Hard Drive Failure *Cries* - Engineers to the rescue?

[Megamox]

Hi All,

Just had my hard drive fail on me out of the blue, no rhyme no reason, not even an early warning symptom. Just died, taking a years worth of data with it. Ouch. Now it sits there, beeping at me and refusing to spin up! Seagate tells me the spindle has failed and they'll happily charge me £1,000 to look at it. Double ouch. I made an incremental back up a while ago and so I do have most of what I've lost but that niggling feeling of forgetting what it is you had...

Just wanted to remind you all to back up if you already havent. I've heard it said unless your data is duplicated in at least 3 places, it doesnt exist. Two backups on site, one off site. Good advice probably.

I've resigned myself to my data not coming back as something inside me is just not willing to hand over that level of payment for my own data. As I've got nothing to lose (literally) with the drive now, I'd welcome any recommendations on what steps you've all taken in the past to rescue date from failing drives, what has worked for you? Also what back up procedures and routines do you all have in place? I obviously need to adjust mine and would welcome the advice. Is backing up to optical media such as DVD and Bluray a good idea? Does it last? How about SSD's? No moving parts to fail right?

The internet seems awash with quite a few stories of whacking the drive (percussive maintenance) to get it going, all the way to opening it up and trying to free the problematic area whilst touching as little as possible to avoid contamination. This is probably the first time in my life I ever wished I owned a clean room. Some people even mention freezing it. Anyone had success with this? I'd love to give this a go if nothing else.

Before I lay it to rest and bury the thing in the back garden, hopefully I can give every possible solution a shot. I've already swapped out the PCB boards in it with another of the same model and same firmware, which didn't work. Next up will be to try the thing in a USB powered caddy to see if It'll work off an independent supply at least and spin up. Failing that, I'll turn the house upside down looking for the one and only torx screwdriver I own and it'll be rubber glove time I'd prefer to keep from opening it as a last resort, as I must admit from what I've read, whether its scaremongering by the data recovery companies or not, I kind of think exposing the thing to air will obviously degrade it if not completely over time, then straight away.

Whatever happens, It'll be a learning experience either way and if any of you have any interesting electrical solutions I'd be happy to give them a try, just for the heck of it! I'm not sure I'm quite willing to run any significant voltage through it though - as fun as that might be. Seagate do tell me from the beeping sound I played for them over the phone, it is either the spindle motor or something called stiction, where the head is stuck to the platter. What if I attached a large signal to the motor to get it going? Or a large signal to the (voice coil I believe) to jolt the head off the platter if it's stuck?

Finally, if any of you skilled steady handed people are in need of a new career, it looks like you could do a lot worse than join the data recovery industry. The lowest quote I've had for an examination (not even to fix it) was £250. Not bad for a few hours work, if you can do it!

Suggestions welcome!

Regards,
Megamox

PS. Have you backed up your files lately?

Edit (6th February): In case you don't feel up to reading the entire thread, I managed to get my data back from the hard drive. It involved making my own tools, swapping out the hard drive head, then swapping out the EEPROM chip. Total cost about £50. Techniques and tools covered in the thread. Good luck!

 

[Mike odom]

back up fanatic. and SSDs do fail... so the 3x rule still applies, even if using SSDs.

 

[cowboybob]

Had a "stuck" Seagate a few years ago that wouldn't spin up. Took it out of the computer (there should be enough cabling to reach) and with it flat on the desk and with the POWER ON, rather viotently twisted the drive back and forth, in a circular motion, a couple of times.

The idea was to use the inertia of the platters (what little there is) to overcome the "friction" of the spindle shaft in its bushing.

It spun up. And then the computer booted up. At that point, obviously, I did a back-up.

Actually did the twisty startup for a few months before I finally got a new one.

As an aside, stiction (of the head to platter type) hasn't really been a problem with hard drives for a very long time. When it was a problem, it almost always resulted in a true head(s) "crash", which destroyed both the heads AND the platters.

 

[unclejed613]

other common "won't spin up" failures are the TVS diodes shorted. remove the TVS diodes, and the drive might work, but plan on backing it up and getting a new one immediately. i have replaced the TVS diodes before and got the drive to work reliably, but i wouldn't run a drive without the diodes for any time longer than it takes to back them up. you may have noise in the power supply that caused the diodes to fail. so it might be prudent to also get a new power supply. usually shorted TVS diodes will keep the power supply from starting.

i remember those old seagate 40 meg IDE drives (the early ones actually were MFM drives with an added IDE controller) and the stiction problems.... one good twist and they would start up, and would continue working unless you turned the computer off long enough for the air to bleed out between the head and the platter.

another place i have experienced an extreme form of stiction was when i was a calibrator in the Army. we had dimensional standards for calibrating micrometers, etc. they were blocks with very precise lengths, and polished optically flat. you had to "wring" them apart, and if you needed to put two together to get a larger dimension, you had to "wring" them together so there was no air trapped between them. you could not pull them apart, because there essentially was a vacuum formed between them, and since they were optically flat, there was no pathway for air to get in to fill the void when you tried to pull them apart. so, yes, there are some very large forces that cause stiction, and the easiest way to overcome it is by shearing the two surfaces, not pulling them.

edit: for a similar reason, never store micrometers closed. not only will you have stiction (usually if they are relatively new), but if they absorb any moisture into the gap (usually if they've been used enough that the ends are no longer flat enough for stiction to occur), they can rust closed, and you will A) never get them open, or B) if you get them open by operating the screw handle, you will damage the threads, and make the micrometer inaccurate

 

[Megamox]

Thanks Guys! It's funny when I told the IT administrator at work about this he laughed and said 'I've only ever owned 3 hard drives - I just replace each one when it dies'. Had to laugh at his humour. Well the 3x suggestion looks like it'll be the plan for the future then.

