TOFD vs Phased Array UT for Weld Inspection: Which Wins

The Question Every Fabrication Shop Eventually Asks

A pressure vessel fabricator in Texas gets a spec calling for volumetric weld examination on a Category A longitudinal seam. The engineer of record's note says "UT in lieu of RT per ASME Section V Article 4, Mandatory Appendix III acceptable." Now someone has to decide: run conventional phased array ultrasonic testing (PAUT), or bring in time-of-flight diffraction (TOFD), or run both. This decision gets made hundreds of times a week across pipeline tie-ins, pressure vessel shops, and refinery turnarounds, and it's usually made by habit rather than by matching the technique to the actual inspection requirement. TOFD and phased array UT are not competing technologies where one simply beats the other — they solve different problems, and the shops that understand the difference get better detection, faster throughput, and fewer code compliance headaches than the ones that default to whichever technique their last equipment purchase happened to include.

How TOFD Actually Works

Time-of-flight diffraction uses a pair of angled-beam transducers positioned on either side of the weld, one transmitting, one receiving, both fixed at a set separation (the probe center spacing, or PCS) calculated from the material thickness and the transducers' refracted angle. Rather than relying on reflected amplitude from a flaw face — the way conventional pulse-echo UT does — TOFD detects the diffracted energy that comes off the tips of a discontinuity. Every crack, lack of fusion, or planar flaw has two tip diffraction signals, and the time delay between them, converted through the known sound velocity in the material, gives an extremely precise through-wall depth and height measurement.

This tip-diffraction principle is TOFD's defining strength: it is, flaw for flaw, the most accurate volumetric method available for sizing the through-wall height of a planar discontinuity — typically within a fraction of a millimeter when properly calibrated per ASTM E2373 or the applicable code case. That precision is exactly what fitness-for-service calculations under API 579-1/ASME FFS-1 need. When an integrity engineer is deciding whether a crack-like indication found during a turnaround inspection is acceptable to run to the next outage or needs immediate repair, the through-wall height TOFD reports is the number that drives that calculation. A phased array sizing that's off by a millimeter or two can be the difference between "monitor" and "shut down and repair."

TOFD's Structural Weaknesses

TOFD has two well-documented blind zones that every technician running it has to actively manage. The near-surface dead zone, caused by the lateral wave and its associated ringing, can mask indications close to the scanning surface — typically the first few millimeters of material, depending on frequency and probe configuration. The zone near the back wall can suffer from a similar effect where the back-wall signal masks indications right at the far surface. Neither zone means TOFD "can't see" flaws there — it means a single TOFD scan pass isn't reliable there, which is why properly written TOFD procedures under ASME Section V Article 4 Appendix III call for a supplementary technique (typically a manual pulse-echo or phased array pass) to cover both dead zones. A TOFD scan run without that supplementary coverage, submitted as a standalone "complete" volumetric exam, is a procedure gap that shows up in exactly the two places most likely to actually matter — the root and the cap.

TOFD is also weak on detecting and characterizing volumetric flaws like porosity and slag inclusions — it's built to find and precisely size planar, crack-like flaws, and it tends to underreport or poorly characterize rounded indications compared to phased array's amplitude-based detection. And TOFD image interpretation is genuinely difficult; a D-scan display full of overlapping hyperbolic diffraction curves takes real training to read correctly, and mistaking geometric reflectors (weld cap, counterbore, root geometry) for flaw indications is one of the most common errors new TOFD analysts make.

How Phased Array UT Actually Works

Phased array uses a multi-element transducer (commonly 16, 32, 64, or more elements) where each element's pulse timing is individually controlled to steer and focus the ultrasonic beam electronically — sweeping through a range of angles (a sectorial or "S-scan") without moving the probe, or focusing at multiple depths within a single scan. This gives phased array two capabilities pulse-echo conventional UT never had at this resolution: full-volume coverage of a weld cross-section in a single probe position, and encoded, recorded data that can be reviewed, re-analyzed, and archived exactly as the scan captured it.

Phased array's real advantage over TOFD is coverage and amplitude-based flaw characterization. Because it's fundamentally a pulse-echo technique (with the added electronic steering), PAUT detects flaws based on reflected amplitude, the same physical principle inspectors have used since the 1960s, just with vastly better resolution and coverage. That makes it strong at detecting and sizing volumetric flaws — porosity, slag, inclusions — where TOFD struggles, and it doesn't carry TOFD's near-surface and back-wall dead zones in the same way, because the sectorial scan covers multiple angles through the full weld volume.

