How Welding Works: The Science Behind the Arc and Weld
Table of Contents
A fabricator taps a stick electrode to a 6 mm steel plate, pulls the electrode back a few millimetres, and a brilliant white arc floods the bay. Ten seconds later there is a bead joining two pieces of metal that did not exist as one piece before. The interesting question is not whether that bead will hold — it is what just happened in the gap between those two pieces, and why a process that melts metal with a lightning bolt can produce a joint stronger than the plate feeding into it.
The Ten Seconds That Fuse Two Pieces of Steel
A fabricator taps a stick electrode against the edge of a 6 mm steel plate, pulls back a few millimetres, and a brilliant white arc floods the bay. Ten seconds later the light goes out, leaving a bead of rippled, blue-grey metal joining two pieces that were, moments earlier, just steel touching but separate. The shop owner drops the joint in a vice, hangs a chain hoist off it, and the load stays put. Nobody is surprised, because this is a Tuesday.
So what happened in that gap? The short answer, and the foundation of this article: welding is a controlled melting and re-solidification of metal that creates a metallurgical bond. The American Welding Society defines welding as a permanent joint formed between pieces of metal by heat, pressure, or both. In a fusion weld, molten base metal and molten filler coalesce, and when the pool cools it becomes one continuous piece of metal with grain structure running across the joint.
Ten Seconds, Four Distinct Events
Slow that ten seconds down and four things happen in order, each depending on the last.
- The strike. The electrode touches the workpiece, and a short circuit heats a tiny contact point enough to vaporize a sliver of metal. Vapour and air ionise, creating a conductive path of plasma. Pull back and the path stretches into the arc column.
- The melt. The arc dumps heat into the plate and the electrode tip. The plate surface melts into a pool. The tip melts too, and in stick welding that molten tip is the filler metal, delivered one drop at a time.
- The travel. As the arc moves along the joint it drags the pool with it: the leading edge melts ahead, the trailing edge freezes behind. Travel speed decides pool size and how much base metal melts. That is penetration.
- The freeze. Pass the last of the puddle and the heat leaves. Within seconds the pool stops glowing and becomes a solid bead. That is the instant two pieces become one.
Stand still and you flood the joint with heat, melting a wide, often ugly bead with deep penetration and a real risk of burning through thin material. Move too fast and the pool shrinks, the bead comes out narrow and cold-looking.
Quick Note: How hot is it really? The arc column in welding reaches roughly 10,000 to 20,000 °C, while the visible surface of the Sun sits near 6,000 K. Lincoln Electric’s arc welding fundamentals put the electrode tip at about 6,500 °F (roughly 3,600 °C). That gap is where the physics of welding gets interesting.
The Three-Part Mental Model
Every welding process, from a farm gate hinge to a robotic cell, does three things:
- Delivers a heat source strong enough to melt the joint. In arc processes that is the plasma column. In friction welding a spinning tool turns friction into heat. In resistance spot welding, current runs through a stack of sheets.
- Supplies a filler and a shield. A consumable electrode or wire melts into the pool to replace metal lost to spatter and oxidation, while a flux coating or shielding gas keeps air away from hot metal. Without that protection the bead becomes porous and brittle.
- Creates a metallurgical bond, not a mechanical one. A bolt holds two pieces together by friction and clamping force. A weld holds them because the grain structure of filler and base metal has grown together into something continuous. There is no interface to slip, the weld metal can be stronger than the plate it came from, and a welded joint can be pressure-tight. You cannot simply unclamp one later.
Why Bead Shape, Sound and Colour Matter
With that model in place, the shop floor starts talking to you. A weld that spits loudly and throws sparks differs from one that hums quietly. Ripple width and height tell you about heat input and travel speed, and a bead that glows orange for a long time before going dark was poured with a lot of heat.
None of this is mysticism. It is the arc and the pool telling you what they are doing, and this article is built around learning to read it: why an arc sounds different on stainless than on mild steel, why a good-looking bead can hide a cold lap underneath, and why colour reveals the atmosphere.
Short Note: This article deliberately sits above the “how do I hold a stick” level. For hands-on steps, electrode handling and first-bead setup, start with the site’s beginner guide to stick welding and its all-position rod guide, then come back to understand what the arc is actually doing.
So: a heat source, a filler and a shield, and a metallurgical bond. The rest of this article takes each one apart, from the arc strike to the freeze, and shows you how to read the result next time you watch a weld being made.
What Welding Actually Is: Joining Families Compared
“We welded it” means different things to different people. A farrier, a refrigeration technician, and a pipeline welder use the same verb and produce joints that behave nothing alike. The taxonomy comes first: what melted, what merely softened, and what was simply stuck together.
The AWS Welding Handbook, Volume 1 splits joining processes into two broad groups. Fusion welding melts the base metal, mixing molten metal with filler until it solidifies as one continuous piece. Solid-state welding never melts the base metal, pressing surfaces together under force until they coalesce while still solid. One question,did the base metal liquefy?, separates nearly every joining decision.
Quick Note: Melt and re-solidify the base metal and you have a weld with a metallurgical bond through the joint. Leave the base metal solid, melt filler between the parts, and you have a brazed or soldered joint with a temperature ceiling below the parent metal’s melting point.
Fusion welding: the family that melts the metal
Most people picture arc welding, and the arc processes dominate the trade.
- SMAW (stick): a coated electrode melts, supplying filler and shielding gas.
- GMAW (MIG): continuous wire melts under a gas shield.
- FCAW: flux-wrapped wire melts and generates its own shield.
- GTAW (TIG): a non-consumable tungsten electrode melts a filler rod or the workpiece.
- SAW: wire melts under a blanket of granular flux.
- Plasma arc: a constricted, very hot arc runs through a copper nozzle.
- Laser and electron beam: focused beams melt metal instead of an arc.
Two older processes still hold niches. Thermite welding melts filler between the parts with the heat of an exothermic reaction, a standard repair on rails and heavy castings. Oxy-fuel welding melts base metal with a burning gas flame, still used for repair and hardfacing far from power. The arc welding fundamentals reference compares the mainstream processes.
Solid-state welding: joining without a puddle
Coalescence happens below the melting temperature.
- Friction welding: one part spins against the other under pressure until the interface plasticizes and bonds, then the flash is trimmed off.
- Friction stir welding: a rotating non-consumable tool pushes through the joint, softening material ahead of it into a bonded seam.
- Ultrasonic welding: stacked parts vibrate together under pressure.
- Diffusion welding: clean surfaces are clamped at elevated temperature in a vacuum so atoms cross the interface.
- Forge welding: the blacksmith’s hammer and upset pressure.
Aerospace, automotive, and rail favor these methods because they skip the melting that causes porosity and distortion. The AWS Welding Handbook, Volume 3 covers solid-state and related processes in depth.
