Floor Levelling Sydney: When Does a Slab Need a Pull-Off Test Before the Pour?

A pull-off test may be warranted before floor levelling when a Sydney concrete substrate appears weak, contains an old screed or topping, has uncertain previous repairs, or could fail at an interface beneath the new levelling compound.

Rather than relying only on appearance, testing measures direct tensile performance at selected locations. It can create a practical hold point before a large pour, particularly where failure would cause costly removal, programme disruption or downstream flooring defects.

The most dangerous concrete floor is not always the visibly damaged one.

A slab can be vacuumed, ground and apparently ready for primer while the upper surface remains too weak to support the materials about to be bonded to it. Alternatively, the concrete itself may be strong while an older screed, skim coat or repair layer sitting above it becomes the weak link.

That distinction matters on Sydney renovation and commercial fit-out projects because self-levelling cement is only one layer in a larger assembly. The completed system may include concrete, a topping or repair layer, primer, levelling compound, flooring adhesive and a rigid or resilient finish. A failure at one interface can require several completed stages to be removed.

A direct tensile pull-off test gives the project team something visual assessment cannot: a measured indication of how strongly the tested substrate or system resists a force pulling perpendicular to its surface.

The Question Is Not Whether Every Sydney Slab Needs Testing

Routine pull-off testing of every residential slab would be difficult to justify. Many substrates can be assessed through the preparation requirements of the selected flooring system, site inspection, mechanical preparation, moisture assessment and product-specific installation requirements.

The more useful question is whether uncertainty about substrate strength has become important enough that pouring without testing transfers too much risk into the next stage of the project.

ARDEX Australia's substrate-preparation guidance specifically discusses tensile testing where substrate properties need verification and recommends concrete exceeding 1.5 MPa tensile strength before the application of a smoothing cement.

Its published system recommendations for certain levelling applications also specify structurally sound concrete with at least 20 MPa compressive strength and 1.5 MPa tensile strength.

Those figures should not be treated as a universal specification for every product or every floor. The nominated levelling, adhesive and floor-covering manufacturers' current requirements remain controlling for the particular system being installed.

What A Pull-Off Test Actually Investigates

A pull-off test is fundamentally different from checking whether the floor is flat. A laser, straightedge or level survey describes geometry. A pull-off test investigates tensile integrity.

In simplified terms, a test disc is bonded to the prepared test area and equipment applies an increasing perpendicular force until failure occurs. The measured load can be converted into a tensile stress result, typically expressed in megapascals.

But the number is only part of the investigation.

Project teams also need to record where the failure occurred.

A test can expose fundamentally different conditions:

Failure within weak concrete near the surface.
Failure within an old screed or topping.
Separation between a topping and the structural slab.
Failure through a repair or skim-coat layer.
Failure at an adhesive or prepared interface during a system test.
Cohesive failure through a sound material at an acceptable test value.

Two areas can therefore record similar numbers while telling the contractor different stories about what is actually happening beneath the proposed floor.

Five Conditions That Can Justify A Testing Hold Point

1. Grinding exposes a friable or powdering surface

Mechanical preparation sometimes reveals concrete that continues to dust, abrade or break down after the obvious contamination has been removed.

More grinding is not automatically the answer. If deterioration extends into the substrate, the project needs to understand what material remains before a leveller is used to bridge over it.

This differs from the issue explored in Elyment's article on concrete grinding after floor removal and mechanically closed concrete.

A closed surface may be too smooth or improperly profiled. A friable surface presents the opposite problem: the material itself may lack sufficient integrity.

2. An old screed or topping is being retained

Sydney apartments and older homes frequently contain floor build-ups accumulated through several renovations. Demolition may expose a sand-cement screed, topping, feathered repair or unidentified levelling layer that the programme assumes can stay.

The critical question is not whether the layer survived beneath the previous floor. It is whether it can accept the stresses generated by the new bonded assembly.

Elyment has separately examined when an old screed should be removed before floor levelling.

Pull-off testing takes the issue one step further where the decision to retain or remove the layer remains genuinely uncertain after inspection.

3. The slab contains extensive historic repairs

A floor can contain several generations of patching without that history appearing on the drawings. After demolition, the contractor may discover different-coloured repair mortars, old levelling compound, feather finishes, filled penetrations and perimeter reinstatement.

That creates a sampling problem. Testing one attractive patch of original concrete may say little about the weakest areas across the room.

For projects where old levelling compound has been discovered beneath an existing floor, selective testing can help turn an uncertain visual assessment into a more defensible retain, remove or investigate decision.

4. A high-value bonded finish will magnify substrate failure

Risk changes when the finished system is expensive to remove.

Consider engineered timber glued over a levelling layer across a large apartment. If the weakest interface is the concrete immediately beneath the leveller, the floor may not simply require a local patch if debonding develops later.

Boards, adhesive and levelling compound can all become part of the rectification scope.

Elyment's analysis of self-levelling cement beneath glue-down timber explains why the weakest interface in the complete assembly matters.

The economic threshold for additional verification may therefore be lower on a premium glue-down timber, resin or specialist commercial installation than on a small, easily accessible area with a less demanding finish.

5. Responsibility for substrate acceptance needs to be documented

Some testing decisions are driven as much by project delivery as by materials.

On a multi-party fit-out, the demolition contractor may expose the slab, another contractor may grind it, a flooring subcontractor may supply the levelling system, and a separate installer may lay the final flooring.

Without an agreed acceptance point, substrate risk can move silently from one contractor to the next.

