Weld Defect Types: A Visual Guide for New NDT Technicians

Why New Technicians Need a Defect Vocabulary, Not Just a Method List

Most new NDT technicians learn methods before they learn defects — they can set up a UT flaw detector or run a PT wash cycle before they can confidently look at a cross-section and say "that's lack of fusion, not slag inclusion." That order gets it backwards. Every weld defect has a specific cause rooted in welding process, joint design, or procedure deviation, and knowing that cause is what lets a technician predict where to look, choose the right method, and — just as important — tell a welding supervisor something actionable instead of just "reject." This guide walks through the weld defect types a Level I or Level II technician encounters constantly, what causes each one, which NDT method actually finds it, and how it gets called against real acceptance criteria like AWS D1.1 and ASME Section IX.

Porosity: The Most Common and Most Misunderstood Defect

Porosity is gas trapped in solidifying weld metal, and it comes in several distinct forms that matter for both cause diagnosis and acceptance:

• Uniformly scattered porosity — small, evenly distributed gas pockets, usually from marginal shielding gas coverage or a slightly contaminated base metal surface
• Cluster porosity — porosity grouped tightly in one location, often from a momentary shielding gas interruption (a draft, a gas valve hiccup) or arc restart
• Piping porosity (wormhole porosity) — elongated, worm-shaped gas pockets that form when gas tries to escape through solidifying metal, common with excessive moisture in flux-cored or SMAW electrodes, or contaminated base metal
• Surface-breaking porosity — open to the surface, detectable by VT and PT; internal porosity requires RT or UT

Root causes trace to shielding gas flow rate and coverage, electrode/flux moisture content (rebaked electrodes lose their low-hydrogen certification if left out of a rod oven too long — a genuinely common shop floor failure), contaminated base metal (mill scale, oil, moisture), or excessive travel speed that doesn't give gas time to escape before solidification. Radiographic testing reads porosity clearly as rounded, distinct dark spots on the film or digital image — this rounded indication signature is exactly why RT and volumetric UT are the standard for detecting subsurface porosity, since PT and MT only catch it if it happens to break the surface. Acceptance is governed by size and clustering density limits in AWS D1.1 Table 6.1 (for statically loaded structural welds) or the tighter cyclically-loaded criteria, and separately by ASME Section VIII Appendix 4 for pressure vessels — the same porosity pattern that's acceptable on a structural gusset plate can be rejectable on a pressure boundary weld.

Slag Inclusions: The Defect That Looks Like Something Else

Slag inclusions are nonmetallic solid material — flux residue — trapped in the weld metal, almost exclusively associated with flux-shielded processes (SMAW, FCAW, SAW). They occur when a welder fails to fully clean slag between passes on multi-pass welds, or when improper welding technique (wrong travel angle, insufficient heat) traps flux ahead of the advancing weld pool. Elongated slag lines running parallel to the weld axis are a classic sign of inadequate interpass cleaning on a multi-pass groove weld.

On radiograph, slag typically appears as irregular, angular dark indications — distinctly different from porosity's smooth rounded shape, which is the primary visual discriminator a Level II film reader uses. This distinction matters practically: porosity's rounded indications are generally less stress-concentrating than slag's angular, sometimes elongated shape, which is part of why acceptance tables often treat linear slag inclusions more conservatively than equivalent-sized rounded porosity. A technician who calls every dark spot on a radiograph "porosity" without checking shape is going to under-call defects that actually carry more fatigue risk.

Lack of Fusion and Incomplete Penetration: Related but Distinct

These two get conflated constantly by new technicians, and the distinction matters for both cause and severity.

Lack of Fusion

Lack of fusion (LOF) is a failure of the weld metal to fuse completely with the base metal or with a previous weld pass — it can occur at the sidewall, between passes (interpass LOF), or at the root. Causes include insufficient heat input, incorrect travel angle causing the arc to "chase" ahead of the puddle, excessive travel speed, poor joint access geometry (a bevel angle too narrow for proper arc manipulation), or contaminated surfaces between passes. LOF is one of the most dangerous defect types because it's planar, sharp-edged, and often oriented in a way that's difficult for standard radiography to detect — RT's detection sensitivity for LOF depends heavily on the flaw's orientation relative to the film/detector and beam angle, which is a major reason ultrasonic testing (particularly phased array and TOFD) has become the preferred method for LOF detection on critical welds. A crack-like planar flaw oriented parallel to the radiation beam can be essentially invisible on film while being clearly detectable by UT approaching from an angle that intersects the flaw plane.

