Study for the CWB Level 2 Welding Inspector exam by rehearsing the full decision chain, not just flaw recognition: observe the weld, measure with the correct gauge, classify the discontinuity by its defining features, compare the measurement against the applicable acceptance criteria, and record the result in report-ready language. Most study time should go into the middle of that chain, where classification and criteria application meet.
Discontinuity vs. defect: the distinction that drives every decision
A discontinuity is any interruption in the normal structure of a weld or base metal; a defect is a discontinuity that the governing acceptance criteria reject. Level 2 judgment depends on classifying first and judging second, never the reverse.
The two terms are not interchangeable, and the difference carries consequences. A weld containing a discontinuity may be perfectly acceptable; a weld containing a defect is not, by definition. When you label an indication a defect before measuring and comparing it against the criteria, you have skipped the evaluation step that the classification system exists to enforce. Practise describing every indication as a discontinuity with measurable properties: type, location, orientation, depth or height, and length.
This ordering matters because acceptance criteria are written for specific discontinuity types with specific dimensions. Undercut rules refer to depth; porosity rules refer to size, spacing, and distribution; crack rules often treat cracks as rejectable wherever they occur. If your classification is wrong, the criteria you apply are wrong, and the accept or repair decision built on them collapses even if your measurement was accurate. Train yourself to say: this is a discontinuity of type X, measured at Y, tested against criteria for X.
Worked example to trace: an indication at a fillet toe is 0.4 mm deep and 60 mm long. If you classify it as minor undercut, your next step is measuring depth against a depth-based allowance. If you instead decide it is a crack, the decision path changes entirely because cracks are typically evaluated on presence, not size. The same visual observation leads to two different inspection actions depending on classification, which is why classification discipline comes before criteria memorization.
Separating the look-alike discontinuities that share one surface appearance
Several distinct discontinuities look nearly identical at a weld toe or surface. Distinguish undercut from toe lack of fusion, surface porosity from slag pockets, and arc strikes from spatter using location, edge character, and confirmatory checks.
Undercut is a groove melted into the base metal at the weld toe or root and left unfilled; it sits in the parent metal and follows the toe line with a sharp, definite edge. Incomplete fusion at the toe is a lack of union between weld metal and base metal; it sits along the joint interface and may show as a crack-like or bridging line rather than a gouge. A 10x magnifier and good raking light help: undercut reads as a smooth depression in the base metal, fusion failure reads as a discontinuity at the interface.
Gas porosity appears as rounded, often shiny-walled voids, frequently distributed or clustered, and can be exposed by surface condition; slag is leftover nonmetallic material, usually elongated along weld passes, dull, and associated with valleys between passes or along the weld edges in multi-pass work. Arc strikes are small localized melt spots on the base metal away from the weld, while spatter is adhered metal droplets sitting on the surface that can often be chipped free. In your notes, record which visual cue separated the pair, because that reasoning is what you are really training.
For each pair, learn one decisive confirmatory check rather than five weak ones. For undercut versus toe fusion failure, the check is whether the feature lies in the base metal surface or along the joint interface. For porosity versus slag, the check is shape and distribution: rounded and scattered versus elongated along a pass boundary. Build a two-column habit: observation, then the check that discriminates the pair.
| Look-alike pair | Discriminating visual cue | Confirmatory check | Why the distinction matters |
|---|---|---|---|
| Undercut vs. toe lack of fusion | Groove in base metal vs. line at joint interface | Raking light plus 10x magnifier at the toe | Depth-based criteria vs. interface integrity concerns |
| Surface porosity vs. slag pockets | Rounded, scattered voids vs. elongated pass-boundary residue | Shape, distribution, and relation to pass edges | Different acceptance rules govern size and distribution |
| Arc strike vs. spatter | Fused melt spot in base metal vs. loose droplet on surface | Attempt non-invasive surface examination; strikes are fused in | Arc strikes on base metal are evaluated as base metal discontinuities |
| Surface crack vs. grinding gouge | Sharp, tight, branched line vs. smooth machined depression | Edge sharpness and orientation under magnification | Cracks and mechanical gouges follow different disposition paths |
Tying the WPS and its essential variables to what you see on the weld
A welding procedure specification states the process, materials, and parameter ranges a welder must follow. Inspection verification means checking the visible weld features against those stated variables, not just against final appearance criteria.
Essential variables are the WPS elements whose change can affect the mechanical properties of the weld: things like base metal grouping, filler classification, process, position, preheat, and heat input drivers such as amperage, voltage, and travel speed. As an inspector, you verify that the weld being produced matches the qualified procedure on record, and that supporting documents trace back through qualified procedure qualification records. The skill to practise is reading a WPS as a checklist of observable and documentable conditions.
Some variables show up in the finished weld. Excessive heat input can accompany a wide, concave bead profile with deep HAZ discoloration; insufficient preheat or interpass temperature control can accompany hydrogen-cracking indications in hardenable steels, often delayed and sometimes subsurface. Others are purely documentary: filler classification and base metal grouping are verified from records and markings, not from the weld face. Knowing which variables are visually checkable and which are paper-only keeps your verification efficient and your conclusions defensible.
Trace this example: a WPS for a multipass groove weld specifies a preheat minimum and a maximum interpass temperature. On the job you observe a narrow, convex bead on a thick section with a chill-ring appearance at the start. You cannot measure preheat from the finished weld, so your verification tool is the temperature record and, on future welds, checking measurement practice before and between passes. The lesson: match each WPS variable to its correct verification method in advance instead of improvising on the weld.
Worked scenario 1: an undercut measurement that tempts a shortcut
A fillet weld shows a toe groove you measure at 0.7 mm deep over 80 mm. The wrong move is accepting it because it looks shallow; the right move is applying both the depth and length conditions of the criteria you are using.
