12.1 Manufacturing Processes and Associated Discontinuities

The ASNT NDT Level II radiographic general exam lists Manufacturing Processes and Associated Discontinuities as official RT topic 4. Topic 3 asked you to tell an indication from a discontinuity from a defect. Topic 4 asks a harder question: given how this part was made, what can that dark mark even be? A Level II who names every blob on film without reading the traveler is guessing. Process history is not background color. It is the filter that makes interpretation legal.

Radiography reports thickness-density change along the beam. A cavity, a gas pore, or a slag pocket that removes metal (or replaces steel with something less absorbing) lets more radiation through and records as a darker region on film or a higher-signal region on a digital detector. A high-density foreign body — tungsten from a GTAW electrode, a copper chill, a lead marker — absorbs more and records lighter. A tight planar separation that presents almost no thickness change, or that sits parallel to the detector, may produce no usable image even when the radiograph meets density and image-quality-indicator (IQI) rules. That physics is why topic 4 is paired with product form, not with a photo album of every possible mark.

Process history before the viewer is switched on

Read the traveler, the drawing, the weld map, and the material specification before you argue about a shape on the image. Ask four questions:

1. What product form is this region? Cast, wrought/forged, welded, or only machined from sound stock?
2. What process wrote this volume? Sand casting, investment casting, SMAW, SAW, GTAW repair, closed-die forge, rolled plate, or a finish machine pass?
3. Where is the last metal to freeze, the fusion face, the working direction, or the machine surface? Discontinuities cluster at those addresses.
4. What is physically impossible here? Slag cannot form in an unwelded pour. A forging lap cannot form in a sand mold. Shrinkage cavities do not nucleate in a finish-milled wrought block that was never recast.

Chapter 3 sorted discontinuities into inherent, processing, and service origin classes. Topic 4 uses that map and then asks how each class looks on a radiograph. If the stem gives a steel sand casting and an irregular dendritic cavity in the thermal center, the answer is shrinkage, not a weld slag line and not a forging burst.

How a discontinuity becomes a radiographic image

Three conditions have to be true at once or the film stays quiet:

• Enough local thickness or density change. A spherical pore a millimetre across in 12 mm of steel is a classic dark spot. A hairline crack 0.01 mm open along a plane the beam only grazes may be invisible.
• Alignment. The beam must travel along a planar gap (or at least not perpendicular to a vanishingly thin plane) for lack of fusion or a crack to darken the image. Sidewall lack of fusion on a steep bevel is a textbook miss when the shot is a single normal-to-film exposure.
• Contrast the technique can record. Energy that is too high flattens subject contrast. Scatter, motion, and geometric unsharpness smear a fine linear indication into the fog.

So interpretation is not “dark equals reject.” Interpretation is “this process can make this shape here, and this shot geometry can show it.”

Castings: solidification writes the list

A casting starts as liquid and ends as a shape. Gates, risers, chills, cores, and the last region to freeze decide where metal is missing or where foreign matter is trapped. Radiography is the classic volumetric map of that freeze.

Shrinkage

Shrinkage is a cavity left when liquid metal contracts and a riser does not feed the thermal center. On the radiograph it is irregular or dendritic, often darker toward the middle of a heavy section, a junction, or a boss. Filamentary (sponge) shrinkage looks like a dark network; a more open cavity has ragged walls rather than a smooth bubble. Location is the giveaway: last-to-freeze metal, not a random spray of spheres. Do not call a riser-neck shrink a weld defect, and do not call it gas because both are dark.

Gas porosity

Gas is trapped mold gas, dissolved gas coming out of solution, or steam from a wet mold or a damp core. Voids are spherical or slightly elongated, with smoother walls than shrinkage. They may cluster under the cope, along a core, or in a turbulent pour path. Isolated pores are rounded dark spots. Wormhole or piping gas can look elongated along the gas-escape path. Surface-breaking pores are also visual or penetrant calls; RT still owns the internal population.

Inclusions

Inclusions are sand, slag, dross, or oxides folded into the pour. They are irregular, often clustered near the cope, a gate, or a turbulent corner. Low-density sand or slag is darker than the parent metal. A high-density foreign body — a chaplet that did not fuse, a piece of tramp metal — can be lighter. Shape plus pedigree separates an inclusion from a gas pore: pores are round; sand is jagged; a chaplet has a manufactured outline.

Hot tears and cold shuts

Hot tears form while the casting is still hot and weak, usually at a section change or a hot spot where contraction is restrained. On film they are ragged linear dark indications, often at fillets and junctions, not a neat centerline. Surface-breaking tears are also visual, penetrant, or magnetic-particle targets; internal tears are RT or ultrasonic work.

Cold shuts form when two metal streams meet and do not fuse, often with an oxidized film. Many reach the surface. An internal cold shut behaves like a planar bond line: it may be faint or invisible on RT unless the plane is open and aligned. Do not promote every faint line in a casting to a hot tear if the location is a stream-meet, not a restrained hot spot.

Castings also carry misrun, core shift, and unfused chaplets. Many of those are visual or dimensional before a film is shot. A stem that says steel sand casting plus irregular dendritic cavity in the last region to freeze is asking for shrinkage.

Welds: a second solidification plus a fusion face

A weld adds weld metal, a fusion line, and a heat-affected zone (HAZ). The names the construction codes use are the names the specific exam will print. Learn them as geometry, not as slang.

