, a fillet is the wrong joint; that needs a groove or partial-penetration detail, as covered in Aerospace TIG Welding: Standards, Techniques, and Training.

The pattern behind every fix is the same: tight fit-up, clean metal, a short arc, and enough filler to build the throat the drawing asks for.

Distortion, Tacking and Heat Management in T Joints

You can lay down a beautiful fillet bead and still end up with a ruined assembly. T joints are distortion magnets. The weld sits on one side of the joint, so its shrinkage shortens that side as it cools and swings the attached plate toward the heat. On a roll cage junction or a machine base, a few degrees at each corner compounds into a frame that rocks on the bench. The good news is that distortion is predictable, and once you know why it happens you can fight it with sequencing, restraint and pedal discipline rather than brute force.

Why T Joints Distort the Way They Do

A fillet weld shrinks as it solidifies and cools. In a T joint, that shrinkage force acts on one face of the vertical plate, not symmetrically through it. The weld side shortens, the far side does not, and the web pivots toward the weld like a door closing on a shrinking hinge. This angular distortion is the dominant mode in fillet-welded T joints, and it grows as the weld gets hotter and bigger relative to the plate.

Restraint is the counter-force, but it is a double-edged sword. Clamping the joint to a rigid fixture or welding it to strongbacks holds alignment while the metal cools. However, restrained joints accumulate shrinkage stress instead of relieving it through movement. On ductile mild steel that is usually fine. On stiff, high-strength or heat-treated materials, heavy restraint without proper sequencing can invite cracking, so the goal is controlled restraint, not maximum restraint.

Sequencing: Let Shrinkage Cancel Itself Out

The most reliable way to reduce net distortion is to spread the shrinkage along the joint instead of concentrating it in one continuous pass. Two proven patterns do this:

• Back-step welding: Divide the joint into short segments and weld each segment in the opposite direction of overall travel. The shrinkage of each segment pushes against its neighbor, and the errors largely cancel.
• Skip (skip-weld) sequencing: Jump around the joint — weld a segment at one end, then one in the middle, then the opposite end, filling the gaps afterward. This distributes heat and keeps any one area from getting hot enough to pull hard.

On a long T joint such as a stiffener rib, a middle-out pattern also works well: start at the center and work toward each end alternately, swapping hands so both halves receive similar thermal cycles. On two-sided T joints or H-shaped assemblies, welding one side then the other keeps the pull balanced.

The Pedal Is Your Distortion Tool, Not the Dial

A frequent habit on benchtop TIG machines is to set the amperage high enough to “get it done” and then fight the puddle. That adds heat for no benefit. Heat input in arc welding is a function of voltage, current and travel speed, and controlling heat input is central to limiting distortion. TIG is uniquely suited to this because the pedal lets you modulate current continuously: back off when the puddle widens, add a quick pulse of current when a corner sucks heat, and coast through crater-fill at the end.

Compared to processes like FCAW or spray-transfer MIG, TIG’s lower current and slower, more deliberate travel speed typically produce lower heat input per inch. That is a genuine structural advantage, and it is why aluminum fabricators often prefer TIG for thin T joints even when MIG would be faster — although Aluminum MIG Welding has its own place on thicker sections.

Preheat, Interpass Temperature and Thick Steel

On thick carbon steel T joints, preheat serves two purposes: it slows cooling to avoid hard, brittle microstructures in the heat-affected zone, and it reduces the temperature gradient between the weld area and the surrounding metal, which softens the shrinkage pull. Interpass temperature matters the same way — keep it consistent, and avoid letting a long T joint run progressively hotter toward the end of the bead. Check the filler manufacturer’s or your procedure’s limits for the alloy family; for unknown or high-carbon material, a modest 150–250°F (65–120°C) preheat is a common field starting point, and AWS D1.1 addresses preheat requirements based on material thickness and composition.

Clamps, Strongbacks and Practical Fixturing

A simple fixture kit pays for itself quickly: C-clips, a flat steel table with slots, a square, and a few pieces of scrap used as backing bars. For a bracket-to-plate T joint, clamp the bracket flat, tack all four sides, unclamp, and check square with a machinist’s square before committing to the full beads. For long joints, tack strongback bars across at intervals, weld, then remove and grind the attachment marks after the assembly has cooled — sequencing and restraint work best together, as AWS technical discussion on distortion control repeatedly emphasizes.

If you work primarily in aluminum, restraint and heat strategy change because aluminum’s conductivity pulls heat out of the corner so fast. That makes cleaning, fixturing and heat placement even more critical, and the same principles show up in Aluminum Boat Welding and Aluminum Cleaner for Welding prep guides.

