“Can you carry out a pull-out test on these anchors?” is one of the most common requests received on construction projects. The problem is that pull-out test can mean very different things depending on who is asking.
A contractor may mean a non-destructive proof test on installed production anchors. An engineer may instead be asking for sacrificial anchors to be loaded towards failure so that the behaviour of an existing concrete or masonry substrate can be investigated. Both involve applying tensile load to an anchor, but they answer completely different engineering questions.
For UK projects, this distinction matters because BS 8539:2012+A1:2021 deals with the selection and installation of post-installed anchors in concrete and masonry, including circumstances where site testing forms part of the fixing strategy.
The key difference: a proof test is normally a controlled, non-destructive check on an installed production anchor that is intended to remain in service. A destructive or ultimate test deliberately investigates performance towards failure using a sacrificial fixing. Asking simply for a “pull-out test” does not define which one is required.
Jump to: Terminology | Proof testing | Destructive testing | Direct comparison | Which test do you need? | Test loads | Site procedure | Understanding results | Common mistakes | UKAS and competence | Checklist | FAQ
Why “Pull-Out Testing” Causes So Much Confusion
The phrase pull-out testing is widely used on UK construction sites because the physical action is easy to understand: a testing rig applies tensile force to a fixing. But the phrase does not, by itself, explain the engineering objective. A testing company could be asked to apply a predetermined proof load to an anchor that must remain in service, or to continue loading a sacrificial fixing towards a defined failure condition so that site-specific performance can be investigated. Those are not interchangeable tests. They involve different anchors, different loading objectives, different acceptance criteria and different uses of the resulting data.
| Term | Practical Meaning | Main Purpose |
|---|---|---|
| Proof testing | Controlled tensile loading of an installed production anchor to a specified proof load. | Installation and workmanship quality control. |
| Pull-out testing | A broad site term for tensile testing of anchors. | Must be clarified before testing because it may refer to proof or destructive testing. |
| Preliminary / site testing | Testing sacrificial anchors in representative site substrate before or during design development. | Provide site-specific evidence where substrate performance is uncertain. |
| Destructive / ultimate testing | Loading a sacrificial fixing towards a defined failure condition. | Investigate resistance, displacement and failure behaviour. |
What Is Anchor Proof Testing?
A proof test is normally carried out on a production anchor that is already installed and is intended to remain part of the completed works. The anchor is subjected to a predetermined tensile proof load rather than deliberately being taken to failure. The objective is generally to provide quality-control evidence that the installed fixing can sustain the specified test load without unacceptable movement, slip or distress.
Proof testing is particularly useful where the project requires independent verification of installation quality. It may identify problems associated with drilling, hole cleaning, resin installation, curing, mechanical anchor setting or other workmanship-related issues. However, a successful proof test does not establish the anchor's ultimate resistance. It confirms the behaviour of the tested fixing up to the specified proof load under the conditions of the test.
Important: if a production anchor is intended to remain in service, the test should not be treated as an opportunity to “see how much it can take”. That would change the purpose of the test and could damage the fixing or surrounding substrate.
What Is Destructive Pull-Out Testing?
Destructive testing has a different purpose. Instead of checking an installed production anchor against a predetermined proof load, a sacrificial anchor is used to investigate how the anchor and substrate behave as the applied tensile load increases towards failure. The test may be appropriate where the existing substrate is uncertain, highly variable or not adequately represented by published performance data. Older brickwork, historic masonry and refurbishment projects are common examples where site-specific evidence may be required before the engineer confirms the fixing strategy.
During destructive testing, the useful information may include the maximum load reached, displacement behaviour and the failure mechanism. Depending on the anchor and substrate, failure could involve anchor pull-out, bond failure, concrete breakout, masonry failure, anchor steel failure or another mechanism. Because the purpose is to investigate failure behaviour, the test fixing is sacrificial. It should not subsequently be treated as a production anchor carrying permanent loads.
Proof Testing vs Destructive Pull-Out Testing: Direct Comparison
| Engineering Question | Proof Test | Destructive / Ultimate Test |
|---|---|---|
| Main objective | Quality-control verification of an installed fixing. | Investigate resistance and failure behaviour in the actual substrate. |
| Anchor type | Production anchor. | Sacrificial test anchor. |
| Intended to remain in service? | Normally yes, if the result is satisfactory. | No. |
| Loading objective | Reach and sustain a specified proof load without unacceptable behaviour. | Increase load towards an agreed failure or test termination condition. |
| Is failure expected? | No. A production anchor is not deliberately tested to failure. | Failure or significant displacement may form part of the intended investigation. |
| Typical output | Test load, behaviour and test outcome. | Load, displacement, failure load and/or failure mode as required by the test brief. |
| Does the result equal allowable working capacity? | No. | No. Engineering interpretation and appropriate safety factors are still required. |
Which Anchor Test Does the Project Actually Need?
