Leak testing to ASME BPVC Section V, Article 10
ASME Section V Article 10 governs leak testing under the Boiler and Pressure Vessel Code. It defines a family of techniques, from bubble solution and vacuum box through halogen diode and three helium mass spectrometer techniques to pressure change, each executed under its own mandatory appendix. Article 10 controls how the test is performed, calibrated and documented. The referencing code, not Article 10, sets acceptable leakage.
The article is structured like every other Section V method article. The T-1000 paragraph series carries scope, general requirements, equipment, technique, calibration, examination, evaluation and documentation, and then the operative numbers are pushed down into mandatory appendices, one for each technique. That structure is why a purchase order clause reading examine per ASME Section V Article 10 is not a specification at all. It does not say which technique, which appendix, at what pressure, at what sensitivity, or against what acceptance. Two vendors reading the same clause will legitimately deliver a soap bubble test and a helium hood test, six orders of magnitude apart in sensitivity, and both will be compliant. The buyer who wants a defined result writes the technique, the required sensitivity, the test pressure and hold time, and the acceptance criterion into the requisition, and then verifies that the vendor procedure names the same appendix.
Source: Sources: ASME BPVC Section V, Article 10 Leak Testing and its Mandatory Appendices; ASME BPVC Section V, Article 1 T-120 personnel qualification; the ASME-adopted standards SE-432 selection of a leak testing method, SE-479 preparation of a leak testing specification, SE-427 halogen leak detector, SE-498 and SE-499 helium mass spectrometer tracer and detector probe techniques; ASME B31.3 para. 345.8 sensitive leak test; ASME BPVC Section VIII Division 1 pressure test requirements; API 650 vacuum box examination of bottom welds; ASNT SNT-TC-1A.
| Technique | Governing mandatory appendix | Order-of-magnitude sensitivity (std cm3/s) | Where it is normally specified |
|---|---|---|---|
| Bubble test, direct pressure | Bubble test direct pressure appendix | 1x10-3 to 1x10-4 | Accessible pressurised welds and fittings, shop and field |
| Bubble test, vacuum box | Bubble test vacuum box appendix | 1x10-3 to 1x10-4 | Tank bottom and shell-to-bottom welds that cannot be pressurised |
| Halogen diode detector probe | Halogen diode detector probe appendix | 1x10-5 to 1x10-6 | Refrigerant and halogenated tracer systems, packaged equipment |
| Helium mass spectrometer, detector probe | Helium detector probe appendix | 1x10-6 to 1x10-7 | Locating a known leak on a pressurised component |
| Helium mass spectrometer, tracer probe | Helium tracer probe appendix | 1x10-8 to 1x10-9 | Evacuated systems, vacuum vessels, jacketed cryogenic lines |
| Helium mass spectrometer, hood | Helium hood appendix | 1x10-9 to 1x10-10 | Total leakage of a sealed assembly, with no location information |
| Pressure change test | Pressure change appendix | Volume dependent, typically 1x10-1 to 1x10-3 | Large fixed volumes, systems already at operating pressure |
| Ultrasonic leak detector | Ultrasonic leak detector appendix | Gross leakage only | Rapid survey of large enclosures and access hatches |
What Article 10 covers, and what it deliberately does not
Article 10 is the leak testing article of ASME Boiler and Pressure Vessel Code Section V. It follows the same architecture as the other method articles: a T-1000 paragraph series carrying scope, general requirements, equipment, procedure and technique, calibration, examination, evaluation and documentation, followed by mandatory appendices that hold the numbers. The general body tells you that a written procedure is required, that personnel shall be qualified in accordance with Article 1, and that the technique selected shall be capable of the sensitivity the job needs. The appendix tells you the pressure, the hold time, the standoff, the scan rate and the calibration regime. Both halves are mandatory once invoked, and a procedure that quotes the article without naming the appendix has not been written to the code.
