14.2 Evaluating Base-Material Product Forms

Official ASNT UT general topic 2 is Evaluation of Base Material Product Forms. The items are not asking you to recite a metallurgy textbook. They ask whether you know where a given process writes a discontinuity and which beam can see that plane or that void. Chapter 3 already sorted welds, castings, forgings, wrought product, and in-service damage as families. This section is the ultrasonic half of that map: 0° for laminations, the bar trio of inclusions, pipe, and seams, forging bursts, corrosion/erosion thickness mapping, a short introduction to cladding/bond checks, and the geometry that steals a backwall and fools a rushed Level II.

Topic 3 (weldments) is the next section. Topic 4 (bonded structures) is the next chapter. Do not drag a 70° weld scan into a plate-lamination question, and do not write a full adhesive-bond essay when the item is a roll-bond clad plate.

ASME Section V, Article 5 is the usual technique article for ultrasonic examination of materials (plate, forgings, castings, tubular products). Article 4 is welds. Acceptance still lives in the construction code or the material specification the traveler cites (an SA/A 20 plate UT supplement, an SA/A 388 forging practice, an owner's remaining-wall table), not in Section V itself. Attribute the article. Follow the procedure.

Product form is the first filter

Before the probe touches steel, read the traveler the way you would before a radiograph:

1. What was this shape before we machined it? Plate, strip, bar, billet, bloom, closed-die forging, open-die shaft, clad plate, or an in-service wall that used to be thicker?
2. Which direction did the metal move? Rolling, drawing, and forging elongate inclusions and open pipe and seams along that direction. Laminations lie in the rolling plane.
3. Which surface is the examination surface? Face of plate, end of bar, OD of a shaft, ID of a vessel?
4. What is physically impossible here? A weld root lack of fusion cannot exist in unwelded plate. A mid-plane lamination is the wrong story for a burst in a pancake forging. Cladding disbond cannot exist if there is no clad.

Ultrasound reports acoustic impedance change and geometry. A planar separation parallel to the entry surface returns a strong 0° echo and kills the backwall. A planar separation perpendicular to the entry surface can return almost nothing at 0° and plenty at 45° or 60°. Product form tells you which plane is likely. The beam choice follows the plane.

Plate laminations with 0°

A lamination in rolled plate is a planar separation parallel to the rolled surface. It is inherited from ingot pipe, inclusion sheets, or incomplete welding-up of voids during rolling. It sits in the rolling plane, often mid-thickness, sometimes stepped, sometimes opening to an edge as a seam.

The beam that sees that plane is straight-beam compression (0°) from the plate face:

• Sound material: a clean back reflection at the known thickness, often with a decaying multiple-echo train if the surfaces are parallel and the instrument range shows them.
• Lamination: one or more mid-wall echoes at a thickness less than T, plus loss or collapse of the true backwall (and of the multiple-echo train) when the lamination is large enough to intercept the beam.
• Small or partial lamination: a mid-wall blip that does not fully kill the backwall. Map it. The procedure's recording table (area, length, remaining ligament) decides whether it is a reportable condition.

Scan pattern on plate is usually a raster or grid on the accessible face, with index small enough that the beam overlaps. Some specifications require both faces, plate ends, or a band along edges that will become weld prep (because a lamination at a beveled edge becomes a weld defect later). Follow the stated coverage. "I waved the probe until I saw a backwall" is not a plate examination.

Why not angle beam for laminations? A 45°, 60°, or 70° shear wave is nearly parallel to the lamination plane. The reflection coefficient toward the probe is poor. You can walk an entire laminated plate with a weld probe and see nothing but grass. Conversely, do not call every mid-wall 0° echo a lamination until you have ruled out a thickness step, a backing bar, a welded doubler, or a second plate in a lap. Geometry first.

Two thickness-trap reminders:

• If the thickness gage locks on the lamination, remaining-wall reads thin. That is a mid-echo, not metal loss. Flip to a full A-scan. Look past the first reflector for the true backwall.
• A very thin lamination packet can ring and look like a cluster of inclusions. Still a rolling-plane story, still a 0° map.

Bar, billet, and rod: inclusions, pipe, and seams

Wrought bar and billet carry three discontinuity families the general exam expects you to separate by address.

Pipe

Pipe is a centerline cavity inherited from primary pipe in the ingot — shrinkage that rolling or forging did not weld shut. It lives on or near the centerline, elongated with the working direction. It is not a surface seam and not a weld-root void.