The answer depends on the engineering question. The testing provider should not choose a destructive test simply because the contractor asked for a “pull test”, nor should a proof test be used where the engineer actually needs evidence about an unknown substrate.
| Project Situation | Possible Testing Approach | Reason |
|---|---|---|
| Production anchors already installed | Proof testing where required by the design or specification. | The anchors are intended to remain in service. |
| Unknown historic brickwork or masonry | Preliminary or destructive testing may be considered using sacrificial anchors. | The engineer may need site-specific substrate evidence. |
| Suspected installation workmanship issue | Proof testing of a defined population may be specified. | The objective is installation quality verification. |
| Trial anchors before production installation | Preliminary or site-specific testing. | Results can inform the design before production anchors are installed. |
| Handover QA check | Proof testing where specified. | Provides evidence that selected installed anchors sustained the specified proof load. |
| Scaffold ties | Follow the scaffold design and applicable NASC testing regime. | Scaffold tie testing has application-specific requirements and should not be reduced to a generic BS 8539 percentage. |
Before the testing team mobilises, the project should be able to answer one question: are we checking the quality of anchors that need to remain in service, or are we investigating the performance of the anchor/substrate system? The answer determines the type of test.
Proof Load Is Not the Same as Failure Load
One of the most serious mistakes in anchor testing is treating every load value as though it means the same thing. A proof load is the predetermined load that an installed anchor is required to sustain during a proof test. An ultimate or failure load is associated with the resistance reached during destructive investigation. Neither figure should automatically be described as the anchor's allowable working load.
For proof testing, the required test load should be derived from the project design or specification. The testing contractor should not look at an unidentified load on a drawing and decide independently what multiplier to apply.
In particular, instructions such as “test it to 1.5 times the load” are incomplete unless everybody understands exactly what the underlying load represents and why that test factor is being used. Where the basis is unclear, the structural engineer, façade engineer, temporary works designer, anchor designer or other responsible specifier should confirm the test load before testing starts.
A test rig reading is not a design capacity. If a sacrificial anchor reaches 30 kN during testing, that does not automatically mean the finished connection has a 30 kN allowable working load. The test result requires engineering interpretation in the context of the design, substrate, anchor system and applicable safety factors.
How the Two Tests Differ on Site
Typical Proof-Test Sequence
- Confirm the anchor identification, substrate, specified proof load, sampling regime and acceptance criteria.
- Check that any bonded anchor has achieved the manufacturer's required curing period.
- Inspect the fixing and surrounding substrate for visible defects.
- Connect suitable calibrated tensile testing equipment.
- Position the reaction arrangement so that it does not artificially confine the potential local failure zone.
- Align the applied tensile force as closely as practicable with the anchor axis.
- Increase the load progressively to the specified proof load.
- Observe the anchor, applied load and substrate for the required test period.
- Unload in a controlled manner.
- Record the result against the exact anchor location and equipment used.
Typical Destructive-Test Sequence
For destructive or preliminary testing, sacrificial anchors are installed at agreed representative locations. After the appropriate installation and curing requirements have been satisfied, tensile load is applied in accordance with the agreed testing procedure.
Unlike a proof test, loading may continue beyond normal service or proof levels towards the defined termination or failure condition. The tester records the load and, where required, displacement and failure behaviour so that the responsible engineer can interpret the results. The tested anchor should then be treated as sacrificial rather than incorporated into the permanent load-carrying installation.
Why the Reaction Arrangement Matters
Both types of test can be compromised by poor rig setup. If the reaction bridge bears too close to the fixing, its compressive force can restrain the concrete, brick or masonry surrounding the anchor and artificially influence the result. Likewise, significant eccentricity can introduce unintended bending rather than a predominantly axial tensile test. The reaction arrangement therefore needs to suit the fixing, substrate and purpose of the test.
Testing is a controlled loading operation. Wider site controls should therefore be reflected in the RAMS and exclusion arrangements. For further guidance, see What a Structural Testing RAMS Must Contain Before Loading Starts.
What Does a Successful Proof Test Actually Prove?