What Article 10 does not do is where the arguments start. It is not a strength test and does not displace the hydrostatic or pneumatic proof test the construction code requires. It does not, by itself, establish an allowable leakage rate. It is not a volumetric examination: a joint that holds helium at 1x10-9 std cm3/s can still contain slag inclusions, lack of side wall fusion, or a crack that has not yet reached the far surface, and none of those will be found by any leak test until the day they do penetrate. Substituting a leak test for radiography or ultrasonics because the schedule slipped is a decision that gets recorded in the file and then read out loud during a failure investigation.
The practical consequence is that Article 10 is always invoked by something else. Before writing a line of procedure, establish which code section or specification is doing the invoking, what leakage that document will accept, at what pressure and temperature the examination is to be run, and whether the result must be quantitative or simply pass or fail. Getting those four answers on paper first is a large part of what ASNT Level III consulting is for on a leak testing scope, and it prevents the far more expensive discovery that the vendor tested to a technique two orders of magnitude coarser than the design intent.
The techniques, and how the appendix binds the procedure
The techniques span roughly seven orders of magnitude of sensitivity. At the coarse end, bubble solution applied to a pressurised joint and the vacuum box drawn over a tank bottom weld both detect in the region of 1x10-3 std cm3/s. Halogen diode detector probe testing sits a decade or two finer. The three helium mass spectrometer techniques, detector probe, tracer probe and hood, occupy the fine end, with the hood technique reaching the 1x10-9 to 1x10-10 region but returning only a total leakage figure with no indication of where the leak is. Pressure change testing is a different animal again: its sensitivity depends entirely on the volume under test, the accuracy of the pressure and temperature instrumentation and the duration, and it is the only technique that scales gracefully to very large fixed systems.
Selection is driven by four questions. What sensitivity does the service actually require? Can the item be pressurised, or must it be evacuated? Do you need to locate the leak, or only quantify it? And is the surface accessible for direct observation? A tank bottom cannot be pressurised, so the vacuum box wins by default. A vacuum-jacketed cryogenic line cannot be observed, so a tracer probe test on the evacuated annulus is the only sensible route. A packaged skid with a hundred threaded joints is a location problem, so a detector probe scan beats a hood test even though the hood is the more sensitive technique on paper.
Once the technique is chosen, the appendix binds the procedure. Changing from the direct pressure bubble technique to the vacuum box technique is not a minor revision; it is a different examination under a different appendix, with different equipment, a different calibration regime and different essential variables. The same applies to altering tracer gas concentration or moving from a hood to a probe. Procedures that list every technique in one document and let the technician choose on site are common, and they fail review reliably, because there is then no way to demonstrate which set of requirements the examination was actually performed under.
The requirements that actually bite
Four requirements generate most of the real non-conformances. First, temperature. The bubble test appendix restricts the surface temperature to approximately 40 F to 125 F, roughly 5 C to 52 C, unless the procedure is qualified elsewhere with the solution in use. Below the band the solution thickens and stops forming bubbles; above it the film flashes off before a slow leak can inflate one. Almost nobody records a surface temperature reading, which means almost nobody can demonstrate compliance after the fact. Second, hold time. The direct pressure bubble technique requires the test pressure to be held for a minimum of fifteen minutes before examination begins, which exists so that a small leak has time to establish a flow path through whatever the joint is packed with.
Third, calibration. The halogen and helium techniques require a traceable calibrated standard leak, a calibration check before the examination and a repeat check at the end, and a maximum interval between checks during long scans. A failed closing check voids everything examined since the last passing check. Auditors know this and go straight for the standard leak certificate date and serial number on the report. Fourth, scanning discipline. The detector probe techniques cap probe travel speed at approximately one inch per second and require the probe tip to be held within about a quarter of an inch of the surface. A technician who scans a thirty metre weld in four minutes has exceeded the cap by a factor of twelve and has produced an examination that cannot detect the leak rate the procedure claims.
Two further items catch people out. The bubble-forming solution must be compatible with the material: halogen-bearing and sulfur-bearing solutions on austenitic stainless steel and nickel alloys are a corrosion problem that surfaces months later, and the solution must produce bubbles without foaming so heavily that it masks them. And Article 10 does not itself state a minimum illumination for observing bubbles, which surprises people who assume the hundred foot-candle figure from the visual examination article carries across. It does not carry across automatically. If the owner wants a defined illumination level for bubble observation it has to be written into the procedure or the specification, and technicians trained to measure and record it, which is a straightforward addition to any NDT training and certification programme.