Incomplete Penetration

Incomplete penetration (IP) is specifically a root-area defect — the weld metal fails to extend through the full thickness of the joint at the root, leaving an unfused gap. It's caused by insufficient root opening, excessive root face dimension, insufficient welding current for the joint thickness, or misalignment preventing the arc from reaching the full root depth. On a single-groove weld inspected from one side, incomplete penetration produces a very recognizable, elongated dark line running along the root on a radiograph — sharp, well-defined edges, centered on the joint root location, distinguishing it from porosity or slag by both its razor-straight geometry and its consistent position relative to the joint.

Both LOF and IP are treated as serious, often zero-tolerance defects under most structural and pressure-retaining acceptance criteria — unlike porosity, which typically has a size and density threshold, planar fusion-type defects in critical service are frequently rejectable at any detectable size under ASME Section VIII and B31.3, because their sharp geometry makes them efficient crack initiation sites under cyclic or thermal stress.

Cracks: The Defect Type Organized by Cause, Not Just Location

Cracks get classified primarily by cause and timing, which is what actually determines the fix:

Hydrogen-Induced (Cold) Cracking

Occurs hours to days after welding, as diffusible hydrogen (from moisture in electrode coatings, flux, or surface contamination) migrates to areas of high residual stress in a hardened microstructure and initiates cracking — most common in higher-strength, higher-carbon-equivalent steels welded without adequate preheat. This is exactly why preheat and interpass temperature control (per AWS D1.1 Annex I or the project WPS) exist as procedure requirements, and why delayed cracking inspection — a re-inspection 24 to 48 hours after welding, specified on higher-hardenability steel joints — catches defects that weren't present immediately after welding.

Hot Cracking (Solidification Cracking)

Occurs during weld metal solidification, typically along the centerline of the weld bead, associated with high-sulfur or high-phosphorus base/filler metal combinations, excessive weld depth-to-width ratio (deep, narrow beads concentrate solidification stress at the centerline), or dilution from certain base metals into the weld pool.

Crater Cracks

Star-shaped or longitudinal cracks at the termination point of a weld pass, from rapid cooling and shrinkage stress concentrated at the arc-stop location without proper crater fill technique — a very common finding on visual inspection of stick welds run by less experienced welders who break the arc without a proper crater-fill motion.

Reheat Cracking

Occurs in the heat-affected zone during post-weld heat treatment (PWHT) of certain alloy steels — a defect specifically relevant to pressure vessel and piping fabricators running PWHT cycles, since the cracking mechanism only activates under that thermal exposure.

PT and MT are the front-line surface crack detection methods, but cracks that don't reach the surface — subsurface hydrogen cracking, in particular — need UT or RT. This is a case where method selection genuinely should be layered: PT/MT immediately after welding to catch surface-breaking cracks, then a delayed re-inspection with the same or a volumetric method 24-48 hours later specifically because hydrogen cracking is time-dependent and a same-day-only inspection program will systematically miss it.

Undercut, Overlap, and Profile Defects

These are geometric weld profile defects, detected by VT and measured directly with a weld gauge:

• Undercut — a groove melted into the base metal at the toe of the weld, left unfilled, from excessive current, too long an arc, or incorrect travel angle
• Overlap — weld metal that has flowed over the base metal surface without fusing to it, from insufficient heat or incorrect electrode angle
• Excessive reinforcement / underfill — weld cap profile outside the acceptable convexity/concavity range specified by the WPS and code

These are purely visual defects with straightforward gauge-based measurement against AWS D1.1 Table 6.1 limits (undercut depth thresholds vary by loading condition — cyclically loaded structures have tighter limits than statically loaded ones), making VT — done to the actual code requirements covered in a proper Article 9 procedure, not a casual look — the correct and sufficient method.

