NDT method selection
PAUT vs TOFD vs Radiography Compared
A method-level comparison of phased array, TOFD, conventional ultrasonics, radiography and eddy current testing — written around what each standard actually scopes rather than around which technique is newest.
Physical basis
Methods differ in what they respond to, not in quality.
A method comparison that ranks techniques by sophistication will select the wrong one. The useful comparison is what physical quantity each method responds to, because that determines which discontinuities it can address and in which orientation. The UK Health and Safety Executive’s inspection and NDT guidance states the same control in regulatory terms: the technique has to be matched to the damage being sought and applied under a procedure that allows capability and repeatability to be assessed.
| Method | Responds to | Scoped by | Principal constraint |
|---|---|---|---|
| Conventional ultrasonics (UT) | Energy reflected from an interface back to the probe | Procedure and applicable weld or component standard | Discontinuity orientation relative to a single configured sound path |
| Phased array (PAUT) | Reflected energy across steered or focused beams from multiple elements | ISO 23865:2021 for FMC/TFM and related technologies | Focal-law design, wedge, aperture and the coverage the scan plan actually achieves |
| TOFD | Signals diffracted from the extremities of a discontinuity, between a transmitter and receiver | ISO 10863:2020, from 6 mm thickness, four testing levels A–D | Scoped to full-penetration joints of simple geometry in low-attenuation material |
| Radiography (RT) | Differential attenuation of penetrating radiation, recorded as a projection image | ISO 17636-1:2022 (film) and ISO 17636-2:2022 (digital detectors) | Planar discontinuity response depends on orientation; a controlled area is required |
| Eddy current (ET) | Distribution of an induced alternating current, analysed in the complex plane | ISO 15549:2019 general principles | Electrically conducting material only, and surface or near-surface depth |
The common pairing
PAUT and TOFD are usually specified together, not against each other.
The two are frequently written into the same procedure because they fail in different places. Phased array is generally interpreted from reflected energy, so its response depends on how a discontinuity is oriented relative to the beams the focal laws produce; the answer to that is coverage design. TOFD interprets diffracted signals arriving between a transmitter and receiver, which makes it well suited to evaluating through-wall extent, and much less dependent on the discontinuity presenting a favourable reflecting face.
ISO 10863:2020 is explicit that TOFD may be used as a stand-alone method or in combination with other NDT methods or techniques, for manufacturing and for in-service inspection. It also sets the boundaries that decide applicability before any equipment discussion: semi- or fully automated testing of fusion-welded joints in metallic materials of minimum thickness 6 mm, applied to full-penetration joints of simple geometry, in material with low ultrasonic attenuation. It specifies four testing levels, A to D, corresponding to increasing inspection reliability.
What to write in a specification
Naming a technique does not define an inspection. State the thickness range, joint geometry, material, expected discontinuity types and orientations, the required detection and sizing outcome, the testing level, and the acceptance criteria. Whether one technique, both, or a different method satisfies that is then an answerable question.
For the distinction between phased array acquisition modes themselves — steered arrays, full matrix capture and total focusing method reconstruction — see the PAUT, FMC and TFM guide. This page stops at the boundary between methods.
UT vs RT
Radiography changes the site, not just the result.
The technical difference is that radiography records a projection through the material while ultrasonics interrogates it along sound paths. That gives radiography a strong response to volumetric discontinuities such as porosity and inclusions, and makes its response to planar discontinuities dependent on their orientation relative to the beam. Ultrasonic methods reverse much of that emphasis, which is one reason procedures for critical welds often call for both rather than treating them as alternatives.
The operational difference is frequently the deciding one. Radiographic examination requires a controlled area and radiation-safety management, which affects scheduling, access for other trades and, on an operating asset, whether the work can be done at all without a shutdown. Ultrasonic methods impose no equivalent exclusion, but require couplant, surface condition and access to a scanning surface.
ISO 17636-1:2022 covers X- and gamma-ray techniques with film; ISO 17636-2:2022 covers digital detectors and states that its procedure provides the minimum requirements permitting acquisition of digital radiographs with sensitivity for detecting imperfections equivalent to film radiography. Digital acquisition changes the recording medium and the workflow; it does not change what radiation responds to.
A boundary that is often crossed by mistake
Surface methods do not substitute for a volumetric examination.
Eddy current, magnetic particle and penetrant testing address surface and, for eddy current, near-surface conditions. They are not reduced-cost versions of a volumetric method. ISO 15549:2019 defines the general principles for eddy current examination and describes the method as based on inducing an alternating electric current in a conducting material, with the measured quantity related to the distribution of the induced currents and represented as a vector in the complex plane.
Two consequences follow directly and are worth stating in a specification rather than discovering on site:
- The material has to conduct. Eddy current is not applicable to a composite blade shell, and its applicability to a coated or clad surface depends on the coating and the configured frequency.
- Depth of interrogation is limited and frequency-dependent, so a surface-method result carries no information about the interior of the section.
Magnetic particle testing requires ferromagnetic material; penetrant testing requires a non-porous surface and an open discontinuity. Where a requirement genuinely concerns the through-thickness condition of a joint or a fastener, a surface method cannot satisfy it regardless of how many indications it reports.
Bolted joints are a separate question again
For a fastener, the requirement is often preload rather than a discontinuity. Detecting an indication and measuring axial force are different examinations with different equipment — see torque vs tension across pretensioning methods and the ultrasonic bolt tension measurement guide.
Evidence record
Sources and limitations
Scope statements and method definitions are taken from the official catalogue entries for the current editions of the standards cited. No vendor comparison material was used.
- ISO 10863:2020 — official scope for the TOFD technique, including thickness and joint-geometry applicability and testing levels A to D.
- ISO 23865:2021 — general provisions for FMC/TFM and related array technologies; does not establish acceptance levels for discontinuities.
- ISO 17636-1:2022 and ISO 17636-2:2022 — radiographic testing of welds with film and with digital detectors.
- ISO 15549:2019 — general principles for eddy current testing, including the conducting-material basis of the method.
- UK HSE inspection and NDT guidance — official guidance on matching technique to damage mechanism, procedure control and demonstrable capability.
Limitation: no detection, sizing or probability-of-detection performance is stated or implied here for any method. Capability is established by a qualified procedure on the actual component, material, geometry and discontinuity type, with qualified personnel and defined acceptance criteria. The standards cited scope how a method is applied; several of them explicitly do not set acceptance levels.
See the source and editorial policy for how supplied product manuals, external method sources and claim limitations are kept separate.
Decision support
Frequently asked questions
What is the difference between PAUT and TOFD?
Phased array uses controlled timing across multiple elements to steer or focus a beam, and generally relies on reflected energy. TOFD uses a transmitter–receiver pair and interprets signals diffracted from the extremities of a discontinuity, producing an image from which through-wall extent can be evaluated. ISO 10863:2020 scopes TOFD to semi- or fully automated testing of fusion-welded joints from 6 mm thickness, and states it can be used alone or in combination with other methods.
Is TOFD better than phased array?
Neither is generally better. They respond to different physics and have different scoped applications, which is why procedures frequently specify both. The choice depends on the joint geometry, thickness, expected discontinuity type and orientation, the required sizing accuracy, and the acceptance criteria the result has to support.
When is radiography used instead of ultrasonic testing?
Radiography produces a projection image and responds well to volumetric discontinuities such as porosity and inclusions, while its response to planar discontinuities depends on their orientation relative to the beam. It also imposes operational constraints — a controlled area and radiation-safety management — that ultrasonic methods do not. ISO 17636-1:2022 covers film techniques and ISO 17636-2:2022 covers digital detectors.
Which method finds surface-breaking cracks?
Eddy current testing responds to surface and near-surface discontinuities, but only in electrically conducting material: ISO 15549:2019 describes the method as based on inducing an alternating current in a conducting material. Magnetic particle testing applies to ferromagnetic material, and penetrant testing to non-porous surfaces of most materials. These are surface methods and do not substitute for a volumetric examination.
Engineering review
Bring the joint and the acceptance criteria, not the technique name.
Tell us the component, material, thickness, expected discontinuity type and orientation, the access available and the standard the result must satisfy. We’ll identify which methods can address it.