Shear Wave NDT: Angle Beam UT Explained for Weld Inspection

Shear wave NDT explained: how angle beam UT turns sound into a transverse wave, angle and skip distance, calibration blocks, and where shear waves fail.

Atlantis NDT · March 2026 · NDT Methods

Reviewed by Anoop Rayavarapu, ASNT NDT Level III (UT, RT, MT, PT, VT). Updated 18 October 2026.

Shear wave NDT, in one paragraph: a shear (transverse) wave is an ultrasonic wave whose particle motion is perpendicular to its direction of travel. In weld inspection it is produced by mode conversion in an angle-beam probe. In steel it travels at a little over half the longitudinal velocity (about 3,230 m/s against about 5,900 m/s), which gives a shorter wavelength and better sensitivity to small flaws at a given frequency. Shear waves do not travel through liquids or gases, which is why couplant carries the sound only as far as the wedge and the steel.

How an angle-beam probe makes a shear wave

An angle-beam probe is a longitudinal-wave transducer mounted on a wedge, usually acrylic or a similar polymer. At the wedge-to-steel interface the sound refracts according to Snell's law and splits into a refracted longitudinal wave and a refracted shear wave. Beyond the first critical angle the longitudinal component no longer enters the steel, so only the shear wave remains. That is why weld probes are specified by their refracted shear angle in steel, typically 45, 60 or 70 degrees, and why the nominal angle must be checked on a calibration block: the actual angle changes with wedge wear and with the material's velocity.

What shear waves find, and what they do not

Shear waves are the workhorse of weld ultrasonics because the angled beam can be aimed so that it strikes a fusion face, a crack or lack of fusion as near to perpendicular as the joint allows. A planar flaw that faces the beam returns a strong echo; the same flaw approached edge-on returns very little, which is why procedures call for more than one angle and scanning from both sides where access allows.

Shear waves are the wrong tool for laminations and other flaws lying parallel to the surface. Those are found with a straight-beam (compression wave) scan, and that is why ASTM E164 and ASME Section V Article 4 call for a straight-beam examination of the base metal before angle-beam work: a lamination in the scanning path blocks or deflects the angled beam and leaves the weld volume behind it unexamined.

Calibration for shear-wave work

Two kinds of block are involved. An IIW-type calibration block is used to find the probe index point, verify the refracted angle and set the screen range. A reference block with side-drilled holes, such as the basic calibration block described in ASME Section V Article 4, sets sensitivity and builds the distance amplitude correction curve. The reference block has to be acoustically comparable to the part, or a transfer correction is needed. Both are explained in UT reference standards and calibration blocks and the calibration block glossary entry.

Tools to plan and practise the technique

Work out half-skip and full-skip distances and sound paths for a given angle and thickness with the free angle beam and skip distance calculator. For the standard that governs contact weld ultrasonics, see ASTM E164 weld UT. To rehearse calibration, scanning and coverage in a browser, practise angle-beam UT in the ultrasonic testing simulator. The guidance below sets out how angle, frequency and coverage change with section thickness.

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 with expiry warnings, certificate storage and ISO 17025 traceability chains — and the free calibration register template covers the accessories that are usually the missing item in an audit.

Atlantis NDT Products & Services

Need ultrasonic testing (UT) done, or learned?

Atlantis provides ultrasonic testing (UT) as an inspection service, trains technicians in it, and lets teams practise it on the Practical NDT simulator.

Refracted angle follows bevel geometry, not preference. The beam must strike the fusion face as near to perpendicular as the joint allows — 70 degrees on thin sections, 60 or 45 degrees as thickness grows, because sound path and attenuation climb with angle. Thick sections need lower frequency, more than one angle, and both half-skip and full-skip passes to cover the whole weld volume.

Two geometric relations control everything at the scanner. Half-skip surface distance from the probe index point is thickness times tan θ, and the sound path is thickness divided by cos θ; full skip doubles both. At 70 degrees the path is 2.9 times section thickness, so on a 100 mm section one leg runs past a quarter metre of steel — attenuation and beam spread wreck signal-to-noise long before the far toe is covered. Dropping to 45 degrees shortens the path to 1.4 times thickness and buys that sensitivity back. Frequency moves the same way: 4 to 5 MHz on thin plate for resolution, 2.25 MHz as the general-purpose choice, 1 to 2.25 MHz on heavy wall where grain scatter dominates. Shear velocity in steel runs near 3230 m/s against 5900 m/s longitudinal, so at equal frequency the shear wavelength is close to half — the resolution gain that pays for the geometric complexity.

