ASME Section V Article 9 — Visual Examination

What Article 9 Covers, and What It Deliberately Does Not

Article 9 of ASME Boiler and Pressure Vessel Code Section V is the method article for visual examination. Its scope covers the act of looking at a surface and recording what was seen: the written procedure, the personnel and equipment, the three recognised techniques of direct, remote and translucent examination, the evaluation step and the documentation that has to outlive the job. It is one of the shortest articles in Section V and one of the most frequently misapplied, because readers arrive expecting acceptance criteria and Article 9 does not carry any.

What it excludes matters as much as what it contains. Article 9 does not define acceptable weld reinforcement, undercut, arc strikes or surface porosity. It does not define the in-service examination categories VT-1, VT-2 and VT-3, which belong to Section XI. It does not cover leak testing, which is Article 10. It does not itself certify personnel; qualification flows from Article 1 and the employer written practice built on ASNT SNT-TC-1A, ANSI/ASNT CP-189 or ISO 9712 as the referencing code permits.

The practical consequence for a QA manager is that an Article 9 procedure is incomplete standing alone. It must name the referencing code and the acceptance paragraph it serves, and it must be reviewed when that reference changes edition. Where the procedure set needs to be technically owned rather than copied from a template, that authority is the substance of ASNT Level III consulting.

Direct Visual: the 24-Inch and 30-Degree Rule People Misread

Direct visual examination under Article 9 is normally performed where access permits the eye to be placed within 24 in. (600 mm) of the surface being examined, at an angle of not less than 30 degrees to that surface. Mirrors are expressly permitted to improve the angle of vision, and magnifying aids are permitted to assist. Both numbers are outer limits. An undercut call made at 24 in. and exactly 30 degrees is a marginal call, and marginal calls are the ones that come back.

The recurring misreading is the angle datum. Thirty degrees is measured from the surface, not from the normal to the surface. At 30 degrees from the surface the examiner is looking along the weld toe at a shallow rake; at 30 degrees from the normal the view is nearly face-on, which is a far more generous geometry than the code allows. An auditor who redraws the angle on a procedure sketch and finds it dimensioned from the normal has a documented finding without ever visiting site.

The second misreading is trading distance against magnification. Article 9 permits aids; it does not permit substituting a 10x loupe at 40 in. for the stated geometry unless the procedure demonstration actually supports that arrangement. Record the geometry and the demonstration together. Examiners trained against the real geometry rather than a slide deck stop generating this finding, which is the practical point of structured NDT training and certification.

Illumination: 100 fc Is a Floor, Measured at the Surface

Article 9 sets a minimum light intensity of 100 fc (1,000 lux) on the surface to be examined, and requires that the light source, the technique used and the light level verification be demonstrated one time, documented and maintained on file. That last requirement is where the audit trail begins and where most programmes are thinnest, because the demonstration is usually done once by someone who has since left and never repeated when the lamp was replaced.

The reading is taken at the examination surface, not at the lamp. A 500 W floodlight at three metres in an open yard at dusk will not put 100 fc into the root of a nozzle-to-shell weld inside a vessel, and a meter held at chest height in the walkway proves nothing about the surface. Auditors ask for three items in sequence: the light meter calibration certificate in date, the reading taken at the point and orientation of examination, and evidence that the same lamp and stand-off distance were used in the recorded demonstration.

Where the referencing code demands more than 100 fc, more applies, and some in-service applications substitute a resolution check for a bare lux figure. Free-text light readings on paper field sheets are the ones that go missing. Capturing the reading against the weld number and the report inside inspection management software means the value is still attached to the examination three years later when a client or a regulator asks how the surface was lit.

Remote Visual: Proving Equivalence to the Naked Eye

The remote visual technique permits aids in place of direct observation — mirrors, telescopes, borescopes, fibrescopes, cameras and other suitable instruments — provided the system has resolution capability at least equivalent to that obtainable by direct visual observation. That single sentence carries the entire crawler, drone and video-probe industry. It is a performance requirement, not a permission slip, and equivalence has to be shown rather than asserted.

Equivalence must be demonstrated with the actual system as deployed: the same working distance, the same lighting, the same lens, the same display, and any video compression sitting in the signal path. A camera that resolves a fine line at 300 mm on a bench under laboratory lighting does not automatically resolve it at six metres inside a dark column through a wide-angle lens with a lossy encoder between the sensor and the screen. Frame rate, monitor size and gamma are all part of the examination system and all belong in the procedure.

The practical discipline is simple. Build a resolution target, photograph it in the field with the deployed system at the extremes of the procedure, and file that image with the report. When a remote examination is challenged two years later, the field image ends the argument. Where remote data is the only surviving record of a surface that has since been coated, lined or buried, independent review of the inspection reports is worth considerably more than a re-examination that is no longer physically possible.

Translucent Visual Examination and Where It Still Applies

Translucent visual examination is a supplement to direct visual examination, using an artificial light source placed behind the object so that light passes through the material and reveals internal discontinuities as variations in transmitted intensity. The direct visual requirements still apply on the viewing side, and the light source must be of sufficient intensity to illuminate and diffuse the light evenly across the area of interest without glare that masks the very variations being sought.

In pressure equipment it is uncommon, but it appears in glass-reinforced plastic tankage, translucent linings and laminate lay-up work where dry patches, voids, delaminations and inclusions show clearly against transmitted light. Where it appears in a procedure copied from a template and no translucent material exists anywhere on the contract, remove it. An auditor reading a procedure that claims a technique will ask to see that technique demonstrated, and an undemonstrated technique in a live procedure is a finding created entirely by cut and paste.

Where the Acceptance Criteria Actually Live

The evaluation clause of Article 9 hands acceptance back to the referencing code section, and the referencing code is where the arguments happen. For unfired pressure vessels, ASME Section VIII Division 1 governs weld finish and reinforcement limits and fillet weld profile. For process piping, ASME B31.3 sets visual examination in para. 344.2 with acceptance in Table 341.3.2, graded by fluid service. For structural steel, AWS D1.1 Clause 8 and Table 8.1 apply, with different criteria for statically and cyclically loaded members. For welder performance test coupons, ASME Section IX carries its own visual requirements at QW-194.

Two consequences follow. First, the same weld can pass under one referencing code and fail under another, so the report must name the acceptance paragraph, not merely the method. Second, a contract that invokes a client specification more restrictive than the code has silently changed the acceptance criteria, and the procedure must reflect that. The most expensive Article 9 disputes are not about whether an indication existed but about which document the parties were measuring it against.

Site teams should keep the acceptance reference on the report line itself rather than in a header note that gets truncated when the record is exported. Where multiple codes run on one project, a single controlled matrix of weld category against acceptance paragraph, maintained by the Level III and reissued when the specification changes, prevents the mixed-code reporting that generates rework at handover.

VT-1, VT-2 and VT-3 Belong to Section XI, Not Article 9

ASME Section XI, IWA-2210, defines the in-service visual examination categories used on nuclear components. VT-1 is directed at detecting discontinuities and imperfections on component surfaces, including cracking, wear, corrosion and erosion. VT-2 is directed at detecting evidence of leakage from pressure-retaining components. VT-3 addresses general mechanical and structural condition — clearances, settings, physical displacements, loose or missing parts, debris, corrosion and wear. Each carries its own resolution demonstration requirements, which are not the Article 9 requirements.

Reports that read VT-2 examination performed in accordance with ASME Section V Article 9 conflate two codes and are corrected on sight by any experienced nuclear auditor. The correct construction names Section XI for the category and its resolution demonstration, and Section V Article 9 only where Section XI invokes it. Getting this wrong in a written practice propagates into every certificate issued under it, which is why the mapping between category, technique and demonstration should be settled once at the procedure level rather than argued weld by weld.

The Written Procedure, Table T-921 and Procedure Demonstration

Article 9 requires a written procedure and Table T-921 lists the requirements it must address, identifying which are essential variables and which are nonessential. Typical entries include the technique used, the surface condition and method of preparation, the method of examination, whether direct or remote, the lighting and viewing aids, the sequence of examination, the data to be tabulated and any post-examination cleaning. A change in an essential variable requires the procedure to be requalified and revised; a change in a nonessential variable requires revision only.

