Proof testing scaffold ties, and what to do after a failure

Scaffold ties are proof tested to check that the anchors were installed correctly. NASC TG4:25 asks for proof tests on every scaffolding project, including projects using anchors that were installed earlier. A proof test does not check whether the anchor suits the base material; that is the job of the anchor's data or of preliminary tests.

Test at least 1 in 20 anchors (5%) on every project, picked at random, and never fewer than 3 in any discrete area.

Treat each of these as its own area, with at least 3 tests: a different installing team, an elevation that has weathered differently, a different anchor, or a change of wall material. On repeated house types, a batch of plots can share a sample if those areas are still respected (NASC TG4:25 §9.6).

Anchors in mortar joints (1 in 10), in header bricks (double) and reused nylon plugs (double) need a higher rate from the start:

• Anchors in mortar joints, where the bricks cannot be drilled: 1 in 10 anchors across the whole job, after preliminary tests.
• Anchors in header bricks, which can be loosened by drilling: double the usual rate.
• Nylon plugs left in the wall and used again: double the usual rate.

See scaffold anchor types.

Scaffold tie anchors are proof tested to 1.25 × the tie duty, taken from the compliance sheet (TG20 or TG30) or the scaffold design. For a 6.1 kN standard-duty tie, that is 7.625 kN. See working out the tie load for every tie class. For jobs specified to BS 8539 instead, see our test-load calculator.

Apply the load directly to the anchor, not through the tie tube or fitting. The gauge should have been calibrated in the last 12 months to an accuracy of 5%. Keep everyone out of the line of fire in front of the tester while it is loaded.

TG4:25 does not set its own pass criterion for proof tests. BS 8539:2012+A1:2021, which TG4:25 lists among its references, counts an anchor as passing when it holds the proof load with no movement and no damage to the anchor or the base material. Record any movement or damage as a failure.

Yes. Tag each anchor you proof test with the test date and the load it held, and mark a failure clearly on its tag. TG4:25 §9.6 describes this as common practice, and the tags let a later inspection pick anchors that have not been tested.

After a failure, investigate and raise the rate: 1 in 10 after one, 1 in 5 after two, every anchor after more.

A failure suggests that selecting or installing the anchor may have gone wrong, so tell whoever specified the anchor and whoever installed it. Failed anchors usually need replacing by a competent installer, after the specification has been confirmed.

In the example above, one of the 6 anchors tested on the front elevation fails.

• After 1 failure: 1 in 10 of the 120 front anchors is 12, which is more than the minimum of 6, so test 12 in all.
• After a 2nd failure: 1 in 5 of 120 is 24, more than the minimum of 12, so test 24 in all.
• After a 3rd failure: test all 120 anchors on the front elevation, and review the specification and installation.

TG4:25 does not say whether the higher rate applies to the area where the failure happened or to the whole job. This example applies it to the area. The scaffold designer or responsible engineer decides.

BS 8539:2012+A1:2021 allows a choice: test at least 2.5% (but less than 5%) of the anchors at 1.5 × the characteristic action, or at least 5% at 1.25 ×. TG4:25's rule for scaffold ties, 5% at 1.25 × the tie duty, matches BS 8539's 5% option. BS 8539's minimum after one failure is also 6. See proof testing explained.