Two findings, both pointing at the crossfeed valve. But the question remained: why was the valve cycling incorrectly? A butterfly valve doesn't just malfunction on its own; something had to be telling it to open every 8 seconds.
Faith asked Dr. Sutter for permission to physically inspect the valve assembly. Dr. Sutter approved it.
Faith pulled on a pair of nitrile gloves, took a penlight from her grandfather's tool roll, and reached into the nacelle access panel. She didn't go for the valve itself; she went for the wiring harness—the bundle of wires that carries the signal from the valve's position sensor back to the engine control computer. She ran her fingers along the harness slowly, inch by inch, the way a doctor palpates for a fracture. She followed it from the valve housing up to the bracket where the harness was supposed to be secured to the nacelle frame.
She stopped. "Neil, hand me a torque wrench, quarter-inch drive."
Neil handed it over. Faith placed it on the bracket's mounting bolt. She turned it carefully until it engaged. She looked at the digital readout.
"3 foot-pounds."
She held it up so the room could see. "Spec for this bracket is 16 foot-pounds. This bolt is finger-tight. Someone put it in by hand and never torqued it."
The room didn't react yet; they didn't understand yet.
Faith explained: "This bracket holds the wiring harness in position. When the bolt is properly torqued, the harness is locked tight: no movement, clean signal. But at 3 foot-pounds, the harness has slack. It shifts under engine vibration. Every time it shifts, it sends a noise pulse through the position sensor wire. The engine computer reads that noise as a signal that the fuel tank is low on one side, so it opens the crossfeed valve to balance the tanks. Then the harness shifts back, the noise clears, the computer closes the valve. Eight seconds later, it happens again."
She laid it out, the full chain, step by step: