Perfect TIG Welding: How to Make Clean, Consistent Welds

Your first TIG bead comes out wobbly, sputtered, and dull grey, and you start wondering whether the problem is the machine, the gas, or you. That moment — a good torch in one hand and a bent rod in the other, watching a puddle you cannot control — is where most welders quit. The good news: a ‘perfect’ TIG weld is not a mystery or a gift, it is the sum of a handful of small, repeatable decisions, and this guide breaks each one down.

What “Perfect TIG Welding” Actually Means (and Why Your First 50 Welds Will Look Bad)

Nearly every TIG beginner remembers the first bead on a scrap plate: the arc sputters, the filler rod bounces off the puddle instead of melting into it, and the bead comes out fat at one end, thin at the other, and a dull, sooty grey. Then the debate starts: machine, argon, or you?

The honest answer: a perfect TIG weld is not a machine setting and not a lucky break. It is the accumulated result of three things you control: metal preparation, machine parameters, and how your hand moves the torch and the rod. Get all three roughly right and the machine gets out of the way. Ignore one of them and no amount of knob-turning rescues the bead.

What the process is actually called

TIG is the shop-floor nickname for gas tungsten arc welding (GTAW), the formal name the American Welding Society uses in its welding terminology and standard AWS C5.5/C5.5M. The nickname comes from the older term “tungsten inert gas,” and although AWS standardized on GTAW decades ago, TIG is what everyone in the shop still says. Both terms mean the same process.

Per Haynes International’s overview of GTAW, heat comes from an arc struck between a non-consumable tungsten electrode and the workpiece, with no consumable filler electrode in the arc. That explains nearly everything about TIG: the tungsten never touches the puddle, there is no spatter, and the filler rod is fed by hand from outside the arc. See the same fundamentals in our guide to aerospace TIG welding standards and techniques.

The three pillars of a repeatable weld

1. Metal preparation. Clean, dry, correctly fitted metal: no paint, oil, mill scale, or leftover oxide.
2. Machine parameters. Amperage, gas flow, electrode size and geometry, torch setup, and foot pedal feel, matched to the joint and material.
3. Hand technique. Torch angle, arc length, and filler feed rhythm, consistent from end to end.

Beginners almost always fix the wrong pillar first: parameters are the easiest thing to change, a knob or a number. The bead is ugly, so you drop 10 amps, then fiddle with the flow meter. Nothing improves. The real culprit was a greasy plate or a 1/16 in root gap you never opened. Start with our guide to prepping aluminum properly before welding: surface condition is the most skipped and most expensive mistake in TIG.

An honest definition of quality

Set aside “looks pretty.” A good TIG bead has four objective markers:

• Consistent width and bead profile from start to finish.
• No porosity and no undercut along the toe or face.
• Minimal heat tint, meaning you controlled heat input instead of melting the metal to a puddle.
• Full fusion at the root, with no cold lap or unfilled gap underneath.

Add the fifth, and it separates hobbyists from fabricators: you can hit the same result again on the next bead, with the same parameters and the same motion. A professional TIG weld is repeatable, not a one-off.

Why TIG is worth the effort, and what it costs you

TIG earns its reputation on thin stainless and aluminum, where heat control decides whether the part survives. It is the standard choice for root passes a fillet cap will bury, for repair work on thin or mismatched sections, and for anything where the finished weld is on show. The payoff: clean, slag-free, paintable, inspectable results.

The price is speed: Haynes International names the low manual deposition rate as GTAW’s major drawback, and it is real. You feed wire by hand at a rate no machine-fed process will match. On a one-inch production fillet that is the wrong trade; on a thin stainless tank seam that must not discolor or distort, it is the right one. Our breakdown of where stick welding still wins shows the same trade-off.

The standards side of the craft is defined too: AWS A5.12/A5.12M covers tungsten electrodes, AWS A5.9 filler metals, and AWS A5.18 argon and argon-helium shielding gases. Browse all three in the AWS standards and publications library.

Where this guide goes next

Next up: arc behavior, polarity, and shielding gas; a setup routine with an amperage-by-tungsten-diameter table; prep work; torch angle, arc length, puddle control, and filler feeding; a material playbook; troubleshooting for porosity, undercut, lack of fusion, and crater cracks; safety around fume, argon, and thoriated tungsten; process comparisons; and a four-week practice plan.

Your first 50 welds will look bad. That is not a sign of a bad machine or a wasted afternoon. It is the process of replacing guesswork with repeatable habits, one bead at a time.

How Gas Tungsten Arc Welding Actually Works: Tungsten, Polarity, and Shielding Gas

TIG is short for gas tungsten arc welding, the AWS designation for GTAW. Before it is a technique, it is a physical event: an electric arc jumping between a sharpened tungsten and the metal you are welding. That arc is both the heat source and the shield.

The Machine, the Arc, and the Non-Consumable Electrode

A TIG power source is constant current, holding amperage steady while voltage floats as arc length changes, so a long arc sags slightly instead of burning hotter. The sharpened tungsten sits in a collet inside the torch, one terminal of the circuit, while your work is grounded to the other. Strike the arc and ionized gas conducts between the tungsten tip and the workpiece.

The tungsten conducts electricity and aims the arc. It is not the heat source and not part of the finished joint, hence the term non-consumable electrode. If it touches the puddle it absorbs molten metal, the tip is ruined, and it must be reground before the next weld, a failure mode Miller walks through in its TIG problem guide. Because the tungsten never melts into the bead, no spatter is thrown off the puddle the way MIG wire or stick rod does, so TIG is prized for appearance-critical work.

Polarity: DCEN, DCEP, and Why Aluminum Needs AC

DCEN (straight polarity, electrode negative) is the default for TIG. Electrons flow from the tungsten to the workpiece, heat concentrates in the work, and the tungsten stays cool enough to carry more current without balling.

DCEP (reverse polarity, electrode positive) reverses that. Electrons bombard the small, high-resistance tungsten, heating the electrode, cutting its current-carrying capacity, and adding a cathodic cleaning action that scatters oxide films, as described in the California Welding Educational Resource Foundation polarity notes.

