Electronic leak detection in depth: low voltage, high voltage and when each works

Low voltage wet testing

Low voltage testing works on a wetted membrane. A conductive wire is laid around the edge of the test area, and around anything earthed inside it, and connected to a pulse generator, with an earth-return lead to a known earth point on the building’s structure. Pulses of current spread over the roof through the water on its surface, and can reach earth only through a breach.

Two hand-held probes then track the size and direction of the current’s flow towards earth, leading the operator to the breach. Detec’s scanning platform does the same with chains that hang from it into the water as it is moved across the membrane, and alerts the operator by sound and light over a breach.

It suits non-conductive membranes such as hot melt, asphalt and single ply, as long as the surface and the structure beneath are well soaked, and the membrane’s thickness makes no difference to it. A wet sponge attachment tests upstands and other vertical surfaces. The whole surface has to stay wet, and Detec advises keeping a hose and a water supply on site, as dew alone might not be enough.

High voltage dry testing

High voltage testing works on a dry membrane. The detector puts a high DC voltage on an electrode, such as a phosphor bronze brush or a metal broom, and an earth lead connects the detector to a known earth on the building’s structure. Carlisle gives the range as 1,000 to 30,000 volts. Swept over the surface, the electrode completes the circuit wherever it passes over a breach: the detector alarms, and a spark shows the spot.

The membrane’s thickness matters here, as the voltage is set from it, to the standard NACE SP0274 in Buckleys’ case. Most non-conductive membranes conduct very slightly, so the detector’s sensitivity is set to ignore that level and alarm only at a breach. The whole surface has to be completely dry, though water under the membrane from a leak can help, by improving the path to earth.

It needs no water and no perimeter wire, so it suits roofs where wet conditions are not present or practical. Carlisle notes more reported false positives, a greater risk of injury from the higher voltage, and some concern about dragging a wire broom across a membrane.

Why conductivity decides everything

Both tests need a conductive layer directly below the membrane to carry the current back. Structural concrete and metal decks provide one. Timber, OSB, cover boards, protection boards, insulation and lightweight concrete do not, and the ASTM guide notes that a membrane laid over insulating foam or a protection board is cut off from any conductive deck below.

The answer is a conductive layer built in directly under the membrane as the roof is laid. Carlisle lists welded stainless steel mesh or a conductive primer for adhered roofs, aluminium screen for loose-laid ones and conductive fabric for loose-laid and mechanically fastened ones, and Detec makes its primer for every high and low voltage method, warning that a conductive layer placed lower, under the insulation or a cover board, gives invalid results.

On an existing roof with an active leak, water already in the build-up carries the current to a conductive deck, and inverted waterproofing laid directly on structural concrete needs nothing added. Neither standard test works where dry boards or insulation separate the membrane from any conductive layer, or on a membrane that conducts itself.

So the deck, membrane, build-up and any coverings are reviewed before a method is chosen, and where records are missing or unreliable, core sampling establishes the build-up and deck.

Membranes that defeat it

Black EPDM contains carbon black, which makes it conduct, and that defeats the standard tests: Carlisle says neither the perimeter wire test nor high voltage testing works on it, or on membranes with the aluminised protective coatings commonly used on modified bitumen.

The specialist method is Detec’s low voltage scanning platform, used with a conductive primer applied directly under the membrane, which Carlisle says works on both. The ASTM practice for scanning also covers certain moderately conductive membranes. It only works where the build-up suits it, as the primer has to go on before the membrane does.

Overburden

A green roof, insulation, a wearing course or a topping slab over the membrane makes leaks much harder to locate, the ASTM guide says, and the ASTM practice for scanning says a membrane that will be covered must be tested for breaches before the overburden goes on.

Results through a covering are less conclusive. Sika lists what can interfere, from lightning protection and steel in topping slabs tied to the building to drains, drainage layers and thick soil, and Detec says valid, repeatable testing can only be done on an exposed membrane. So to test a covered membrane to that standard, the covering over the area has to be lifted, at the owner’s cost, Sika notes.

A roof meant to be tested once it is covered needs that planned in: on some roofs a trace wire is installed during construction so the membrane can be tested later, and Sika says the designer must make sure the overburden design will not interfere with testing.

Integrity testing at handover

Many membrane manufacturers accept or require electronic testing before a warranty is issued. Sika’s roofing division requires it as the main quality assurance method over the completed, exposed membrane on all its waterproofing projects, to help confirm the system is watertight before the overburden is placed and a warranty is issued. Carlisle tells specifiers to require it as field quality control, to the ASTM practice, with the conductive layer installed directly below the membrane.

A test certificate records a result, which covers only the membrane tested, by the method used and in the conditions the report sets out. No test is completely accurate, Sika notes, and a result cannot cover damage done afterwards, a breach the current could not reach, such as one under a covering or where the path to the conductive layer was broken, or how far water has spread inside the roof.

Damage by later trades is a known cause of failure, so the membrane is tested again once plant, solar or services installers have worked on it, and before any covering goes on.

Reading a leak detection report

Each breach is marked on the roof itself and on a plan of the roof, with photographs. The report also sets out the method used and the surface conditions, the estimated size of each defect, and recommendations for repair or further investigation.

The contractor uses the marks and the plan to repair each breach where it was found, so the work is limited to the damaged areas rather than the whole roof, and the repaired areas can then be tested again. Where a leak appears after handover, the report is the evidence of the membrane’s defects.

Questions about electronic leak detection