MPI Test for Welding: How Magnetic Particle Inspection Works

The weld passed visual inspection, the fit-up was clean, and the fabricator was ready to ship. Then the inspector ran an electromagnetic yoke along the toe of the cap, dusted on a fine cloud of particles, and a jagged dark line appeared out of nowhere. That moment, when an invisible toe crack suddenly becomes impossible to ignore, is exactly why the MPI test exists.

Why a Weld That Looks Perfect Can Still Fail an MPI Test

Picture a fabrication shop finishing a run of structural brackets. The cap profiles are uniform, the toes are clean, and visual inspection signs off without a comment. Weeks later, one bracket fails in service and the investigation finds a tight crack running along the weld toe. It was there the day the part shipped. It was simply too narrow, too tight, and too well disguised by the cap profile to see with the naked eye.

That gap between “looks good” and “is good” is exactly why magnetic particle inspection exists.

What an MPI Test for Welding Actually Is

An MPI test for welding, also called magnetic particle testing (MT), magnetic particle examination (MPE), or magnetic particle inspection (MPI), is a nondestructive examination method that magnetizes a ferromagnetic part and applies iron particles to its surface. As ASNT explains, the particles are attracted to places where magnetic flux leaks out of the part, and those clusters of particles make discontinuities visible. In welding, that usually means cracks, lack of fusion that reaches the surface, and undercut.

The catch hides in the word ferromagnetic. MPI works on carbon steel, low-alloy steel, and the ferritic or martensitic stainless steels. It does not work on austenitic stainless steel, aluminum, copper alloys, or any other nonmagnetic material. Aluminum in particular needs entirely different processes and inspection logic, which our aluminum MIG welding guide covers in detail. So before anyone wheels a yoke onto a joint, the first question is not “is the weld good?” but “is this material magnetic?”

Why Surface Flaws Matter More Than Most Welders Expect

A surface-breaking defect is often the most damaging kind in a fatigue-loaded structure. It acts as a stress concentrator from day one, and every load cycle works on it. A toe crack only a few thousandths of an inch wide can grow into a through-wall failure while the weld still looks respectable from arm’s length.

Visual inspection depends on contrast, access, eyesight, and lighting. A tight crack can close up under residual stress and effectively disappear. MPI does not rely on the crack being open enough to see. It relies on the crack disturbing a magnetic field, which is a far more sensitive test for this specific family of flaws.

There is an economic argument too. MPI is fast, relatively inexpensive, and gives immediate results, which is why it remains one of the most common nondestructive testing methods in fabrication. You are not waiting days for film to be developed or a report to come back from a lab.

Who This Guide Is For

This article is written for welders who want to know what happens after they lay a bead, fabricators pricing and scheduling inspection, QC staff building checklists, inspectors refreshing technique, engineers writing or reviewing acceptance requirements, and hobbyists who have just bought their first yoke. No prior NDT background is assumed. The goal is practical understanding: what the method does, where it stops working, and how to avoid the mistakes that cause missed defects.

The Vocabulary You’ll Meet on the Floor

• MT, MPT, MPE, MPI — the same method under different abbreviations. MT is the standard shorthand; MPI is the term most welders hear.
• Yoke — a portable electromagnet that straddles the weld and induces a magnetic field between its poles.
• Flux leakage — the magnetic field that spills out of the surface where a discontinuity interrupts the flux path.
• Discontinuity, defect, and indication — a discontinuity is any interruption in uniformity. It becomes a defect only when it exceeds the applicable acceptance criteria. An indication is the visible particle buildup itself, which still has to be evaluated.
• Acceptance criteria — the written limits in the contract or design code that decide whether an indication is acceptable, repairable, or rejectable.

What’s Ahead

The rest of this guide walks through the physics of flux leakage, the step-by-step inspection sequence, magnetization methods and field direction, wet versus dry and visible versus fluorescent particles, the codes that govern acceptance, what MPI can and cannot find, how it compares with penetrant, ultrasonic, and radiographic testing, the mistakes that cause false calls, and a practical checklist for your next weld inspection. The ASME and AWS document sets that shape most of this work get their own section, because knowing the method is only half the job. Knowing what the contract actually requires is the other half.

By the end, you should be able to look at a weld and judge whether MPI is the right tool, what it will realistically reveal, and what a trustworthy inspection actually involves.

What Is Magnetic Particle Inspection? The Physics Behind Flux Leakage

Magnetic particle inspection (MPI), also called magnetic particle testing (MT) or magnetic particle examination (MPE), is a nondestructive method for finding surface and near-surface flaws in ferromagnetic materials. It does not rely on sharp eyes or guesswork. It uses magnetism as the flaw detector, then makes the result visible with iron particles. The American Society for Nondestructive Testing classifies it among the core surface examination methods for good reason: when the setup is correct, a crack that is nearly invisible can produce a clear, repeatable indication.

Ferromagnetism, Domains, and the Invisible Field

Steel and many alloy steel welding electrodes produce ferromagnetic weld deposits, meaning their atomic structure can be magnetized. Inside the metal are tiny magnetic regions called domains. In unmagnetized steel, those domains point in different directions and largely cancel each other out. When an inspector magnetizes the part, the domains line up and the weld becomes a temporary magnet.

The magnetic field travels through the steel along paths called flux lines. Picture those lines as traffic moving through the material. As long as the path is continuous and the steel is reasonably uniform, the flux stays mostly inside the part. It exits and re-enters at the poles of the magnet, which is why a yoke placed on a plate creates a field through the steel between its legs.

The key number behind all of this is permeability, or how easily a material carries magnetic flux. Steel is thousands of times more permeable than air. That contrast is the entire trick of the method. As long as the field is happy inside the steel, nothing shows on the surface. Give it a reason to escape, and the escape becomes visible.