Aluminum is why AC TIG exists. Each AC cycle has two half-cycles: the electrode-positive half breaks down the stubborn aluminum oxide skin, the electrode-negative half puts heat into the base metal and melts it. Miller’s technical guidance makes the balance explicit, and the same cleaning action shows up in our aluminum MIG welding guide. Single-polarity DC will not reliably weld aluminum without both halves. DC is fine for brazing aluminum and some cleaning passes, but the fusion pass needs AC.

Tungsten Types and Their Real Trade-Offs

Electrode classification and sizes are set out in AWS C5.5.

Type	Strengths	Watch-outs
Pure (EWP)	Cheapest; fine for AC and low current	Low current capacity; contaminates easily
Thoriated 1% / 2% (EWTh-2)	Best arc start and current capacity; very stable	Radiation and inhalation hazard from grinding dust; being phased out by some governing bodies
Ceriated (EWCe-2)	Non-radioactive; excellent arc start	Very little downside
Lanthanated (EWLa-2)	Non-radioactive; comparable performance on AC and DC	Very little downside
Zirconiated (EWG)	Mid-range price, good on AC	Slightly lower current capacity than thoriated

Haynes International notes that ceriated and lanthanated electrodes perform comparably to thoriated without the thoriated hazard, and that thoriated is being phased out. Ceriated or lanthanated is the sensible default today. Shops still running thoriated should grind with dedicated extraction. If your work is spec-driven, our aerospace TIG welding standards guide shows how electrode control is documented.

Electrode Geometry Changes Your Bead

Grind to a 30 to 60 degree included angle with a 0.040 to 0.060 in (1.0 to 1.5 mm) flat at the tip, per Haynes. Grind longitudinally, never in a circle, so the tip stays even. The taper decides everything: a blunt point spreads the arc into a wide, soft bead with shallow penetration, while a sharper cone focuses the arc, narrows the bead and drives penetration deeper. Too sharp and the tip becomes fragile and breaks off.

Shielding Gas and the Gas Lens

Argon is the standard shield because it is inert and heavier than air, so it blankets the puddle. Welding-grade argon at 99.996 percent minimum purity suits most GTAW work. Flow is critical both ways: too little leaves the pool unprotected, giving oxidation and porosity; too much turns the flow turbulent, entraining air and whipping the arc around. A gas lens, a screen of fine holes in the cup, diffuses the gas into a uniform curtain, cuts turbulence and allows a longer tungsten stick-out. Supply questions are covered in our AllGas welding supply guide.

Pulse TIG

Modern machines add pulse TIG, cycling between a high peak current and a low background current. The pool forms during the peak and cools between pulses, giving far tighter heat control on thin material and stainless than a single DC setting. The arc stays lit throughout, so clean metal and steady gas still matter. The TIG and GTAW process overview shows where pulse fits.

Machine Setup: Amperage, Gas Flow, Torch Cup, Tungsten Size, and the Foot Pedal

Most TIG machines that “run bad” trace back to a setup shortcut: gas set by feel, whatever tungsten was in the collet, amperage guessed at on a Friday. Fix the setup and the same machine lays consistent beads. This routine takes ten minutes.

The repeatable setup routine

1. Check the machine and work lead for frayed insulation, loose connections, and a dirty clamp face. A poor ground causes a wandering arc.
2. Confirm polarity and current type. GTAW runs DCEN for most work; aluminum needs AC or a proven DCEN-to-AC cleaning cycle. Set polarity before striking.
3. Set the flowmeter upright. Angled units read wrong, as do those above the regulator’s outlet side. Vertical, float centered in the scale.
4. Set pre-flow and post-flow. A second of pre-flow clears air from the cup; several seconds of post-flow protects the hot bead.
5. Calibrate the foot pedal per the manufacturer’s instructions, then confirm it starts and ends the arc cleanly.
6. Run a test bead on scrap with the same joint, filler, and settings. Change one variable, not four.

Tungsten size and amperage

An electrode too small for the current balls up, wanders, and contaminates the weld, so match it to the load. Typical DCEN bands, from the Lincoln Electric TIG reference:

Tungsten diameter	Thoriated / lanthanated (DCEN)	Pure tungsten (DCEN)
1/16 in (1.6 mm)	70–150 A	approx. 50–80 A
3/32 in (2.4 mm)	150–250 A	approx. 80–150 A
1/8 in (3.2 mm)	250–400 A	approx. 150–250 A
1/4 in (6.4 mm)	500–750 A	not recommended in DC

Push the electrode firmly into the collet so it cannot wobble, grind so 1/8 to 1/4 in of taper protrudes, and never let it touch the collet nose or the work. Pure tungsten runs cooler and needs lower amperage than thoriated, and is the practical choice for AC aluminum below roughly 150 A, where it resists breakdown on the electrode-positive half-cycle. Above that, step up to thoriated, lanthanated, or ceriated and the next diameter up. Electrode geometry and grading are covered in the GTAW electrode reference.

Set the amperage to the lowest value that still fuses the thickness you are welding. Extra amps are wasted heat: they widen the bead, drive undercut and tungsten balling, increase distortion, and burn the heat tint off stainless. No flow or unmelted edges, add heat; a rippled, wide, stained bead, take heat away.

Cup size, gas lens, and shielding gas flow

Match the cup to the amperage band: small ceramic cups (roughly 5/8 in, 3/4 in) suit low current, 1 in cups handle mid-range work, and 1 1/4 in or larger cups are for high amperage and long stick-out. Haynes International recommends the largest practical cup so gas arrives at lower velocity, plus a gas lens, whose screen diffuses the gas column, cuts turbulence, and allows about an inch of stick-out.

Argon flow scales with amperage and cup size. CK Worldwide’s guidance runs roughly 5–8 CFH at very low amperage, up to 20–35 CFH in the 250–350 A range with larger cups. Haynes International recommends 20–30 CFH (9–14 L/min) of high-purity argon for typical GTAW. Adjust for joint, position, and wind.

• Too little gas: bright, buzzy arc, oxidation and black pepper, puddle skins over.
• Too much gas: lazy, fuzzy arc, churned puddle, air pulled in from the cup rim, rough orange-peel bead.

A leak in the torch, cup, or hose upstream of the flowmeter starves the arc while the gauge still reads fine. Check O-rings and the cup seal when you change cups. The welding equipment and gas supply guide covers regulators, flow-control regulators, and cylinder handling.

