I can also imagine that they may have internal built-in controller firmware stored in flash and it becomes corrupted.Or something like this.When flash write is in progress and power is lost, with poorly designed internal hardware logic it could happen that internal flash programming logic fails miserably due to too low voltage and flash write can actually hit WRONG place.This is known reason of some firmware corruptions in microcontrollered devices where microcontroller IC has poor reset\brownout logic design.

Okay, take a look on scheme and please explain then how this happens and why, then.Both cards are attached to the same IC named "Menelaus" which does performs power regulations(this activity can be seen in dmesg log as well).It is pretty doubtful it has non-symmetric properties for same functional pins.So I guess electrical properties are near the same (plus or minus minor issues due to PCB layout).Of course corrections from Nokia staff are OK.Guy who posted this not seems to be Nokia staff.Or I'm wrong?