A friend recently asked me about this, and today I thought to pose the question here.

He has a USB stick which was pulled out during a copy without asking the OS nicely. Usually that would just mean a little corruption if there was a write in progress. However, stick stick doesn't work when plugged in. He has tried it on several machines to confirm its the stick itself.

The interesting part is what happens when its detected. Instead of coming up as a USB memory stick, it appears now as a bridge with no media present. I've seen other types of USB devices have similar problems when their sreial ROM is killed by ESD and the controller chip comes up in a default state. But there is no such ROM in this device, its just one big NAND and one of the common controllers, a SM3255. I have plenty of experience with raw NAND, but essentially none with the controllers used in these cheap USB sticks.

Anyone with some experience with these controllers and their failure modes care to take a guess if the culprit is the controller or the flash? I am unfamiliar with what could cause the controller to behave this way, but my pessimistic side thinks the flash is dead and thus the controller presents itself as a bridge with no media rather than as a storage device. For his sake, I'm hoping that there's a chance the NAND flash is good and the controller is just not reading it proper as then I might have a chance of reading the contents sometime. Then would be the fun part, reconstructing a filesystem from out of order blocks.

It might just be the crystal oscillator. Try swapping it out.