Resurrecting Dead Harddrives?

Broue Master asks: "The main harddrive of a friend's computer stopped working. He described to me that the computer began by emitting strange 'scratching sounds', and after a while, it made a 'loud *tock* sound' and stopped. He tried to reboot it but soon realized that the harddrive wasn't spinning anymore. He asked me if I could revive it, at least long enough so that he could retrieve at least his "my documents" folder. The computer was running XP. I did a little googling(tm) of my own to find out that the most recommended solution out there seems to be 'freezing' the harddrive for a day in a ziplock bag. I'd like to know what fellow Slashdot readers have done in the past to try and resurrect dead harddrives and if the freezing method would still be a good idea, today. The harddrive is a Samsung 30Gb." A good 95% of the time, once an HD is gone, break out the shovel, because it's time to bury it. Still, it would be interesting to note, if only from an anecdotal standpoint, if any of you have managed to perform such miracle hardware resurrections. Have you managed to revive a dead and decaying drive from the dead long enough to pull data off of it? If so, what did you do?

Comments

Freeze first, then

[SpaceLifeForm]

try finding an identical drive and swap the electronics.

Re:Freeze first, then

[DA-MAN]

Cryogenically freeze it now and maybe in thirty or fourty years it will be cheap to take it to a data recovery company, cure what ails it and what not.

Re:Freeze first, then

[XenonOfArcticus]

I've done the brain swap successfully before (about 10 years ago). I have heard anecdotally since then that this may not work with more modern drives that may store calibration data in non-volatile storage on the electronics board. I can't confirm if this is true.

I've never tried freezing one. I think I'd try a brain swap first, as it's unlikely to cause physical harm. I can't say the same for the freeze operation for sure. If you do freeze it, put a bunch of dessicant (silica gels) in with the drive for a few days beforehand. You don't want moisture in with the drive freezing, expanding and damaging something.

Re:Freeze first, then

[twilightzero]

I can say authoritatively that swapping boards on newer drives does NOT work - I used to do tech support for Western Digital. The problem becomes that a chip on the board holds the other half of the calibration data that's embedded on the platter at factory low level format time. So in a sense every drive is individually tuned to a point where only that controller card will work with that body. I've heard about it being done successfully with drives having the same model, sub model, firmware rev, very close s/n's, etc. I wouldn't reccommend it at all. Also if you brain swap the drive, you lose all warranty on both the drives (yes they can tell). I agree with the first poster, get a shovel and bury it because it's dead or send it out for data recovery. If the data means ANYTHING to you then don't try home methods, they're liable to get you into a far worse situation drive-wise.

Re:Freeze first, then

[technos]

Unfortunatly you're wrong. I've done board swaps on at least two dozen drives in the last five years, mostly antique Seagate Barracuda (swapped at least 8 of those I can think of off hand), but the line up included everything from a Micropolis full-height MFM through a slew of more modern 2-20Gb IDE drives..

Success rate? 100% on the Barracudas, 100% on the MFM and RLL volumes, probably only 60% on the rest. A lot of the drives will not tolerate a logic board swap, but its always worth a shot if you're not going to warranty RMA the drive. (One of the successes was a recent vintage Western Digital 20G drive, which is why I was compelled to respond)

Of course, I've also de-stuck probably 50 drives with the old "power it up, tap it against the desk during POST" trick.. That nearly always works when it won't spin up.

PLEASE NOTE:

My success rate is tempered by having 15 years of experience. My first recovery efforts were on Seagate 10 meg drives WHEN THEY WERE NEW. Fully 20-30% of the drives I come across are UNRECOVERABLE by any means you or I can do. Send the damn drive to Ontrack if you value your data.

Mine is dead right now

[whodunnit]

Heh,

My main drive died yesterday, turned on the computer.. wirrr, click.... wirrr, click. so i took it out of the computer and turned it upside down, gave it a little shake, and walah it spun up. I've managed to back up my data off of it now, and hopefully it'll last till the replacement drive shows up tommarow.

Re:Mine is dead right now

[angst_ridden_hipster]

Perhaps he meant "wallah" or "walla," the Hindi suffix for "one who does" or "one who deals with." Its' a good suffix for creating descriptive terms: the taxiwallah is a synonym for "cabbie," the rickshawwallah is the poor guy who carries the rickshaw, while the datawallah is the guy who flips bits for you when you write code.