Distortion Control Checklist

1. Fit up tight, square and clean before anything is tacked.
2. Tack symmetrically in at least four places, then verify square.
3. Pick a back-step or skip pattern before you strike the arc.
4. Set amperage conservatively and manage heat with the pedal.
5. Keep travel speed steady and the puddle width constant.
6. Clamp or strongback long joints, then release after cooling.
7. Check alignment on a flat surface before the assembly goes cold and locked.

Distortion cannot be eliminated, but on a T joint it can almost always be kept inside what your assembly tolerances will accept. The welders who get it right rarely use special equipment — they simply plan the heat before they use it.

Codes, Inspection and Quality: What an Acceptable T Joint Looks Like

A T joint that looks decent under shop lights still has to pass somebody else’s judgment. On real work that judgment is not a matter of taste. It is a documented code with numeric limits, a qualified inspector, and a paper trail. Learning to weld the corner is one skill; learning what “acceptable” actually means is another, and it is the skill that keeps your work from getting chipped out and rewelded the next morning.

Which Code Applies to Your Joint

In the United States, most carbon and low-alloy steel T joints fall under the AWS D1.1 Structural Welding Code—Steel, which covers welding requirements for structural steel including fillet welds in T, lap, and corner joints. Two scope details decide almost everything downstream:

• Thickness. D1.1 applies to base metal 1/8 inch (3.18 mm) and thicker. Sheet thinner than that falls under AWS D1.3, Structural Welding Code—Sheet Steel, which has its own fit-up, filler, and acceptance rules for light gauge work.
• Service and industry. Pressure piping, vessels, bridges, and aerospace each carry their own specification. Airframe and precision TIG work, for example, is governed by aerospace standards rather than D1.1, as covered in our Aerospace TIG Welding: Standards, Techniques, and Training guide.

Outside the US, ISO 5817 is the common reference. It defines three quality levels for fusion welds in steel, copper, nickel, and titanium: B (most stringent), C, and D. Each sets its own limits on weld size, undercut, concavity, convexity, overlap, and porosity. The practical point: the contract or drawing must state which level applies, because “it looks fine” means something different at level B than at level D.

Measuring the Weld: The Fillet Gauge

Fillet welds are verified with a fillet weld gauge, usually a flat combo gauge with notched cutouts that let you read both leg length and the corresponding throat directly against the bead. You check each leg on both plates, because a weld carrying a full 1/4 inch leg on one side and 3/16 inch on the other is out even if the average looks right.

Two rules catch most arguments in the shop:

• Undersize is a rejection. A leg length below the specified size fails AWS acceptance criteria, full stop. The effective throat, not the visual mass of the bead, carries load, and for a 45-degree fillet that throat is only about 0.707 times the leg.
• Oversize is generally not a rejection. A fillet larger than specified is usually acceptable unless it creates a profile defect, overlaps the toe, restricts fit-up, or pushes the joint past a stated maximum.

Visual Acceptance: What Actually Gets Rejected

Visual inspection is the first and most common check, and visual examination follows a defined procedure rather than a glance: adequate lighting, cleaned surfaces, and an inspector qualified to a written practice with demonstrated near-vision acuity.

The usual rejection triggers on a T joint are:

• Undercut at the toes. D1.1’s visual acceptance table limits undercut to 1/32 inch (0.8 mm) for members in static service and 1/16 inch (1.6 mm) where cyclic loading applies. CK Worldwide’s breakdown of undercut limits in TIG welding explains why travel speed and torch angle, not just amperage, drive this defect.
• Insufficient size. Either leg below the specified minimum.
• Excessive concavity or convexity. D1.1 caps convexity by leg size in the same table; concavity is judged by whether the remaining effective throat still meets the specified size.
• Overlap at the toe. Filler that rolled onto the base metal without fusing is rejected, not ground smooth and passed.
• Cracks of any length, porosity beyond the stated limit, and incomplete fusion.

Welder qualification reflects this reality: AWS D1.1 performance tests include T joint fillet coupons covering single-bead large fillet welds and minimum-size multi-pass fillets, then sectioned or otherwise examined for soundness. If formal certification is part of your plan, a structured Aluminium Welding Course: What You Learn, Costs & Careers shows how code-based coupon testing and inspection training fit into a welding program.

