Data recovery from a failed Seagate?

Bluefront · Post by **Bluefront** » Thu Aug 16, 2007 3:56 pm

My sister's main back-up drive failed......an external USB drive. She has some important data that needs to be extracted. The motor failed......I carefully took the top off and watched the two platters. They try to turn when powered on, but barely move. Giving the platters a spin doesn't do anything.

Data recovery places have quoted $500+ to attempt a fix. Way too much. Has anyone tried to fix such a problem? I'm thinking to buy another (working) Seagate drive of the exact size/type, and attempting to swap platters. Never tried such a thing.....but I have enough small tools to do so. What do you think?

nick705 · Post by **nick705** » Thu Aug 16, 2007 4:52 pm

arrghh... sorry to say it, but unless you've got a Class 4 clean room, I reckon any chance you may have had went south the minute you popped the cover.

It could be worth a shot, but without specialist tools I wouldn't bet the farm on it. $500 is pretty reasonable for this kind of recovery BTW... if the data isn't worth that I think you'll probably just have to write it off to experience...

jhhoffma · Post by **jhhoffma** » Thu Aug 16, 2007 5:49 pm

Well, it'll cost you, but if the data is THAT critical, then you may want to try this tool. It'll cost you about $90 bucks, which is cheaper than the cost of professional recovery. I don't know that the odds of actual data recover is any better or worse (especially considering you've opened the case.

If it's not worth throwing $100 at a chance or $500 on a little better chance, buy a new drive and move on.

fjf · Post by **fjf** » Fri Aug 17, 2007 12:53 am

If you take apart the platters and mount them again in a different drive, you'll face 2 problems:

1-If the platters are not PERFECTLY aligned as they were originally, the new drive wont be able to read them.

2-If the firmware is not EXACTLY the same, again no luck. This has to be solved getting a drive manufactured in a date close to yours.

Watch this one:

http://www.youtube.com/watch?v=Kx-D1nJcv0k

There are like 7 parts, very informative.

jaganath · Post by **jaganath** » Fri Aug 17, 2007 1:49 am

seriously, watch that Youtube video before you do anything else.

I'm thinking to buy another (working) Seagate drive of the exact size/type, and attempting to swap platters. Never tried such a thing.....but I have enough small tools to do so. What do you think?

according to the vid, not only does it have to be the exact same model, it also has to be manufactured at almost exactly the same time. I'm thinking it's going to be pretty hard to find such a drive.

Matija · Post by **Matija** » Fri Aug 17, 2007 2:03 am

If you haven't f*cked up the platters by opening the drive (why the hell did you do that??!?!), data recovery companies can probably cheaply scrape the data off.

In the future,

0) Do not open hard drives!!!!!!!!!
1) Use RAID-1.
2) Use a backup drive daily and keep it *stationary*. No carrying allowed!
3) Burn the important stuff to two DVD media (followed by doing a surface scan) weekly.

Bluefront · Post by **Bluefront** » Fri Aug 17, 2007 2:23 am

Thanks for that video link.....can't watch it at home due to my slow connection. I'll try at work today. I didn't think it would be easy, but I might be able to locate another Seagate of the same pro. date.

As far as that "clean room" thing......this would be a temporary fix to retrieve data, not a permanent fix for the drive. I know others have done this, so there's some hope. I have some experience with very small devices. And I probably have enough tools for the operation. I am a "mechanic" after all.

Bluefront · Post by **Bluefront** » Fri Aug 17, 2007 2:32 am

Oh.....this is not my drive. My sister cannot handle difficult back-up procedures. Most of her important data is on her hard drive, duplicated on an external USB drive she carries around, and backed up on CDs frequently. That's about all she can handle. For whatever reason, there happened to be some data on the USB drive she cannot find on her main computer drive, or the back-up CDs.

We got a bunch of quotes on data recovery before I touched the thing. $500 is too much......that was the cheapest.

nick705 · Post by **nick705** » Fri Aug 17, 2007 2:38 am

jhhoffma wrote:Well, it'll cost you, but if the data is THAT critical, then you may want to try this tool. It'll cost you about $90 bucks, which is cheaper than the cost of professional recovery. I don't know that the odds of actual data recover is any better or worse (especially considering you've opened the case.