I like the idea of spinning/twisting the drive to see if that gets things going while it's on power. I've had a few troubles with the PSU of the laptop. It looks like this might be a contributing factor if not THE contributing factor then. I just assumed if there was a voltage spike, it'd affect more than just one hard drive in the laptop. Everything else seems fine. I've ordered a new PSU just to be safe anyway, better spend the money on this than a whole new laptop I say.

I'm liking the idea about the TVS diodes, I'm not sure where they are though or how many of them to look for. If anyone can point me in the right direction I'll get the multimeter out and check for a short. Hopefully if this is the issue, the TVS's stepping in to divert the spike and data is still salvagable maybe I can snip them off and give operation data rescue another go.

Thanks,
Megamox

 

[kubeek]

I´d say the TVS is that black thing on the top right, but you will be better off looking up the pinout of the power connector and checking resistances there. Also check that white part close to the large hole, I think it is a fuse.

 

[()blivion]

In the future, I recommend getting several identical drives and do RAID1 or RAID5 or RAID10 for Data striping + mirroring (or parity). This also has the happy side effect of substantially increasing performance. As for SSD's, IMO they are currently not worth the price, and also have problems with failure.

 

[Megamox]

Thanks guys. I assume for a fuse, I'd just short it? The top right component seems to have the letters KEX on it. At least that's what I think without a magnifying glass! It's funny you mention RAID because when I bought the laptop it came with a RAID configuration as it has two hard drive bays. I cant remember which set up it's called, but it's where the same data is identical on both hard drives at the same time? Anyway you know how it goes, one day you need a bit more data and foolishly you decide to split the drives up and then.. well here we are. Luckily, or so I thought, because the laptop came with two identical hard drives, same model, same firmware revision, I could just swap out the logic board PCBS and get this one working again, but that didnt work. The good board in the bad drive - still not recognised and still beeping. For fun,the bad board in the good drive - the drive didnt make any sounds but it wasnt recognised any longer. Put the good board back in the good drive and it worked fine again. So obviously the drive that's broken has an issue with the board for sure. I cant say for sure whether the mechanics of the platter and head are okay or not too. I understand that even a similar board in another hard drive sometimes doesnt work even if the firmware is a good match, something about some adaptive memory on the original PCB that must be copied over or something. But I'm in no way an expert, but I feel like I've read enough to at least realise how fragile my data has been for the last two years.

Megamox

 

[Timescope]

Do a Google search for DIY HDD repair. You may be able to open it up and get it working long enough to retrieve your data.

Timescope

 

[MrAl]

Hi there,


You should always make sure to follow the three golden rules of storing your data on a computer:
1. Backup.
2. Backup.
3. Backup.

But i see from this thread that you've learned your lesson now

Mine came when i was working on a hard drive controller software routine. It was silly of me to do that on my main computer but who knew it would take out data. Lost a few programs that at the time where quite large and backup storage media wasnt as abundant as today.

Today you have lots of options. CD or DVD is supposed to be the most durable over the years. I dont know about Blue Ray though. But if you dont have hundreds of gigabytes to back up then CD's or DVD's are a very reliable option, provided you use a secondary program to test the data immediately after burning. That's also a very good idea because some backup programs 'miss' some data even if they have a verify function.

Also popular today are USB thumb drives. You can get 64 gigabyte thumb drives today for 35 dollars (USD) so it has become a economical option for backing up. It's easier than DVD but not quite faster if you get the low end 64GB drives. The cheaper ones (35 dollar range) are much slower, and take much longer to store data, but once it is there you can read it fast enough to still make this a good option for backing up. Large files generally transfer much faster than a bunch of small files. So if you have a huge number of small files expect longer storage backup times. But then if you use a program you can just let it run.

And also once the data is stored, no matter how it is stored or what media or what program is used, you should also use a program that is made for comparing the files on the backup to the files on the computer. This tells you if the backup software missed anything or wrote a bad file that the backup software verify did not catch. Strange as this sounds, it happens. If you write software yourself you can write your own little program to check the files. It's an extra step but it's worth it.

Once in a while it also helps to check the data integrity of the backup by comparing it to the main data source again. You might think about what frequency you'd like to do this, once a month, once a year, etc. Data can change without intervention on a storage media so 5 years from now when you go to get your data part of it might be corrupt simply because a bit or two flipped states. This is also where the 3x data backup storage comes in handy, because it is unlikely that 3 backups that have random corrupt files will all have the same corrupt files, so at least one of them should have a good copy.

