Phased Array Ultrasonic Testing (PAUT) Training

Focal laws: the arithmetic that replaces the wedge chart

A focal law is the delay table the instrument applies to each element so the wavefronts from all of them arrive in phase at one point in the part. Change the angle and you change every delay. Change the wedge and you change the entry point, the sound path and the exit point of every element. Conventional UT training hands you a wedge stamped sixty degrees and a chart of skip distances. Phased array training hands you the parameters — element count, pitch, aperture start, angle range and angle step — that generate hundreds of those wedges in software.

The parameters entered wrong most often are wedge velocity, first element height and probe-to-wedge-front distance. All three are printed on the wedge, all three get typed from memory, and all three shift the computed depth of every indication without producing any visible error on screen. A technician who cannot re-derive depth from time-of-flight by hand has no way to catch it. That is the specific reason a UT Level II foundation is a hard prerequisite rather than a recommendation, and why method training hours by method are structured the way they are.

Good instruction makes you build a focal law from a blank setup on a calibration block, verify it against a known side-drilled hole at three separate depths, and then deliberately corrupt one parameter to see what the resulting error looks like on screen. If a course only walks you through loading a saved setup file supplied by the instructor, you are being taught to operate one particular instrument rather than to apply the technique.

Wedge and probe selection is a coverage decision, not a catalog pick

Wedge choice is a coverage argument. The wedge natural angle sets the center of your usable sweep, and steering more than roughly twenty degrees either side of it collapses sensitivity. Element pitch sets how far you can steer at all, because a coarse pitch throws grating lobes at high angles. Frequency trades penetration against resolution: 5 MHz for general carbon steel weld work, 7.5 or 10 MHz for thin wall and small-diameter tubing, 2.25 MHz where attenuation or coarse grain structure defeats anything higher.

The decision is driven by the joint, not by what happens to be in the case. A 37.5-degree single-V on 25 mm plate demands a different aperture and angle range than a compound bevel on 12 mm pipe with a counterbore. Training that never puts a weld prep drawing in front of you produces technicians who select probes by habit, and habit is how root coverage disappears on the first job that is not a flat plate butt weld with clear access on both sides.

Standards name the technique for exactly this reason. ASTM E2700, Standard Practice for Contact Ultrasonic Testing of Welds Using Phased Arrays, governs manual and semi-automated contact work. ISO 13588 covers the use of automated phased array technology on welds and defines graded testing levels. ASTM E2491 and ISO 18563 cover how the instrument and array themselves are characterized and verified. Knowing which of the four your client's specification invokes is part of the job, and it changes the setup you are permitted to run.

The scan plan is the deliverable employers actually check

The scan plan is the drawing that proves the weld will be covered before anyone squeezes couplant onto the surface. It shows the joint geometry to scale, the probe and wedge sitting on the surface at the index offset you will actually use, and the beam envelope for the lowest and the highest angle in the sweep, first leg and second leg. Every zone of the bevel — root, fill, cap, and the fusion faces on both sides — has to fall inside a beam at some angle in the law.

This is the single element most conventional UT technicians have never produced, and it is the first thing a reviewer looks for. When a phased array report gets rejected by a third party, coverage is almost always the reason: the scan was run, the data looks clean, and nothing in the file demonstrates that the far fusion face was ever insonified. What makes an NDT report defensible works through the identical failure from the reviewer's side of the table.

Training should have you draw scan plans by hand before it lets you near software that draws them for you. The hand version forces you to reason about index offset, skip distance, and the difference between what the first leg reaches and what the second leg reaches after the beam has spread and lost amplitude. Software will happily render a beautiful, fully covered plan for a geometry that does not exist on the job.

TCG, not DAC: calibrating amplitude across the whole sweep

Conventional UT compensates for attenuation with a DAC curve built on one angle. Phased array cannot do that, because the metal path, the wedge path and the transmission coefficient at the interface all change as the angle sweeps. Time-corrected gain applies a separate gain-versus-time correction at every angle in the law, built from the same reference reflector acquired at several depths. Done properly, a 3 mm side-drilled hole reads the same amplitude at 45 degrees and at 70 degrees, and at 15 mm deep and 40 mm deep.

