Key takeaways
- The non-destructive test validates the actual bond of the fixing without impairing the integrity of the anchor tested.
- The proof load corresponds to a controlled fraction of the characteristic resistance given in the technical assessment.
- Checking commonly covers a targeted sample representing at least five per cent of the fixings installed.
- The test report records the applied force and confirms conformity for the as-built file.
Defining the proof load and the sampling from the technical assessment
On a metalwork construction site, fixing a rail or a face-fixed balustrade to the edge of a concrete slab affects the stability of the balustrades and fall protection. While the design calculation sizes the metal anchors or resin-bonded threaded rods, only the actual installation guarantees mechanical performance under tensile forces. The European Technical Assessment, referred to by the acronym ETA, is the manufacturer's technical reference. This official document gives the characteristic resistance of the anchor for a given concrete class, a specific spacing and a defined edge distance. However, the ETA describes behaviour under ideal installation conditions: on raw slab, the presence of micro-cracks, the proximity of passive reinforcement or insufficient cleaning of the drilled hole considerably reduce these theoretical values.
- Proof load
- Axial tensile force temporarily applied to an installed fixing, calibrated to check the quality of the anchorage without exceeding the elastic limit of the steel or initiating failure of the concrete cone.
To validate the anchors without destroying the work, the metalworker carries out a direct non-destructive tensile test, known as the proof load test. This control value is not intended to cause pull-out. It generally lies between 1.2 and 1.5 times the permissible service load calculated for the balustrade, while remaining strictly below the characteristic resistance specified in the ETA. The aim is to verify that the anchor has not slipped in its hole and to make sure that torque tightening has not fractured the substrate on its underside.
| Total number of anchors on the work | Recommended sampling rate | Minimum number of fixings tested | Action required in case of slip |
|---|---|---|---|
| Fewer than 20 anchors (one-off construction site) | 15% to 20% | 3 anchors, spread out | 100% check of the batch and removal of the grouting |
| 20 to 100 anchors (landing or balcony) | 10% | 5 representative anchors | Immediate doubling of the tested sample |
| More than 100 anchors (terrace run) | 5% | 10 anchors per homogeneous zone | Stop installation and analyse the supporting concrete |
The selection of test points must not be left to chance. The foreman gives priority to the anchors in the least favourable configurations: balcony edges with a small edge distance, concrete construction joints, or end plates supporting the end posts of a handrail. Testing a methodical sample takes about one hour of work with a calibrated mechanical or oil-pneumatic pull-out tester. That hour spent on the slab prevents any later dispute about the resistance of the balustrades or the quality of the drilling.
Carrying out the pull-out test on site and drawing up the test report
The pull-out test on an installed anchor is carried out using a tensile pull-out tester fitted with a follower-needle pressure gauge or a calibrated electronic sensor. So as not to distort the result, the device must rest on a wide-base tripod. The support feet of the frame must be set apart from the anchor by at least twice the embedment depth so as not to compress the concrete in the volume where a possible failure cone would form. The connection to the anchor or the grouted rod is made by means of a threaded adapter suited to the nominal diameter of the metalwork fixing.
- Positioning the pull-out testerScrew the threaded adapter onto the anchor without applying any parasitic torque. Adjust the tripod feet to align the axis of the jack perfectly with the mechanical axis of the drilled hole.
- Gradual load increaseOperate the hydraulic pump continuously to raise the tensile force without jolts until the proof load prescribed by the technical file is reached.
- Holding under load and observationHold the tension stable for two full minutes. Watch the pressure gauge to check that there is no pressure drop and visually inspect the concrete around the washer.
- Release and dimensional checkRelease the hydraulic pressure. Using a dial gauge or a depth gauge, check that there is no permanent displacement or residual axial slip.
