RT Training Built Around What Manufacturing Actually Radiographs

What manufacturing actually radiographs

Production radiography splits into three families that behave quite differently. Welded fabrication, meaning pressure vessels, piping spools and structural connections, where the construction code sets both technique and acceptance criteria. Castings, where the discontinuities are volumetric and diffuse and acceptance is judged by comparison against reference radiographs rather than against discrete criteria. And assemblies or components, where radiography confirms that something is present, correctly positioned and unobstructed rather than sizing a flaw at all.

Each family changes the technician's task completely. On a weld the geometry is known and repeatable, and the work is technique discipline: source-to-film distance, IQI placement, location markers, film overlap and identification. On a casting the section thickness changes across the part, so a single exposure will not hold density across the whole field and multi-film or multiple-exposure techniques become normal rather than exceptional. On an assembly, the question is orientation, and a radiograph that answers it perfectly may be worthless for flaw detection.

The shop economics matter too, and pretending otherwise produces training that nobody applies. Manufacturing radiography is production work with takt time attached; a technique that adds ten minutes per part can cost a shop a shift a week. That pressure is precisely why shortcuts appear in radiographic records, and why the audit findings are so consistent from shop to shop. Training that acknowledges the pressure and shows where the time can honestly be recovered is worth more than training that pretends the schedule does not exist.

Source and energy selection

The first decision is X-ray or gamma, and it is usually made by geometry and access rather than by preference. An X-ray tube gives lower energy, better subject contrast and a controllable output that switches off between exposures, but it needs power, physical space and a directional beam that must be aimed. Isotopes are portable, work in confined access, and are frequently the only practical answer inside a fabricated assembly, at the cost of continuous emission, a fixed spectrum you cannot tune, and a source that decays on a schedule whether you track it or not.

Among isotopes the working ranges are well established. Selenium-75, with a mean energy near 200 keV and a half-life of about 120 days, gives noticeably better contrast on thin steel than iridium, roughly from a quarter of an inch to a little over an inch. Iridium-192, with a half-life of about 74 days, is the general-purpose choice from around three quarters of an inch through two and a half inches of steel. Cobalt-60, emitting at 1.17 and 1.33 MeV with a five-year half-life, reaches heavy sections but delivers poor contrast and demands substantially more shielding and larger boundaries.

A technician who does not track source decay produces underexposed film and blames the processor. Exposure calculations must use the current source strength, not the activity printed on the certificate six months earlier, and on iridium that is a material difference inside a single quarter. On X-ray equipment, kilovoltage is the contrast control: raise it and you gain penetration and latitude but lose subject contrast, which is exactly the wrong trade when the acceptance criterion turns on fine cracks or tight lack of fusion rather than on gross porosity.

IQI, density and the sensitivity you must be able to prove

An image quality indicator does not measure the size of the smallest detectable flaw, and technicians who believe it does draw the wrong conclusions from it. What it demonstrates is that the radiographic technique achieved the required sensitivity under the actual conditions of that exposure, and it is the only objective evidence on the film that the technique was adequate at all. Under ASME Section V that means a hole-type IQI to ASTM E1025 or a wire IQI to ASTM E747, selected for material thickness plus weld reinforcement, placed on the source side unless the code permits otherwise, and shimmed so it rests on material of equivalent thickness.

Density is the second proof. ASME Section V requires transmitted film density through the area of interest of at least 1.8 for X-ray and 2.0 for gamma radiography, with 4.0 the maximum for single-film viewing, and it limits density variation through the region of interest to within minus fifteen and plus thirty percent of the density measured at the IQI. A densitometer with current calibration against a step wedge comparison film is what turns those numbers into evidence rather than assertion, and an uncalibrated densitometer quietly invalidates every film it certified.