A proof test provides evidence about the behaviour of the specific fixing under the specified test load. For the result to be meaningful, the anchor should reach the specified proof load and remain stable after the initial seating or bedding-in of the testing arrangement. Progressive anchor movement, permanent slip, cracking, spalling or other substrate distress should be recorded and assessed against the acceptance criteria.
Initial mechanical seating of couplers, threads and reaction components should not automatically be confused with movement of the anchor itself. This is why experienced testing personnel and appropriate test procedures matter.
| A Proof Test Can Demonstrate | A Proof Test Cannot Automatically Demonstrate |
|---|---|
| The tested anchor sustained the specified proof load. | The anchor's ultimate failure resistance. |
| Quality-control evidence for that installation. | That every untested anchor on the project is satisfactory. |
| Behaviour during the defined site test. | Long-term resistance to corrosion, creep, fatigue, fire or environmental degradation. |
| Evidence forming part of an agreed sampling regime. | That an unsuitable design or wrongly selected anchor has become acceptable. |
What Does a Destructive Test Result Mean?
A destructive test can provide information about site-specific resistance and failure behaviour, but the highest number shown on the gauge should not simply be transferred onto a drawing as the allowable capacity of the fixing. The engineer needs to consider the number and distribution of tests, substrate variability, observed failure mechanisms, anchor configuration and the design framework before deriving any value used for the permanent works. This is particularly important in existing masonry. Two visually similar sections of historic wall can behave very differently because of changes in brick strength, mortar condition, voiding, previous repairs or construction build-up.
Common Mistakes When Ordering Anchor Testing
| Common Request | Why It Is a Problem | Better Instruction |
|---|---|---|
| “We need a pull-out test.” | The purpose of the test is undefined. | State whether proof, preliminary or destructive testing is required. |
| “Pull it until it fails.” | A production anchor may be permanently damaged. | Use designated sacrificial anchors for destructive investigation. |
| “Test it to 1.5 times the load.” | The basis of the underlying load and test factor has not been defined. | Provide the engineer-approved test load and testing regime. |
| “It reached 20 kN, so its capacity is 20 kN.” | A test result is not automatically an allowable design resistance. | Have the test data interpreted by the responsible designer. |
| Testing chemical anchors immediately after installation. | The adhesive may not have achieved the required cure. | Follow the manufacturer's curing requirements for actual site temperature. |
| Using one successful test to approve every fixing. | Different substrates, installers and anchor populations may behave differently. | Use the project-specific sampling and acceptance regime. |
UKAS, Calibration and Tester Competence
Calibration and accreditation are related quality issues, but they are not the same thing. Testing equipment should have appropriate calibration and traceability for the work being undertaken. The final report should clearly identify the equipment used so that the measured result can be traced back to the relevant equipment records.
UKAS accreditation, where required, relates to an organisation's competence for a defined accredited activity and scope. BS 8539 should not be described as imposing a universal statutory rule that every routine site proof test must be carried out by a UKAS-accredited organisation.
However, a client, principal contractor, infrastructure authority, consultant or project specification may make accreditation a contractual requirement. This should therefore be checked before the testing provider is appointed. Competence remains important regardless of accreditation status. Anchor testing requires the technician to understand rig setup, load application, substrate behaviour, equipment limitations, test validity and reporting, not simply how to operate a hydraulic pump.
What to Send Before Requesting a Quote
The clearest testing enquiries contain enough information to determine what the project is actually asking for.
- Test purpose: proof testing, preliminary/site testing or destructive testing.
- Anchor specification: manufacturer, product, diameter, grade and fixing type where known.
- Substrate: concrete, brick, block, natural stone or other material.
- Number of anchors: total installation population and proposed test quantity where specified.
- Required test load: engineer-approved proof load or destructive-test brief.
- Acceptance criteria: project-specific pass/fail or reporting requirements.
- Anchor locations: drawings, elevations, grid references or marked positions.
- Anchor status: production anchors that must remain in service or sacrificial test anchors.
- Resin information: product and installation time where chemical anchors are involved.
- Access: working height, scaffold requirements and available clearance around each fixing.
- Reporting requirements: photographs, equipment records, location schedules or other project QA requirements.
- Accreditation requirements: confirm any contractual UKAS or project-specific requirements before mobilisation.
Need Anchor Testing?
For an accurate quotation, provide the anchor type, substrate, number and location of tests, required test load, whether the fixings are production or sacrificial anchors, and any project-specific acceptance or accreditation requirements. If the engineering test regime has not yet been defined, it should be confirmed by the responsible designer or specifier before site testing begins.
