Ultrasonic Testing: The Ultimate Guide (UT, PAUT, TOFD) — 2026

Ultrasonic Testing (UT) is the workhorse of industrial NDT: high-frequency sound waves (typically 1–5 MHz) detect internal defects and measure remaining wall thickness in metals up to 300 mm thick. This guide — written under ASNT NDT Level III authorship — covers the physics, the six UT method families, weld inspection, thickness measurement, governing standards, UT vs RT, and the Level II certification path.

What Is Ultrasonic Testing?

Unlike surface-only methods (MT, PT), UT penetrates deep into material. A transducer converts electrical pulses into sound; boundaries and defects reflect echoes, and echo timing reveals depth. UT dominates weld inspection, pipeline thickness surveys, tank integrity assessment, and pressure vessel examination: it finds cracks, voids, laminations, and lack of fusion with no radiation hazard, portable equipment, and single-side access.

How UT Works: Physics and Principles

The core relationship is simple: Thickness = (Sound Velocity × Echo Travel Time) / 2. Steel carries sound at roughly 5,850 m/s, so a 3.4-microsecond round trip equals 10.0 mm of wall. Four concepts govern real-world performance: acoustic impedance (Z = ρ × v — impedance mismatch causes reflection), attenuation (coarse-grained and austenitic materials absorb sound, limiting depth), coupling (sound will not cross an air gap — couplant gel or water is mandatory, and poor coupling loses 50–90% of signal), and frequency selection (higher frequency finds smaller flaws but penetrates less).

The 6 Ultrasonic Testing Methods

Method	Frequency	Penetration	Best For
Conventional Pulse-Echo (A-scan)	1–5 MHz	100–300 mm	Thickness measurement, defect depth sizing
Through-Transmission	1–5 MHz	100–300 mm	Composite disbond and delamination screening
Phased Array UT (PAUT)	2–5 MHz	50–200 mm	Fast multi-angle weld scanning and characterization
TOFD	1–2 MHz	50–150 mm	High-accuracy depth sizing, tight crack detection
Automated UT (AUT)	1–5 MHz	50–200 mm	Tank floor scanning, pipeline production inspection
Immersion UT	1–10 MHz	10–150 mm	Aerospace components, complex geometry, lab work

PAUT fires 32–256 elements in timed sequences to steer and focus the beam electronically, inspecting at multiple angles (typically 35–70°) from one probe position — 5–10x faster than conventional scanning with superior small-flaw detection. TOFD uses diffraction from flaw edges instead of face reflection, delivering depth accuracy around ±0.5 mm on 25 mm sections and finding cracks tighter than 0.1 mm. Modern pipeline and pressure-vessel projects routinely pair PAUT + TOFD.

Ultrasonic Weld Inspection

UT detects the internal weld defects that decide fitness for service: porosity, lack of fusion, lack of penetration, slag inclusions, tungsten inclusions, and cracks including delayed hydrogen cracking. Acceptance criteria come from the governing code: ASME Section V for pressure equipment, AWS D1.1 for structural steel (zero tolerance for cracks), API 1104 for pipeline girth welds, and ASME Section VIII for vessels. Surface-breaking flaws remain better served by MT or PT — UT and surface methods are complements, not substitutes.

Wall Thickness Measurement and RBI

Pulse-echo thickness gauging achieves ±0.1–0.5 mm accuracy, works through paint, and feeds directly into Risk-Based Inspection. Best practice: grid scan patterns at 25–50 mm spacing, calibration on a reference block before and after each survey, and corrosion-rate trending against prior baselines. Remaining Safe Operating Life = (Current Thickness − Minimum Allowable) / Corrosion Rate, per the API 579 fitness-for-service framework — the calculation that turns thickness data into inspection intervals.

UT Standards and Codes

Key references: ASME Section V (pressure equipment NDE), ASTM E164 (contact UT of weldments), API 1104 (pipeline girth welds), AWS D1.1 (structural welds), API 650 (storage tanks), ASTM E494/E2375 (composites/aerospace), and ISO 17640/16810.

