Study for CSWIP 3.1 by classifying every defect along two axes: location in the joint and timing of formation. A defect you can place, explain by cause, and describe by appearance is one you can inspect for, report, and refer for the right repair decision.
Lack of Fusion vs Lack of Penetration: One Question of Location
Ask where the planar flaw sits. Lack of penetration lies along the joint preparation face or root; lack of fusion lies between weld metal and unwelded surfaces beside or beneath it.
Incomplete penetration means the weld metal never reached through the full depth of the preparation or failed to bridge the root gap, so the unwelded land or closed root remains at the joint line itself. Incomplete side-wall fusion, by contrast, describes a run that failed to melt into the bevel face or the previous run, leaving a sharp, planar interface off the joint centreline. Because the two defects sit in different places, their causes differ too: penetration responds to root gap, root face size, current, and travel speed, while side-wall fusion responds more to heat input, arc angle, and technique on thick sections.
Worked scenario: on a single-V butt in 15 mm plate, the inspector records a planar discontinuity and the welder assumes incomplete penetration because the root gap had closed slightly. The better decision is to trace where the indication runs: along the root face confirms penetration problems, but a sharp line partway up the bevel face indicates side-wall fusion between the second and third run. The distinction matters because the fixes diverge: penetration problems point to preparation and root parameters, while a fusion problem points to interpass control, arc positioning, and possible cleaning of the joint faces before any welding continues.
Use sketches rather than memory alone. Draw a cross-section of a single-V joint, mark the root face, root gap, bevel face, and each run boundary, then place each fusion-related defect where it physically belongs. When you can annotate the drawing from recall, the written definitions stop blurring together, and you can justify a classification to anyone reviewing your report.
Hot Cracks vs Hydrogen-Induced Cracks: Timing Decides the Cause
Solidification cracks form as the weld metal freezes, within the weld bead; hydrogen-induced (cold) cracks appear after cooling, often in the heat-affected zone or weld metal.
Solidification cracking occurs during freezing of the weld pool when low-melting films, often enriched in impurities such as sulphur, separate the shrinking solid. It shows as a crack along the centreline or between dendrites, visible at or soon after solidification. Hydrogen-induced cracking is a delayed mechanism: hydrogen absorbed during welding diffuses and concentrates at stressed, hardened microstructures, and cracking can appear hours or even days later. It is therefore sometimes called cold cracking or delayed cracking, and it typically follows the fusion boundary or the toe of the weld rather than running down the centreline of a fresh bead.
Worked scenario: a fillet weld on a thick structural tee shows a crack near the toe inspected a day after welding. The plausible mistake is calling it a solidification crack and adjusting welding current, which would address a problem that forms during freezing. The better decision notes the delayed appearance and the heat-affected-zone location, which points toward hydrogen-induced cracking driven by moisture, hydrocarbons on the joint, insufficient preheat, or a consumable that was not low-hydrogen. The repair decision differs sharply: the hydrogen route demands dried consumables, clean surfaces, and preheat and interpass control, not merely a parameter tweak.
Build the habit of recording when a crack was first seen relative to welding completion, and where it sits relative to the weld and HAZ. Those two observations separate the crack families more reliably than crack appearance alone, and they drive the preheat, cleanliness, and consumable controls that prevention depends on.
Porosity and Slag Inclusions: Tracing Each Back to a Cause
Porosity is gas trapped as the pool freezes; slag inclusions are non-metallic residues left between runs or along edges. Their causes point to gas supply, cleanliness, and technique.
Porosity arises when dissolved or generated gases cannot escape before the pool solidifies, producing rounded gas pores: uniformly scattered, clustered, or an elongated wormhole. Candidates for cause include disturbed or leaking gas shielding, draughts blowing shielding away, damp or damaged coated electrodes, contaminated joint surfaces, or excessive arc length. Slag inclusions, meanwhile, are solid non-metallic material trapped in the weld, typically from incomplete slag removal between runs, poor interrun profile that traps slag in grooves, or too low a heat input for the slag to float out. Distinguishing the two matters because rounded gas shapes and angular slag debris respond to entirely different controls.