Three Anchor Testing Myths to Avoid
Myth 1: “Pull-out testing always means pulling the anchor out.”
No. “Pull-out testing” is commonly used as a broad site description for tensile anchor testing. Many site requests using this phrase actually require a non-destructive proof test.
Myth 2: “If the anchor reached 30 kN, its safe working capacity is 30 kN.”
No. The load reached during a proof or destructive test is test data. It must be interpreted within the relevant design framework before any allowable design resistance is established.
Myth 3: “An ETA means site testing can never be required.”
An ETA provides technical assessment information for the anchor system and defined conditions. Whether site proof testing is required still depends on the project design, substrate, installation controls, supervision and specification. The testing requirement should therefore be confirmed by the responsible specifier rather than assumed from the ETA alone.
Evidence-Based Summary
Proof testing and destructive pull-out testing answer different engineering questions.
Proof testing is normally performed on production anchors intended to remain in service. The test checks their behaviour against a specified proof load.
Destructive testing uses sacrificial anchors and investigates performance towards a defined failure or termination condition.
The phrase “pull-out test” is too broad to form a complete technical test instruction on its own.
Neither a proof-test load nor a destructive peak load should automatically be treated as the allowable working capacity of the finished fixing.
The test type, load, sample regime and acceptance basis should be defined before the testing equipment is connected.
FAQ: Anchor Proof Testing vs Pull-Out Testing
Is a proof test the same as a pull-out test?
Not necessarily. “Pull-out test” is commonly used as a broad description of tensile anchor testing. A proof test is specifically intended to apply a defined load to an installed anchor without deliberately taking it to failure.
Does proof testing damage the anchor?
A properly specified proof test is intended to be non-destructive. If the anchor displays unacceptable movement, slip or substrate distress, the result should be investigated before the fixing is accepted for service.
Does a proof test establish ultimate anchor capacity?
No. A proof test demonstrates performance up to the specified proof load under the test conditions. It is not an ultimate capacity test.
When would destructive anchor testing be required?
It may be required where the designer needs site-specific resistance or failure information, particularly where the substrate is unknown, existing or highly variable. The requirement should be defined by the responsible engineer or specifier.
Can a destructive test be carried out on a production anchor?
A fixing deliberately loaded towards failure should be treated as sacrificial. Production anchors intended to remain in service should not normally be used for destructive capacity investigation.
Who decides the anchor test load?
The engineering test requirement should normally be defined through the structural engineer, façade engineer, temporary works designer, anchor designer or other competent specifier responsible for the connection. The site tester should not invent the design basis.
If an anchor reaches 20 kN during a destructive test, is its working capacity 20 kN?
No. The measured test load is evidence for engineering interpretation. Appropriate design methods, safety factors, substrate variability and the wider fixing configuration still need to be considered.
Can resin anchors be tested immediately after installation?
No. The manufacturer's required curing time must be achieved before loading. Cure times can vary substantially with product and substrate temperature.
Does an ETA remove the need for all site testing?
Not automatically. The need for site testing depends on the design, substrate, installation controls, supervision and project specification. The responsible specifier should confirm the testing requirement.
Does BS 8539 require every test to be UKAS accredited?
BS 8539 should not be treated as creating a universal statutory requirement for every routine anchor proof test to be undertaken by a UKAS-accredited organisation. However, UKAS accreditation may be required by the project specification or contract.
Primary Technical References
The principal reference for post-installed anchors in UK construction is BS 8539:2012+A1:2021 — Code of practice for the selection and installation of post-installed anchors in concrete and masonry. Practical site-testing guidance is also published by the Construction Fixings Association. Where scaffold anchors are involved, the applicable NASC guidance and scaffold design requirements should also be checked. The official licensed British Standard and applicable manufacturer technical documentation should be consulted for detailed requirements. This article does not reproduce the complete test procedures, calculations or acceptance provisions contained in those documents.
Source Context and Technical Note
This article is a STRUCTinspect technical explainer intended to clarify the practical difference between proof testing, preliminary testing and destructive anchor testing on UK construction projects. Anchor testing does not replace appropriate fixing selection, structural design, competent installation, inspection or manufacturer requirements. The usefulness of a test depends on the engineering question being clearly defined before testing begins. This article does not provide structural design, temporary works, legal or contractual advice. Test type, test loads, sampling regimes, acceptance criteria and interpretation of destructive test data should be confirmed by the engineer, designer, specification author or other competent professional responsible for the fixing system.