Brazing, soldering, and everything that is not welding
According to the AWS definition, brazing requires three conditions at once: the base metals are not melted, the filler has a liquidus above 840 to 842 °F (450 °C), and the filler wets the base metal and is drawn into the joint by capillary action. Soldering follows the same logic with one change, a filler liquidus at or below 840 °F (450 °C). That 450 °C line is not arbitrary: it sits below the eutectic temperature of most common solder alloys, so the base metal never reaches its own liquidus.
A brazed joint can approach base metal strength, yet its maximum service temperature stays near the filler’s liquidus. Heat that thickens a weld bead will soften and let a brazed joint go. That is why torch brazing with an air-acetylene setup remains the traditional choice for thin tubing and non-ferrous metals, and why it is a different operation from post-weld brazing.
Two more families get lumped in with welding by mistake. Adhesive bonding uses a cured polymer to carry the load; mechanical fastening relies on bolts, screws, or rivets. Both are legitimate answers, and often the better one. A bolted flange is serviceable, removable, and handles fluctuating loads better than a weld.
Why the label changes the rules
Codes, filler specifications, and inspection requirements pivot on the family. AWS D1.1/D1.1M:2025, the 25th edition approved by ANSI on March 19, 2025, sets requirements for welded structures of carbon and low-alloy constructional steels. A brazed joint built to a brazing specification is never accepted in place of a code weld. Filler metals are classified separately by family, and so is inspection: a code weld demands qualified procedures, qualified welders, and documented results, while a brazed assembly may fall under a lighter specification.
“Will this joint hold?” is the wrong question until you know whether you are looking at a fusion weld, a solid-state bond, a brazed joint, or a bolt.
| Family | What melts | Typical strength | Heat | Common uses |
|---|---|---|---|---|
| Fusion welding (arc, laser, electron beam) | Base metal and filler | Equal to base metal | High, local | Structural steel, pipe, pressure vessels |
| Solid-state welding | Nothing; interface plasticizes | Equal to base metal | Moderate to high | Aerospace, automotive, rail |
| Brazing | Filler only (above 450 °C) | Moderate to high | Below base metal liquidus | Heat exchangers, refrigeration, tubing |
| Soldering | Filler only (450 °C or below) | Low | Far below base metal liquidus | Electronics, PCB assembly |
| Adhesive bonding | Nothing; adhesive cures | Varies by adhesive | Cure temperature only | Composite panels, laminates |
| Mechanical fastening | Nothing | Joint design governs | Installation tooling only | Flanges, demountable assemblies |
Keep this map in mind: the arc plasma, weld pool, heat-affected zone, and defect troubleshooting that follow all belong to the fusion family, while solid-state processes skip the pool, the arc, and the shielding gas entirely.
The Physics of the Arc: How Electricity Makes a Plasma
The arc between two pieces of steel is not a flame and not a spark. It is a plasma: a gas so hot that its atoms have been stripped of electrons, leaving free electrons and positive ions swirling together. Because those particles carry charge, the gas conducts electricity, and that is the whole trick behind welding. You cannot hold an arc in ordinary air, but you can hold one in ionized gas.
Ignition starts at the cathode tip, where the electric field is fierce enough to tear electrons off atoms. They accelerate across the gap, knocking more electrons loose from neutral gas, an avalanche that ionizes the column almost instantly. Molecular gases break apart too. That is dissociation: the carbon dioxide shielding gas used in MIG splits into carbon monoxide and free oxygen atoms at arc temperature. Those oxygen atoms oxidize the puddle and raise porosity risk, which is why argon-rich mixes have displaced pure CO2.
Electrons come back down, and when a fast one meets an ion the pair can recombine and release their energy as a photon, while excited atoms emit light as they relax. That recombination radiation is the white glare of the arc, and its color is diagnostic: gas mixes tint it, and welders read those shades.
The Three Parts of Every Arc
- The cathode, where electrons are emitted: the tungsten in TIG, the contact tip and wire stub in MIG.
- The arc column, the glowing plasma between electrodes and the hottest part of the process.
- The anode, which collects electrons and, in DCEN, receives most of the arc heat.
Voltage along the column runs at roughly 1 volt per millimetre of arc length. At 100 amps, that gradient costs about 100 watts per millimetre, so a long arc is harder to sustain and costly. Short arcs are hotter and more focused; long arcs wander, spread heat and eventually blow out.
Heat splits very differently at each end of the arc, and that is the entire reason polarity exists:
| Where the arc heat goes | Anode (+) | Cathode (−) |
|---|---|---|
| Into the workpiece | ~40% | ~90% |
| Into the electrode and surroundings | ~60% | ~10% |
The University of Alberta’s arc welding notes and Eagar’s MIT work set out the current and temperature fields in full if you want to go deeper.
Polarity in Plain Language
Electrons flow from negative to positive and carry most of the arc heat, the opposite of ordinary wiring where the metal conductor carries. Flip the polarity and you flip where the heat lands.
DCEN (electrode negative, workpiece positive) makes the workpiece the anode, so the base metal absorbs the larger share of arc heat. It is the default for TIG on steel, stainless and nearly everything else, the most efficient way to put energy into the joint, and it keeps the tungsten cooler and longer lived.
DCEP (electrode positive, workpiece negative) reverses that. The workpiece becomes the cathode, only about 10% of the heat goes into the base metal, and most goes into the electrode, which runs hotter and erodes faster. But cathodes do something extra: at the cathode spot, the electrostatic field breaks down the thin oxide layer and scrubs the surface clean. You can watch it as bright sparkle points across the puddle. This is cathodic cleaning.
Cathodic cleaning is why welding aluminium with an arc process works at all. Aluminium forms alumina, an oxide that melts near 2,050 °C while the metal melts around 660 °C. Strip the film or the oxide floats on the puddle and the bead refuses to fuse. Cathodic cleaning removes it as you weld, but less power is left for fusion, so you run more amperage. Prep still matters: our aluminium cleaning and prep routine does half the work, and our aluminium MIG welding guide covers the setup.
Quick Note: DCEN is the default for TIG on nearly everything except aluminium, where DCEP provides cathodic cleaning. If you are unsure of a machine’s polarity, read the labels on the machine rather than guessing from cable colors, since those are not standardized.
The Plasma Jet and Why Arcs Feel Stiff
The arc is not a static glow. The plasma expands violently as it leaves the nozzle and forms a jet travelling at on the order of 300 m/s, faster than a cruising airliner and roughly the speed of sound at room temperature.
That jet does several jobs. It blows molten metal aside, creating the depression that lets the arc force metal downward and produce penetration. It shields the pool, carries spatter, and pushes the arc tip toward the work, which is why an arc feels stiff and a TIG arc holds in a corner joint.
Arc force is not only the jet. Current through a plasma creates its own magnetic field, and that field interacts with the current, pinching the column so it stays narrow, hot and concentrated. That coupling of fluid and magnetic field is magnetohydrodynamics, the field in which Hannes Alfvén won the 1970 Nobel Prize in Physics. The NASA note on plasma arc welding covers the jet and shielding in detail, and Lincoln Electric’s arc welding fundamentals is a practical companion.