Why a Crack Creates a Leakage Field

Now introduce a crack. If the discontinuity cuts across the flux path, the magnetic lines are forced to detour. Air cannot carry the same amount of flux as steel, so some of the field bulges out of the surface. That local disturbance is called flux leakage. The escaping field creates small magnetic poles along the crack edges, and those poles attract nearby iron particles.

The result is a visible buildup: a dark line, a fuzzy cluster, or a bright fluorescent concentration depending on the particles used. The particles do not detect the crack directly. They respond to the leakage field that the crack creates. This is also why orientation matters so much. A tight crack running parallel to the flux may barely disturb the field, while a crack crossing it at a good angle produces a much stronger signal. The rule inspectors work by is simple: the flux should cross the suspected flaw at roughly 90 degrees, which is why two perpendicular magnetizing directions are often used on the same weld.

Detection depth is limited. MPI is excellent for flaws that break the surface and useful for some very near-surface defects, but its depth capability drops quickly. It cannot match ultrasonic or radiographic testing for volumetric inspection. A flaw buried deep in a thick weld can be serious and still remain completely invisible to MT.

Discontinuity, Defect, and Indication: Three Different Words

These terms are often used interchangeably in casual conversation, but they mean different things in inspection work.

Term	What it means
Discontinuity	Any interruption or irregularity in the material, such as a crack, porosity, slag, undercut, or a geometry change. It is not automatically a defect.
Defect	A discontinuity that exceeds the acceptance criteria of the applicable code or specification. Rejection is a code decision, not a personal opinion.
Indication	The visible particle pattern produced during examination. It may come from a real flaw, a geometry effect, residual magnetism, or poor technique.

That last row is where most arguments start. An inspector who finds an indication has not automatically found a defect. The indication must be evaluated: cleaned, examined closely, sometimes re-tested with a different field direction, and compared with the acceptance criteria. Only then can the weld be accepted or rejected.

Relevant vs Nonrelevant Indications

Not every particle pattern comes from a crack. A relevant indication is one caused by a genuine discontinuity that requires evaluation against the code. A nonrelevant indication comes from something else, usually geometry or residual magnetism.

Common examples include fillet weld toes, keyways, sharp section changes, threaded features, and machined shoulders. These shapes concentrate flux even when the metal is sound, producing a line of particles that can look suspicious. Magnetic writing is another culprit. A magnetized tool, grinding, lifting, or handling can leave a residual field that attracts particles in a streak. The fix is usually to clean and demagnetize the area, then retest. Heat-affected zone permeability changes can also create misleading patterns in some steels.

An experienced inspector learns to recognize these false calls. That is one reason procedure, technique, and judgment all matter. The method is sensitive, but sensitivity without interpretation produces confusion rather than confidence.

Why the Method Endured

Magnetic particle inspection has industrial roots going back to the early twentieth century. It spread because it was fast, portable, and far more sensitive to surface cracks than visual inspection alone. Today the technique is standardized through documents such as ASTM E709 and E1444, and referenced by construction and pressure equipment codes around the world.

Its endurance is easy to understand. A yoke, a can of particles, and a trained eye can find a toe crack in minutes, often in places where larger inspection systems cannot reach. MPI is not a universal answer, but for surface-breaking flaws in ferromagnetic steel, it remains one of the most practical tools available.

How an MPI Test on a Weld Works: Step-by-Step Procedure

A magnetic particle inspection is not a single action. It is a chain of small decisions, and one weak link can turn a cracked weld into a accepted one. The inspector’s job is to control that chain from the paperwork to the final cleanup. Here is the practical sequence used on real welds.

1. Review the Procedure and Confirm the Material

Before anyone energizes a yoke, the inspector reads the applicable code, the written procedure, and the drawing. The procedure should state the magnetization method, particle type, field strength checks, lighting, and acceptance standard. It should also match the material. Carbon and low-alloy steels are ferromagnetic and suitable for MPI; austenitic stainless steels, aluminum, and copper alloys generally are not. Material certificates matter here, and filler selection for low-alloy work is covered in our guide to alloy steel welding electrodes, but for MPI the key question is whether the base metal will carry magnetic flux.

2. Clean, Dry, and Verify the Equipment

Surface preparation is where many inspections quietly fail. Scale, slag, spatter, paint, oil, and moisture must be removed because loose material can mask a leakage field and trap particles in places that look like indications. The surface should be dry, and cleaning should not smear metal over a tight crack. A light grinding or wire brushing may be needed, but the procedure should define what is allowed.

Next comes equipment verification: yoke lifting power, cable condition, particle concentration, and UV lamp output. Wet baths need regular concentration checks with a settling tube, and a dirty bath can cause both missed defects and false calls. A contaminated bath is not a minor housekeeping issue; it directly affects sensitivity.

3. Magnetize and Apply Particles in the Correct Sequence

The inspector chooses continuous or residual magnetization. In continuous magnetization, the current flows while the particles are applied. In residual magnetization, the current is stopped first and the remaining field is used. Continuous magnetization is generally more sensitive for weld examination, which is why it is the common choice for critical joints.

Particles are applied gently while the field is on, usually as a light cloud rather than a flood. On a weld, the field should be applied in two directions roughly 90 degrees apart, with overlapping strokes, because a crack lying parallel to the flux may produce little or no indication. The basic principles of magnetic particle testing explain why direction control matters so much.