UT approach: 0° from the end of a cut bar (a disk-like view down the axis) or 0° from the OD aimed through the radius, looking for a persistent centerline reflector. Rotary or multi-probe bar testers do the OD version at production speed. A single contact probe can do the same physics on a sample length: peak at mid-radius depth, walk the length, see whether the echo stays on the axis.

Pipe that has been partly worked shut can leave a string of centerline reflectors rather than one cavity. Still pipe. Still centerline.

Seams

Seams are elongated, often surface-breaking or near-OD discontinuities running along the length. They come from rolled-in cracks, folds, or overfills. They are the bar cousin of a plate-edge lamination that reached the surface.

UT approach: a beam that intersects the seam from the OD — 0° with a small-diameter probe walking the circumference, a surface-wave or shallow-angle technique when the procedure allows it, or the rotating-probe / spinning-tube arrangement used in bar mills. A single 0° shot from the end can miss a shallow OD seam entirely. Address matters.

Seams are easy to confuse with handling scratches and with machined oil grooves. A relevant seam persists, has depth, and is not explained by the drawing. Visual and, where required, a surface method sit beside the UT call.

Inclusions and stringers

Nonmetallic inclusions elongate into stringers in the working direction. They may sit at mid-radius, in subsurface bands, or as scattered reflectors. Amplitude is often modest compared with an open cavity. They can be individual or clustered.

UT approach: 0° from the OD or from the end, plus a second direction if the procedure is a two-axis bar exam. Do not promote every stringer to pipe. Pipe is a cavity on the centerline. A stringer is a solid impedance mismatch that can live off-axis. The report should say which.

Forging bursts

A burst is an internal rupture from working the metal when it was too cold, too heavily reduced, or too dirty. Bursts are irregular, often clustered, and they sit where the forge strain was highest — commonly the core of a shaft, the web-to-flange junction of a die forging, or the center of a pancake. They do not form a single tidy mid-plane the way a plate lamination does.

Related but not identical:

• Flakes (hydrogen flakes) are thin, disc-like ruptures, often in a band, often roughly parallel to the local working plane. They can look more "laminar" than a classic burst.
• Laps are surface-folded metal. They are primarily a surface-method problem; UT may see a near-surface planar echo if the lap has depth.
• Unhealed pipe in a forged bar is still pipe by origin, even though the product form is now a forging.

UT approach: 0° through the accessible faces to find the cavity or the cluster, plus angle beam when the burst plane is not parallel to the entry surface or when the geometry (a stepped shaft, a bore) hides the 0° backwall. Map the cluster. Do not average it into one equivalent hole.

A burst versus a lamination, in one sentence: a lamination is a rolling-plane sheet in plate; a burst is a rupture in worked metal that does not owe you a single plane. If the traveler says "forging" and you write "lamination," you have the wrong product-form word even if both kill a backwall.

Heavy forgings often follow a practice such as SA/A 388 (technique family) with an acceptance table in the purchase order. The general exam will not ask you to recite a table you were not given. It will ask whether a central irregular cluster in a shaft is more burst than plate lamination, and whether 0° from the end plus 0° from the OD is a sensible pair.

Corrosion and erosion thickness mapping

In-service corrosion and erosion do not write a manufacturing plane. They remove wall. The UT job is remaining-thickness mapping, not "find the crack."

What has to be true before a number is a number:

• Velocity matches the remaining alloy (and not the coating, and not a clad layer you did not mean to include).
• Zero / delay is set for this probe on a known step or the instrument's documented zero block.
• You know whether you are reading first backwall, echo-to-echo (coating ignored), or a dual-element interface that is optimized for remaining wall and pitting.

Technique choices:

• Dual-element (TR) probes are the corrosion-mapping workhorse. The roof angle creates a pseudo-focus a short distance under the surface, which is kind to pitted ID and to rough OD. They are weaker at very thick, clean walls than a large single-element contact probe.
• Single-element contact works on smooth, parallel remnants (machined ID, new plate).
• Delay-line probes help on thin remaining wall so the interface ring-down does not swallow the backwall.
• Echo-to-echo (multiple-echo) through paint can ignore a uniform coating if both echoes are metal backwalls. A thick, uneven coating or a disbonded coating breaks that assumption.

Mapping is a grid the procedure can reproduce: mark-out, probe-center coordinates, and a record of the minimum in each cell. Isolated pit peaks need a smaller probe or a tighter grid; a large dual-element can bridge a pit and read optimistic. A lamination or a mid-wall inclusion can steal the gate and print a fake thin spot — watch the A-scan, not only the digital thickness.

