What a UT Level III Actually Signs Off On, and Where That Authority Stops

What the Level III Signature Actually Covers in Ultrasonics

In ultrasonics the Level III approves four things, and they are separable. The written procedure, which names the referencing code, the acceptance criteria and the range of every essential variable. The technique sheet or scan plan, which pins that procedure to a specific joint geometry, thickness, material and access condition and shows that sound actually reaches the surfaces you claim to have examined. The calibration basis: the block, the reference reflector, the DAC or DGS reference level, and the transfer correction between block and part. And the personnel: the general, specific and practical examinations that put a certified Level II in front of that procedure with defensible evidence they can run it. A signature on the procedure alone, with no control over the other three, is decoration.

What the Level III does not do is take over the Level II's job of performing and interpreting the examination, or the API 510, 570 or 653 inspector's job of authorising and signing an in service inspection. Atlantis supplies NDT technical authority: written practice, procedure development and qualification, personnel certification within that practice, technical representation when a client or registrar questions the programme, and independent review of inspection data. It does not act as your inspector of record and it does not audit your process safety management programme.

The practical consequence for a buyer is a traceability test you can run yourself. Take any accepted or rejected weld and walk it backwards: report, technique sheet, procedure revision, Level III approval date, and the technician's certification status in that method and technique on the day of the scan. In most Houston shops the chain breaks at the technique sheet, because it is the one document nobody owns. Engineering assumes QA writes it, QA assumes the Level III issued it, and the crew works from a copy that has no revision number on it at all.

Straight Beam, Angle Beam, and the Limits of a Single Angle

Straight beam work carries more of the load than its reputation suggests. It measures wall, finds laminations and inclusions in plate and forgings, checks bond in clad and overlay, and, critically, clears the scanning surface before any angle beam scan. A laminar reflector sitting in the path of a shear wave truncates that beam and produces a clean, quiet, entirely false result: the operator sees nothing, and nothing is exactly what a lamination guarantees. Procedures require the straight beam pre scan for that reason. In practice it is the first step dropped when a turnaround falls behind, and the report gives no clue that it was skipped.

Angle beam brings its own geometry argument. Lack of fusion on a prepared bevel is a specular reflector: it returns energy strongly when the beam strikes it near normal and weakly otherwise. On a conventional single V bevel a forty five or sixty degree shear wave usually has a workable angle of incidence. On narrow gap, J bevel or near vertical fusion faces, common in heavy wall vessel and subsea work, the favourable angle may be seventy degrees, a tandem arrangement, or a phased array sectorial sweep that covers the face across a range. A procedure that offers one angle for all joint preparations has quietly decided which flaws it will not find.

The other half of interpretation is knowing what is not a flaw. Root geometry, counterbore, weld crown, misalignment, mode converted signals and surface waves all produce repeatable indications at predictable sound paths. The procedure must state the rules for distinguishing geometry from a discontinuity, usually by comparing sound path and surface distance against the joint drawing and by probe movement characteristics. Where those rules are absent, two competent technicians produce two different reports on the same weld, and the client learns about the disagreement at the worst possible moment.

DAC, DGS and the Transfer Correction Nobody Records

A distance amplitude correction curve is built on side drilled holes in a basic calibration block whose material, product form, heat treatment, surface finish, thickness range and curvature represent the part. Each of those qualifiers is a control, and each is routinely violated: a curve built on a normalised carbon steel block used on quenched and tempered material, or a flat block used on a small diameter pipe where curvature changes the contact and the beam. DGS avoids block dependency by using theoretical curves tied to the probe's own parameters and an equivalent flat bottom hole, but it assumes the probe still behaves like its published diagram. A worn wedge, a chipped element or a hot shoe means it does not.

Transfer correction is the bridge between block and part, and it is the single most under recorded number in ultrasonic testing. Measured pitch and catch with two probes on both surfaces, it captures the difference in attenuation and coupling. The arithmetic is unforgiving: six decibels is a factor of two in amplitude. An uncorrected six decibel loss puts a fifty five percent DAC indication on screen at around twenty seven percent, comfortably below a recording threshold set at fifty percent. Nothing in the report will ever show why the indication was not written down, because the transfer correction field is blank.

Treat that number as a per component, per shift measurement rather than a one off. Wedge wear, couplant substitution, surface temperature, paint, blast profile and weld crown condition all move it. A procedure that requires transfer correction and a report format that has nowhere to record it is a documented control with no evidence behind it, and any competent auditor will spot the gap in the first five reports they open.

Thickness Readings That Read Thick: the Temperature Trap

Sound velocity in carbon steel falls as temperature rises, by roughly one percent for every hundred degrees Fahrenheit. The instrument does not know that. It multiplies the measured transit time by whatever velocity it was calibrated with, usually on a step wedge at ambient, so a hotter part with a lower true velocity displays more thickness than it has. On a six hundred degree line, that is about five percent of over reading. On a nominal three eighths inch wall it is roughly twenty thousandths, which in many services exceeds a full year of real corrosion. The direction of the error is the dangerous part: it always makes the wall look healthier than it is.

The second trap is arithmetic rather than physics. Corrosion rates under API 570 and API 653 are computed from thickness differences, and a difference between two readings each carrying five thousandths of repeatability is not a reliable estimate of a four thousandths per year loss rate. The rate is then used to set the next inspection date, so measurement noise propagates directly into an interval decision. The controls are unglamorous and effective: permanently marked condition monitoring locations, the same probe type and instrument setup, a written instruction on couplant, surface prep and mode, and a standing rule that a corrosion rate is never revised on a single pair of readings.

