6.2 Types of Discontinuities Detected by MT

Types of Discontinuities is official MT general topic 8 on the outline administered prior to 15 December 2026. The exam is not asking you to recite a metallurgy textbook. It is asking which conditions produce a usable leakage field at a ferromagnetic surface, and which conditions magnetic particle testing will never see no matter how good the yoke is.

A discontinuity is an interruption in the typical structure of the material. Magnetic particle testing (MT) can reveal it only when three things are true at once:

1. The material is ferromagnetic enough to carry flux.
2. The discontinuity is at or near the examination surface.
3. The magnetic field has a component across the discontinuity, so flux leaks and particles collect.

Break any one of those rules and the method is the wrong tool. ASTM E1444, ASTM E709, and ASME Section V, Article 7 all assume those limits. They do not turn MT into radiography.

Surface versus near-surface

Surface-breaking openings — fatigue cracks at weld toes, grinding cracks on a journal, seams that reach the bar surface — leak strongly when the field crosses them. Near-surface conditions — a shallow underbead crack, a forging burst that almost breaks out, a stringer just under a light machine cut — can leak if the ligament is thin and the current type drives flux deep enough. Direct current, half-wave direct current, and prod or central-conductor circular fields reach deeper than alternating current. They still do not inspect mid-wall.

If the stem places a shrinkage cavity in the center of a 4-inch (100 mm) casting wall, MT is the wrong method. If it places a quench crack on a hardened spline, MT is the right method.

Crack families the general exam expects you to name

Fatigue cracks

Fatigue cracks start at a stress raiser — a weld toe, a thread root, a fillet radius, a corrosion pit — and grow under cyclic load. They are almost always surface-initiated, tight, and roughly perpendicular to the principal stress. An AC yoke with dry or wet particles is the classic in-service tool. The particle pattern is usually sharp and linear. Fatigue is the discontinuity that keeps MT on in-service structural and pressure equipment.

Grinding cracks

Grinding cracks form when abusive grinding overheats a hardenable surface. They are fine, often shallow, and may appear as a network or as lines perpendicular to the grinding direction. Ground weld caps, repaired journals, and hard-faced seats are typical homes. Because they are tight, they favor finer wet fluorescent particles on a smooth surface. Heavy dry powder on a still-rough grind can skip them.

Heat-treat (quench) cracks

Heat-treat cracks, often called quench cracks, open during rapid cooling of a hardenable steel. They prefer section changes, hole edges, and other stress concentrations. They are usually surface-breaking, can be deep, and may be filled with oxide from the heat-treat atmosphere. A filled crack is a detection problem: clean first, then magnetize. Do not assume a black oxide line is already a rejectable MT indication until the procedure's evaluation rules are applied.

Crater cracks

Crater cracks (star cracks) form in the weld crater when the arc is broken and the last freeze-shrinks. They are surface-breaking, often star-shaped, and sit at weld stops and tack ends. MT and visual testing both find them when the crater is accessible. They are a workmanship discontinuity, not an in-service growth mechanism, but they are stress raisers if left in a fatigue-critical weld.

Heat-affected-zone and toe cracks

Heat-affected-zone (HAZ) cracks and toe cracks sit at the fusion-line / toe region. Hydrogen cracks (cold cracks) in carbon and low-alloy steel welds often break the surface in the HAZ after the weld cools. In-service toe cracks are frequently fatigue. Both are prime MT targets on ferromagnetic welds. Orientation is usually along the weld (longitudinal) or slightly off-axis; a field across the weld finds them. A field along the weld can miss them.

Processing defects: laps, seams, bursts, and stringers

These four words are product-form language. Learn them as MT targets, not as synonyms for "crack."

A seam on bar stock is why shops use a circular field (head shot or central conductor) along the length: the seam runs longitudinally, so a circular field is perpendicular to it. A lap on a forging often follows the die line or the flow. A burst that never approaches the surface is an ultrasonic or fracture problem, not an MT problem.