AC balance, pulse, and arc start

On AC, the balance control sets how much of each cycle is electrode positive (cleaning) versus electrode negative (penetrating). Too much electrode-positive time balls the tungsten and etches the work, the biggest aluminum complaint, and Miller warns of exactly this effect. Too little and the oxide skin stays intact, the puddle refuses to wet out, and black pepper specks appear, meaning more cleaning time, not more amperage.

Pulse TIG is the modern heat-control tool: a background current keeps the arc stable while the peak current pulses on and off. On stainless and aluminum it holds a small, steady puddle at high travel speed, easier than feathering the pedal all day.

A high-frequency start unit ionizes the gas and establishes the arc without striking the tungsten on the work. As CAERT explains, the high voltage and frequency carry current along the conductor’s surface, so there is no contact and no tungsten contamination at the start. Up-slope and down-slope timers ramp current up gently and taper it down at the end, the first defense against crater cracks.

The three-quarter pedal technique

Set the amperage with the pedal about three-quarters depressed, leaving a quarter of the travel in reserve for striking the arc and trailing off at the bead’s end. Tom Bell of UC Davis recommends exactly this: set the machine, then modulate the last quarter of travel for the start and the finish. The amperage you set is then the amperage you weld at.

Once this routine is second nature, the machine stops being the mystery. The aerospace TIG welding guide shows the same setup discipline scaled to work where repeatability is not optional.

Prep Work That Decides the Weld: Cleanliness, Fit-Up, Joint Geometry, and Filler Metal

Most bad TIG welds are not bad because of amperage, gas flow, or a cheap machine. They are bad because the joint was dirty, the parts did not fit, the root gap wandered, or the filler was the wrong alloy. Parameters can fix heat problems. They cannot fix contamination.

Cleaning: The Part Everyone Skips

Clean every surface the puddle touches: joint faces, the land, root faces, and about an inch of metal either side, since heat spreads past the puddle line. Strip oil, grease, paint, rust, mill scale, and oxides. Grease and oil vaporize into the arc, turn the shielding gas column sooty, and leave porosity no parameter tweak can remove, so the bead has to be cut out, cleaned to bright metal, and redone. On carbon steel, grind or blast away mill scale, a hard refractory oxide that resists melting: a puddle sitting on it instead of fusing in is the textbook definition of lack of fusion. On aluminum, treat aluminum cleaner for welding as a careful step of its own, not a wipe with a rag.

Aluminum grows an oxide skin almost instantly. Tom Bell of UC Davis notes that 2024 and 7075 alloys are not weldable, while 5052, 3003, and 1100 are soft, bendable alloys that weld easily and forgive beginners. Oxide has to come off right up until you weld, so mark a stainless wire brush for aluminum only and never use an aluminum brush on stainless or carbon steel, which deposits iron particles that rust and contaminate the bead. On stainless, scrub heat tint with a dedicated stainless brush and keep carbon steel wheels away.

Joint Geometry: What You Are Welding Into

Most TIG work falls into three joint families: butt, lap, and T-joint. Butt joints get the root pass you fill later, lap joints a single fillet, and T-joints need both sides accessible.

Thickness drives the decision. Per CAERT training materials, metal under 3/16 inch may need no edge preparation, TIG is seldom used on metals thicker than 1/4 inch except aluminum and magnesium, and thicker edges should be ground or machined for full bead penetration. On heavy plate the geometry must let the arc reach the bottom of the groove: a square butt on 1/2 inch steel is a fight you will lose.

Thick plate is also where TIG stops being the economic answer. Filling a V-groove on 1/2 inch plate and grinding it flush can cost more than a MIG pass that goes in four times faster. TIG wins on thin material, stainless, aluminum, repair work, and appearance-critical parts.

Root gap needs to be consistent more than perfect. A uniform 1/16 inch land gives a repeatable puddle and root bead; a drifting joint produces root fusion problems no matter what the machine is doing. Measure it, use a spacer wedge, and hold the gap open as the parts heat.

Fit-Up and Tacking

Fit-up is where most root problems begin. Parts not held in alignment shift as they heat, and you get a root bead that fuses on one side and misses on the other. Tack before you weld: CAERT recommends tacking lap and T-joint flat assemblies every 3 inches, which keeps plate from bowing or rotating as the weld shrinks. On a butt joint, a single tack at each end is often enough if the clamps are good. Tacks are temporary, so use small, easy-to-cut filler and do not worry about how they look, but place them where the torch reaches the joint and where they will not interfere with your bead, since every tack gets ground later. Skip tacking on a T joint and burn through one leg, and that burnthrough is a fit-up problem before it is an amperage problem.

Filler Metal Selection

Filler should match the base metal, the service conditions, and the position. For stainless, ER308L and ER308LMo are the common choices, with the Mo grade for higher corrosion service. For aluminum, ER4043 works well for many 3000 and 6000 series applications, while ER5356 and ER5183 are standard picks for 5000-series alloys. Bell notes that 5000-series filler (5356, 5183, 5556) suits parts that will be anodized because it takes the anodizing layer cleanly, while 4000-series welds turn black when anodized. For carbon steel, ER70S-6 is the standard mild steel rod. GTAW filler metals are governed by AWS A5.9/A5.9M, and shielding gases fall under AWS A5.18.

Diameter controls bead width: CAERT states a TIG bead should be two to three times the filler rod diameter in width, so a 3/32 inch rod fills roughly a 1/4 to 5/16 inch wide bead. Small rods feed precisely but demand a steadier hand, so start with 3/32 or 1/8 inch rod and step down for narrow beads.

Match the alloy to the service, not just the part name. When the base metal is unknown or the job is code work, pull the mill certificate so heat and alloy numbers stay traceable to the weld record. Check the aluminum before you fill: if it turns out to be 2024 or 7075, stop.

Pre-Weld Prep Checklist

• Degrease joint faces and surrounding metal, then let it dry.
• Strip oxides and mill scale to bright metal right before welding.
• Dedicated tools: stainless brush for stainless, marked brush for aluminum.
• Consistent land and gap, held by spacers or tacks; lap and T joints tacked every 3 inches.
• Filler matched to base metal and service, rod diameter matched to bead width.

I’ll rewrite this section to fit the 640–960 word range, keeping the heading, facts, and all links intact while tightening filler and repetition.

Core Technique: Torch Angle, Arc Length, Puddle Control, and Feeding the Filler

Everything up to this point has been preparation. This is where the weld actually happens, and where most beginners lose the battle against their own hands. A good TIG weld is not one clean movement. It is three things happening at once without any one of them taking over: you are managing torch angle, arc length, and filler feed simultaneously, making tiny corrections to all three while the puddle walks along the joint.

The GoWelding TIG technique guide describes the same three-part approach most experienced operators use: a forehand torch position tilted roughly 10 to 15 degrees off vertical, a short arc of about 1/8 inch, and filler dipped into the leading edge of the puddle rather than melted in the arc itself.

Arc Length: Short Means Everything

Hold the tungsten about 1/8 inch (roughly 3 mm) off the workpiece and no more. That is close enough for a tight arc column, concentrated heat, and clean shielding gas flow around the molten metal. Lengthen the gap and everything degrades at once. The arc wanders and hunts because it loses stability. Heat spreads wider, so penetration drops even though the amperage setting has not changed. The shielding gas gets aspirated, pulling air in from outside and diluting the argon blanket, which is exactly how you get a grey, oxidised bead.

A long arc is also hard to read. At 1/8 inch you can clearly see the edge of the puddle and the two bright spots where the arc contacts the metal. Stretch to 3/16 or 1/4 inch and you are squinting at a fuzzy blue haze, guessing where the pool actually is.

Torch Angle: Push for Flat Work, Pull for Tight Corners

For most flat and horizontal work, run the forehand (push) method: the torch leans forward in the direction of travel, tilted about 10 to 15 degrees off vertical. Pushing gives you the best view of the puddle, a naturally slightly convex bead profile, and good access to the joint as you advance.

Use backhand (pull) when you need to reach inside a corner, work a tight radius, or fill a joint where the torch has to stay on the far side of the bead. Pulling puts the heat slightly behind the puddle and produces a flatter, wider bead, useful for building reinforcement but harder to read, so it takes practice.

One nuance from UC Davis welding professor Tom Bell: on aluminium, adding filler makes the pool rise. If the torch is not positioned properly, the swelling pool can climb toward the tungsten and contaminate it. On aluminium, keep the torch slightly further from the work and slightly more tilted than you would on steel, and watch the pool itself, not the puddle line.

Puddle Control: Watch the C, Not the Blob

A healthy TIG puddle has a C-shaped profile. You should see two distinct bright spots where the arc meets the metal, with the darker centre of the pool between them. Those spots are the actual fusion zone, the points where base metal is melting. If you see only one, you are either not centred over the joint or the puddle is too small to read.

Narrow puddles come from low amperage or fast travel and produce a tall, slender bead with poor sidewall fusion. Wide puddles come from high amperage or slow travel and produce a fat bead that spreads heat and burns through thin material. Your job is to keep the puddle the same size from the first inch to the last, which means adjusting amperage as the joint changes and being honest about travel speed rather than fixing everything with the pedal.

As Haynes International recommends, use only enough current to melt the base metal and allow proper fusion of the filler, and prefer stringer or narrow weave beads over wide weaving.

Feeding the Filler: Lead Edge, Not Centre

This is where the biggest gains hide, and it is more subtle than most instruction videos admit.

• Dip into the leading edge of the puddle. Just in front of the pool, in the boundary zone where the arc is still forming a bead.
• Do not feed into the centre. Filler pushed into the middle of the pool physically shoves the arc away from you. Your instinct is to pull back and re-centre, which lengthens the arc and destabilises it. An unstable arc bounces, and a bouncing arc eats your tungsten.
• Do not feed right at the leading edge either. Too close and the rod hangs up, drags across the hot bead on stainless steel, and gets blown off and oxidised on aluminium, where it ends up floating in the pool as dross.
• The sweet spot is between the leading edge and the centre. Close enough to melt cleanly, far enough back that the rod does not touch the arc.
• Hold the rod at roughly 90 degrees to the torch angle and feed it in line with travel, not across it.
• Never melt the filler in the arc. The rod can swing back, touch the tungsten, and contaminate it. Tap the rod into the edge of the pool and let the puddle do the melting.
• Keep the hot end of the rod inside the shielding gas. If the glowing end wanders out of the gas column, it oxidises and your weld fills with inclusions.

Start with larger filler, 3/32 or 1/8 inch, as Tom Bell advises. Thin rod is genuinely harder to feed and control. Bell also recommends thin gloves on the feeding hand so you can feel the rod touching the pool rather than only seeing it.

Walking the Cup, Switching Hands, and Closing the Bead

For pipe and tube, the most repeatable technique is walking the cup. Rest the ceramic cup on the workpiece and rock it in a steady cadence as you move. That constant, visible reference fixes arc length for you and produces the classic stacked-dime beads you see in qualified 6G pipe work. GoWelding describes it as one of the most consistent ways to get a repeatable bead.

On long joints, alternate which hand feeds the filler and which holds the torch. It halves fatigue and keeps both hands sharp. Resting your non-torch hand on the workpiece gives you a stable platform and kills the wobble that produces erratic bead width.

Finish every bead with a proper crater fill. Rather than snapping the arc off and leaving a hole that cracks open, ramp the current down with the foot pedal or up-slope/down-slope control while feeding a small amount of filler into the shrinking pool, then step out over the crater. To build up a cap, the back-step method, welding over the start of the previous bead, gives a cleaner result than trying to fill a deep crater.

Master arc length, angle, puddle size, and filler placement and you have the whole craft. The operators who look effortless at TIG are simply making these four corrections so small that nobody can see them. For more on feeding and filler selection, see our all-position welding rod guide, and for the precision demands of critical work, our look at aerospace TIG welding standards and technique.

The Material Playbook: Stainless Steel, Aluminum, and Carbon Steel

The same torch, foot pedal, and rod give three very different beads on stainless steel, aluminum, or carbon steel. Each metal conducts heat differently, and that decides whether a weld succeeds or turns into a repair job. Perfect TIG welding is a material-specific playbook that starts with a number.