Building TCG takes real bench time. You place the block, acquire the reference reflector at each depth in turn, and let the instrument solve the curve across the whole aperture. Skipping depths, or building TCG on a block with a different velocity or surface condition than the component, produces a calibration that silently undersizes flaws at the extremes of the sweep — which is precisely where the fusion faces of a bevel are examined and where lack of fusion lives.

Verification matters as much as construction. Re-checking TCG at shift change, after a temperature swing, and after any wedge change is a procedural requirement in most written practices, and it is the evidence auditors ask to see. If your instruction never covers what a failed verification obliges you to do with the data already acquired that shift, it has skipped the part of the technique that protects both you and your employer.

Encoders turn a scan into evidence

An unencoded phased array scan is a live picture. An encoded one is evidence. The encoder wheel reports position along the weld, the instrument indexes each A-scan to that position, and the resulting file can be reopened months later with every indication tied to a station number that a fitter can find with a tape measure, without the technician who acquired it standing on the job to interpret from memory.

Encoder calibration is a two-minute task that gets skipped and then invalidates a full day of data. You roll the wheel a known distance — 500 mm along a straight edge is the common choice — and correct the counts-per-millimetre until the readout matches. Slip on a wet or scaled surface, a wheel dragging against a weld cap, or a magnetic wheel jumping a tack will stretch or compress the C-scan, and the resulting position error is invisible in an otherwise clean image.

Encoded data is what makes phased array competitive with radiography on a shutdown. The argument is not only that there is no source and no exclusion zone; it is that the client receives a re-interrogable data file rather than a film they cannot revisit. A technician who only ever scans unencoded cannot make that argument to a client, and cannot deliver the record that justifies replacing a radiographic technique with an ultrasonic one.

Reading an S-scan: separating a flaw from geometry

A sectorial scan paints every angle in the sweep as a wedge-shaped image with the part geometry overlaid on it. The first skill is separating real reflectors from geometry: the root bead, the counterbore, a backing bar and the far surface all produce strong, repeatable indications that sit exactly where the overlay says the geometry sits. The second skill is recognising which angle a real indication responds best at, because the answer tells you how the reflector is oriented in the weld.

Lack of fusion on a bevel face responds strongly at the angle normal to that face and falls away sharply on either side of it. Porosity responds weakly across the whole sweep and moves erratically as the probe moves. A crack shows tip diffraction — a second, weaker signal above the corner reflector — and that tip is what you measure height from. Slag gives a broad, ragged response with no tip at all. None of those signatures is available on an A-scan alone.

Sizing is where phased array earns its premium and where it is most often done badly. Amplitude sizing carried over from conventional UT overcalls long, low-amplitude planar flaws and undercalls short, well-oriented ones. Tip diffraction and the 6 dB drop applied along the correct axis are the techniques that hold up under review. Practise on flawed samples with known answers — PAUT Level II practice questions cover the theory, but sizing skill only comes off blocks.

Prerequisites, hours and experience: the real timeline

The route is UT Level I, then UT Level II, then phased array. SNT-TC-1A recommends 40 hours of formal instruction at UT Level I and a further 40 at Level II, with method experience accumulating alongside. ASNT's own Phased Array Level II course is specified at 80 hours, and most US written practices adopt that figure for the technique endorsement. That is 160 classroom hours before anyone can call you a credentialed PAUT Level II, and the hours have to be documented.

Experience is the longer pole. Written practices commonly require three months of supervised phased array work on top of the underlying UT method experience, and that supervision has to be documented by a Level III certified in ultrasonic testing. Companies with no UT Level III on staff cannot sign the certification at all, which is why a training purchase frequently fails to unblock a program that actually has a certification-authority problem rather than a knowledge problem.

Two details cause most rejected certifications. First, the written practice must name phased array explicitly and state its training hours and experience requirement; a document that says only "UT" does not cover the technique endorsement. Second, the specific examination has to be written against the equipment and procedures the technician will actually use — a generic phased array exam bought off the shelf does not survive a client audit. Written practice development closes both gaps in a single pass.