The procedure must also be demonstrated to be adequate, and the demonstration record is the document most often missing. The demonstration should be run at the least favourable conditions the procedure permits — the maximum viewing distance and the minimum illumination — and should establish that a fine detail on the order of 1/32 in. (0.8 mm) can be resolved on the surface in question. Demonstrating on a clean bench under bright light and then working at arm's length in a dark vessel is exactly the mismatch auditors probe.

Keep the demonstration record physically attached to the procedure revision it supports. When the procedure is revised for a new referencing code or a different remote camera, the old demonstration no longer covers it. Programmes that treat the demonstration as a one-off exercise from the year the procedure was first written accumulate a quiet and growing gap between what is written and what is done.

Documentation, Records and the Findings That Recur

Article 9 requires a written report and the records that support it, and the referencing code frequently adds a checklist or a specific report format. A defensible visual examination record identifies the procedure and revision, the technique, the examiner and certification level, the date, the lighting used and its verification, the surface condition and preparation, the extent and location of coverage, the referencing code acceptance paragraph, the results and the disposition of every indication. A report that says only accepted cannot be defended once the surface is painted.

The findings that recur are consistent across industries. Light readings recorded at the lamp rather than the surface. Light meters out of calibration. The 30-degree angle dimensioned from the normal. No procedure demonstration record, or a demonstration run under conditions more favourable than the procedure permits. Remote camera systems with no resolution equivalence evidence. Near-vision examinations expired, which is a written practice requirement rather than an Article 9 one but voids the examination all the same. Acceptance criteria cited as Article 9 instead of the referencing code.

None of these are examination failures. They are record failures, and they are cheap to prevent and expensive to remediate after the fact, because remediation usually means re-examining a surface that is now insulated, coated or in service. A gap review of the procedure set, the demonstration records and a sample of completed reports, run before the client audit rather than after it, is the least costly point at which any of this can be fixed. If that review is useful to you, start with a consultation.

Frequently asked

Does Article 9 require a certified Level II visual examiner?

Article 9 does not itself set the certification level. Personnel qualification flows from Section V Article 1 and the referencing code section, which normally requires certification under the employer written practice based on ASNT SNT-TC-1A, ANSI/ASNT CP-189 or ISO 9712. Many referencing codes and client specifications require Level II for interpretation and Level III approval of the procedure. Check the contract, then make the written practice match it, rather than assuming a default.

How often must the light meter be calibrated and readings taken?

Article 9 requires the light source, technique and light level verification to be demonstrated once, documented and kept on file. Calibration interval and the frequency of field readings come from your quality system and the client specification, and annual meter calibration with a reading recorded at each examination location is the practice that survives audit. Programmes that record light levels only at the demonstration and never again are the ones that struggle to defend a challenged examination.

Does Article 9 apply to in-service inspection or only new construction?

It applies wherever the referencing code invokes it. Section V is a methods manual; the referencing code decides when Article 9 is used. It is commonly invoked for new construction under Section VIII and B31.3, and also for in-service examinations where the owner-user specification calls for it. In nuclear in-service work under Section XI, the examination categories and their resolution demonstrations govern, and Article 9 applies only where Section XI invokes it.

What is the difference between Article 9 and AWS D1.1 visual inspection?

They answer different questions. Article 9 describes how the examination is performed and documented. AWS D1.1 Clause 8 and Table 8.1 give the acceptance criteria for structural welds, with separate limits for statically and cyclically loaded connections. On a structural contract the acceptance criteria are D1.1, and the technique controls may be D1.1 or Article 9 depending on the specification. Mixing the two on one report is a common source of disputed rejections.

Can magnifiers and mirrors be used under Article 9?

Yes. Mirrors may be used to improve the angle of vision and magnifiers or other aids may be used to assist the examination. The condition is that the aids are named in the written procedure and covered by the procedure demonstration. An examiner who introduces a borescope or a magnifier that the procedure never mentioned has stepped outside the qualified technique, which is a finding independent of whether the call itself was correct.