An undercut in welding is a groove or channel melted into the base metal along the edge of a weld that doesn’t get filled back in with filler metal. It looks like a narrow ditch running parallel to the weld bead, right where the weld meets the parent material. This defect thins out the base metal at the joint, creating a weak point that can lead to cracking under stress or repeated loading.
How Undercut Forms
During welding, the arc melts both the filler material and a thin layer of the base metal along the edges of the joint. Normally, molten filler flows outward and fills those melted edges as everything cools. Undercut happens when that process breaks down: the base metal melts away, but filler metal doesn’t replace it.
The physics behind this involve surface tension and how the molten pool moves. As temperature rises, the liquid metal’s surface tension drops, which makes it less likely to “wet” or adhere to the solid base metal at the edges. At high arc currents, the arc pressure can push the molten pool downward, leaving only a very thin film of liquid touching the sides of the weld. If that thin film solidifies too quickly, the filler metal never reaches the edges, and you’re left with a groove carved into the base material.
What Undercut Looks Like
Undercut appears as a visible groove or depression in the base metal right next to the weld toe (the line where the weld face meets the base metal). It’s typically wider at the surface and narrows as it goes deeper. The defect can run continuously along the full length of a weld or appear in short, isolated segments. It can show up on both the top surface and the root side of a joint, though external undercut is far easier to spot during a visual inspection.
Don’t confuse undercut with other profile defects. Overlap is the opposite problem, where excess weld metal spills over onto the base metal without fusing to it. Incomplete fusion leaves gaps inside the joint. Undercut is specifically material missing from the base metal surface along the weld edge.
Common Causes
Undercut rarely has a single cause. It usually comes from a combination of parameter settings and technique problems working together.
- Too much heat. Excessive voltage, amperage, or both melt more base metal than the filler can replace. This is one of the most frequent culprits, especially for beginners who crank settings up trying to get better penetration.
- Traveling too fast. Moving the torch or electrode along the joint too quickly means the molten pool doesn’t have time to flow into the edges before solidifying. The arc outruns the puddle.
- Wrong electrode angle. Tilting the electrode more than about 20 to 25 degrees from vertical pushes molten metal away from the edges of the joint instead of letting it settle evenly. This blows the puddle to one side and starves the other.
- Not pausing at the toes. When weaving (moving the electrode side to side), skipping the pause at each edge is a classic mistake. The base metal melts at the edges, but the filler pools up in the center because the electrode moves back toward the middle too quickly.
- Arc length too long. Holding the electrode too far from the workpiece spreads the arc wider, reducing its focused heat while still melting the edges. This makes it harder for filler to reach and fill the weld toes.
Why Undercut Is a Serious Problem
The groove left by undercut creates a sharp geometric change at the weld toe. Under load, stress doesn’t distribute evenly across a smooth surface. Instead, it concentrates at that notch, sometimes dramatically. Research on welded joints shows that undercut is particularly damaging under repeated or cyclic loading, the kind structures experience from wind, traffic, vibration, or thermal expansion. The stress concentration at the groove accelerates crack formation, and once a crack starts in that notch, it propagates quickly.
Studies have found that the depth of the undercut is the single most important factor controlling how much fatigue life is lost. Continuous undercuts along a weld toe reduce fatigue life to a degree comparable to having actual cracks of similar length. Even relatively shallow undercuts change the stress pattern enough to matter in structural applications. The radius at the bottom of the groove also plays a role: a sharp, V-shaped undercut concentrates stress more severely than a rounded one of the same depth.
Acceptable Limits in Structural Work
Not every trace of undercut means a weld fails inspection. Structural steel codes set specific depth and length limits. Under the widely used AWS D1.1 standard for structural welding, the rules work like this:
- Continuous undercut running the full length of a weld cannot exceed 1/32 inch (about 0.8 mm) deep.
- Intermittent undercut on connection plates, stiffeners, or along groove welds can be up to 1/16 inch (about 1.6 mm) deep, but only if no single segment exceeds 2 inches in length within any 12-inch stretch of weld.
These aren’t suggestions. An inspector measuring 1/16 inch of undercut running 4 inches straight will reject the weld, even if the total weld is 36 inches long. The 2-inch segments must be separated by enough clean weld to maintain the 2-in-12 ratio. Fatigue research supports these limits: maximum allowable depths of 0.5 to 0.7 mm (roughly 1/32 to 1/16 inch depending on joint type) align with the thresholds needed to maintain reliable fatigue performance.
How Inspectors Measure Undercut
Visual inspection catches most undercut, but measuring its depth requires a gauge. The most common tool is the bridge cam gauge (sometimes called a WG-4), a pocket-sized instrument that sits across the weld and uses a sliding pointer to measure the depth of the groove. It reads undercut depth to 1/32 inch and can also check fillet leg length, throat size, crown height, and joint misalignment. Some inspectors use a simple fillet gauge or a dedicated undercut depth gauge for quick pass/fail checks on the shop floor.
How to Prevent Undercut
Prevention comes down to controlling heat input and making sure filler metal reaches the edges of the joint.
Start by checking your voltage and amperage. If you’re seeing undercut, try reducing one or both. For flux-cored wire on structural steel (a common scenario), small adjustments of a volt or two can make the difference. Travel speed matters just as much. Slow down enough to let the puddle fill the edges, but not so much that you overheat the base metal, which causes its own set of problems.
Keep your electrode angle within about 15 to 20 degrees of vertical. When you’re weaving on vertical-up passes or wider joints, pause deliberately at each edge of the weave pattern. That dwell time lets filler metal flow into the toes of the weld where undercut forms. If you rush through the edges and spend most of your time in the center, the middle of the bead builds up while the edges starve. For flat and horizontal positions, stringer beads (straight passes with no weave) generally produce less undercut than weaving because they keep the heat concentrated and the puddle manageable.
Preheat can also help on thicker or higher-carbon steels. A warmer base metal stays liquid slightly longer at the edges, giving filler time to wet the toes before everything freezes.
How to Fix Undercut
Grinding alone doesn’t fix undercut. Since the defect is missing material, blending it with a grinder just makes the thinned area smoother without restoring the lost metal. It may remove the sharp notch, but it leaves the base metal thinner than it should be.
The correct repair is to grind out the affected area to clean metal, then lay a small stringer bead along the ground-out groove to replace the missing material. On structural work, this repair weld needs to meet the same code requirements as the original, and depending on the specification, it may need to be documented on the weld map. For cosmetic or non-structural welds, a light grinding pass to blend the notch is sometimes accepted, but on anything load-bearing, the material has to be replaced.