A magnetic particle tester is the spec-first choice when the job is surface and slightly sub-surface crack detection on ferromagnetic steel, with handheld AC/DC yokes widely listed at 1-set MOQ supply up to 8,000 sets/year [S1] and dedicated magnaflux yokes like the XXD-310DC offered at US$560–600 in 1–4 piece quantities (DC12V/AC220V supply) [S2].
An ultrasonic flaw detector is the spec-first choice for sub-surface volumetric defects, thickness sizing, and coded welds, with general-purpose UFD parameter sets governed historically by GOST 23049-84 and measurement methods by GOST 23667-85 [S3]. For ferromagnetic welds on rail, pressure vessel, and in-service structural work, the two methods are complementary, not interchangeable.
Decision rule: pick by material, defect depth, and recordable output
MT works only on ferromagnetic material because the method relies on flux leakage at a discontinuity; austenitic stainless, aluminium, copper, and most non-ferrous alloys are excluded, which immediately rules it out for many chemical, marine, and aerospace jobs [S4]. UFD works on any material that conducts ultrasound with reasonable attenuation — carbon steel, alloy steel, aluminium, copper, and many composites — but cannot resolve tight surface-breaking cracks when the transducer cannot couple to the surface or when the crack is oriented parallel to the beam [S3].
Defect depth separates the two methods cleanly: MT sees surface-open and near-surface (typically within roughly 1–2 mm of the surface, depending on magnetising force and material permeability) cracks; UFD sees sub-surface laminations, inclusions, lack of fusion, and hydrogen cracks at depths of millimetres to metres, with sizing accuracy depending on frequency, probe diameter, and DAC/TCG calibration [S4][S3]. For in-process QA of welds on rail axle, pressure vessel nozzle, and offshore node, UFD is therefore the only method that produces a recordable depth reading; MT gives a binary "indication present / not present" at the surface.
Recordable output is the second gate: digital UFDs with B-scan, TOFD, and phased-array add-ons output amplitude-vs-depth traces that feed directly into weld-record packages under ISO 9712 / ASME V [S3]. MT records a visible indication pattern on the surface, supported by before/after photos and demagnetisation logs, which is enough for many code accept/reject decisions but is qualitative, not quantitative [S4].
Comparison: magnetic particle vs ultrasonic on five decision criteria
On defect class, MT is restricted to surface and near-surface open-to-air cracks on ferromagnetic parts, while UFD covers surface, sub-surface, and through-thickness defects in any workable alloy [S4][S3]. On measurable output, MT is binary (indication or no indication, with location only); UFD is quantitative, giving depth, reflector size, and amplitude in dB or % DAC [S3]. On material coverage, MT excludes austenitic stainless and most non-ferrous alloys, while UFD covers ferritic, austenitic, and non-ferrous metals with appropriate couplant and probe selection [S4].
On operator skill and cost, MT needs a trained Level II but the hardware is low — handheld magnetic yokes start at sub-US$1,000 retail and OEM supply is broad (8,000 sets/year capacity on a single supplier line, 1-set MOQ) [S1][S2]. UFD hardware is higher, requires skilled set-up of gain, gate, and probe stand-off, and integrates with phased-array and TOFD add-ons that push system cost well above MT [S1][S3]. On throughput, MT continuous method is fast on small parts (weld overlay, fillet welds) because the part can be magnetised and read in seconds; UFD requires scanning, couplant, and per-zone calibration, so throughput drops but recordability rises [S4][S3].
Use-case map: where each method is the right tool
Magnetic particle testing is the right tool for in-process inspection of welded ferromagnetic assemblies during fabrication, for crack detection on crane hooks, axle journals, and lifting gear, and for field survey of weld toes on carbon-steel pipe where speed matters more than recordable depth [S4]. The continuous method (magnetic field applied while magnetic slurry or dry powder is dusted on) is the workshop default; the residual method (magnetise, remove field, then apply medium) is used where the part geometry makes a fixed yoke hard to hold [S4].
