5 ms·

Does anyone have any advice (other than sending to a professional data recovery service) for how to access data on an HDD whose controller card stopped working,

Does anyone have any advice (other than sending to a professional data recovery service) for how to access data on an HDD whose controller card stopped working, and refuse to accept a spare controller card taken from a brand new identical HDD?

danparsonson 3y ago
Is it not the case that the controller boards store track data specific to an individual drive? I don't think they can be swapped out.
- daneel_w 3y ago
  I fear this is the case...
- ericbarrett 3y ago
  Historically they did contain a map of physically unusable sectors particular to the physical platters in the drive, and I'd be surprised if they didn't now. So the consequence of a controller swap was making the assembly unstable because you were using a foreign sector map; usually it was still enough to recover as long as you weren't writing new data. But nowadays a drive controller is far more complex, e.g. it might implement transparent hardware AES encryption, in which case swapping the board loses the key. And I've no doubt there are many modern manufacturing-related tricks for yield that go into them as well, any of which might make a different set of platters than what was shipped unreadable.
- jaclaz 3y ago
  Yes, they (the PCB's) cannot be swapped unless you also swap the chip containing the adaptive data.
londons_explore 3y ago
How do you know for sure it's the controller? Note that many controllers store parts of their firmware/config on the disk platter, so without the disk platter the board may not show up via sata.
- daneel_w 3y ago
  It's my suspicion, at least. In fact I have two of those drives, bought at the same time, and both died in the same sudden way with no more than two weeks between.
  - londons_explore 3y ago
    Wasn't there a firmware bug in some drives that made them fail after a certain number of seconds of uptime? Was a workaround/fix published?
    - daneel_w 3y ago
      Doesn't ring a bell for me, but given the highly questionable tiering/marketing methods from HDD manufacturers these days and the fact that they have been reducing warranty durations by some 50%, that type of planned obsolescence wouldn't surprise me. The drives are 3.5" 2TB WD Green purchased around 2012, and had been in use for about 1 year when they both died.
      - justinclift 3y ago
        > 2012 If that's not a typo, then it seems like those drives have been powered off for about 10 years. I think powered off hard drives are commonly said to retain data for um... maybe 3 years (from rough memory). So, your drives have probably lost their magnetism (and thus the data). :(
        daneel_w 3y ago
        Not a typo, they've been powered off for around a decade. But HDDs (just like floppies) don't lose their magnetized property in just 3 years, or even 10 years, barring the very rare case of an extreme fluke or environmental exposure. Flash memory (memory cards, USB sticks, SSDs) loses electrical charge relatively fast, however. That might be what you're thinking of.
londons_explore 3y ago
Cool it down to freezer temperatures and see if it works. Heat it up to ~60C and see if it works. This applies to both hard drives and SSD's. A good half of "failed" drives can be made to work for at least a few hours longer with that method. Unlike OP, If you see signs of life, don't mess with clamps or reflowing. Just leave it in the freezer/oven while you take data off it, with longish power/sata cables to a machine just outside the freezer/oven door. I recommend GNU ddrescue for getting data off - when you only have a few hours of service life left till it is dead-dead, it maximizes data recovery in a given time. There are various ways to generate a mapfile to skip recovering free blocks, which are worth using if you suspect the drive is mostly empty.
- lacrimacida 3y ago
  Couldn’t this method risk making it completely irrecoverable if it fails?
  - sgarland 3y ago
    Freezing / heating? Other than the risk of humidity as mentioned, not really. Unless the head crashes into the platter, it's unlikely to do anything that would permanently kill the drive. That said, if you ever have a drive with absolutely critical, must-have data, don't bother with any of this and just ship it to professionals. You'll pay dearly, but they'll get your data out.
- theolivenbaum 3y ago
  And for the freezer method, be careful with humidity when you take it out. Worth putting it in a bag with some dehumidifiers and just take the cable out
kwstas 3y ago
There are bios chips on the controller board than need to be transfered to the new one. In a reasonably modern drive that would be pretty challenging and rewuire a hot air station. https://www.hddzone.com/hard_drive_pcb_replacement.html https://www.hddzone.com/hard_drive_pcb_replacement.html
h2odragon 3y ago
Double check the numbers on the controller boards; HDD's are complicated little computers and the manufacturers change things during production of a particular model of drive fairly often. If you've got the controller board your drive wants and still nothing, then its time for professional help or considering the data lost, imo.
- daneel_w 3y ago
  The PCBs are of the exact same model. My fear is a DRM type scheme connected to serial numbers etc. stored on the platters having to match those of the controller.
  - h2odragon 3y ago
    doesn't even have to be DRM, just "media mapping" where its adjusting tot he individual platters and such. like factory low-level format stuff. I don't know how complex HDDs have got, but I recall giggling at someone installing linux on an HDD controller board several years ago. So I bet its much worse now.
jaclaz 3y ago
Since many years (like 15 or more) the PCB of a hard disk contains a chip (essentially an eeprom or flash device) containing so-called "adaptive data", a set of data written in the factory that are specific to the disk drive (head/platters) the PCB is mounted on. There is specialized hardware (and software) to be able to extract the data and save on another board's memory, but the poorman's way is that of transferring the actual chip from the old board to the new (identical) one. This, commonly referred to as "ROM swap" is not particularly difficult[1] as the chip is usually a rather simple 8 pin one, if you are not into this kind of things a hardware repair shop (like a phone repair one) will normally make this work for you. However newish hard disks may have not this separate chip, it has to be seen which model yours is. Here is a site with some more info: https://hddpcb.eu/gb/content/how-to-swap-hdd-pcb https://hddpcb.eu/gb/content/how-to-swap-hdd-pcb [1] meaning that it can be done DIY if you are familiar enough with soldering/desoldering components
justsomehnguy 3y ago
> other than sending to a professional data recovery service Ah, so you don't really need that data...
- daneel_w 3y ago
  Why the snark? I will survive without the data, but I'm curious about alternatives to throwing $4000 at the problem and forfeiting my privacy and my data's integrity.
  - justsomehnguy 3y ago
    There are many people out there who thought what they could recover the data by themselves, only to swallow a bitter pill later, that's why. Because if you really need the data then you go to people who makes a living by recovering data. But if you are okay to lose the if unsuccessful then it's okay to try, but you should know/tell that beforehand. Reading through the other comments - you have a very low chance to succeed, because if you want to swap controller boards then you need to move adaptive data too, as other had said. But I'm curios what exactly happened, WD Green from 2012 are not the worst drives out there. How exactly they failed, what happens now when you power them on, with SATA connected, without? Did you try external USB2SATA converters/boxes?
    - daneel_w 3y ago
      >"Reading through the other comments - you have a very low chance to succeed, because if you want to swap controller boards then you need to move adaptive data too, as other had said." I'm pretty decent with soldering iron and hot-air gun. Migrating the SMD flash/EEPROM chip shouldn't be too hard. >"But I'm curios what exactly happened, WD Green from 2012 are not the worst drives out there. How exactly they failed, what happens now when you power them on, with SATA connected, without? Did you try external USB2SATA converters/boxes?" The discs spin up but the controller no longer communicates with the host. The computer doesn't see that the drives are attached to the SATA bus. There were no signs of problems coming, they just suddenly stopped working from one power-up to the next. I tried with different motherboards and a couple of SATA-USB bridges, all same result.
toast0 3y ago
Just a double check... you said in thread you have two of these drives that died at 1 year old. Did you have a third, working, drive you're transplanting boards from?
- daneel_w 3y ago
  I have a third drive, working and unused, with same PCB model number. Others in this thread have said that my chances are good if I bring the DIP-8 EEPROM from the broken drives over to the working PCB.