An arc has a negative resistance curve: as current rises, voltage falls. Your machine handles this with a constant-current source that holds the set amperage, which is why the arc feels unshakeable as arc length changes. It is also why a long arc on a machine set too high runs away and spits metal.
Put it together: electricity makes a plasma, the plasma becomes a hot stiff jet, and the jet plus polarity decide how heat divides between electrode and metal. Every bead shape starts here.
From Arc to Weld: Heat Flow, the Molten Pool and the Solidifying Grain
The arc only delivers the heat. The weld is a moving puddle of liquid metal: heated, stirred, fed from a rod or wire, then frozen into a continuous crystal.
Three ways heat reaches the joint
The AWS Welding Handbook puts it plainly: in fusion welding, heat reaches the joint by conduction, convection and radiation, and nearly every fusion process applies that heat externally.
- Conduction diffuses heat through solid metal. Slowest of the three, but it governs how far heat spreads into the plate beyond the bead, how fast interpass temperature climbs, and how much parent metal softens.
- Convection is heat carried by moving liquid or gas. Inside the pool it dominates transport, driven by churning metal.
- Radiation needs no medium. At arc temperatures it delivers a real share of the energy to the workpiece and warms surrounding metal, which is why a bead stays hot after the arc moves on.
The thermal cycle happens every inch of travel
The temperature distribution across a joint is always uneven and changing, and AWS notes that metallurgical changes occur progressively as the weld advances. Each point runs the same small cycle: heats, expands, melts, mixes with filler, cools. Metal 200 mm behind the arc is already cooling while metal 200 mm ahead is still untouched.
Inside the pool: four forces moving metal
The pool is not still. Four forces push liquid metal around, and their balance decides penetration depth and bead shape.
- Plasma-jet drag. The arc is a jet of ionized gas that drags molten metal backwards toward the arc, hollowing out a depression, and it grows stronger as the plasma focuses. Most arc welding fundamentals overviews describe it.
- Buoyancy. Hot metal is less dense than cold, so hot liquid rises and cool liquid sinks toward the root. Natural convection always helps.
- Electromagnetic force. Current through the molten column pinches it inward and drives metal down the axis and out beneath the arc. Higher current, stronger push, which is why amperage changes penetration so much.
- Marangoni flow. Surface tension varies with temperature and with dissolved oxygen and sulfur. Flow runs from low to high surface tension, so for clean steel it runs hot to cold, pushing metal to the edges. Sulfur reverses it, causing erratic penetration.
The AWS Welding Handbook makes the practical point: in TIG welds of steel, convection current in the pool often dominates heat transport, and the four forces together move the metal. Change the balance and you change the weld. Deep, narrow finger penetration comes from metal driven downward and inward; wide, flat beads come from metal pushed outward to the toes.
Quick Note: Bead shape is a readout, not a style choice. A flat, lazy bead usually means heat is going into the metal rather than down into the joint. On aluminium, where conductivity is high and the oxide film must be stripped back properly before fusion can happen, pool forces need taming.
The keyhole effect
When the heat source is concentrated enough, it blasts metal out of the way rather than just melting it, opening a cavity that exposes base metal far below the surface. As the source moves on, the cavity collapses behind it and metal flows in to fill it. That is a keyhole, and it is why a focused beam welds deeper than a broader arc.
Plasma arc welding is the accessible example. A NASA technical note on plasma arc welding describes a keyholing arc that pushes molten metal a centimetre or two deep, producing a weld deeper and narrower than a conventional GTAW bead of comparable size. Laser and electron beam welding take it further. The trade-off is control: a keyhole that collapses too early or blows out too violently is unstable, so keyhole modes want tight travel speed, clean parts and steady gas flow.
What happens as the pool freezes
Solidification is a competition between two kinds of grains. Columnar grains nucleate on partially melted grain boundaries of the base metal and grow toward the centre of the pool, following heat flow. Where the pool is wide or shallow enough, equiaxed grains nucleate in the undercooled centre, producing the columnar-to-equiaxed transition metalspeople look for in sound welds.
The other event is elemental segregation. Elements that do not fit comfortably in the crystal lattice are pushed ahead of the solidification front into the remaining liquid, so sulfur, phosphorus and carbon collect in the last metal to freeze, along the centreline and at grain boundaries, forming low-melting eutectics. That is the chemical root of hot cracking and centreline weakness, and it is why filler chemistry and heat input matter. Careful aluminium MIG welding practice is largely about controlling that reactive, conductive pool.
The three zones of a finished weld
| Zone | What happened there | Why it matters |
|---|---|---|
| Fusion zone | Base metal melted, mixed with filler, froze as a new alloy. | Strength comes from the mix and grain structure, and it inherits the filler’s inclusions. |
| Heat-affected zone | Never melted, but cycled through sub-zones: overheated (coarse-grained), normalized (fine-grained), intercritical and tempered. | Sub-zones differ in hardness and toughness. AWS D1.1 requires HAZ toughness because this band fails first. |
| Unaffected base metal | Heated, but not enough to change its microstructure. | Reference properties. Post-weld heat treatment changes this zone and the HAZ too. |
Why the HAZ is the weak link
Strip the joint down and the logic is clear. The fusion zone is a fresh, well-mixed casting, as strong as or stronger than the plate it joins. The base metal is unchanged. It is the narrow band in between that was cooked without being remade: grains coarsened, phases partly transformed, properties degraded, often only a few millimetres wide.
On top of that, the joint is under stress while it freezes. As weld metal solidifies and shrinks, it exerts stress on the surrounding weld metal and HAZ, building as the weld cools to ambient temperature, sometimes reaching the yield point of the base metal. Shrinkage restraint turns that stress into distortion and cracking in the weakest band of the assembly. That is why “the weld is only as good as its HAZ” exists. The arc is the easy part.
The Main Welding Processes, Compared Honestly
Picking a welding process isn’t about finding the “best” one, it’s about finding the right one for the joint in front of you. Two questions narrow the field the most: where does the heat come from, and what protects the molten pool while it melts?
As the arc welding fundamentals overview from AWS explains, most of the processes below fall under the arc welding family. These use an electrical arc to create fusion between metals, often with a consumable electrode supplying filler. For a full breakdown of each, Volume 2, Part 1 of the AWS Welding Handbook covers them in their own chapters.
