Every field an NDT report must carry, with the code that requires it

Block one: the universal header

Fourteen fields apply to every report, whatever the method. Client and job or purchase-order number. Site and physical location. Component identification with the drawing, isometric or weld-map reference. Material specification and thickness. Item or weld identification exactly as it appears on the map. Joint configuration and welding process. Heat-treatment condition. Surface condition and preparation as found. Stage of examination. Extent of examination as a percentage. Date and time. Report number. Revision. Page x of y.

Two of these carry more weight than the rest. Stage of examination — before or after post-weld heat treatment, before or after hydrostatic test — determines whether the examination satisfies the code requirement at all, and a date alone does not establish it. Extent as a percentage matters because partial examination is a contractual and code question: a report covering 10% of welds must say so on its face, or a reader will assume complete coverage.

Page numbering as x of y is the field most often skipped and the easiest to exploit. A report issued as five separate unnumbered pages can lose a page without anyone knowing. Numbered as page 3 of 5, a missing page announces itself. Every page also carries the report number and the item identification, because pages get separated in transmittal.

Block two: procedure, code edition and acceptance criteria

The results block is meaningless without the standard it was judged against. Record the procedure number and its revision — the revision matters, because procedures change and a report referencing "UT-004" without a revision cannot be tied to the technique actually in force. Where the code requires procedure qualification or demonstration, reference the qualification record.

Then the governing code with its edition and addenda, and the specific acceptance clause. Codes are revised on a cycle and clause numbering moves between editions, so a report citing an acceptance paragraph with no edition sends a reviewer to whichever edition they happen to hold. Cite ASME VIII Division 1 UW-51 or UW-52, ASME B31.3 paragraph 341 with the table applied, AWS D1.1 Clause 8 with the connection type, or API 1104 with the applicable paragraph — each with the edition.

Where a client specification tightens the code, both are recorded and the tighter of the two is identified as applied. This is where reports quietly go wrong: the examination is judged against the code while the contract required a stricter client standard, and the discrepancy surfaces during handover review when re-examination is expensive. A program audit and gap assessment finds these systematically.

Block three, part one: UT, PAUT and TOFD records

Ultrasonic reports fail on equipment and calibration detail. Record instrument make, model and serial number; search unit frequency, element size and nominal angle; wedge identification; cable type and length, because both change instrument response; couplant type and batch. Then the calibration record: block identification, reference level and gain, DAC or TCG construction, and the times of the initial calibration and every re-check. Re-checks are required at intervals not exceeding four hours, when examination personnel change, and at completion of the examination.

Add the physical conditions that determine what the technique could detect: scanning surface and its finish, surface temperature, scanning technique and directions, and coverage achieved against coverage required. Indications are recorded with location referenced to a datum, depth, length, amplitude relative to the reference level, and the evaluation against the acceptance clause.

Phased array and TOFD add a layer. The scan plan with its coverage demonstration, focal law parameters, element-activity check results, encoder calibration distance and measured error, index offsets, and — critically — identification and archive location of the raw data file. Encoded data that cannot be located twelve months later reduces an advanced examination to an unverifiable assertion. Name the file on the report and store it where the report can reach it.

Block three, part two: RT records

Radiography carries the longest technique record and the most consequential single omission. Source details first: isotope and activity, or tube kV and mA. Then geometry — source-to-film distance, source size, and calculated geometric unsharpness. Film manufacturer, type and number of films in the cassette. Screen material and thickness, front and back. Exposure time. Processing method.

The image-quality-indicator entry is the field that decides the report. Record the IQI type and designation, its placement source side or film side, the justification where film side is used, the essential hole or wire that must be visible, and confirmation that it was. Alongside it, density readings taken in the area of interest and through the IQI body, and the density variation limits applied. A radiograph without an identified and demonstrated IQI has no established sensitivity, and no established sensitivity means no established result.

Complete the record with location markers, the technique sketch showing source position relative to the weld, weld and component identification appearing in the image itself, film-side identification, and the number of exposures with overlap. Interpretation follows: each indication by type and location, the acceptance clause applied, and the accept or reject decision. Delivering all of this consistently across a crew is a workflow problem before it is a documentation problem, which is what inspection reporting software addresses.

