• So I haven't been in the forum in a few days, outside of a couple minutes here and there at work. I have been working on recovering from a hard drive crash on my computer. It was just a couple weeks ago I attempted to do a backup of my home folder to my google drive, only to find out that it missed some things because I ran out of the 15GB of space that google allows pretty quick. I didn't think that it was going to copy hidden folder stuff too. Well, since then I hadn't figured out a decent solution for filtering what gets saved or whether I even wanted to use up my google drive space for that. So needless to say I didn't have a full backup of the drive. Which gets me into why I am posting this in here. It is not a MySensors question really, but more of an electronics question. The drive was a 128GB SSD. I had heard stories of people that have recovered things from SSD drives by reflowing the components on the board using an old toaster oven, in the event that it's death was due to a possible failed solder joint. Since I don't have a proper reflow oven, and the fact that I didn't have much to loose, I decided to try it. I picked up an old toaster oven from a local second hand thrift shop for $5 US. The one article I read on the subject said to put it in and run it at 450° to 475° for 8 minutes with no preheating, and then open the door and let it cool for 30min. I tried that at both 8 minutes and 10 minutes with no luck. Does anyone in here have any thoughts on that process, or any suggestions for a better one to try to revive the drive? I only need it long enough to copy the data. One of the main things that I want to save is my 3D printing folder as it has all of my OpenSCAD drawings and .stl files. Some of the major stuff I have out on my thingiverse page, so I am not completely out, but enough where it is REALLY frustrating.

    I now have 2 - 250GB SSD drives in a mirrored raid 1 configuration. Needless to say even though I now have a raid set up, once I get things fully set up I will be pushing a backup of my home folder to an offsite server nightly. I would have liked migrating from my old Fedora 23 setup to Fedora 27 under different circumstances, but this "forced" migration was probably a good thing.

  • Mod

    That sucks, I am with you on these circumstances. I can't help much on SSD recovery since I never had an SSD drive failing that hard even at work ( I only saw very few with problems or damaged sectors but they were still able to provide access to data). As backup I am using a account as I have 50GB space for free, where I keep my projects synced between computers. Does the HDD SMART reporting any weird values or the drive is completely not detected by the system?

    As other cheap solution for reflowing, I saw using a 55W car headlight bulb like it was a hot air gun.

  • @gohan Not even recognized by the system. You mention about the car headlight bulb. I didn't really see it for reflowing necessarily, but for board desoldering and component removal. A coworker had emailed me about this a couple weeks ago. This was the video that he sent me:
    I actually purchased a car lamp, but haven't tried it out yet. Maybe this is the time.

  • Mod

    according to a guy that has a youtube channel about electronics repair, the reballing of the chips (usually are graphics chips that undergo a lot of thermal stress) it is actually useless, while in most cases it seems a solution, but it is actually the heating of the chips that restores some internal connections that may got ruined. On a SSD drive I doubt it can be the case since they normally don't get very hot, so personally I think it is most likely a NAND chip just gone bad earlier than others (did you usually kept the disk very full?).

  • @gohan said in Recovering:

    I think it is most likely a NAND chip just gone bad earlier than others (did you usually kept the disk very full?).

    I figured it was probably a bad NAND chip, which is why I figured trying to reflow wasnt going to hurt too much. I never intentionally kept the disk full, but there were a few times where I had an issue where lightdm would fill up the log files with "[' '][' '][' '][' '][' '][' '][' ']...." to the point where I couldn't boot the machine other than in recovery mode. In those cases, I would just go and delete that log file and I was back up and running.

  • Mod

    I was asking because the wear leveling algorithm of the disk firmware work well if there is a 10-15% minimum free space, so keeping disk nearly full for long time does wear some cells quicker than others.

  • @gohan It's not like the drive stayed full for long. I usually ran with between 20GB and 40GB of free space.

Log in to reply

Suggested Topics

  • 273
  • 19
  • 12
  • 2
  • 1066
  • 14
  • 242
  • 1