The Most Common NDT Program Audit Findings, Ranked, With the Fix for Each

Why the ranking looks the same in every shop

Findings cluster because they all trace to the same two artifacts: the written practice and the certification file. The document promises hours, intervals, grading rules and signing authority. The file either evidences those promises or it does not. Everything an auditor writes is a statement about the distance between them, which is why programs in aerospace, refining and shipyard work throw up the same seven items in roughly the same order despite working to different codes.

The sequence auditors follow is stable enough to rehearse. Read the written practice and note what it committed to. Pull three or four personnel files, weighted toward technicians whose reports are already in the sample. Check every date in those files against the intervals the document promised. Then compare the signatures on released reports against the certificates behind them, and the procedure approvals against the Level III's method list. ASNT's own guidance on NDT audits describes the same pattern: verify the procedure aligns with customer requirements, then verify the practice is being followed and documented.

That order matters for preparation. Fixing records without fixing the document leaves the promise intact and the gap open. Fixing the document without fixing records creates a fresh discrepancy between a current revision and a historical file. Programs that close findings permanently work both together, which is what an NDT program audit gap assessment is scoped to do.

Finding 1: the written practice no longer describes actual practice

This heads every list because it generates the others. A written practice is a set of commitments — these methods, these hours, this grading rule, this interval, this certifying authority — and once the shop moves and the document does not, every record in the file gets measured against a standard the company stopped applying. The three common forms are a method added without a revision, hour figures nobody applies in practice, and a named Level III who left the company two years ago.

Auditors find it in the first thirty minutes without opening a personnel file. The method list in the document is compared against last quarter's job files. A phased array report with no phased array in the scope section is the finding, and it arrives before any question about training records. The same comparison works on the codes cited: a document referencing a superseded edition while the crew works to the current one is the same failure in a different clause.

The fix is a reconciliation and a cycle. Reconcile the document against the current service menu, the current certificates and the record forms actually in use, then set a review cycle with a named owner and a date. Section-by-section guidance on what each clause has to state is in the written practice walkthrough, and firms rebuilding the document rather than patching it run written practice development against their own records instead of a template carrying somebody else's method list.

Finding 2: training records that record attendance, not hours

The written practice states training hours by method and level. SNT-TC-1A offers Table 6.3.1A for the employer to consider — 40 organized training hours and 210 experience hours in the method for UT Level I, a further 40 training hours and 630 cumulative experience hours at Level II — and whatever figures the employer adopts become the number the file has to evidence. What the file usually holds instead is a certificate saying a course was completed and a sign-in sheet saying somebody attended.

Auditors ask a specific question: show me the hours. A course completion certificate with no syllabus, no hour breakdown by topic and no instructor credentials answers none of it. The same problem appears on the experience side, where "employed since 2023" is offered as evidence of hours in the method. Employment is not experience, and an auditor who has seen the substitution once looks for it everywhere in the file.

Rebuild the record format rather than chasing missing paperwork. A compliant training record carries the syllabus, hours allocated by topic, the instructor's qualification, the dates, and a signed completion. A compliant experience record carries job type, method, hours and a supervising signature. Classroom hours delivered through corporate NDT training produce the first; the second has to be captured on the job, on a form designed before it is needed. Method-by-method hour figures are broken out in the NDT training hours guide.

Finding 3: vision examinations past their annual date

This is the fastest finding an auditor writes and the one that recurs most. SNT-TC-1A (2020) requires near-vision acuity permitting the candidate to read a minimum of Jaeger Number 2 or equivalent at the distance designated on the chart and not less than 12 in. (30.5 cm), with an Ortho-Rater minimum of 8 accepted as an alternative, examined annually. Color contrast differentiation demonstrates the ability to distinguish the colors or shades of gray used in the method, at initial certification and at five-year intervals afterwards.

