Technical7 min read

How Many Roof Anchors Should You Test on a 40-Point System?

AT
Anchor Testing Australia

Strata managers and asset owners regularly ask a version of the same question: if a building has 40 roof anchor points, do all 40 need to be proof-load tested, or is a sample sufficient? The answer is not a fixed percentage. It depends on the anchor type, the substrate, the installation record, the applicable specification, and what protocol governs the scope when a sample point fails.

Getting this wrong in either direction carries real consequences. Under-testing leaves unverified anchors in service. Over-specifying 100 percent testing on a well-documented, uniform system can add cost without proportionate risk reduction. The right scope comes from understanding the frameworks that actually govern this decision.

Why There Is No Universal Sampling Percentage

Australian practice draws on several overlapping frameworks: AS/NZS 1891.4 for industrial fall-arrest systems, AS 5216:2021 for post-installed and cast-in fasteners, AEFAC Technical Note TN05, and for civil and infrastructure work, VicRoads Section 680. None of these documents prescribe a single percentage that applies across all building types, anchor configurations, and substrates.

The reason is straightforward. A 40-point system on a 1980s precast concrete roof with no installation records is a fundamentally different risk profile from a 40-point system installed two years ago by a documented contractor into a known-strength concrete slab with torque records on file. Treating both with the same sampling fraction would either under-test the first or waste resources on the second.

What the frameworks do provide is a logic for determining scope, and that logic is worth understanding before you engage a testing contractor.

The AEFAC TN05 Approach

AEFAC Technical Note TN05 addresses anchor testing for height-safety applications and is the most directly relevant guidance for roof anchor programmes. The TN05 volumes distinguish between situations where installation quality is known and situations where it is not.

Where installation records exist, the anchors were installed by a competent party, the substrate is well characterised, and the system is uniform in type and embedment, a sampling approach is defensible. The sample is not arbitrary; it should be statistically representative of the population and should include points that are more likely to be marginal, such as those near edges, in repaired sections of the substrate, or in areas with visible surface deterioration.

Where installation records are absent or the substrate is variable, the presumption shifts toward a higher proportion of testing, and in some cases toward 100 percent. The logic is that sampling only produces useful information when the sample can be taken as representative. If you cannot characterise the population, you cannot draw valid inferences from a sample.

For a 40-point system with no installation history, a testing contractor who proposes sampling 10 points and certifying the rest by inference is making an assumption the data does not support.

VicRoads Section 680 and Infrastructure Practice

VicRoads Section 680 is a specification that engineers reference for anchor testing in civil and infrastructure contexts, particularly for post-installed anchors in concrete. It is not a roof-anchor document specifically, but its sampling logic is instructive because it is one of the more detailed Australian treatments of the question.

Section 680 sets out acceptance criteria and specifies that where anchors are installed in lots, a defined number from each lot must be tested. A lot is a group of anchors installed under the same conditions: same installer, same substrate batch, same embedment depth, same anchor type. The lot concept is important because it defines the boundary of what a passing sample actually certifies.

Applying this thinking to a 40-point roof system: if all 40 anchors are the same type, installed by the same contractor in the same substrate zone at the same time, they may reasonably constitute a single lot and a proportional sample applies. If the system was installed in stages, by different contractors, into different substrate zones, or with different anchor types at different points, the lot structure fragments and the sample size for each sub-group increases relative to its population.

This is one reason a desktop review of installation records before mobilising to site is worth doing. It shapes the testing scope before anyone climbs onto the roof.

What Happens When a Sample Fails

This is the part of sampling logic that is most often misunderstood. A sampling protocol is not just a way to reduce the number of tests. It is a decision rule that specifies what happens when the sample does not pass.

In most frameworks, a failure in a sampled anchor does not mean the job is done and the anchor is tagged out. It triggers an expansion of the test set. The expansion logic varies by specification, but the general principle is that a failure in a sample raises doubt about the entire lot or sub-group from which that sample was drawn. The response is to test more anchors from that group, not to stop.