Acceptance criteria and where they really come from
The evaluation paragraph of Article 10 is short because the criterion normally lives elsewhere. Absent a stated allowable rate, the component is unacceptable when leakage is detected, and that default is stricter than most purchasers realise. Where the referencing document does state a rate, for example an allowable helium leakage for a sealed enclosure or a permitted pressure decay over a defined period, the procedure must show that the technique selected can actually resolve that rate against the background present on site. That is a demonstration, not an assertion, and the demonstration record is what an auditor asks for.
For bubble testing, the discriminating skill is telling a leak from an artefact. A leak produces continuous bubble growth at the same point that re-establishes itself after the solution is wiped and reapplied. Entrapped air escaping from a lap, solution outgassing, and bubbles migrating from an adjacent joint all look convincing for the first thirty seconds. The correct response is to clean, reapply and observe again from a fixed viewpoint, and to mark and record the location either way. For helium work the criterion is instrument response above the recorded background, and the background itself must be recorded, because a facility that has been flooding helium into an adjacent bay all morning will have a background that swallows the reject level.
Repairs bring their own trap. A joint that leaked, was repaired and then retested must be re-examined by the same technique and to the same sensitivity, and the re-examination has to appear in the record as a separate examination with its own calibration checks, not as an annotation on the original report. Where a leak indication is dispositioned as acceptable against a stated allowable rate, the quantitative basis for that disposition belongs in the file. Independent review of inspection reports and data before the package leaves site is the cheapest place to find these gaps, because the alternative is finding them when a certifying body samples the file a year later.
How Article 10 interacts with the referencing code
ASME B31.3 is the clearest example. Its sensitive leak test provision requires the test to be performed in accordance with Section V Article 10, at a pressure of at least the lower of one hundred and five kilopascals gauge, about fifteen psi, or twenty five percent of the design pressure, and with a sensitivity of not less than 1x10-3 std cm3/s under test conditions. That sensitivity figure is the operative number: it eliminates the coarsest approaches and effectively points the designer at a properly executed bubble test or better. For Category M fluid service the sensitive leak test sits on top of the normal code leak test, and treating it as an alternative is a classic misread that surfaces during pre-commissioning.
On the pressure vessel side, Section VIII Division 1 governs the proof test in its own paragraphs, and leak testing appears where the user design specification calls for it, typically on jacketed vessels, sealed enclosures and equipment in toxic service. The obligation to specify sits with the user, not the fabricator, so a vessel purchased without a leak test requirement will not receive one. On storage tanks, API 650 calls for vacuum box examination of bottom welds and specifies its own partial vacuum values and technique detail, which are not identical to the Section V appendix. Where a purchase order invokes both documents the procedure must satisfy the stricter of the two, and someone has to make that comparison explicitly rather than assume they agree.
The general rule is that the referencing document wins on acceptance and on when the test is required, while Article 10 wins on how the examination is performed. Where the owner specification adds requirements, such as a defined illumination level, a specific tracer concentration or a maximum background, those additions are contractual and must appear in the procedure. Keeping that mapping visible across a fleet of procedures, so a reviewer can see which code drives which clause, is exactly the sort of thing that belongs in a controlled inspection management system rather than in a folder of loose files on a shared drive.
The misreadings that produce audit findings
The most common finding is a procedure written to the article without naming a technique or appendix, which makes the examination unauditable. Close behind is the multi-technique procedure that lets the technician pick on the day. Then come the record gaps: no surface temperature, no hold time, no standard leak serial number or certificate date, no background reading, no acceptance criterion cited to a source document. Each of these individually looks trivial. Collectively they mean the file cannot demonstrate that the stated sensitivity was achieved, and an auditor who cannot verify the sensitivity has to treat the examination as not performed.
A second cluster concerns personnel. Article 1 requires qualification and certification in accordance with the employer written practice, and leak testing is a method where the written practice frequently lags reality. Practices that certify people in leak testing without differentiating the techniques leave the employer unable to show that the individual who ran the helium tracer probe test was qualified for that technique. Vision requirements are another soft spot, because bubble testing is fundamentally a visual examination and the annual near-vision record has to be current on the date of the examination, not merely on the date of certification.