Lack of fusion

Lack of fusion (LOF) is a planar failure of weld metal to fuse to a bevel, a land, or a previous pass. It sits on a fusion face. Radiography finds it only when the plane is open enough and aligned with the beam. A single wall shot normal to the plate often misses sidewall LOF on a steep groove. When it does show, it is a straight or gently curved dark line along the bevel or interpass boundary, not a cluster of round pores. Angle-beam ultrasonic testing is the stronger primary method for tight LOF; RT is not a promise that fusion was complete just because IQI wires are visible.

Incomplete penetration

Incomplete penetration (lack of penetration) is failure to fill or fuse through the required root or throat. On a single-wall radiograph with the beam along the root land it is often a straight dark line on the weld centerline at the root, sometimes with a sharp land shadow. A visible root can also be a visual call. Do not confuse a permitted root gap image, a mismatch shelf, or a high-low density step with incomplete penetration — those are geometry. The procedure and the referenced code tell you which root images are discontinuities.

Slag

Slag inclusions are trapped flux. They are irregular or elongated, often following a pass boundary or the weld contour, and they are darker than steel because slag is less absorbing. They image well on RT, which is why slag is a classic film call. A continuous slag line along a bevel can mimic LOF; location (inside a pass versus exactly on a fusion face) and raggedness help, but some images stay ambiguous and the report should say what you can defend.

Porosity

Weld porosity is trapped gas in weld metal — scattered, clustered, or piping. RT is the reference map: rounded dark spots in the weld envelope. Piping porosity is elongated. Clustered porosity sits in one pocket of dirty or moisture-laden weld metal. Surface pores are visual or penetrant calls. Do not call shrinkage dendrites in a casting “weld porosity,” and do not call a chain of aligned pores a crack unless a linear opening is actually there.

Cracks

Weld cracks are planar. Hot (solidification) cracks prefer the centerline, a crater, or a fusion-line tear while the metal is still hot. Cold cracks, including many hydrogen-assisted and HAZ toe cracks, form after cooling. On a radiograph a crack is a fine, often jagged linear dark indication. Many cracks are RT-quiet when tight or poorly aligned. Crater cracks may be star-shaped at a stop. Surface-breaking cracks are also magnetic-particle or penetrant work; do not claim a clean film proves a toe is uncracked.

Other weld images you must not misname

Undercut is a groove melted into base metal at a toe or root. It is primarily visual. On film it can appear as a dark line at the toe. Tungsten inclusions from GTAW are bright (high density). Burn-through and excess penetration change the root envelope. Icicles and root convexity are geometric density changes. None of those is slag, and none is a forging lap.

Forgings and wrought product: RT is not the default

Forging, rolling, and extrusion work solid metal. Discontinuities elongate in the working direction and often lie parallel to the surface. That orientation is why radiography is rarely the first method on a forging or on plate.

Laps are oxidized folds at the surface. Seams run with bar length. Laminations are mid-plane separations in plate. All three present little thickness change to a beam shot through the thickness, so the film can look clean while the part is badly laminated. Magnetic particle or penetrant testing owns surface laps and seams. Straight-beam ultrasonic testing owns laminations.

Bursts are ruptures from working that is too severe, too cold, or performed on a dirty or piped ingot. An internal burst can be irregular and crack-like. RT may see a burst when there is a real cavity or an open rupture and the geometry gives a through-thickness path — a bored forging, a machined window, a relatively simple section. A massive closed-die forging with a burst buried in a plane parallel to the die face is still an ultrasonic problem. If the stem says “forging” and “first method,” do not reach for RT unless the stem also gives access and a volumetric opening.

Pipe leftover from an ingot that was not cropped, then flattened by rolling, is a lamination, not a casting shrink in the finished plate. Origin may be inherent; morphology is wrought.

Machined parts: cutting does not make RT targets

Finish machining removes metal. It does not pour a mold and it does not deposit weld metal. Grinding cracks, abusive-machine tears, and smeared laps are surface conditions. Liquid penetrant or magnetic particle testing (on ferromagnetic stock) is the method family. Residual stress and a bad surface finish are not radiographic cavities.

RT becomes reasonable on a “machined part” only when the pedigree still contains a cast volume or a weld — a machined valve body that was a casting, a finish-machined weld neck, a repair weld under a blended contour. The machine stamps do not create shrinkage. If the drawing says wrought bar, no weld, no cast feature, a request for film as the primary exam is the wrong method, not a thorough Level II.

Putting the map to work on the exam

Work every topic-4 stem in this order:

1. Name the process at the location of interest.
2. Strike impossible families. No slag without flux or a weld. No hot tear in cold-worked bar that was never recast at that spot. No incomplete penetration without a joint that required penetration.
3. Match shape and address on the image to the remaining family: dendritic and last-to-freeze → shrinkage; round and in weld metal → porosity; linear on a bevel → suspect LOF; bright speck in a GTAW weld → tungsten.
4. Ask whether this shot could show that family. A single normal shot of a U-groove sidewall is a weak LOF test. A through-thickness shot of plate is a weak lamination test.
5. Do not over-call geometry. Root mismatch, a chaplet outline, a permitted land, and a density step at a counterbore are not automatically defects.

A production radiograph of a SAW long seam that shows irregular elongated dark streaks between passes is slag until proven otherwise. The same shape in an unwelded steel casting is not slag. A faint straight line exactly on a steep fusion face, with excellent IQI sensitivity and no open gap, may be lack of fusion that this shot cannot prove — which is a method-limitation answer, not a “film is clean so fusion is complete” answer.

Topic 4 rewards slow language: process, then possible names, then whether the beam can see them. Speed that skips the traveler is how a shrinkage cavity becomes a “weld crack” on a report, and how a forging lap is sent to the darkroom instead of the magnetic-particle bench.