Welding joins metal parts by heating their surfaces to melting point with an electric arc, a blowpipe, or another heat source and fusing them together, often with filler metal. Steel fabrication depends on it: bolts handle the connections that must come apart, while welds carry loads everywhere else. A single defective weld can cut the capacity of a joint below what the design assumes, and repairing it costs far more than making it correctly the first time. When defects slip through unnoticed, they can resurface months later as warranty claims for construction defects against the builder, which is why competent shops document every pass and every joint.
Why Welding Defects Form
Most welding defects trace back to quality control and quality assurance failures rather than to the welding process itself. Incorrect practices, unskilled operators, damp electrodes, and rushed procedures account for the majority of rejected joints. Understanding where defects come from is the first step toward preventing them, and the same discipline applies whether a crew works in a shop or runs portable arc welding on the jobsite.
Root Causes Behind Defective Welds
Defects concentrate where procedure control breaks down. The same causes repeat across fabrication shops and site work:
- Weak quality control and quality assurance processes that let defective work pass inspection
- Unskilled or unsupervised welders working beyond their qualifications
- Excessive current or excessive arc length that overheats the base metal
- Damp electrodes and contaminated base metal surfaces
- Incorrect electrode angle and travel speed
- Joint preparation that does not match the welding procedure specification
Each cause leaves a characteristic signature. High heat melts the base metal edges and drains them into the weld, low heat leaves gaps between passes, and moisture releases gas that gets trapped as the pool freezes. Reading those signatures is how inspectors trace a defect back to the practice that created it.
Undercut and Slag Inclusions
Undercut and slag inclusions are the defects inspectors see most often, and both are preventable when caught early.
Undercut: A Groove That Cuts Into Capacity
Undercut is a groove melted into the base metal along the toe of the weld. It reduces the cross-section thickness of the parent plate, so the member loses strength exactly where the weld transfers load. Excessive current is the usual trigger: the heat melts the edge of the base metal and lets it drain into the weld pool. Poor technique adds to the problem, including insufficient filler metal, an incorrect electrode angle, and an excessively long arc. Damp electrodes widen the temperature gradient between the center of the weld and its surface, which makes the edge melt faster.
Operators who hold a steady travel speed and aim the electrode to fill the toe produce clean profiles. The habits carry across processes: the MIG welding tips from a pro cover the same ground, with consistent travel speed, correct gun angle, and wire feed matched to heat input.
Slag Inclusions in Multipass Welds
Slag is the non-metallic residue left by the flux that protects the molten pool. When it is trapped inside the weld metal, it acts like a void and reduces the strength of the weld considerably. Slag inclusions appear most often in multipass welds, where each pass covers the one below it. An inclusion that spans a large continuous cross-section does the most damage, because it interrupts the load path across the whole joint.
Why Slag Traps in Tight Spaces
Narrow grooves and deep joints give slag nowhere to escape. If cleaning between passes is inadequate, the residue from the first pass is buried under the second. A skilled welder avoids most inclusions by chipping and brushing every pass before laying the next, and by keeping the arc rooted so the slag floats to the surface instead of freezing inside.
Incomplete Penetration and Lack of Fusion
Incomplete penetration and lack of fusion both leave unfused gaps inside the joint, but they form at different locations. Penetration problems occur at the root of the weld, while fusion problems occur along the faces of the joint.
Incomplete Penetration at the Root
Incomplete penetration means the filler metal did not reach the full depth of the joint preparation. It shows up in double-sided welds when the passes from each side meet but do not overlap, leaving a plane of unfused metal in the middle. Too low a current is the most common cause, because the arc lacks the energy to melt down to the root. A shallow electrode angle has the same effect, as does a root gap too narrow for the electrode to reach. Geometric problems produce the same result: a groove with an acute included angle leaves a wedge the filler cannot fill.
The fix starts before the arc: correct the joint geometry, set the root gap to the procedure, and run the first pass with enough current to penetrate. Rework after welding means gouging out the root and starting again.
Lack of Fusion Along the Joint Faces
Lack of fusion leaves weld metal sitting on top of the base metal without metallurgical bonding. The joint looks filled from the outside and fails only under load or in testing. Contaminated surfaces, cold base metal, and a wandering arc all prevent the surfaces from melting together. Preheat and cleaning are the standard remedies.
Weld defects are not unique to steel. Every trade has its own failure modes: defects in concrete formwork systems appear when joints are poorly detailed or concrete is placed too fast, and timber members fail at knots and splits. The pattern repeats across materials: defects form where preparation and procedure break down.
