clause5.io/ welding/ weld-defects

Discontinuity vs Defect

D1.1 uses “discontinuity” as the neutral, technical term for any interruption in the expected structure of a weld or base metal. A crack, a pore, an undercut, an inclusion — all are discontinuities. The term carries no judgment about whether the weld passes or fails.

A discontinuity becomes a “defect” only when it exceeds the acceptance criteria defined in Table 8.1. This means the same physical condition — say, a small amount of undercut on a fillet weld — can be an acceptable discontinuity on one connection and a rejectable defect on another, depending on thickness, loading, and the specific limits in the table.

This distinction matters for inspection reports. An inspector who calls something a “defect” is saying it exceeds the code limit and must be repaired. Using “discontinuity” correctly signals that the condition has been evaluated against the acceptance criteria and may or may not require action.

Table 8.1 Discontinuity Categories

Table 8.1, titled “Visual Inspection Acceptance Criteria,” organizes weld discontinuities into eight categories. Each category has separate acceptance criteria for statically loaded nontubular connections and cyclically loaded nontubular connections. An “X” in the table indicates that the category applies to that connection type.

(1) Crack Prohibition

Any crack shall be unacceptable, regardless of size or location. This is the only discontinuity in Table 8.1 with an absolute zero-tolerance criterion. It applies to both statically and cyclically loaded connections. There is no minimum length, no depth threshold, and no exception — if a crack exists, the weld fails.

(2) Weld/Base Metal Fusion

Complete fusion shall exist between adjacent layers of weld metal and between weld metal and base metal. Incomplete fusion — sometimes called “lack of fusion” or “cold lap” — is unacceptable for both connection types. Like cracks, this is a zero-tolerance criterion.

(3) Crater Cross Section

All craters shall be filled to provide the specified weld size, except for the ends of intermittent fillet welds outside of their effective length. An unfilled crater at a weld termination reduces the effective throat and creates a stress concentration. Both connection types require this.

(4) Weld Profiles

Weld profiles shall be in conformance with Clause 7.23, which defines acceptable convexity, concavity, and reinforcement limits. Excessive convexity creates stress concentrations at the weld toe. Excessive concavity reduces the effective throat below the design minimum. Applies to both connection types.

(5) Time of Inspection

Visual inspection of welds in all steels may begin immediately after the completed welds have cooled to ambient temperature. For ASTM A514, A517, and A709 Grade HPS 100W steels, acceptance criteria shall be based on visual inspection performed not less than 48 hours after completion of the weld. This delay allows delayed hydrogen cracking to manifest before the inspection is finalized. Applies to both connection types.

(6) Undersized Fillet Welds

The size of a fillet weld may be less than the specified nominal size without correction by limited amounts: up to 1/16 in for welds 1/8 in to 3/16 in, up to 3/32 in for 1/4 in welds, and up to 1/8 in for welds 5/16 in and larger. In all cases, the undersize portion shall not exceed 10% of the weld length. On web-to-flange welds on girders, underrun is prohibited at the ends for a length equal to twice the width of the flange. Applies to statically loaded connections only.

(7) Undercut

Undercut limits depend on material thickness and loading type. For statically loaded connections, material less than 1 in thick allows undercut up to 1/32 in; material 1 in and over allows up to 1/16 in, with specific accumulated-length exceptions. For cyclically loaded connections, undercut on primary tension members is limited to 0.01 in; all other cases allow 1/32 in. See the full breakdown at weld undercut acceptance criteria.

(8) Piping Porosity

Porosity limits vary by weld type, connection type, and loading. For statically loaded CJP groove welds in tension, no visible piping porosity is permitted. Fillet welds and other groove welds have specific frequency and diameter limits — for example, the sum of visible piping porosity 1/32 in or greater shall not exceed 3/8 in per linear inch of weld. Cyclically loaded connections have tighter limits. See the detailed criteria at weld porosity acceptance criteria.

The Inspection Sequence

Visual testing (VT) is required on all production welds under Clause 8.9. Every weld on the project — not just a sample — must pass the acceptance criteria in Table 8.1 before the work is accepted. VT is the baseline inspection method for all D1.1 work.

Radiographic testing (RT) and ultrasonic testing (UT) are not automatically required. They are specified only when the contract documents call for them, per Clause 8.15. When RT is specified, the acceptance criteria are found in Clause 8.12. When UT is specified, the acceptance criteria are in Tables 8.2 and 8.3. These methods detect internal discontinuities that VT cannot see — subsurface porosity, slag inclusions, lack of fusion buried within the weld cross-section.

The practical sequence on most structural projects is: the welder completes the weld, the weld cools to ambient temperature (or waits 48 hours for A514/A517/HPS 100W steels), the inspector performs VT against Table 8.1, and if the weld passes VT and the contract specifies additional NDT, the weld proceeds to RT or UT. A weld that fails VT is already a rejectable defect — it does not proceed to RT or UT until the visual condition is corrected.

When a Defect Requires Repair

When a discontinuity exceeds the limits in Table 8.1, it becomes a defect and must be repaired. Clause 7.25 governs the repair of defective welds. The general sequence is:

First, the defective portion is identified and marked based on the inspection results. The inspector specifies the extent of the defect and the acceptance criterion it violated. Second, the defective weld metal is removed — typically by grinding, air carbon arc gouging, or chipping — to sound metal. The cavity must be cleaned and inspected to confirm all defective material has been removed before re-welding. Third, the repair weld is made using an approved WPS. The same essential variables (process, filler metal, preheat, interpass temperature) apply to the repair weld as to any production weld. Fourth, the repaired area is re-inspected using the same acceptance criteria that identified the original defect.

If the original defect was found during VT, the repair is re-inspected by VT against Table 8.1. If it was found during RT, the repair is re-examined by RT against Clause 8.12. The repair must meet the same standard as the original weld — there is no relaxed criterion for repaired areas.

Repair is almost always preferred over complete removal and replacement. Replacing an entire weld introduces additional heat cycles, distortion risk, and cost. Clause 7.25 allows targeted repair of the defective portion while leaving the sound portions of the weld intact.