When a Fillet Weld Is Not Enough

A fillet weld fuses only the surfaces of the joint. If the design requires load to transfer through the root of the T, you do not specify a bigger fillet; you specify a groove weld. AWS D1.1 distinguishes three cases:

1. Complete joint penetration (CJP) groove weld. Required for full-strength, fatigue-critical, or dynamically loaded T joints, using a backing bar, gas backing, or a double-sided joint preparation so the root fuses completely.
2. Partial joint penetration (PJP) groove weld. A grooved joint that intentionally fuses only part way through, with a stated design throat. It is a different weld type from a fillet and must be drawn and inspected as such.
3. Fillet weld. Fine wherever code, design throat calculation, and service allow it, including combined fillet-plus-groove details where each element carries its share of load.

Writing or Requesting a WPS for T Joint Fillet Welds

Every accepted weld traces back to a Welding Procedure Specification. For a T joint fillet in GTAW, a usable WPS states the base metal group and thickness range, filler classification under AWS A5.1 (steel) or A5.10 (aluminum), joint configuration, welding position, current type and polarity (DCEN for steel and stainless, AC for aluminum), shielding gas and flow, preheat and interpass temperature, any heat input limits, and the acceptance criteria to be applied. It also lists essential variables: change one, and the procedure or the welder’s qualification may need requalification.

Under D1.1 you can write or accept a WPS two ways: as a prequalified procedure, where every variable falls inside the code’s prequalified conditions and no test coupon is needed, or as a qualified procedure supported by a procedure qualification record with test results. Ask your fabricator or inspector for the WPS number before production starts on anything structural, and check that visual criteria, weld size, and any required NDE appear on it. A drawing calling for a “1/4 inch fillet” with no code, no position, and no acceptance level is an incomplete instruction, and the rework argument that follows costs more than the clarification would have.

Practice Drills, Practical Tips and Your Path to Consistent T Joints

Knowing what a good fillet looks like and producing thirty identical ones are two different skills. The gap closes through deliberate practice, not by burning scrap for the sake of it. What follows is a progression you can run at home or in a shop, from your first coupon to a routine that survives inspection.

The Plate Thickness Progression

Start on 2 mm or 3 mm mild steel. The puddle is large enough to see clearly and steer, but the plates will not collapse the instant you hold half a second too long. Thinner sheet punishes hesitation before you have built muscle memory, while thicker plate eats tungsten and gas while you are still figuring out puddle control.

• 3 mm flat beads on plate until your ripples are evenly spaced over a 3 inch run.
• 3 mm T joints, stringer fillets, run both left-to-right and right-to-left.
• 6 mm T joints, where you will need slower travel and noticeably more filler to build throat.
• Stainless, then aluminum, once your steel fillets repeat reliably. Stainless forgives fusion but punishes heat control; aluminum is a different process entirely, as covered earlier in this guide.

Buy your practice metal from a single batch so thickness and alloy stay consistent, and pick up your gas, tungsten and filler from a reliable welding supply source rather than mixing mystery rods into practice. Material consistency is what makes your results interpretable.

Four Coupon Drills Worth Your Arc Time

1. Stringer beads. Ten on flat plate with no filler, then with filler. You are training travel speed and arc length only. Aim for a bead 1/4 inch wide with straight, evenly spaced ripples.

2. Corner ties. Run 3 inch stringers into an open corner joint, tying both legs together. This is the closest drill to a real fillet, and it teaches you to watch the puddle wet into both plates rather than riding one side.

3. Step-over welds. Weld a 2 inch segment, stop, let it cool, then restart on the next. Long continuous runs are where beginners drift into bad angles and blown corners, and segment welding forces the clean restarts that production T joints demand. Miller’s troubleshooting guidance flags restarts and crater management as recurring sources of cracking and porosity in TIG work (Miller Welds, Common TIG Weld Problems).

4. Filler cadence drills. Weld a fillet while saying “dip, move, dip, move” out loud in rhythm. Consistent dab rhythm is the single strongest predictor of uniform fillet appearance. When your beads look like a caterpillar, the fix is almost never amperage; it is rhythm.

On rod choice and feed rate, our all-position welding rod guide breaks down how diameter and deposition rate change the way you feed.

Section and Etch: The Honest Self-Test

Visual inspection lies. A fillet can look perfect and still have no fusion at the root. The definitive self-training check is to cut a coupon through the weld throat, grind and polish the cut face, and etch it with ferric chloride or ammonium persulfate solution. The etched face reveals the fusion line between weld and parent metal in plain sight. If the weld sits on top of the corner instead of tying into both plates, you will see it immediately, and no amount of grinding will hide it. Etch one coupon every session for your first month; it accelerates learning faster than any video.

