11.3 Radiographic Indications, Discontinuities, and Defects

ASNT's radiographic general exam treats Indication, Discontinuities and Defects as official RT topic 3. The items are not asking you to memorize every weld-fault nickname. They are asking whether you can read a negative (or a digital gray-scale that behaves like one), tell less absorber from more absorber, name the common weld and product-form pictures, and refuse to call every dark smudge a defect.

ASME Section V, Article 2, ASTM E94, and ASTM E1742 tell you how to produce a readable image. They do not, by themselves, decide accept or reject. The construction code or specification named in the procedure — ASME Section VIII, ASME B31.3, AWS D1.1, API 1104, an aerospace table, a customer specification — supplies the acceptance sentences. Those documents are industry and code references, not ASNT exam publications. Attribute every length or porosity-cluster rule to the named standard. The procedure wins if it is tighter.

Chapter 12 takes manufacturing processes and in-service damage in more depth. This section is the viewing vocabulary: dark versus light, the usual weld indications, and the three-word ladder indication / discontinuity / defect.

Dark versus light on a negative

Industrial radiographs are viewed as negatives. Radiation that passes through the part exposes the detector.

• More radiation through a location → more silver (film) or a higher exposure signal (digital) → darker on the negative.
• Less radiation through a location → less exposure → lighter (whiter) on the negative.

Translate that into absorbers:

A one-line memory aid the written exam rewards: darker = less metal (or less-dense metal) in the path; lighter = extra metal or high-density metal in the path.

Two traps sit next to that sentence.

First, scatter, undercut of the beam at a sharp edge, and a density gradient from changing thickness can paint a dark halo that is not a cavity. Read the part geometry before you invent a pore.

Second, digital windowing can invert a display or stretch a tiny gray difference into a theatrical black. Interpret from a calibrated window the procedure allows, not from a screenshot that someone "enhanced." Film viewed on a dim illuminator has the opposite problem: a legal 3.5 density weld looks black and every light inclusion pops; that is a viewing error, not a sudden epidemic of tungsten.

Indication, discontinuity, defect

These three words are not synonyms. Topic 3 punishes anyone who treats them as such.

• An indication is what you see on the radiograph — a dark spot, a light speck, a line, a density step. Artifacts from 11.2 are indications too until you classify them out.
• A discontinuity is an interruption in the typical structure of the material or weld: porosity, slag, lack of fusion, a crack, incomplete penetration, a tungsten inclusion, undercut. Geometric conditions the procedure treats as weld profile (reinforcement, permitted concavity) may be indications without being discontinuities the acceptance table cares about — or they may be listed as rejectable profile, depending on the document.
• A defect is a discontinuity that failed the acceptance standard named in the procedure. A 1 mm isolated pore can be a discontinuity and still be acceptable. An 8 mm crack-like line in the same weld can be a defect. The table decides.

Classification is a process:

1. Confirm the image is valid — density or digital QC, essential IQI, no processing artifact (11.1–11.2).
2. Decide whether the mark is in the part. Reshoot or change cassettes if 11.2 says you must.
3. Describe type, location, length or diameter, and orientation (weld metal, HAZ, toe, root, parent; longitudinal, transverse, clustered).
4. Compare that description with the named acceptance paragraph.
5. Only then use the word defect — or acceptable discontinuity.

A Level II who writes "defect" on every dark indication has not evaluated the radiograph. A Level II who writes "porosity, OK" without a length, a cluster rule, or a document name has not evaluated it either.

Common weld indications the general exam expects you to name

Product form predicts the family. This section uses welds as the teaching set because that is where most general-exam stems live. Castings, forgings, and in-service damage return in Chapter 12 with the same dark/light physics.

Porosity and other gas cavities

Gas porosity is a rounded cavity left by gas that did not escape the weld pool. On a negative it is a rounded dark spot with a relatively sharp boundary. Isolated pores, clustered pores, and linear or piping porosity (elongated, sometimes tapered wormholes) are the usual pictures. Spherical gas is less dense than steel, so it is a textbook less-absorber / darker indication.

Rounded is not automatically acceptable. Acceptance tables limit diameter, spacing, and accumulated length in a weld increment. A chain of pores can become an aligned condition the standard adds together. Do not call a linear slag line "porosity" because the first glance was a blob, and do not call a tight crack "elongated porosity" because you would rather use the rounded table.

Slag

Slag inclusions are trapped nonmetallic welding flux or oxides. They are usually less dense than the parent steel or nickel alloy, so they print dark, but the outline is irregular, often elongated along a pass, a toe, or an interpass boundary — not a neat sphere. Slag can follow a wagon-track pair of lines in a poorly cleaned SAW or SMAW joint.

Slag is a discontinuity the moment it is real. Whether it is a defect depends on length, location (root versus cap), and the table. A thin, irregular dark line at an interpass boundary may be slag or may be lack of fusion; the difference is often sharpness, straightness, and whether grinding or a second angle confirms a planar unfused land. When the stem does not give you a second angle, describe what you see and apply the stricter linear rule if the procedure treats unresolved linear indications that way.

