Which Document Governs? Precedence, Coverage and Digital Radiography
A shop builds a 25 mm SA-516 Gr.70 vessel to ASME Section VIII Div. 1. The purchase order carries an order-of-precedence clause: PO, then owner specification NDT-100 Rev. 4, then the code, then referenced standards. NDT-100 requires 100% RT of all Category A and B butt welds, IQI sensitivity 2-2T, and permits computed radiography 'where qualified.' The fabrication estimate assumed the code default - full RT on Category A, spot RT on Category B - and the design was priced at the joint efficiency that goes with that. Site RT cites the 2017 edition of Section V with a 1.8–4.0 film density window; the shop's CR scanner is quoted at 130 µm basic spatial resolution; the Authorized Inspector has not seen a digital technique. During the work the PM asks which document actually controls.
1. The PM argues that because the PO itself does not mention NDT-100, and the PO ranks first in the precedence clause, the code default applies. Is he right? Work through the precedence and the joint-efficiency consequence.
Read the worked answer
He is wrong on both counts.
- The PO incorporates NDT-100. The precedence clause names NDT-100 as a document of the contract; a document cannot be ranked second in a list and simultaneously be excluded from the contract. The clause tells you which wins when two applicable documents conflict - it does not delete the specification. Here there is no conflict on coverage: NDT-100 is simply more stringent than the code, and a specification may always be more stringent than the code.
- Precedence cannot be used to go below the code. The code of construction is a legal/regulatory requirement in most jurisdictions. A precedence clause that purported to reduce examination below the code minimum would be unenforceable; it can only allocate which of the higher requirements governs.
Joint-efficiency consequence - the direction matters:
- Type 1 double-welded butt joints: E = 1.00 with full RT, 0.85 with spot RT, 0.70 with none.
- The estimate was priced at spot RT, i.e. E = 0.85. Requiring 100% RT raises E to 1.00, which is the more favourable assumption for design. So the design remains valid; the fabricator cannot claim a design change, only a cost and schedule change in NDT. That claim is legitimate and should be handled as a variation, not as a technical dispute.
- The trap to check: if the design had been run at E = 1.00 and the shop then executed spot RT without a design review, that would be a genuine nonconformance requiring either full RT or a re-rated design. Verify which way the drawings were calculated before conceding anything.
2. What must be resolved before the CR system can replace film on this job, and what parameters would you actually verify rather than accept from the vendor's datasheet?
Read the worked answer
The vendor's 130 µm SRb is a scanner property, not an examination result. The governing question is whether the image achieves the required quality on this thickness, geometry and material.
Verify by demonstration on a production-representative test piece containing known flaws:
- Achieved IQI sensitivity - must meet NDT-100's 2-2T, including the 2T hole, on the actual 25 mm section with the actual source-to-object distance and geometry used.
- Achieved basic spatial resolution measured with a duplex-wire IQI in the image, not quoted from the spec sheet; the resolution must be sufficient to resolve the essential wire and the 2T hole.
- Signal-to-noise ratio in the image, and the scanner's calibration/normalisation and bad-pixel map, with a defined re-calibration interval.
- Grey-level/density equivalence - the film 1.8–4.0 window does not transfer to digital; a digital procedure specifies grey-level limits and SNR instead. The site procedure citing the 2017 edition of Section V and a film-density window must be rewritten for the digital technique and re-approved.
- Personnel - the radiographer, the interpreter and the Level III must be qualified for the digital technique, not just for film.
- Approval chain - a written digital procedure demonstrated and accepted by the AI and the owner before production use. The AI's 'not yet seen' is the controlling fact: without that acceptance the CR image is not a code examination.
If the demonstration fails 2-2T, the CR system is not usable on this job regardless of its price or its datasheet.
3. Assume the code default had applied and the shop was performing spot RT on a Category B weld. One spot shows a slag inclusion above the acceptance limit in a weld made by welder #7. What is the required response, and what does it tell you about the rest of #7's work?
Read the worked answer
The spot has failed, and the consequence is escalation, not repair-and-move-on.
- The defective spot must be repaired (excavate and re-weld with a qualified WPS, then re-examine).
- The spot examination is then extended: the code's spot rule requires additional spots in the same weld, by the same welder/operator, to determine whether the defect is local or characteristic of that welder's work on that joint. If the additional spots also fail, the examination escalates to full radiography of that weld.
- Re-examine the design assumption. Spot RT corresponds to E = 0.85. If the weld is not fully examined, the design must remain at 0.85 or lower; you cannot repair the defect and quietly keep a 1.00 assumption.
- #7's other work is now a population at risk. One failed spot is evidence about a process, not just a joint. The sound engineering response is to examine a sample of #7's other production welds in the same period and on the same WPS, and to review the WPS variables, the flux/electrode lot and the joint preparation. If the sample also fails, stop the welder and requalify.
- Document all of it - the failed spot, the extension, the repair, the re-examination - in the RT report and the nonconformance record; the AI will ask for the trail, and the joint efficiency assumed in the design must be traceable to the examination actually performed.