Fixing Common Weld Defects: Porosity, Undercut & Warping
Most welding troubleshooting advice fails in the same way: it lists every possible cause of a defect without telling you which one it usually is. That turns a five-minute fix into an afternoon of changing variables at random. What follows is organised the other way round — by what you are looking at, in the order the causes actually turn up, with the specific adjustment that addresses each one.
Porosity
What it looks like: small round holes in the weld face, or a scattering of pinholes revealed when you grind the cap back. Sometimes it is entirely internal and only shows up when the weld is cut or broken.
What it is: gas that was dissolved or trapped in the molten puddle and did not escape before the metal solidified. The gas came from somewhere — finding where is the whole diagnosis.
| Check | What to look for | Fix |
|---|---|---|
| Gas coverage | Is gas actually reaching the puddle? | Check cylinder content, regulator setting, hose for leaks, solenoid operation, and a nozzle clogged with spatter |
| Draught | Fan, open door, breeze across the work | Shield the work area — a light breeze is enough to strip the gas envelope |
| Flow rate | Too high as often as too low | Excessive flow causes turbulence that entrains air; set to the manufacturer's stated range |
| Contamination | Oil, paint, rust, mill scale, cutting fluid, galvanising | Clean back to bright metal well beyond the joint |
| Moisture | Damp electrodes, humid air, condensation on cold steel | Store electrodes dry; let cold parts reach shop temperature |
| Technique | Excessive stick-out or arc length | Shorten to specification — long stick-out puts the puddle outside the gas envelope |
Blocked nozzles and worn contact tips cause more porosity than most people credit. Keeping a consumables kit and anti-spatter nozzle gel on the bench turns a recurring quality problem into a thirty-second maintenance step.
Undercut and overlap
These two are opposites and they sit at the toe of the weld — the line where weld metal meets base metal.
Undercut is a groove melted into the base metal alongside the weld and left unfilled. It reduces the effective thickness of the part and concentrates stress exactly where you least want it. Causes, in order of frequency: current too high; travel speed too fast for the current; work or gun angle throwing the arc onto the base metal rather than into the joint; and on weaving passes, failing to pause at the edges long enough to fill.
Overlap (sometimes called cold lap) is the reverse — weld metal spilling over the base metal without fusing to it. It looks like a bead sitting on the surface rather than blended into it. Causes are typically current too low, travel speed too slow, excessive wire feed for the voltage, or a work angle that lets the puddle run ahead of the arc.
- Undercut: reduce amperage first, then slow travel slightly, then check your angle. Pause at the toes on a weave.
- Overlap: raise amperage or reduce wire feed, speed up travel, and keep the arc on the leading edge of the puddle rather than letting the puddle get ahead of it.
- Both: a weld pool that is running downhill on a vertical or overhead joint is telling you the heat input is too high for the position.
Lack of fusion and penetration
This is the defect that matters most and shows least. Lack of fusion is weld metal adjacent to base metal, or to a previous pass, without a metallurgical bond. Lack of penetration is weld metal that has not reached the root of the joint. Both can produce a weld that looks entirely acceptable from the outside and carries a fraction of the intended load.
Causes cluster around insufficient energy delivered to the right place: current too low; travel too fast; the arc aimed at the filler or the previous bead rather than the joint root; joint preparation too tight or with the wrong included angle; and welding over mill scale, slag or a previous pass that was not cleaned between runs.
Warping and distortion
Warping is not really a defect in the weld; it is the part responding to heat. Metal expands where the arc heats it, is restrained by the surrounding cold material, and then contracts as it cools — pulling the part with it. You cannot eliminate this, only manage it, and the management is mostly about sequence and heat input rather than technique.
- Reduce total heat input. Smaller beads, faster travel, appropriate current. More heat is more movement.
- Use intermittent welding where the design permits, rather than one continuous pass.
- Alternate sides and ends. Balancing where the contraction pulls is the single most effective technique available.
- Back-step. Weld short segments in the opposite direction to overall travel to distribute heat along the joint.
- Pre-set the joint so the part pulls into alignment rather than out of it — this is a skill that comes from knowing your own material and joints.
- Clamp and fixture properly, accepting that heavy restraint reduces movement but increases residual stress and, on some materials, cracking risk.
- Let it cool between passes. Impatience is a leading cause of distortion.
A common instinct is to cool the part quickly with water or compressed air to stop the movement. On many carbon steels this is at best unhelpful and at worst introduces hardness and cracking risk in the heat-affected zone. Check what your specific material requires before quenching anything.