Quick Note: Filler metals are not interchangeable between processes. A5.1 governs SMAW electrodes, A5.18 covers GTAW and GMAW filler metals, A5.20 covers FCAW, and A5.14 covers aluminium. Pick the process first, then match the filler spec to it.
| Process | Heat source and consumable | Pool protection | Deposition rate | Heat control | Positions | Where it wins | Equipment and consumable cost |
|---|---|---|---|---|---|---|---|
| SMAW (stick) | Constant-current arc; coated electrode melts to make the bead | Flux coating decomposes into shielding gas and slag | Low | Coarse: set by amperage and arc length | All, including vertical and overhead | Field repairs, structural steel, thick plate, dirty or rusty surfaces | Lowest capital cost, electrodes are cheap, slow labour |
| GMAW (MIG) | Continuous solid wire feed; wire carries current and supplies filler | Shielding gas (argon, CO2 or mixes) | High | Fine: wire feed speed, voltage and travel | All, best flat and horizontal | General fabrication, production, thin to medium plate | Moderate capital, good wire availability, fast cycle time |
| FCAW | Continuous wire with flux core; gas-shielded or self-shielded | Flux core, with added gas on the gas-shielded types | High to very high | Good, but less refined than GMAW | Flat and horizontal mainly | Shipyard and site work, heavy plate, mill scale, outdoor use | Moderate, self-shielded types need no gas bottles |
| GTAW (TIG) | Non-consumable tungsten electrode; filler added by hand | Inert gas (argon or helium) | Low | Excellent: current and gas flow independently | All, best in any position | Thin gauge, stainless, aluminium tube, precision and repair work | Highest capital and slowest labour, high quality results |
| SAW | Continuous bare wire consumed under a flux blanket | Molten flux layer covering the joint | Very high | Good but directional, mostly flat and horizontal | Flat and horizontal | Heavy plate, long seams, tanks, pipe and ship sections | Higher capital, low consumable cost per metre |
| Plasma arc | Constricted arc between a tungsten and the workpiece | Inert or mixed gas | Medium | Good, with a highly focused arc | All for cutting, limited for welding | Cutting, bevelling and keyhole welding of thicker plate | Moderate to high, plasma torches cost more |
| Laser | Focused beam, usually fibre or CO2 | Gas assist or contact tip shielding | High | Very precise, narrow heat affected zone | All on automated cells | High-volume automated work, thin to medium section, low distortion | High capital, low consumable cost |
| Solid state | Mechanical energy: friction, ultrasonic, spin or current | None, no molten pool | Very high for resistance spot welding | Fixed by tooling and machine settings | Fixed shapes or projected joints | Automotive panels, aluminium structures, hermetic sealing | High capital, dies and tooling costs |
SMAW: The Process That Survives Everything
Stick welding only needs a machine, a pair of leads and a box of electrodes, which is exactly why it’s still the go to for site work, farm repairs, and structural erection. Its flux coating provides both shielding and a protective slag, and the arc remains stable in almost any position. The trade off is speed and adjustability. Amperage, arc length and travel are what govern the pool, with no separate gas or filler feed to fine tune it. For a full rundown of its strengths, it’s worth looking at why SMAW still wins.
GMAW and FCAW: The Shop and Site Workhorses
Although they look nearly identical, MIG and flux core are built for two different worlds. GMAW feeds a clean solid wire with shielding gas, allowing for precise control through wire feed speed and voltage. FCAW swaps in a flux filled wire, and its self-shielded varieties require no gas bottles at all. This makes them far more practical for outdoor jobs, whilst also giving them more tolerance for mill scale and rust. For aluminium in particular, the two split once again, with our guides for aluminium MIG welding and aluminium flux core welding breaking down where each applies.
GTAW: Control Over Throughput
TIG welding uses a non-consumable tungsten to establish the arc, with filler added entirely by hand. This grants independent control over current, gas flow and filler deposition, making it ideal for thin gauge stainless, aluminium tube, and precision repairs where even a small heat affected zone matters. With no wire feed to fall back on, it’s slow and unforgiving on consumables, however the results are unmatched. The level of precision can go all the way to aerospace, which is covered in our aerospace TIG welding guide.
High Deposition and Automated Processes
For long seams, submerged arc welding hides the arc under a blanket of flux, achieving some of the highest deposition rates found in common shop work. Plasma arc welding constricts that arc into a tight jet, excelling at cutting and bevelling, with the ability to keyhole weld thicker sections. Laser welding takes this focus even further with an extremely narrow heat affected zone, though it only becomes cost effective at volume. Our Alpha laser welding guide explores these systems in more detail. For repairs or removing defects, carbon arc gouging is still used, with our article on air arc welding and gouging explaining the process. Electron beam welding goes even further, however its vacuum chamber keeps it out of all but the most specialized facilities.
Solid State: Welding Without An Arc
Not every joint requires melting metal. Friction welding bonds parts by mechanical friction heat, whilst friction stir welding uses a rotating tool to join them in a solid state with a narrow, predictable heat affected zone. For sheet metal, resistance spot welding can produce tens of thousands of welds per day on a single assembly line. The limitation across all of these is flexibility. These processes are dictated by tooling, making them unbeatable for automation, but unpractical for freehand repairs.
Skill Over Stereotypes
The idea that MIG is strictly for production and TIG is strictly for hobby work couldn’t be further from the truth. Skilled TIG welders are a staple in stainless tanks, food and dairy equipment, pressure parts and aerospace assemblies, whilst GMAW is relied on for speed and filler deposition across countless production lines. Regardless of preference, structural work must still meet qualification requirements set by the current AWS D1.1 structural welding code, which treats both processes as equals. In the end, the badge doesn’t matter. The joint does.
Argon dominates because it is the most abundant noble gas, chemically inert and cheaper than helium, which is why it gives the cleanest arc and the best control of a small pool. A 25 % CO2 mix shifts the transfer character: the arc behaves more like a spray than a dribble, the bead smooths out, spatter drops sharply and the arc tolerates dirty or rusty steel better. Pure CO2 still cuts deeper but leaves a rougher bead and more spatter, so it belongs on heavy fabrication rather than finish work.
One piece of chemistry is worth remembering. In a CO2 arc the gas does not simply wait to be ionized, it dissociates into carbon monoxide plus atomic oxygen first, and that reactive oxygen oxidizes the pool and raises the risk of porosity. That is why CO2-shielded wires carry extra deoxidizers such as silicon and manganese, and why a contaminated cylinder makes things worse. For gas selection and cylinder handling, see the air gas welding supply guide on gases, cylinders and safety.
Flux: the gas and slag factory
In stick welding the coating is a small chemical plant. As the electrode burns it decomposes into shielding gas (largely CO2, hydrogen and water vapour) that blankets the pool, and a molten slag that floats on top and is chipped away between passes. The slag also carries deoxidizers into the deposit, and the coating type sets the bead’s character: titania for smooth, fast, all-position work, basic for strength and low diffusible hydrogen, cellulose for deep penetration and pipe.
FCAW moves the same chemistry into the core of the wire, which burns to produce its own shielding gas, slag and deoxidizers. Self-shielded flux-cored wire needs no gas bottle at all, which is why it is the go-to for field pipelines and structural work in bad weather, while metal-cored wire keeps most of the slag and deposits almost as cleanly as solid wire.
Surfacing and cladding
Sometimes the weld is not meant to match the base metal, it is meant to cover it. Hardfacing and build-up welding lay down a hard, wear-resistant layer over a worn cutting edge, die or rail, so the filler is chosen for hardness and wear resistance rather than ductility. Cladding does the same job for corrosion, laying a thin layer of stainless or nickel alloy over cheaper carbon steel. In both cases the consumable carries the property you are buying, and the joint is only as good as the bond between the two layers.
