I am a high school student, and I basically do simple troubleshooting and attempt to fix things when things go wrong. As my experience is based on real problems, my teacher asked me to attempt to fix this hard disk drive that is considered "dead".

I've fixed at least two "dead" hard drives from my teacher's hard drive collection, and I've used SpinRite 6 to fix them with success.

But this particular one is the strangest of them all, as it can not boot, and it generates a continuous series of two beeps, and the program remains in the BIOS.

I checked what the beeping sound meant in the article Computer POST and beep codes.

To my understanding, the hard drive might have a "parity circuit failure" which sound pretty bad. I plugged this hard drive on the other computers and it appears to be the same problem. I am certain that this hard drive is at fault, as I've tested the other computers with working hard drives, and they had no problems.

My next step to fixing this hard drive is apparently SpinRite 6. After I got Spinrite running, and select level to 2 (for recovery) I went to the screen to select the drive and it displays "DOS A: Undetermined Format". I looked at the details and it said that "SpinRite is unable to read sector zero, which is the first sector of this removable drive".

Ignoring the error, I attempted to try SpinRite anyways, but it couldn't start, which now leaves me without any solutions.

Is this hard drive absolutely dead with no solution in fixing it? Or are there some things I've not tried yet?

If it's absolutely dead, I'm going to take it apart : )

  • ddrescue, spinrite etc are great program when you can communicate with the drive. However reading your post I get the impression that you have a hardware problem. You can not solve that with software. Instead check the drive for damaged copper traces, blown up capacitators etc. – Hennes Dec 8 '13 at 17:51

Your diagnosis of this hard drive is a bit faulty. Two beeps, and your corresponding page which shows errors, points to a parity bit error in the memory, not the hard drive.

In order for SpinRite to operate on the hard drive, it needs to be recognized in the BIOS. Since you said you can't get past the BIOS screen there is no way for SpinRite to access the drive.

Unless there is some sort of data you need off of these drives, or the two that you 'fixed', I would trash them. Once they start getting errors they are likely to get more and more. So even if they are 'fixed' - by marking bad sectors - there is a good chance they will fail in the near future.

If there is something you need off of them, you could try to replace the controller card on the hard drive, which has worked for me in the past (using a working controller from the exact same model). However, if that doesn't get you anywhere, the drive is likely dead.


Just want to point out that SpinRite is usually not the way to go for data recovery. Data recovery is about getting as much data as possible off the drive before it is no longer accessible. Unlike SpinRite, data recovery software usually does not write the data back to the original source drive. SpinRite often trashes the drive in the process of trying to save it.

First use ddrescue (Antonio Diaz's GNU ddrescue, not dd_rescue) to try to create an exact copy of the original drive. The copy is done at block level. Therefor you need a drive that is just as big as the damaged drive (or bigger). First grab the areas free from errors as fast as possible: drescue -n /dev/damageddisk /dev/newdisk log.txt

Next step is to let ddrescue try to recover the problem areas: ddrescue -r 1 /dev/damageddisk /dev/newdisk log.txt

All rescue attempts (usually on a file system level) can now be done to the newdisk. Try something like GetDataBack.


Some thoughts ...

1- Assuming you are correctly interpreting the startup issues to the hard drive, I would consider it dead and do something interesting with it.

2- Since you are using SpinRite (which I HIGHLY recommend), perhaps you could get more technical info from the SpinRite newsgroup. A webportal to it is here. You might find more info there, although digging will be required since there is no search.

3- I would consider Dave Drager's recommendation that you remove the repaired drives from service. If you can't, back them up frequently. As drives age the errors start to come more quickly.


It sounds like you're not horribly attached to the data, but I figured I'd chime in anyway and let you know that even the expensive data recovery companies like Ontrack cannot recover your data if Sector 0 is bad. I sent a drive to them for recovery, which cost me $100 just to have them give me the same diagnosis I had already arrived at on my own: "Sector 0 is faulty; the data cannot be recovered."

Rather than using SpinRite, I'd try to dupe the drive using a Linux live CD and the "dd" utility (or a utility built around it, like dd_rescue). Also, you might want to think twice before using SpinRite as your first recovery option. I'd suggest trying other recovery options first, since SpinRite may do more harm than good on an already-dying drive.

As a last-ditch effort at recovering your data, you might also want to freeze the hard drive and plug it into an external enclosure. It might only run for a few minutes per freezing, but you can try to recover the more important files if you already know where to look.

As the others have already said, a defective hard drive will only degrade more over time. You should remove any defective drives from service permanently, because SpinRite doesn't actually fix physical defects.


While it can be a bit of fun and even be educational to try to revive hard drives, once you get over the novelty value the only time you might do so is to try to recover data that wasn't backed up. After all, it's always that last job someone was working on that is needed yesterday and exists only on the dead drive. Once you recover what you can, if that is even necessary, do yourself a favour and just replace the drive. Once a drive starts to go bad it's not going to get better.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy