Symptoms and initial checks
A returned display would not turn on. The PC could not detect it over USB, even when forced into bootloader mode. Power supply voltages checked out, so the failure was not explained by a missing supply rail.
Fault isolation
While preparing to connect a JTAG programming lead, I discovered that the microcontroller reset line was held low. The line was shared with an FM33256B companion IC. Earlier failures of this part had produced periodic resets, making it a plausible suspect.
Result and limitation
Removing the companion IC released the reset line, and the display powered on again. That isolated the faulty part and recovered the board. The original cause of the companion IC failure was not determined, so this case should not be presented as a complete failure-mechanism analysis.