In ISO practice the same job is done through different documents, and shops exporting to European customers routinely run both systems on one shop floor. ISO 17636-1 covers film techniques and 17636-2 covers digital, wire IQIs follow ISO 19232, and the technique classes A and B distinguish basic from improved sensitivity. Acceptance then comes from ISO 10675-1 tied to the ISO 5817 quality level named on the drawing. Technicians must read the drawing to know which system governs, because the two are not interchangeable even where the numbers look similar.

Geometric unsharpness, and the calculation people get wrong

Geometric unsharpness is the penumbra cast by a source of finite physical size: Ug equals the source size multiplied by the object-to-film distance, divided by the source-to-object distance. Two mistakes recur constantly. The first is using source-to-film distance in the denominator instead of source-to-object distance. The second is measuring the object-to-film distance from the film side of the material rather than from the source side of the section being imaged. Both errors make the calculation look compliant on the technique sheet when the actual exposure is not.

ASME Section V sets the limits by material thickness: 0.020 inches for material under two inches, 0.030 inches from two to three inches, 0.040 inches from three to four inches, and 0.070 inches above four inches. The practical lever is source-to-object distance, and the practical constraint is exposure time, which rises with the square of that distance. A technician who genuinely understands the inverse square relationship can negotiate that trade sensibly against the shop schedule. One who does not either shoots too close and fails the geometry, or backs off needlessly and burns hours.

Small focal spot X-ray equipment and physically small sources exist precisely to relax this constraint inside tight shop geometry. Selenium sources are typically physically smaller than iridium sources of comparable activity, which helps the unsharpness arithmetic as much as their softer spectrum helps contrast, and that combination is why selenium has displaced iridium on a lot of thin-wall production work where access is restricted and the geometry cannot be improved.

Sampling, joint efficiency, and why RT is not an AQL problem

In a fabrication shop the extent of radiography is usually a commercial decision embedded in the design rather than a quality decision taken later. ASME Section VIII Division 1 assigns a joint efficiency to a butt weld according to how it is examined, with full radiography, spot radiography and no radiography carrying different values, and the allowable stress used in the wall thickness calculation is multiplied by that efficiency. Choosing spot radiography over full radiography does not simply reduce inspection cost; it thickens the shell. Technicians who understand this stop treating the radiography scope as an arbitrary imposition.

Spot radiography under UW-52 is a triggered scheme rather than a statistical one. Spots are taken at a defined frequency of weld length, and when a spot fails, two additional spots are radiographed on the same weld increment at locations away from the original. If either of those additional spots fails, the increment represented is fully radiographed or the weld is repaired. The logic is escalation on evidence, not estimation of a defect rate. ASME B31.3's random radiography for normal fluid service operates on similar reasoning at a percentage of welds per welder and per procedure.

This is exactly where technicians arriving from a production quality background mis-calibrate. An AQL sampling plan assumes independent, randomly distributed defects across a homogeneous lot, which is a reasonable model for a machined dimension and a poor one for a weld. Weld and casting flaws cluster: one welder, one position, one bad heat, one afternoon when a purge failed or a rod got damp. A failed spot is a signal about a population, and writing it up as a single non-conformance with a rework note misses the point the code is making about the welds nobody radiographed.

Acceptance standards: welds, castings, and the ones nobody reads

For welds, the acceptance criteria live in the construction code rather than in the radiographic standard. ASME Section VIII Division 1 sets them for pressure vessels, ASME B31.3 by fluid service category for process piping, and AWS D1.1 for structural steel, with separate criteria for statically and cyclically loaded connections that catch people who learned one table and assumed it was the table. In ISO practice the acceptance levels come from ISO 10675-1, tied to the ISO 5817 quality level, typically B, C or D, named on the drawing or in the welding specification.

For castings the judgement is comparative. ASTM E1030 sets the practice for radiographic examination of metallic castings, and the assessment is made against reference radiograph plates: E446 for steel castings up to two inches of section, E186 for heavier walls from two to four and a half inches, E280 for heavier sections still, and E155 for aluminium and magnesium castings. The technician grades the observed severity of gas porosity, shrinkage, inclusions and hot tears against the plates, and the drawing states the maximum acceptable severity level for each discontinuity type. Grading against the wrong set for the section thickness produces a defensible-looking report that is simply wrong.