Perhaps the grandparent reference is to the "platterwallah," who spins up hard drives for the upper castes. Incidentally, while they may sound similar, the platterwallahs don't like being confused for dishwallahs, as they don't clean dishes.

Drop trick

[HalfFlat]

Back in the days of 1GB SCSI drives ...

Sometimes they'd get 'stuck' if they were left on for a long time (like a year or two), then turned off. At this point they wouldn't spin up, or make a half-hearted attempt to.

If they couldn't be coaxed into moving, taking it out the enclosure and letting it drop four centimetres or so flat onto a wooden table often got them unstuck enough to grab the data and back it up.

That said, have had some success with the same trick with newer drives with different modes of failure. Of course try the least invasive approaches first and work up, but if the drive is otherwise dead, then there's little left to lose. Unless you want to spend big dollars with a professional data recovery mob.

Re:Drop trick

[innosent]

True, if the data is important, don't touch it, send it to the professionals. If it's just your porn collection, break out the sledgehammer, at best, you'll get your data back, at worst, and most likely, you'll have fun hitting a drive with a hammer.

Re:Drop trick

[Molina the Bofh]

True, if the data is important, don't touch it, send it to the professionals. If it's just your porn collection, break out the sledgehammer

Who said my porn collection is not important, you insensitive clod ???!!?!!

Re:Drop trick

[Oinos]

Remember kids, if you're going to drop your hard drive, drop it on it's SIDE. Dropping it flat on it's top or bottom is only going to embed the heads into the platters.

if it's the Circuit board

[jaredmauch]

(Which i've seen before), you find an identical drive and swap it. Now the problem i've seen is that the PCB's have gotten more and more reliable over the years. Most drive failures I've seen recently are all physical failures inside the sealed environment. Depending on how the drive stopped working pray for no physical damage.

hard drives are so cheap these days it might be worthwhile to do a daily rsync to help save your data. This is what I do, rsync/tar over to another system for my backups. It's nice to have a backup copy on spinning media nearby.

Give it a whirl

[eakerin]

I had to get data off a dead drives a few times before. The drives didn't make much noise, except for not spinning up.

To bandaid it I un-screwed the cover, and gave the platters a quick spin (make sure to only touch the SIDE of the platter, not the surface).

I put the drive back in the PC, and it started up just fine. I then quickly copied the most important data off of that drive, and then made a copy of the entire drive to another known good one.

Re:Give it a whirl

[narratorDan]

The first time I did this trick my friend's HD had stopped working, it never spun up and the heads never moved not even a *click-click*. After checking the drive with an o-scope it was determined to be a hardware issue rather than a PCB issue. So we tried dropping it on it's side to dislodge the heads but it didn't work. So as a last ditch effort I suggested that we open it up and manually free the heads or spin the platters. Since I had disassembled many dead HDs for the nice magnets on the head arms it was decided that I would be the one to do it. Once I got it open we discovered the heads were tightly wedged up onto the platter spindle! The head arm had somehow passed the stop and gotten stuck. Gently twisting the platters (put a screw into axis of the spindle and twisting it with a screwdriver) while pushing the arm back away from the spindle dislodged it. After checking to make sure that all parts were moving freely and that everything was secure we powered it up and bingo! Kinda neat watching a HD operate. Put the lid back on and he backed up all his data and eventually used that drive for another two years problem free! Pretty amazing since the whole operation took place on his bedroom floor not a clean room in sight.

NarratorDan

Dead eh.

[GigsVT]

Well if it's dead, and the data isn't very important, time for drastic measures.

It sounds like a mechnical problem, a head crash that went as far as pulling the head right off the arm, and wedging it between the platters, or the bearings died.

Freezing it won't hurt anything, but beware condensation when you take it out of the freezer.

I'd recommend getting in the cleanest room you can, preferably with high humidity to reduce dust.

Take off the cover and look for metal shavings in the inside. If there are no metal shavings visible, then the bearings have gone, and you might as well give up.

If there are metal shavings, then there's a chance you can recover data. Try to move the platters/heads, if the heads aren't in the landing zone, then this WILL damage the platters (a little).

If the top head is the bad one, you are lucky. Try to rig it in a way so the platters can turn again. If it's a head between other platters, you are going to have to pull the platters out, which is not easy at all.

Anyway, if the data is worth more than $1000 to you, then send it to a professional recovery service. I don't think an electronics board swap will help anything in your case. It would only waste time and money.

Doing any and all of the above things may damage the disk more than it already is. You've been warned.

