Start with the repair decision
Membrane adhesion pull-off testing can help decide whether an existing waterproofing layer is a suitable base for a bonded repair or overlay. Its value depends on asking a precise question: are you investigating the attachment of the old membrane, or checking the bond of a proposed new system? Those questions require different test arrangements and can lead to different repair decisions.
Consider a roof membrane where it turns up against a parapet wall. Before authorising another coating over a lifting edge, establish which layer is separating and whether the proposed repair can bond to what remains. Commission testing when its result could change the scope: retaining sound material, removing defective layers, changing preparation or reconsidering the overlay. A number without an agreed interpretation offers little help.
Define the layer and detail being tested
In a pull-off test, a small loading fixture called a dolly is bonded to the test surface and pulled perpendicular to it. The measured stress describes what that particular assembly resisted under the test conditions. Elcometer explains that preparation, alignment and complete curing of the dolly adhesive matter to the procedure. See its pull-off testing guidance.
Testing the existing membrane addresses the layers already present. Testing a cured trial application introduces the proposed preparation, primer and overlay into the assembly. Ask the investigator to identify the intended interface on a simple layer sketch before testing begins.
At a parapet, distinguish the accessible flat upstand face from the corner, reinforcement and upper termination. A result from the face should not be presented as proof that the termination or corner detail works. If the actual question concerns a sheet seam or mechanical attachment, ask whether a different investigation is needed.
Agree the method and acceptance basis first
Request a written test proposal identifying the membrane, supporting material, test method, locations and acceptance basis. It should explain the dolly size, whether the membrane will be cut around it, and which layers any cut will isolate. These choices belong to the selected procedure and system guidance; they should not be improvised after seeing the readings.
Avoid accepting a generic minimum value copied from an unrelated product sheet. A requirement for bare concrete is not automatically an acceptance limit for an aged flexible membrane or a new overlay. Ask who supplied the criterion, which assembly it covers and whether the observed failure mode is acceptable.
For procurement, a useful deliverable is a clear recommendation linked to the results: suitable for the specified trialled repair, unsuitable, or inconclusive pending further investigation. Agree how inconsistent readings and failed test-glue bonds will be handled. This is practical commissioning guidance, not a statement of a Cyprus statutory testing requirement.
Make the trial reflect the intended work
A repair trial should reproduce the proposed work closely enough to inform it. Record the cleaning and preparation method, primer identity, application sequence and curing period. An unusually intensive preparation on one small trial patch is weak evidence for a quotation that allows only a quick surface clean.
Sika's single-ply refurbishment guidance specifically stresses representative samples, realistic preparation, appropriate primers and correctly interpreted pull-off results. It also identifies material compatibility as a separate consideration. That supports trialling the actual proposed system rather than assuming that any liquid membrane will suit an unidentified existing surface.
For a Cyprus property, ask the test plan to record exposure, shade, surface condition and relevant weather conditions. Confirm the applicable preparation and curing limits with the selected system supplier. Choose locations that address observed differences, such as an old patch beside apparently intact membrane. Do not let convenient access become the only reason for selecting a test point.
Read the fracture alongside the number
The report should show where each specimen separated. An adhesive failure occurs at an interface between materials; a cohesive failure occurs within a material. Failure in the dolly glue concerns the test connection itself. Elcometer's guidance for coatings on concrete explains these distinctions and the need to record both the pull-off result and failure mode.
Their practical meanings differ. Separation between the proposed overlay and old membrane directs attention to that interface. Fracture within an older layer means that layer limited the tested assembly. Neither observation, by itself, establishes the original cause of deterioration. Photographs of both fracture faces help make the interpretation reviewable.
Ask for individual readings rather than only an average, with locations and descriptions of mixed fractures. An average can conceal a locally poor result at the very detail being repaired. Where glue separation prevents the intended interface from being assessed, the investigator should explain whether the selected procedure permits a limited conclusion or requires another test.
A hypothetical parapet repair decision
Imagine a building manager considering a liquid overlay on an existing roof upstand. There is a lifting patch near the parapet corner, while the adjacent face looks intact. This is a hypothetical example, not a Waterproofed.cy project. The investigator proposes comparison locations and a trial using the repair supplier's specified preparation and primer.
Suppose the trial beside the intact area satisfies the agreed criterion, but testing near the lifting patch separates within an older underlying coating. The reasonable next step is to investigate the extent of that weak layer and define removal boundaries. Simply applying more primer would not answer why the retained assembly broke beneath the new bond.
After preparation is revised, further trials may support a local repair. If similar weakness appears across additional representative locations, broader removal may need consideration. If the specimens instead separate in the test glue, the manager should not classify the roof as having failed adhesion without the investigator resolving that limitation. Each outcome changes the next action; none automatically justifies replacing the whole roof.
Keep adhesion separate from whole-roof performance
A pull-off result is local evidence about the tested assembly. It does not, by its measurement alone, establish watertightness, locate a leak or approve every junction. Keep drainage, visible openings, terminations and the condition of concealed layers within the wider investigation rather than treating a satisfactory adhesion result as closure.
Attachment also matters. Sika warns that an overlay on a mechanically fixed single-ply roof can remain dependent on the original fixings. Good surface adhesion therefore does not settle the condition of those fixings or the suitability of the complete refurbishment.
For an owner, the useful distinction is between evidence supporting a particular bond and evidence supporting the entire repair proposal. Ask the written recommendation to state which conclusions were tested, which came from observation and which remain assumptions.
Include reinstatement and a usable report
Treat test locations as part of the repair scope from the outset. Agree how disturbed membrane and any exposed supporting material will be reinstated, which compatible products will be used and who will document completion. Avoid commissioning destructive investigation without a workable plan for closing the openings.
The final report should connect the location map, individual results, fracture photographs and trial preparation to a defined repair recommendation. It should also identify untested areas and unresolved compatibility questions. That gives the building manager something meaningful to compare against the proposed works.
For help assessing the detail within the wider problem, explore Waterproofed.cy's waterproofing services, then request an assessment of the waterproofing detail. Describe the membrane, lifting or blistering, previous repairs and proposed overlay. Confirm whether specialist pull-off testing, interpretation and reinstatement are included in the agreed scope.
Frequently asked questions
What pull-off value should a waterproofing membrane achieve?
Use the criterion agreed for the particular membrane assembly, test method and proposed repair system. A substrate-strength requirement from an unrelated product sheet should not be treated as a universal membrane acceptance limit.
Does a successful pull-off test mean the roof is waterproof?
No. It provides local evidence about the tested assembly's resistance to perpendicular loading. Leaks, seams, terminations, drainage and other details require their own assessment.
Can the existing membrane be tested before choosing an overlay?
Testing can investigate the existing assembly, but it cannot establish the bond of an overlay that has not been trialled. A proposed overlay may need a separate cured trial using the intended preparation and primer.
What if the dolly glue fails?
The investigator must interpret it under the selected procedure. It may prevent assessment of the intended interface and require another test; it should not automatically be reported as membrane bond failure.
Is pull-off testing included in a Waterproofed.cy assessment?
The assessment page describes a site visit and written assessment, but does not explicitly list specialist pull-off testing. Confirm testing, reporting and reinstatement arrangements before commissioning them.
Sources
Sources checked 2026-10-15.