Quick Note: If a bead cracks right after you start, suspect moisture before you suspect the wire. A damp spool, a wet electrode or base metal that was never dried after wet cutting can put more hydrogen into the weld than your entire preheat plan can handle.
Heat Input, Preheat, Distortion and Residual Stress
Two welds off the same machine at the same amperage can behave completely differently: one leaves a narrow, deep bead and a part that sits flat, the other a wide, bulging bead and a job that twists overnight. The difference is heat input, the energy the joint receives per millimetre of travel, and it governs penetration, bead profile, grain size, dilution, cooling rate, distortion and residual stress.
Heat input: the number that governs the weld
The formula is short enough to memorize: H = (60 × V × A) / T. H is heat input in kJ/mm, V is arc voltage, A is amperage, T is travel speed in mm/min, and the 60 converts watt-seconds into kilojoules per millimetre.
A 150 A, 22 V weld travelling at 100 mm/min gives H = (60 × 22 × 150) / 100 = 1,980 J/mm, or 1.98 kJ/mm. At 250 mm/min the same settings drop to 0.79 kJ/mm: less than half the energy per millimetre, a flatter bead, shallower penetration.
Quick Note: For SMAW, GMAW and GTAW, essentially all the electrical energy becomes heat, so arc efficiency is 1.0. FCAW and SAW differ because some arc heat goes into the wire, so AWS applies a spool arc-on factor instead. For flux-cored or submerged arc, use the procedure’s factor, not 1.0.
| Heat input | Bead appearance | What goes wrong |
|---|---|---|
| Too low | Narrow, shallow or inconsistent penetration | Hard brittle HAZ, cold cracking, lack of fusion |
| Right for the joint | Evenly filled, clean tie-in | Nothing to worry about |
| Too high | Wide, bulging bead, heavy spatter, dilution | Coarse-grained HAZ, softening, distortion |
In AWS D1.1/D1.1M:2025, prequalified fillet welds are commonly capped at a maximum heat input of 20 kJ/mm, since going beyond it can soften the heat-affected zone. That is a limit, not a target: structural fillets on ordinary mild steel usually sit well below it, often 1 to 3 kJ/mm, though a 6 mm fillet in heavy fabrication may run higher.
Preheat and interpass temperature
Preheat is the base metal temperature before the arc is struck; interpass temperature is the weld metal temperature when you return for the next pass. Codes set a minimum preheat and a maximum interpass: preheat slows first-layer cooling and drives moisture out of the joint, while the interpass maximum stops you stacking energy into earlier passes.
- Slower HAZ cooling means less hard martensite and fewer crack-prone structures
- Hydrogen has a better chance to escape before the weld metal locks around it
- A smaller temperature gradient across the joint, cutting distortion and residual stress
- A more consistent arc start and better wetting of the parent metal
AWS D1.1 Table 5.8 lists prequalified minimum preheat and interpass temperatures by material group and thickness. The 2025 edition adds normative Annex B, offering two alternative methods based on steel composition, thickness, filler metal diffusible hydrogen level and degree of restraint. Use Annex B outside the prequalified table or for crack-sensitive joints; the AWS Welding Digest guide to preheat and interpass shows how.
The supporting tools are carbon-equivalent formulas. AWS D14.8M presents four methods for avoiding cold cracks, and AWS D1.1 Annex B builds on them:
| Method | Origin | What it adds |
|---|---|---|
| CE | International Institute of Welding | The baseline, used in most preheat charts |
| CET | European | Refines CE, weights boron more heavily |
| CEN | Japanese | Very sensitive to boron, good for very low carbon steels |
| Pcm | Japanese (Yoshida) | Predicts critical HAZ cooling rate, used for linepipe steels |
Higher carbon equivalent means a more hardenable steel, a harder and more crack-sensitive HAZ, and so more preheat. Quenched and tempered steels and the alloy families in our alloy steel welding electrode guide are where that reasoning gets tested.
The trade-off, stated plainly
Push too much heat and the HAZ grain coarsens and softens, the bead widens, dilution, distortion and spatter rise; on quenched and tempered steels, extra heat can temper a hardened HAZ into a brittle condition. Push too little and the HAZ cools fast into hard martensite, tensile residual stress climbs, and hydrogen has a crack to open. Preheat is the compromise.
Distortion: the joint fights back
Distortion has a simple cause: the weld region heats and then contracts on cooling, but not evenly, and the shrinkage forces do not act through the centre of the section. Those eccentric forces stretch the weldment while it is hot, leaving it permanently out of shape once cool.
- Longitudinal shrinkage, the joint shortening along its own axis
- Angular distortion, the parts rotating about the weld line, common on T-joints and thin flange work
- Bowing, the out-of-plane bending of a plate or tube
- Buckling, the wrinkling of thin plate driven by compressive stress along the weld axis
Weld size is the lever most people miss. In a 4 mm fillet weld, doubling it gives a total weld area of 16 mm², so a 2 mm increase more than doubles the metal you add, and with it the filler metal, the electricity, the labour hours and the shrinkage causing the distortion. Over-welding to be safe bends a part and burns consumables.
- Sequencing, from the most restrained area toward the least
- Balanced welding, alternating between opposite sides or passes
- Back-step, welding a short length then stepping back
- Tacking with a matching preheat, so tacks are not the weak point
- Strongbacks and clamps to hold geometry during welding
- Preheat, reducing the temperature difference driving shrinkage
- Post-weld heat treatment for stress relief where the code allows it
Residual stress: what clamping really costs
Residual stress is the stress left in a weldment after it has cooled, created by the same non-uniform heating and restrained shrinkage that cause distortion. Where it is tensile, it works against the joint and adds to any crack-driving force; where it is compressive, it helps.
Clamps deserve a second look: often sold as the answer to distortion, they quietly worsen the stress problem. Research on A516 Gr. 70 plate found that preheating reduced longitudinal weld metal residual stress by 5 to 15 MPa and coarse-grained HAZ residual stress by 110 to 130 MPa, while rigid clamping increased residual stress in both weld metal and HAZ. Clamps control shape during welding and add stress that stays in the part; where the code allows, post-weld heat treatment relieves what remains.
Five numbers to write on the job sheet
Before striking an arc, put these in writing:
- Amperage and voltage
- Travel speed, so heat input is calculated, not guessed
- Preheat temperature and how you will verify it
- Maximum interpass temperature
- Sequence, tacking plan and strongback or clamp locations
Position matters too, because travel speed and heat control change when the joint stops being flat and horizontal. Our all-position welding rod guide covers electrode angles and technique changes, and why stick welding still wins for field work where preheat discipline and a controllable arc beat deposition rate. Master this section and you weld to a specification, not by feel.
Defects, Bad Beads and How to Fix Them
Most welding defects are not mysteries. A porosity breakout shows where the gas came from. An undercut shows what the arc was doing when it wandered off the puddle. A cold crack shows you the day it happened. Read a bead the way a mechanic reads an engine and a bad weld becomes a list of fixable causes, not guesswork.
