Proving the procedure works on a block with flaws in it

Written, demonstrated, qualified — three states codes never use interchangeably

Three different conditions get called procedure qualification on a shop floor, and only one of them leaves data an auditor can re-test. A procedure is written when it carries every requirement listed in the Section V variables table for its method — T-421 for ultrasonics, T-621 for penetrant, T-721 for magnetic particle, T-921 for visual. It is demonstrated when the Authorized Inspector has watched it work and is satisfied. It is qualified when it has found flaws in a specimen whose contents were fixed and recorded before the demonstration began.

The distinction carries money. A supplier that quotes qualification and hands over a well-drafted document has delivered the first state and invoiced for the third. Writing the procedure is real work in its own right, and our NDT technical procedure development practice exists because the variables tables punish vague drafting. But the moment a client contract or a construction code uses the word qualified, a block with flaws in it enters the scope of supply, and so does a named witness.

Read the sequence in reverse and the dependency becomes obvious. Qualifying a procedure that has not been written is impossible, because the qualification record has to state which essential variable ranges it locks in. Demonstrating a procedure to an Inspector who has never seen the document is equally impossible. Each state rests on the one below it, and a skipped rung surfaces later as a rejected weld map, a rewritten technique sheet, or a hold point nobody can clear.

The referencing construction code decides, not Section V

ASME Section V is a toolbox that construction codes reach into. T-150 says that when required by the referencing Code Section, examinations shall follow a written procedure, and that the procedure shall be demonstrated to the satisfaction of the Inspector. Qualification is a separate switch, thrown by a separate sentence in a different book. Practitioners on the ASME forums put it bluntly: procedure qualification is required by the reference code only. Section V supplies the method for qualifying; it never orders you to do it.

Section VIII Division 1 shows the mechanism working. UW-51(a)(4) permits ultrasonic examination in place of required radiography and routes it to Section V Article 4 Mandatory Appendix VIII, which uses fracture-mechanics-based acceptance criteria and carries its own procedure qualification requirement through Mandatory Appendix IX. Appendix VIII also constrains the technique: automated or semi-automated scanning with computer-based data acquisition and analysis. Manual straight beam, manual angle beam and manual phased array are not accepted as the substitute for required radiography.

That chain is also volatile, so read the edition in front of you. The 2025 edition of Section V initially dropped the Appendix IX reference from VIII-421.1, leaving procedure qualification pointing only at Article 1 T-150(d). ASME then issued Errata R#25-1636 restoring the Appendix IX requirement. A QA manager who built a scope of work from the uncorrected print would have bought the wrong blocks. The first question in every qualification job is which paragraph of which construction code put you here — a question our outsourced Level III consulting work starts with rather than assumes.

What T-150(d) actually demands of the specimen

The operative text is short and unusually specific. When qualification of the written examination procedure is required by the referencing Code Section, it is qualified by performing a demonstration examination on a minimum of one test specimen having flaws whose size, location, orientation, quantity and characterization have been determined prior to the demonstration and are known only by the supervising Level III Examiner. Five attributes, fixed in advance, held by one person. Every word in that sentence is load-bearing.

Known only by the supervising Level III is the clause that turns paperwork into a test. The operator who will run production work approaches the block without the answer key. The Level III who wrote the procedure watches somebody else prove it. This is a semi-blind trial, not a walkthrough, and firms that hand the technician a flaw map beforehand have run a training exercise and filed it as a qualification. Auditors ask who held the flaw data, and the answer sits in the record.

Representativeness carries the rest of the weight. The specimen has to stand in for the production article: thickness within the range the procedure claims, matching product form, comparable weld configuration, comparable surface condition and material group. Common shop practice sizes the demonstration block inside the thickness band being qualified rather than at one extreme, and holds it to the same finish the production surface will present. A block that is easier to inspect than the work qualifies nothing.