Source: ASME BPVC Section V (2023 Edition), Article 4 — Ultrasonic Examination Methods for Welds; AWS D1.1/D1.1M Structural Welding Code — Steel, which tabulates the required search-unit angle against material thickness; ISO 17640 for ultrasonic testing of welded joints; ISO 10863 for time-of-flight diffraction where thickness defeats amplitude-based pulse-echo.

Technically reviewed by Anoop Rayavarapu — ASNT NDT Level III (UT, RT, MT, PT, VT)

Angle-beam setup by section thickness
Section thickness	Refracted angles	Frequency	Coverage strategy	Limiting factor
Under 13 mm (1/2 in)	70 degrees	4–5 MHz	Full skip from one side; the whole volume sits inside one leg	Dead zone and near-field length relative to thickness; wedge noise
13–38 mm (1/2–1-1/2 in)	70 degrees, 60 degrees to confirm	2.25–5 MHz	Half skip and full skip, both sides of the weld where access allows	Root geometry echoes confused with lack of penetration
38–75 mm (1-1/2–3 in)	60 and 45 degrees	2.25 MHz	Multiple angles, both sides, both surfaces; weld volume split into depth zones	Sound path length; attenuation across the far third of the volume
75–150 mm (3–6 in)	45 degrees, 60 degrees for the upper zone	1–2.25 MHz	Half skip only, scanning from both surfaces; transverse scan for cracks along the weld axis	Beam spread and attenuation; index-point drift on curved surfaces
Over 150 mm (6 in)	45 degrees plus tandem pairs or PAUT sectorial	1–2 MHz	Tandem for vertically oriented fusion-face flaws; TOFD for through-wall sizing	Amplitude response no longer sizes reliably at that path length
Austenitic or clad, any thickness	Shear wave unreliable	1–2 MHz	Dual-matrix transmit-receive longitudinal, or TOFD	Coarse columnar grain scatters and skews the shear beam

These angles are the engineering starting point; the applicable code tabulates the required search-unit angle against thickness and that table governs. Thickness bands are cumulative — a 100 mm weld gets the thick-section coverage strategy, not just the thick-section angle.

Why not use 70 degrees on everything?

The sound path at 70 degrees is 2.9 times section thickness against 1.4 at 45 degrees. On heavy wall that multiplies attenuation loss and widens beam spread, dropping far-side flaw signals below reference level. The 70 degree probe earns its place on thin sections, where a shallower angle would put the full-skip index point beyond the reachable scanning surface.

How far back from the weld do I scan?

Half-skip distance from the index point is section thickness times tan θ; full skip doubles the thickness term, and half the weld cap width is added to reach the centreline. At 45 degrees on 50 mm that is 50 mm to half skip and 100 mm to full skip. Confirm scanning-surface clearance before writing the technique — obstructions decide angle as often as physics does.

Do I need to scan from both sides of the weld?

Both sides where access allows. A planar flaw on one fusion face reflects poorly to a probe approaching from the same side at an unfavourable angle, so single-sided coverage leaves a real detection gap. Where nozzle welds or pipe-to-shell joints force one-sided access, record it as a technique limitation; phased array sectorial scanning is what recovers the missing angles.

Why does shear wave fail on austenitic stainless and clad welds?

Coarse columnar grains growing along the solidification direction scatter and refract the shear beam. The beam skews off its nominal angle, the noise floor climbs, and a real flaw signal cannot be separated from grain noise. Low-frequency dual-matrix transmit-receive longitudinal probes or time-of-flight diffraction replace it, both chosen because longitudinal waves scatter less in that structure.

Amplitude says the flaw is rejectable — does it tell me the height?

Not on thick sections. Amplitude response is governed by flaw orientation and face roughness as much as by size, and a badly oriented planar flaw returns very little. Through-wall height for fitness-for-service input comes from tip-diffraction technique or time-of-flight diffraction, which measure arrival time rather than echo strength. Workmanship acceptance uses amplitude; engineering critical assessment does not.

What calibration proves the technique before scanning starts?

Three things, all recorded. Refracted angle and probe index point verified on a calibration block at the wedge temperature you will scan at. A distance-amplitude correction curve or reference level built from the specified reference reflectors. Then transfer correction measured on the component itself, covering the difference in surface roughness and attenuation between block and part.