Ultrasonic Testing vs Radiography

UT wins on tight-crack detection, exact depth sizing, speed, safety (no radiation), and single-side access. RT wins on porosity-pattern visualization, permanent image records, and very thin or contractually film-mandated work. In modern practice UT (increasingly PAUT) is the primary weld method with RT as confirmation — see the full RT vs UT comparison.

UT Certification and Training

ASNT Level II UT under SNT-TC-1A requires formal instruction (typically 120–160 hours), supervised field experience, and written, practical, and specific examinations. PAUT and TOFD add specialized training on top. The career math is compelling: a UT Level III with PAUT and API credentials commonly earns $120,000–$180,000 per year, and PAUT skills alone carry a substantial pay premium — see the global NDT salary guide. Atlantis NDT's ASNT certification training, led by an ASNT Level III, covers UT Level I/II plus PAUT and TOFD specializations; explore eddy current testing for the complementary surface method.

Frequently Asked Questions

What is ultrasonic testing used for?

Weld inspection, wall thickness measurement, corrosion mapping, tank floor scanning, forging and casting inspection, and composite testing — anywhere internal flaws or remaining thickness must be quantified without damaging the part.

What is the difference between UT and PAUT?

Conventional UT uses a single beam from one transducer; PAUT electronically steers and focuses a multi-element array, covering multiple angles simultaneously. PAUT is faster, characterizes defects better, and reduces operator dependency, at the cost of more training and higher-end equipment.

What is TOFD?

Time-of-Flight Diffraction uses two angled probes bracketing a weld and times diffracted signals from flaw tips, giving exceptional through-wall sizing accuracy. It is the preferred technique when precise crack height drives an accept/reject or fitness-for-service decision.

What materials can UT inspect?

All common metals (carbon steel, stainless, aluminum, titanium, copper) and many composites. Coarse-grained castings and austenitic welds attenuate sound and need lower frequencies or specialized probes; most polymers and rubbers scatter sound too heavily.

How accurate is UT thickness measurement?

Typically ±0.1–0.5 mm depending on surface condition and frequency. Smooth, calibrated surveys reach ±0.1 mm — sufficient for corrosion-rate trending and API 579 remaining-life calculations.

How do I become UT certified?

Complete formal training hours, log supervised field experience, and pass written and practical exams under your employer's written practice (SNT-TC-1A) or a central scheme. Atlantis NDT runs the full path from Level I to PAUT/TOFD specialization — request details via a free consultation.

Atlantis NDT Products & Services

When the scanning is done, the data still has to be managed. Atlantis NDT supports inspection teams end-to-end with NDT inspection management software — Atlantis ERP, our digital twin platform for asset integrity, and purpose-built NDT reporting software. Our NDT training & certification programs are authored by an ASNT NDT Level III, while ASNT Level III consulting and 3D laser scanning services cover the full inspection lifecycle. Everything is affordable, accessible, and fully customizable — book a free consultation for a tailored quote.

Calibration control at company scale

Instrument, probe, wedge and reference-block calibration is the second thing a client audit tests after personnel qualification. Calibration management covers interval control, certificate storage and ISO 17025 traceability chains, including hard lockout so an out-of-calibration instrument cannot be dispatched — and the free calibration register template covers the accessories that are usually the missing item in an audit.

UT technique selected by the question the examination has to answer
Question the examination must answer	Technique	Frequency	Sizing accuracy	Code reference
How much wall is left?	Pulse-echo thickness, dual-element on corroded backwall	2–5 MHz	±0.1–0.5 mm	ASTM E797, API 570
Is this weld acceptable?	Angle beam at 45, 60 and 70 degrees, or encoded phased array	2–5 MHz	Length and amplitude against DAC or DGS	ASME Section V Article 4; acceptance in Section VIII, AWS D1.1, API 1104
How tall is this crack?	TOFD, paired transmit-receive probes bracketing the weld	1–2 MHz	±0.5 mm on 25 mm sections	ASME Section V Article 4 TOFD appendix; feeds API 579
Is the whole tank floor sound?	Automated UT or MFL floor scanning with follow-up prove-up	1–5 MHz	Area coverage rather than point accuracy	API 653
Has this composite disbonded?	Through-transmission, or pulse-echo with a delay line	1–5 MHz	Disbond area and boundary	ASTM composite practices
Can I inspect this austenitic or dissimilar-metal weld?	Low-frequency dual-matrix transmit-receive longitudinal phased array	1–2 MHz	Reduced by beam skew and attenuation	ASME Section V Article 4 with procedure demonstration on a mock-up

UT technique selected by the question the examination has to answer

Question the examination must answer

Technique

Frequency

Sizing accuracy

Code reference

How much wall is left?