Buying, building or borrowing the demonstration block

Three routes exist and each has a cost profile. Buy implanted-flaw specimens from a specialist manufacturer, fabricate blocks in-house with deliberately introduced weld defects, or use a client-owned mock-up already built for the project. The purchased route is fastest and carries documentation the auditor recognises. The fabricated route is cheapest in cash and most expensive in lead time, because establishing flaw truth on a block you made yourself is a project of its own.

Manufacturer specifications tell you what a code-built kit contains. Flawspec publishes an ASME Section V Article 4 Appendix IX qualification kit of six carbon steel specimens holding eighteen flaws, three per specimen randomly located, with tolerances of plus or minus 0.5 mm on flaw height and 1.0 mm on flaw length, manufactured to paragraph IX-435.1 alongside Article 4 paragraphs T-434.1.2 through T-434.1.7, and supplied with as-built CAD drawings, blank test sheets, a certificate of conformance and material test reports. Sonaspection supplies comparable flawed specimen sets built to the ASME Section XI Appendix VIII supplements.

Flaw truth is the part firms underestimate. A flaw is only known if its dimensions were established by a route independent of the technique being qualified — destructive sectioning, or a documented reference examination performed and recorded separately. Without traceable flaw truth, a successful demonstration proves that two unverified numbers agreed. The as-built drawing and the material certification are what let an auditor three years later confirm the block in your cabinet is the block in the record.

What counts as a successful demonstration

Detection is the floor, not the finish line. For a workmanship-criteria procedure, the trial asks whether every flaw the specimen holds was found and reported, and whether the operator called flaws where none exist. False calls are counted, because a technique that reports everything detects nothing useful. The acceptance criteria are written into the qualification plan before the block is unwrapped, so nobody negotiates the pass mark after seeing the data.

Flaw sizing work raises the bar again. Under Article 4 Mandatory Appendix IX the demonstration extends past detection into measuring through-wall height and length on each flaw, and into categorizing each one as surface-connected or subsurface. Those are the two outputs fracture-mechanics acceptance criteria consume, which is why the appendix exists at all. The comparison is against the flaw truth held by the supervising Level III, evaluated against the criteria the appendix states.

A failed demonstration is a result, not an emergency. Record it, identify the variable that caused it — probe angle, gain, index offset, scan increment, surface preparation — change that one variable, revise the procedure and run again. What destroys a programme is the undocumented re-run: three attempts, one filed record, and a technician who remembers the first two. Keep every attempt in the file and the qualification becomes defensible instead of merely favourable.

Who witnesses, and why the Level III cannot be everywhere at once

Two roles have to be occupied by two people. The supervising Level III controls the flaw truth, sets the acceptance criteria and directs the trial. The Authorized Inspector witnesses and accepts the demonstration for code purposes. On owner-driven qualifications a client representative or third-party inspection agency joins, and on nuclear work an independent programme administers the whole event. Forum practitioners state the arrangement plainly: the demonstration sits under the control and supervision of an NDE Level III and the Authorized Inspector.

The separation exists for the same reason it exists in personnel certification. A Level III holding the answer key who also stands at the technician's shoulder has run a coached exercise. SNT-TC-1A takes the identical position on examinations, forbidding anyone from administering their own examination or a subordinate's, and the independence argument behind our Level III examination oversight service transfers directly to procedure demonstrations.

Small firms hit the arithmetic problem here. One in-house Level III who wrote the procedure, holds the flaw truth, employs the technician and signs the record has collapsed four roles into one signature. Bringing in an outside Level III for the demonstration alone costs a day and removes the single largest objection an auditor or a client's technical authority will raise against the resulting record.

The procedure qualification record: what has to be inside it

The record is the product. It names the procedure number and revision qualified, the specimen identification and its as-built drawing, the instrument make, model and serial number, the probes, wedges, cables, couplant, film or consumable batch as the method requires, the software version where data acquisition is computerised, the operator's name and certification level, the date and place, and the raw data captured. Then the results: what was found, what was missed, what was called that was not there.