Heat Input: The Number Behind Every Good Weld

Heat input is the energy delivered per millimetre of travel, and it governs penetration, distortion, burn-through, heat tint, and grain growth

Heat input is the energy delivered per millimetre of travel, and it governs penetration, distortion, burn-through, heat tint, and grain growth at once:

Heat input (kJ/mm) = (arc voltage × amperage × 60) ÷ (travel speed in mm/min × 1000)

Apply an arc efficiency factor of roughly 0.6 for GTAW, since not all arc energy lands in the workpiece; reference values in AWS C5.5 assume a similar range. At 11 volts:

Amperage	Travel speed	Gross heat input	Typical job
50 A	150 mm/min	0.22 kJ/mm	Thin stainless or aluminum
90 A	100 mm/min	0.59 kJ/mm	18 gauge carbon steel, 1/8 in stainless
120 A	80 mm/min	0.99 kJ/mm	3/16 in carbon steel butt joint
150 A	60 mm/min	1.65 kJ/mm	Heavy plate, fillet repairs

Low amperage with fast travel defends against distortion, burn-through, and heat tint, and travel speed matters as much as amperage. Haynes International lists manual GTAW speeds of 4 to 6 in/min (100 to 150 mm/min) for nickel and cobalt alloys, slower than for carbon and stainless steel because those alloys demand more heat.

Stainless Steel: Heat Is the Enemy

Stainless has no oxide to break down, so it welds on DCEN with a simple, stable puddle. The challenge is heat. The surface walks through a color sequence as temperature climbs: pale straw, bronze, deep blue, then dull grey. Those tints act as a live thermometer and a warning. According to Miller, overheating stainless changes its color, reduces corrosion resistance, and weakens mechanical properties, and that discoloration cannot be reversed. Treat a bright blue or grey band as a defect before welding another inch.

Watch the backside too. Oxidation on the reverse of a stainless weld is called sugaring, and Miller notes it weakens the joint and compromises corrosion resistance. Prevent it with argon back purging or lower amperage so the root does not melt through, and keep TIG beads narrow rather than wide and flat.

Filler should match the base metal and service conditions: ER308L for standard 300-series work, ER308LMo where molybdenum-bearing base or higher corrosion resistance is needed. Some alloys need their own procedures, such as alloy 20 welding procedures, so match the rod to the alloy. The joint must be clean, dry, and free of sulfur-bearing contamination: oil, paint, grease, and cutting oils cause cracking and porosity that no parameter change can rescue. Aerospace work, where heat input is tightly controlled, shows how precise this gets: see our aerospace TIG welding guide.

Aluminum: Different Machine, Different Rules

Aluminum welds on AC only. Miller is explicit on why: DCEN fails to break down the aluminum oxide layer, so filler mixes with partially melted oxide and produces a contaminated bead. The electrode-positive half-cycle cleans, the electrode-negative half-cycle melts.

Cleanliness is everything, and it takes dedicated tools: a stainless brush on aluminum leaves iron particles on the surface, and a carbon steel brush on stainless causes rust. See our guide to prepping aluminum for welding properly.

There is a second contamination trap: the filler itself. Miller’s guidance is to wait until the aluminum puddle shows a shiny dot before adding filler, since adding it before the oxide breaks contaminates the deposit.

Aluminum also needs far more heat than steel for the same thickness. Tom Bell of UC Davis puts a 1/8 in aluminum open corner joint at roughly 130 amps on AC, and his point is worth memorizing: a “hot” machine with a fast travel speed means less heat input, less distortion, and less burn-through. Aluminum pulls heat away from the puddle faster than steel, so you need more power at the arc but less time on the joint.

For filler, ER4043 is standard for many 3000 and 6000 series applications and forgives slightly contaminated metal. ER5356 or ER5183 fit 5000-series alloys and anodized work, where higher strength and anodizing response matter. Watch for crater cracking: fill the crater with filler as current tapers down, use downslope, and never break the arc abruptly. For repair work where heat control is everything, see our alloy wheel welding guide.

Carbon Steel: The Forgiving One, If You Prep It

Carbon steel forgives the most, but only on clean metal. Mill scale is a hard, dark oxide that melts far above the steel’s melting point, so welding through it creates ragged, inclusion-filled beads. Remove it with a grinder or wire wheel before striking an arc.

ER70S-6 is the workhorse filler, and joint geometry changes everything. A fillet joint bridges a solid corner, so the heat is more forgiving. A butt joint puts the arc on the root gap, where you need enough heat to fuse the root without collapsing the plate. Our all-position welding rod guide shows how rod behavior changes out of flat position.

Controlling Distortion

Distortion is heat management, so the same heat input logic applies, with structure on top:

• Tack sequence: work from the ends inward, outside toward the middle, so shrinkage balances out.
• Strongbacks: a stiff bar clamped along the joint absorbs the weld’s pull and holds alignment.
• Balanced sequences: alternate joints, weld both sides of a T-joint, or use back-step and split welds.
• Preheat: on heavy or high-carbon steel, slow cooling, shrink the hardened zone, and let the structure relax.

Get the heat input number right and most distortion problems never start. Ignore it and the best technique in the world will still pull your plate out of flat.

The Most Common TIG Mistakes and the Fix for Each

TIG is unusually honest metalwork. A gas-shielded arc does not hide what you did wrong: the bead records gas coverage, heat input, travel speed, and joint prep all at once.

Go symptom first. Name the defect, work backwards to the parameter behind it, change one variable at a time, run another coupon. Nearly every early TIG defect falls into gas, heat, motion, or cleanliness.

Symptom	Likely cause	First move
Porosity (pits, specks)	Contamination, moisture, poor shielding, wind, long arc	Clean and dry, check gas flow, shorten the arc
Undercut (groove at the toe)	Current too high, travel too fast, arc too long	Lower current, pause at the toe, shorten the arc
Lack of fusion	Current too low, travel too fast, wrong angle, poor fit-up	More current, slower travel, fix angle and fit-up
Crater cracks	Arc broken off, crater left unfilled	Fill it with a back-step, downslope current
Black pepper on aluminum	Too little AC cleaning, dirty metal	Clean, shift AC balance to cleaning, wait for a shiny puddle
Sugaring on stainless	Backside oxidation	Back-purge with argon, lower heat input
Bead too wide or narrow	Current, travel speed, arc length, electrode taper	Balance all four, one change at a time
Arc or tungsten wanders	Long arc, dirty or undersized tungsten, loose ground clamp, too much electrode-positive time	Shorten the arc, resize the electrode, tighten the clamp, adjust AC balance

Porosity: fix the metal, not the machine

The GoWelding troubleshooting table traces porosity to contamination, moisture, inadequate shielding gas, wind disturbing the shield, and an arc too long. Take that list literally.