Worked scenario: radiography of a multipass butt reports both scattered rounded indications and a linear, jagged indication along the fusion line of an earlier run. The plausible mistake is treating every indication as porosity and investigating the gas system. The better decision classifies the linear jagged feature as a slag inclusion trapped where the previous run's profile left a groove, then links it to interrun cleaning and run placement rather than shielding. This matters because the corrective action is procedural: dress the offending run, remove slag completely, and place the next run so it wets out the groove, rather than changing gas equipment that was never at fault.
Practise reading cause chains backwards. Take an indication, name its morphology, name the mechanism that produces that morphology, and only then name the likely contributing condition. Reversing that order, from suspected cause straight to conclusion, is where classifications go wrong.
Undercut, Overlap, and Profile Faults: Parameter fingerprints
Profile defects come from how the arc interacts with the plate. Undercut is a groove melted at the toe and left unfilled; overlap is metal that spread but never fused.
Undercut appears as a groove melted into the parent metal at the weld toe or between runs and left unfilled, often linked with excessive current, excessive travel speed, or an incorrect electrode angle. Overlap is the opposite appearance: weld metal that has flowed over the surface without fusing to it, leaving a sharp notch, commonly associated with too low a travel speed or excessive metal deposition. Excess convexity, excessive penetration bead, and misalignment belong to the same profile family. Each profile fault is a geometry problem you can describe with measuring tools: a fillet gauge, a rule, and a straight edge, which is why visual inspection carries so much weight in welded fabrication.
Worked scenario: a fillet shows a sharp notch along one toe and the trainee's report says undercut caused by too much current. Checking the other toe shows weld metal sitting proud with an unfused edge there instead. The better decision reports undercut on one toe and overlap on the other, and attributes them jointly to travel technique and electrode angle rather than current alone, because a current that is too high for the plate cannot be simultaneously melting a groove on one side and failing to fuse on the other. The report then drives a technique review, which addresses both features at once, instead of a parameter change that fixes neither.
When you describe a profile defect in practice, pair every observation with the measurement that supports it: depth of the undercut groove against plate thickness, leg length against the specified fillet size, angular misalignment against the allowed tolerance. A report with numbers attached is a report a welding engineer can act on.
Before, During, and After: The Three Stages of Weld Inspection
Inspection divides into pre-weld checks of materials and preparation, in-process checks of parameters and interpass condition, and post-weld checks of the finished weld and documentation.
Pre-weld inspection covers everything that becomes expensive to fix once the arc strikes: correct base material identification, consumable type and condition, joint preparation dimensions against the drawing, cleanliness and freedom from moisture, fit-up, root gap and alignment, and the welding procedure specification available and understood. In-process inspection follows the first pass and each subsequent one: amperage and voltage against the WPS, preheat and interpass temperature, run sequence, interrun cleaning, slag removal, and the profile of each run as it is placed. A fault found at this stage costs a dressing or grinding pass; the same fault buried under four more runs costs excavation and rewelding.
Post-weld inspection includes the full visual examination of the completed weld: profile, toe condition, surface defects, root condition where accessible, dimensional checks such as leg length and alignment, and verification that required testing, heat treatment, or documentation has been carried out. It also covers identity and traceability: that the weld you are examining is the one the paperwork describes, made to the procedure on file. Treating the three stages as one connected record, rather than three separate chores, gives you the trail of evidence that supports a sound accept-or-refer decision on the finished joint.
A practical exercise: pick any welded assembly available to you, even a training coupon, and write a three-column checklist for it before you see the finished weld. Column one lists everything checkable before welding, column two during, column three after. Fill it in as the work proceeds and note which checks were skipped in practice. Reviewing which column went unchecked tells you where defects would have been caught earliest, and sharpens the staging instinct that inspection work depends on.
Choosing Between MT, PT, UT, and RT for a Suspected Flaw
Match the method to the flaw's depth and the material's properties. Surface methods find what opens to the surface; volumetric methods find what hides inside.
