Toe Line on a Coated Nozzle Weld

The evidence points to a nonrelevant (geometry/entrapment) indication, but that is a provisional classification, not a disposition.

Evidence for nonrelevant:

• It follows the toe contour exactly, i.e. it tracks a surface feature rather than a defect path.
• It persists after wiping with the field off - the classic signature of particles held mechanically in a crevice, not held by leakage flux.
• Grit is visible in the crevice, which is a ready reservoir.

Sequence before reporting:

1. Clean the toe properly - brush and, if the code permits, remove the primer locally - so the grit and entrapment reservoir are gone.
2. Re-magnetise in two directions at about 90° and re-observe under black light.
3. If a line reforms only after cleaning and with the field on, and is sharp and has a distinct end, treat it as a linear indication and evaluate it.
4. Note that the toe may also carry genuine undercut, which is a weld flaw in its own right even when the particles are only trapped; measure depth or profile it, and confirm with PT or a light blend.

One caution: the weld cap was not ground, so a toe crevice is expected - but two identical nozzles show nothing, so 'expected' does not make it acceptable without evaluation.

Not automatically - but it is now a documented variable, not a background detail.

• For an AC yoke, thin, continuous, well-bonded coatings up to roughly 0.25 mm are commonly accepted, provided the lifting power is demonstrated on the coated surface itself. That was done: 4.5 kg at 150 mm pole spacing on the primer.
• What governs is not the nominal thickness but whether the coating is uniform, dry and tightly bonded, and whether the lift test was done at the same coating condition and pole spacing used in production. A thick, flaking or uneven coating bridges the gap and diverts flux around the defect.
• Practical traps here: the nozzle contour forces a different pole spacing and leg attitude than the flat plate used for the lift test; and a 0.10 mm nominal coat can be 0.2 mm at a fillet. Re-verify lifting power on a coated mock-up or a coated area of the same component at the spacing actually used.
• The primer is not what produced this indication - the trapped grit is - but the coating does raise the field requirement, so record the thickness and the verification in the report.

A 45° pass is not a substitute for a planned two-direction inspection. Particles align along the leakage field, so an indication is strongest when the magnetising field is roughly perpendicular to the long axis of the defect. Two passes at about 90° to each other therefore cover all orientations, because any defect axis is at worst 45° from one of them - adding a third pass 45° off gives no orientation that the first two do not already cover.

For this nozzle-to-shell weld, plan:

• Pass 1: poles straddling the weld, field across the weld, to find indications running along the toe or the fusion line (the likely crack orientation).
• Pass 2: field along the weld axis, to find transverse indications.
• Overlap each pass by at least 25 mm and keep the legs on clean, primed surface with the spacing verified.
• On the curved nozzle-to-shell junction, the two legs will not sit at a fixed 150 mm; check that the effective spacing still gives adequate field, because lifting power falls as the spacing is reduced on a tight contour.