Roof Leak Detection in Troy, MI

The sun heats the roof all day. After dark, wet insulation holds that heat longer than dry insulation, and a thermal camera shows the warmer patches. The Snell Group, a thermography trainer, sets the conditions: a clear, calm night, an hour or two after sunset, on a roof whose surface, gravel included, is dry. A thermographer and a helper can cover 100,000 to 200,000 square feet in a night, which makes infrared the usual first pass on a big warehouse or distribution roof.

It has blind spots. Ballasted single-ply roofs and newer roofs with low-absorbency insulation are hard to read, and the camera only reads surface temperature; it never sees the water itself. ASTM C1153, the infrared standard, requires every suspect area to be verified, and Professional Roofing notes that verification has to include cores. The suspect areas get painted outlines and numbers on the membrane; make sure the same numbers appear on the report's roof plan.

ELD finds the hole itself. A charge is put across the membrane, and where water passes through a breach to a grounded conductive layer underneath, the circuit closes and the operator can pinpoint the spot. ASTM D7877 lists four methods:

• Low-voltage scanning platform. A small platform with hanging chains is pushed over a wetted membrane.
• Low-voltage vertical roller. The same idea for flashings and walls.
• Electric field vector mapping (EFVM). The roof is dampened, a bare wire loop goes around the test area, and the operator follows the current to the breach with two probes.
• High-voltage spark testing. A charged metal broom at 1,000 to 30,000 volts DC passes over a completely dry membrane and sparks at a hole. Carlisle notes more false positives and more injury risk with this one.

ELD has hard limits, and they decide whether it can be used on your roof at all. Every method needs an exposed membrane and a conductive layer directly beneath it, such as a steel deck. EFVM and high-voltage testing don't work on black EPDM, which conducts because of its carbon black, or on modified bitumen with an aluminized coating. ASTM D8231, the low-voltage scanning practice, allows forensic use on existing roofs with stated limitations and says it doesn't replace visual inspection.

Both methods lay the roof out on a grid, typically with points 5 to 10 feet apart, and take a reading at each point. A nuclear gauge follows ANSI/SPRI/IIBEC NT-1; an impedance scanner follows ASTM D7954. The readings are relative. Both standards say they have to be tied to weighed core samples before anyone can say how wet the insulation is.

Be wary of a "moisture survey" done with a pin meter alone. There is no standard for pin meters on roofing; the one people cite, ASTM D7438, was written for the wood industry. Seasonal vapor drive also leaves a thin film of condensation near the membrane in winter and near the deck in summer, so a pin can read wet when the insulation facer isn't. Professional Roofing warns that false positives like these lead to roofing being torn off for no reason.

The core is the proof. The plug is weighed wet, dried and weighed again, and moisture content is (wet weight minus dry weight) divided by dry weight, times 100. On nuclear surveys, cores are usually taken at low, middle and high readings, but not at the very highest one. Every opening has to be patched properly afterward.

Flood testing gets suggested more often than it should. ASTM D5957, the flood-test guide, is written for plaza and parking-deck waterproofing and says it isn't meant for building roofs. Carlisle adds that if a serious leak exists, a flood test can put a lot of water into the roof assembly. On an occupied building, that's a hard trade.