The CWI exam asks you to reason the way an inspector does at a joint: what must be measured, which document or code clause governs it, and what decision follows. Build preparation around that habit instead of isolated fact recall. For every topic you study, attach a question like 'what would I check on this joint, and against what?' Work through the scenarios, the measurement exercise, and the document-tracing sequence below, and keep a short note linking AWS's certification pages for administrative details.
Applying acceptance criteria instead of memorizing them
Study each discontinuity as an applied decision: what must be measured, which document sets the limit, and whether the weld passes, needs repair, or needs further examination. A criterion recalled without its conditions cannot produce a defensible answer.
Acceptance criteria differ from workmanship descriptions. A workmanship section tells you what good practice looks like; an acceptance clause tells you the measurable limit that decides conformance. Many criteria also carry conditions — joint category, loading type, material thickness, or welding position — so the same undercut depth can be acceptable on one joint and rejectable on another. Train yourself to read a criterion as a sentence with variables, not a single number to memorize.
A practical drill: for undercut, porosity, cracks, incomplete fusion, and excessive convexity, write three lines — the property measured, the governing clause, and the decision pathway. Then test the drill against practice questions and compare your rationale with the answer explanation. The typical self-created error is stopping at the first number you find and skipping a second clause that qualifies it. If your written rationale names every clause that touched the decision, the drill is working.
Tracing a joint back through the WPS, PQR, and WPQ
The WPS instructs production; the PQR proves the procedure by test; the WPQ proves the welder by test. Build document-to-problem mapping as a study exercise: for each practice question, write down which document answers the doubt it raises.
Keep the three documents distinct by their verbs. A Welding Procedure Specification directs; a Procedure Qualification Record evidences; a Welder Performance Qualification attests to a person. Essential variables — base metal grouping, process, thickness range, position, and similar items — connect them: change one beyond its qualified limit and the qualification no longer covers the work. An inspector checks production conditions against the WPS, and checks the WPS's supporting PQR when the procedure itself is in question.
Rehearse the trace as a chain on paper: production joint, then WPS in use, then PQR supporting that WPS, then each welder's WPQ. At each link ask what could break it — a parameter outside the WPS range, a base metal outside the PQR's qualified range, a position the welder never qualified in. The table below pairs a question you might face at a joint with the document that answers it; practice rebuilding that pairing from memory until each doubt you can name points to a source document.
| Inspection question | Document to consult | What you verify |
|---|---|---|
| Are the machine settings correct for this joint? | WPS | Amperage, voltage, travel, process against stated ranges |
| Has this procedure been proven by test? | PQR | Test results and qualified ranges supporting the WPS |
| Is this welder qualified for this work? | WPQ | Process, position, and material coverage of the welder's test |
| Does this discontinuity meet the limit? | Governing code or contract criteria | Measured size, length, or accumulation against the applicable clause |
| Is this the specified material? | Material test report and drawing | Grade, specification, and identity of the item delivered |
Measuring discontinuities the way scenarios expect
For each indication, decide whether size, length, or accumulation applies. Undercut is judged by depth and extent; porosity by individual size and clustering; cracks by the criterion's treatment of linear indications.
Measurement method matters as much as the limit. Undercut depth is measured from the base metal surface down into the gouge, using a bridge-type gauge or fillet weld gauge suited to the geometry — not estimated with a scale held at an angle. Porosity requires both the size of the largest individual pore and the pattern across a length of weld, because many criteria limit accumulation separately from single indications. Linear indications such as cracks are often treated more strictly than rounded ones; the applicable criterion tells you how.
Exercise: take three photos or drawings of flawed test plates — one with edge undercut, one with scattered porosity, one with a surface crack — and for each, write the measurement you would take, the tool you would use, and the clause you would cite. Self-check rubric: two points if you named the measurement dimension correctly (depth, diameter, or length), one point if you named the tool, one point if you identified an accumulation or pattern requirement. A score of six or more of eight per set is a reasonable milestone that your measurement thinking is forming; it is not a prediction of any exam result.
Checking a weld in three stages, not just one
Organize inspection knowledge by stage: before welding (materials, WPS availability, joint preparation, consumables), during welding (parameters, interpass condition), and after welding (visual, dimensions, NDT oversight). Stage discipline prevents late discovery of early faults.
Hold points make the stages concrete. A fit-up check before striking an arc can catch wrong root opening, missing bevel angle, or an unidentified material before any welding effort is spent on it. During welding, the inspector verifies that the conditions actually match the WPS — the parameters on the machine, cleaning between passes, and interpass temperature where the procedure specifies one. After welding, visual examination and arrangement of nondestructive testing come into play, which is why the CWI role includes overseeing NDT rather than performing every method personally.