Two intervals on two clocks is what breaks it. Firms track the five-year color date with the certification cycle and let the annual near-vision date float, so the file ends up with a current color record beside a near-vision form dated fourteen months back. The auditor reads one date and writes the finding. There is nothing to argue about, which is why it appears in reports that otherwise find little.

The control is a calendar attached to the certification file rather than to the technician. Put the next-due date for both examinations on the certificate cover sheet where the reviewer sees it, run a lookahead monthly, and name the role authorized to administer the test in the written practice so there is no ambiguity about who books it. Where a technician wears correction, the record notes the correction — an uncorrected result on a corrected technician is the same finding with an extra step.

Finding 4: certifications expired against the recertification interval

SNT-TC-1A (2020) sets a recommended maximum recertification interval of five years for all levels, measured from the date the certifying authority signed the certification record. CP-189 works to the same five-year ceiling with tighter routes: Level I and II recertify by reexamination, or on documented experience with reexamination at least every ten years, and Level III recertification is the employer verifying that the individual's ASNT Level III certificate is current. The auditor checks a signature date against today's date, and arithmetic decides the outcome.

Two things push firms past the line. The first is a written practice citing "SNT-TC-1A" with no edition year, which leaves the interval ambiguous and lets the auditor choose. Earlier editions carried different intervals for Level I and II, so the document has to name the edition it invokes. The second is choosing the recertification route without building the evidence for it — SNT-TC-1A permits recertification on evidence of continuing satisfactory technical performance, but that evidence is the documented technical performance evaluation the Level III is required to perform periodically on Level I and II personnel. Firms that pick the evidence route and never run the evaluations have picked a route they cannot walk, and the finding lands on both clauses at once.

Run a ninety-day lookahead on every certificate in the file and treat the evaluation record as a standing obligation rather than a renewal formality. Where a technician has been out of the method, the interrupted service clause governs, and the written practice has to state the consecutive-month threshold and which examination portions are repeated. SNT-TC-1A leaves that number to the employer, so a document that never sets it has no defensible answer when the auditor asks.

Finding 5: examination records not retained

The certification record shows a composite grade. The auditor asks for the papers behind it, and the papers are gone — graded, transcribed to a summary sheet, and discarded because the sheet looked like the record. It is the single most avoidable finding on this list, and it converts a certification that was fully compliant when made into one nobody can verify.

What has to be retained is the whole set: the graded general examination, the graded specific examination, the answer keys, and the practical checkpoint sheet showing which checkpoints were evaluated on which specimen. SNT-TC-1A structures Level I and II qualification around those three written and practical parts, with the composite as the simple average and no individual part below 70 percent against a composite of at least 80. Without the papers, an auditor cannot confirm the composite was calculated the way the written practice describes, and cannot confirm the grading rule was applied at all.

Retention rules also cover who may have administered the examination. SNT-TC-1A states that examinations shall not be self-administered nor administered by a person the candidate supervises, and that in-house examination questions require Level III approval before use. The retained set has to show the administering signature and the approval trail. Firms that route Level III examination and question-bank approval through ASNT Level III consulting services keep the independence visible in the record rather than having to assert it in the meeting.

Finding 6: a Level III signing outside their qualified methods

This is where documentation findings start reaching the product. The Level III is designated in the written practice for named methods, and the designation is the basis of every procedure approval and certification record they sign. Approve a radiographic procedure on a certificate that names UT and MT and the approval is void, along with the certifications issued under it and the reports released against it. CP-189 states the requirement without room: the employer's Level III shall hold current ASNT NDT Level III certification in the method, verified by the employer at each five-year recertification.

Auditors find it by cross-reference, not by inspection. The Level III's certificate is placed beside the procedure approval sheets and the certification records, and the method columns are compared. It surfaces most often in firms that added a method for a customer — phased array or TOFD — and had the existing Level III approve a procedure a senior Level II wrote, because nobody treated the addition as a certifying-authority question.

