GMAW-P sidewall lack-of-fusion after a travel-speed increase
A fabricator is producing 8 mm S355J2 butt joints in the PC position with a single-V preparation (60° included angle, 2 mm root gap, ceramic backing). The qualified WPS uses GMAW-P, 1.2 mm G3Si1 wire, Ar/CO2 82/18, 210 A, 26 V, 12 mm wire stickout, and 35 cm/min travel. To meet a tight delivery date, the supervisor raises travel speed to 50 cm/min and reduces the root gap to 1 mm to improve fit-up. The first 12 m of production passes visual inspection, but RT on the next 20 m shows intermittent lack-of-fusion (LOF) up to 15 mm long at the fusion line on both sidewalls, mainly in the root and second pass. The machine’s pulse programme was not changed. The welder reports that the arc ‘sounds harsh’ and the toes look peaked. The customer requires ISO 5817 level B and full RT.
1. Identify the most probable causes of the sidewall lack-of-fusion and recommend a parameter window to restore fusion. Support your answer with arc-energy reasoning.
Read the worked answer
At 35 cm/min the arc energy was approximately 0.6 × 26 × 210 / 5.83 = 0.56 kJ/mm. At 50 cm/min it falls to approximately 0.6 × 26 × 210 / 8.33 = 0.39 kJ/mm. The higher travel speed reduces arc dwell time and weld-pool volume, so the arc does not wet the sidewall; the reduced 1 mm gap further lowers the root volume and encourages bridging rather than root fusion. The ‘harsh’ sound and peaked toes point to an arc length that is too short for the new speed: 26 V is low for 50 cm/min, and the pulse programme was optimised at 35 cm/min. A 12 mm stickout also reduces effective current. The LOF at both sidewalls in root and second pass is therefore consistent with insufficient sidewall wetting and poor root opening.
Restore 35–40 cm/min, 27–28 V, 10–12 mm stickout, and a 2–3 mm root gap. Check the pulse peak/background currents and wire-feed calibration; aim the electrode at the sidewall with a small weave and pause at each toe. Increasing current alone would raise deposition but not necessarily sidewall fusion, and could cause undercut. Verify the revised window with a production test piece, macro sections at the root and second pass, and RT before restarting.
2. Explain why reducing the root gap to 1 mm is not a valid productivity measure and what fit-up controls should replace it.
Read the worked answer
The root gap controls the volume available for the root bead and therefore root penetration. Reducing it from 2 mm to 1 mm makes the joint effectively tighter, so the root pass bridges the gap instead of penetrating to the backing. That creates incomplete root penetration and a sharp unfused land that can extend into the second pass. The apparent gain in fit-up speed is false economy: it converts a welding problem into a repair problem and can produce planar defects that are difficult to size by RT.
Fit-up should be controlled by the WPS, not by the welder’s convenience. Use a jig or strongbacks, verify bevel angle, root face and gap before tacking, and use the specified 2–3 mm gap with ceramic backing. Tacks should be tapered and fully fused. If gap variation is a chronic problem, improve cutting and assembly tolerance rather than closing the gap. Productivity should come from process optimisation, not from reducing joint preparation.
3. Describe the verification and WPS/PQR actions you would require before production restarts.
Read the worked answer
Quarantine the 20 m produced after the change and RT it fully, including the root and second pass; do not rely on visual inspection. Cut macro sections from a test piece made with the revised parameters at the same position and thickness, and check root penetration and sidewall fusion on both sides. Re-qualify the WPS by a new PQR under ISO 15614-1 with the revised current, voltage, travel speed and gap range, then confirm welder qualification under the revised WPS.
Check the practical causes as well: wire-feed calibration, liner condition, gas flow and mixture, contact-tip condition, and whether the pulse programme was actually selected. Record heat input and stickout in the WPS. Restart with first-off RT and a macro section, then reduce inspection frequency only after a stable run. The customer’s ISO 5817 level B requirement means any confirmed LOF is likely to be unacceptable and must be repaired before despatch.