File 04: A Roof Rack Strike in a Parking Structure

The visible damage was a shattered fan lid, a rack crossbar bent about fifteen degrees, one rack foot lifted at an angle, and a solar panel with a crack running diagonally across the cells. The owner had covered the fan opening with tape and driven the unit here rather than leaving it in a structure, which was sensible.

What the photographs could not show was what happened where the rack foot lifted. Rack feet on a conversion of this kind are bonded and through bolted into the roof panel, sometimes into a reinforcing plate and sometimes not. When the crossbar took the beam impact, that load went into the foot, and the foot went into the metal.

Removing the foot brought a section of roof panel with it, roughly the size of a credit card, torn around the bolt hole and deformed upward for several inches around it. The panel was dented independently at the strike point, a shallow crease about eight inches long, which was the straightforward part of the metal work.

The tear was not straightforward. Roof panel metal on a Sprinter is thin and it is a large unsupported span, which means both that it dents easily and that it cannot simply be filled. The affected area was worked back to contour with the panel supported from beneath, then the bolt area was reinforced with a properly sized backing plate so the next rack foot loads into structure rather than into sheet.

The cracked panel was a model the converter had used for two seasons and then dropped. Nothing with the same footprint was available from any channel. Available panels were either wider or longer, and none of them lined up with the existing mounting rails on the roof, which had been placed for that specific frame.

So mounts were fabricated. A panel of similar output and a workable footprint was selected, brackets were made to bridge from the existing rail positions to the new frame, and the assembly was mounted with the correct standoff for airflow underneath. That is fabrication work rather than parts replacement, and it is the reason nine of the thirty six hours sat in a different department.

When the panel cracked, its lead was pulled and the conductor strands had begun to separate inside the insulation about four inches from the junction. Splicing mid run on a roof is a poor idea: any joint up there is a future entry point for water and a future resistance point in a circuit that runs at low voltage where resistance matters.

The correct answer was to take the run back to the junction and remake it there, inside the enclosure, with a properly sealed gland. That added a small amount of labor and removed a joint from the exposed section. The array was then load checked at the controller before the roof was resealed, so nobody had to open anything twice.

Thirty six hours: twenty seven body at $210 per hour for the metal work, refinishing and reassembly, nine mechanical and electrical at $260 for the array, the harness and the load check. Eleven shop days, five of which were the search for a substitute panel and the bracket fabrication. Every roof penetration was resealed and the unit was water tested before delivery.

The transferable point is that rack feet transmit load into the roof. If anything on your roof takes an impact, the estimate has to include lifting the mounts and looking underneath, not just replacing what is visibly broken. Ask for that explicitly. On this unit it was the difference between a straightforward repair and a supplement.