Keeping the Arc Stable

The most common instability in practice T joint work is tungsten contamination: touching the filler rod to the electrode. When it happens, stop, regrind the tungsten, reset, and do not keep welding while blaming the machine. Three habits cover most other arc problems: correct cup size and gas flow, a sharp correctly ground electrode, and a joint free of oil and mill scale. For aluminum, dedicated aluminum cleaning and prep supplies are non-negotiable, since oxide contamination ruins fillets invisibly.

The Five Habits That Decide Your Results

Strip everything away and consistent T joints come down to five things, in order of impact:

1. Fit-up. A gap you did not close before tacking cannot be fixed with technique.
2. Cleanliness. Oil, oxide and mill scale sit behind most porosity and lack of fusion.
3. The right amperage. Cold arcs wander and refuse to wet; hot arcs undercut the toes.
4. Enough filler. On fillets, rod is your heat sink; adding filler cools a hot puddle faster than dropping amperage.
5. Slow enough travel. If the puddle has not wet into both plates before you move, it never will.

If you would rather be taught than self-taught, an aluminium welding course covers fillet technique formally, and tight-tolerance shops publish useful process references; our aerospace TIG welding guide summarizes the standards involved.

Conclusion: How a T Joint in TIG Welding Is Actually Made

To answer the question this guide opened with: a T joint in TIG welding is made by depositing a fillet weld into the inside corner where two members meet, and the result is judged by fusion at the root and correct throat geometry, not by how the bead looks from above. Fit the joint honestly, clean it properly, set the amperage your material actually needs, feed enough filler at a steady rhythm, and travel slowly enough that the puddle ties both plates together before it solidifies. Do those five things consistently and the lumpy caterpillars disappear, replaced by fillets that a gauge will pass and an etch test will confirm. That is the whole craft, learned one coupon at a time.

Frequently Asked Questions

What is a T joint in TIG welding?

A T joint is a corner joint where two members meet at approximately 90 degrees and are joined by a fillet weld along the inside corner. TIG welders build it by running a controlled fillet bead, adding filler in rhythm, so the weld fuses into both the horizontal and vertical plates.

What amperage should I use for a T joint with TIG?

A common starting point is about 1 amp per 0.001 inch (0.025 mm) of thickness — roughly 40-60 A for 16-gauge mild steel and 80-110 A for 1/8 inch steel. Adjust based on how the puddle behaves: too cold and the bead sits on top, too hot and the toes undercut.

Why is my T joint TIG weld not fusing into the corner?

Lack of fusion at the root usually comes from improper fit-up, holding the torch too far away, a wrong torch angle, travel speed too fast, or insufficient heat. Increase heat slightly, tighten the fit-up, keep the arc close and add filler at the leading edge of the puddle.

Do I push or drag a TIG fillet weld?

For most T joint fillet welds, a slight drag (trailing) travel angle of 5-15 degrees works well on steel, giving good shielding coverage and puddle control. Some welders push on aluminum for better gas coverage over the puddle; consistency matters more than the exact direction.

What filler rod should I use for a steel T joint?

For mild steel, ER70S-2 is the standard general-purpose choice; for stainless, ER308L matches common 304 stainless. Choose rod diameter by weld throat size rather than plate thickness — larger fillet welds need thicker rod to fill the joint.

How do I weld an aluminum T joint with TIG?

Clean the aluminum thoroughly with acetone and a dedicated stainless brush, use AC with enough electrode-positive balance for oxide cleaning, and expect the corner to heat much faster than steel because of aluminum’s thermal conductivity. Add filler generously — it acts as both strength and a heat sink.

How is a T joint fillet weld measured for acceptance?

Fillet welds are checked with a weld gauge measuring leg length and effective throat, and evaluated against the drawing or code requirements. AWS D1.1 visual inspection rejects undersized legs, excessive undercut and unacceptable profiles.

How do I stop undercut on T joint TIG welds?

Undercut at the toes comes from excessive heat, travel speed too fast, or not adding enough filler. Slow down, add more filler into the toes, and reduce amperage slightly — on fillet welds, more filler almost always fixes undercut faster than changing settings.

Does a T joint need full penetration?

Not usually — most T joints use fillet welds, which are surface welds with a defined throat. When the design requires structural root fusion, a partial or full penetration groove weld detail is specified instead, and the joint prep changes accordingly.

How do I prevent distortion when welding a T joint?

Use balanced tack welds, clamp the assembly to a rigid fixture, and sequence welds to distribute heat (back-step or skip welding rather than running one long continuous bead). The TIG pedal lets you modulate heat without changing machine settings, which helps keep distortion low.