ISO 898-1

What the test does

ISO 898-1 is a specification with its test methods built in. It does not describe one test: it defines the property classes a bolt can be sold as, then sets out the tests that prove a bolt belongs to the class stamped on its head.

A full-size fastener is pulled to failure in the tensile test, which runs to ISO 6892-1. The crosshead speed shall not exceed 25 mm/min, and a lower limit of 10 mm/min applies where yield strength or offset yield is being determined — slower, because a yield reading taken too fast reads high.

Alongside it the standard carries a proof load test, head soundness, hardness by Vickers, Brinell or Rockwell, a torsion test and a notched impact bending test. Testing is at ambient temperature, 10 °C to 35 °C.

The proof load test is the one most often run in production, and it is not a strength test at all. The fastener is loaded to its specified proof load, held for 15 seconds, and released. It passes if it has not permanently stretched — the permitted measurement uncertainty being ±12.5 µm. A bolt that grows by more than that has yielded, whatever load it might eventually have reached.

What it measures, and why it matters

The property class is the whole point. It is written as two numbers separated by a dot, and both halves carry information: the number left of the dot is the nominal tensile strength in megapascals divided by 100, and the number to the right is ten times the ratio of nominal yield strength to nominal tensile strength. So a class 8.8 bolt has a nominal tensile strength of 800 MPa and a nominal yield at 80 % of it. Class 10.9 means 1000 MPa with yield at 90 %.

That second digit is what an engineer is really buying. Two bolts can reach the same breaking load and behave quite differently in a joint, because one yielding at 90 % of its ultimate has almost no plastic reserve between working load and failure.

Wedge loading is where head quality shows. The bolt is pulled through an angled washer, so the head is loaded off-axis as a slightly non-parallel joint face would load it; angled mounts come in 4°, 6° and 10°. A sound head-to-shank radius survives it. One with a forging defect snaps at the underhead fillet well below its axial capacity — the failure a straight pull misses.

Specimen

The fastener itself, full size and as manufactured — the standard tests the product, not a coupon machined from the same bar.

Form

Full-size bolt, screw or stud, as manufacturedThread form, head geometry, underhead radius and whatever heat treatment and coating did to them are all inside the specimen, because they are all inside the failure.

Machined test pieces

Permitted in circumstances the standard sets outThey answer a narrower question: they characterise the steel rather than the bolt.

Temperature

10 °C to 35 °CAmbient testing. The range is stated in the standard's own scope.

Test speed

Two limits on the tensile test, and a hold on the proof load test.

Tensile

Shall not exceed 25 mm/minThe tensile test runs to ISO 6892-1. Speed inflates yield readings first and ultimate readings second, which is why there is a ceiling at all.

Yield determination

10 mm/min maximumLower than the general limit, because a yield or offset-yield reading taken too fast reads high.

Proof load hold

15 sThe load is applied, held, and released. The bolt passes if it has not permanently stretched.

Permanent set allowed

±12.5 µm measurement uncertaintyGrowth beyond that is yield, whatever load the bolt might eventually have reached. It is measured on the bolt with a length gauge, not from crosshead travel.

How the test runs

1. 01Condition to ambient, between 10 °C and 35 °C.
2. 02For proof load: measure the bolt's length with a gauge resolving better than 12.5 µm.
3. 03Apply the specified proof load, hold 15 s, release, and re-measure. No permanent set means a pass.
4. 04For tensile: pull to failure at no more than 25 mm/min, or 10 mm/min if yield is being determined.
5. 05For wedge tensile: fit the angled washer — 4°, 6° or 10° — so the head is loaded off-axis, and pull to failure.
6. 06Record where the bolt failed, not only the load it reached.

What travels with an ISO 898-1 result

A class is a claim about two properties, so a single load figure never proves one.

• The property class claimed, and both tests that support it.
• Tensile strength and yield or offset yield, with the speed each was run at.
• Proof load result as pass or fail, with the measured permanent set.
• For wedge tests, the washer angle used.
• Where the failure occurred — thread, shank or underhead fillet.

What the machine must be capable of

Force to break the largest fastener in scope, which for high-class bolts in structural sizes runs well into the hundreds of kilonewtons, and a crosshead that can be held below 25 mm/min and below 10 mm/min for yield work.

Fixturing matters as much as the frame. Axial testing needs threaded adapters that grip without crushing the thread; wedge testing needs the angled washer set. Proof load work needs a machine that holds a set load steady for 15 seconds, and a length gauge on the bolt itself resolving 12.5 µm — not a crosshead reading.

What goes wrong in practice

Running the tensile test too fast is the commonest — 25 mm/min feels slow on a long production run. Speed inflates yield readings first and ultimate second.

Measuring proof-load elongation on the machine rather than on the bolt is the second. The permitted uncertainty is 12.5 µm; a load frame deflects by orders of magnitude more than that.

Gripping the thread crest without support is the third: it damages the thread, moves the failure into the grip, and gives a low result that looks like a weak bolt.

Finally, reading a class from a breaking load alone. The class is a pair of numbers, and the second — the yield ratio — cannot be read from an ultimate load.

Axial or wedge tensile

Both pull the bolt to failure. Only one of them finds a bad head.

A bolt with a forging defect at the underhead fillet can pass an axial pull at full load and snap under a wedge test well below it.

Questions we are asked about this test

Related and equivalent standards

ISO 898-2 covers nuts and ISO 898-5 set screws. ISO 3506-1 is the corresponding specification for stainless fasteners, where the class notation differs. The tensile test defers to ISO 6892-1, which is why a laboratory equipped for metals testing is most of the way to this one.