4. Inspect Under Proper Lighting and Evaluate Indications

Lighting is part of the test, not an afterthought. For visible color-contrast examination, ASME Section V, Article 7 requires a minimum of 100 footcandles on the surface being examined. Fluorescent examination requires a darkened area, controlled ultraviolet light, and ambient white light typically limited to about 20 footcandles during inspection. The UV lamp should be warmed up and checked before use.

An indication is not automatically a defect. The inspector evaluates its shape, size, orientation, and location against the acceptance criteria, then marks it and records it. Documentation should include the part and weld identification, procedure and revision, equipment, particle type, magnetization method, lighting conditions, results, and inspector certification. That traceability is what makes the report defensible later.

5. Demagnetize, Clean, and Protect the Part

Demagnetization is not always required, but it may be necessary when residual magnetism could interfere with machining, welding, or service operation. It is usually done with an AC coil or yoke and then checked with a field indicator. After inspection, particles are cleaned off, the surface is dried, and any required protective coating is applied.

A disciplined inspector treats every step as part of the result. Clean surface, verified equipment, correct field direction, proper particles, adequate light, and honest recording are what separate a meaningful MPI test from a quick visual scan with extra equipment.

Magnetization Methods and Directions: Getting the Field Right

Magnetic particle inspection only works when the magnetic field crosses a defect at a useful angle. A discontinuity that runs parallel to the flux produces little or no leakage, so the particles have nothing to cluster around. That one rule explains why an inspector often magnetizes the same weld twice, with the field turned roughly 90 degrees between passes.

Why Direction Matters More Than Brute Force

Think of magnetic flux lines as stretched ropes running through the steel. When a crack cuts across those lines, the field is forced out of the surface and back in, creating a small magnetic pole on each face of the crack. That leakage field is what attracts iron particles and forms a visible indication. If the flux runs along the crack instead, it simply travels around the opening. No poles, no leakage, no indication, even when the crack is wide enough to catch a fingernail.

Weld defects do not arrive in one neat orientation. A toe crack usually follows the weld direction, a transverse crack crosses it, and a crater crack may radiate in several directions. Magnetizing only along the weld can miss a transverse crack; magnetizing only across the weld can miss a toe crack. That is why two roughly perpendicular directions are standard practice for weld examination. The method is described clearly by ASNT’s magnetic particle testing overview.

On a typical butt weld, one pass may run along the joint and the second across it. Complex geometries, such as nozzle connections or fillet welds, may need additional directions. More passes are not automatically better. Each one has to be controlled, checked, and documented so the coverage is genuinely complete.

The Main Magnetization Methods

Electromagnetic yokes are the workhorse for field weld inspection. An AC yoke gives a strong surface field and is easy to demagnetize, which suits surface-breaking cracks. A DC yoke reaches somewhat deeper, making it useful for near-surface flaws. Permanent magnets need no power supply, but their fixed strength and awkward handling limit their flexibility. Prod or contact methods pass current directly through the part, while coils and central conductors induce a circular field around a long component. Direct magnetization uses clamps or contacts to send current through the workpiece itself.

Method	Typical weld use	Limitations
AC electromagnetic yoke	General weld toes, caps, fillets, field work	Shallow penetration; needs power and cable access
DC or permanent-magnet yoke	Near-surface defects, remote locations	Heavier, harder to demagnetize, fixed strength
Prod/contact	Local areas, large fabrications, repair excavations	Arc burns, cleaning burden, operator skill required
Coil or central conductor	Long shaft-like parts, longitudinal defects	Geometry dependent; limited access on large assemblies
Direct magnetization	Small parts or controlled bench work	High current, contact damage, less practical on site

Checking That the Field Is Strong Enough

Field strength cannot be judged by the sound of the yoke or the confidence of the operator. ASME Section V, Article 7 specifies a minimum lifting power of 10 pounds for AC electromagnetic yokes at maximum pole spacing. DC or permanent-magnet yokes commonly must demonstrate higher lifting power, often in the 30 to 50 pound range, because their fields penetrate deeper. The check uses a calibrated weight, not a guess.

Pie-shaped field indicators and Ketos ring test blocks verify field direction and adequate strength before production examination. A pie gauge shows whether the field is running where it should; a Ketos ring gives a repeatable check on sensitivity. Prod methods typically use about 90 to 110 amperes per inch of prod spacing for common thicknesses, following the applicable procedure. Yoke pole spacing should stay within the manufacturer’s limit and the written procedure, since spreading the poles too far weakens the field at the weld.

Common Direction Mistakes

• One pass only: the most common shortcut, and the easiest way to miss a toe or transverse crack.
• Field aligned with the likely flaw: a longitudinal crack can disappear when the flux runs along the weld.
• Poles too far apart: the field spreads out and surface sensitivity drops.
• Poor prod contact: arcing damages the surface and can create indications that are not real defects.
• Unverified equipment: a weak yoke or damaged cable may pass a visual check but fail a lifting test.

Before any of this, confirm the base metal is ferromagnetic. Austenitic stainless steel, aluminum, and copper alloys will not respond to magnetic particle inspection. On alloy steel welds, the base metal and filler chemistry still matter, so it helps to understand alloy steel welding electrodes before selecting a technique. Get the field direction and strength right, and MPI becomes a fast, revealing surface examination. Get them wrong, and the weld can look clean while the crack stays hidden.

Wet vs Dry, Visible vs Fluorescent: Choosing Particles and Equipment

Magnetization creates the leakage field, but the particles tell the story. Pick the wrong particle type, carrier, or color and a good magnetic field can still produce a clean-looking weld that hides a crack. Particle choice changes sensitivity, coverage, speed, and how easily an inspector separates a real indication from background noise.

