Vertical-Up Fillet With Undercut and Root Lack of Fusion
A welder is making a 10 mm fillet weld in the vertical-up (3F) position on a column stiffener using 1.2 mm ER70S-6 wire and 75/25 gas. Settings are 24 V, 210 A, gas 14 L/min, stick-out 20 mm. The torch is angled about 45° to the plates. He uses a 20 mm wide weave and travels fast to avoid the pool dropping. A roller door 3 m away is open and there is a noticeable draft. The finished weld shows undercut up to 1.5 mm on the vertical plate, a convex bead with lack of fusion at the root, and scattered spatter. The welder says vertical-up always looks like that and wants to grind the undercut and paint over it.
1. Diagnose the causes of the undercut and the root lack of fusion. Which causes are parameter, technique and environmental?
Read the worked answer
Undercut (1.5 mm on the vertical leg): 24 V/210 A is high for vertical-up on 10 mm; it produces a large, hot pool that wets the sidewall but drains away before it freezes. The 20 mm weave is too wide, so the arc dwells at the toes and then moves quickly, leaving the toe underfilled. A 20 mm stick-out also reduces effective current control and makes the arc wander.
Root lack of fusion: fast travel and a wide weave mean the arc does not penetrate the root corners; a long stick-out and a push angle can direct heat at the face rather than the root. If the root gap is tight or fit-up is poor, the root never opens.
Environment: the open roller door creates a draft that can blow shielding gas away, causing porosity and an unstable arc; it does not directly cause undercut but makes the weld harder to control.
Technique: pushing upward with a wide weave and fast travel is the main operator factor. Spatter is consistent with high voltage and long stick-out.
2. Give the corrected technique and settings, and explain why each change works.
Read the worked answer
- Reduce heat input: set a vertical-up range, typically about 18–21 V and 110–150 A for 1.2 mm wire, and match WFS to the WPS. Lower current lets the pool freeze before it runs.
- Shorten stick-out: 10–12 mm, not 20 mm. This restores current control and keeps gas close to the arc.
- Narrow the weave: use a stringer or a weave no wider than about 15 mm, with a short pause at each toe to fill the undercut tendency, then move quickly across the centre.
- Travel: slower and more consistent than the current fast pass; let the arc do the work rather than racing the pool.
- Torch angle: work angle about 45° to both plates; travel angle close to perpendicular, with a slight drag (0–10°) so the arc and gas cover the root.
- Gas and draft: 15–18 L/min at the nozzle, and close the roller door or erect a windbreak. Verify with a flow test.
- Sequence: a root pass that is fully fused before the cover passes; inspect each pass, not just the final one.
3. The welder proposes grinding the undercut and painting the joint. Evaluate this disposition.
Read the worked answer
Grinding and painting is not an acceptable disposition. Undercut of 1.5 mm on 10 mm is about 15% of thickness and exceeds normal acceptance limits (often 0.5 mm or 10%, whichever is smaller). Grinding it smooth may make the profile look better while reducing the effective throat, and it does nothing for the root lack of fusion, which is a planar defect that can act as a crack starter. Painting hides both defects from visual inspection. The correct action is to mark the joint, remove the defective weld to sound metal (gouge/grind), verify removal - for a crack-like indication, use PT or MT - and re-weld to the WPS with the corrected technique. Then re-inspect, including the root if accessible. Record the repair, the welder ID and the reason (parameters and technique non-conformance), and retrain or requalify if the same error repeats.