The third cluster is scope creep in the wrong direction: using a bubble test to demonstrate weld quality, running a pressure change test on a system whose volume was never measured, or accepting a vacuum box examination where the box placements were not overlapped and no placement map exists, so there is no evidence that the whole weld length was covered. All three are recoverable if caught during fabrication and expensive if caught at handover, which is the argument for a technical review of leak testing procedures and records at the start of a project rather than at the end of it.
Building a procedure that survives review
A defensible Article 10 procedure states the technique and its appendix on the front page, lists the essential variables and what happens when each changes, defines the equipment down to the standard leak and its calibration interval, sets the test pressure, hold time and temperature window, defines the scanning pattern or box placement and overlap, states the acceptance criterion with the document it comes from, and specifies exactly what the report must record. It also states what the procedure does not cover, which prevents it being stretched onto a job it was never demonstrated for.
Demonstration matters more in leak testing than in most methods, because sensitivity is a property of the whole system rather than of the instrument. A helium test that passes a bench calibration can still fail to detect the target rate on a large vessel with a long response path. Demonstrating the procedure on a representative geometry with a calibrated leak installed at the worst-access point is the only honest way to establish that the claimed sensitivity is real, and it produces a record that answers the auditor question before it is asked.
Atlantis writes and qualifies leak testing procedures under ASNT Level III technical authority, trains and certifies technicians within the client written practice, and performs independent review of completed leak test packages. The work is affordable, accessible and fully customisable to the referencing code and the owner specification in play. If a leak testing scope is coming up, or a completed package needs a second set of eyes before it goes to a certifying body, request a consultation and describe the equipment, the referencing code and the sensitivity you have to demonstrate.
Does ASME Section V Article 10 set the allowable leak rate?
No. Article 10 sets out how the examination is performed, calibrated, evaluated and recorded, but the quantity of leakage that may be accepted comes from the referencing code section, the design specification or the purchase order. In the absence of a stated allowable rate, evaluation defaults to no detected leakage, which is a far harsher criterion than most buyers intend and one that vendors will price accordingly when they see it.
What surface temperature range applies to bubble solution testing?
The bubble test appendix restricts the temperature of the surface under examination to a band of roughly 40 F to 125 F, that is about 5 C to 52 C, unless the procedure has been qualified at other temperatures with the solution actually being used. This bites twice a year: winter shutdowns in northern climates and summer daylight work on dark-painted vessels both push surfaces outside the band without anyone recording a reading.
Can a sensitive leak test replace the hydrostatic test?
No. A leak test proves tightness at the pressure applied. It says nothing about the strength of the component, and it is not a substitute for the hydrostatic or pneumatic proof test required by the construction code. Where a sensitive leak test is required, as in ASME B31.3 for Category M fluid service, it is an additional requirement layered on the normal code leak test, not an alternative to it.
Which technique should be specified for an evacuated cryogenic jacket?
The helium mass spectrometer tracer probe technique is the normal choice, because the annulus can be evacuated and connected to the spectrometer while helium is played over the outer surface. That configuration reaches the 1x10-8 to 1x10-9 std cm3/s region. Detector probe testing is the wrong tool here: it needs the tracer inside the component at pressure and cannot see through the outer jacket at all.
How often must the instrument be checked against a standard leak?
The halogen and helium appendices require calibration against a traceable calibrated standard leak before the examination begins and again at the end, with a stated maximum interval between checks during long tests and an immediate recheck whenever the instrument is disturbed. If the closing check fails, everything examined since the last passing check is void and must be re-examined. Record the standard leak serial number, its rate and its certificate date.
What must a leak test report record to survive an audit?
Technique and the appendix it was performed under, procedure number and revision, test pressure and hold time, part and ambient temperature, tracer gas and concentration, solution or instrument used, calibrated standard leak identity and calibration status, scan rate or box placement map, examiner name and certification level, acceptance criterion with its source document, results, and any repair and re-examination. Missing temperature and standard leak identity are the two commonest gaps.