Lack of fusion and incomplete penetration

Lack of fusion (LOF) is a planar failure to fuse to a sidewall, to a previous pass, or to the land. It prints as a linear dark indication, often straighter than slag, sitting at a fusion boundary. Sidewall LOF is offset from the weld centerline; interpass LOF sits between beads.

Incomplete penetration (IP) is a root condition: the joint did not fuse through. It is typically a dark line on or near the weld centerline at the root, sometimes with a blunt or slightly gapped look.

Both are less absorber along a plane. Both can be very tight. A closed LOF land that is nearly perpendicular to the beam may show only a faint density change or may not show at all — the same orientation problem as a tight crack. A second angle, a tighter energy, or a complementary method (UT) is a technique answer, not a reason to invent a white LOF on the film.

Cracks — including the ones you will not see

Cracks are the discontinuity everyone expects radiography to catch and the one it is worst at when the opening is tight.

A crack that presents a measurable gap in the thickness direction — a crater crack opened at the surface, a wide hot tear, a crack whose faces are slightly separated and aligned with the beam — prints as a fine linear dark indication, sometimes jagged, sometimes straight. Location predicts the story: crater cracks at stops, longitudinal centerline cracks, transverse cracks, HAZ or toe cracks, underbead hydrogen cracks.

A tight crack whose opening does not change the transmitted intensity enough — faces closed, or the plane nearly parallel to the film so the beam sees almost the full metal thickness — can be radiographically invisible. Seeing a crisp 2-2T hole does not refute that sentence. The IQI is a drilled void. A fatigue crack can be a fraction of a millimetre open and still be structurally important. This is why codes restrict RT's role, why some joints require UT or surface methods for crack-sensitive details, and why a Level II must not write "no cracks" when the honest report is "no crack-like indications on this exposure."

If the stem gives you a tight toe crack found later by PT or MT and a clean RT film with a valid IQI, the correct teaching is orientation and tightness, not "the film was light-struck so the crack vanished."

Burn-through, icicles, undercut, concavity, extra metal

These are geometry conditions. They still follow dark/light physics.

Burn-through and icicles confuse candidates because both extra metal (light) and a hole (dark) can appear in the same centimetre. Read the pair: a white carrot on the ID is extra metal; a black hole next to it is missing metal. Neither is a processing water spot if it is locked to the root on a reshoot.

Tungsten inclusions (GTAW) are the classic light metallic indication: sharp, bright, often rounded or chunky white specks where electrode metal broke into the puddle. Copper from a backing bar or a melted fixture can similarly print light. Do not call a tungsten speck porosity. Porosity is dark.

From description to disposition

Once the mark is a discontinuity, measure what the procedure says to measure. Typical industry tables speak in length of a linear indication, diameter of a rounded indication, spacing, accumulated length in a weld increment, and sometimes proximity to a free edge or to another indication. Some documents reject any crack or any incomplete fusion; others allow limited slag or porosity. Do not invent a shop number. If the stem prints a limit, use that limit. If it names a document and does not print a number, the exam is testing the vocabulary (crack versus pore, dark versus light), not your memory of UW-51's latest sentence.

Worked logic the specific exam likes:

• Rounded dark spots, isolated, each 1.5 mm, table limit 3 mm, spacing met → acceptable discontinuities, not defects.
• Linear dark indication 12 mm at the sidewall, document says no lack of fusion → defect (LOF).
• Sharp white 2 mm speck in a GTAW root, table rejects tungsten → defect (tungsten inclusion), even though it is light.
• Fine linear dark at a toe that vanished on a second cassette and matched a crescent → not a discontinuity; go back to 11.2.
• Tight crack later found by PT, RT film valid, beam nearly parallel to the crack face → limitation of the method, not proof the PT is false.

Viewing practice that belongs with evaluation (Chapter 12.3 expands it): use an illuminator that can push light through the legal density, mask glare around the film, let your eyes adapt, hold a densitometer, and do not interpret from a phone photo of a light table. Digital evaluation uses the procedure's window, zoom, and measurement tools — not a random gamma slider.

Realistic exam scenarios

A cluster of rounded dark spots sits in the weld metal. The IQI is legal, density is 2.4, a reshoot shows the same cluster. Classification: porosity (discontinuities). Disposition: apply the rounded / clustered paragraph in the named standard. Do not call the cluster tungsten and do not call it a defect until the table says so.

A sharp white speck sits in a GTAW root pass and is locked to the weld on two shots. Classification: tungsten inclusion (high-density extra absorber). Light does not mean "ignore."

A dark line follows a weld toe, the adjacent cap is full, and a profile snapshot in the stem shows a groove in the parent metal. Classification: undercut, a thinning indication. Compare with the undercut depth/length rule, not with the crack table, unless a crack-like extension leaves the groove.

A light tear-drop hangs from the root ID on a 2G pipe weld. Classification: icicle (extra metal). A dark valley along that same root without extra metal is concavity. Mixing those two words is a standard stem trick.

A procedure requires 2-2T. The 2T hole is crisp. A later surface exam finds a tight toe crack that never appeared on the film. The RT was not "wrong" and the crack is not an artifact. Tight cracks may be invisible. Report what RT can support; do not claim a crack-free volume the physics did not inspect.

Topic 3 language is short: dark or light, what absorber that implies, the discontinuity name, then the table. Anything else is storytelling.