If it's not worth throwing $100 at a chance or $500 on a little better chance, buy a new drive and move on.

If the data's critical and there's no backup, you really don't want to be dicking around with it yourself at all, as you'll likely make things worse - just hand it over to the pros with their specialist tools and knowledge. It's only worth attempting it yourself as a shot to nothing, and if it won't be a complete disaster if things don't pan out.

Provided there's no physical damage to the platters (they *will* have been contaminated due to exposure to a non-clean environment, and the bill will probably go up accordingly), there's a very good chance all the data can be retrieved intact. It's really just a question of whether it's worth it...

nick705 · Post by **nick705** » Fri Aug 17, 2007 2:42 am

Bluefront wrote: As far as that "clean room" thing......this would be a temporary fix to retrieve data, not a permanent fix for the drive. I know others have done this, so there's some hope. I have some experience with very small devices. And I probably have enough tools for the operation. I am a "mechanic" after all.

hmmm... OK, but we're dealing with tolerances of microns here - I think the most you can hope for is an interesting (and catastrophic) series of head crashes.

Look forward to seeing how it turns out...

whiic · Post by **whiic** » Fri Aug 17, 2007 4:08 am

Bluefront: "The motor failed......"

If platters don't spin, it doesn't necessarily mean the motor has failed. That's only one out of many possible reasons.

"I carefully took the top off and watched the two platters. They try to turn when powered on, but barely move. Giving the platters a spin doesn't do anything."

Did you actually expect opening the cover to help? You could have taken the (unopened) HDD in you hand and twist it axially while it tried to spin up and achieve pretty much the same effect as trying to rotate spindle by hand. You only contaminated the HDD by opening it, making even professional data recovery pricier and more likely to fail. Of course twisting the drive while powering it on isn't good for the drive but compared to opening the top cover, it's relatively harmless.

"Data recovery places have quoted $500+ to attempt a fix."

Is that before or after you opened the top cover or did you not bother to mention that you have contaminated it with dust? Because I'm afraid 500 bucks don't yet include a clean room recovery which is probably necessary as the drive has already been contaminated. If you hadn't opened it, it might be possible to get it working without one.

"I'm thinking to buy another (working) Seagate drive of the exact size/type, and attempting to swap platters."

That is doomed to fail. Swapping PCBs is both more likely to succeed and would not kill the replacement drive (meaning after recovery attempt (whether it's successful or not) you can return the PCB to your replament and use it normally (unless it got zapped by ESD or something like that).

While PCB swap is much more likely to work, easier to perform (and doesn't need a clean room), and would not need you to permanently lose a perfectly operational donor HDD, you would still need a donor that's manufactured within weeks of your original.

"Never tried such a thing.....but I have enough small tools to do so."

No, you don't have the right equipment.

jhhoffma: "Well, it'll cost you, but if the data is THAT critical, then you may want to try this tool. It'll cost you about $90 bucks, which is cheaper than the cost of professional recovery."

How would a piece of software solve electrical or mechanical failure? Complete waste of money.

peteamer · Post by **peteamer** » Fri Aug 17, 2007 10:11 am

whiic wrote:While PCB swap is much more likely to work, easier to perform (and doesn't need a clean room)

(Bold mine...)

Have you managed to find a supply of HD's that only have one, external, PCB?
I've never seen one, not even the behemoth 40MB, 5 1/2", 1KG drive I took apart for the magnets (and yes they do/can hurt if not very carefull

).....

Bluefront · Post by **Bluefront** » Fri Aug 17, 2007 3:38 pm

Of course I'm only guessing the motor failed.....it moves the platters slightly. I can hear a clicking noise, at which time the platters move a little.

Before I took the top off I tried various things, like holding it at different angles, tapping it as I powered it on, freezing the drive.....trying to start it cold, etc. I thought there might be something loose inside jamming the platters.....or at least I might see a problem. As it is, everything inside looks ok.

Believe me....unless there is a special puller needed, I have enough tools here to take anything apart. And I can make tools if needed. I am actually a machinist turned mechanic. I never had a hard drive apart before.....but somebody assembled it. It can come apart......now if it will be able to retrieve data after being re-assembled is the big question mark.

