Why Most Anchor Points in SA Fail the Design Test | WAHS

There is a dangerous assumption embedded in how fall protection anchor points are installed across South Africa: that if an anchor holds under a pull-out test, it is fit for purpose. This assumption is wrong — and in a fall arrest event, it can be fatal.

Pull-out testing verifies one thing only: that the chemical bond between the stainless-steel anchor rod and the concrete substrate can resist a prescribed axial tensile load for a defined period. What it does not test — and what most installations never consider — are bending forces, shear forces, substrate suitability, free-length calculations, and anchor point geometry. These are not optional engineering niceties. They are the variables that determine whether an anchor point actually holds a falling person.

At Work at Height Solutions, we see this gap constantly. Buildings are handed over with anchor points installed by contractors who understand the drill and the resin but have never performed a bending moment calculation or assessed the effect of a cantilever distance on anchor rod stress. The building owner receives a “compliance” file, a certificate, and a false sense of security.

This post unpacks the engineering behind a properly designed anchor point — and explains why design, not just installation, is the foundation of real compliance.

What SANS 50795 Actually Requires

SANS 50795 is the South African standard for anchor devices used in fall protection and rope access applications. It specifies performance requirements and test methods for anchor devices across several classes — from Class A1 single-point anchors bonded into concrete, through to Class D rigid horizontal rails.

A critical point that is routinely overlooked: SANS 50795 requires that manufacturers supply installation instructions and that installers confirm the suitability of the base material into which anchors are fixed. The standard does not treat installation as a simple drilling exercise. It treats it as an engineering activity that must account for the actual loads the anchor will experience during a fall arrest event.

Class A1 anchors — the most commonly installed type in South Africa — are subjected to a type-test axial load of 10 kN for three minutes during manufacturer qualification testing. The soundness-of-fixing pull-out test applied on-site is 6 kN axially for 15 seconds. Passing this site test confirms the chemical bond is sound. It says nothing about bending, shear, or the adequacy of the anchor rod under combined loading.

What Happens in a Real Fall Arrest Event

To understand why design matters, you need to understand the forces that act on an anchor point when a person falls and is arrested.

The starting point is momentum. An 80 kg person — the internationally accepted standard mass used in EN 1808 and EN 892 — falling freely for just two metres reaches a velocity of approximately 6,3 metres per second. Their momentum at that point is over 500 kg·m/s. An energy absorber in the lanyard system extends the stopping time to roughly 0,08 seconds, which limits the arrest force to the international maximum of 6,0 kN. This 6 kN figure is not arbitrary — it is the threshold beyond which the force on the human body becomes injurious.

That 6,0 kN does not act purely in one direction. Depending on where the anchor is positioned relative to the worker’s fall trajectory, the force resolves into a combination of axial tension and shear. On a parapet wall, for example, where the rope runs at an angle from the anchor point down to a worker on the façade, both a horizontal shear force and a vertical tensile force act simultaneously on the anchor rod. These forces must be calculated using the geometry of the actual installation — not assumed to be manageable.

The combined loading formula from the design guidelines is clear: where an anchor must simultaneously resist both design shear and design tensile forces, the combined utilisation ratio must not exceed 1,0. Exceeding this limit means the anchor is overloaded, even if each individual force component appears acceptable in isolation.

The Bending Moment Problem

This is the most commonly ignored failure mode in South African anchor installations — and arguably the most dangerous.

In most post-fixed anchor installations — that is, anchors installed after construction is complete — the anchor rod penetrates layers of waterproofing membrane and plaster render before reaching the structural concrete. These surface layers are soft and non-load-bearing. They do not support the anchor rod laterally. When a force is applied to the anchor point, the unsupported length of rod above the structural concrete acts as a cantilever, and a bending moment develops in the anchor rod.

The free length “L” — the cantilever distance that governs the bending moment — is not simply the thickness of the plaster. It includes the geometry of the anchor point device (the hook-on unit), the thickness of all non-load-bearing surface coverings, and a correction factor equal to half the nominal diameter of the anchor rod, which accounts for concrete damage at the mouth of the drilled hole during installation. A poorly prepared hole mouth can increase this correction factor substantially.

The consequence of ignoring free length is significant. Consider a hook-on device that protrudes 8 mm from its base plate, installed through 5 mm of waterproofing on an M16 stainless steel anchor rod. The free length calculates to 21 mm, and the resulting bending moment under a 6,0 kN applied force is 126 Nm. An M16 A4-grade stainless anchor has a maximum allowable bending moment of 104 Nm. The installation fails the bending stress check even though it will pass a pull-out test.