Ultrasonic flaw detection is the right tool for volumetric inspection of plate, forging, and casting in pressure-equipment fabrication under ASME V / EN 10160, for rail weld and bolt-hole inspection using B-scan techniques, and for in-service corrosion mapping on tank shells and offshore nodes [S3]. Phased-array and TOFD variants extend the same transducer head into full volumetric coverage of thick-wall welds, with Russia-origin general-purpose UFDs cited in the technical literature as supporting B-scan, amplitude, and TOF data channels in a single instrument [S3].
For a worked example: a 25 mm fillet weld on a S355 carbon-steel bracket calls for MT on the toe (fast, sensitive to toe cracks) and UFD on the weld throat (sizing of lack of fusion). For a 316L austenitic nozzle weld, drop MT — only UFD with shear-wave probes at low frequency (typically 2–4 MHz) is workable, and even then sensitivity is reduced versus ferritic [S4][S3]. For rail-end bolt-hole cracking, only UFD with B-scan and dedicated small-footprint probes gives the size + location data the railway code requires [S3].
Limitations, failure modes, and what each method will miss
MT will miss sub-surface defects deeper than the magnetic field can leak to, will miss tight fatigue cracks oriented perpendicular to the field (rotate the part or use a multidirectional yoke), and will not work on austenitic stainless without a separate cross-magnetisation technique that is rarely used in general QA [S4]. False positives in MT come from surface scratches, magnetic writing from prior magnetisation, and from a part that was not properly demagnetised after the test [S4]. UFD will miss cracks that are parallel to the beam, will miss near-surface defects in the dead zone of the probe (first 1–3 mm depending on frequency and element size), and will give wrong depth readings if the velocity is set for the wrong material or the couplant is dry [S3].
Couplant hygiene, probe calibration, and DAC/TCG setup are the three most common UFD failure modes in field use; documented instrument parameter checks against a known block (GOST 23667-85-style calibration protocol) are the standard control [S3]. For MT, the most common failure is insufficient magnetising force — a handheld AC yoke must pull at least 4.5 kg (ANSI/ASTM E709 general workshop guidance) on a 25 mm spacer; below that, indication sensitivity collapses.
Standards, codes, and sourcing notes for a 2026 buyer
Magnetic particle testing in the international QA chain sits on ASTM E709 / E1444, EN ISO 9934-1, and ASME V Article 7; ultrasonic flaw detection sits on ASTM E164, EN ISO 17640, EN 10160, and ASME V Article 4 — referenced here as the standard families that govern acceptance, not as a substitute for reading the latest revision of the cited code [S4][S3]. For Russian-origin UFDs, the historical parameter baselines are GOST 23049-84 (basic parameters and technical requirements) and GOST 23667-85 (measurement methods), which are still cited in technical literature as the reference baseline for general-purpose instrument parameter sets [S3].
Sourcing-side data points worth pinning: a single Beijing-based OEM lists ultrasonic flaw detectors, magnetic flaw detectors, and ultrasonic flowmeters under one product line, with stated supply ability of 8,000 sets/year and a 1-set MOQ — useful for small-batch distributors and service-company buyers [S1]. A separate Made-in-China listing for the XXD-310DC AC/DC magnetic yoke detector prices at US$600 for 1–4 pieces and US$560 for 5+ pieces, with T/T / Paypal / Western Union accepted, indicating where volume pricing kicks in for handheld MT hardware [S2]. For deeper coverage of magnetising method trade-offs and 2026 buyer checklists, see this magnetic-particle-tester selection guide.
Adjacent method for the same defect family is eddy current testing, which pairs with MT for ferromagnetic surface cracks and with UFD for sub-surface defect sizing in non-ferrous tube and aerospace structure; cross-reference the 2026 eddy current tester cost and spec map when the workpiece is austenitic or non-conductive-coating-restricted. Two final trackable signals for a 2026 spec-first purchase: phased-array UFD channels and TOFD add-on cards are now standard in the general-purpose UFD class, not premium options [S3]; and dual-purpose AC/DC yokes at the 1–4 piece tier remain the best value entry into MT for a service-lab that runs both methods on carbon-steel weldments [S2].
The underlying component specifications are covered under dust particle meter.