Dry Powder: Portable and Forgiving on Rough Surfaces

Dry magnetic powder is exactly what it sounds like: fine ferromagnetic particles dusted or clouded onto the surface. It is highly portable, forgiving on rough weld caps and dirty field conditions, and usable at elevated temperatures where liquid carriers would evaporate or boil. That makes it a practical choice for field welding, heavy structural work, and temporary inspections. Dry powder also comes in color-contrast shades, so it can be matched to the background. The trade-off is sensitivity: dry powder is usually less effective on very fine cracks than a wet suspension, and moisture on the surface can make particles cluster instead of migrate.

Wet Suspensions: Water or Oil Carriers

Wet methods suspend much finer particles in water or a light oil carrier. The liquid helps particles migrate to leakage fields and gives more uniform coverage on smooth surfaces. Water baths are common and need conditioners for wetting and corrosion protection. Oil baths wet the surface well and leave a protective film, which suits machined or finished parts. Wet inspection is generally more sensitive than dry powder, especially on smooth weld caps and machined surfaces where fine toe cracks are the concern.

Water baths need more care than most people expect. Concentration, contamination, pH, and corrosion control all need checking. Oil baths avoid some of that but still need particle concentration and cleanliness monitoring. Neither carrier compensates for poor surface preparation.

Visible Color Contrast vs Fluorescent Particles

Visible color-contrast particles are usually red or black, read against a light surface or a thin white background coating. They are simple, fast, and need no dark adaptation. ASME Section V, Article 7 requires a minimum of 100 footcandles at the inspection surface for visible magnetic particle examination. That is brighter than many field work areas provide, so lighting has to be checked, not assumed.

Fluorescent particles are viewed under ultraviolet light in a darkened area. Wet fluorescent MPI is generally the most sensitive combination and is widely used for critical weld examination in fabricated and pressure equipment. Fluorescence makes fine indications stand out, but the method depends on proper dark adaptation, controlled white light, and a healthy UV lamp. A degraded lamp or a filter that leaks visible light can make an indication effectively invisible, which is why particle testing standards and written procedures call for intensity and white-light leakage checks. The American Society for Nondestructive Testing has more background on the method. Expect extra cleanup afterward, since oil-based fluorescent baths leave residue that must be removed.

Particle / carrier	Sensitivity	Best surface	Portability	Typical weld use
Dry powder	Moderate	Rough, hot, or dirty	Excellent	Field structural, heavy fabrication
Wet visible	Good	Smooth to moderately rough	Good	Shop fabrication, general welds
Wet fluorescent	Highest	Smooth, clean	Fair to good	Pressure vessels, critical welds
Dry fluorescent	Good	Clean, portable setups	Good	Specialty field work

Bath Concentration and Contamination

Particle concentration in a wet bath must be checked regularly. Too many particles create background noise that hides fine indications; too few weaken detection and make the inspector work harder for a weaker signal. Contamination from dirt, grease, lint, rust, or mixed particle types causes false indications and missed cracks. A bath that looks fine in the tank can be badly out of specification. Settling tests, concentration checks, and routine replacement are not optional housekeeping.

Lighting and UV Equipment Care

Visible inspection needs enough white light at the surface, not just in the aisle. Use a calibrated light meter and check the weld toe and cap, not only the brightest spot. Fluorescent inspection needs a dark environment, adequate UV-A intensity, and minimal white light leakage. Clean the lamp lens, inspect the filter, let the lamp warm up, and verify output with a meter. A lamp that still glows is not necessarily a lamp that still works.

Matching the Combination to the Job

For field welding, a portable yoke with dry powder or wet visible particles is often the practical answer. In shop fabrication, wet visible is a good general-purpose choice. Pressure vessels and critical fabricated equipment usually justify wet fluorescent because fine crack detection matters. Structural steel may use dry or wet visible methods depending on the code, surface condition, and contract. The governing specification decides the method; convenience does not. Ferromagnetic alloy steels are the usual candidates, and electrode and material selection can affect surface condition and inspectability, as covered in this guide to alloy steel welding electrodes.

MPI Codes and Acceptance Criteria: ASME, AWS, ASTM, and ISO

An MPI procedure can be technically perfect and still cause an argument on the shop floor. The inspector sees an indication, the welder insists the weld is fine, and someone reaches for the drawing. The missing piece is usually a clear separation between two different questions: how the test must be performed, and what result is acceptable. Codes answer those questions in separate documents, and confusing them is one of the most common reasons an inspection report gets challenged.

The method itself is well established. ASNT describes magnetic particle testing as a surface and near-surface method for ferromagnetic materials. What varies from project to project is the written rulebook that says which indications matter and how large they may be.

Examination Standards Tell You How to Test

ASME Section V, Article 7 defines how magnetic particle examination is performed. It covers surface preparation, magnetization methods, particle application, field-strength verification, and the way indications are recorded. What it does not do is tell you whether a weld passes. That distinction trips up plenty of fabricators who assume a Section V reference on a drawing is an acceptance standard. The ASME code library is the place to confirm which document applies to a given construction type.

ASTM documents serve a similar supporting role. ASTM E709 is a guide to magnetic particle testing technique, while ASTM E1444 is a practice covering equipment and procedure details. Internationally, the ISO 9934 series covers general principles, detection media, and equipment. None of these documents invents a project acceptance limit. As the ANSI standards overview makes clear, they provide the technical framework that other codes reference.

Acceptance Standards Tell You What Passes

Acceptance criteria live in the design or construction code. ASME Section VIII Division 1 contains acceptance standards for pressure vessels and references Section V for the examination method. ASME B31.3 does the same for process piping. AWS D1.1 is the governing structural welding code for many steel fabrications in the United States and includes magnetic particle examination requirements within its inspection provisions. The American Welding Society publishes that code, along with commentary that helps readers interpret intent.