The solution in that scenario is either to specify a larger anchor diameter — an M20 anchor has an allowable bending moment of 202 Nm — or, more effectively, to eliminate the soft surface layer between the anchor plate and the structural concrete. Bolting the anchor device directly against the concrete removes the cantilever contribution from the plaster, reduces free length to the device geometry only, and drops the bending moment to within acceptable limits for an M16 anchor. It also prevents water ingress through the cracked render around the anchor — an important durability consideration for permanently installed fall arrest systems.

None of this is calculated in a pull-out test. None of it is visible to the person certifying the anchor after installation unless the design has been done correctly upfront.

Substrate Suitability: Not All Concrete Is Equal

SANS 50795 Class A1 anchors are designed for installation into in-situ cast concrete. The standard assumes cracked concrete class 20/25 as the design substrate — a conservative assumption that accounts for the weakest realistic condition in a structural concrete element.

Several substrate types that appear suitable are not. The following are critical exclusions that must be understood:

Brick walls

South African brick wall construction is fundamentally different from the cavity wall systems used in the United Kingdom, where SANS 50795 pull-out testing on masonry is referenced. UK cavity walls are tied at intervals with steel flat straps, creating a composite structure with measurable load capacity. South African brick walls are generally single-skin or unfilled cavity construction that offers no equivalent structural integrity for point loads. Installing anchors into a South African brick wall without a written engineering certificate confirming suitability is not compliant. A pull-out test on a single brick does not test the wall — it tests the resin bond to that brick alone. The entire wall can fail at the anchor load while the brick-to-resin bond holds perfectly.

Post-tensioned concrete

Modern commercial buildings frequently use post-tensioned concrete slabs and beams to achieve longer unsupported spans. High-tensile steel cables are threaded through plastic ducts cast into the concrete and stressed after curing. Drilling into post-tensioned concrete without first confirming cable positions through the building engineer can sever a prestressing cable, causing catastrophic structural failure. Written confirmation from the building owner that marked anchor positions are clear of post-tensioned cables is a non-negotiable pre-installation requirement.

Steelwork

Anchors fixed to steel beams or columns must be verified by calculation from a registered engineer. The visible size of a steel member is not confirmation of its capacity to carry a fall arrest load, particularly at connection points that may have corroded. Drilling through hollow steel sections introduces oxygen and moisture, initiating internal corrosion that is invisible from the outside.

Rendered surfaces

Render, plaster, and waterproofing membranes are not structural. An anchor installed with its plate bearing against a rendered surface and not against the structural concrete beneath it is relying on a non-load-bearing layer to transfer forces. This is one of the most common and least visible installation deficiencies in the South African market.

Anchor Selection: The Hook-On Unit Is Not Interchangeable

The anchor rod chemically bonded into the concrete and the hook-on unit attached to it are two distinct components with different engineering requirements. Both must be selected and verified for the specific installation.

Eye nuts and eyebolts are frequently used as hook-on devices because they are inexpensive and widely available. Their limitation is critical: they are designed for axial loading only. Any deviation from pure axial load reduces their capacity, and no lateral load should be applied beyond 45 degrees from the anchor axis without written confirmation from the eye nut manufacturer that the capacity remains adequate. On a rooftop installation where the rope runs at an angle across a parapet, a standard eye nut may be entirely unsuitable as the hook-on device.

Purpose-designed anchor plates that bear flat against the concrete surface can accept loads at right angles to the anchor rod axis — the typical loading direction on a vertical façade. However, they must be pressed firmly against the structural concrete, not against a render layer, and the correct torque must be applied to the fixing nut to ensure the plate is clamped tight. Loose plates create additional free length and increase the bending moment in the anchor rod.

Expansion anchors and hammer-set mechanical anchors must not be used for fall arrest or rope access applications. Only chemical anchors — resin-bonded stainless-steel rods — are acceptable for permanently installed fall arrest anchor points.

Load Testing After Installation: What It Confirms and What It Does Not

On-site proof load testing of anchor points is required to confirm the soundness of the chemical bond after installation. This is typically conducted as an axial pull-out test at 6,0 kN for 15 seconds.

This test has value. It identifies installation defects — insufficient resin, incorrectly prepared holes, inadequate embedment depth, green concrete — that would cause the anchor to fail before reaching its design capacity. These are real failure modes that testing catches.