Then there are the instrument level errors that survive because the display looks clean. Single echo mode adds the paint or coating to the wall; echo to echo mode removes it but needs a usable second backwall echo, which heavily corroded internal surfaces often do not give. Thin wall readings can double when the gate latches the second backwall, and heavily attenuating material can make the instrument lock onto a later echo entirely. A Level III's job is to write the mode selection rule into the procedure rather than leaving it to whichever menu the last technician left the instrument in.

Certification in UT Is Not Certification in Every Ultrasonic Technique

SNT-TC-1A and CP-189 both certify by method. Ultrasonic testing is the method. Phased array, time of flight diffraction, encoded corrosion mapping, immersion, guided wave and high temperature techniques are techniques within it, and none of them arrive automatically with a UT Level II certificate. If your written practice does not define a technique, set out its additional training and examination content, and mark the resulting limitation on the certificate, then every job performed with that technique rests on a certification that never contemplated it. That is not a theoretical exposure: it is the finding that most often invalidates completed work retrospectively.

Limited certifications cut the other way too. A technician whose practical examination was a thickness survey holds evidence of competence in thickness measurement, not weld interpretation, and the certificate should say so. Where it does not, the dispatcher has no way to know, and the sales team will happily promise weld scanning. An auditor who pulls that practical examination has a finding whether or not the field work was sound, because the record does not support the scope the company sold.

Scheme boundaries matter as much as technique boundaries. Houston shops feeding the aerospace and space supply chains around Clear Lake work to NAS 410 and the prime's own requirements, which handle the Level 3 role and examination administration differently from SNT-TC-1A. Work that crosses to Canada or Europe usually needs ISO 9712 or CGSB certification, which is issued centrally rather than by the employer. A US employer certificate does not travel across those borders, however good the technician is.

Houston's Industrial Base and What It Asks of a UT Level III

The Houston Ship Channel corridor, running from Pasadena and Deer Park through La Porte, Baytown and down to Texas City, is in service territory. The work is corrosion mapping, weld scanning during turnarounds, screening for damage mechanisms described in API 571, and examination in support of API 510, 570 and 653 programmes. The acceptance basis is the owner user's procedures and, where a flaw is left in service, a fitness for service assessment under API 579. Construction workmanship criteria have no standing there, yet they are copied onto in service reports constantly, because that is what the template said.

Move inland and west and the picture changes to fabrication: pressure vessel and heat exchanger shops, pipe spool fabricators along the northwest corridor and out toward Katy, and structural and modular yards. Here Section VIII of the ASME Code drives examination through Section V Article 4, AWS D1.1 governs structural welds with its own acceptance table, and the ultrasonic in lieu of radiography route, which entered the Code through Code Case 2235 and now sits inside Section VIII, demands encoded techniques with demonstrated flaw sizing. That is not the same procedure as shop weld screening and should never be the same document.

The offshore and subsea supply chain along the Ship Channel and down the coast toward Ingleside adds client specifications on top of code, frequently with their own qualification demonstrations, sizing tolerance requirements and data storage rules. LNG work on the upper Texas coast brings cryogenic materials and austenitic and nine percent nickel welds, where coarse and anisotropic structure scatters and steers the beam and a carbon steel procedure simply does not apply. A Level III who knows only Section V will be exposed at the first client audit in any of these three markets.

Audit Findings That Recur Across Houston Shops and Plants

The same handful of findings appear year after year. Technique sheets with no revision number, no approval signature and no link to a procedure revision. Calibration blocks with no material certificate, or a certificate for a different heat than the block stamp. Instrument linearity checks required by the procedure at a stated interval, with no record that any check was ever performed. Transfer correction fields left blank. Acceptance criteria taken from the construction code and applied to in service degradation. And reports from a Level II whose certification in the technique used had lapsed weeks before the scan.

One finding deserves separate mention because it is almost universal and almost never volunteered. Procedures require a calibration verification at a stated interval and after any change, and they require that if the check fails beyond the allowed tolerance, everything examined since the last valid check is void and must be re examined. Ask any contractor to produce a single record of a failed calibration check and the resulting re examination. In ten years of files, most cannot. Either no calibration check has ever drifted, which is not credible, or failures are being quietly corrected and the affected material never revisited.

None of this is fixed by buying better instruments. It is fixed by a technical authority who writes documents that can be followed, report formats that force the controls to be recorded, and an examination programme that produces technicians who understand why the controls exist. That is unglamorous work, and it is exactly what survives a client audit at eight in the morning during a turnaround.

How Atlantis Works as Your Outside Ultrasonic Level III

The engagement usually begins with a read of what you already have: the written practice, the ultrasonic procedures and their revision history, a sample of technique sheets, the calibration block inventory and certificates, and three or four completed report packages picked at random. That review produces a gap list mapped to the codes and client specifications that actually apply to your work, in priority order, with the exposures that would invalidate completed work listed first. Nothing existing is discarded; procedures are revised and extended rather than replaced, so your history stays intact.

From there the work is concrete. Procedure development and revision against ASME Section V Article 4, Section VIII, AWS D1.1 or an owner user specification. Technique sheets and scan plans for the joints you actually fabricate or inspect. Calibration block specification and transfer correction protocol. Examination writing, administration and grading for general, specific and practical examinations in the ultrasonic method and its techniques, within your written practice, with the employer retaining certification authority. Independent review of encoded data and A scan records when a client disputes a result.

Atlantis is a US and India based NDT technology and services company led by an ASNT NDT Level III in multiple methods who also holds API 653 authorised inspector certification, and it supplies inspection management and reporting software, digital twin asset visualisation, NDT training and report validation alongside consulting. Positioning is affordable, accessible and fully customisable. To discuss a Houston scope, request a consultation and a written proposal at info@atlantisndt.com.

Request a consultation