What a PAUT Level II is expected to do unsupervised

The employer test is not whether you can operate the instrument. It is whether you can walk up to a joint you have never seen, read the weld prep from the drawing or the welding procedure, choose the probe and wedge, build the focal law, produce a scan plan, calibrate TCG and the encoder, acquire the data, and then interpret and report it without a Level III standing behind your shoulder correcting the setup.

That is the entire reason for the pay gap. A conventional UT Level II is paid to execute a technique somebody else designed. A phased array Level II is paid to design the technique on site and defend it to a client inspector who may know the governing standard better than they do. Companies that treat phased array as a button on the machine end up dispatching a Level III to every job, which erases the economics of owning the equipment in the first place.

The functional checklist contractors actually use has four items: build a setup from scratch in under twenty minutes; explain your coverage from the scan plan without notes; produce a report another technician can reproduce from the data file alone; and name the specification clause you are working to. A technician who clears all four moves permanently to the top of the call-out list, and the call-out list is where the money is.

The pay differential, in published numbers

ZipRecruiter's US aggregate for "NDT Level II" stood at $61,725 a year on 13 February 2026, with the 25th percentile at $48,000, the 75th at $71,500 and the 90th at $90,000. Its aggregate for "Phased Array Ultrasonic Testing" stood at $87,706 on 22 June 2026 — about 42 percent above the general Level II figure, and above the 75th percentile of that band entirely. The premium is not a rumour; it is visible in the posting data.

Read title-matched aggregates carefully before quoting them. The same source's "Phased Array Technician" listing averaged $70,578 on 20 August 2026, and its "PAUT Technician" hourly listing averaged $21.53 with a median of $19.71, because those two titles sweep in helper, trainee and assistant postings that are not certified phased array work at all. The distance between $21.53 an hour and $87,706 a year is the distance between a job title and the certification standing behind it.

The differential compounds upward. Glassdoor's US aggregate for ASNT NDT Level III sat at $100,205 a year in 2026, and a UT Level III who is also the phased array authority for a company sits above that, because they own procedure approval as well as interpretation. NDT Level II salary and NDT Level III salary break the bands out by region and industry. For a scoped cohort on your own instruments and procedures, request a demo or a quote.

Frequently asked

Is PAUT a separate ASNT method or a technique under ultrasonic testing?

It is a technique under the ultrasonic method. There is no standalone ASNT phased array method certification; your certificate reads ultrasonic testing, with phased array recorded as a technique endorsement carrying its own training hours, its own practical demonstration and its own specific examination. This matters when you change employers, because the new written practice has to recognise the endorsement before your phased array work is covered.

How long does it take to go from UT Level II to working phased array unsupervised?

The instruction is 80 hours. The experience most written practices require on top of it is three months of documented supervised phased array work. In practice, technicians who scan daily reach genuine independence — building setups and scan plans without help — around six months in, and technicians who touch the instrument once a month do not reach it at all. Frequency of use matters more than elapsed time.

Should I train on the same instrument my employer owns?

Train on the technique, verify on the instrument. Focal law construction, TCG, scan plan geometry and sizing transfer across every manufacturer. Menu structure, setup file handling and calibration wizards do not. A course that teaches only one interface leaves you stranded the first time a client supplies different equipment, so the useful sequence is generic technique first, then a short familiarisation on the specific instrument you will carry.

Is a practical examination required for a phased array endorsement?

Yes, where your employer's written practice follows SNT-TC-1A. The practical requires you to demonstrate the technique on a specimen and report the results, and it is scored alongside a general and a specific written examination. A written practice that omits the practical for phased array is the finding an auditor writes up, which is why the endorsement should be added to the document deliberately rather than assumed under the existing ultrasonic entry.

Can Atlantis deliver phased array training on our own equipment and procedures?

Yes. On-site delivery uses your instruments, your calibration blocks, your written practice and your job geometries as the worked examples, so exam preparation doubles as a program review. That combination also produces the specific examination and the documented hours your written practice needs. Delivery is affordable, accessible and fully customizable to your shift pattern — request a demo or a quote for a scoped cohort.