Product-form mapping

Map the discontinuity to the thing you are holding. The specific exam writes stems that way.

Welds. Look for crater cracks at stops, HAZ and toe cracks, underbead cracks that approach the face (DC or HWDC may help), and surface-breaking incomplete fusion or overlap. Geometric collectors such as undercut and the toe radius are not automatically defects; they are evaluation problems (next section). Deep slag pockets and mid-wall lack of fusion are not MT work.

Castings. Look for hot tears and shrinkage cracks that reach the surface, and for cold shuts that are open. A central shrinkage cavity, gas porosity deep in a boss, and many nonmetallic inclusions never leak at the surface. Prods plus DC on a rough ferromagnetic casting are a classic technique pairing when the procedure accepts the arc-burn risk.

Forgings. Laps and bursts dominate. Examine the flash line, the parting line, and section changes. Grain flow can create nonrelevant particle collection; that is classification, not a reason to skip the exam.

Wrought bar, rod, and pipe. Seams and stringers dominate. Circular magnetization finds longitudinal seams. End faces and cut edges can show stringers that did not appear on the OD.

In-service components. Fatigue, corrosion fatigue, and grinding cracks from maintenance are the usual MT finds. Stress-corrosion cracking is method-relevant only in ferromagnetic alloys and only when it breaks the surface. Austenitic stainless chloride cracking is a penetrant or eddy current problem, not an MT problem.

What magnetic particle testing misses

Memorize the miss list. Topic 8 items are often written as "which condition will MT not detect?"

• Deep internal voids. Mid-wall porosity, shrinkage, and many bursts never disturb surface flux.
• Nonmagnetic materials. Austenitic stainless steels (304, 316), aluminum, copper alloys, titanium, and most nickel alloys will not support the flux a leakage-field test needs. A 300-series weld on carbon steel is not an MT weld. Some 400-series stainless and duplex structures can be mixed; the procedure, not hope, decides.
• Tight cracks filled with contaminant. Paint, scale, grease, peening smear, and oxide can choke the opening so little flux leaks and particles cannot enter. Surface preparation is part of detectability.
• Discontinuities parallel to the flux. A seam parallel to a longitudinal field, or a toe crack parallel to an AC-yoke field that was placed along the weld, produces little leakage. That is why procedures require two field directions about 90° apart, or a multidirectional technique.
• Laminations parallel to the examination surface. A mid-plane lamination in plate does not leak to the rolled face. Straight-beam ultrasound is the usual tool.
• Subsurface planar lack of fusion that never approaches the face. Choose ultrasonic testing.

Orientation and contaminant fill — why a known crack can disappear

A heat-treat crack that is real can still produce no particles. Two Level II explanations cover most stems:

1. The first field ran parallel to the crack. Reinspect with a field roughly perpendicular to the suspected length.
2. The opening is packed with scale, oil, or smear. Clean to the procedure, then reinspect. Residual magnetism, magnetic writing, and over-application of dry powder can also hide or fake a pattern; those are evaluation and process-control issues, but they start from the same physical fact: particles report leakage, not the metallurgist's name for the discontinuity.

Realistic exam scenarios

A carbon-steel crane-girder toe shows a sharp linear buildup after an AC-yoke shot across the weld. That is the textbook fatigue or HAZ/toe crack indication — a relevant surface condition until the acceptance standard says otherwise.

A quenched-and-tempered shaft shows a network of fine lines on a ground journal after wet fluorescent MT. That pattern is consistent with grinding cracks, not with a single fatigue origin.

A rolled bar is examined with a longitudinal coil shot only. A long seam can hide because it runs parallel to the coil field. The missing examination is a circular head shot or central conductor.

A Type 316L stainless nozzle weld is offered for yoke MT. There is no valid ferromagnetic path. The discontinuity type does not matter; the material already took MT off the table.

Topic 8 is a capability map. Name the discontinuity, name the product form, then say whether a leakage field can exist.