Reviewing an aerospace NDT report against the technique, the operator and the process record

Why aerospace NDT reports fail differently from industrial ones

In refinery work an inspection report is evidence about a specific piece of equipment. In aerospace it is evidence about a process. The part either conforms or it does not, but the finding is only usable if the process that produced it was the approved process, unchanged, run by a qualified operator, on a system whose performance was verified that day. Nadcap defines special processes as those whose results cannot be confirmed by later inspection and whose parameters are influenced by component geometry — which is precisely why the record, and not the readout, carries the conformity claim.

The failure modes follow from that definition. An industrial report fails because a number is wrong. An aerospace report fails because a link in the chain of evidence is missing: a technique sheet at the wrong revision, an operator whose annual vision examination lapsed three weeks before the shift, a penetrant system performance check with no recorded result for that day. The part may be perfectly good. The report cannot prove it, and a prime cannot accept it. The general standard is covered in what makes an NDT report defensible.

The frozen-process concept, stated precisely

Nadcap does not, by itself, freeze anything. Nadcap accreditation certifies that a supplier does what its customers and the applicable specifications require. The freeze comes from the prime, through quality clauses in the purchase order: the prime approves a specific written NDT procedure and technique for a part number, and the supplier may not change a listed parameter without going back for re-approval. Typical clause language requires the source performing NDT to hold a written procedure approved by the prime customer, and requires the certification supplied with the parts to indicate that customer approval.

The audit then verifies the freeze held. AC7114 and its method checklists ask whether the supplier followed what the customer imposed — not whether the parameters chosen were technically ideal. A supplier that improves its technique and gets better results has still broken the process if the improvement was not approved first. This is the mental model that catches most escapes: on a frozen process, an unapproved improvement and an unapproved degradation are the same finding, and both stay discoverable from the records for years afterwards.

For the reviewer, the practical question is which document is the frozen one. The customer-approved procedure, the technique sheet derived from it, and the shop traveller are three artefacts that drift apart over time. When the traveller cites a technique sheet revision the approved procedure does not authorise, the shop is running an unapproved process while every individual document looks valid on its own. The scope limitations of your Level III's method authority determine who can even sign the change.

Check one: technique sheet against part number and revision

Start at the part, not at the procedure. Take the part number and engineering revision from the traveller, find the technique sheet the report cites, and confirm the sheet names that part number at that revision. Then confirm the sheet's own revision is the one the customer approved, and that the approval is on file with a date preceding the examination. Three matches, in that order. A break at any one of them means the examination was performed to a technique the customer has not accepted for that configuration.

The second pass is parameter by parameter. Every value the technique sheet fixes — penetrant family and type, dwell time, emulsifier contact time, magnetising current and direction, transducer frequency and reference standard, kilovoltage, source-to-film distance, film or detector class, IQI designation — must either appear in the report's process record or be traceable to an equipment log for that shift. A technique sheet specifying a value that appears nowhere in the record is an unverifiable claim, and a source inspector treats it as one.

Check two: NAS 410 currency on the date of examination

NAS 410 is the aerospace personnel qualification standard, and it moved recently. AIA and Accuris released NAS410 Revision 6 on 1 December 2025, introducing a new Appendix D that creates an additional route to Level 3: Level 2 personnel accumulate points for designated NDT activities, and once the point thresholds are met across the activity categories they become eligible to sit the initial Level 3 examination. Suppliers whose written practice still cites Revision 5, with no documented transition plan, are carrying a finding into their next audit.

The report-level check is narrower and unforgiving. For each operator who signed, confirm the method and level, confirm the certification was in force on the examination date rather than the report date, and confirm the annual vision examination — near vision acuity and colour or contrast differentiation as the method requires — was current on that same date. Fluorescent methods add the requirement that the operator's dark adaptation and the booth conditions were controlled. A lapsed vision record invalidates every examination the operator signed inside the lapsed window.

Level matters as much as currency. A Level 1 performs specific calibrations and tests under direction and records results; the accept-or-reject decision on a part belongs to Level 2. A report where a Level 1 signature appears in the disposition field is a finding regardless of whether the call was correct. Employer-based qualification systems under SNT-TC-1A follow the same logic, and suppliers running both frameworks must keep the boundary visible in the record itself.

Check three: penetrant process control records

Penetrant is where auditors find the most, because the process carries the most daily checks. Materials must come from the qualified products list under AMS 2644; Nadcap verifies compliance with the customer's specification rather than approving materials itself. Magnaflux's published reading of AC7114/1 puts the system performance check in Section 5.9, requiring verification at least prior to use and at least once per day where the system runs daily, against a baseline the facility established for each known defect standard and material on that specific system — with the result recorded, not merely performed.

The known defect standards themselves are controlled items. The same reading of AC7114/1 Section 5.10 requires maintenance procedures that ensure adequate cleaning between uses and detect physical changes making a panel unsuitable, a degradation check at least annually, and a stated tolerance of plus or minus 30 % of the baseline measurement for that check. A panel outside tolerance invalidates the system checks performed against it, which is how a single missed annual check propagates into a year of unsupported examinations across every part that ran on that line.