Here is a working troubleshooting manual: what each defect looks and sounds like, the physics behind it, and the corrective action.
Porosity
Looks like: pinholes, small volcanic craters on the crown, or internal voids found only by radiography or sectioning a coupon. Sounds like: a sharp spitting crackle instead of a steady sizzle, an audible warning before the visual one.
The physics: a weld absorbs only so much gas while molten. Hydrogen is the most soluble gas in steel and its solubility drops by orders of magnitude as the metal freezes, so hydrogen still dissolved in the pool is expelled. Nitrogen and oxygen form pores too. Contamination on the plate is the most common culprit, which is why OSHA advises cleaning welding surfaces of solvent residue, paint and rust before welding. Solvent residue burns in the arc and produces hydrogen right in the pool.
Causes and fixes:
- Moisture in the consumable or gas. Wet electrodes, damp wire, a wet gas line, a dirty bottle. Bake electrodes per the maker’s guidance, keep wire dry and sealed, dry the lines.
- Insufficient shielding. Air brings nitrogen and oxygen into the arc zone. Check flow and nozzle condition, tighten hoses, never weld on a leaking line.
- Arc too long in a gas-shielded process. A long arc entrains air instead of shielding gas. Keep stick-out short, use the voltage the WPS calls for.
- Contaminated base metal. Cutting oils, rust, mill scale, zinc or aluminium. Grind back to bright metal; for aluminium the oxide must come off chemically or mechanically, as described in how to prep aluminium properly for welding.
- Nitrogen from air. Often a leaky gun or poor gun-to-work cable connection rather than the shielding gas.
Undercut, Incomplete Fusion, Slag and Spatter
| Defect | What you see and hear | Why it happens, and the fix |
|---|---|---|
| Undercut | A groove melted into the joint at the toe, about the width of the arc, with a lip of unfused metal. Hissing, spitting crackle, and the most common cause of a rejected fillet weld. | The arc column throws heat forward, so when it licks the joint side it melts base metal without adding filler and cuts the fusion line. Causes: excess current, travel too slow, long arc, wrong work or electrode angle. Fix: tilt into the joint, travel steady, check arc length on aluminium and stainless. |
| Incomplete fusion and lack of penetration | Fusion shows a dark seam line or unbonded side in cross-section; lack of penetration means the root was never melted, so the bead sits on top. Sounds normal, which is what makes it dangerous. | Fusion needs the pool to wet and bond to base metal, so a cold pool or an oxide film stops them merging: both are heat-management problems. Causes: low heat input, fast travel, a root gap too wide, dirty surfaces, poor fit-up. Fix: add heat, slow down, clean and re-fit. For aluminium, see the complete guide to aluminium MIG welding. |
| Slag inclusion | Dark pockets of solidified flux trapped in the weld metal or along the toe, common in SMAW and FCAW. Silent, so inspection finds it, not your ear. | Slag floats on molten steel, but if it is not cleaned between passes it gets trapped and rolled into the bead. Causes: poor interpass cleaning, a travel angle that lets slag run ahead of the puddle, too little heat. Fix: chip and grind every pass, keep the angle consistent, add heat where the bead runs cold. |
| Excessive spatter | Metal droplets scattered around the joint and stuck to the bead, a rougher surface, a chattering arc. Heavy, irregular spitting instead of a fine mist. | Spatter is molten metal thrown out by arc force and droplet transfer. Pure CO2 and high wire feed speed make it far worse than argon-rich mixes. Causes: CO2 or a high-CO2 mix, wire feed too high for the current, long arc, moisture in the gas line. Fix: shorten the arc, balance wire feed to amperage, move to a higher argon mix if allowed. |
Hot Cracking and Cold Cracking
Hot cracking happens while the weld is solidifying. The crack runs along the centreline or in the toe and is visible as the weld cools. The cause is low-melting eutectics from sulphur, phosphorus and other impurities that segregate to the grain boundaries as the metal freezes. A concave bead profile, high restraint and wide, deep puddles make it worse. Fix: reduce sulphur and phosphorus in the base material where possible, keep beads slightly convex, use slightly lower heat input, and increase preheat so the pool freezes more slowly.
Cold cracking appears hours or days after welding, usually in the heat-affected zone rather than the weld metal. It needs three things at once: hydrogen, high tensile stress, and a susceptible hard HAZ microstructure, so break any one of them. Use low-hydrogen consumables, maintain preheat and interpass temperature per the procedure, control restraint, and avoid quenching the joint. That third element is why preheat and interpass temperature sit at the top of the AWS D1.1/D1.1M:2025 essential variables, where a change outside the qualified range can invalidate the procedure.
Overlapping Beads, Blow-Through and Burn-Through
Overlapping means a new bead is laid partly on top of the previous one instead of alongside it, leaving a lumpy, poorly fused stack. Fix: mind the torch or electrode position, keep bead spacing consistent, plan the sequence so each pass lands beside the last.
Blow-through is molten metal falling out of the bottom of the joint, usually on thin material or with too much heat and too large a root gap. It shows as burn marks underneath and a hole in the bead. Fix: lower the heat, reduce the root gap, add a backing bar or consumable insert, slow down slightly.
Burn-through is the same failure further along, where the pool punches through the metal completely. On thin sheet, travel faster, use a smaller electrode and a backing bar. Stainless and aluminium are far more prone, as noted in what actually works with aluminium flux core welding.
Quick Note: Start the diagnosis with your ears. A steady, even hiss usually means a good arc. Spitting means gas or a dirty pool. Hissing with a wandering arc means poor arc control. Listen for two minutes before you touch a setting.
Inspection: How Defects Get Found
Visual inspection catches roughly two-thirds of everything and is the first line of defence on every code job. Beyond that:
- Dye penetrant finds fine surface-breaking cracks on clean, non-porous materials, but nothing under the surface.
- Magnetic particle finds surface and near-surface defects in ferromagnetic materials, the go-to for carbon steel.
- Ultrasonic locates internal flaws and, with phased arrays, tells the inspector exactly where they are and how deep.
- Radiographic produces an image of internal structure and is the standard for pressure piping and heavy joints.
These methods and their acceptance criteria are laid out in AWS quality and inspection guidance, including the AWS B2.1 guide referenced in the AWS Welding Handbook. Qualification matters too: under AWS D1.1/D1.1M:2025, Clause 6.8 sets out the essential variables, including preheat and interpass temperature, that determine whether a qualified WPS stays valid, and the 2025 edition split the prequalification tables into four and updated the macroetch acceptance criteria. The AWS Welding Inspector’s Guide and AWS Welding Handbook, Welding Metallurgy cover the framework in depth, and AWS Welding Digest on preheat and interpass temperature is a useful refresher on the variables that most often invalidate a procedure.