Setup, with simplified exercise limits stated for teaching only and not quoted from any code: undercut is acceptable up to 0.5 mm deep; undercut deeper than 0.5 mm but not over 1.0 mm is acceptable only where its length does not exceed 50 mm, and deeper indications require blend grinding and re-inspection. The candidate measures the groove with a fillet weld gauge and bridge-type depth gauge, gets 0.7 mm, and accepts the weld because the depth is visually modest and the weld otherwise looks good.
The mistake is stopping at depth. The stated conditions combine depth and length: 0.7 mm exceeds the 0.5 mm unconditional allowance, and the 80 mm length exceeds the 50 mm cap on the extended allowance, so under these exercise limits the indication is not acceptable as-welded. The better decision is to reject as-welded, recommend blend grinding to remove the groove, and record the gauge reading, exact location, and length so the repair can be re-inspected against the same criteria.
Why it matters: inspection decisions are reproducible only when the full condition set is applied, including every dimension the criteria reference. A decision that rests on visual impression cannot be defended to a contractor, a reviewing inspector, or an auditor. Rehearse this exact sequence on practice welds: measure depth, measure length, apply each condition in the clause, then write the disposition sentence.
Worked scenario 2: misclassifying buried indications and choosing the wrong clause
A groove weld shows elongated dark indications along the fusion line on a film. Calling them porosity routes the evaluation to porosity rules; recognizing possible incomplete fusion or slag routes it to interface-discontinuity evaluation, a materially different disposition.
Setup: film review of a multipass butt weld shows a line of dark indications running parallel to and along the weld interface, each slightly elongated, unevenly spaced, with a couple clustered near a pass start. The reviewing candidate evaluates them under rounded-porosity distribution rules, finds the projected sizes individually small, and drafts an accept note.
The problem is the classification, not the arithmetic. Rounded porosity is characteristically discrete, roughly equidimensional, and scattered; these indications are elongated and aligned along the fusion boundary, which is consistent with incomplete fusion or elongated slag. The better decision is to hold the accept note, describe the indications by their defining geometry and location, and route the report toward the appropriate discontinuity family, including a recommendation for an additional evaluation angle or a supplementary examination method where the interpretation remains ambiguous.
Why it matters: acceptance rules are indexed by discontinuity type, so a wrong classification means a wrong clause and potentially an unsound weld in service, particularly for interface discontinuities that reduce effective throat or load-bearing fusion area. The transferable habit is a geometry check before any rule application: are the indications rounded or elongated, scattered or aligned, at a pass boundary or inside a pass? Write that classification sentence into the report before any size evaluation.
Inspection records: writing the report a reviewer can act on
An inspection record must let a reader who never saw the weld reconstruct the decision: what was examined, what was found, how it was measured, against which criteria, and what disposition follows. Vague descriptions undo accurate findings.
Compare two entries. Weak: minor undercut noted, acceptable. Strong: undercut along toe, 150 mm from tack, maximum depth 0.7 mm by depth gauge over 80 mm length; evaluated against [applicable criteria reference]; as-welded disposition: blend grind and re-inspect. The strong version names the discontinuity, quantifies it, states the measuring method, references the evaluation basis, and commits to a disposition. Every one of those elements is a field you can practise completing.
Traceability is the second discipline. A record should connect the weld to the joint, the joint to the procedure, and the finding to a location the reader can find: drawing reference, weld identification, and a position description such as distance from a datum or clock position on pipe. Date, inspector identification, and examination stage (before, during, after welding) complete the picture. Practise writing a full record for a single practice weld until all fields appear without prompting; that completeness habit is what carries into any evaluation format.
Self-check exercise with expected observations: take ten photographs of fillet and groove welds from any practice material. For each, write a four-line record: discontinuity type with discriminating cue, measurement with named gauge, criteria-based verdict under your stated exercise limits, and disposition. Score yourself on this rubric: correct classification, 7 of 10 minimum; a named measurement method present, 9 of 10; verdict consistent with the criteria you cited, 8 of 10; disposition matching the verdict, 10 of 10. Repeat the set weekly until all four thresholds hold on a fresh set of images.
A six-week preparation sequence built around the decision chain
Sequence study in three phases: build classification fluency, then apply measurement and criteria to scenarios, then drill full records under time. Each phase has a concrete output you can grade against the rubric from the previous section.
Weeks one and two: terminology and classification. Learn the standard discontinuity vocabulary and, for each term, the discriminating features and one confirmatory check. Output: the ten-image classification set from the previous section, scored. Weeks three and four: measurement and criteria. Practise gauge use on plate and pipe practice pieces or high-quality images, then write accept or repair verdicts under clearly labeled exercise limits you set for yourself. Output: five verdict sentences that each cite a dimension, a method, and a criteria condition.
Weeks five and six: integration. Combine WPS verification questions, look-alike pairs, and measurement decisions into single-page scenarios, then complete them under a self-imposed time limit so classification and recording happen at working speed. Finish with readiness checks rather than a guessed score: you can state the discontinuity versus defect distinction and apply it unprompted; you can name the correct gauge for depth, length, and profile measurements; you can write a complete record from a photograph without missing fields; and your scenario verdicts never cite the wrong discontinuity family. For administrative details about the credential itself, including registration and examination arrangements, refer to the CWB Group directly.
Adaptation note: compress the sequence if you already work in inspection by shortening phase one, but do not skip the record-writing drills in phase three. Classification knowledge and recording discipline are separable skills, and the second is the one most often left untrained.
References and further reading
Use these references to explore the concepts and check the latest information from the relevant organizations.