In short
A BS 8539 proof test checks that anchors were installed correctly. The proof load is the characteristic action N_Ek times 1.5 where at least 2.5% of anchors are tested, or 1.25 where at least 5% are. Test at least 3 anchors in each discrete area. An anchor passes if it holds the load without movement or damage.
This guide explains how site tests are planned and carried out under BS 8539:2012+A1:2021. It does not replace the standard, the anchor manufacturer's data, or the instructions of the specifier or engineer responsible for the job, who decides the test plan, the loads and the factors. A test records what happened on the day; it does not approve a fixing for use.
Key figures
| Proof load, at least 2.5% (under 5%) of anchors tested | N_p = 1.5 × N_Ek |
|---|---|
| Proof load, at least 5% of anchors tested | N_p = 1.25 × N_Ek |
| Upper limit on the proof load | 1.5 × N_rec, the manufacturer's recommended resistance |
| Smallest sample | 2.5% of the anchors, and at least 3 in each discrete area |
| After 1 failure in an area | 5% of that area's anchors, and at least 6 |
| After more than 1 failure in an area | Test 100% of the anchors (the area or the whole job: the specifier decides) |
| Time to reach the proof load | 30 seconds, applied steadily |
| Closest bridge support, proof tests in concrete | L at least 0.5 × h_ef, which suits most proof tests |
What is a proof test?
A proof test is a site test that loads an installed anchor to a set proof load to check that it was installed correctly. BS 8539:2012+A1:2021 calls up proof tests in Clause 9.3 and sets out the method in Annex B.3.
A proof test checks the quality of installation only. It does not show that the anchor suits the base material: that comes from the anchor's European Technical Assessment (ETA) or from tests to find an allowable resistance. See when you need to site test.
How is the proof load worked out?
The proof load is the characteristic action multiplied by a proof factor: Np = NEk × ν_P,test (equation B.11 of BS 8539).
- N_Ek, the characteristic action, is the load the design applies to one anchor, without safety factors. It is sometimes called the unfactored load. The specifier supplies it.
- ν_P,test, the proof factor, is chosen by the engineer responsible for specifying the fixing. BS 8539's recommended values are 1.5 where at least 2.5% (but less than 5%) of the anchors are tested, and 1.25 where at least 5% are. Testing more anchors allows a lower load.
The proof load should not be more than 1.5 times the manufacturer's recommended resistance, Nrec, for that anchor and size in that base material, where Nrec is known. N_rec must take account of any reduction for close spacing or edge distance.
Our test-load calculator page sets out this method beside its preliminary test calculator.
How many anchors should be proof tested?
Proof test at least 2.5% of the anchors on the job, and never fewer than 3 in any discrete area. Pick them at random.
Split the job into discrete areas wherever something changes that could affect how the anchors went in: a different installing team or a different anchor, a change of wall or slab material, or an elevation that has weathered differently from the rest. Each area gets its own minimum of 3.
Worked example: 160 anchors on two elevations
| Step | 2.5% sample, factor 1.5 | 5% sample, factor 1.25 |
|---|---|---|
| Characteristic action N_Ek (from the specifier) | 2.4 kN | 2.4 kN |
| Proof load N_p = N_Ek × ν_P,test | 2.4 × 1.5 = 3.6 kN | 2.4 × 1.25 = 3.0 kN |
| Limit 1.5 × N_rec, with N_rec = 2.8 kN | 4.2 kN, so 3.6 kN is allowed | 4.2 kN, so 3.0 kN is allowed |
| North elevation, 60 anchors | 2.5% is 1.5, so the minimum of 3 applies: 3 | 5% is 3: 3 |
| South elevation, 100 anchors | 2.5% is 2.5, rounded up: 3 | 5% is 5: 5 |
| Anchors to test in total | 6 | 8 |
How is a proof test carried out?
A proof test is carried out by loading the anchor slowly and steadily to the proof load, reaching it in 30 seconds, and watching the anchor and the base material.
- Check the anchor is the one specified, and confirm the proof load with whoever asked for the test.