A short applied question to drill: for a multi-pass groove weld, list one check per stage and name the document each check relies on. Before: joint dimensions against the drawing and WPS. During: machine settings against the WPS ranges. After: surface condition against the code's visual criteria. The mistake to avoid is treating inspection as an end-of-line activity — by then, a wrong consumable or a mis-prepped joint is buried under sound-looking passes and costs far more to resolve.
Worked scenario: judging undercut on a groove weld
A groove weld shows a continuous undercut along one edge. The plausible mistake is quoting a limit from the wrong loading category. The better decision confirms the contract's criteria, measures depth and extent properly, and documents the basis.
Simplified example: the contract adopts a structural steel code and caps undercut at 1/32 inch for the material and category in question. The inspector measures with a scale held freehand, reads about 1/32 inch, sees the weld 'looks acceptable,' and releases it. The mistake is twofold: the freehand measurement is not a defensible reading, and the inspector never confirmed which loading category or joint classification the contract invokes, though that classification can change the applicable limit. Surface appearance and conformance are separate judgments.
The better sequence: confirm the contract and its invoked criteria, measure depth with a bridge gauge, record the length of the undercut along the joint, and check whether the criterion also limits accumulated extent. If the depth is at the limit, 'at the limit' means conforming only if the criterion is exclusive — which is exactly the kind of wording to verify in the clause text rather than assume. Why it matters: an undercut call drives rework cost on one side and, if missed, a repair in service on the other, so the documented measurement is the inspector's protection in either direction.
Worked scenario: production parameters drift from the WPS
A welder runs a joint at settings outside the WPS range, and the weld's appearance is good. The plausible mistake is accepting on appearance. The better decision verifies the reading, holds the joint, and raises a documented nonconformance.
Simplified example: the WPS states 250 to 300 amperes for the fill passes; the panel meter shows 320 on the pass just completed, and the bead looks full and clean. Accepting because the weld 'looks fine' skips the actual question: does the work follow the qualified procedure? A parameter outside the stated range means the weld was not made under the procedure as written, independent of how the bead appears. The inspector's first check is instrument validity — is the meter calibrated and connected correctly — because an erroneous reading changes the response entirely.
If the reading holds up, the better decision is to hold the joint, document the observation factually — pass number, parameter indicated, WPS range cited — and escalate per the contract, rather than quietly advising the welder to adjust and moving on. Why it matters: the essential-variable logic means the weld lacks the procedure's qualified basis, which is a conformance question for the responsible parties, not a coaching moment for the inspector to resolve informally. Practice this as a written exercise: one paragraph of factual observation, one line citing the WPS clause, one line stating the action taken.
Keeping the inspector's role, records, and reporting clear
The CWI determines whether weldments meet acceptance criteria, manages qualification records, oversees nondestructive testing, and documents objectively. The role reports conformance facts; it does not design welds, direct repair methods, or absorb pressure to reinterpret criteria.
Boundaries protect both the work and the credential. Determining conformance is the inspector's job; choosing how to fix a rejected weld belongs to fabrication and engineering. Documentation should record observations and citations — what was measured, where, against which clause — not opinions or characterizations. When a contractor disputes a call, the resolution path runs through the contract and the governing code's interpretation mechanisms, not through private renegotiation at the joint. Keeping these lanes separate is itself tested material under professional standards.
For records practice, take one practice scenario and produce the paperwork: an inspection report line with the finding, a nonconformance note citing the document breached, and a note of what verification you requested before final disposition. Readiness checks before exam day: you can trace any practice joint to its WPS, PQR, and WPQ without notes; you can state measure-what, compare-what, decide for five common discontinuities; you can write a factual nonconformance paragraph; and your practice-set self-check scores sit at the milestones you set. Those are learning milestones only. For exam format, eligibility, application, and fees, rely on AWS's certification pages rather than summaries.
- Trace check: pick any practice joint and, on paper, name the WPS, its supporting PQR, and the welder's WPQ coverage in two minutes.
- Criteria check: for undercut, porosity, crack, incomplete fusion, and convexity, state the measurement dimension and the governing clause type from memory.
- Documentation check: write one factual nonconformance note with an observation, a document citation, and an action — no adjectives, no blame.
- Milestone check: set self-check score targets on practice sets and treat them as progress markers, not pass predictions.
- Administrative note: verify current exam logistics, eligibility, and fees directly with AWS, since application and renewal rules change.
References and further reading
Use these references to explore the concepts and check the latest information from the relevant organizations.