Pulse-echo thickness, dual-element on corroded backwall

2–5 MHz

±0.1–0.5 mm

ASTM E797, API 570

Is this weld acceptable?

Angle beam at 45, 60 and 70 degrees, or encoded phased array

2–5 MHz

Length and amplitude against DAC or DGS

ASME Section V Article 4; acceptance in Section VIII, AWS D1.1, API 1104

How tall is this crack?

TOFD, paired transmit-receive probes bracketing the weld

1–2 MHz

±0.5 mm on 25 mm sections

ASME Section V Article 4 TOFD appendix; feeds API 579

Is the whole tank floor sound?

Automated UT or MFL floor scanning with follow-up prove-up

1–5 MHz

Area coverage rather than point accuracy

API 653

Has this composite disbonded?

Through-transmission, or pulse-echo with a delay line

1–5 MHz

Disbond area and boundary

ASTM composite practices

Can I inspect this austenitic or dissimilar-metal weld?

Low-frequency dual-matrix transmit-receive longitudinal phased array

1–2 MHz

Reduced by beam skew and attenuation

ASME Section V Article 4 with procedure demonstration on a mock-up

DAC or DGS — which sizing curve should the procedure use?

DAC where the code names it and reference reflectors matching the expected flaw type are available; the curve is drawn from side-drilled holes at increasing sound path in a block matching the part. DGS uses a theoretical diagram tied to the probe's own beam characteristics, needs a single reference reflector, and holds up better on thick sections where a full DAC block becomes impractical to machine and carry.

Why does UT miss flaws breaking the scanning surface?

The initial pulse and the probe near field create a dead zone directly under a zero-degree probe, where a surface-breaking flaw hides inside the main bang. Angle-beam probes solve it by inserting sound at 45 to 70 degrees and reaching the near surface on a skip, and magnetic particle or penetrant testing covers the accessible surface directly. UT and surface methods are complements, not substitutes.

What limits ultrasonic testing on austenitic and dissimilar-metal welds?

Coarse, columnar, anisotropic grain structure. Sound velocity changes with direction through the weld metal, so the beam skews off its nominal angle, splits, and attenuates, while grain boundaries scatter energy back as noise that buries small reflectors. The answer is low-frequency dual-matrix transmit-receive longitudinal phased array, plus procedure demonstration on a mock-up containing the flaw type being sought.

What is transfer correction and when is it required?

Transfer correction is the decibel adjustment compensating for the attenuation and surface-roughness difference between the calibration block and the actual component. It is measured by comparing the same reflector response through both, then added to the reference level. It is required whenever component surface finish, curvature or grain structure differs materially from the block. Skipping it undersizes flaws on rough or attenuative components.

Does ASME allow ultrasonic examination in place of radiography?

Yes. ASME Code Case 2235 permits UT in lieu of RT on welds from 1/2 inch (13 mm) thickness upward, subject to a qualified procedure demonstrated on specimens containing flaws representative of those being sought, and to personnel qualified for the technique. Encoded phased array with permanent data recording is what makes the substitution auditable years after the weld was made.

How many training and experience hours does ASNT recommend for UT Level II?

ASNT SNT-TC-1A recommends 40 hours of UT training and 210 hours of method experience for Level I, then a further 40 hours of training and 630 cumulative hours of experience for Level II. Phased array and TOFD sit on top as separate technique training. ISO 9712 sets 40 hours at each level, with three months of Level 1 and nine months of Level 2 industrial experience.

Having ultrasonic testing performed on your equipment

This guide explains the method. If what you actually need is the examination carried out — thickness surveys, weld examination, corrosion mapping and phased array performed to a qualified procedure — that is the ultrasonic testing service. Teams mobilise to your site under Atlantis procedures with ASNT Level III oversight; findings are evaluated against the acceptance criteria your contract names, and records are structured to survive a client audit years later. Scope an examination.