Oil, paint, rust, mill scale, cutting fluid, or a fingerprint of grease outgas in the arc and spit holes into the puddle. No parameter setting fixes that. Cleaning aluminum before welding makes the same case for a tougher oxide skin, chemically hostile to the weld.

Argon is heavier than air, so a wet joint makes a steam bubble that escapes through the melt. Wind strips the shielding envelope and pulls air in. A long arc means the shield travels further. Check gas, weather, then the gap between tungsten and work.

Undercut and lack of fusion: a heat and motion problem

They look like opposites but behave like a pair. GoWelding lists undercut causes as current too high, travel speed too fast, and arc length too long; the fixes are turning current down, pausing at the toe, and shortening the arc. For lack of fusion it lists low current, fast travel, and a wrong electrode angle; the fixes are more current, slower travel, and a corrected angle.

Together they give one rule: heat input and travel speed are the same dial. Too much heat for your speed gouges the toe; too little gives a puddle that slides over the surface without melting it. Miller notes that improper fit-up, a torch held too far from the joint, and incorrect filler rod feeding also cause root lack of fusion, and that modern inverters with adjustable frequency and pulsing produce a narrower, more focused arc cone.

When the tungsten becomes the problem

A dark, pitted, blobby tungsten is contaminated, and the usual cause is that the electrode touched the puddle or the filler. Keep it 1/16 to 1/8 inch off the work, size it so it is not at its limit, confirm DCEN on steel and stainless, and grind it back to shape. Touch contaminated metal to a tungsten and you can ruin a pass.

Wobbling is different. Tom Bell explains in his TIG tips that a tungsten which starts to wiggle is near its current capacity, and moving the square-wave balance toward penetration reduces heat on the tip. Wobble means an overloaded electrode, not a bad machine, so step up to a larger diameter. When regrinding thoriated tungsten, handle the dust as the radioactive material it is, per AWS Z49.

Black pepper, sugar, and bead shape

On aluminum, a dark, grainy surface like coarse pepper is oxide that never got broken down or burned into the puddle. Clean the metal, shift AC balance toward cleaning, and wait for a shiny puddle before adding filler. On stainless, oxidation on the back of the joint is called sugaring, and the answer is a back-purge with argon plus lower heat input. Miller recommends pulsing on stainless to reduce heat input and improve puddle control, and lowering amperage or raising travel speed to prevent discoloration.

Bead width comes down to four variables in tension: current, travel speed, arc length, and electrode taper. Too wide means too much current, too little travel speed, or too much stick-out with a wide taper; too narrow means the opposite. Flux-assisted aluminum welding is worth knowing for dirty or badly oxidized aluminum, but it is a separate process from TIG and does not rescue a contaminated TIG setup.

Safety: Welding Fume, Argon, Thoriated Tungsten, and the PPE You Actually Need

TIG looks clean: no slag, no spatter, a bead like brushed metal. It still pours metal fume into the air and runs an arc that can damage your eyes in a fraction of a second. Technique is the rest of this guide’s job. Safety discipline is what keeps you practising it.

What is actually in the fume

The mix depends on the metal. Stainless is the worst offender because its chromium oxidises into hexavalent chromium, a confirmed carcinogen. Mild steel fume carries manganese, nickel alloys add nickel compounds, and galvanised coatings produce the oxide behind metal fume fever. NIOSH’s Pocket Guide to Chemical Hazards lists welding fumes as a potential occupational carcinogen, with metal fume fever, dyspnea, cough and interstitial pneumonitis among the effects.

Acutely, OSHA fact sheet FS-3647 lists eye, nose and throat irritation, dizziness and nausea, and warns that prolonged exposure may cause lung damage plus lung, larynx and urinary tract cancers. Long-term manganese fume exposure can produce Parkinson-like symptoms.

One number changed practice. In 2018 ACGIH cut the hexavalent chromium threshold limit value from 50 µg/m³ to 0.2 µg/m³, a 250-fold drop, as AWS Z49 records, so exposure at or above the TLV is likely wherever ventilation is missing. OSHA’s own limit is 5 µg/m³ as an 8-hour time-weighted average, under 29 CFR 1910.1026 and 1926.1126. That gap is why “the shop feels fine” is not a control strategy.

Ventilation that actually works

• Keep local exhaust ventilation close to the plume, not across the shop. Fume hoods and extractors belong near the source.
• Stay upwind outdoors so the breeze carries fume away from your face.
• Strip the work before striking an arc: paint, solvent residue, rust and galvanised coating all create toxic exposure.
• Use a respirator when engineering controls fall short, matched to the hazard and fit tested. A dust mask will not stop chromium fume.

Argon and the asphyxiation trap

Argon is not toxic, which is why it is dangerous. It displaces oxygen and can cause suffocation without warning, especially in confined spaces, as OSHA and the Health and Safety Executive both note. Argon’s relative density is 1.38 against air (helium is 0.14), so it settles into floor pits, tank bottoms and pipe interiors. The HSE warns shielding gas can build up fast inside a fabrication during internal welding.

So purge before you weld, cut a vent hole at the far end of the joint, and never enter an unventilated container without testing oxygen concentration with a calibrated meter. Never sweeten the atmosphere with oxygen. That is prohibited, and it turns an oxygen-deficient space into an ignition hazard.

Thoriated tungsten and invisible dust

The principal risk with 2% thoriated tungsten is not the arc, it is inhaling radioactive dust while grinding the electrode, as Haynes International explains. Controls that matter:

• Grind on a dedicated grinder with local exhaust ventilation capturing dust at the wheel.
• Use wet methods so dust is captured rather than airborne.
• Wear a respirator suited to the dust and the metal being ground.
• Collect dust in sealed containers and dispose of it under local radioactive waste rules.

Cerated and lanthanated electrodes perform comparably without radioactive thorium, making them the sensible default. Choose thoriated only when you need it and have the controls in place.

PPE you actually need

Helmet shade is where new welders get burned. AWS F2.2 guidance for GTAW:

Arc current	Comfortable shade	Minimum shade
Below 50 A	8	10
50 to 150 A	8	12
150 to 500 A	10	14

Wear the shade you can actually see the puddle through, not the darkest you can get away with. Side-shield glasses cover the arc your helmet leaves exposed. TIG gloves are thinner on the back for heat sensing, with a flame-resistant leather palm and cuff for spatter. Wear flame-resistant clothing, add a leather apron for stainless work, and use safety glasses with side shields for grinding and chipping.

