A crater in welding is the depression that forms at the end of a weld bead when the arc is stopped. It looks like a small bowl or pit in the surface of the weld, and it forms because the molten metal pool shrinks as it solidifies without enough filler material to replace the lost volume. Craters are one of the most common weld defects, and if left unfilled, they can crack and compromise the strength of the entire joint.
How a Weld Crater Forms
Every weld involves a pool of molten metal that travels along the joint as the welder moves the torch. When you stop welding, that pool no longer has new filler material feeding into it, but the metal still needs to solidify. As it cools, it shrinks in two ways: the solid material contracts from thermal cooling, and the liquid metal loses volume during the transition from liquid to solid. With no fresh material to compensate, the surface of the pool sinks inward, leaving a concave depression at the termination point.
The faster you pull the arc away, the more pronounced this effect becomes. A sudden stop gives the molten pool no time to fill itself in, resulting in a deeper, more noticeable crater. The shape varies, but most craters are roughly circular and sit slightly below the profile of the rest of the weld bead.
Why Craters Are a Structural Problem
An unfilled crater isn’t just a cosmetic issue. The depression creates a point where stress concentrates under load, much like how a nick in a piece of glass makes it easier to snap. In welded joints, stress concentration at defect points like toes, roots, and craters dramatically reduces fatigue life. Research on aluminum alloy lap joints found that the fatigue strength of welded joints dropped to just 27% of the base metal’s strength, largely because of stress concentration at vulnerable points in the weld geometry.
The concave shape of a crater also means the weld is thinner at that spot than anywhere else along the bead. Under repeated loading, this thin section becomes the weakest link. Cracks that start in a crater can propagate down the length of the weld, turning a small end-of-bead defect into a full longitudinal failure.
Crater Cracks and How They Spread
Craters are especially prone to cracking because the conditions inside them are ideal for it. As the metal in the crater solidifies, shrinkage stresses pull on the cooling material from all directions. If the crater is too shallow or doesn’t contain enough filler metal, those stresses exceed what the solidifying metal can handle, and cracks form.
Crater cracks often appear as star-shaped patterns radiating outward from the center, though they can also form as a single line running along the weld’s centerline. The real danger is that a crater crack doesn’t always stay put. It can propagate backward along the weld bead, becoming a longitudinal crack that undermines a much larger section of the joint. This is a form of hot cracking, meaning it happens while the metal is still cooling rather than from external forces applied later.
Certain materials are more vulnerable. Aluminum alloys, stainless steels, and nickel-based alloys tend to have wider solidification temperature ranges and higher shrinkage rates, making crater cracks more likely in these metals than in mild carbon steel. Structural welding codes reflect this concern. AWS D1.2, which covers aluminum structures, explicitly requires that all craters be filled to the full cross section of the weld for both groove and fillet welds. Leaving an unfilled crater is a rejection-worthy defect in code work.
Manual Techniques to Fill Craters
The simplest and most effective way to prevent a crater is to never leave one behind. When you reach the end of a weld bead, instead of pulling the torch straight away, reverse direction and step back into the weld about half an inch (roughly 12 mm). Hold there for a second before breaking the arc. This backstep technique feeds extra filler into the termination point, filling the depression before it can form.
Another approach is to move the arc slightly to the side of the weld bead and finish there. This shifts the crater off the main weld profile, where it’s less likely to cause problems. Some welders also use a series of short, overlapping dabs at the end of the bead, gradually reducing the size of the molten pool so it solidifies more evenly. The right technique depends partly on the welding process and partly on the joint geometry, but the principle is always the same: don’t let the pool collapse on itself.
Crater Fill on Modern Machines
Most modern welding power sources include a built-in crater fill function that automates the process. When activated, the machine gradually reduces the welding power and wire feed speed at the end of the weld, giving the pool time to fill as it shrinks. The result is a smooth, full termination without requiring the welder to manually manipulate the torch.
How you trigger this depends on the machine’s control logic. With a two-trigger (2T) setup, releasing the trigger starts the crater fill sequence, which runs for a preset duration. With a four-trigger (4T) setup, pressing the trigger during welding initiates the sequence, and the machine holds at the reduced power level until you release. The 4T option gives you more control over how long the fill lasts, which is useful on thicker materials or wider beads where more fill time is needed. Dialing in the right crater fill settings for your wire, gas, and material combination is worth the few minutes it takes. It eliminates one of the most preventable defects in production welding.
Repairing a Crater After the Fact
If inspection reveals an unfilled crater or a crater crack, the repair process follows a straightforward sequence. The defective area is ground out to remove the crater and any cracking beneath it. Grinding should extend slightly beyond the visible defect to ensure no microscopic cracks remain in the surrounding metal. The contour needs to be smooth and flowing, with no sharp edges that could trap slag or create new stress risers.
Once the area is clean, a new weld pass fills the ground-out section. The restart should overlap into sound weld metal on both sides of the repair to ensure full fusion. The repaired area then gets the same crater fill treatment at its own termination point, because a repair that ends with a new crater hasn’t actually solved the problem. On code-governed work, the repaired section is re-inspected to the same acceptance criteria as the original weld.