The periodic checks around it come from the aerospace penetrant practice. Magnaflux's published overview of ASTM E1417 lists water content of Method A water-washable penetrant checked weekly by refractometer against a 5 % ceiling, black light intensity daily and black light integrity weekly, visible light intensity weekly, examination area cleanliness daily, and water wash pressure and temperature at the start of each working shift — with penetrant brightness moved from quarterly to semi-annual in the current edition. The reviewer matches examination dates against those logs.

Check four: magnetic particle field verification

Magnetic particle claims are claims about a field that no longer exists by the time anyone reads the report, so the verification artefact carries the entire evidential weight. ASTM E1444/E1444M is the aerospace practice — E1444/E1444M-22a in the version checked — and it requires test parts or shims manufactured to specified requirements to verify field direction, strength and balance where multidirectional magnetisation is used. The record must show which artefact was used, when it was used, and what indication pattern it produced, tied to the specific equipment and the specific shift.

Demagnetisation is a separate acceptance criterion, not a housekeeping step. The practice sets a residual field ceiling read with a field indicator or strength meter after demagnetisation — 3 G in absolute value anywhere on the part, in the version checked. Parts shipped without a recorded residual reading carry an open claim into the customer's assembly line, where a magnetised part attracts debris in service. Reviewers check for the reading, the instrument identity, and the calibration status of that instrument on that date.

Around those two sit the supporting logs: ammeter and timer accuracy verification, light meter calibration, bath concentration and contamination checks, and the current or field values the technique sheet fixed. Each must exist for the shift the parts ran, not merely for the month. Auditors sample by date, and a log carrying monthly entries cannot answer a question about a Tuesday.

Check five: radiographic image quality evidence

Radiography leaves the best evidence and offers the most opportunities to lose it. ASTM E1742/E1742M covers minimum requirements for radiographic examination and requires hole-type IQIs to Practice E1025 or the alternate design in its Annex A1, or wire-type IQIs to Practice E747, with wire IQIs correlated to hole-type quality levels under E747. IQIs must be procured or fabricated to those requirements with a manufacturer's certification of compliance for alloy and dimensions, and the design, material designation and thickness identification must appear in the written procedure or on a drawing the procedure references.

The report-level check is that the IQI visible on the image matches the IQI the technique sheet specified, that the required hole or wire is discernible, and that the shooting sketch places the IQI where the procedure requires it. Then density or grey-level readings, film or detector class, and the identification markings on the image itself. Digital adds a layer of its own: AC7114/6 covers digital radiography using a digital detector array and AC7114/8 covers computed radiography using photostimulable luminescence, each carrying its own performance evidence.

Remote interpretation carries a checklist of its own — AC7114/9 covers remote interpretation of radiographic films and digital images — because the interpreter, the display and the viewing conditions become process variables the moment the person reading the image is not in the room where it was made. Reports naming an interpreter without recording the workstation identity and the controlled viewing conditions leave that variable entirely undocumented.

What a prime's source inspector challenges first

Source inspectors work from probability, and they know where records break. The first request is almost always the personnel file for the operator who signed the report — certification, level, method scope, and the vision examination date. It costs one minute and it fails often, because vision examinations run on an annual clock that drifts out of sync with certification cycles. The second request is the technique sheet at the revision cited, checked against the part revision on the traveller.

The third is the process control record for the shift, pulled by date rather than by month. The fourth is the calibration status of every instrument named in the report on that date. Only after those four does anyone look at the indications themselves. A supplier that answers the first four inside ten minutes, from the report package rather than from a search of the quality system, ends the visit early. A supplier that cannot has already shown the inspector where to dig.

The escape scenario is worse than the finding scenario by an order of magnitude. A finding is closed with root cause and corrective action inside the response cycle. An escape means non-conforming or unverifiable hardware reached the prime, which triggers containment across every lot processed under the same broken link — usually months of production, sometimes across multiple part numbers. The choice between correcting a report and re-examining hardware is covered in third-party review versus re-inspection.

How the review fits the Nadcap cycle

Nadcap runs on a schedule that rewards clean records. PRI administers the programme on behalf of the subscribing primes; the normal accreditation cycle is 12 months, extending to 18 months and then to 24 months on merit, with merit gated on prior accreditations of at least 18 months each and no major non-conformances. After an audit the supplier has 21 calendar days for the initial response to findings and seven calendar days for each subsequent response, before a task group of industry members reviews the pack and votes on accreditation.

That arithmetic makes pre-audit report review a scheduling decision as much as a quality one. Every major finding avoided protects the merit interval, and every merit interval protects a year of audit cost and disruption. Independent review of live report packages, against the checklists your accreditation scope actually carries, finds the broken links while they remain correctable — before an auditor samples the same shift and turns a paperwork gap into a non-conformance report.

Atlantis performs independent NDT report validation for aerospace suppliers, and supplies outsourced ASNT Level III authority where a written practice or a procedure needs a signature your organisation does not hold internally. Affordable, accessible, fully customizable — request a quote or a sample review.

Request a consultation