PAUT, TFM and PCI: How to Specify an Advanced UT Trial

An advanced ultrasonic testing demonstration can be visually impressive while leaving the purchasing decision unresolved. A sharper image is useful only when the team understands what produced it, which inspection volume it represents, and how the result will support an acceptance decision. For asset owners and service companies, the central question is whether a proposed system produces defensible results under the constraints of the job.

This guide sets out an original trial framework for comparing phased array ultrasonic testing, total focusing method and phase coherence imaging. It is intended for inspection managers, buyers and technical reviewers in the United States and Canada. It is a procurement framework, not a qualified examination procedure or a substitute for the governing construction or in-service requirements.

What the technology names actually describe

PAUT electronically controls groups of array elements to form ultrasonic beams. Full matrix capture, or FMC, is a data acquisition sequence in which individual elements transmit and the array receives. Plane wave imaging, or PWI, uses a different acquisition sequence. TFM reconstructs focused images from acquired signals; PCI forms an image using phase coherence rather than summed signal amplitude. Acquisition and image reconstruction are therefore different decisions. Evident’s technical comparison describes these distinctions and demonstrates application-dependent tradeoffs. 1

Ask a supplier to write the complete combination on the trial sheet. “TFM included” is less informative than a record of the acquisition approach, probe, wedge, wave set, reconstruction area, scan increment and software version. The inspector who repeats the trial six months later needs that configuration, not simply the name printed on the instrument case.

Avoid turning the terminology into a league table. A technique can be valuable for one inspection task and unnecessary for another. Your trial should identify where an additional mode changes the decision, where it confirms an existing result, and where it adds review time without providing useful information.

A specific development to watch in 2026

On January 12, 2026, Evident described a TFM grid verification tool introduced in OmniScan onboard software version 6.3. Its explanation connects grid spacing, signal type, probe characteristics and material wave speed to amplitude variation, and describes experimental verification using reference reflectors. This is a documented software development, not an announcement that TFM was invented in 2026. 2

The purchasing implication is practical: request the verification record associated with the demonstrated setup. Ask which groups were checked and whether the final production configuration differs from the demonstrated one. If the supplier changes a reconstruction setting after the trial, have the responsible technical authority decide what must be repeated. A saved verification record makes that conversation much more concrete than a statement that the software supports a standard.

Start with the inspection decision

Write a one-page trial brief before inviting equipment demonstrations. Describe the component, material, thickness range, surface condition, access and expected damage. State whether the decision concerns fabrication acceptance, in-service screening, characterization of a known indication or monitoring a previously located feature. Those are different objectives and should not share an unexplained pass criterion.

Then identify the smallest set of results that would change your purchasing decision. Examples include resolving an indication close to a geometric reflector, reproducing a location after removing and replacing the scanner, or completing the required coverage within an available access window. The responsible inspection authority should establish the technical requirements and applicable acceptance criteria before results are available.

Include the people who inherit the data. An owner’s reviewer may need to reopen the inspection on another workstation. A contractor may need to train additional operators. A maintenance planner may need an unambiguous location reference. If only the demonstrator can use the result, the trial has tested a presentation rather than the complete workflow.

Build a comparison that controls the variables

Choose representative specimens and document their limitations. A smooth calibration block can demonstrate setup behavior; it cannot reproduce every surface, geometry or material condition in a plant. Keep the reference information under the control of an agreed coordinator when a blind comparison is practical. Record which targets were known to the operator and which were not.

Use a common reporting form across suppliers. Allow each to propose a technically justified configuration, but require the differences to be recorded. An identical probe is not always the fairest comparison if the proposed solutions are designed differently. What must remain comparable is the inspection task and the evidence used to judge it.

This table is a suggested evaluation structure. It does not impose code acceptance limits or a universal scoring system. Decide which failure would disqualify a solution before giving weight to convenience features.

Separate detection, sizing and interpretation

Give each task its own result field. Record whether an indication was found, how it was located, what dimensions were reported and how the reviewer characterized it. A useful detection image does not by itself establish a qualified sizing method. Similarly, a correct location does not establish that the indicated feature has been correctly classified.

