Fillet Weld Size on a Stiffener Under Cyclic Loading
A fabricator is welding 12 mm stiffener plates to the web of a 25 mm thick built-up girder using a 6 mm leg fillet, single pass, GMAW short-circuit transfer with ER70S-6 wire and a 75% Ar/25% CO2 shield. The drawing calls for a 6 mm fillet, minimum. The inspector measures a concave profile on several welds; the actual throat measures 3.8 mm on the worst. The weld toes show a slight undercut, and the girder will see 2 million cycles of service in a bridge application. The welder argues the leg length meets 6 mm at the root, and the concavity is cosmetic. The WPS specifies 19-21 V and 180-210 A at 6.4 mm/s travel.
1. Is the concave fillet acceptable as a 6 mm fillet under typical structural code requirements? Show the reasoning.
Read the worked answer
A fillet weld size is defined by the leg length for a convex or flat profile, but a concave fillet must have a throat at least 0.7 times the specified leg. For a 6 mm fillet, the minimum theoretical throat is 6 x 0.707 = 4.24 mm. The measured 3.8 mm throat is below that, so the weld does not deliver the specified capacity even if the root leg measures 6 mm. The concavity is not cosmetic; it reduces the effective throat and concentrates stress at the toe. The correct disposition is to reject the undersized welds, require repair or reinforcement per the applicable structural welding code, and re-verify the throat after repair. The inspector should document the measured throat, not just the leg, on the report.
2. What process adjustments likely caused the concave profile, and what would you require before production resumes?
Read the worked answer
Concavity in short-circuit GMAW fillets usually comes from excessive voltage relative to wire feed speed, high travel speed, or a long arc that spreads the puddle without filling it. The stated 19-21 V at 180-210 A with 6.4 mm/s travel is on the hot, fast side for a 6 mm single-pass fillet. Before production resumes I would require: a revised WPS with lower voltage or slower travel, or a slight weave to fill the throat; a procedure qualification test (PQR) on the same thickness and position; and a first-article inspection with throat measurement. I would also check the wire feed calibration and gas flow, since a wandering arc can mimic concavity. The point is that the fix must be demonstrated on a test piece, not asserted.