Spatter, slag and surface problems
Spatter rarely compromises the weld itself but it wastes filler, sticks to everything, and signals that some parameter is off. The usual causes are an incorrect relationship between voltage and wire feed speed, wrong polarity for the wire type, dirty base metal, and shielding gas choice — carbon dioxide inherently produces more spatter than argon-rich mixes.
Slag inclusions are a different problem, specific to stick and flux-cored processes. Slag that gets trapped between passes because the previous run was not chipped and brushed clean becomes a permanent defect inside the weld. On multi-pass work, cleaning between passes is not housekeeping, it is part of the procedure.
Cracking
Cracking is the defect worth being conservative about, because the causes are metallurgical rather than purely technique-based and the consequences are the most serious. Broadly, cracks that appear while the weld is still hot tend to relate to weld metal composition, joint restraint and unfilled craters; cracks that appear hours after the job is finished tend to relate to hydrogen, hardenable material and cooling rate.
- Fill every crater. An unfilled crater at the end of a bead is a ready-made crack starter.
- Match filler to base material properly — on alloy steels and aluminium this is a genuine engineering decision.
- Keep hydrogen out: dry electrodes, clean surfaces, no oil, no damp.
- Respect preheat and interpass temperature requirements where the material specifies them.
- Reduce restraint where you can, and be aware that heavily clamped joints in thick section concentrate stress.
How to repair a defect properly
- Identify the cause first. Repairing without changing the parameter that caused the defect reproduces the defect.
- Remove the defective metal. Grind, gouge or cut back to sound material, with generous margin at each end of the affected area.
- Clean thoroughly. Remove slag, oxide, grinding debris and any contamination from the prepared area.
- Reprofile the joint so the repair area has a shape you can actually weld — a narrow gouged slot with vertical walls invites fresh lack of fusion.
- Reweld with the corrected parameters, and inspect again.
The habit worth building is running test coupons on offcuts before committing to real work, then destroying those coupons. Bend them, break them, look at the fracture surface. Ten minutes on scrap tells you more about your settings than an hour of inspecting finished welds you are reluctant to cut apart.
Frequently asked questions
What causes porosity in welds?
Gas trapped in the solidifying weld metal. In practice the source is almost always one of four things: shielding gas not reaching the puddle, contamination on the base metal, moisture, or a gas flow rate so high it creates turbulence and pulls air in. Work through them in that order because that is roughly the order of frequency.
How do I stop undercut?
Undercut is a groove melted into the base metal at the toe of the weld that has not been filled. It usually means too much current, travel speed too fast for the current, an incorrect work angle, or insufficient dwell at the toe. Reducing amperage and pausing slightly at the edges of the weave fixes the majority of cases.
Why is my weld warping the part?
Because welding puts localised heat into a restrained part and the metal contracts as it cools. Warping is not a mistake so much as an unmanaged consequence. Control it by reducing total heat input, using shorter intermittent beads instead of one long pass, alternating sides of the joint, sequencing welds to balance the pull, and letting the part cool between passes.
What is lack of fusion and how do I know if I have it?
It is weld metal sitting against the base metal without actually melting into it. It is dangerous precisely because it can look fine from the outside. Common causes are insufficient current, travel speed too fast, incorrect gun angle directing the arc at the filler rather than the joint, and welding over mill scale or a previous pass that was not cleaned. Bend testing a coupon is how you find out on practice work.
Should I grind out a bad weld or weld over it?
Grind it out. Welding over a defect traps it and usually adds a second defect on top. Remove the affected metal back to sound material, clean the area properly, and reweld with the parameter change that addresses the cause. Adding filler on top of porosity or slag produces a weld that looks repaired and is not.
What causes spatter and does it matter?
Spatter is molten metal ejected from the arc. It is usually cosmetic rather than structural, but heavy spatter signals a problem worth fixing — typically wrong voltage-to-wire-speed relationship, wrong polarity, dirty base metal, or the wrong shielding gas for the process. Carbon dioxide produces more spatter than argon-rich mixes by nature.
Why do my welds crack after they cool?
Cracking has more possible causes than the other defects and is the one worth taking seriously. Common contributors include excessive restraint in the joint, wrong filler for the base material, hydrogen from damp electrodes or contaminated surfaces, unfilled craters at the end of beads, and cooling too fast on materials that need controlled cooling or preheat. On anything structural, get qualified advice rather than guessing.
How do I tell a bad weld from an ugly one?
Appearance and soundness are different questions. A weld can look uneven and be perfectly sound, or look uniform and hide lack of fusion. Visual inspection catches surface defects — undercut, overlap, surface porosity, crater cracks, insufficient throat. It cannot see internal defects. On practice coupons, bend and break tests tell you what the surface cannot.