A software program certainly won't fix this problem.....no matter how good it is.

nick705 · Post by **nick705** » Sat Aug 18, 2007 2:19 am

Bluefront wrote: Believe me....unless there is a special puller needed, I have enough tools here to take anything apart. And I can make tools if needed. I am actually a machinist turned mechanic. I never had a hard drive apart before.....but somebody assembled it. It can come apart......now if it will be able to retrieve data after being re-assembled is the big question mark.

Taking something apart is one thing... putting it back together so that it works is quite another. IIRC, you don't have a broadband connection, so you might not have seen those YouTube clips linked above - to be honest, it really looks like you're stuffed from the word go. Not necessarily because of contamination (although that's a factor), but in a multi-platter drive the individual platters are apparently free to spin independently once the spindle cover has been removed, and as the data is stored in "cylinders" which span all the platters, once they've become misaligned the data is gone for good. You'll need to realign the platters *exactly* (to within small fractions of a micron) to make the data recoverable, and without a reference point I can't see any way on earth you could do it, even if you had the right tools.

That's just *one* of the major problems you'll encounter - there are more, but since that one is a showstopper in itself, there's not much point in boring you with them...

If you hadn't cracked the drive open, there may have been some chance, but by doing so I think you've put even professional recovery well beyond the bounds of what's financially realistic.

Having said all that, I'd love to see you do it and prove us all wrong, but I just don't think it's going to happen...

klankymen · Post by **klankymen** » Sat Aug 18, 2007 2:48 am

Bluefront wrote:For whatever reason, there happened to be some data on the USB drive she cannot find on her main computer drive, or the back-up CDs.

Well, if it's not on the main drive, you may have to consider the possibility that it's not on the back-up either, in which case investing any money in recovery would be none other than a waste of money.

Bluefront · Post by **Bluefront** » Sat Aug 18, 2007 3:53 am

That lost data story is getting clearer to me....after some questioning. My sister works at a big hospital in the city. She's in research.....the data is apparently a section of a new medical text she's writing. This area of the city is so dangerous, nobody wants to be seen carrying anything of value. People get mugged all the time.

Also that research field is so competitive, data is constantly being stolen from inside the building. She's afraid to connect to the LAN, for fear of data loss. So she keeps all data on this one drive, and on CDs which are locked in a safe at night. On Friday she gets someone to walk her back to her car with the drive.....and takes it home to transfer the data to her home computer. Then she clears the data on the USB drive, and takes it back to work on Monday.

I told her to use a small flash drive she could carry in her pocket.....that's probably what she'll do from now on. I couldn't work in those conditions....but she's been doing it for years. It's only in the last ten years or so that St Louis has become this dangerous.

whiic · Post by **whiic** » Sat Aug 18, 2007 12:07 pm

peteamer: "Have you managed to find a supply of HD's that only have one, external, PCB?"

Well, there might be some pre-amps attached to moving actuator arm itself, to amplify the weak signals before they are transmitted to external PCB, but other than that, there is just one PCB and it's on the outside.

I don't really understand you question. Do you mean one PCB or one type of PCB?

If you mean the number of PCB attached to the HDD, you might have misundestand the terminology: PCB = Printed Circuit Board = that (typically green coloured) glass-plastic-fibre composite board that mounts various IC chips as well as passive components. Typically there's just one of them. ICs: many. PCB: just one. (Not including a very small one that may or may not be inside the case.)

What did you mean one type of PCB? And that "supply of", did you mean whether there is distributors that specialize on selling HDDs of specific old variants, for purposes of using as donors? You'll probably see them on Ebay... and find them by Googling. You may also notice that retailers for old computer hardware know the price of a perfectly matching HDD (not for the general public but for the one who really needs it) and take advantage of it. Be ready to pay extra unless you have a good poker face. Companies would probably keep the same price (whether it's low or high) regardless of whether they think you're in distress but individual would more likely lift the price up.

If you want to make sure you don't give away too easily, you can always tell you need identical drive for a RAID array and you're afraid of performance loss with non-identical drives, thus hiding the true magnitude of distress and giving an excuse that appears valid for knowing things like subvariant model number, firmware code version, plant of manufacture and manufacture date.

Bluefront: "Of course I'm only guessing the motor failed.....it moves the platters slightly."

Since the platters don't spin properly, it may be spindle motor controller chip that is busted (I have personally encountered on WD that had one) or motor coils (but if that's the case, you can't do anything about it). Bearings are probably OK if the platters move even a bit when trying to power up. I've heard of some stuck bearings on some Seagates, but if they get stuck, they REALLY get stuck. I mean: they won't twitch when powered... they won't even move with excessive external force applied.

Fixing the PCB with a replacement chip is possible too, and might not need a perfect match of donor and receiver. It should be the same type of chip, though, and even here, closer match is always better. Changing the controller IC is useless because that chip is responsible for most incompatibilities so changing the whole PCB is much easier. Chaging just the spindle motor controller chip (which is a smaller square-like IC that typically read "SMOOTH" on it if the drive is a Seagate) might be a solution. Especially if there's any visible damage to it, it'd be the safest bet.

"I thought there might be something loose inside jamming the platters.....or at least I might see a problem. As it is, everything inside looks ok"

1) there's not much that can become loose and jam the spindle. Heads can literally fall of the arm due to an extreme head crash but they cannot jam the spindle. They are very small. Only bearings can jam a spindle and those cannot be replaced or and likely not unjammed either.

2) you don't see the damage that causes spindle not to operate, whether ist's burnt out motor coils or stuck bearings. You can only notice some damage on the inside only if extreme damage has already occured. That means you can see a head fallen off, top-most platter scratched or dust inside the HDD case (especially in the filter but in extreme case, everywhere) which would indicate that some of the lower platters are scratched even if not visible. There's not much you can do about it.

"Believe me....unless there is a special puller needed, I have enough tools here to take anything apart. And I can make tools if needed."

The trickiest part no doubt is to unload and load the heads without damaging them in the process. After you have removed the top magnet and the limiter, you can pull the heads out of the platter, but you need to make certain they don't touch each other... and they will try touching each other because there's a spring load creating downforce (that is cancelled dut to aerodynamic lift when platters are spinning fast enough). You need something to keep them apart. Each one of them.

Loading heads back to the platters is reversal but at least as tricky. On a drive with one head, platter swap may be "easy". On a drive with two heads and one platter, platter swap is more difficult (because head may touch each other). One drives with 2 or more platters, platter swap is absolutely impossible. (You'd need to swap head stacks. That is the actuator arm and the heads attached.)

I don't think it's a trick that is likely to work for a first-timer. Your donor drive would be as dead as the current drive too.

nick705: "in a multi-platter drive the individual platters are apparently free to spin independently once the spindle cover has been removed"

No. You misunderstood the video. The platters spin independently from the spindle (and worse: of each other) when the "hub cap" on top of the spindle is removed (or even loosened). Just opening the top cover wouldn't affect the spindle because majority of HDDs don't have their spindle motor attached to top cover any more. (Old ball-bearing drives had (and so do certain new 5-platter drives: 7K400, 7K500, 7K1000), but that's just for additional support. They'd still work without the top cover... that is, until contamination kills them.)

I've heard rumours that there is a manufacturer that's difficult to fix after top cover has been opened even once (even in clean room): WD. They have apparently been low-level formatted with the top cover tightened and rely for some extra support of actuator arm (not the spindle!) to remain the condition they were when low-level formatted at the factory. Making a WD work after bolting the top cover on after repair would require fineadjusting actuator arm either through a screw hole by some small screwdriver or by altering the tightness of the screw.

If the data on it is valuable (but not valuable enough to pay thousands) a replacement PCB would be a good thing to try. At least try your luck to find a similar drive. If it doesn't work, return the working PCB back to the working drive and you have a ... working drive. You wouldn't lose anything extra by trying. You do need a new drive after losing that old one (especially considering you can't return that for warranty replacement any longer) so what's there to lose? Just time... if you can keep a poker face, that is.

Good luck.

EDIT1: can => cannot

Das_Saunamies · Post by **Das_Saunamies** » Sat Aug 18, 2007 7:29 pm

Quite a thriller. But if it's seriously such valuable data, why not just pay the professionals? It'll sting, sure, but it's probably the only way.

