First movement, load relaxation and the ways anchors fail

First movement is the load at which an anchor begins to shift in the base material, and on site it is usually read at 0.1 mm of movement. It matters because an anchor that moves under its working load may be strong enough but still unsuitable, so BS 8539:2012+A1:2021 asks for it in every site test. (Site tests to a European Assessment Document don't ask for it.)

You don't need to pin the exact moment down. Once an anchor starts to move you can usually see it, and the load barely changes between 0.1 and 0.2 mm.

Expect a high figure where the anchor clamps something to the surface. On a bracket bolted tight to a wall, for example, a sound anchor usually shows no significant movement until the load is well past the manufacturer's recommended resistance.

On most tests, watch the joint where the pulling plate meets the base material, by eye or with a feeler gauge, and note the load when it opens. If you are pulling from a coupler above a nut tightened to its installation torque, watch the washer under that nut instead: the load at which it goes slack is your first movement.

When the engineer wants movement measured properly, set up a dial gauge:

1. Choose a gauge accurate to 0.02 mm or better.
2. Stand it on its own base, clear of the bridge (which flexes under load) and of any concrete that might lift. Never fix it to the bridge.
3. Point it at the anchor as directly as you can; a pulling frame with the gauge inside, bearing on the bolt head, works well.
4. Note the reading at zero load, then load in at least 10 steps up to the expected failure load.
5. At each step, wait for load and movement to settle, then write both down together.

A digital tester with a displacement module can draw the load and movement graph for you. Detailed measurement makes each test much slower, so it is only done when the engineer asks for it, as some safety-critical work does, so the movement can be compared with the anchor's European Technical Assessment.

Load relaxation is the reading drifting down on a sound anchor once you stop loading. It is normal, and it is why, when you load in steps, you wait for the reading to settle and then read load and movement together. The cause is stress around the anchor spreading into the base material; it slows over the anchor's life but never quite stops.

Reload to the same level: a relaxing anchor barely drops again, while a failing one drops again and slips further out.

This check matters most with resin anchors. They do not fail suddenly: after the bond has gone, it can still take a large force to pull the anchor out, so a failed bonded anchor can look as if it is still holding.

Apply the load steadily, reaching a proof load in about 30 seconds or a failure load in 60 to 90 seconds, unless movement is being measured in steps.

Agree the stop rule before a test to failure or a preliminary test. Once the load passes the clamping force from tightening, some movement is normal; the engineer says whether to record it. The engineer may also set a movement limit that counts as failure: stop there, and state the limit in the report. When the 5 preliminary test anchors are loaded to failure after one has failed, stop at about 1.0 mm to limit damage, and take that load as the failure load. See preliminary tests.

Anchors fail in one of three places: the anchor itself, its grip in the hole, or the base material around it. Shear tests add one more mode, pry-out, where the anchor is levered out. Name the failure by what gave way.

Record the failure mode, the load and, where a cone of material broke away, its size. Describe what the anchor and the surrounding material look like afterwards, cracks included. If the anchor did not fail, release the load and record any movement that remains.

Remove or mark every anchor that was tested to failure, or that failed to reach its proof load, so that it is never put into service. The test report records what happened. It does not say why an anchor failed or what to do next unless the tester is qualified to judge that, and it does not approve the fixing. See what goes in a site test report.

No movement is allowed in a proof test. Any movement seen must be recorded, and it counts as a failure. A visual check is usually enough. Our test-load calculator page shows how proof and preliminary test loads are worked out. See proof testing explained.