Hot work in an enclosed area needs a fire watch, fire blankets over combustibles and a clear escape path. Keep cylinders upright and secured, away from heat, with the valve capped when idle and the correct regulator and flowmeter fitted. With a water-cooled torch, check grounding, inspect coolant lines for leaks and keep connections dry. Electrical safety falls under 29 CFR 1910 Subpart Q, and eye protection must meet ANSI/ISEA Z87.1.

Ventilation near the plume, gas under control, tungsten dust contained, correct shade before the arc starts. Get those four right and you can spend the rest of your time on the weld itself.

TIG vs MIG vs Stick vs Laser — and How to Prove the Weld Is Right

A beautiful TIG bead is not automatically the right answer. If a 1/2 in structural fillet has to ship tomorrow, a 1/16 in filler rod and a 70 amp foot pedal are the wrong tools. Choosing the right process is half of welding well. Proving the weld met the standard someone is paying for is the other half.

Where TIG genuinely wins, and where it does not

Haynes International characterizes GTAW as a very versatile all-position process used in production and repair, with precise control of welding heat, routinely applied to thin base metal and to root passes of thicker section welds. TIG burns no consumable electrode, so there is no spatter transfer, the puddle stays small, and the heat-affected zone stays narrow. The case for TIG is clear: thin material, stainless, aluminum, root passes, repair work, and anything where appearance counts.

MIG wins on throughput with heavier steel and lower labor cost per foot. SMAW still wins outdoors in the field, on rusted or dirty material, and where power is rough. Laser welding wins on repeatability and very narrow, deep welds, but demands serious capital and setup. On aluminum, TIG is not the only path: aluminum MIG is faster and cheaper for many structural jobs, while TIG stays the choice for thin, cosmetic, and repair work.

Process	Heat control	Deposition	Best at	Weakest at	Capital cost
TIG (GTAW)	Excellent, current and voltage set independently	Low	Thin material, stainless, aluminum, root passes, repair, cosmetics	Thick production fillets, high volume, tight schedules	Low to mid
MIG (GMAW)	Good, coupled to wire feed and voltage	High	Carbon and stainless production, structural fillets	Very thin material, critical aluminum and stainless cosmetics	Low to mid
Stick (SMAW)	Good, arc voltage tied to electrode length	Moderate	Field work, outdoor, dirty or rusted steel, overhead	Thin material, cosmetic finishes, aluminum	Lowest
Laser	Excellent, programmable and highly repeatable	High, narrow HAZ	Automation, repeatability, narrow deep welds, volume production	Thick sections on some setups, small shops	Highest

TIG’s consumable cost is near zero because the electrode is non-consumable, so its real cost is labor. MIG pays for wire and gas but moves metal fast. Stick has cheap consumables and slow travel. Laser pays in equipment, maintenance, and often positioners, which is why aerospace TIG welding and laser work live in volume shops. Spatter and fume follow suit: TIG produces the least of both, MIG and stick produce more, and laser still needs fume extraction since metal is vaporized anyway.

Proving the weld is right

Visual inspection comes first and is the cheapest quality control available. Under AWS and ASME code requirements, inspectors look for undercut, overlap, porosity, excessive reinforcement, insufficient throat or leg, and incorrect bead profile. Learn to read them before claiming a weld is good; grinding the bead crown flat fixes none. They are design and process problems, not cosmetic ones.

For surface-breaking defects, dye penetrant testing per AWS C5.5 guidance and state guidance is the standard next step. Testing to ASTM E1417 and AWS Z244.1 finds cracks too fine for the eye. It only works on clean, uncoated, non-porous surfaces, so degrease the part and remove heat tint where the procedure allows.

Then comes the paperwork. AWS publishes the framework: welding procedure specifications, performance qualification records, and welder performance qualification requirements. AWS D1.1 governs structural steel welding, and ASME BPVC Section IX governs qualification for ASME Code work. A WPS says how the weld is made. A PQR proves the procedure was tested. A WPQ proves the welder can make it. No controlled shop accepts a good-looking bead without that trail.

That trail means more than an amperage number. Miller and AWS materials both stress that heat input and travel speed records matter as much as the amperage setting. A complete TIG record includes filler lot, gas type and flow rate, tungsten type and diameter, amperage, travel speed, and preheat and interpass temperature.

Not every weld needs a certification package. Hobby work, non-code repair, and appearance-only fabrication do not. Code work, customer work, pressure work, and regulated shops do. Find where that line sits before you assume a clean bead ends the job. For most fabricators: TIG for the details you want to be proud of, and the right documentation to prove the rest.

A Practice Plan for Consistently Better TIG Welds (and What “Perfect” Should Mean to You)

Knowing what a good TIG weld looks like and making it every time are different skills. Only the second comes from repetition on a schedule. Here is a four-week plan you can run on coupons at a bench, 30 to 45 minutes a day. The goal is a weld you can reproduce the hundredth time, not a trophy bead on the first try.

Week	Focus	Key drill	Pass mark
1	Setup and puddle control, no filler	Pedal drills, ten-inch flat beads	Puddle size changes less than 20% along the bead
2	Filler feeding rhythm	Five-second dry feed, bead-on-plate coupons	Rod touches the leading edge, never the pool center
3	Position work	Flat, horizontal, vertical up, vertical down	One current setting works in all four
4	Joints and repeatability	Butt, lap, T-joint, tube, then repeat one bead	Two beads match in width, penetration and color

Week 1: Setup and Puddle Control, Without Filler

Week 1 is deliberately boring, and that is the point. Run your setup routine every time: machine check, polarity, electrode seated in the collet, cup and gas lens tight, flowmeter upright, pre-flow and post-flow timers set, pedal calibrated. Then run beads with no filler, since filler distracts you until you can hold a puddle.

Drill the pedal, where most heat control lives:

• Clean start. Strike the arc, hold a steady puddle, then release the pedal so the arc breaks and the crater fills without spitting. Twenty reps.
• Mid-bead restart. Stop halfway, restart on the same bead and finish it, learning where the puddle re-forms so no cold crater is left.
• Steady puddle. Run ten inches without letting the puddle grow or shrink.