All results are recorded in a pull-out test report drawn up on the spot by the person in charge of installation. This document sets out the precise identification of the construction site, the trade reference and the diameter of the anchor, the corresponding ETA approval number, the strength class of the host concrete, the calibration number of the tensile device, and the values measured for each anchor in the sample. In the event of a failure at one point, the protocol requires recording the observed failure mode, doubling the sample size in the same zone and, if necessary, adapting the fixing method.
In accordance with the rules of the art of finishing work and the execution practices of the SIA 118 standard, this report is countersigned by the site management and filed in the as-built file. It provides indisputable proof that the light metalwork fixings were checked in accordance with the required specifications. In the event of a safety inspection or a claim after acceptance of the work, this technical traceability demonstrates the diligence of the metalwork workshop and relieves it of operational responsibility.
Defining the sampling and calculating the proof load according to the ETA
For the building metalworker installing roof balustrades, balcony railings or light stair strings, the resistance of the fixings does not rest solely on the manufacturers' data sheets. If the metal or chemical anchor holds a European Technical Assessment (ETA), this certification guarantees the mechanical performance in standardised laboratory concrete. On the construction site, the reality of the substrate often calls for verification: cracked concrete at the slab edge, heterogeneous aggregates, reinforcement restricting drilling, or ageing of the substrate in renovation work. Carrying out a non-destructive pull-out test on a representative sample of installed anchors makes it possible to validate the conformity of the installation and the load-bearing capacity of the substrate without altering the work.
The sizing of the sample follows a rule of technical and economic proportionality. Testing every fixing would hold up the installation work on the construction site unreasonably, while insufficient checking exposes the company to a challenge at acceptance. On an access balcony or a balcony receiving several dozen metalwork posts, the sample size is agreed with the site management during the preparation phase. The test is carried out directly on the anchors intended to take the base plates, provided that a strictly calibrated proof load is applied so as not to exceed the elastic range of the steel or damage the concrete-resin bond.
| Total number of anchors | Sample size to be tested | Proof load applied | Purpose of the test |
|---|---|---|---|
| Fewer than 20 anchors (one-off construction site) | At least 2 fixings | 1.2 to 1.5 times the permissible service load | Validation of the substrate-anchor pairing and of the cleaning method |
| 20 to 100 anchors (collective balconies) | 5% of the batch (minimum 3 fixings) | 1.2 to 1.5 times the service load according to the ETA | Check of installation consistency and absence of premature slip |
| More than 100 anchors (long runs) | 3 to 5% spread over different zones | Value defined by the civil engineer or the specifications | Detection of possible local variations in concrete compactness |
The proof load must never be confused with the ultimate failure load given in the ETA. The test on an installed anchor is a non-destructive check test. The axial tension exerted by the hydraulic pull-out tester remains bounded between the recommended service load and the characteristic load divided by the applicable safety factors. This proof tension is held for a minimum of one to two minutes in order to verify the stability of the fixing under continuous tension.
Pass criteria and traceability in the test report for acceptance of the work
Passing the test rests on precise criteria measured during the axial tensile force. A compliant anchor must take the full prescribed proof load without failure of the concrete cone, without pull-out of the threaded rod and without visible cracking around the anchor point. The residual displacement after return to zero load is the second decisive indicator: it must remain below the tolerances permitted by the anchor's technical documentation, thus confirming that no internal slip has degraded the mechanical expansion mechanism or the chemical bond.
- Setting up the calibrated equipmentConnect the pull-out cup or the threaded adapter to the sample anchor. Wedge the feet of the load-spreading bridge on sound concrete and position the micrometer dial gauge to measure the vertical displacement.
- Gradual load increaseOperate the manual hydraulic pump in regular steps until the target proof load is reached. Stabilise the pressure and hold the load for at least 60 seconds, reading the deflection on the pressure gauge.
- Checking the criteria and unloadingCheck that there is no creep or pressure drop. Release the proof load, record any residual displacement and visually inspect the concrete surface around the washer.
- Drawing up the check reportComplete the site document immediately, stating the location of the balustrade post, the ETA reference of the product, the actual drilling depth, the load reached and the joint signature of the technician and the site management.