Porosity and Cracking
Porosity and cracking matter most for structural safety, because they create voids and stress raisers inside the joint. The table below summarizes the four defects that show up most often in fabrication and how to deal with each.
| Defect | Typical cause | Detection method | Prevention |
|---|---|---|---|
| Porosity | Moisture, rust, or oil on the joint surface | Radiographic or ultrasonic testing | Clean and dry the joint, store electrodes in a rod oven |
| Hydrogen cracking | Hydrogen from damp consumables with delayed onset | Visual inspection after 48 hours | Dry electrodes, preheat, control the cooling rate |
| Hot cracking | High sulphur or carbon content under high restraint | Visual check of the weld centerline and craters | Control filler chemistry and joint restraint |
| Lamellar tearing | Through-thickness strain on inclusions in rolled plate | Ultrasonic testing | Improved joint detailing or higher grade plate |
Porosity: Gas Trapped in the Pool
Porosity is caused by gas bubbles trapped as the weld pool solidifies. Moisture, rust, oil, and paint on the base metal release hydrogen and other gases into the pool, and when the metal freezes faster than the gas can escape, spherical voids remain. Porosity lowers the effective cross-section of the weld and gives fatigue cracks a place to start. Keeping the joint dry and clean and storing electrodes in rod ovens eliminates most porosity.
Hydrogen Cracking in the Heat-Affected Zone
Hydrogen cracking, also called cold cracking or heat-affected zone cracking, appears hours or days after welding, which makes it easy to miss. Hydrogen from damp flux or contaminated surfaces diffuses into the steel, and under restraint the hardened heat-affected zone cracks. The delayed onset is the danger: the weld looks sound at handover and fails later under service load. Waiting at least 48 hours before final inspection catches most cases.
Steel is not the only material with hidden damage. In wood construction, timber defects such as shakes and decay reduce member capacity without being obvious at first glance; in steel, the equivalent is cracking below the surface that visual checks cannot see.
Hot Cracking and Lamellar Tearing
Hot cracking, or solidification cracking, happens while the weld metal is still cooling. High sulphur and carbon content make the metal brittle during the final stage of solidification, and high restraint pulls the joint apart as it contracts. Cracks run along the weld centerline and often start in the crater at the end of a pass.
Lamellar tearing is different: it occurs in the rolled plate, not the weld. Through-thickness strains from a restrained joint pull the plate apart along inclusions in the steel, typically under T-joints and corner joints in thick plate. Improved joint detailing and plate with better through-thickness properties prevent it.
Inspection and Testing of Welded Joints
Catching a defect at the bench costs a fraction of repairing it in the field. Inspection follows a clear sequence, and each step has a purpose.
Visual Inspection Checklist
Visual inspection is the first line of defense and catches most surface defects. A systematic check covers:
- Confirm the welder qualification matches the procedure, position, and process
- Check joint preparation, root gap, and fit-up before welding starts
- Examine the weld profile for undercut, overlap, and excessive convexity
- Look for surface porosity, crater cracks, and slag residue between passes
- Measure weld size against the drawing at regular intervals
- Recheck the joint after 48 hours for delayed hydrogen cracking
Non-Destructive Testing Options
When visual inspection is not enough, non-destructive testing finds the defects that stay hidden. Radiographic testing records the internal weld structure on film and reveals porosity and lack of fusion. Ultrasonic testing measures reflections from internal discontinuities and is the standard method for detecting lamellar tearing and crack-like defects. Magnetic particle testing works on ferromagnetic steel and finds surface and near-surface cracks, while dye penetrant testing covers non-ferrous materials.
Surface condition matters at handover too. A weld that is dimensionally correct can still be rejected because of painting defects on coated steelwork, so grinding spatter and sharp toes before coating saves a full rework cycle.
Quality Control That Keeps Defects Out
Prevention beats repair. Shops that control the process produce welds that pass inspection the first time, and the controls are straightforward.
Procedure and Welder Control
Every weld should follow a written welding procedure specification that fixes the current range, travel speed, electrode type, and preheat. Welders should hold qualification tests for the positions and processes they run. Damp electrodes belong in a rod oven, and joint surfaces need to be clean and dry before the arc starts.
Documentation and Traceability
A weld that cannot be traced to its welder and procedure is a liability. Marking each joint with the welder’s stamp, recording consumable batches, and keeping non-destructive test reports on file create a traceable record for the life of the structure.
Classification matters when a dispute arises. Defects found during inspection are patent defects, visible at handover, while those that show only under load are latent defects. Understanding the difference between latent and patent defects helps owners and contractors decide who fixes what, and it follows the same logic that governs builder obligations in new home warranties.