Erosion (flow-accelerated thinning, elbow extrados, pump impingement) is the same measurement with a different map shape. The exam point is not the corrosion mechanism. It is that you are measuring remaining ligament with a calibrated 0° system, and that geometry (a counterbore, a weld root, a taper) can look like metal loss if the gate is lazy.

Cladding and bond check — introduction only

Clad plate and weld-overlaid vessels put a corrosion-resistant layer on a carbon-steel substrate. Topic 4 in the next chapter owns bonded structures. Topic 2 only needs the interface picture so you do not misread a clad plate as "thin carbon steel" or as a lamination.

At 0°:

• Bonded cladding with a modest impedance step: a small interface echo at the clad thickness, plus a backwall at the total thickness (clad + substrate). Energy went through the bond.
• Disbond: a strong interface echo and loss of the substrate backwall. The interface is now acting like a free surface.
• Wrong velocity: if you load carbon-steel velocity and gate the clad interface, you will print a "thickness" equal to the clad only and think the vessel has no wall.

Roll-bond, explosion-bond, and weld overlay do not have identical interface amplitudes even when the bond is good. The procedure's reference (a known bonded standard, a transfer from a sound region) matters. For this section, remember the binary: interface plus total backwall versus interface without substrate backwall. Leave adhesive honeycomb, rubber-to-metal, and immersion C-scan bond maps to the bonded-structures section.

Geometry that steals the backwall

Loss of back reflection is a symptom, not a diagnosis. Laminations, large inclusions, coarse grain, cladding disbond, a probe that is not flat, and geometry all kill a backwall. Topic 2 items will offer you a taper or a T-joint and dare you to write "lamination."

Common thieves:

• Tapers and tapers into a weld bevel. The back surface is no longer perpendicular. The 0° beam reflects away from the probe. Mid-wall is quiet. This is not a lamination unless a mid-wall echo appears and survives a scan that keeps the probe normal to the local face.
• T-joints and fillet-welded attachments. There is no backwall opposite the stem in the usual plate sense; the stem is a third member. A 0° probe on the flange opposite the stem sees a geometric reflection from the far fillet or sees nothing useful. Angle beam from the flange, or 0° from the stem end, is the examination — and that is already bordering on weld/attachment work.
• Counterbores, ID tapers, and weld-prep lands on pipe. A thickness probe just off the land reads the full wall; on the land it reads the counterbore. Both numbers can be true. Neither is corrosion until you know where you sat.
• Contours, nozzles, and saddles. The back surface walks out of the beam. Use a smaller probe, a shoe that matches the OD, or accept that this patch is a limitation.
• Coarse grain and austenitic welds in the base-metal scan path. Scatter raises grass and eats the backwall. That is attenuation, not a planar lamination. A second frequency or a different location on the same heat may confirm it.

The disciplined sequence when a backwall dies:

1. Re-couple. Dirt, a rock in the couplant, and a tilted probe are the most common "laminations" on earth.
2. Look for a mid-wall echo. No mid-wall, no lamination story.
3. Move. If the backwall returns as you leave a taper, a toe, or a counterbore, you have been measuring geometry.
4. Change direction or face if the procedure allows. A real mid-plane lamination is still there from the other face; a one-sided geometric miss is not.
5. Only then map area and remaining ligament against the material specification.

How topic 2 items are really written

Expect a product form, a probe angle, and a screen story:

• 0° on plate, mid-echo at 0.5 T, backwall gone over a 200 mm patch. Lamination (or a deliberately bonded doubler — check the drawing) , not a 70° toe crack.
• 0° from the end of a rolled bar, persistent reflector on the axis. Pipe.
• OD scan of bar, long reflector at 2 mm depth running the length. Seam, not pipe.
• Irregular cluster in the core of a shaft forging, no single plane. Burst (or flake band), not plate lamination.
• Grid of dual-element readings on an elbow, A-scan shows a clean backwall walking thinner toward the extrados. Erosion/corrosion mapping, not a burst.
• Clad vessel, strong echo at 3 mm, no echo at 25 mm. Bond-line problem or a gate locked on the clad — not a 22 mm lamination in monolithic plate.
• 0° on a tapered transition, backwall gone, no mid-echo. Geometry.

What topic 2 is testing

Evaluation of base-material product forms is address plus beam. Laminations live in the rolling plane and answer to 0°. Pipe lives on the centerline of bar. Seams live along the OD. Bursts rupture forged cores and do not owe you a single plane. Corrosion mapping is a calibrated thickness system on a grid. Cladding introduces the interface echo. Tapers and T-joints steal backwalls without writing a discontinuity. If you keep those six sentences, the product-form items on the general exam have nowhere to hide.