Block three, part three: MT, PT, VT and ET records

Magnetic particle records the technique — yoke, prods or coil — with current type and amperage, pole or prod spacing, and evidence that two substantially perpendicular magnetisation directions were applied. Then the particle system: dry or wet, visible or fluorescent, with batch identification. The lift-test date for the yoke, referencing a verification performed at least every six months and after any damage. Lighting measured at the surface. Demagnetisation and post-examination cleaning where required.

Penetrant records surface preparation and cleaning method, penetrant family and type with batch number, application method, penetrant dwell time and surface temperature as measured values, excess removal method, developer type and batch with its own dwell time, and evaluation time. Dwell times are the field reviewers check first, because a short dwell produces a clean-looking part that was never properly examined.

Visual examination records whether it was direct or remote, viewing distance and angle, the illumination measured at the surface, and any equipment or aids used. Eddy current records frequency, probe or coil type, fill factor or lift-off, the reference standard identification, calibration and re-check, and phase and gain settings. Across all four methods the pattern is identical: record what was measured, not that it was adequate.

Block four: results, disposition and signatures

Results are stated against the acceptance clause, not in general terms. Each indication carries location referenced to a datum, dimensions, type, the clause applied, and an explicit accept or reject. An overall statement for the item follows. Where nothing was found, say so explicitly — "no recordable indications" against a named acceptance standard is a result; a blank field is an omission.

Rejects open a loop that the record must close. Disposition — repair, replace, or use-as-is under a documented concession — then the re-examination after repair, cross-referenced to the original report number and covering the repaired area plus the required margin. A file containing a reject report and no closing re-examination report is the single most common finding when a job package is reviewed at handover.

Signatures close the report. The examiner's name, certification level and method, and signature. The reviewing Level III where the code, written practice or contract requires it. The certification behind each signature must have been current on the examination date — which is exactly why certification records are kept as dated history rather than a current-status flag. Clients increasingly verify signatures against certification currency through a client portal.

Running the checklist without slowing the crew down

A checklist applied after the fact catches errors when correction is most expensive — the technician has left site, the part has moved, and the missing dwell time cannot be recovered. Applied at capture, the same checklist costs seconds. The mechanism is field-level validation on the form the technician fills: required fields that cannot be left empty, equipment identifiers pre-populated from the equipment register at issue, batch numbers selected from currently valid consumables, and calibration re-check prompts that fire on the four-hour clock.

This is the operational argument for structured capture over free-text templates. A word-processor template accepts any content, including nothing. A structured form knows that a magnetic-particle report requires a lighting measurement and a lift-test reference, and that a radiographic report requires an IQI designation with a placement statement. The method chosen at the top of the form determines the fields required below it.

Field work happens without connectivity, so the validation must run on the device rather than on the server. Offline field capture that validates locally and syncs later gives you complete reports at source instead of correction cycles afterwards. Atlantis runs this on an Odoo-based platform, cloud or on-premise, configurable to your report formats and acceptance standards.

Using this checklist on someone else's reports

The checklist works in reverse as an incoming-review tool. When you receive reports from a subcontracted inspection vendor, run the universal header and procedure blocks first — those two catch most problems in under a minute per report. Missing code edition, missing stage of examination, missing procedure revision and missing acceptance clause together account for the majority of reports that cannot be defended later.

Then sample the method block. For radiography, check the IQI entry and the density readings. For ultrasonics, check the calibration re-check times against the shift length. For penetrant, check dwell time and temperature. For magnetic particle, check the lighting measurement and the lift-test date. Sampling five reports across a package tells you whether the vendor's process produces complete records or whether the whole package needs review.

Where the volume is large or the exposure is high, independent report validation applies this systematically across a package and reports findings by field and by frequency, which turns a stack of individual corrections into a single conversation with the vendor. Compare how reporting systems handle these fields in the reporting software comparison, see the wider inspection software overview, and find the companion templates in /tools.

Request a consultation