If a fixing fails or slips abnormally during the test, the sampling must immediately be doubled on the same batch to determine whether the problem stems from an isolated drilling error (insufficient dust removal, presence of formwork release oil) or from deficient resistance of the edge concrete. The project's civil engineer is then consulted to validate a compensating solution, such as an increase in anchor diameter, the use of an injection resin for cracked concrete, or a greater embedment depth.
The test report is the key document of the review and safety file handed to the client at acceptance of the work under the SIA 118 standard. The hour invested on the construction site to calibrate and test the representative anchors relieves the metalworker of any presumption of defective execution in the event of a later dispute. This document formally certifies that the balustrades installed meet the usage and safety requirements for the users of the building.
Handling non-conformities: what to do in case of failure or slip
When an anchor gives way under the proof load or shows an axial displacement greater than the criterion set by its European Technical Assessment (ETA), the whole batch represented by the sample is deemed non-compliant. In building metalwork, the slip or pull-out of a balustrade shoe or base plate anchorage compromises the safety of users against falls from height. It most often reveals an execution fault, a poorly cleaned hole or a local weakness in the host concrete. The installation team must immediately suspend assembly in that area and record the failure in the site logbook.
- Systematically double the check sampleIn order to isolate a local heterogeneity of the substrate rather than a systematic installation defect, immediately test two additional anchors for each fixing that failed. If a single additional anchor does not hold the proof load, the entire series installed in the zone must be rejected.
- Diagnose the root cause of the defectExamine the condition of the extracted anchor: hole not brushed, residual dust reducing the chemical bond, nominal tightening torque exceeded with the torque wrench, or premature cracking of the cover concrete along the slab edge.
- Agree the corrective method with the site managementSubmit the analysis of the defects to the site management. Solutions generally consist of increasing the embedment depth, replacing mechanical anchors with a high-performance chemical anchoring system, or changing the spacing of the fixings on a wider base plate.
After complete correction of the fixing process, a new series of non-destructive pull-out tests is carried out on the new batch. The final report, countersigned by the installer and the site management, details the proof loads reached, the holding times and the slip values measured. This official report must be attached to the as-built file in accordance with the requirements of the SIA 118 standard, guaranteeing the legal and technical conformity of the metalwork delivered.
Frequently asked questions
At what point on the construction site should the pull-out test on the anchors be scheduled?
The test takes place immediately after the first base plates are installed and before the balustrades or handrails are fully assembled. For chemically grouted anchors, the full curing time stated on the data sheet, according to the ambient temperature of the concrete, must be respected. Carrying out this test beforehand avoids committing the fixing of all the light elements to a substrate that would show structural or drilling weaknesses.
Is there a risk that the non-destructive pull-out test will damage the anchor tested?
No, if the proof load is correctly calculated. It remains between the permissible service load and the elastic limit of the steel or the slip limit of the concrete defined by the European Technical Assessment (ETA). The tensile device applies a controlled axial force with a bridge support wide enough not to compress the failure cone of the concrete. A compliant anchor that has undergone the test remains perfectly usable for fixing the balustrade.
How many anchors need to be tested along a run of balustrade?
In routine metalwork, the usual sampling covers 3 to 5% of the anchors installed in the batch, with a strict minimum of three points per substrate or homogeneous concreting zone. If the construction site includes different types of slab or concretes of different ages, each distinct configuration requires its own test sample. Testing at least three fixings guarantees a minimum statistical representativeness to validate the adequacy of the installation with the check report.
What specific information must appear on the pull-out test report?
The report records the identification of the construction site, the exact reference of the anchor and its ETA number, the date of installation and the weather conditions. It states the drilling diameter, the embedment depth, the proof load applied in kilonewtons, the duration of the holding step, the displacement measured on the dial gauge and the conformity opinion. The document must be signed by the technician who operated the pull-out tester and endorsed by the site management.
Texts and standards cited
- SIA 118General conditions for the execution of construction work