For a controlled specimen with independently established reference dimensions, report the individual sizing differences. Do not hide a large discrepancy inside a favorable average. For example, three illustrative errors of minus 1 mm, plus 1 mm and plus 6 mm average plus 2 mm, but the third result deserves its own explanation. These figures are arithmetic examples, not observed equipment performance or recommended tolerances.

Make the uncertainty visible when the reference itself has limits. If a flaw boundary is not precisely known, label the reference accordingly. A disagreement between two instruments is a reason to investigate the measurement chain; it does not automatically identify which instrument is correct.

Test repeatability under realistic changes

Repeat the acquisition after removing and replacing the probe or scanner. Have a second appropriately qualified operator restore the saved configuration and explain the result. Introduce only those changes that the trial plan permits, then record what was changed and why. This tests whether the process can survive an ordinary shift change.

Add a deliberate handover exercise. Give a reviewer the dataset, component drawing and procedure reference without the demonstrator’s verbal explanation. Ask the reviewer to locate a nominated indication, identify the relevant acquisition settings and list any missing information. Treat unresolved questions as trial findings rather than correcting the record silently.

Where operating conditions vary, specify a separate trial condition instead of extrapolating from the easiest specimen. Access restrictions, curvature and material behavior may require a different setup. A limited successful demonstration should produce a correspondingly limited approval statement, with the remaining conditions listed for follow-up.

Calculate throughput from accepted work

Measure the complete workflow: preparation, setup, verification, acquisition, interruption, repeat acquisition, interpretation and reporting. Instrument scan speed is only one term. A fast pass that produces unusable data does not improve delivery, and a slower acquisition may be acceptable if it makes review substantially clearer.

For an illustrative shift, suppose a team acquires 100 m of weld data but only 80 m meets the agreed coverage and data-quality requirements after review. Record 80 m as accepted coverage and explain the disposition of the remainder. Do not present the acquired distance as completed inspection. This example defines a bookkeeping practice, not a forecast of productivity.

Separate one-time implementation costs from recurring work. Training, procedure development, customer demonstration and software configuration belong in the purchasing estimate. So do the recurring costs of data storage, review licenses, maintenance and backup. Request a quote against the same scope from each supplier so commercial comparisons do not reward omitted work.

Make data ownership part of the trial

Ask exactly which files the customer receives and which software can reopen them. Preserve acquisition settings, component coordinates and the final report together. If an export contains only rendered images, determine whether that is adequate for the intended future review. A PDF can be a useful signed deliverable while still being insufficient for reanalysis.

Perform a sample export during the demonstration. Move it to the intended review environment and reopen it there. Check units, axes, indication identifiers and the relationship between the report and the source record. Document any vendor-specific dependency, including the license needed after a subscription ends.

For projects using a digital asset model, start with stable identifiers. An attractive three-dimensional location marker is useful only if the recipient can trace it back to the relevant record. Our digital twin reporting guide develops that handover question, while the DICONDE radiography guide explains a separate imaging interoperability workflow.

Turn the trial into a usable purchasing record

End with a short decision document containing the application tested, the exact configuration, the evidence reviewed, unresolved limitations and the conditions of any approval. Attach the underlying results. Name the roles responsible for technical acceptance, training, data administration and commercial approval so implementation does not depend on one enthusiastic demonstrator.

Use three outcomes where helpful: suitable for the defined task, suitable after specified follow-up, or insufficient evidence. A supplier should be able to understand what additional evidence would change the conclusion. Keep the record available for the next software revision or materially different component instead of starting the evaluation from memory.

For the underlying method, begin with the ultrasonic testing method guide. For a broader view of current developments, read the 2026 NDT technology briefing. When seeking a service quotation, attach your trial brief or its relevant requirements to an inspection request. That gives providers a clearer basis for proposing equipment, personnel and deliverables.

Sources and reporting notes

Sources were checked on October 11, 2026. The manufacturer material supports the descriptions of its techniques and its dated software update. The trial framework, example calculations and purchasing questions are NDT Connect editorial guidance; they are not independent instrument benchmark results.

1. Evident: Comparing PAUT, PWI, PCI and TFM Technologies.
2. Evident: A New Tool for Experimental TFM Grid Resolution Verification, January 12, 2026.