The Bead as Evidence
A finished bead records its own history — porosity, undercut, and lack of fusion each point to the specific thing that failed. · 12 min
A weld cannot keep secrets. Whatever happened while the puddle was molten — lost gas, excess heat, a cold plate — freezes into the bead and stays there, readable. This folio teaches three defects, flaws that make a weld weaker than it looks: porosity, undercut, and lack of fusion. Each has a distinct appearance, a specific cause, and a cure you can apply on the very next bead.
Guess before you learn
You run a bead with the shop door open and a stiff breeze crossing the bench. The finished MIG bead comes out peppered with small holes. What failed?
The shielding. Those holes are porosity: air reached the molten puddle when the breeze stripped away the gas envelope, and the trapped gas froze in place as bubbles. Every defect in this folio works like that — a cause you can name, printed into the metal.
9–12
3–5
Three clues to read on any weld line. Little holes: the protecting gas got pushed away and air sneaked in. A ditch carved along the edge: the arc was too hot, or moved too fast to fill its own groove. A bead lying on top like a rope: the metal under it never melted.
6–8
Porosity: gas trapped in the freezing puddle, left as round holes on or under the surface. The shielding failed — an empty cylinder, wind, too little flow, or dirty, oily, damp metal boiling off vapor. Undercut: a groove melted into the plate at the toe — the line where weld face meets base metal — and never refilled. Too much heat, too long an arc, or too fast a pass.
Lack of fusion: the bead solidified against base metal that never melted, so the two merely touch. Too little heat, too long a stickout, or an angle that let the puddle roll ahead of the arc and cushion it. It is the most dangerous of the three, because the bead can look perfect from outside.
9–12
Each mechanism is worth owning. Molten steel dissolves gases that solid steel cannot hold; as the puddle freezes, rejected gas nucleates bubbles, and any bubble that cannot float out in time is porosity. Undercut is a melting-rate imbalance: the arc cuts a channel at the toe faster than surface tension can pull filler back into it.
Fusion requires the base metal itself to reach melting at the interface. Deposit filler onto plate that stays solid and you have two surfaces pressed together — mechanically, a crack already in place, waiting for load. That is why lack of fusion sends inspectors to bend tests: it hides under clean-looking caps.
K–2
A finished weld line can show three clues. Little holes mean air got into the melt before it froze.
A small ditch along the edge means the heat dug a groove and did not fill it back.
A line that just sits on top, not sunk in, means the metal under it never melted.
Undergrad
Gas solubility drops sharply at solidification — hydrogen and nitrogen most consequentially in steel — so pore formation is a nucleation-and-escape race governed by solidification rate and puddle viscosity. Undercut correlates with high travel speed and high voltage: arc-side melting outruns mass transport of filler to the toes.
Lack of fusion is an energy-balance failure at the fusion boundary — insufficient local heat input, often disguised by adequate average heat input. Reading defects is reading the welding-procedure variables in reverse: current, voltage, travel speed, and shielding each leave a distinct signature in the solidified bead.
Postgrad
The codes formalize this reading. AWS distinguishes a discontinuity — any interruption in structure — from a defect: a discontinuity exceeding acceptance criteria for the service. Porosity is tolerated to a density limit; planar discontinuities like lack of fusion are rejected near-categorically, because fracture mechanics treats them as sharp cracks with full stress-intensity amplification.
Pore nucleation follows the supersaturation set by Sievert's-law partitioning at the advancing front; undercut has been modeled as competition between arc-pressure-driven displacement and Marangoni-driven backfill at the trailing puddle edge. In this register a bead is a data record of the process window, and defect statistics are process-control charts.
toe
The line where the weld face meets the base metal. Undercut lives here — a melted groove the puddle failed to refill.
Diagnosis is only half; the discipline is tracing cause to cure before the next bead. A defect names the variable to change. Porosity says fix the shielding — check flow, close the door, grind the metal clean and dry. Undercut says take heat out, or slow down and let the puddle fill its toes. Lack of fusion says put heat in — higher settings, shorter stickout, and keep the arc on the puddle's leading edge. One defect, one correction, one test bead.
Note
Holes plus a white, acrid smoke usually mean zinc — a galvanized coating. Stop and reread folio 2 before welding it: zinc fume is a health matter first and a porosity matter second.
Read this bead: narrow, ropey, with a groove along the top toe — the steps fade as you master them
Undercut — a melted groove left unfilled
Travel speed too fast — the puddle never had time to fill
Slow the travel; leave the settings alone
One test bead on scrap: toes filled, ripples even
Reading beads is the skill that makes practice cheap: every flawed bead now pays you information instead of costing you confidence. You will use it on every remaining folio, and hardest of all on folio 16, where a bend test brings hidden defects to the surface. Next: the five ways two pieces of steel can meet.
Practice — new ink and old, interleaved
1.From memory: the four variables of a stringer bead, with a number for each.
Work angle near 90°; travel angle a 5–15° drag; arc length one rod diameter; travel speed steady — the pace that keeps width at about twice the rod diameter.
How close were you? Grade yourself honestly — it sets your review date.
2.Re-striking mid-bead with a half-used rod: where does the new arc start?
3.Where does each of the three defects live in a cross-section?
Undercut at the toe on the surface; porosity in the bead body; lack of fusion along the fusion boundary.
How close were you? Grade yourself honestly — it sets your review date.
4.Above roughly what crosswind, in miles per hour, does MIG shielding start to fail?
5.Match each defect to its cure.
6.Sound stringer width with a 5/32-inch rod, roughly — in inches?
7.A bead sits proud on the plate, and a chisel pops a whole section off cleanly. What was it?