Lamellar Tearing: The Base Metal Defect That Isn't the Welder's Fault

Lamellar tearing is a defect category new technicians often misdiagnose as a welding error when it's actually a base metal problem. It occurs in rolled steel plate that has low ductility in the through-thickness (short-transverse) direction, due to elongated nonmetallic inclusions — manganese sulfide stringers, primarily — flattened during the rolling process. When a weld imposes significant through-thickness shrinkage stress on this plate, typically at T-joints, corner joints, and heavily restrained connections, the plate can tear along these inclusion planes in a characteristic stepped, terrace-like pattern, usually in the heat-affected zone or just beyond it, parallel to the plate surface.

Lamellar tearing is largely a design and material-selection issue rather than a welding-technique issue — control measures include specifying through-thickness tested plate (Z-grade steel per EN 10164 or equivalent through-thickness ductility testing), joint redesign to reduce restraint, buttering techniques that redistribute shrinkage stress away from the susceptible plate, and controlled interpass temperature. Detection is primarily UT, using angle-beam techniques specifically aimed at the plate's through-thickness plane near heavily restrained joints — a standard thickness grid aimed straight through the plate can miss a tear running parallel to the surface, similar to the geometric blind spot HIC presents in refinery plate inspection. On heavy structural steel and offshore fabrication work, where highly restrained T- and K-joints are common, lamellar tearing risk assessment during design review is often a bigger lever than any inspection technique applied after the fact.

Arc Strikes and Mechanical Surface Damage

Arc strikes — accidental, momentary arcing between the electrode or ground clamp and the base metal outside the intended weld joint — create small, localized heat-affected zones that can produce hard, brittle microstructure and even micro-cracking in susceptible steels, despite looking cosmetically minor. Codes like AWS D1.1 treat arc strikes as a defect requiring evaluation and typically removal (by grinding, followed by MT or PT verification that the strike location is fully removed) rather than something a welder can simply grind flush and ignore, precisely because the underlying heat-affected zone can extend deeper than the visible discoloration suggests.

Mechanical surface damage — gouges, chisel marks, grinding damage from other operations, or dents in Q&T (quenched and tempered) high-strength steels — carries similar concern, since these steels can be notch-sensitive enough that a mechanical dent or gouge becomes a legitimate stress concentration and crack initiation site, not just a cosmetic blemish. Both categories are purely visual findings caught during VT, and both illustrate why VT's scope in a rigorous inspection program extends well beyond "look at the weld bead" to the full surrounding base metal surface.

Building the Judgment, Not Just the Checklist

The real skill a Level II technician develops over years isn't memorizing this list — it's the pattern recognition that connects a defect's shape and location back to a probable welding cause, which is what makes an inspection report useful to a welding engineer trying to fix a recurring problem rather than just a pass/fail stamp. A technician who can look at repeated cluster porosity across a shift and tell the welding supervisor to check the shielding gas regulator, rather than just rejecting weld after weld, is delivering real value beyond the inspection itself. This pattern-recognition competency, along with the code literacy to apply the right acceptance table to the right service condition, is exactly what Atlantis NDT's NDT training and certification programs and Level II/III development paths are built to instill, and it's the foundation ASNT SNT-TC-1A qualification is designed to verify before a technician signs reports independently.

For shops managing this training pipeline across a growing technician roster, tracking who's qualified on which defect categories, which methods, and at what certification level is its own operational challenge — one that Atlantis NDT ERP handles alongside calibration tracking and project scheduling, and one where accurate field documentation through structured reporting software — capturing defect type, location, size, and photographic evidence consistently — turns individual inspection reports into a real dataset a welding engineering team can actually use to fix root causes instead of just re-inspecting the same recurring defect indefinitely.

Atlantis NDT Products & Services

Atlantis NDT pairs field expertise with software: NDT inspection management software — Atlantis ERP, a digital twin platform for asset integrity, and NDT reporting software. Build your team with NDT training & certification (ASNT SNT-TC-1A) and ASNT certification pathways, or bring in ASNT Level III consulting. Affordable, accessible, fully customizable — book a free consultation.