In short
Under BS 8539, anchors are site tested to find an allowable resistance where the maker has no data for the base material, or to proof test installation, on at least 2.5% of anchors and at least 3 per area. An anchor with a European Technical Assessment (ETA), used within it and installed by a competent person under supervision, needs neither.
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
| Purposes of a site test | 2 (allowable resistance; quality of installation) |
|---|---|
| Anchors in a preliminary test series | 5, installed only for the test |
| Smallest proof test sample (BS 8539) | 2.5% of anchors, and at least 3 in each discrete area |
| Smallest proof test sample (NASC TG4:25, scaffold ties) | 5% (1 in 20), and at least 3 in each discrete area |
| Ω where the anchor is designed for concrete in tension (cracked concrete) | 0.7 |
Why are anchors tested on site?
Anchors are tested on site to find an allowable resistance or to check the quality of installation, and BS 8539:2012+A1:2021 treats the two separately.
- To find an allowable resistance. The manufacturer has no load data for the base material on the job, so tests decide what load the anchor may carry there. BS 8539 calls these up in Clause 9.2, with the methods in Annex B.2. Read preliminary tests and the test load, or work one out with our test-load calculator.
- To check the quality of installation. Anchors already installed on the job are loaded to a proof load, to show they were set correctly. BS 8539 calls these up in Clause 9.3, with the method in Annex B.3. Read proof testing explained.
A proof test does not show that an anchor suits a base material. It only shows that the anchors tested were installed properly. Choosing an anchor is the specifier's job, and the specifier, or the engineer responsible for the job, decides which tests are needed.
When is no test needed to find an allowable resistance?
No test is needed to find an allowable resistance when the anchor has an ETA and the base material meets the ETA's conditions. The load data in the ETA then applies.
For masonry, "meets the conditions" means two things. The units on site must be at least as strong as the ETA's figure, and no smaller than the sizes it lists. If the units are weaker or smaller, tests are needed, under BS 8539 Annex B.2.2.
Concrete rarely needs these tests, because BS 8539 recommends anchors with an ETA there. The exceptions are a job where no suitable ETA anchor exists, and concrete whose condition puts the ETA's figures in doubt.
When is no proof test needed?
Under BS 8539, a proof test is not needed when an anchor with an ETA has been designed and selected for the job and installed as the ETA requires, by a competent, supervised installer.
Other guidance can ask for more. For scaffold ties, NASC TG4:25 asks for proof tests on every scaffolding project, including those with anchors that were already in place. See scaffold tie proof testing. Follow the specification for the job: if it calls for proof tests, they are needed.
Which anchors are not usually tested?
Three kinds of fixing are not usually site tested, and some of them cannot practically be proof tested at all:
- anything cast in: anchor bolts and channels set in the concrete when it was poured;
- fixings driven or screwed straight in: gas- or powder-actuated pins, self-drilling screws and small concrete screws;
- light-duty plastic plugs used for non-structural work.
If you are unsure whether a fixing on your job needs a test, the structural engineer and the anchor manufacturer are the people to ask.
What can a site test not show?
A site test cannot show how an anchor will behave in cracked concrete. Tests on site prove performance only in uncracked concrete. Where the anchor is designed for concrete in tension, where cracks are expected, the specifier applies a condition factor Ω of 0.7 in tests to find an allowable resistance. It raises a preliminary test load, or lowers a resistance worked out from tests to failure. Proof loads do not use Ω.
Shear is a separate question. A preliminary test load is set for tension alone, so if the anchor will also carry shear, ask the responsible engineer whether to raise the test load to leave margin for both. Shear tests themselves are rare: see shear testing.
Which test does the job need?
| Situation on the job | Test needed | BS 8539 reference |
|---|---|---|
| Anchor with an ETA, base material within the ETA | No test to find an allowable resistance | None needed |
| Anchor with an ETA, masonry weaker or smaller than the ETA states | Tests to failure: 15 or more to the EAD and its Technical Report, or 5 to 15 with statistics | Annex B.2.2 and B.2.2.1 |
| No anchor with a relevant ETA for the base material, or concrete in doubtful condition (the anchor type must be approved by its maker for that base material) | Preliminary tests on 5 anchors, or 5 to 15 tests to failure with statistics | Annex B.2.3.1 and B.2.3.2 |
| Anchors installed, quality of installation to be checked | Proof tests on a sample | Clause 9.3 and Annex B.3 |
| ETA anchor designed, selected and installed to its ETA by a competent person under supervision | No proof test under BS 8539 | None needed |
Who decides whether tests are needed?
The specifier, or the engineer responsible for the job, decides whether tests are needed, which type, how many and to what load. The tester carries out the tests and reports what happened. A test report records facts. It does not approve the fixing for use, and the tester should not draw that conclusion unless qualified to do so. See what goes in a site test report.
More questions
Yes. An anchor with an ETA needs tests to find an allowable resistance when the base material is outside the ETA's conditions, for example masonry units that are weaker or smaller than the ETA states. It may also need proof tests if it was not installed by a competent person under supervision, or if the job specification asks for them.
No. A proof test loads an installed anchor to a set proof load and checks it holds without movement or damage. A test to failure, sometimes called a pull-out test, loads an anchor until it fails, to measure its resistance. Tests to failure are used to find an allowable resistance, never on anchors that will be used on the job.
No. A proof test shows only that the anchors tested were installed properly. Whether an anchor suits a base material is shown by its ETA or by tests to find an allowable resistance.
BS 8539 recommends that every part of anchor use, testing included, is done by competent people. Industry training schemes assess and certify testers. Check that the tester is qualified for the type of test asked for. Training schemes can require a higher grade of tester for tests to find an allowable resistance than for proof tests.
Scaffold ties follow NASC TG4:25, which lists BS 8539 among its references but sets its own factors and sample sizes for scaffolding. For example, it asks for proof tests on at least 5% of tie anchors on every project. See scaffold tie anchors.
Sources
- BS 8539:2012+A1:2021 Code of practice for the selection and installation of post-installed anchors in concrete and masonry, Clauses 9.2 and 9.3 and Annex B, BSI
- NASC TG4:25 Anchorage systems for scaffolding, Sections 9.5 and 9.6, NASC
Test anchors with the right kit
Staht's t25 and t60 digital pull testers proof test anchors up to 25 kN and 60 kN, with bridges for wider support spacing.