Porosity Cluster at the Root of a Class A Weld
A D17.1 Class A circumferential GTAW weld on 0.090 in (2.3 mm) 304L tubing is radiographed. The film shows a scattered cluster of rounded indications at the root, with the largest individual pore approximately 0.020 in (0.5 mm) and a cluster length of about 0.080 in (2.0 mm). The weld passes visual and penetrant inspection. The shop's draft disposition is 'accept, porosity is rounded and small.' The drawing invokes D17.1 Class A and requires radiographic acceptance per the standard.
1. Determine whether the porosity is acceptable under Class A and explain the reasoning.
Read the worked answer
Rounded morphology does not make porosity acceptable by itself. Class A limits for porosity are expressed in terms of individual pore size, cluster size, and distribution relative to wall thickness. At 0.090 in wall, a 0.020 in pore is roughly 22% of thickness, and an 0.080 in cluster is nearly the full wall thickness in length. Those figures are large enough that the cluster is likely to exceed Class A limits even though each pore is rounded. The correct reasoning is to compute the size ratios against the D17.1 Class A table rather than judging 'rounded equals benign.' The draft disposition is unsafe. The likely outcome is reject and repair, with root cause examination of shielding gas flow, travel speed, and joint fit-up, since root porosity in thin tubing usually traces to gas coverage or contamination.
2. The shop wants to re-radiograph at a different angle and accept if the cluster looks smaller. Is that a valid path?
Read the worked answer
No. Re-shooting to find a more favorable projection is not a disposition method; it is selecting the evidence to fit the desired answer. Radiographic interpretation must use the required technique and the film that properly reveals the discontinuity. If the original film is technically acceptable, the discontinuity is real and its size is what it is. A second view can confirm extent or location but cannot be used to shrink a measured indication. The valid path is to apply the Class A limit to the measured cluster, reject if exceeded, repair, and re-radiograph the repair per the drawing. The exam is testing whether the candidate recognizes evidence-shopping as a nonconformity in reasoning, not just in paperwork.