Two entries decide whether the record ages well. First, the essential variable ranges the qualification covers — thickness band, material group, weld configuration, scanning surface, angles and modes — stated as ranges rather than single values, so the procedure travels as far as the demonstration honestly supports. Second, the witness signature with the witness's role, because a record without a countersignature is a self-assessment.

The common audit gaps are dull and repeat everywhere. An unsigned qualification record. A procedure that never cites the record qualifying it, so the two documents live in separate binders. A block referenced by a number that matches nothing physical in the shop. A revision to the procedure issued after qualification with no assessment of whether the change touched an essential variable. Each of these is a finding, and each is cheaper to close before an auditor writes it than after.

Qualifying a procedure is not qualifying personnel

These are two different objects governed by two different documents. A procedure is qualified under ASME Section V, against a construction code trigger, and the record is a procedure qualification record. A technician is certified under the employer's written practice, built on SNT-TC-1A or ANSI/ASNT CP-189, through general, specific and practical examinations, and the record is a certification file. Different standard, different signature, different expiry. The two get confused because both involve a block with flaws in it.

The personnel side has its own numbers. A practical examination under SNT-TC-1A uses a minimum of ten different checkpoints covering the NDT variables and the employer's procedural requirements, and the candidate is required to detect the specified discontinuities in the approved specimens. That specimen is not the procedure qualification block, it is not scored against fracture-mechanics criteria, and passing it says nothing about whether the technique itself works on the production geometry. Our SNT-TC-1A compliance guidance covers that side of the wall.

One code deliberately knocks the wall down, and the exception proves the rule. ASME Section XI Mandatory Appendix VIII qualifies procedure, equipment and examiner together as a single examination system, which is why the nuclear industry calls it performance demonstration rather than procedure qualification, and why the US utilities funded the Performance Demonstration Initiative to administer it. Outside Section XI, qualifying the procedure leaves every technician still needing their own certification file.

Essential variables: what quietly voids the qualification you paid for

Each Section V method article carries a table of procedure requirements split into essential and nonessential variables. Table T-421 does it for ultrasonics, T-621 for penetrant, T-721 for magnetic particle, T-921 for visual. A change to a requirement marked essential, away from the specified value or range, requires requalification of the written procedure. A change to a nonessential requirement needs a revision and reissue, and nothing more. The table is the contract between what you demonstrated and what you are allowed to do.

The traps are mundane and they arrive through procurement rather than engineering. A probe model discontinued and replaced with a near-equivalent. A thickness range extended to win one job. A scanner substituted because the original is on another site. A surface condition changed from ground flush to as-welded. An acquisition software version updated across the fleet overnight. Each of these lands on the essential side for ultrasonics, and each turns a qualified procedure into an unqualified one without anybody noticing.

The fix is governance, not cleverness. Attach a one-page variables register to every qualified procedure listing the essential variables and the exact ranges the demonstration covered, and route equipment purchases and software updates through it. Firms that carry this register close change-control findings in minutes, and it is one of the first artifacts we build during an NDT programme audit and gap assessment.

How we run a demonstration for a small NDT firm

The engagement starts at the construction code, not at the block. We identify the paragraph that triggered qualification, confirm which appendix chain applies in the edition your client contract names, and establish whether you actually need a T-150(d) qualification or a demonstration to the Authorized Inspector. That determination alone reroutes a large share of the enquiries we receive, because the majority of shop work needs the cheaper state and has been quoted the expensive one.

From there the work is linear. Repair or write the procedure against the method's variables table. Specify the demonstration block — thickness, product form, weld configuration, flaw population — and source or supervise its fabrication. Run the trial under a Level III certified in that method, with the flaw truth held apart from the operator. Produce the procedure qualification record, the variables register and the linkage back into the procedure. We cover UT, PAUT, TOFD, RT, MT, PT and VT.

Atlantis is affordable, accessible and fully customizable, and the scope is built around the code paragraph you are answering rather than a fixed package. Send us the client specification and the construction code reference and we will tell you which of the four states in the table above you need to reach — request a quote or a demo and we will scope the demonstration from there.

Request a consultation