For a 40-point system, suppose the programme calls for testing 10 anchors as an initial sample and one fails. The appropriate response under most frameworks is to test additional anchors from the same lot, often doubling the sample, and to assess whether the failure is isolated or indicative of a systemic problem. If further failures appear, the scope may expand to 100 percent of that lot.

This expansion logic matters for budgeting and programme planning. Asset owners who plan for the minimum sample cost without allowing for expansion testing can find themselves mid-programme with a failed anchor and no contingency for the additional scope. Building the expansion protocol into the contract scope upfront avoids that situation.

When 100 Percent Testing Is the Right Starting Point

There are circumstances where sampling is not appropriate and 100 percent testing is the correct starting scope.

Anchors with no installation records and no way to establish a lot structure fall into this category. So do anchors in substrates with known variability, such as older lightweight concrete, aerated autoclaved concrete, or masonry where block and mortar strength can vary significantly across a facade. Anchors that have been in service for an extended period without any prior testing, particularly in coastal or industrial environments where corrosion is a factor, also warrant a higher testing fraction.

Systems where a single anchor failure would have serious consequences for a connected horizontal lifeline or static line also warrant closer scrutiny. A single-point anchor that one person clips onto is a different risk profile from an anchor that forms part of a system where multiple workers may be connected simultaneously.

For some asset classes, the certifying engineer or the RPEQ reviewing the programme will specify 100 percent testing as a condition of certification. That is not conservatism for its own sake; it reflects a considered view that the available information does not support a valid inference from a sample.

Applying This to a Real 40-Point System

To make this concrete without inventing a universal rule: a 40-point system might be scoped as follows, depending on what the pre-test review reveals.

  • Well-documented, uniform installation, known substrate:: A sample of 10 to 15 anchors, selected to include edge cases and any points with visible concern, may be defensible. Expansion protocol applies if any fail.
  • Partial records, mixed installation dates, or substrate variability:: The system should be divided into sub-groups by lot characteristics, and each sub-group sampled proportionally. Total tested anchors may be 20 or more.
  • No records, unknown substrate, or visible deterioration:: 100 percent testing is the appropriate starting position. The certifying engineer should confirm this in writing before the programme is scoped.

These are not rules; they are illustrations of how the frameworks translate into practice. The actual scope for any specific building should be determined by a competent person with access to the installation records, a substrate assessment, and an understanding of the applicable standards.

The Certification Scope Follows the Testing Scope

One practical point that asset owners sometimes miss: an anchor-point certification report can only certify what was tested. If 15 of 40 anchors were proof-load tested and 25 were not, the certification covers 15 points. The remaining 25 are not certified by inference unless the certifying engineer has explicitly assessed the lot structure and documented the basis for extending the finding.

This matters when the certification is used to satisfy a PCBU's duty to provide a safe system of work, or when a principal contractor needs to demonstrate compliance to a client or regulator. A certification that covers a sample without a documented basis for extending findings to the untested population is not a certification of the system; it is a certification of the tested points only.

Before engaging a testing contractor, it is worth asking how the certification report will handle the untested anchors and what documentation will support any extension of findings. If the answer is vague, the scope needs to be clarified.

Getting the Scope Right Before Mobilising

The testing scope for a 40-point system is not a number to be picked from a table. It follows from a structured assessment of installation history, substrate conditions, anchor type uniformity, and the applicable frameworks. AEFAC TN05 and VicRoads Section 680 provide the logic; the certifying engineer applies it to the specific building.

Anchor Testing Australia works with strata managers, asset owners, and principal contractors across Australia to determine the appropriate testing scope before mobilising, not after. If you have a multi-point system and need clarity on what a defensible programme looks like, contact us at [email protected] or call +61 7 3132 2534, or visit anchortesting.au to learn more about our anchor-point certification services.

Need anchor testing for your project?

Send us your drawings, anchor schedules, and substrate details. We'll respond with the right test pathway and a scope within 24 hours.

Send Test Brief