What the pull-out test does not confirm is whether the combined loading capacity, bending stress, substrate structural capacity, and hook-on unit suitability are adequate for the specific installation geometry. An anchor can pass the pull-out test and still be inadequate for the actual fall arrest forces it will experience in use.

A properly issued compliance certificate for an anchor point installation must be based on an engineering design, not solely on the results of a pull-out test. The design must precede the installation and must address all relevant load cases for the specific anchor position, substrate, and hook-on device geometry.

Competence: Who Is Qualified to Design and Install Anchor Points?

The Institute for Work at Height (IWH), the SAQA-recognised professional body for the work at height industry in South Africa, has developed a structured qualification framework for persons working with anchor points. The relevant unit standards are:

SAQA Unit Standard 229998 — Work in a fall risk position. This is the entry-level certification confirming that a person can work safely at height using fall protection systems. It does not qualify someone to install or design anchor points.

SAQA Unit Standard 229994 — Fall Protection Planner. Qualifies a competent person to compile Fall Protection Plans under Construction Regulation 10. Relevant to the site planning of anchor point systems but not to their engineering design.

Anchor point design — particularly the calculation of free lengths, bending moments, combined forces, and substrate suitability — falls within the domain of engineering professionals. For installations into steelwork or masonry, SANS 50795 explicitly requires verification by a qualified engineer. For concrete installations, the design parameters discussed in this post must be addressed by someone with the technical competency to perform the calculations correctly.

This is not a barrier to the industry — it is the minimum standard the industry must meet. Altramed, the accredited training partner within our group, delivers IWH-aligned training across all three unit standards. If your team handles anchor point installations and inspections, proper certification is the starting point — not the finish line.

What a Properly Designed Anchor Point Installation Looks Like

A compliant, properly designed fall protection anchor point installation at minimum includes the following:

A pre-installation engineering design that establishes anchor positions, substrate suitability, anchor rod size and grade, embedment depth, edge distances, hook-on unit selection, free-length calculations, and bending moment verification for each anchor position. For Class A1 concrete installations, this design must assume cracked concrete and must confirm that the combined shear and tensile utilisation ratio does not exceed 1,0 under the design fall arrest load of 6,0 kN.

An installation by a competent person who understands the difference between drilling into a waterproofed parapet surface and bonding into the structural concrete below it — and who specifies hole preparation, resin type, cure time, and correct torque to the anchor plate.

Post-installation proof load testing at the prescribed axial load, conducted in the direction of the anticipated applied force — not just vertically upward, which is easier to rig but may not reflect actual loading geometry.

Documented installation records that include the address and location, installer’s name and company, anchor positions, substrate details, anchor rod specifications, hook-on unit specifications, resin product and batch number, test loads applied, and the date of installation. This documentation must be handed over to the building owner and retained on-site for future inspection.

Annual inspection and re-certification by a competent person, identifying corrosion, mechanical damage, movement in the substrate, or any change in condition that affects the anchor’s integrity.

The Liability Consequence of Getting This Wrong

OHSA Section 8 places a general duty on every employer to provide and maintain a working environment that is safe and without risk to health. Where that environment includes working at height, the employer’s duty extends to ensuring that fall protection systems are engineered, not just purchased and bolted down.

An anchor point that fails during a fall arrest event — because bending stress was never calculated, because the hook-on plate was torqued against waterproofing and not concrete, because nobody confirmed the substrate was not post-tensioned — is not a product failure. It is an engineering and management failure. The liability sits with the employer who commissioned the installation and the persons who signed off the compliance documentation.

A pull-out test certificate is not engineering compliance. It is evidence that one force component, in one direction, held for 15 seconds on a specific day. The building owner who believes that piece of paper represents a safe anchor system is carrying a risk they do not know about.

WAHS Anchor Point Installations: Engineered from the First Drill

Work at Height Solutions designs and installs SANS 50795-compliant anchor point systems with engineering at the centre of every installation. Our RS Anchor Device is fabricated from 6 mm stainless steel and chemically bonded to concrete via stainless-steel rods — engineered for the loads, the substrates, and the geometries of South African buildings.

Every installation includes a pre-installation design, compliant installation by trained personnel, post-installation proof load testing, and documented certification. Annual inspection and re-certification keeps your compliance file current and your liability exposure managed.

If your building carries anchor points that were installed without engineering design documentation, or if you cannot confirm that bending moments and substrate suitability were assessed, contact us. We can inspect, assess, and either certify or remediate your existing installation — and issue you a compliance file you can rely on.

Published by Jaco Coetzer | Work at Height Solutions (Pty) Ltd