Keep a log of every weld: the joint, the parameters, the weather if you are outside, and what went wrong. Within a year you will know more about welding than most people learn in a year of classes, because defects teach the physics faster than any textbook.
Safety and Health: What the Arc Does to You
Welding safety is usually sold as a list of rules to memorise: wear this helmet, clear that area, keep the cylinder upright. But every rule exists because the arc is physically doing something to you. The arc makes hot metal fume, the fume rises, the plume finds your lungs, and the arc light cooks your cornea. Understand the mechanism and the rule stops being arbitrary.
Watch a welder running hot, dirty stringer beads on plate and you see a visible column of smoke lifting off the joint and tracking along the ceiling. That column is condensed vapour of every element in the workpiece and the consumable, oxidised in the arc and fine enough to reach the alveoli. Put that same welder on a positioner working overhead and his head sits at the top of the plume, the worst place in any shop to breathe.
What the fume does to you
Short term, most welders eventually meet metal fume fever: flu-like symptoms such as fever, chills, chest tightness and nausea, often appearing the same evening or the next morning, and classically on a Monday after a weekend away from the arc. It clears within a day or two, which is precisely why people stop taking it seriously.
The long-term picture is heavier. OSHA’s welding fact sheet states that prolonged exposure to welding fume may cause lung damage and cancers of the lung, larynx and urinary tract. Manganese in the fume is a neurotoxin, and chronic exposure produces Parkinson-like symptoms: tremor, slowed movement, gait disturbance. That is why fume extraction matters more on stainless and high-alloy work than on clean mild steel, even though the stainless job looks tidier.
Hexavalent chromium has its own rule
Hexavalent chromium, or Cr(VI), is generated when you weld stainless steel, chrome-plated parts or chromate-coated steel. It is a confirmed human carcinogen and a skin sensitiser, so coated structural steel cannot be treated like bare plate. Under 29 CFR 1910.1026, OSHA sets these airborne limits, and at the action level the employer must begin exposure monitoring and take corrective action before anyone reaches the limit.
| 29 CFR 1910.1026 limit | Airborne Cr(VI) |
|---|---|
| Permissible exposure limit (8-hour TWA) | 5 micrograms per cubic metre (5 µg/m³) |
| Action level | 2.5 µg/m³ |
Quick Note: If you weld stainless or chromate-primed steel, a fume extractor with a filter rated for Cr(VI) is not optional. It is the difference between a compliant job and a violation.
Ventilation is the law, and oxygen is never the answer
OSHA’s welding standard, 29 CFR 1926.353, requires local exhaust ventilation with movable hoods placed as close as practicable to the work, so fume is captured before it drifts into anyone’s breathing zone. The same section explicitly prohibits using oxygen for ventilation or cleaning. An oxygen-enriched atmosphere makes ignition far easier, and shop fires usually begin with a good intention and a torch.
Enclosed and confined spaces get stricter treatment. Welding or cutting zinc, lead, cadmium, chromium-bearing or mercury-bearing metals in an enclosed space requires the specified ventilation; where ventilation cannot be obtained, the welder must wear an air-line respirator with an attendant stationed outside and a rescue plan ready. Beryllium is the most demanding of all: 29 CFR 1926.55 calls for local exhaust ventilation and air-line respirators together, with no cartridge-mask shortcut.
Eyes, skin, fire and cylinders
The arc emits intense ultraviolet and infrared radiation. Catching it even briefly causes photokeratitis, and because the pain arrives hours later, plenty of people go back to work before the damage shows. A helmet with the correct shade rating, plus safety glasses underneath for grinding and chipping, is the minimum.
Hot work needs a fire watch wherever a fire hazard exists, nothing combustible within reach, and a plan for a hose or fitting that fails under pressure.
Cylinders need their own discipline: upright, capped, restrained against impact, and away from heat and traffic. Fitting a regulator to an oxygen cylinder with oily hands or an oil-contaminated fitting is a well-documented way to start a fire. And the gas you rely on every day is not the gas you breathe. This guide to welding gases, cylinders and safe handling covers the details most shops get wrong.
Shielding gases are asphyxiants. Helium, argon and carbon dioxide all displace oxygen, and they are heavier than air, so in a pit or a confined vessel the gas pools at floor level long before anyone notices. The OSHA fact sheet is blunt about the consequence: suffocation can follow quickly, with no warning.
Short Note: No PPE makes fume safe at the source. Extract it at the joint, ventilate the space, and treat respirators as the last line of defence, not the first.
None of this is meant to make welding sound dangerous. It makes it mechanical. The fume carries manganese and sometimes Cr(VI) into the air you share. Control the plume at the joint and the remaining hazards reduce to ordinary shop discipline.
Putting It Together: Choosing a Process and Building Skill
All the physics in the previous nine sections collapses into one practical question every fabricator faces: for this job, right here, which process do I reach for and why? The answer is never “whatever I have on the bench.” AWS’s own guidance is blunt about this in its Handbook of Fusion and Solid-State Welding, where the fusion and solid-state chapters make the point that the physics of the bond differs so markedly between the two families that process selection cannot be safely guessed. And the underlying logic is straightforward: process selection for any joining operation depends on what the product has to do, what it is made of, and what shape it has to be.
The Six Questions to Ask Before You Pick
Before you choose a process, you should be able to answer six questions. If you cannot answer them, you do not have a job description yet, and any process you pick is a guess.
- What is the material? Carbon steel behaves nothing like stainless, and aluminium behaves nothing like either. Galvanised coating, aluminium oxide, high-silicon castings, and unknown alloys each punish a process that would be perfect on clean mild steel.
- How thick is it? Gauge sets the heat window. Sheet under about 3 mm wants a focused, low-heat arc and clean gas coverage. Plate over 25 mm wants penetration and preheat, not finesse.
- What is the production volume? One-off repair work and a 10,000-piece production run are different engineering problems entirely.
- What position is the joint in? Flat, horizontal, vertical, overhead. A process that lays down a beautiful flat bead can be a miserable overhead process.
- How much distortion and appearance can you tolerate? Code-governed structural work often cares about strength. A visible railing or a stainless handrail cares a great deal about look, and heat input drives both.
- Which code governs? This is the question that ends arguments. If AWS D1.1/D1.1M:2025 applies, you are not free-rolling. It is the 25th edition of the structural welding code for carbon and low-alloy constructional steels, approved by ANSI on March 19, 2025, and it turns an opinion into a qualified procedure.
Those six answers also explain why process selection is not a shortcut. The AWS Welding Handbook fusion welding chapters cover a range of processes with different heat sources, consumables and shielding arrangements, and each one behaves differently on each of the six variables above.
