Root-Bend Failures on a 12 mm CJP Groove
A 12 mm A36 plate is joined with a single-V groove and FCAW-G using 1.2 mm E71T-1 wire. The procedure calls for a 2 mm root gap, ceramic backing tape and three passes. Four root-bend specimens are cut per AWS D1.1. Two specimens open on the bend face with a 4 mm elongated slag inclusion, and a macro from the same region shows 1.5 mm of lack of fusion at the root sidewall. Tensile specimens reach 480 MPa and fail in the base metal. The welder reports the root pass ran at 180 A, 22 V and 3.5 mm/s, and the interpass temperature was not measured between passes. The joint is a complete-joint-penetration groove weld in tension.
1. Interpret the bend and macro results against AWS D1.1 acceptance criteria and decide whether the joint is acceptable.
Read the worked answer
Bend specimens are judged by the size and type of discontinuity exposed on the tension face. A 4 mm elongated slag inclusion exceeds the 3 mm (1/8 in) limit used for bend-test discontinuities in D1.1, so those two specimens fail regardless of the other two passing.
The macro adds a second, independent rejection: 1.5 mm of lack of fusion at the root sidewall is a fusion discontinuity and is not acceptable in a CJP groove weld. For a complete-joint-penetration groove weld in tension, the joint must be sound through the thickness; passing tensile specimens do not offset that, because the tensile test samples only a fraction of the length and base-metal failure simply means the weld was stronger than A36 in that location.
The result is a failed procedure qualification or failed production test, depending on which the specimens came from. That requires investigation, corrective action and re-testing, not a retest of the same welds.
2. Using the reported root-pass parameters, assess whether they explain the slag inclusion and lack of fusion.
Read the worked answer
Heat input = (22 V x 180 A) / 3.5 mm/s = 3,960 / 3.5 = about 1,131 J/mm, or 1.13 kJ/mm. That is high, yet sidewall fusion still failed, which tells us the problem is placement and bead shape rather than total energy.
At 22 V and 180 A, the arc is short and the wire is not directed into the sidewall. Travel of 3.5 mm/s is slow, so a large volume of molten metal rolls ahead of the arc and freezes over the sidewall before it can fuse. That both traps slag and leaves lack-of-fusion planes at the root. With only three passes in 12 mm, each pass is large and the root pass is likely to be wide and shallow, which worsens slag entrapment against the ceramic tape.
Unmeasured interpass temperature is a separate risk: if it ran too high, bead shape flattens and slag is harder to float out; too low, and hydrogen and fusion problems increase. The presence of 4 mm slag plus 1.5 mm LOF fits a slow, large, poorly directed root pass.
3. What corrective actions and retests would you require before accepting the procedure?
Read the worked answer
Procedure changes
- Split the joint into more passes, with a controlled root pass and stringer fill passes rather than three large passes.
- Adjust voltage and travel so the arc wets the sidewall; consider a slight weave or a split weave with a dwell at each sidewall, and confirm the wire is aimed at the root.
- Verify polarity and gas flow, and clean between passes to remove slag.
- Measure and record interpass temperature at each pass against the qualified range.
Retest
- Re-qualify with new test plates: two tensile and four bend specimens, plus a macro from each test piece.
- If the original specimens were production, cut additional specimens from the same joint after repair to confirm the repair is sound.
- Record actual parameters on the test coupon, not the nominal ones, and compare with the WPS.
- Only release production after the new bend and macro results pass and the welder has been briefed on the revised technique.