ISO 15330 — preloading test for the detection of hydrogen embrittlement

Scope — and what it does not cover

ISO 15330 describes a preloading test used to detect internal hydrogen embrittlement in finished, coated fasteners: samples are held under a defined tensile load in a fixture with parallel, non-wedging bearing faces for a defined time, and delayed fracture during the hold is the failure signal. It is a lot-based detection test. It does NOT qualify a plating process (ASTM F1940 is written for that), does not address embrittlement that develops in service, and a pass does not certify the parts that shipped — the parts that were tested were consumed by the test.

Standards are copyrighted documents. We explain scope, intent and practical application in our own words; obtain the current edition from ISO, ASTM, DIN or your national standards body.

What actually bites in production

• 1The bearing faces are parallel by design. A wedge angle would add bending and distort the delayed-fracture behaviour the test is looking for.
• 2The load is a percentage of the fastener's minimum ultimate tensile load — 75 % is the common value, some customer specifications go higher — and the hold is commonly 48 hours or more.
• 3Timing matters. The test starts after coating and after baking, with a defined waiting period, because hydrogen redistribution is a function of time.
• 4Sample size is a statistics question, not a metallurgy question. A handful of parts out of a 500 000-piece lot detects only gross embrittlement.
• 5The primary defence is process control — pickling time and chemistry, current density, prompt baking with logged time and temperature. The test is confirmation, not protection.

What to write on the drawing or PO

• Whether the test is required at all, on which lots, and who performs it — customer, coater, or an independent laboratory.
• The applied load as a percentage of the minimum ultimate tensile load, the hold time, and the acceptance rule.
• The waiting period between coating or baking and the start of the test.
• Sample size per lot and what happens on a single failure: lot rejection, 100 % re-bake, or re-test.

What goes wrong when it is cited loosely

• Using a wedge-loading fixture and calling it ISO 15330.
• Running the test immediately after plating, before baking and before the specified waiting time — the result then means very little.
• Reading a pass as a certificate for the whole shipment.
• Specifying the test without the load percentage or the hold time, so two laboratories produce two different answers.