Week 2: Filler Rhythm and Bead-on-Plate Coupons

PacificarcTig’s five-second dry feed drill is a smart warm-up: with the torch running a dry pass, feed the rod out from a stationary hand for five seconds, rest, repeat. It builds filler coordination before you commit to a live weld.

Then lay real beads on coupons matching the alloy and thickness of your production work, which Tom Bell of UC Davis calls the fastest route to muscle memory in his TIG tips handout. Cut finished coupons in half to check penetration, and pick your filler from the all-position welding rod guide before striking an arc.

Week 3: Position Work

Run the same joint flat, horizontal, vertical up and vertical down, keeping current and arc length identical. If your machine pulses, remember that pulsing alternates high and low amperage automatically, controlling heat input on thin material and laying a uniform ripple. Miller makes the same case: pulsing reduces heat input and improves puddle control, exactly what thin-gauge and stainless coupons need.

Week 4: Joint Configurations and the Repeatability Test

Butt, lap, T-joint and tube, with coupons sorted so every session starts with the hardest joint. Then run the repeatability test: pick one bead, note every parameter, and make it again from cold without changing a thing. If the two beads match in width, penetration and color, you have a process.

Bell also recommends experimenting with torch grips and tungsten profiles so you learn firsthand how each changes bead shape, width and penetration. Three grips and two tapers teach more than an hour of reading.

How to Measure Progress

• Consistency, not beauty: identical width, penetration and color.
• Defects per foot of weld: porosity, undercut, rework.
• Time per pass: steady and slightly slow beats fast and erratic.
• Repeatability: same parameters, same result on a different day.

Write Down What Works

Log amperage, tungsten size and taper, cup and gas lens, gas flow, filler alloy and diameter, and travel speed for every weld that behaves. Bell’s advice is blunt: write down what works and what does not.

The Ten-Item Pre-Weld Checklist

1. Joint cleaned, degreased
2. Fit-up tight, root gap even
3. Filler matched to base alloy
4. Tungsten ground, right size and taper
5. Electrode centered in collet, cup tight
6. Gas flow set, flowmeter upright
7. Polarity and current type confirmed
8. Ground clamp on work, machine grounded
9. Amperage set for material and thickness
10. PPE on, ventilation running

What “Perfect” Should Mean to You

A perfect TIG weld is preparation, the right parameters, and technique that holds when you get tired. It is a skill you build, not a machine you buy. The aerospace TIG welding guide shows these fundamentals in regulated work, the aluminium welding course breakdown covers what a structured program looks like, and the AWS guides and handbooks remain the standard source for qualification requirements.

• Why does my bead wobble? Hand tension and too long an arc, not the machine. Shorten the arc and rest a hand on the work.
• Why does my filler keep sticking? You are hitting the leading edge, not the sweet spot between leading edge and pool center.
• When is TIG the wrong process? Heavy production steel, where MIG and flux core processes win on throughput.

Frequently Asked Questions

What amperage should I use for TIG welding?

Use the lowest amperage that gives full fusion in your joint. A common starting point is roughly 1 amp per .001 inch of material thickness for stainless and carbon steel on DCEN, with aluminum on AC typically needing about 1.5 times that figure, but joint geometry, filler size and travel speed all change the answer. Verify on a scrap coupon of the same alloy and thickness rather than trusting a chart alone.

Why is my TIG weld full of porosity?

Porosity almost always comes from gas, contamination or moisture — not from amperage. Clean and dry the metal and filler, check that the gas cup is not cracked or too small for the current, shut off drafts and fans that disturb the shield, and shorten your arc length. Remove the defective metal and reweld only after the cause is fixed, or the porosity will simply repeat.

Should I use DC or AC for TIG welding?

Use DCEN (straight polarity) for stainless steel, carbon steel, nickel alloys and copper. Use AC for aluminum and magnesium, because the electrode-positive half-cycle is what breaks down the aluminum oxide skin — DCEN will not clean it and the filler will pick up oxide and contaminate the bead. On AC machines, adjust the balance setting to control cleaning versus penetration.

What size tungsten electrode do I need?

Choose the smallest electrode that can carry your amperage without overheating — an oversized tungsten gives a wandering, unstable arc. As a guide, 1/16 in (1.6 mm) handles roughly 70–150 A on DCEN with thoriated or lanthanated, 3/32 in (2.4 mm) about 150–250 A, and 1/8 in (3.2 mm) about 250–400 A, with pure tungsten rated lower. If the tungsten wobbles or balls up, it is undersized for the current.

How much shielding gas flow should I set?

For most GTAW work, 10–25 CFH (about 5–12 L/min) of argon is the working range, rising with amperage and cup size, and Haynes International recommends 20–30 CFH for many applications using the largest practical cup. Too little flow leaves the pool unprotected and causes oxidation; too much creates turbulence that draws air into the shield and can make the weld dirty.

Can I TIG weld without adding filler rod?

Yes — autogenous TIG is used on very thin material and for close-fitting joints where the base metal flows together under the arc. It is also common as a cleanup pass. On thicker material you need filler to make up the joint volume, and remember that autogenous welding needs a very short arc because there is no filler to keep the pool stable.

Why does my aluminum TIG weld look dirty and contaminated?

Almost always oxide. The aluminum skin must be removed before filler enters the pool, which on AC means waiting until the puddle shows a shiny dot, and on any metal means cleaning with a dedicated stainless brush and a degreaser before you strike an arc. If the AC balance is set too far to the penetration side, the electrode-positive cleaning time is too short and the oxide stays in the puddle.

Is TIG harder to learn than MIG or stick welding?

It is a different kind of difficulty rather than a simple ranking. Stick welding is forgiving because the electrode does much of the work, and MIG is fast because the wire feeds automatically; TIG asks you to control heat, puddle size and filler by hand simultaneously. That makes the early learning curve steeper, but TIG gives finer control and produces the cleanest results on thin stainless and aluminum.

Do I need certification to TIG weld?

For hobby and general fabrication work, no formal certification is required. For code work — pressure vessels, boilers, pipelines and structural steel — you will need a qualified procedure and a welder performance qualification under ASME BPVC Section IX or AWS D1.1, and documented parameters such as filler lot, gas and amperage. Ask the customer or code authority which qualification they require before you start.