A test regime requested by the relevant consultant, manufacturer or flooring specification can create a formal hold point:

Demolition exposes the substrate.
Contamination and weak material are removed.
The floor is divided into representative substrate zones.
Nominated test locations are prepared.
Testing is completed and failure modes recorded.
Results are reviewed against the specified flooring system requirements.
Remediation is completed where required.
The levelling pour is released only after the substrate is accepted.

A Test Result Without A Failure Location Can Be Misleading

The test report should not become a single number forwarded through an email chain.

Suppose a test reaches a project-specified value but the failure occurs inside an old topping. That result means something different from failure occurring deep within sound structural concrete.

Equally, a low result from one isolated damaged corner should not automatically be presented as representative of an entire commercial floor.

Good investigation records should connect the reading with location and substrate condition.

Test location

Allows the result to be mapped against substrate zones and repairs.

Surface condition

Shows whether the area was original concrete, screed, repair mortar or another layer.

Preparation method

Provides context for the state of the surface when tested.

Measured tensile result

Allows comparison with the nominated system requirement.

Failure mode

Identifies which material or interface became the weak point.

Photograph and floor-plan reference

Makes the evidence usable by consultants, contractors and project managers.

Resulting action

Records whether the area was accepted, retested, repaired or removed.

Representative Testing Matters More Than Testing The Easiest Spot

Large Sydney floors are rarely uniform after demolition.

One tenancy can include original structural concrete, old kitchen screed, infilled service penetrations, construction joints, previous wall lines and patches left by earlier fit-outs. Testing only the most convenient open area risks answering the wrong question.

Where testing is considered necessary, test locations should be selected to reflect meaningful changes in substrate condition and the requirements of the responsible consultant, system manufacturer or testing professional.

Particular attention may be appropriate around:

Visibly weak or powdering zones.
Boundaries between old and new concrete.
Retained screeds and toppings.
Large historical patches.
Areas affected by previous adhesive or coatings.
Locations exposed to water or prolonged moisture.
High-traffic commercial zones.
Areas where failure would be difficult to rectify after joinery or flooring installation.

Pull-Off Strength Does Not Replace Moisture Or Surface Preparation

Passing a tensile test does not prove that a slab is ready for every flooring system.

A floor can have adequate tensile integrity and still be unsuitable because of excess moisture, curing compounds, oil, adhesive residue, incompatible coatings, insufficient surface profile, movement, cracking or contamination.

Likewise, primer should not be used as a repair strategy for a weak substrate. ARDEX's published substrate-preparation guidance specifically states that priming is not intended to compensate for poor surface preparation or a weak base.

The practical release decision therefore considers several different questions:

Is the substrate mechanically sound?
Is its tensile performance appropriate for the nominated system?
Has contamination been removed?
Is the required surface profile present?
Are moisture conditions within the applicable system limits?
Have cracks, joints and unstable repairs been dealt with correctly?
Is the primer compatible with both the substrate and levelling compound?

The Sydney Programme Problem: Testing Has To Happen Before The Pour Crew Arrives

The technical test is usually easier to manage than the programme consequences of ordering it too late.

Consider a commercial refurbishment where flooring removal finishes on Monday, grinding is allocated Tuesday and forty or fifty bags of levelling compound are scheduled to be mixed on Wednesday morning.

If questionable substrate strength is first raised late Tuesday afternoon, the project may already have labour, material delivery, access bookings and downstream installers committed.

A more controlled programme identifies possible testing before the critical path is locked:

Pre-start: Identify old toppings, uncertain repairs and high-risk substrate history.
After removal: Expose enough of the floor to verify what is actually present.
After initial preparation: Determine whether weak material is local or widespread.
Hold point: Conduct specified testing before primer and levelling conceal the substrate.
Decision: Accept, remediate, remove or seek technical advice.
Release: Confirm the substrate is ready before the pour crew and materials are mobilised.

Elyment's earlier examination of the remove, grind and level project sequence shows why substrate verification belongs between exposure and application rather than after the new floor has already been installed.

Grinding For Testing Still Brings NSW Silica Obligations

Preparing concrete for testing or further levelling can involve drilling, cutting or grinding. These activities should not be treated as minor simply because the test area is small.

SafeWork NSW's current code of practice identifies mechanical cutting, drilling, grinding, sanding and abrasive polishing of crystalline silica substances as processes capable of generating respirable crystalline silica.

Appropriate work planning and controls therefore remain relevant during substrate preparation and testing activities.

See SafeWork NSW guidance for cutting, drilling and grinding concrete and masonry products for current requirements and guidance.

Testing Can Also Clarify Variations Before They Become Disputes

Unexpected weak substrate can materially change a renovation scope.

A quotation may have allowed for ordinary removal, grinding, priming and levelling. It may not have allowed for removing a failed 30 mm topping across an apartment, repairing deteriorated concrete or delaying the flooring installer while further investigation is completed.

For NSW residential building work, NSW Fair Trading emphasises clear written contracts, plans and specifications and advises that contract variations be recorded in writing. It also recommends keeping photographs and records as work progresses.

That makes substrate evidence operationally useful. A documented investigation can demonstrate why the original floor-preparation assumption changed and why additional work became necessary.

Owners and project managers can review NSW Fair Trading's building and renovation contract guidance when managing residential building scopes and variations.

The Cost Decision Is Really A Consequence Decision

Pull-off testing adds time, coordination and testing cost. The relevant commercial comparison is therefore not simply the price of a test against the price of doing nothing.

It is the testing cost against the consequence of discovering a weak interface after the project has moved several stages further.

Before the pour

Testing access remains relatively straightforward.