There is a third category that regularly overrides both: the customer specification. A great many manufacturing contracts carry purchase order clauses tightening an acceptance level, mandating full coverage on a particular feature, or requiring retention and submission of the radiographs themselves. Technicians are trained on codes and then handed a drawing. Reading the notes block and the applicable specification list before writing the technique sheet is the habit that separates shops with clean audit histories from shops that are merely busy.

Radiation safety is a separate qualification, not a module

NDT method certification does not authorise anyone to make a radiographic exposure. Industrial radiography is a licensed activity, regulated under 10 CFR Part 34 in NRC jurisdiction or under the equivalent state radiation control programme in an Agreement State, and the individual must be a certified radiographer under that licence. In practice that means passing a recognised certifying entity examination and being trained and tested on the employer's specific equipment, operating procedures and emergency procedures, with records to prove it.

The operational requirements are concrete and non-negotiable. A calibrated survey meter used and logged for every exposure and every source return, an alarming ratemeter, a direct-reading pocket dosimeter alongside a personnel monitoring badge, restricted area boundaries posted and surveyed, source utilisation logs, periodic leak testing of sealed sources and physical inventory, and additional personnel present for certain field operations. Occupational dose limits under 10 CFR Part 20 run to an annual total effective dose equivalent of five rem, with far tighter limits at unrestricted area boundaries.

The finding that recurs in this area is documentary rather than physical. Surveys performed but not recorded with the time, instrument serial number and calibration due date. Utilisation logs missing the source return survey entry. Dosimetry results received but never reviewed and initialled. A shop can be performing radiography entirely safely and still fail an inspection on its records, and in radiography, unlike most of NDT, the body reviewing those records holds enforcement powers that a customer does not.

Digital radiography changes the training, not the physics

Computed radiography using storage phosphor imaging plates, and digital detector arrays, are now ordinary in manufacturing, and the code framework has followed them with mandatory appendices in ASME Section V Article 2 and with ISO 17636-2 on the international side. The geometry is unchanged, the IQI logic is unchanged, and the unsharpness arithmetic is unchanged. What changes is how image quality is proven and how the record is kept, and those are the two areas where shops moving from film get into difficulty.

Instead of film density, the technician demonstrates signal-to-noise ratio and basic spatial resolution, conventionally with a duplex wire gauge, and must control detector calibration and offset, bad-pixel maps, scatter and the ever-present temptation of post-processing. Contrast enhancement can manufacture the appearance of a discontinuity or bury a genuine one, which is why the codes constrain what processing is permitted and require the original unprocessed data to be retained. Technicians trained on film who move to digital without retraining over-process almost without exception, because the tools make it easy and the screen makes it look better.

The archival question is equally real. A radiograph is its own record; a digital image is a record only if the storage medium, file format, metadata and audit trail are controlled and verified. Shops that adopt digital detector arrays and keep no verified backup or format migration plan discover the gap three years later when a customer requests a re-read on a vessel that has just failed in service, and the images are unreadable or the settings that produced them are gone.

What Atlantis delivers for manufacturers

Radiographic testing training to ASNT SNT-TC-1A and ISO 9712 at Levels I, II and III, delivered as classroom, on-site corporate or blended programmes, with film, computed radiography and digital detector array content matched to what the shop actually runs rather than to a generic syllabus. Specific examinations are built on the employer's own procedures, IQIs, technique sheets, product forms and acceptance standards, so the certification file reflects the work the technician will genuinely be asked to do.

For manufacturers building a programme or repairing one after a customer audit, our ASNT Level III consulting covers the written practice, the examination bank, the Level III of record function, technique sheet development and the record structure that stands up when a customer or regulator opens the file. Contact info@atlantisndt.com for a consultation or a scoped quote. Programmes are shaped to the product and the codes in play, and the positioning is affordable, accessible and fully customisable.

Request a consultation