RV roof repair, reseal and replacement

Water enters at a penetration, runs along a rafter or a seam, and appears somewhere else entirely. A stain above the dinette very often starts at a vent, a skylight or an antenna base several feet away. Patching the stain end of that path fixes nothing.

Tracing it means starting at every penetration on the roof, checking the sealant condition at each, and following the framing. It is unglamorous and it is why a roof repair that skips it comes back the next wet season.

A roof with sound decking and failed sealant is a recoat and reseal, at the bottom of the range. A roof with soft decking is a replacement, because a new membrane over rotten substrate is a cosmetic repair on a structural problem.

Deciding between them takes walking the roof and pressing it, not looking at it. Softness underfoot near a seam or a vent is the signal, and once it is present the question is how far it has travelled rather than whether to replace.

EPDM is a rubber membrane, usually white, chalky as it ages. TPO is also a membrane but a different chemistry with different compatible sealants. Fiberglass is a hard shell, so it cracks rather than tears. Aluminium is panels and seams, which means the seams are the risk.

Sealant choice follows from that. Self levelling goes on horizontal surfaces, non sag on vertical, and using a product incompatible with the membrane is a common cause of a reseal that fails within a year.

This climate is not a freeze problem, it is an ultraviolet and thermal cycling problem. Sealant hardens, shrinks and lets go at the edges long before it looks obviously bad, and it does it faster on a coach stored outdoors than the maintenance interval in the manual assumes.

A six month sealant inspection is not a marketing number in this region, it is roughly how long a bead lasts between checks when it spends summer at surface temperatures well above air temperature.

By the time a brown ring appears on a ceiling panel, water has been in the structure for a season or more. The earlier signals are duller and easier to miss, and every one of them is cheaper to act on.

Smell is the most reliable of them. A coach that has been closed up for two weeks and smells faintly of damp cardboard has wet substrate somewhere, and no amount of ventilation fixes the reason. The second most reliable is a fastener that will not stay tight, because wood that has been wet no longer holds a screw.

• A damp smell after the coach has been shut for a fortnight
• Screws in a ceiling trim or a cabinet that keep backing out
• A wall panel that has gone slightly wavy near a seam
• Sealant that has pulled away at one edge, even where the bead looks whole
• A soft or springy patch underfoot near an exterior wall
• Interior lights or a vent fan behaving oddly after rain

Carriers separate sudden damage from gradual damage, and roofs sit right on the line. A branch through a membrane, hail, or something torn off at speed is usually a covered loss. Water that has been entering slowly through failed sealant is usually not, on the grounds that maintenance is the owner's responsibility.

Which means the interesting cases are the ones where a covered event opened a path that then leaked for months. The claim then turns on documentation: whether there is a record of the event, and whether the sealant condition before it can be established. An owner who photographs the roof twice a year has that record. An owner who does not is arguing from memory against an adjuster with a checklist.

Nobody enjoys hearing that a five figure repair is not covered. It is worth knowing the rule before the leak rather than during the phone call.

A failed bead at a vent is a small job while it is only a failed bead. Once water has been through it for a wet season, the decking under the membrane has usually softened in a patch, and the repair now includes structure. After a second season the water has generally reached a wall cavity and the floor at the edge of it, and the interior comes out to reach any of it.

The progression is not dramatic and that is exactly the problem. Nothing visibly fails between those stages, so there is never an obvious moment to act. The trigger has to be the calendar rather than a symptom.

Inspecting is entirely reasonable. Walking the roof carefully where the structure supports it, photographing every penetration, and noting where a bead has lifted is useful work and it is the part that catches problems early. Cleaning a membrane with a product compatible with it is also reasonable.

Sealing is where it goes wrong. A tube of the nearest product from a general hardware aisle, applied over a bead that has already released, traps water rather than excluding it and hides the evidence of where it was entering. Self levelling and non sag are not interchangeable, and neither is compatible with every membrane. If you are unsure which membrane is up there, that is the first thing to establish rather than the thing to guess at.