Part of your choice of backup media is how much data you have to store. If you only have a few gigabytes then you might go with several USB thumb drives, copying all files to *all* the drives so you have several copies.

Good luck with it

P.S. If you open up the hard drive dont get any dust inside of it or it could crash just because of that.

 

[Sangoma]

I have managed to get data off several non functioning drives by cooking them in the oven for an hour at 150C.
Before you bin it or break it, (ie the very last resort) it may be worth a try.

It is almost certain that opening it will result in it not working again unless you have the right environment and skills.

 

[unclejed613]

DO NOT OPEN THE DRIVE! even 1 microscopic speck of dust will make the surface unreadable if it gets caught under the head.

it doesn't look like there's a TVS diode on the board, and i'm assuming you took a picture of that side of the board because there was nothing but traces on the other side. the power connector is the 4-pin connector in the upper left corner, and USUALLY, the TVS diodes are within an inch or two of the connector. sounds like the drive is mechanically dead (maybe from stiction, or a seized bearing) or bad motor windings. btw, i know people who have had Seagate drives last a long time, but Seagate drives have a bad reputation for reliability. Hitachi and Fujitsu drives seem to have very good reliability. Seagate, Hitachi and Fujitsu all have 5 year warranties on their drives (but that won't help you get your data back...)

 

[ronv]

Yep, the beep beep is the firmware vibrating the spindle because sometime the heads will stick and it won't spin up.
But, that doesn't happen much now days because the disks have little laser bumps to reduce the surface area. The guy is probaly right and the spindle froze up. Once your done trying everything else you could open it and see if the spindle will move.

 

[Timescope]

You could also try cleaning the pads at the bottom of the circuit board that connect the heads etc to the circuit. They look tarnished.

Timescope

 

[MrAl]

Hi,


You could also try opening up the case and cleaning the heads with alcohol and a Q tip

Just kidding of course, but on the more serious side the SSD drives are not too bad at all. Their advantage is shock resistance so you dont get a crash just because the computer case got bumped hard or lots of vibration. Also less power demand from the power supply or battery. These two features are especially good for laptops.

 

[ElectroMaster]

Seems to be the month for it!

The solution I have come up with for this site is to synchronise the data to the cloud with the company ibackup.com. They have a really good special on and it's easy to use. They have a windows application that automates the process. The great thing is that once the initial backup is complete it only uploads files that have changed. They also keep 30 revisions of files.

Best of luck!

EM

 

[MrAl]

Seems to be the month for it!

The solution I have come up with for this site is to synchronise the data to the cloud with the company ibackup.com. They have a really good special on and it's easy to use. They have a windows application that automates the process. The great thing is that once the initial backup is complete it only uploads files that have changed. They also keep 30 revisions of files.

Best of luck!

EM

Hi EM,

Wow that's really good news. So next time there is a crash 99.99 percent of the data will be recoverable right? That's great. I didnt realize how important the images are in a thread until i try to read one that 'lost' some of it's images. The thread almost makes no sense when that happens.

 

[olly_k]

I have just had my second Samsung hard drive fail on me after about 2 months - the first was a 3.5" in my main PC and this one was a 2.5" in my laptop :-( Not doing Samsung again!!! Both times they stopped spinning without warning, the lappy example gets quite warm so this is clearly a seized spindle or similar... Now I can't find the bloody receipt arghhhh!!!!

So, I just wanted to drop a tip. If you do decide to pull the top off, there is a method you can use to minimise dust. Do it in the bathroom with a cold shower running! The idea being that the water catches any dust that is airborn. Still not perfect solution and you must pay attention to static but if I go down this route that is where you will find me

 

[geko]

+1 for RAID. Good for protecting against disk failures - just make sure you have the necessary monitoring/alerting in place to know when a disk has failed in the RAID set. Doesn't have to a cost a lot either - I picked up a VIA PCI RAID card for about five quid for my PC at home - mirrored disks - did it's job when a disk failed in that.

Backups are essential for recovering corrupt files, deleted files or files overwritten with old data etc. as well as hardware failures. Also for finding students who claim their assignment work has been deleted from the network but actually the previous six months of backups show they haven't done any work

At work all our servers, disk arrays and SANs have RAID of one level or another and we've never lost data due to a disk failure in 16 years. However, we have had to recover from a backup when a disk array controller failed so even with RAID backups are essential.

As others have mentioned, check you can restore from the backups regularly.

Back in the late eighties/early nineties I was a support engineer and customers with backups that wouldn't restore and RAID protected disk arrays that lost everything because they didn't know a disk had failed until another one in the RAID set also failed was quite common. Both these issues usually came to light at the same time

 

[kubeek]

the power connector is the 4-pin connector in the upper left corner...

No that´s not it, this is a SATA drive, so the power connector is the wide on the right side. The one on the left is used by the manufacturer presumably. On a dead toshiba drive I have here one pin from the left connecotr is ground and the three remaining are some data pins.