Repair or Recover: Backed-Out Fasteners on a 20-Year-Old Through-Fastened Roof

Back-out and overdriving usually come from installation and movement, not from the fastener alone. An overdriven screw crushes the washer and leaves the panel loose, so every wind and thermal cycle works the fastener; an underdriven screw leaves the panel riding on the shank. Panels restrained at both ends by tight fasteners convert thermal movement into hole elongation and gradual withdrawal. Cracked neoprene after 20 years is normal weathering, and once the seal is gone the hole leaks regardless of the fastener's grip.

Before choosing, investigate: pull-out tests on the actual purlin gauge at several bays, since screw thread engagement of at least three threads through the steel is what develops capacity; the condition of the panel laps and whether leaks are at fasteners, laps, or ridge closure; corrosion at the cut edges and fastener heads; and whether the wet insulation is localized or widespread. If the panels are sound and the leaks are fastener-driven, a targeted repair is credible. If the panel laps are corroded or the holes are enlarged and ovalized, new screws in the same holes will not seal or hold, and the repair becomes a temporary patch rather than a permanent fix.

Option A is only permanent if the panels are structurally sound, the holes are not enlarged, and the leaks are truly at the fasteners. Larger-diameter screws in the same holes may split or strip the thin purlin steel, and the new washer still has to seal against a distorted pan. A better version uses oversize sealing screws or moves to new holes, replaces the ridge closure, and removes and replaces the wet insulation in the affected bays. It is inexpensive and can be done in phases without shutting down the plant, but it inherits the existing panel's remaining life and the same thermal movement problem.

Option B eliminates the through-fastened penetrations and gives a new finish and a new warranty, but it must be engineered. New sub-purlins add dead load, typically in the range of 1.5 to 2 psf or more, and the existing purlins, joists, and connections must be checked for that load plus snow, wind uplift, and any drift. Attachments must reach the structure, not just the old panel. A thermal break and a vapor retarder are needed so the old roof does not become a condensing surface between the two layers, and wet insulation must come out rather than be sealed in. Confirm the new panel manufacturer will warrant a recover over an existing metal roof.

The added load question cannot be answered by the roofing contractor alone. A licensed structural engineer should evaluate the existing open-web joists, purlins, and connections for the added dead load plus the governing snow, wind uplift, and drift, and should check ponding and deflection limits for the new assembly. If the joists are already near capacity or have corroded members, the recover is not viable without reinforcement, and the repair option becomes the practical answer even if it is less permanent.

The submittal should include the engineer's sealed analysis, the panel manufacturer's span and uplift tables with the actual clip and sub-purlin spacing, the attachment details showing how the sub-purlins reach the existing structure, and the thermal and vapor control details for the cavity between the old and new roof. For a food facility, include a sanitation plan for debris and fasteners during work over production areas. Documenting the existing wet insulation and its removal is also part of the record, because a recover over saturated insulation will fail and the owner will look to the contractor who recommended it.