Bad Hard Drives

[raptor1956]

In my 35 years of PC ownership, give or take, I've had perhaps 10 HD's fail on me but my track record over the last decade or so has been much better than the 35 year record.  I was just going through some of my retired PC's including a Sony Vaio laptop of 2004 vintage and a desktop PC of 2003 vintage and neither one will boot.  I doubt either have been fired up since perhaps 2008 but I was thinking about junking them but wanted to check for important data before drilling holes in the HD's and tossing. 

So, I open up my old desktop and pull out the four HD's, two of which are SATA drives and the other two, being older and dating to about 2000, are IDE drives.  I then popped the drives into external enclosures to see if any of them come to life and two of them did -- guess which ones -- that's right, the two oldest IDE drives.  The newer SATA ones are both doing the 'click-click-click' thing and even a round of slapping them onto a table was unsuccessful in bringing them back.

Besides the slap method which only occasionally works is there any other thing I can try -- I mean, the worst that can happen is that they're still dead...

Interestingly, in the 35 years I've owned PC's I've updated and modified pretty much all of them but only ever built two PC's from scratch.  For the most part it's cheaper to buy an off-the-shelf PC/laptop, but if you have a specific need and usually that means performance, then it's often cheaper to build rather than buy and wind up paying huge mark-ups to get the higher spec version.  In 2003 I was getting more serious with digital photography and at the time running Photoshop was pretty power demanding so I built a desktop PC to do that.  After about 2007 I was travelling a lot (85% of the time) so a laptop was more practical and by then a performance laptop was faster than my desktop.  Fast forward to early 2016 and as I'm now doing a lot more 4K video editing I once again needed to build a new PC with the i7-55820K uP as the central component.

OK, so if there's some tricks to revive a dead 'clicking' SATA HD let me know and I'll give it a shot -- couldn't hurt!


Brian

[sleemanj]

Since you have nothing to lose, it's certainly possible to pop the lid off and unstick the head manually to get it able to spin up, reassemble, and pull off as much data as you can.

[hans]

My motto with old storage media/backup archives is that if I haven't had the need to pull files from them for years, then I can safely assume to delete them without trouble. There is nothing of current interest in it; stuff I did 10 years back is barely relevant for me (I got into computers by website building and PHP scripting as a 13 year old)

I might even forget completely what was on the storage media or in the archives (some of my older ones, like 10 years ago, are a mess).

Nevertheless, I do feel a moment of nostalgia and satisfaction if I wander through my old stuff. Like.. jeez did I really go through this much hell to make this? Glad I know better now

[XOIIO]

Since you have nothing to lose, it's certainly possible to pop the lid off and unstick the head manually to get it able to spin up, reassemble, and pull off as much data as you can.

Jesus christ do not do that unless you want to guarantee your data loss, doing that at home is one of the stupidest things you can do if you have anything remotely important. Even the tiniest specs of dust can get between the heads and the platters scratching them and destroying your data.

One trick that can work is putting them in a freezer for a few hours, if anything has seized, they may work long enough to get your data back

Slapping is also a poor idea if the heads have not parked properly, that too can cause damage to the platters.

I would also try them in the old system again, there could be some sort of issue with the power delivery of the enclosure/it's adapter which they are just not liking, I have not noticed that with an external enclosure yet but I have a laptop that refuses to let the hard drive work using one power adapter, but another near identical one works fine. Unfortunately I don't have a scope to see if it's some odd ripple on the output or something that could be causing it.

[rrinker]

 Don't just chuck hard drives - at LEAST salvage the magnets!

The most fun way I ever permanently erased one was when I found out they will spin up with the cover removed - the platter motor has no real torque to speak of, but I was able to lightly drag a screwdriver on the spinning platter until it polished off all the magnetic coating, and I was left with a shiny aluminum disk. But, pounding the platters with a hammer is therapeutic in its own way, and a lot faster. Then there are the commercial shredding machines that can take the whole hard drive and render it into tiny bits of metal, plastic, silicon, and what have you.

But always - always, take the magnets. 

[nctnico]

I have had some luck rotating the drive quickly (hold in hand and use wrist to rotate) while it is powered to get the platters to spin again but usually the platters get stuck again when the drive has been left unpowered for a while.

[madires]

Scott Moulton gave several talks about hard drive recovery at DefCon and Blackhat. Most talks should be available on YouTube. His channel is https://www.youtube.com/user/SuperFlyFlippingA/videos .

In my data center @ home I'm replacing desktop HDDs every 3-4 years and server HDDs every 5-6 years. So far I had no issues with bad drives in operation. Only one or two decommissioned drives died. I've given away some old drives to relatives and friends and they usually got another few years out of the drives. And of course, I'm running backups regularly. Old SCSI drives got the most amazing lifetime. Last year I got rid of a pair of SparcStations  with nearly 20 years old SCA drives. The drives were still fine.

