> Thus you're going to need a lot of energy to create the heat required to cook the silicon. Your chip may desolder itself from the board before it's been appreciably damaged.
Yeah, this is a really weird and unreliable method for killing the flash. I would be more inclined to go the route of super high voltage.
> Encrypt the bulk flash memory, store the key in supercap backed SRAM, and then zero out and short the RAM when triggered.
Probably the easiest method for sure is a decryption key. Older style flash could also be erased using UV, but it seems like this is now not the case.
I'm not entirely sure about using moisture as a method for ensuring the USB device is interacted with in a certain way. I would probably consider some other options:
* A few holes with light sensors that check for a binary code as the stick is inserted.
* A covert fingerprint sensor.
* DIP switches that are somewhat hidden. Perhaps on first entry an LED flashes, and you must set the switches in response to the sequence. In this case somebody could watch you do it and still not understand what they must do.
* Perhaps even something as simple as unplugging it and plugging it in several times.
All of these seem somewhat more reliable that the resistance you happen to create when licking you fingers, and are harder to replicate.
As you say, having it so that just the encryption code is lost means that you could recover the device by loading in the correct key again.
Bonus points for a USB stick that with the correct knock algorithm (plug in 3 times, unplug for 10 seconds, plug in again) presents a secret partition, otherwise presents a dummy partition.
Yeah, this is a really weird and unreliable method for killing the flash. I would be more inclined to go the route of super high voltage.
> Encrypt the bulk flash memory, store the key in supercap backed SRAM, and then zero out and short the RAM when triggered.
Probably the easiest method for sure is a decryption key. Older style flash could also be erased using UV, but it seems like this is now not the case.
I'm not entirely sure about using moisture as a method for ensuring the USB device is interacted with in a certain way. I would probably consider some other options:
* A few holes with light sensors that check for a binary code as the stick is inserted.
* A covert fingerprint sensor.
* DIP switches that are somewhat hidden. Perhaps on first entry an LED flashes, and you must set the switches in response to the sequence. In this case somebody could watch you do it and still not understand what they must do.
* Perhaps even something as simple as unplugging it and plugging it in several times.
All of these seem somewhat more reliable that the resistance you happen to create when licking you fingers, and are harder to replicate.
As you say, having it so that just the encryption code is lost means that you could recover the device by loading in the correct key again.