# Inspect Lap Sealant After the First Long Heat Run Up the I-15 Corridor > A dark membrane in inland sun runs far above air temperature, then drops after dark. Membrane, roof edge and bead each move at a different rate. Source: https://ocrv.info/blog/inspect-lap-sealant-after-a-long-heat-run Published 2026-02-23. Category: Seasonal Maintenance. About 6 minutes. Sustained inland heat drives a dark roof membrane well above air temperature, then it cools sharply overnight. The membrane, the aluminum or fiberglass edge and the sealant bead all expand and contract at different rates, so the bead works itself apart at the flange shoulder. Inspect every penetration after a long hot run. ## The short answer Heat cycling, not water, is what destroys lap sealant. A dark EPDM or TPO membrane sitting in full inland sun runs far hotter than the air around it, then sheds that heat fast after dark. The membrane, the aluminum or fiberglass roof edge under it, and the polymer bead bridging the two all change dimension at different rates on every one of those cycles. The bead is the weakest link in the stack, and it loses at the shoulder where it meets a vertical flange. ## What one heat cycle does to three materials at the same time Consider a plumbing vent boot on a laminated roof. The boot flange is plastic. The screws through it go into a wood backer. The membrane under the flange is EPDM. The bead over the top is a self leveling polyurethane or a similar formulation. That is four materials with four different coefficients of thermal expansion, mechanically joined, sitting under a heat load that swings across a wide range twice a day. Run that stack up to a roof surface temperature that can sit forty to sixty degrees above the air on a still afternoon, then let it fall to near ambient overnight. The plastic flange grows and shrinks more than the membrane. The membrane moves more than the aluminum edge trim it terminates into. The bead has to absorb every one of those differences by stretching and relaxing. Fresh sealant does that easily because it is elastic. Sealant that has lost its plasticizers to a few seasons of heat cannot, so the movement gets transferred into the bond line instead of the body of the bead. That is the whole mechanism. It does not require standing water, poor workmanship or a hailstorm. It only requires time and sun, which a coach that runs the I-15 corridor in the summer gets a great deal of, and which a coach parked on an open pad off Limonite Avenue gets just as much of without moving at all. ## Self leveling on horizontal, non sag on vertical, and what goes wrong when they trade places These two products look similar in the tube and behave nothing alike. Self leveling lap sealant, the Dicor type most people know, is formulated to flow. Applied to a horizontal roof plane it spreads out, wets the substrate, feathers its own edges and self heals small tooling marks. That flow is the entire point, because it means the bead makes intimate contact with both surfaces instead of bridging over a void. Non sag sealant is the same chemistry family with rheology modifiers that stop it from moving after it leaves the nozzle. It belongs on vertical surfaces: front and rear cap seams, corner extrusions, window flanges, anything where gravity would otherwise pull the bead down and thin it at the top. Put self leveling product on a vertical seam and it slumps. It leaves a thin, starved section at the top of the joint and a fat useless ridge at the bottom, and the thin section cracks first. Put non sag product on a horizontal roof seam and it does the opposite problem: it holds its tooled shape and never flows into the joint, so it bridges the gap like a tent instead of filling it. Under it, an air pocket forms. Water gets in through a pinhole and has nowhere to go. ## The failure signs a bead shows before it lets water through None of these require a moisture meter. They require getting close enough to touch the bead. - Chalking. A white powdery surface that comes off on a finger. The surface layer has oxidized and is no longer part of the flexible mass. - A bead that has migrated. It sits fully on the flange or fully on the membrane, with a visible gap where it used to bridge. The joint has been pulling in one direction and the bead followed it. - A hairline running the length of a seam. Straight, thin, and consistent, which is the signature of thermal movement rather than impact. - No tack under a fingernail. Healthy sealant gives slightly and feels faintly grabby. Dead sealant feels like hard plastic and clicks when tapped. - A glossy, rubbery layer over an older matte one. That is silicone applied over cured sealant. Silicone does not bond to most roof sealants or to EPDM, and it is difficult to remove, which makes the next proper repair harder than it needed to be. ## Why layering fresh sealant over a failed bead hides the split The common shortcut is to run a new bead over the top of the old one and call it done. It looks correct for a season, which is the trouble. The split is in the old bead, at the bond line, usually where the old material meets the flange. A new layer applied over the top bonds to the old sealant surface, not to the flange. Underneath, the crack keeps moving with every heat cycle, and it is now sealed off from view. Water still enters at the original discontinuity, travels between the old bead and the substrate, and emerges somewhere else entirely. You have not stopped the leak. You have made it harder to find and given yourself a false clock. Doing it properly means removing the old material back to clean flange and clean membrane, wiping the surface with the solvent the sealant manufacturer specifies, and laying a fresh bead onto bare substrate. That is slower, dirtier, and it is the only version that lasts. The [roof repair](/services/roof-slide-and-awning/roof-repair) process we run starts with removal for exactly this reason. ## Membrane condition is a separate question from sealant condition A perfect reseal on a spent membrane buys very little. These are two independent variables and they need to be judged separately. Rubber roofing wears from the top down. The factory surface layer erodes, and once it is gone the material underneath weathers quickly, going thin, powdery and eventually translucent over the seams. Press a thumb into a suspect area. Healthy membrane resists and returns. Spent membrane feels like wet paper over the substrate, and if there is any softness in the decking below it, the problem is no longer on the surface at all. Walk the roof and feel for depressions between the rafters. A dip you can feel underfoot means the substrate has taken water and the laminate bond has released, which puts the job into [delamination repair](/services/fiberglass-and-delamination/delamination-repair) territory. When enough of the plane reads that way, resealing penetrations is spending effort on the wrong layer, and [full roof replacement](/services/roof-slide-and-awning/full-roof-replacement) becomes the honest recommendation. A high sided Class A coach parked with its long side into the afternoon sun near Eastvale tends to reach that point on the curbside first, because that side takes the longest daily exposure. ## What opening the roof up will and will not reveal A teardown answers some questions precisely and leaves others open, and it helps to know which is which before the work starts. Removing sealant and lifting a flange will tell you exactly where water entered, what condition the butyl or putty tape under the flange is in, whether the wood backer has gone soft, and whether the fasteners still have thread engagement. Pulling a section of membrane will tell you how far the substrate damage extends laterally and whether the laminate bond has failed beyond the visible dip. What it will not tell you is where the water went after it entered. Water travels along rafters, down wiring chases and behind wall paneling, and it can surface a long way from the flange it came through. Opening a roof establishes the source. Establishing the extent still takes a moisture survey of the walls and ceiling from the inside, and sometimes an interior access cut. Expect the roof work to define the beginning of the job rather than the whole of it. --- OCRV Center. 23281 La Palma Ave, Yorba Linda, CA 92887. (949) 799-3387. info@ocrvcenter.com. Serving Eastvale, California in Riverside County, about 20 miles from the shop. All work is performed in shop at the Yorba Linda facility. No mobile, roadside or on-site service is offered. Scope is body, paint, structural, fiberglass, interior and vehicle systems work. Engine rebuilds, transmission rebuilds, drivetrain work, DOT inspections and emission testing are not performed.