[dannyf]

about 10 years ago, I ran through quite a few bad hard drives over a 1-2 year period. Mostly IBM/Hitachi, WD and to a lesser extent Seagate. All are PATA drives.

Since then, I haven't had any bad hard drives - knock on wood, especially sata ones with 1TB or bigger capacity. I do have a few 80GB-250GB sata drives that are SMART-reporting reallocated sectors but they have logged on-time of close to hundred of thousands of hours. No out right failures, however.

With the proliferation of ssd, I think hd reliability will only improve.

[System Error Message]

your laptops not booting, does it post? If not try the discharge trick first. It has brought a few motherboards back to life for me.

[raptor1956]

your laptops not booting, does it post? If not try the discharge trick first. It has brought a few motherboards back to life for me.

Didn't talk much about the laptop but what do you mean by the "discharge trick"?  When first powered up yesterday the battery was deeply dead and the laptop booted into BIOS where I had to reset the real time clock etc.  So, it started from a discharged state to begin with...


Brian

[raptor1956]

Since you have nothing to lose, it's certainly possible to pop the lid off and unstick the head manually to get it able to spin up, reassemble, and pull off as much data as you can.

Jesus christ do not do that unless you want to guarantee your data loss, doing that at home is one of the stupidest things you can do if you have anything remotely important. Even the tiniest specs of dust can get between the heads and the platters scratching them and destroying your data.

One trick that can work is putting them in a freezer for a few hours, if anything has seized, they may work long enough to get your data back

Slapping is also a poor idea if the heads have not parked properly, that too can cause damage to the platters.

I would also try them in the old system again, there could be some sort of issue with the power delivery of the enclosure/it's adapter which they are just not liking, I have not noticed that with an external enclosure yet but I have a laptop that refuses to let the hard drive work using one power adapter, but another near identical one works fine. Unfortunately I don't have a scope to see if it's some odd ripple on the output or something that could be causing it.

I think I'll give the freezer method a shot, if that fails its drill time!

Beyond the HD's I still need to dispose of the laptop and desktop PC's.  I plan to harvest the ATX PS from the desktop and the power brick for the laptop but other than that and maybe some of the heatsinks and heatpipes the rest is going in the trash.  Actually, I probably ought to check for a recycling center in the area or for a proper disposal of e-waste.


Brian

[rob77]

if you just want to make sure no-one will ever read that drive after you dump it, then dump it with the platters removed there is no-one on this planet who will be able to mount the platters back in the correct order and correctly aligned

regarding reading the data before you dump it.... if you didn't miss the content for the last few years, then you won't miss it.

[nctnico]

if you just want to make sure no-one will ever read that drive after you dump it, then dump it with the platters removed there is no-one on this planet who will be able to mount the platters back in the correct order and correctly aligned

I wouldn't be so sure about that. There is ones & zeroes which can be read from the disk and based on that I'm pretty sure they can reconstruct whatever is on it. Your best best is to physically destroy the platters.

[TerraHertz]

Or, to dispose of unwanted PCs and laptops, you could just run Bleachbit or any other HD scrubber utility, then put the machine out on the street or your front fence intact with a note saying "free for the taking, works but OS-wiped." Include any manuals, driver CDs, etc.
Someone who isn't as well off as you will probably appreciate it.

[ Added: HD scrubber software works by repeatedly overwriting all sectors with random bit patterns. After even 3 or 4 cycles there's no way any original data is going to be recovered by anyone. And the software overwrites as many hundreds of times as you like. Physically smashing working drives is just barbaric. If the drive is dead and you can't run a scrubber, then sure, recover magnets then crush or burn the platters. Recycle the metals.]

Likely failure causes of long-idle drives:
 - Stuck heads - easily fixed by tapping and quickly rotating the drive case back and forth, using the platter inertia to break the sticktion.
 - Firmware flash memory data loss. This is guaranteed to happen eventually, and as flash cells get smaller the timescale must be shrinking. Very unlikely any modern drive is going to work after 15 to 20 years storage. It might be a good idea to reflash the firmware every 5 years.
 - Decay of the magnetic domains on the platter, also the servo tracks. As head, track, domain size and media thickness shrink, this timescale is bound to shrink too. Does any drive manufacturer publish figures on expected drive data retention time? Not sure if reformatting and rewriting your data regularly would help, as some things like servo tracks (and sector headers?) are not rewritten.
 - The point-contact connections between stuff inside the drive case, to the PCB. The usual story of metal-to-metal contacts that are not pure gold, developing an oxide insulation layer able to withstand the low voltages. Sometimes just taking the PCB off, cleaning those contacts, and putting it back can resuscitate a drive.