I'm paranoid so I always have at least three copies of even just my presentations that I'm supposed to do at the university. Optical RW or flash drive, ISP-provided FTP and e-mail. E-mail's obviously not an option with sensitive data, and even encrypted transfer is still less secure than fully offline backup... but with the strongest encryptions the chances of cracking are extremely small, and will require some brutal calculating power. A physical, non-encrypted data storage unit(a flash drive) is much less secure: could get lost, stolen or plain "borrowed". Network transfer cracking takes a lot more effort than the good old sticky fingers routine, and I bet the hospital has some IT security already in place to thwart that.

The upside of encrypted transfer, FTP or otherwise, is that you don't have to carry anything. It all sounds very James Bond, but it's a simple matter, and just about everything of value is encrypted nowadays. A flash drive backup to compliment that and it's a double-safety system, no more data loss. Even if flash drives do have their shortcomings, they're still more rugged than their mechanical counterparts. Some even offer on-the-fly encryption by default, Kingston's selection is worth checking out.

Bluefront · Post by **Bluefront** » Sun Aug 19, 2007 3:08 pm

I want to thank you guys for all the information. I was never able to view the video link, but I will eventually. The good news is the lost data was found, recorded on a CD, but never removed from the drive.

I still will attempt to remove the platters from the Seagate.......just to see if it is physically possible for me with a generic set of small metric tools, no experience with this sort of thing, but with some confidence that hard drives can at least be taken apart, and reassembled. I am not going to acquire another Seagate for a transplant......so I won't know if the data could be saved, assuming the dis-assembly is successful. I was thinking that if I could get the thing spinning, some software data recovery program might be able to get the data.

I still have one more thing to try before taking the drive apart.... I'm going to try to raise the drive input voltage slightly, to see if it might help to start the motor. I have done that with some success on other devices.....notably a set of stuck new fuel injectors. The installation took about six hours, after-which none of the new injectors would even click. As a last resort, I momentarily gave each injector an extra three volts or so....worked. Not the same as this drive problem, but I've got nothing to loose now.

Oh....you'd think my sister could transfer data over the internet direct to her home computer. Nope. The hospital complex requires you to install a special program before you can connect to their LAN.....from within the building. No exceptions are allowed, even for doctors who have worked there for years. This is a piece of unknown software....no explanations are given. Apparently most people use it without problems, none they know about anyway. It will not work with Zone Alarm. My opinion.....this is a piece of spyware, designed to steal data. My sister agrees, and won't use it. Bad deal....

Das_Saunamies · Post by **Das_Saunamies** » Mon Aug 20, 2007 12:25 am

Excellent news! Looks like she's pretty good at it after all.

Good luck with the platter experiment, I for one am interested in hearing how it goes. Never had the chance to take one apart myself. Home-brew full recovery is impossible for precisely the reasons already presented, so no need to go back to the how and why, but it's interesting just to see the internal mechanics, maybe have a description of the problem.

I'd like to hear the name of this mystery software as it intrigues me. If she's using the hospital's workstation she has no admin rights to, I'd say just leave it as it's difficult if not impossible to get that secure, but if it's her own laptop or she has admin rights then it's possible to encrypt and protect everything even prior to connecting anywhere, then just sending the encrypted packets. Main problem with not having admin rights: you can't turn off remote connections and you can't secure your Administrator account, which means someone can connect remotely with full access(paranoia rules). If it's important scientific study of value, then one shouldn't trust even the IT administrators and they should understand that.

There should be no need for helper apps in this day and age, but maybe the hospital is worried about unauthorized access and needs to keep tabs on the machines that connect. That is doable without any installations on the client, but having a mandatory client program makes it a whole lot easier.

whiic · Post by **whiic** » Mon Aug 20, 2007 5:24 am

"I still will attempt to remove the platters from the Seagate....... [clipclip] so I won't know if the data could be saved, assuming the dis-assembly is successful. I was thinking that if I could get the thing spinning, some software data recovery program might be able to get the data."