Document	What It Governs	What It Does Not Do
ASME Section V, Article 7	MPI technique and examination requirements	Set acceptance limits
ASME Section VIII Div. 1	Pressure vessel acceptance criteria	Teach magnetization technique
ASME B31.3	Process piping acceptance criteria	Replace the contract specification
AWS D1.1	Structural steel fabrication and inspection	Apply automatically to every steel weld
ASTM E709 / E1444	Technique guidance and equipment practice	Define project acceptance
ISO 9934 series	International MPI principles, media, equipment	Override the contract

How Common Criteria Treat Weld Discontinuities

The details vary, but the themes repeat. Cracks are generally unacceptable regardless of size or orientation under common pressure and piping acceptance standards. Linear indications are evaluated by length and alignment, while rounded indications are judged by size, count, and distribution because they often point to porosity or slag. Undercut may be permitted within depth and length limits, but inspectors often use MT to confirm that a sharp toe is not hiding a crack. Surface porosity is assessed against size and frequency rules that differ between codes.

The Contract Is the Final Word

In practice, the hierarchy runs from the contract to the design code, then to the referenced examination standard, and finally to the written procedure. The inspector’s job is to apply those limits consistently and record what was found, not to invent a pass or fail threshold on the spot. If the acceptance criteria are missing or contradictory, the correct move is a request for information before testing begins.

One practical limit sits above every code: MT only works on ferromagnetic material. Aluminum fabrications fall outside its scope entirely, which is why their inspection routes look different. If your work involves aluminum, our Aluminum MIG Welding: The Complete Practical Guide explains the processes and quality checks that actually apply.

Good documentation closes the loop. The report should identify the governing code, the acceptance criteria used, the technique, the equipment, and each relevant indication with its location and disposition. That record is what protects the fabricator, the inspector, and the end user when a weld is reviewed years later.

What MPI Finds in Welds and What It Cannot Find

Magnetic particle inspection is often called a surface method, but that label both undersells and oversells it. MPI is extraordinarily sensitive to a crack that breaks the surface of ferromagnetic steel, sometimes finding a defect too tight for the eye to see. It is also completely blind to a plane of lack of fusion sitting 10 mm below the cap. Knowing that boundary is the difference between a useful inspection and a false sense of security.

What MPI Reliably Finds

The method works best when three conditions line up: ferromagnetic material, a discontinuity open or nearly open to the surface, and a magnetic field crossing the flaw at a useful angle. Under those conditions, magnetic particle testing reliably reveals:

• Toe cracks. The classic MPI find: a tight fatigue or restraint crack running along the weld toe, often invisible under shop light.
• Crater cracks. Star-shaped cracks at a stop-start point or arc termination, where shrinkage stress concentrates.
• Longitudinal and transverse cracks. Cracks running along the weld or across it, provided they reach the surface and cross the field.
• Hydrogen-assisted cold cracks. Delayed cracks in thick, restrained, or high-strength steel welds, which is why inspection timing matters.
• Surface-breaking lack of fusion. Sidewall or interpass lack of fusion that intersects the weld face or root.
• Undercut. A sharp, deep toe groove can produce a strong indication, though shallow undercut may be judged visually.
• Surface porosity and grinding cracks. Open pores show as individual dots; abusive grinding can leave fine cracks that MPI picks up.

Where MPI Struggles or Fails

Depth is the hard limit. A discontinuity deeper than a few millimeters is generally beyond reliable magnetic particle detection, and AC yokes are especially surface-focused because the alternating field concentrates near the surface. DC fields reach a little deeper, but MPI is still not a volumetric method.

• Mid-wall lack of fusion. It is only reliably detected when it intersects the surface or lies very close to it.
• Internal slag and deep porosity. Buried flaws do not disturb the surface field strongly enough to pull particles into a readable indication.
• Defects parallel to the flux. A crack aligned with the magnetic field may produce little or no leakage, which is why two field directions are used.
• Tight laminations. These planar flaws run roughly parallel to the surface and may not create a useful leakage field.
• Nonferromagnetic material. Austenitic stainless steel welds are not ferromagnetic in the usual sense, and aluminum and copper cannot be examined with standard magnetic particle techniques at all.

For aluminum fabrications, MPI is not the right tool; process selection and prep are covered in Aluminum MIG Welding: The Complete Practical Guide.

Coatings and Surface Barriers

Paint, primer, galvanizing, and thermal spray create a barrier between the particle and the flaw. A thin, tightly bonded coating may be acceptable under a qualified procedure, but most procedures set a maximum coating thickness because sensitivity drops as coating thickness rises. A thick or poorly bonded coating can bridge over a fine crack and prevent particle accumulation at the leakage field. Metallic thermal spray is especially troublesome because the coating itself can create magnetic indications that have nothing to do with the weld.

How Weld Cap Profile Affects Interpretation

The weld cap is not just cosmetic. A convex, ropey cap with deep ripples, spatter, or a sharp toe can trap particles mechanically and create nonrelevant indications. Wet particles can pool in a concave cap or a valley between beads, and a heavy layer of particles can hide the sharp, well-defined line that signals a real crack. A genuine crack indication usually persists after light wiping and follows a metallurgical feature; loose powder sitting in a groove tends to scatter or wash away.

The toe deserves special attention. It is a stress concentrator, and a tight toe crack can be nearly invisible. A good inspector positions the field so it crosses the toe rather than running along it, then checks the same area in a second direction.

Using MPI Within a Larger Inspection Plan

MPI is a complement, not a replacement. It is fast and sensitive for surface-breaking flaws, which makes it useful after visual inspection on fatigue-critical welds, restraint-heavy repairs, and thick sections where hydrogen cracking is a concern. It cannot confirm internal fusion or deep porosity; those need ultrasonic or radiographic testing. Use MPI where the risk is at the surface, and use volumetric methods where the risk is inside the joint.

MPI vs Penetrant, Ultrasonic, and Radiographic Testing

Choosing a weld NDT method is not about finding the “best” test. Each method answers a different question. MPI asks whether something is breaking the surface of ferromagnetic steel. Penetrant asks a similar surface question on more materials. Ultrasonic and radiographic testing look inside the weld. That division prevents wasted time, missed defects, and unnecessary inspection cost.

What Each Method Can and Cannot See

Visual testing (VT) is the fastest, cheapest option. It catches misalignment, poor cap shape, undercut, spatter, and obvious surface cracks. It cannot resolve a tight toe crack that is nearly closed, which is why VT is usually a first pass rather than the final word.

Liquid penetrant testing (PT) works on nonporous materials, including austenitic stainless steel and aluminum. Penetrant enters a surface opening, dwells, then bleeds out onto a developer. PT excels on complex shapes and nonmagnetic alloys, but it detects only flaws open to the surface. A slag-filled or hammered-shut crack may produce no indication at all. Cleanliness matters enormously, and aluminum often needs its own prep routine, as covered in our guide to aluminum cleaner for welding.

Magnetic particle testing (MT or MPI) is restricted to ferromagnetic materials such as carbon steel and many low-alloy steels. The part is magnetized and dusted or flooded with iron particles. Flux leakage at a surface or near-surface discontinuity attracts those particles and forms a visible indication. MPI is fast, portable, and sensitive to tight cracks that VT misses. The ASNT overview of magnetic particle testing is a useful primer on its capabilities.

Ultrasonic testing (UT) sends high-frequency sound into the weld and reads the echoes. It detects internal flaws and is especially sensitive to planar defects oriented roughly perpendicular to the beam. That orientation dependence cuts both ways: a crack parallel to the beam may barely respond. UT also demands skilled interpretation, good surface contact, and a suitable couplant.

Radiographic testing (RT) passes radiation through the weld onto a detector. It reveals volumetric flaws such as porosity and slag but is relatively insensitive to tight planar cracks unless the beam aligns with the defect. RT often needs access to both sides of the joint and always requires strict radiation controls. The NRC guidance on industrial radiography explains why exclusion zones and trained radiographers are nonnegotiable.

Side-by-Side Comparison

Method	Best at finding	Material limits	Surface prep	Speed and cost	Key limitation
VT	Profile, undercut, misalignment	Any	Clean, good access and light	Fastest, lowest	Tight cracks can be invisible
PT	Surface-breaking cracks, porosity	Nonporous, incl. stainless and aluminum	Clean, dry, no paint or scale	Moderate	Surface-only; contamination causes false calls
MPI	Surface and near-surface cracks, surface lack of fusion	Ferromagnetic steel only	Clean, dry, coating usually removed	Fast, relatively low	No interior data; direction-dependent
UT	Internal planar flaws, lack of fusion, laminations	Most metals, with grade and geometry limits	Smooth contact, couplant	Slower, higher skill cost	Orientation-sensitive; interpretation-heavy
RT	Porosity, slag, volumetric flaws	Most materials; thickness and access limits	Less finish-sensitive, but access matters	Slower, higher cost, radiation controls	Misses tight cracks unless aligned

Why Codes Often Require a Combination

Surface and volumetric flaws are different problems, so one method rarely closes the loop. Many pressure equipment codes require surface examination by MT or PT in addition to volumetric examination by RT or UT. The rules come from the governing code and contract, not personal preference. ASME’s code finder is the place to confirm which sections apply to vessels, piping, or other equipment. Structural work under AWS D1.1 follows a different acceptance framework, and international projects may use ISO 9934 for magnetic particle testing.

A typical pressure vessel butt weld might get RT or UT through the weld body plus MT or PT at roots, toes, and attachment welds. The volumetric test finds buried slag or porosity; the surface test finds the tight toe crack radiography may never see. Neither is redundant.

A Practical Decision Guide by Scenario

• Field weld on carbon steel structure: VT first, then an MPI yoke for surface and near-surface cracking. Portable, quick, and suited to irregular conditions.
• Austenitic stainless or aluminum: Skip MPI. Use PT for surface flaws and UT or RT for internal inspection.
• Pressure vessel or process piping butt weld: Plan RT or UT volumetrically, then add MT or PT where the code requires surface coverage.
• Thick-section nozzle or suspected lack of fusion: UT, often phased array, is usually stronger because the flaw may sit well below the surface.
• Repair verification: VT, then the surface method that suits the material. Re-test after excavation and repair rather than re-inspecting the original indication.
• Coated or painted weld: Remove the coating locally before MPI or PT. A thick coating can bridge a crack and mask the indication.

MPI earns its place because it is fast, portable, and unusually good at finding the surface-breaking cracks that matter most in fatigue-loaded steel. It is not a substitute for volumetric testing. The right plan uses MPI where surface integrity is the question, PT where the material is nonmagnetic, and UT or RT where the weld interior must be proven.

Common MPI Mistakes, Safety Risks, and Fixes

Most failed magnetic particle inspections don’t fail because the inspector lacked talent. They fail because someone skipped a boring step. A surface that looked dry but wasn’t. A yoke dragged in one direction only. A particle bath topped up for months without a concentration check. The flaw was there the whole time; the method just never got a fair chance to find it.

The Mistakes That Cause Missed Defects and False Calls

Nearly every problem traced back through an MPI report falls into one of eight buckets. Here’s what each one does to your results, and the fix that actually works.

Mistake	What Goes Wrong	Practical Fix
Inadequate cleaning	Oil, scale, paint, or moisture blocks particles from reaching the leakage field	Degrease, dry fully, and wire-brush or grind only where the procedure allows
Wrong magnetization direction	Cracks parallel to the flux path produce almost no leakage	Magnetize in two directions roughly 90° apart
Insufficient field strength	Weak leakage fields leave fine toe cracks invisible	Verify lifting power or use a pie gauge or Ketos ring before you trust the field
Poor particle concentration	Too dilute misses flaws; too heavy buries them in background	Check bath concentration with a settling tube at the stated interval
Wrong particle type	Dry powder on a smooth machined toe, or visible particles where fluorescent sensitivity is required	Match particle and carrier to the procedure, surface finish, and lighting available
Bad lighting	Shadows and glare hide faint, nonrelevant, and relevant indications alike	Measure white light and UV irradiance instead of eyeballing it
Inspecting too soon	Hydrogen-assisted cracks haven’t opened yet, so the weld passes now and fails later	Respect the cooling or hold time written into the procedure
Over-cleaning after magnetization	Excess particle removal wipes away the very indication you were hunting	Remove excess gently and interpret before any aggressive cleaning

Timing and Background: The Two Quiet Killers

Two of these deserve extra attention because they rarely announce themselves.

Timing. Testing a weld before it has cooled and finished delayed cracking can produce a clean report on a joint that cracks hours later — during fit-up, at final inspection, or in service. Hydrogen-assisted cold cracking in higher-strength steels is a time-dependent phenomenon, and many procedures build in a hold period for exactly that reason. If the drawing, the governing code, or the written procedure specifies a waiting time, that waiting time is part of the test, not an optional delay.

Background. Excess background from a dirty particle bath or poor lighting is one of the most common reasons fine indications are missed. A contaminated bath leaves particles clinging to the whole surface, so a hairline toe crack has to compete with noise everywhere. The ASNT overview of magnetic particle testing is a useful refresher on why particle condition and illumination are treated as method parameters rather than housekeeping details.

Safety Risks That Catch People Off Guard

• Electrical hazards. Prods and yokes run on mains power and can deliver a genuine shock, especially in damp work areas or on grounded structures. Inspect leads, use residual current protection, and never let anyone lean on the part during magnetization. OSHA’s welding and cutting guidance covers the broader electrical and hot-work context.
• Arc strikes and surface damage. Prod contact can cause arc strikes and localized overheating. On critical welds those marks may themselves become rejection points — you can create a defect while looking for one. Poor contact pressure, dirty contact tips, and careless prod placement are the usual culprits.
• UV exposure. Unfiltered ultraviolet light from fluorescent inspection lamps can harm eyes and skin. Shield the work area, wear UV-rated eye protection, and keep bystanders out of the beam.
• Aerosol inhalation. Solvent cleaners, developer sprays, and carrier mists all put something in the air. Ventilate, and follow NIOSH guidance on welding-related exposures where fumes or mists accumulate.
• Skin irritation. Prolonged contact with oils and solvent-based carriers aggravates skin. Gloves and barrier cream are cheap insurance.
• Housekeeping. Oversprayed particles, slippery floors, tangled cables, and unlabeled baths turn a tidy inspection into a hazard. Clean as you go.

Who Should Be Doing This Work

Competence is not self-declared. In the United States, personnel certification commonly follows ASNT SNT-TC-1A or ANSI/ASNT CP-189, with the employer responsible for the written practice and for certifying its own inspectors. That employer-written procedure is what turns a generic standard into a specific test for your weld. Before you accept an MPI report, ask which written practice the inspector was certified under and who signed off.

Pre-Inspection Checklist

• ☐ Written procedure and applicable code on hand
• ☐ Material confirmed ferromagnetic; coating thickness acceptable
• ☐ Surface clean, dry, and free of loose scale
• ☐ Equipment calibrated; yoke lifting power verified
• ☐ Particle concentration and carrier condition checked
• ☐ White light or UV irradiance measured
• ☐ Required cooling or hold time completed
• ☐ Two magnetization directions planned, roughly 90° apart
• ☐ Indications marked, measured, and recorded before cleaning
• ☐ Repair excavation, re-test, and documentation flow agreed

When a repair is needed, excavation is often done by gouging before re-welding and re-inspection — the same discipline covered in this guide to air arc welding and carbon arc gouging. Get the small habits right and MPI will keep finding the flaws that visual inspection never will.

Cost, Preparation, and a Practical Checklist for Your Next Weld MPI

An MPI test can take minutes. Getting a useful, defensible result takes preparation. Most inspection problems are not caused by a mysterious defect. They come from scale left on the weld toe, a procedure nobody read, or a report that cannot prove what was covered.

How to Prepare a Weld for MPI

Start with the drawing and the contract. Confirm the acceptance standard, how much of the weld must be examined, and whether repairs need a different technique. Then give the inspector a surface particles can actually reach.

• Let the weld cool to a safe, stable temperature and allow the code-required delay for delayed cracking in susceptible steels.
• Remove slag, spatter, scale, rust, paint, oil, and loose coatings from the examination area. Grinding should be smooth and controlled, not a gouge that creates a false indication.
• Dry the surface completely. Moisture dilutes wet particles and disrupts dry powder.
• Provide safe access to the full joint, including toes, cap, and any area the procedure defines as coverage.
• Confirm the material is ferromagnetic. MPI does nothing useful on austenitic stainless steel, aluminum, or copper alloys.
• Have the approved procedure, acceptance criteria, and repair history on site.