[System Error Message]

some boards if they dont boot up need to be discharged by pressing the power button for a long time. In your case it booted to bios so you have a working laptop.

There are utilities that show the detailed S.M.A.R.T. info about the drive so you can see the health of it.

Seagate has a very good live boot utility for hard drives. I suggest you check it out.

[raptor1956]

if you just want to make sure no-one will ever read that drive after you dump it, then dump it with the platters removed there is no-one on this planet who will be able to mount the platters back in the correct order and correctly aligned

regarding reading the data before you dump it.... if you didn't miss the content for the last few years, then you won't miss it.

There are data recovery outfits that would laugh at that statement!  There are government agencies that would laugh hard!


Brian

[Halcyon]

if you just want to make sure no-one will ever read that drive after you dump it, then dump it with the platters removed there is no-one on this planet who will be able to mount the platters back in the correct order and correctly aligned

regarding reading the data before you dump it.... if you didn't miss the content for the last few years, then you won't miss it.

Australia's state and federal Police organisations have the resources to do exactly this and it's not particularly complicated or expensive. Most people know that Governments are usually behind the times when it comes to technology. You'd be surprised by what the average nerd can achieve at home. A lot of data can be recovered from a single platter. There is a reason secure destruction processes exist and they usually include a combination of erasure by way of a deguasser and physical shredding.

Even at home I never ever just throw a hard disk in the bin. It's always over-written several times with random garbage then physically destroyed into several pieces (bolt cutters work great for 2.5" drives). I also never sell my old phones because data can still be recovered if you do a factory erase or delete the encryption key. I either hang onto them for nostalgic reasons or they get the same treatment as the hard disks (paying particular attention to the memory chips).

Our retail phone shops have devices which can read and bypass security on many phones (up to an including Apple iPhone 4) so they can "backup" customer data if they forget their password or buy a new phone.

[XOIIO]

Not going to quote to save time but I'd also salvage any small heatsinks from vrm or onboard graphics, those can come in handy. Nice to see you will be recycling some of it.

[HighVoltage]

When I had drives in an external housing and it made click click during start up , one time the current supply was not high enough and a different type of external connector worked much better and the drive started right away.

As for destroying data, I always open up the drives, selvage good parts like magnets and use a hammer on the disk.

[raptor1956]

Well, I tried the freeze method and it didn't help at all with the first HD, but the second one actually spun up but it to would not respond beyond that.

So, I pulled them apart and wanted to harvest the spindle motors but I could only do that with one of them as the other motor was press fit into the frame and without the tools to press it out it's not going to be of any use to me.

But, I was able to get one of the spindle motors and now I need to figure out how to drive it.  It looks to be a BLDC motor in Y-configuration as I measure about 2.7 ohms between all three pins -- there is no other pins and the frame is insulated.

I have run some Neodymium magnets across the platters and will follow that up with my drill.


Brian

[dannyf]

then dump it with the platters removed

as our ex-secretary of state can tell you, hammer is the most effective approach. Bleachbit is no math.

there is no-one on this planet who will be able to mount the platters back in the correct order and correctly aligned

I wouldn't bet on that.

[hasithvm]

Well, I tried the freeze method and it didn't help at all with the first HD, but the second one actually spun up but it to would not respond beyond that.

So, I pulled them apart and wanted to harvest the spindle motors but I could only do that with one of them as the other motor was press fit into the frame and without the tools to press it out it's not going to be of any use to me.

But, I was able to get one of the spindle motors and now I need to figure out how to drive it.  It looks to be a BLDC motor in Y-configuration as I measure about 2.7 ohms between all three pins -- there is no other pins and the frame is insulated.

I have run some Neodymium magnets across the platters and will follow that up with my drill.


Brian

Small blowtorch should also do the trick. You can also reuse the blowtorch to make a perfect crème brûlée crust afterwards!

[madires]

Forget about the hard drive. I strongly second the crème brûlée

[Whales]

I've heard stories of heating them in ovens to soften seized bearings.  I'm not sure if this is an issue much nowadays.

[macboy]

...
 - The point-contact connections between stuff inside the drive case, to the PCB. The usual story of metal-to-metal contacts that are not pure gold, developing an oxide insulation layer able to withstand the low voltages. Sometimes just taking the PCB off, cleaning those contacts, and putting it back can resuscitate a drive.

I have recovered a few drives this way. The contacts had become nearly black with corrosion. The drives all had different symptoms ranging from excessive CRC errors showing in SMART, to not showing up during BIOS POST, to click of death. Cleaning with a rubber pencil eraser removed all the corrosion, revealing shiny contact pads again. The drives have worked for well over a year since this fix.  I advocate this as a potential fix for a seemingly failed drive, especially one that had been working but has failed after a period of storage.