A Simple Decision Guide
| Situation | First choice | Why |
|---|---|---|
| Thin sheet, repair work, clean-looking welds | GTAW (TIG) | Total heat control, no filler needed on light gauge, clean bead |
| General fabrication and production throughput | GMAW (MIG) | Fast, cheap, easy to automate, good on most steels |
| Outdoor or field work, no shielding gas | SMAW (stick) or FCAW | Self-shielding flux, portable, tolerant of rough surfaces |
| Heavy plate, long seams, flat and horizontal | SAW | Highest deposition rate, flux blankets the pool, less fume |
| Thin stainless or aluminium tube | GTAW | Heat control and inert gas protect the reactive alloys |
| High-volume automated production | Robot GMAW or laser welding | Repeatability, cycle time, consistency of weld parameters |
Quick Note: Aluminium changes the map. Real-world practice on aluminium often lands on MIG with the right gun configuration, or flux-core where the base metal is dirty and the part cannot be cleaned properly. See our guides to aluminium MIG welding and aluminium flux core welding before assuming a carbon steel rule carries over.
And a caution about process prejudice. Stick welding’s advantages are real and worth defending: it works in the rain, it does not care about wind, and it is forgiving. But “GMAW is for production and GTAW is for hobbyists” is a mistake. TIG is chosen because it offers precise heat control and clean results on thin and reactive metals. GMAW wins on speed and cost. Both processes appear in professional shops every single day, and plenty of professional shops run both on the same job.
Why Procedures Exist at All
Here is the connection between all the physics and the paperwork. Weld performance is not a matter of luck or instinct. It is controlled by documented, repeatable parameters: voltage, amperage or wire feed speed, travel speed, shielding gas, preheat, and interpass temperature. That is exactly what a Welding Procedure Specification exists to capture, and exactly what a welder qualification then proves the operator can do.
When the parameters drift, the physics drifts with them. More heat means more grain growth in the HAZ. Faster travel means less time for the pool to degas, so porosity goes up. Longer arc means more nitrogen and oxygen get into the pool. You have already seen these mechanisms; the WPS is simply the written version of them.
A Learning Path That Actually Builds Skill
Want to genuinely understand welding rather than just copy settings? Work through this sequence.
- Pick one process and stay there. Learn GMAW properly, or TIG, or stick. Switching constantly keeps you a beginner at everything.
- Learn to read your own bead. Sound, width, ripple, colour of the spatter, where the bead humps. This is the diagnostic skill the physics makes possible, and arc welding fundamentals is a good place to sharpen it.
- Keep a weld log. Date, material, thickness, settings, position, and what you found when you cut it. This is how a hobbyist starts thinking like a fabricator.
- Measure with calipers. Before and after. Track how much the part moved; that is distortion data, not opinion.
- Find the HAZ on your own work. A light pickle or etch on a sacrificial coupon reveals the transition zone that was invisible while you were welding it.
- Read a code. AWS D1.1 for structural steel, D1.6 for stainless. Even slowly, the language of essential variables, fillets and groove dimensions changes how you set a bevel.
- Work toward certification. AWS Certified Welding Inspector, and ASME Section IX for the qualification path used in pressure and boiler work.
For perspective on how the credential route is structured, our overview of aerospace welding school and careers shows how much documentation, procedure discipline and inspection ride along with the torch work in regulated industries.
The Honest Limits of Knowing This
Three caveats matter more than anything else in this article.
Welding is a manual skill. Theory tells you what should happen. Feedback and inspection tell you what did happen. The gap between the two closes with practice, not reading. This article can explain the physics, but it cannot replace a qualified procedure, and a page of prose cannot substitute for a WPS reviewed by someone qualified to review it.
Codes and consumable specifications change. Preheat and interpass requirements evolve between editions, and filler metal classifications get revised. Always work from the current edition of the code and the current manufacturer data sheet, not from a blog post, including this one.
Nothing here substitutes for inspection. A bead can look excellent and still be rejected. A bead that looks rough can be perfectly acceptable under a code, because the code measures properties, not appearance.
The Whole Chain in One Line
Electricity becomes plasma. Plasma becomes heat. Heat melts and mixes metal. The metal solidifies into a continuous grain. That continuous grain is what “welded” actually means, and it is why a properly made joint can outlast the plate that fed it. Everything else in this article is just the detail of how that happens, and how to do it deliberately, repeatably, and safely.
Pick your process from the six questions. Write your parameters down. Read the bead. Inspect the result. Then repeat it well enough that anybody can follow your paperwork and get the same weld you got.
Frequently Asked Questions
How hot does a welding arc actually get?
The arc column reaches roughly 10,000–20,000 °C, far hotter than the visible surface of the Sun, while the tip of the electrode and the weld pool sit closer to 3,500–6,500 °F. Most of the visible brightness is intense ultraviolet and visible radiation, which is why welding requires a filtered helmet and why arc eye is a real injury.
Is a welded joint stronger than the base metal?
It can be, and it often is. A properly made fusion weld with matching filler metal and controlled cooling can match or exceed the strength of the parent metal in the weld zone. The weak link is usually the heat-affected zone, where the base metal’s microstructure has been altered by the thermal cycle, and the heat input that created the strong weld metal may also have softened the HAZ.
What is the difference between welding, brazing and soldering?
In welding, the base metal melts along with the filler. In brazing and soldering, only the filler melts and flows into the joint by capillary action, with the base metal staying solid. AWS distinguishes them at 840–842 °F (450 °C): filler above that is brazing, at or below it is soldering.
Why does some steel crack right after welding?
The most common cause is hydrogen-assisted cold cracking: diffusible hydrogen from the consumable or moisture combines with a hard, susceptible heat-affected zone while high residual tensile stress is still locked in. The controls are low-hydrogen consumables, correct storage and baking, appropriate preheat and interpass temperature, and low restraint.
What is the heat-affected zone and why does it matter?
The HAZ is the band of base metal next to the weld that was heated enough to change its microstructure without melting. Depending on peak temperature it can be hardened, softened, or embrittled, and it often governs the fatigue life and crack resistance of the whole assembly.
Why is my weld full of holes, and how do I fix it?
Those are porosity bubbles, usually from gas trapped during solidification. Common causes are contaminated or damp base metal, dirty shielding gas or a leaking hose, drafts blowing away the shielding, an arc length too long for the process, and — with CO2 — too much reactive oxygen. Clean the joint, check gas flow and bottle pressure, fix leaks, and shorten the arc.
Can I weld aluminium with a stick electrode?
Yes, with AC or DC reverse polarity, because the cathodic cleaning action breaks up aluminium’s tenacious oxide layer. But it is unforgiving of technique, and most fabricators use TIG or MIG with a spool gun or a push/pull gun on a dedicated aluminium wire for better control.
How do I know if a weld is good without destroying it?
Start with visual inspection for the surface defects you can see — undercut, porosity, overlap, poor profile — then use dye penetrant for surface-breaking cracks on non-porous metals, magnetic particle on ferromagnetic steel, and ultrasonic or radiographic methods for internal flaws. Under AWS D1.1, the procedure and welder must also be qualified and documented before the joint is considered code-compliant.
Nicholas Harper
Editorial TeamCovering welding processes, equipment selection, and shop-floor practice for WeldZen — with every claim checked against the governing code.
Related Studies
Get The Next Benchmark First
New field studies and calibration sheets, once a week. No spam.