Tick Tick Tick BOOOM Dead

[Unholy_Kingfish]

Once you get the ticking, clicking, and screeching sounds you might be out of luck. Mechanical failures tend to destroy the actual data on the platters. Heads crash, bearing explode or shred leaving metal shaving on the platters and cause physical damage. Some recovery is possible, but rarely total recovery.

Most big data recovery operations keep new units of every type and model of hard drives around for recovery. If you had an armature go, they pull your platters out of your drive, and put them in EXACTLY the same make and model HD in their clean room and try to recover the data with standard software. You could try and do this yourself but buying EXACTLY the same make model and series of drive. But opening up one of these units while not in a clean room is not a good idea.

For sector damage, non physical damage, there are tons of tools like this and this out there that might help. But sometimes the damage to the MBR and backup MBR are so bad that recovery tools might not be able to make sense of the bits. I have one sitting right here that is like that. Somehow the bits got shredded over the ENTIRE disk. I assume there was a physical malfunction that dragged the head across the platters and made Swiss cheese of it.

Identical drive and swap platters.

[Wardish]

From the sound of it it was a nasty head crash.

If the value is high enough send it to a qualified recovery company. If your willing to risk it and you have the tools, swap the platters from the bad drive to an identical known good drive.

Odds of getting it running with cold or hot is low considering the reported noises.

Qualified recovery company figure 100% they get data and probably about 90% of it. Odds of switching platters yourself and getting most of your data figure 60%, odds of using cold (freezer) or heat (it can work...) 30% or so.

BTW if you do the freezer make sure and bag it. You don't want a lot of nice humid air on your drive when it's nice and chilly.

Now back to my Thorazine...mmmm thooraaaszzzzzhhhh....

Try this.

[Tooxs]

Take the drive out and have cables long enough to set the drive flat on a smooth non conductive surface, like a bench or table top. Turn the system on and give the drive a quick spin about a quarter turn around the axis of the platter, then listen to see if it spins up. If it was just a sticky spin motor it might let go, if it does, try and get your stuff asap. I've recovered data on a few drives like that. I've had a couple were this worked once but not twice, so you shouldn't press your luck. It's a long shot, but it gives quick easy results if it works.

Re:Try this.

[DaoudaW]

I had a laptop which had been too long in a monsoon climate. I opened the case and took out the drive and gave it the jerky 90 degree turn a few times, quickly plugged it back in and voila it spun up. Not having a replacement drive I just kept using it. After opening the case several times I realized that I could just give the laptop the same spin and the drive would spin up. Eventually I opened the foil cover on the drive case and stuck the drive in a dry place (a ziplock bag with several fresh silica gel packets) for a few days. Last time I tried it was still working.

DriveSavers

[gabe]

www.drivesavers.com

Re:DriveSavers

[afidel]

Very good (98+% recovery from a fleet of dead Hitachi drives in one departments Thinkpad laptops) but NOT cheap. I think the typical bill was a couple thousand dollars to have couple day service with DVD's shipped back to us.

don't freeze it...

[theIG]

a friend of mine said that he has saved the data of many people in the same situation by leaving the HD in the backseat of his car while he is at work on a sunny day. When he gets home, sometimes the heat will let it spin up just long enough to pry away any salvagable data. I know for sure that it has worked for him at least 3 times.
-kyle

Re:Unlikely

[alienw]

Wrong on both counts. Cooling it and then reheating it back to room temperature causes small mechanical movements which could make the mechanism work a little longer (though it is not very likely).

The electronics, of all things, won't give a flying crap -- every IC I have seen, including delicate DSPs and such, is rated to at least -40 degrees Celsius (datasheet for a TI DSP, commonly used in HDDs, look at page 130). Unless you immerse it in liquid nitrogen, it won't be a problem. I'd be more worried about water condensing on internal surfaces and such.

Freezer Believer

[candl]

I had a dead drive that I kept around for a year just in case some miracle recovery cure came up. Sure enough I heard about the freezing thing and six months ago broke out the HD, stuck it in a plastic bag (in hopes of tricking any condensation into adhering to the other surface) and threw it in the freezer while I went and "prepped my system" for getting a drive in and fired up asap.

On the first attempt I had put the drive in the freezer for only about 20 minutes (hey, I was anxious) and it fired up. I browsed to the drive's directory before it crapped out on me, but I knew I was close. So the drive went back in the freezer while I thought of plan B.

