LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   General (http://www.linuxquestions.org/questions/general-10/)
-   -   Broken SD card experiment (http://www.linuxquestions.org/questions/general-10/broken-sd-card-experiment-4175477013/)

H_TeXMeX_H 09-13-2013 12:16 PM

Broken SD card experiment
 
I recently had an SD card come up with I/O errors when I try to read and write to it. So, I did what I usually do and wiped it with random data, and then created a partition on it. However, unlike other SD cards that I have recently wiped, after wiping and partitioning I mount it only to find 62 GB of randomly titled files. I try to format it again, but it doesn't help. I keep trying a few times, but it just doesn't work.

Instead of throwing it away, I remember an article where it said that they came up with new SSDs that last longer because they internally refresh the cells by heating them up.

DISCLAIMER: DO NOT TRY THIS AT HOME, BECAUSE IT MAY NOT BE SAFE.

So, I decided to remove the plastic casing carefully with a screwdriver, take the circuit board out and hold it over a fire (gas stove) using some pliers. I tried it first with low heat (a foot above the fire), that didn't work. I kept trying closer to the fire, and that didn't work either. So, I finally tried putting it right above the fire until it smelled like melting solder and the board was close to melting. Well, it may have actually worked. After putting the board back in the plastic casing, wiping it and partitioning it, I can read and write to the card fully and without error.

I'm pretty sure I tried everything to fix it before heating it up, so it may be that the heat fixed it ? I'll have to try it on one of the old USB sticks I have at my other place.

qlue 09-13-2013 01:34 PM

Interesting! Although I would rather try using the heat gun from a solder reflow station.
Of course, everything I know about electronics tells me that intentionally heat I.C. packages up is not a good idea to begin with!

H_TeXMeX_H 09-14-2013 03:15 AM

I'm not 100% sure if this actually is the reason for the fix. I'll have to do some more tests in the future.

I was just wondering if it is theoretically possible.

Here's the article I mentioned:
http://www.bbc.co.uk/news/technology-20579077

qlue 09-14-2013 05:21 PM

That article talks about microseconds. I really cannot say if there is any real truth to it.
It's still interesting, and it's worth keeping in mind as an experiment.

ntubski 09-14-2013 05:56 PM

Quote:

Originally Posted by qlue (Post 5027757)
That article talks about microseconds.

It talks about hours too:
Quote:

Heat has long been known to help heal degraded materials in old flash memory. But because the heat healing process meant baking the memory chip in an oven at 250C for hours, few saw it as a practical solution.

H_TeXMeX_H 09-15-2013 01:42 AM

I was thinking of using an oven, but I have to figure out on what to put the circuit board when it is in the oven. Probably on a small metal plate.

k3lt01 09-15-2013 01:55 AM

Quote:

Originally Posted by H_TeXMeX_H (Post 5027858)
I was thinking of using an oven, but I have to figure out on what to put the circuit board when it is in the oven. Probably on a small metal plate.

With baking paper on top and then the card. If you just sit it on a metal plate (baking tray?) you may get spots of burn through the contact. If you use baking paper it should be non-stick but will still allow the heat to do its job.

jamison20000e 09-15-2013 12:10 PM

Interesting. I have an SD card my sister returned to me after stating her two year old was chewing on it, only miner bite marks but I've been waiting so it would dry internally before hooking up one of my old PCs to test it. If it dose not work I will have to try this.

Off-topic and different concept: this is reminding me of "How to Hack a laser pointer into a burning laser" "by destroying the circuit board limiter that keeps the laser from being as powerful" (using heat)

qlue 09-15-2013 07:43 PM

I don't think you can safely heat the circuit board to 250C without melting the solder. And I suspect that while that heat may heal the flash chips, it may well destroy the controller chips.

Of course, if the flash drive is dead already, you have nothing much to lose.

P.S: The maximum output of a laser crystal is determined by the crystal's quantum properties. All, that hack will succeed in doing is over-currenting the pump diode. (which will probably just let all the magic smoke out!) :p


All times are GMT -5. The time now is 11:07 AM.