danparsonson 3y ago

Is it not the case that the controller boards store track data specific to an individual drive? I don't think they can be swapped out.

daneel_w 3y ago
I fear this is the case...
ericbarrett 3y ago
Historically they did contain a map of physically unusable sectors particular to the physical platters in the drive, and I'd be surprised if they didn't now. So the consequence of a controller swap was making the assembly unstable because you were using a foreign sector map; usually it was still enough to recover as long as you weren't writing new data. But nowadays a drive controller is far more complex, e.g. it might implement transparent hardware AES encryption, in which case swapping the board loses the key. And I've no doubt there are many modern manufacturing-related tricks for yield that go into them as well, any of which might make a different set of platters than what was shipped unreadable.
jaclaz 3y ago
Yes, they (the PCB's) cannot be swapped unless you also swap the chip containing the adaptive data.

londons_explore 3y ago

How do you know for sure it's the controller? Note that many controllers store parts of their firmware/config on the disk platter, so without the disk platter the board may not show up via sata.

daneel_w 3y ago
It's my suspicion, at least. In fact I have two of those drives, bought at the same time, and both died in the same sudden way with no more than two weeks between.
- londons_explore 3y ago
  Wasn't there a firmware bug in some drives that made them fail after a certain number of seconds of uptime? Was a workaround/fix published?
  - daneel_w 3y ago
    Doesn't ring a bell for me, but given the highly questionable tiering/marketing methods from HDD manufacturers these days and the fact that they have been reducing warranty durations by some 50%, that type of planned obsolescence wouldn't surprise me. The drives are 3.5" 2TB WD Green purchased around 2012, and had been in use for about 1 year when they both died.
    - justinclift 3y ago
      > 2012 If that's not a typo, then it seems like those drives have been powered off for about 10 years. I think powered off hard drives are commonly said to retain data for um... maybe 3 years (from rough memory). So, your drives have probably lost their magnetism (and thus the data). :(
      - daneel_w 3y ago
        Not a typo, they've been powered off for around a decade. But HDDs (just like floppies) don't lose their magnetized property in just 3 years, or even 10 years, barring the very rare case of an extreme fluke or environmental exposure. Flash memory (memory cards, USB sticks, SSDs) loses electrical charge relatively fast, however. That might be what you're thinking of.