Phased Array's Real Limitations

Sizing accuracy for planar flaw height is PAUT's weak spot relative to TOFD. Amplitude-based sizing techniques (6dB drop, 20dB drop, maximum amplitude) are inherently less precise for through-wall height than TOFD's tip-diffraction timing, especially for tight, tightly-closed cracks that don't reflect strong amplitude even though they diffract measurably. This is exactly backwards from what many junior technicians assume — PAUT's rich, colorful S-scan display looks more "complete" and authoritative than TOFD's sparse diffraction lines, but for the specific job of telling an integrity engineer exactly how tall a crack is, TOFD is more trustworthy.

PAUT procedures also require more calibration rigor to get right — wedge delay and sensitivity calibration across every angle in the sweep, verified against reference reflectors in a calibration block matching the actual material and geometry, per ASME Section V Article 4 or AWS D1.1 Annex. A shop running PAUT with a calibration block that doesn't match the production weld's thickness and curvature is generating S-scans that look clean but are miscalibrated in ways that won't show up until a missed flaw fails in service.

Head-to-Head: Which Wins for What

• Precise through-wall crack sizing for fitness-for-service: TOFD wins clearly. This is its designed purpose.
• Detecting porosity, slag, and volumetric flaws: Phased array wins. Amplitude-based detection is built for this.
• Near-surface and root/cap coverage on a single pass: Phased array wins, due to TOFD's dead zones.
• Scan speed on long, straight seams (pipeline girth welds): TOFD wins — a single linear scan covers the full weld length efficiently, which is why TOFD dominates pipeline construction inspection under API 1104 Appendix A.
• Complex geometry (nozzle-to-shell welds, branch connections): Phased array wins — its electronic beam steering adapts to geometry far more easily than TOFD's fixed-PCS probe pair setup.
• Permanent, code-defensible archived record: Both techniques produce encoded, reviewable digital data when run correctly, a major advantage over conventional manual UT for both. This is precisely the kind of dataset that belongs in a structured NDT reporting software system rather than a folder of loose scan files, because a client re-review request six months after a turnaround needs the original raw data, calibration records, and technician certification tied together, not scattered across USB drives.

The Combined Answer Most Codes Are Moving Toward

The honest, code-literate answer to "TOFD vs. phased array, which wins" is increasingly "both, run together." ASME Section V Article 4 Mandatory Appendix III and most major pipeline and pressure vessel specifications now explicitly favor a combined TOFD-plus-PAUT technique for critical welds: TOFD provides the primary volumetric coverage and precise sizing backbone, PAUT provides the near-surface, root, and cap coverage TOFD's dead zones miss, plus better characterization of volumetric indications. This combined approach isn't redundancy for its own sake — the two techniques' blind spots and strengths are almost perfectly complementary. A crack sitting right at the root, which TOFD's near-surface dead zone might mask, is exactly where a focused PAUT sectorial scan is strongest. A tight, tip-diffracting crack mid-wall that PAUT's amplitude-based detection might undersize, TOFD nails within a fraction of a millimeter.

For shops making equipment and staffing decisions, this means the real question isn't "which technique should we standardize on" — it's "do we have qualified technicians and instrumentation for both, and do we know which code clause on which weld class calls for which combination." That's a training and procedure-writing problem as much as an equipment problem. Building that dual competency is exactly the kind of gap Atlantis NDT's NDT training programs and ASNT Level III consulting engagements address — reviewing a shop's actual weld classes against their current procedures and identifying where a TOFD-only or PAUT-only approach is leaving a documented gap against the governing code.

Cost, Throughput, and the Practical Decision

Equipment cost narrows the choice less than it used to — a modern phased array flaw detector with TOFD capability built into the same instrument is now standard across major manufacturers (Olympus, GE, Zetec, and others), so most shops aren't choosing between two separate capital purchases anymore. The real cost differential is training and procedure development time. TOFD interpretation training runs longer and requires more supervised scan-reading hours to reach competent independent analysis than PAUT's more visually intuitive S-scan display, even though both require substantial documented experience hours under ASNT SNT-TC-1A guidelines before a technician should be signing off UT Level II or III reports independently.

Throughput favors TOFD on long, geometrically simple welds — a pipeline contractor running hundreds of girth welds a day needs the linear scan speed TOFD provides. Throughput favors phased array on complex, one-off geometry — a refinery turnaround crew inspecting a dozen different nozzle configurations in a shutdown window needs PAUT's adaptability more than TOFD's raw scan speed on any single weld.