- Set the bridge over the anchor. For proof tests in concrete, the closest support can be as close as 0.5 × hef to the anchor's centre line, which suits most proof tests. hef is the effective embedment depth. In masonry, set the feet on the neighbouring units where you can; with large units, the closest support can come in to 0.5 × h_ef. See choosing a bridge.
- Connect the tester to the anchor so the load is applied along the anchor's axis. See connecting the tester.
- Apply the load progressively. Note the load if the anchor starts to move.
- At the proof load, record the load and look for movement, cracks or damage.
- Release the load, remove the tester and record the result.
Detailed measurement of movement is not usually needed in a proof test, because any movement at all is a fail. A visual check, or a feeler gauge at the joint between the anchor and the base material, is usually enough.
What counts as a pass or a fail?
An anchor passes a proof test when it holds the proof load without movement and without any damage or deformation to the anchor or the base material. Any anchor that moves or is damaged is recorded as a failure.
Mark or remove any anchor that fails, so it is not put into service.
What happens when an anchor fails a proof test?
When an anchor fails a proof test, find out why and test more anchors. A correctly chosen and correctly installed anchor should not fail its proof load, so treat any failure as a sign that something in the choice or the installation may be wrong.
- 1 failure in an area: investigate the cause, and raise the number tested in that area to 5% of its anchors, and at least 6.
- More than 1 failure in an area: test all the anchors, find the reasons, and reconsider the specification. BS 8539 says 100% and does not limit it to the area, so ask the specifier whether it means the area or the whole job.
Tell the specifier and the installer what you found, so they can put it right. Before a failed anchor is replaced, the specifier confirms what goes in, since the specification may change; a competent installer then fits it. Drilling new holes is the norm, in positions the specifier sets. Re-using the old hole needs the manufacturer's or a specialist's agreement first.
The tester reports what happened. The report records facts and does not approve the fixing. See what goes in a site test report.
More questions
BS 8539 recommends a proof load factor ν_P,test of 1.5 where at least 2.5% (but less than 5%) of anchors are tested, and 1.25 where at least 5% are. The engineer who specifies the fixing chooses between them. The proof load should not be more than 1.5 times the manufacturer's recommended resistance.
Anchors with a projecting threaded stud can be. A threaded coupler connects to the stud above the nut, and the nut stays tightened to the installation torque. When the load reaches the clamping force, the washer under the nut comes loose, which marks first movement. If the proof load is above the anchor's clamping force, the washer will come loose during the test, so agree with the specifier before testing how that is judged. Anchors with a bolt head cannot be proof tested without taking the bolt out, so consider specifying the projecting-stud version of the same anchor.
Under BS 8539, proof tests are not needed when an anchor with a European Technical Assessment has been designed, selected and installed to that assessment by a competent person under supervision. Many specifications ask for them anyway. For scaffold ties, NASC TG4:25 asks for proof tests on every project.
No. The "proof load" in BS 8539 has nothing to do with the proof load tests in Section 3.2.3 of EOTA Technical Report TR 053, which belong to a separate procedure for metal injection anchors in masonry.
No. In concrete, the closest bridge support can be as close as 0.5 × hef to the anchor in a proof test, which suits most cases and lets you watch the joint between the anchor and the concrete. The reduction factor αdist is not used in proof tests.
Any movement during a proof test is a failure and must be recorded as one. A correctly installed anchor should show no movement at its proof load.
Sources
- BS 8539:2012+A1:2021 Code of practice for the selection and installation of post-installed anchors in concrete and masonry, Clause 9.3 and Annex B.3, BSI
- EOTA TR 053:2022-07 Recommendations for tests of metal injection anchors for use in masonry to be carried out on construction works, Section 3.2.3, EOTA
- NASC TG4:25 Anchorage systems for scaffolding, Section 9.6, NASC
Proof test to 25 kN or 60 kN
The Staht t25 and t60 show live and peak load on one screen and connect by Bluetooth to the Staht Connect app.