Removing the platters from spindle assembly and reassembling it has
- 0 percent likelyhood to succeed because there's no markers on the platters on how you should reinstall them. And even if there was markers, you couldn't put them in position accurately enough. We are talking about millionths of an inch here after all.
- has no chance of fixing the spindle motor. Even if you managed to remove the platters, you will not be able to open the spindle motor. You could likely be able to unbolt the motor from the case but you would notice that any attempts to separate the spindle from the stationary part is impossible without destroying the motor. There's no screws. Here's a nice, descriptive picture:
http://common.ziffdavisinternet.com/uti ... 262,00.gif
That picture includes a cut-through of a BB motor and a regular FDB motor used in most modern drives. Notice that only the yellow colored parts are moving. How the hell would you disassemble that? Note that to remove the spindle from cast you'd need to separate those two yellow pieces from each other. I don't think there's any screws but that they are forced into each other using extreme force and precision. Same applies to green and orange part (these are the parts visible from the outside) which use no screws. You would have to open one of these joints to see motor coils.

And I'm not that sure motor itself is the problem. Maybe 10% motor, 90% motor controller chip (on PCB). (Of all the failures, 85% is software (not the case with HDDs not spinning up, obviously), 15% is hardware. Of those 15%, 10% is PCB, 4% head/disk, 1% spindle motor. And since it doesn't spin up, it's probably not head/disk related. 10% versus 1% gives around 90% and 10%... well, not exact: 90.9% vs 9.1% would be more accurate.)

If you insist on having any chance of spinning that drive up again, you shouldn't do anything to the platters. If you on the other hand want to have fun with no chance at all to recover files, open it again and dismantle everything. But you can't combine these two possibilities.

If you want to recover anything inspect the condition of the PCB (both from the upside and the downside). Of course there's little you can do without a donor board (either a complete replacement or something that could donate components (latter doesn't necessarily require a perfect match)). You can of course remove the PCB and clean the contacts on the case and on the PCB to remove any traces of oxidation. But that wouldn't probably solve a drive that doesn't spin up, since those contacts are for head stack.

"just to see if it is physically possible for me with a generic set of small metric tools, no experience with this sort of thing, but with some confidence that hard drives can at least be taken apart, and reassembled."

It can be taken apart. I have taken them apart (for fun, not for recovery). Not damaging them during the process is a bit trickier, though. Reassembling them is the reversal and again: not damaging them in the process makes it quite a trickier. With some care you can assemble it back to a condition that may look visually normal but still will not work due to platters being (microscopically or macroscopically) misaligned. There's no way you can position them correctly.

If you want to learn something "useful", just attempt to remove the head stack intact, not allowing the heads to touch each other, then replace the head stack without damaging the head or media on platters. Replacing the platter is a useless skill to have, unlike replacing the head stack. Even when you actually wanted to replace the platter (only possible with one platter drive), you'd still need to move the head stack out of the way, making head stack removal a common procedure in both head stack replacements (more common) and platter replacements (less common).

Also, if you want to actually learn something, you should make the "repair" procedure in an environment you'd use if it was real DIY recovery, in a "clean room" (=toilet) or "clean box". There's tutorials on how to build a "clean box". Basically it's a plastic box with a transparent top cover and holes where there's two plastic gloves taped through and sealed airtight from the glove and top cover openings. Before each use, the clean box (including the gloves inside) and tools such as screwdrives (which are to be inserted into the box before opening the HDD) has to be cleaned.

The easier alternative is to use a "clean room" is to take a hot shower, blocking all air intakes of the toilet (including the gap below the door you used to enter the bathroom). The idea is to produce moist air and cause condensation. Condensation removes dust from air onto surfaces. Of course the HDD has to have it's top cover on at this point. Read a book while waiting for most of the moistness to condensate or otherwise leave the room (but if it leaves too fast, you should suspect that you have not sealed the room well enough). Take the HDD out of a plastic bag it was in. Open the top cover and make the repairs. If you see water condense on the platter, you're screwed and should have waited longer.

"I am not going to acquire another Seagate for a transplant......"

Yeah, that'd be waste of money since you found the necessary files from a CD. Just take it apart for fun. But do acknowledge that what you are about to do would make even professional recovery absolutely posively irreversibly impossible. So, never do it again with a drive you need to be recovered.

Eadward · Post by **Eadward** » Mon Aug 20, 2007 6:10 pm

As someone said stop dicking around and use a data recovery service. I just used kroll ontrack and looks like I'm getting all data back. Cost a LOT however. An expensive lesson in backup.