The fix is structural. Build a method matrix with methods sold on one axis and certifying authority plus certificate expiry on the other, keep it current quarterly, and name an alternate Level III in the written practice for the methods the staff Level III does not hold. Firms weighing whether to close the gap by examination or by engagement work through it in the staff versus contract Level III comparison; the arithmetic usually favors naming an alternate, because each added method is a further ASNT Method examination and a further wait.

Finding 7: NDT performed by uncertified personnel

The most serious finding on the list, and the one most likely to be written as major, because it reaches the product directly. SNT-TC-1A requires personnel performing NDT to be qualified and certified under the employer's written practice. A report signed by someone with no current certificate in the method means the examination behind it has no qualification standing, and the customer holding that report has a problem the auditor has to record.

It rarely arrives as deliberate. The common route is an expired certificate nobody caught, a new hire working while paperwork is processed, a technician covering a method they hold at Level I under an instruction that calls for Level II interpretation, or a technician returning after a long absence with an interrupted service clause nobody applied. All four are administrative failures that produce a technical finding, which is why finding 7 usually appears in the same report as findings 3 and 4.

Gate the release, not the intent. Report release should require a current-certificate check against the method and level the report calls for, run as a control rather than a habit. Where a method has no valid certifying authority or no currently certified technician, work in that method stops until it does — that is a commercial decision made once, against a finding that costs far more. Firms that have already taken this finding usually restructure the whole control set, and a gap assessment covering the document, the files and last quarter's released reports is what tells them where the remaining exposure sits.

Major, minor, and what turns one into the other

Auditors classify findings by consequence. A minor finding is a documentation gap with no demonstrated effect on product — a missing hour breakdown, a form completed on the wrong revision. A major finding indicates the program failed in a way that could reach the product, and it can fail the audit outright. Findings 6 and 7 on this list start as major. Findings 1 through 5 start as minor and get promoted.

Promotion happens two ways. The first is accumulation: several minor findings of the same type across multiple files stop reading as isolated slips and start reading as a control that does not exist, which gets rewritten as one major against the system. The second is linkage. An expired vision record is minor on its own. An expired vision record on a technician who released a fluorescent penetrant report during the lapsed period links a documentation gap to a product decision, and the classification moves.

This is why closing findings one at a time under-performs. Corrective action that fixes the individual record leaves the control that allowed it, and the same finding returns at the next audit with an aggravating history. Corrective action that fixes the control — a lookahead calendar, a release gate, a review cycle with a named owner — closes the class rather than the instance, and auditors reading a repeat visit look specifically for which one you did.

The pre-audit pass that closes all seven

Every finding on this list is discoverable in a day by anyone willing to run the same sequence the auditor runs. Read the written practice and write down what it committed to: methods, hours, grading rule, intervals, named certifying authority with methods. Pull four personnel files spanning the methods you sell most. Check every date against the promised intervals. Compare the signatures on last quarter's released reports against the certificates behind them and the procedure approvals against the Level III's method list. What that produces is your finding list, before it is anyone else's.

Do it with someone outside the reporting line. A program reviewed by the person who wrote it returns the answers the writer already believes, and the two findings that reach the product — signing outside qualified methods and uncertified personnel — are the two an internal reviewer is least likely to write against their own signature. Independence is a requirement inside the standard for examinations, and it is a practical necessity for review. An NDT program audit gap assessment is scoped exactly this way, against the document and the records together.

Then fix the controls rather than the instances, and set the cycle. A lookahead calendar closes findings 3 and 4. A record format closes finding 2. A retention rule closes finding 5. A method matrix with a named alternate Level III closes finding 6, and a release gate closes finding 7. Finding 1 closes only with a review cycle and a named owner, which is why it sits at the top: it is the one that comes back the moment attention moves elsewhere. Firms with method coverage gaps behind findings 6 and 7 work through the coverage decision in the staff versus contract Level III comparison, because no calendar fixes a method nobody is certified to sign.

Request a consultation