A Field Scenario: Getting the Decision Right on a Tie-In Weld

Consider a realistic case: a midstream operator is tying a new 24-inch, 0.500-inch wall API 5L X65 pipeline segment into an existing compressor station header. The girth weld is a single joint, but it's a high-consequence location under 49 CFR 192 — near a road crossing, in a Class 3 location — so the inspection stakes are higher than a typical cross-country girth weld. The crew has both TOFD and phased array capability on the truck. Which do they run, and why?

The right answer starts with what's actually being verified. API 1104 Appendix A, the standard governing this weld's acceptance, permits both UT techniques but requires specific scanning coverage and reference reflector calibration appropriate to the technique used. Given the high-consequence classification, most experienced crews run a combined scan: TOFD as the primary volumetric pass for its superior sizing of any planar indication (since a missed or undersized crack-like flaw at this location carries outsized consequence), plus a supplementary phased array pass specifically covering the root and cap regions where TOFD's dead zones are weakest. This isn't overkill — it's the standard combined approach API 1104 Appendix A increasingly expects for Class 3 and Class 4 locations, and it reflects exactly the complementary logic described above: TOFD's sizing precision where it counts, phased array's near-surface coverage where TOFD structurally can't see.

A crew that ran TOFD alone here, trusting its single linear scan pass for speed, would have a documented gap at the root — precisely where a girth weld hydrogen crack or a root-pass lack of fusion is statistically most likely to occur, since root passes are typically laid with less heat input and are more susceptible to fast cooling and hydrogen entrapment than fill passes. That gap might never surface in a clean weld. It becomes a serious liability the one time it doesn't.

Common Procedure Mistakes That Undermine Either Technique

Beyond the TOFD-vs-PAUT decision itself, a handful of procedural errors show up repeatedly in audits and account for more missed or misreported flaws than the choice of technique itself:

• Calibration block mismatch: using a reference block with a different wall thickness, curvature, or material grade than the production weld, which shifts both TOFD's time-based depth calibration and PAUT's amplitude-based sensitivity in ways that don't show up until a real flaw is missed or mis-sized.
• Skipping delayed re-scan on hydrogen-susceptible material: higher-carbon-equivalent steels welded without confirmed preheat can develop hydrogen cracking hours after the original scan — a single same-day UT pass, TOFD or PAUT, won't catch a crack that hasn't formed yet.
• PCS drift on TOFD: probe center spacing that isn't re-verified after a probe swap or wedge wear shifts the entire depth-timing calibration, producing systematically wrong height sizing without any obvious sign in the display.
• Treating S-scan color intensity as a sizing tool: PAUT's visually rich sectorial display tempts less experienced analysts into eyeballing flaw size from color/amplitude alone rather than applying the actual specified sizing technique (6dB drop, maximum amplitude, or a code-specified alternative).
• Under-qualified analysts signing off complex geometry scans: nozzle and branch connection PAUT interpretation requires substantially more supervised experience than straight-seam scanning, and ASNT SNT-TC-1A written practices should reflect that distinction in an inspection body's qualification requirements, not treat all PAUT experience hours as interchangeable.

Every one of these is a procedure-discipline problem, not a technology limitation — which is exactly why the technique itself matters less than how rigorously it's calibrated, documented, and matched to the actual code requirement for the joint being examined.

What This Means for Procedure Writers and QC Managers

The practical takeaway for anyone writing or approving an inspection procedure is to stop treating "TOFD vs. PAUT" as a single site-wide standard and start treating it as a per-weld-class decision driven by three questions: what does the governing code actually specify or allow (ASME Section V Article 4, API 1104, AWS D1.1), what's the consequence of a missed or mis-sized flaw on this specific joint (does it feed a fitness-for-service calculation, or is it a pass/fail acceptance check), and what's the joint geometry. Getting this decision right the first time — rather than discovering during a client audit that the technique specified doesn't match what the code actually requires for that weld category — is the difference between an inspection program that holds up under scrutiny and one that generates costly rework and re-inspection. It's also exactly the kind of procedure gap that surfaces during a structured code review, and closing it before a job starts costs a fraction of what it costs to fix after a failed audit.

Atlantis NDT Products & Services

Atlantis NDT pairs field expertise with software: NDT inspection management software — Atlantis ERP, a digital twin platform for asset integrity, and NDT reporting software. Build your team with NDT training & certification (ASNT SNT-TC-1A) and ASNT certification pathways, or bring in ASNT Level III consulting. Affordable, accessible, fully customizable — book a free consultation.