Magnetic particle testing locates surface and slightly subsurface discontinuities in ferromagnetic materials, so it suits toe cracks, surface porosity, and near-surface inclusions in steels, but it cannot be used on non-magnetic materials such as austenitic stainless steels or aluminium. Liquid penetrant testing finds discontinuities that open to the surface in any non-porous material, regardless of magnetism, but it sees nothing below the surface. Radiographic testing reveals volumetric internal flaws such as porosity and slag well, and can show certain planar flaws depending on orientation and access, while ultrasonic testing excels at internal planar flaws such as lack of fusion and cracks, though its interpretation depends heavily on operator skill and calibration.
The inspector's job is usually not to perform these methods but to understand what each can and cannot find, so that a crack hiding below the surface is not passed because only a surface method was used, and a volumetric flaw is not missed because only a planar-optimised method was chosen. That understanding also explains why codes and procedures often specify combinations of methods for critical joints. When you read a specification, ask which flaw types matter for that joint and whether the specified methods can actually detect them.
A scenario to reason through: a root crack in an austenitic stainless steel pipe butt, not visible externally. Magnetic particle testing is ruled out by the material, radiography needs access and shows orientation-dependent results for tight planar cracks, and ultrasonic testing can find the planar flaw but demands a qualified operator. The sound decision is to recommend penetrant testing for surface openings plus ultrasonic testing by qualified personnel for the internal extent, and to document that reasoning in the report.
| Method | Finds best | Key limitation |
|---|---|---|
| Magnetic particle (MT) | Surface and near-surface flaws in ferromagnetic steels | Cannot be used on non-magnetic materials |
| Liquid penetrant (PT) | Surface-opening flaws in any non-porous material | Blind to subsurface flaws |
| Radiography (RT) | Volumetric flaws such as porosity and slag | Access, safety controls, orientation limits for tight planar flaws |
| Ultrasonics (UT) | Internal planar flaws such as lack of fusion and cracks | Heavily dependent on operator skill and calibration |
A Practice Sequence and Self-Check Rubric You Can Run This Week
Work in three passes: definitions with drawings, scenario classification with timing and cause, then full report-style justifications. Score yourself against a fixed rubric each round.
Pass one, over roughly the first third of your preparation time: draw cross-sections and place every defect you have learned on them, and write one-sentence cause chains linking each defect to its mechanism and likely contributing conditions. Pass two: write short scenarios of your own, such as the crack-on-a-thick-tee case, and force yourself to state timing, location, classification, cause, and action for each. Pass three: take a complete joint, from drawing to finished weld in imagination, and produce the pre-weld, in-process, and post-weld checklists plus an NDT recommendation with justification. Repeat the cycle, shrinking the time you allow yourself.
A realistic adaptable sequence: one week on fusion and penetration defects with annotated drawings; one week on cracks with a timing table separating hot from hydrogen mechanisms; one week on gas and profile defects with cause chains; one week on the three inspection stages; one week on NDT selection with the decision table; then rotating mixed rounds of scenario classification for the remainder of your preparation. Adjust the weighting toward whichever defect family your rubric scores weakest, and repeat the whole cycle rather than moving on from any family you cannot yet place confidently on a drawing.
Use this self-check rubric per scenario, each item worth a point, five points per scenario: correct defect named; location justified against the joint cross-section; mechanism stated, not just a guess at a cause; timing stated where relevant, such as hot versus delayed; and a reasoned action that follows from the classification, not from the defect name alone. Treat eight of ten across two scenarios as a learning milestone to move on, not as a prediction of any exam outcome. Readiness checks before you finish: you can fill in an annotated cross-section from memory; you can separate the two crack families by timing and location without hesitation; you can write a three-stage inspection checklist for an unseen joint type; and you can justify an NDT recommendation for a material and flaw pairing you have not seen before.
- Rubric item 1: the defect is named correctly from the description alone
- Rubric item 2: the location is justified against a drawn cross-section, not asserted
- Rubric item 3: the mechanism is stated, so the cause follows logically
- Rubric item 4: timing is stated where it matters, such as hot versus delayed cracking
- Rubric item 5: the recommended action follows from the classification and its cause
References and further reading
Use these references to explore the concepts and check the latest information from the relevant organizations.