I, too, agree that the various manufacturers go through periods of excellent quality and periods of pushing piss-poor products out the door. Unfortunately, the same manufacturer can have multiple products in their portfolio, some of which are great, others of which reliability is a major problem. It is always the case that one poor product will affect a company's reputation for many years after the fact, even after they've cleaned up their act.
Right now the HDD industry pushes out something well over 60 million hard drives each quarter. That means about 250 million hard drives each year. Folks who look at thier own limited experience with the various manufacturers likely do not have enough statstical data to fairly rank the various reliability standards of each company. That said, I do the same things when it comes to PC hardware and other electronics in my life. If a Sony DVD player craps out of me unexpectedly, I'll swear off Sony goods for life (and not just their DVD players!).
I know that each of the various HDD manufacturers have laid eggs in the past few years (my own employer included). I have a pretty good feel for the various products that are known (in the biz) as being real dogs. The fine folks over at
www.storagereview.com have a reliability index (you need to sign up for their site to see it) that tries to rank the various drive families in order of best reliability to worst. Its an imperfect system, but as a gross indicator, it's probably the best thing avialable on the web. I wouldn't let 3% difference between models sway my thoughts on a drive (the survey isn't nearly granular enough), but I'd probably let a 30% difference do so.
Right now the hard drive manufacturers shoot for a 5-year drive lifetime. Our hope is that a hard drive, kept in a favorable operating enviornment, will last on average five years. That means that for every drive that lasts seven years, there'll have to be one that fails after three. It also means that if you operate a drive in a non-favorable enviornment (dirty power supply, not well mounted to your PC rails, elevated temperatures, improper shut-downs, etc.) that your expected life for an average drive will be less than five years. Each and every drives undergoes a significant suite of tests when it comes off the manufacturing line. Roughly 9 of every 10 drives will pass these test and will get shipped out the door. The tenth drive, depending on what the testing says is wrong with it, could be reworked to see if we can make it good enough to pass the tests, or it may be completely scrapped. This tenth drive may only be off by a few percentage points on the important metrics to be determined inappropriate for sale. Conversely, there may be drives that are only marginally better than this drive that get stickers slapped on them (gently!) and sold to the masses. A perfectly viable product line can see the occassional failure 90 days into the life of the drive (which sucks royally for the end customer).
"Funny clicking noises" generally means the servo system that controls the actuator arms and determines the head positioning has failed. The heads don't know where they are, and they flail about uselessly trying to get themselves synched back up with the system. The clicks are likely from the actuator arms banging into the crash stops. I can't speak for all manufacturers (I've only worked for two of the three domestic desktop HDD companies), but my experience shows that there are only certain places on the drives "Oh Fuck Something's Not Right Flow Chart" that will see it get into one of these loops in the code where it decides that it doesn't know where it is and then starts banging on the crash stops. If it clicks like this right away when you give the drive power, then your hosed. But if its an intermittant thing, then you may be able to get your data off of it before its finally all-the-way dead. Time is of the essense, however, as it's extremely unlikely that things will magically get better inside the drive.
You got good advice for your attempts to get data off the drive. Hopefully that website had some software and you had some karma and you've made it happen.
Kasey Chang wrote:There are two main portions of the HD that can fail. The spindle motor... The central spinner that kicks it up to 7200 RPM or whatever RPM your HD uses, and keep it there precisely, and the stepper motor, which moves the read/write heads in precise measurements to follow the tracks.
Kasey, your knowledge of hard drives would seem to be a bit out of date. We haven't used stepper-motors in 10-15 years. Right now the actuation system that positions the heads uses a voice coil motor (VCM) which is analogous to a set of speakers. Instead of linearly driving heads across the surface of a disk, however (which is a stepper motor), the VCM rotates an actuator body on a pivot with an electromagnet to position the heads. My understanding is that stepper motors SUCKED. The VCM system is pretty elegant (but far from perfect).
It's fair to say that the two major mechanical components which could lead to failure are the motor that spins the rotating magnetic media (we just call this the 'spin motor,' or more likely the 'motor') and the VCM-driven actuator. There are lots of subsystems associated with each of these components, and there are other possibilities for hard drive failure than these two large systems. But mechanically, they're the prime culprits.
If you can't get it repaired, try this completely unscientific method... Raise it off the desk by about an inch, and drop it, giving it a bit of a shock (put a bit of cloth or napkins underneath to limit the shock). That sometimes MAY let you get the HD spinning enough to get the data off.
This may help you out if you've got heads stuck on the non-textured zone of the media. Knocking the drive has a very small chance of actually breaking the "stiction" of the heads over your data zone. It's a one-in-a-thousand fix to a 5% failure mode. The clicking that nan describes does not sound to me to be heads parked over data. In that case, dropping the drive like this has a zero percent chance of making anything better.
Holy crap. The animated graphic on that page looks like a drive with a very early version of the VCM system -- but one so convoluted and impractical at its very core that it made me shudder in fear just looking at it. HORRIBLE! That's got to predate my time in the industry by 5-10 years (I started working on hard drives in 1998).
If you're pulling top VCM plates apart (our electromagnetics), be careful you don't pinch yourself. We use *strong* magnets (especially in seek-optimized SCSI drives), and you can end up bloody.
~Neal
That 40 meg HD is still there, I think. 