Inspectors should work to a written procedure qualified or approved for the applicable code and material. Borrowing a procedure from another job is a shortcut that can invalidate the whole examination. The relevant requirements live in documents such as ASME’s code and standards system and project specifications.

What to Ask Your Inspection Provider

Before the crew arrives, ask direct questions. Which procedure and acceptance standard will be used? What certification does the inspector hold, and is it through a recognized scheme such as ASNT SNT-TC-1A or CP-189? How will yoke or prod field strength be verified? Which particle type and carrier, and why? What is included in coverage, and what will the report contain? A provider who cannot answer clearly is a risk, however competitive the quote looks.

What a Good MPI Report Contains

An MPI report should identify the procedure used, equipment and particle type, magnetization technique, coverage, indications found, and the acceptance standard applied. It should also record the weld identification, surface condition, lighting method, date, inspector, and the disposition of every relevant indication. Photographs help but do not replace measurements and descriptions.

What Actually Drives the Cost

Access, surface preparation, and documentation often drive inspection cost more than the examination itself. Magnetizing and particle application are fast. Getting a lift to a high joint, scaffolding a tank, blasting a contaminated surface, or producing a traceable report takes time.

• Access: ladders, lifts, confined-space entry, out-of-hours work.
• Surface preparation: cleaning, grinding, drying, especially on site.
• Coverage: full weld length versus spot checks, and two magnetization directions.
• Lighting: white light or UV floodlights, plus intensity verification.
• Documentation: reports, photographs, mapping, qualified review.
• Mobilization and re-test: travel, setup, returning after repair.

Do not choose an inspector on day rate alone. Cheap coverage usually costs more after a failed audit.

Practical Pre-Inspection Checklist

Item	Confirm before the test
Drawing and scope	Weld locations, extent of examination, and repair areas marked.
Code and acceptance criteria	Contract names the standard, revision, and limits.
Procedure	Written, approved, matched to material and joint.
Equipment calibration	Yoke, prod set, light meter, UV lamp in calibration.
Personnel certification	Qualification current and appropriate for method and level.
Surface condition	Clean, dry, smooth, free of masking coatings.
Lighting	White-light or UV intensity checked and recorded.
Recording	Report forms, weld maps, camera, markings ready.
Repair and re-test flow	Excavation, repair approval, re-inspection agreed.

When an Indication Appears: Evaluate, Mark, Excavate, Repair, Re-Inspect

An indication is not automatically a defect. The inspector evaluates it against the acceptance standard, measures it where required, and marks the area. If it is rejectable, the defect is excavated to sound metal, repaired under an approved welding procedure, and re-examined with the same method and criteria as the original weld. That closed loop keeps a repair from becoming an unverified patch.

The Bottom Line

MPI is fast, portable, and remarkably good at finding surface-breaking cracks and near-surface flaws in ferromagnetic steel. It is also limited: it cannot see deep internal defects, and it struggles with tight laminations or flaws aligned with the flux. That is why critical joints usually get MPI or penetrant for the surface plus ultrasonic or radiographic testing for the volume. For aluminum, MPI is not an option at all; our aluminum MIG welding guide covers methods that suit nonferromagnetic material. For steel, treat MPI as one part of a planned examination: clean surface, qualified procedure, and a report that proves what was done.

Frequently Asked Questions

What is an MPI test for welding?

An MPI test, also called magnetic particle inspection or magnetic particle testing, magnetizes a ferromagnetic weld and applies iron particles to reveal surface and near-surface discontinuities. Where a crack or lack of fusion interrupts the magnetic field, particles cluster into a visible indication. It is fast and effective, but only for ferromagnetic materials.

Can MPI detect lack of fusion in a weld?

Only when the lack of fusion reaches the surface or lies very close to it. A mid-wall lack of fusion that runs roughly parallel to the magnetic field may produce no detectable leakage. Ultrasonic testing is usually the better choice for internal planar defects.

Does MPI work on stainless steel?

Standard MPI does not work on austenitic stainless steel because it is essentially nonmagnetic. Ferritic and duplex stainless steels may be inspectable depending on their magnetic permeability, but the procedure must be validated. For austenitic stainless, use liquid penetrant testing instead.

How soon after welding can you perform an MPI test?

The weld must be cool enough to inspect safely and the surface must be clean. For crack-sensitive steels, the governing code or procedure may require a delay of 24 to 48 hours so hydrogen-assisted cracks have time to develop. Always follow the specified procedure rather than testing immediately for convenience.

What is the difference between MPI and dye penetrant testing?

MPI uses magnetic fields and iron particles and works only on ferromagnetic materials, but it can detect some near-surface flaws. Dye penetrant testing uses a liquid dye drawn into surface openings and works on any nonporous material, including stainless steel and aluminum, but detects surface-breaking flaws only.

What does a magnetic particle indication look like?

With visible color-contrast particles, a relevant indication usually appears as a sharp dark line or cluster against a lighter background. With fluorescent particles under ultraviolet light, it glows bright green-yellow. Indications from geometry or magnetic writing can look similar, so evaluation by a qualified inspector is essential.

Is an MPI test expensive?

MPI is generally one of the lower-cost nondestructive examination methods per unit of surface area. Total cost depends on access, surface preparation, extent of coverage, lighting requirements, documentation, and mobilization. It is usually cheaper and faster than ultrasonic or radiographic testing for surface examination.

Can MPI be done on painted welds?

Thin, tightly adherent coatings may be acceptable under some procedures, but thick paint, galvanizing, and thermal spray can bridge fine cracks and block particle attraction. In most code work the weld and adjacent heat-affected zone must be bare, clean metal before examination.