londons_explore 3y ago

Cool it down to freezer temperatures and see if it works. Heat it up to ~60C and see if it works. This applies to both hard drives and SSD's. A good half of "failed" drives can be made to work for at least a few hours longer with that method. Unlike OP, If you see signs of life, don't mess with clamps or reflowing. Just leave it in the freezer/oven while you take data off it, with longish power/sata cables to a machine just outside the freezer/oven door. I recommend GNU ddrescue for getting data off - when you only have a few hours of service life left till it is dead-dead, it maximizes data recovery in a given time. There are various ways to generate a mapfile to skip recovering free blocks, which are worth using if you suspect the drive is mostly empty.

lacrimacida 3y ago
Couldn’t this method risk making it completely irrecoverable if it fails?
- sgarland 3y ago
  Freezing / heating? Other than the risk of humidity as mentioned, not really. Unless the head crashes into the platter, it's unlikely to do anything that would permanently kill the drive. That said, if you ever have a drive with absolutely critical, must-have data, don't bother with any of this and just ship it to professionals. You'll pay dearly, but they'll get your data out.
theolivenbaum 3y ago
And for the freezer method, be careful with humidity when you take it out. Worth putting it in a bag with some dehumidifiers and just take the cable out

kwstas 3y ago

There are bios chips on the controller board than need to be transfered to the new one. In a reasonably modern drive that would be pretty challenging and rewuire a hot air station. https://www.hddzone.com/hard_drive_pcb_replacement.html https://www.hddzone.com/hard_drive_pcb_replacement.html

h2odragon 3y ago

Double check the numbers on the controller boards; HDD's are complicated little computers and the manufacturers change things during production of a particular model of drive fairly often. If you've got the controller board your drive wants and still nothing, then its time for professional help or considering the data lost, imo.

daneel_w 3y ago
The PCBs are of the exact same model. My fear is a DRM type scheme connected to serial numbers etc. stored on the platters having to match those of the controller.
- h2odragon 3y ago
  doesn't even have to be DRM, just "media mapping" where its adjusting tot he individual platters and such. like factory low-level format stuff. I don't know how complex HDDs have got, but I recall giggling at someone installing linux on an HDD controller board several years ago. So I bet its much worse now.

jaclaz 3y ago

Since many years (like 15 or more) the PCB of a hard disk contains a chip (essentially an eeprom or flash device) containing so-called "adaptive data", a set of data written in the factory that are specific to the disk drive (head/platters) the PCB is mounted on. There is specialized hardware (and software) to be able to extract the data and save on another board's memory, but the poorman's way is that of transferring the actual chip from the old board to the new (identical) one. This, commonly referred to as "ROM swap" is not particularly difficult[1] as the chip is usually a rather simple 8 pin one, if you are not into this kind of things a hardware repair shop (like a phone repair one) will normally make this work for you. However newish hard disks may have not this separate chip, it has to be seen which model yours is. Here is a site with some more info: https://hddpcb.eu/gb/content/how-to-swap-hdd-pcb https://hddpcb.eu/gb/content/how-to-swap-hdd-pcb [1] meaning that it can be done DIY if you are familiar enough with soldering/desoldering components

justsomehnguy 3y ago

> other than sending to a professional data recovery service Ah, so you don't really need that data...

daneel_w 3y ago
Why the snark? I will survive without the data, but I'm curious about alternatives to throwing $4000 at the problem and forfeiting my privacy and my data's integrity.
- justsomehnguy 3y ago
  There are many people out there who thought what they could recover the data by themselves, only to swallow a bitter pill later, that's why. Because if you really need the data then you go to people who makes a living by recovering data. But if you are okay to lose the if unsuccessful then it's okay to try, but you should know/tell that beforehand. Reading through the other comments - you have a very low chance to succeed, because if you want to swap controller boards then you need to move adaptive data too, as other had said. But I'm curios what exactly happened, WD Green from 2012 are not the worst drives out there. How exactly they failed, what happens now when you power them on, with SATA connected, without? Did you try external USB2SATA converters/boxes?
  - daneel_w 3y ago
    >"Reading through the other comments - you have a very low chance to succeed, because if you want to swap controller boards then you need to move adaptive data too, as other had said." I'm pretty decent with soldering iron and hot-air gun. Migrating the SMD flash/EEPROM chip shouldn't be too hard. >"But I'm curios what exactly happened, WD Green from 2012 are not the worst drives out there. How exactly they failed, what happens now when you power them on, with SATA connected, without? Did you try external USB2SATA converters/boxes?" The discs spin up but the controller no longer communicates with the host. The computer doesn't see that the drives are attached to the SATA bus. There were no signs of problems coming, they just suddenly stopped working from one power-up to the next. I tried with different motherboards and a couple of SATA-USB bridges, all same result.

toast0 3y ago

Just a double check... you said in thread you have two of these drives that died at 1 year old. Did you have a third, working, drive you're transplanting boards from?

daneel_w 3y ago
I have a third drive, working and unused, with same PCB model number. Others in this thread have said that my chances are good if I bring the DIP-8 EEPROM from the broken drives over to the working PCB.