On the Second attempt I left the drive chill for 45 minutes and this time when I took it out, I brought it out on a frozen gel pack. I was able to get all my data w/o condensation or other complications.

I'd like to say that my data was fine, but somewhere along the line my poetry and stories really started to really suck. It must have been corrupted at some point because I swear it was better when I wrote it. Hopefully your data will retain it's value...

Re:Hard drive identity problems

[cybotix]

remember you learned about thermal expansion in high school. the more you heat it, the more gigs you have on that drive.

Re:Hard drive identity problems

[Bananatron]

So you mean if I pop my HD in the nuker I'll have 80 gigs instead of 60? Off to the kitchen I go!

it's dead jim...

[joelja]

The sound described is highly indicative of either spindle seizing or a head crash. neither is something you're likely to recover from.... it's possible if the spindle has seized that a strong lateral motion when it's powered on might break the disk free and allow it to spin up again.

So you can power it back on and clonk it into something but I wouldn't hold out a whole lot of hope.

A comment on freezing.

[stvangel]

This can be a (very) temporary fix with a drive that's having problems with the electronics. Often if components are flaking out but haven't actually fried, they'll run when they're cold but die when they heat up. Get it cold and then power it up and work fast. You'll probably only have a few minutes at a time though. The same applies to motherboards, chips, and memory. To give you more time, you should probably set everything to as low speed and low voltage as you can get away with. I actually did this with a machine outdoors in 15 degree weather once. The machine had been crashing during boot and I couldn't get another machine to recognize the drive's data format ( it was a strange integrated controller on the motherboard ). Outside it booted and ran for two hours while I copied all the data over a long ethernet cable I'd ran out a window. Turned out to be the motherboard. After a replacement with something a little more generic and a reformat, I copied everything back to the drive and it was fine.

Granted this probably has nothing to do with your current drive problems. It sounds like it blew chunks with physical problems. Even if you could get it working again I'd bet you've got significant platter and/or head damage and any data you could get off it would have serious corruption issues. Scratching noises and loud thumps coming from hard drives are never good things.

fiery re-entry

[Anonymous Coward]

no, no: "voila" is the *metric* spelling. in america it is "walah".

it's people like you who cause spacecraft to fail.

Re:stiction

[richie2000]

This gets the heads sliding across the platters to overcome the static friction

Um, no. The heads never touch the platters. The stiction comes from the ball bearing grease solidifying over time, not heads sticking to the platters. Newer drives have female bearings (they're full of fluid with no balls) and don't have this problem.

External HD copier

[Bushcat]

I use an external hardware-based HD copier made by Century. It's OEM'd to someone in the US but I don't know who. Basically, plug any drive into one side, plug any into the other, push a button and it rapidly clones one to the other. It also has a USB2 connection. When connected to a PC, it can be used to mirror an internal drive, clone it or simply be an external drive.

Since it's standalone, it can clone a non-bootable drive. It also seems to be able to clone drives that are too damaged to spin up in a PC.

Recent rescues:

60GB 2.5-inch drive would spin up intermittently. Attached it to the external box, where it had the same problem. So I removed the lid, and got the drive to spin up with a thumb twist on the central boss. I got the drive cloned in 20 minutes, and the drive continued to work for another 40 minutes.

Fujitsu with the (in)famous circuit board problem: Got a replacement drive. Cloned an identical functional drive from another machine in the office onto it. Swapped the circuit board on the functional drive to the non-functional one. Drive started, so cloned it to the original functional drive.

The Century unit has been worth its weight in gold to me over the years. The newest one is smaller & lighter. Around $150.

Stuck bearings.... the cure

[SomethingOrOther]

If an old drive has been running for a long time and is switched of, we've known the bearings to stick (drive wont spin up at all)

The fix is to hold the drive between the palms of your hands (like praying... good analogy :-)) with the axis of the drive between the heals of your hands. Then violently flick your wrists downwards untill your fingertips point to the ground.

The idea is to spin the hard drive casing whilst causing the platters to stay still, so giving the bearings plenty of torque to free them off. It would often be enough to get this thing going again long enough to get the data off it.

But in this case, it sounds like drive mechanics.
Yes this does sound like heads hitting the disk. I've known people physicsly open the drive and poke at the heads (they had skipped off the disk somehow) and get it going long enough to recover data.

Anyway, if the dude has important stuff on this disk I suggest he takes it to a profesional companny who knows what they are doing. You do have backups, dont you?