Defective hard drive: Rock me gently

Posted by Pierre Igot in: Macintosh
February 10th, 2008 • 11:30 am

This has to be the strangest hard drive rescue operation that I have ever experienced. Since it was ultimately successful, I thought it’d be worth sharing with Betalogue readers.

A week ago, my sister-in-law in Ottawa phoned me to tell me that her three-year-old PowerBook G4 15″ was acting up, with freezes, random shut downs, and weird noises coming from the internals. It didn’t sound too good. The problems were somewhat intermittent, because she was sometimes able to boot normally and run her system for a while, but ultimately she would experience one form of failure or another. At times the laptop did not even recognize the internal hard drive as a bootable volume, so the prime suspect had to be the hard drive.

I sent her a copy of DiskWarrior to try and boot from the utility’s CD and repair the hard drive. The laptop booted from the CD, but DiskWarrior could not even see the internal hard drive. We let it rest for 24 hours and the next day, DiskWarrior was able to see the drive and my sister-in-law was able initiate the directory repair, but the process was soon stuck.

There was clearly little hope for that hard drive. Since she had to travel to Nova Scotia this week anyway, I told her to bring the machine and I would have a look in order to confirm the diagnostic and possibly rescue what could be rescued. She had backed up a certain amount of stuff in recent weeks so she was not about to lose everything, but, like most people, there were still things for which she did not have the freshest, most recent backup.

She had also resigned herself to the prospect of having to buy a new machine, so the issue was mostly to try and determine whether any part of this PowerBook G4 could be salvaged.

When I first turned the laptop on, everything appeared to be working fine. The system booted instantly, and I couldn’t hear any weird noises coming from the machine. But I had already established, through her descriptions over the phone, that this was mostly a matter of minutes, before the machine had a chance to warm up.

I took advantage of the fact that the laptop was running to write down a few things and then load the original system CD. I booted from the CD to run the hardware test. The RAM and the logic board passed with flying colours, but there was a cryptic error for the “mass storage,” i.e. the internal hard drive, which confirmed that this was where the problem was located (although I was also able to verify that her power adapter was seriously defective, which an on-and-off power supply coming to the laptop, depending on how I was bending the cable coming out of the power adapter, and with ominous electrical noises coming from inside the power adapter).

I ran the hardware test a second time and this time the “mass storage” passed as well as the rest. But I suspected that it wouldn’t last long and, sure enough, a few minutes later, the laptop started making all kinds of humming and whining noises, and I was no longer able to boot successfully.

I then started the laptop in target disk mode and connected it as a FireWire volume to my Mac Pro. I was able to mount the hard drive on my desktop. I started copying her files. I was able to get through one or two gigabytes of data, and then the copying stalled, with the noises coming from the laptop intensifying.

I turned everything off and let it cool down for ten minutes. I booted in target disk mode again and was able mount the volume and continue to copy data, but again the process stalled after a few minutes. I went through this cycle a dozen times over the next few hours, and in the process was able to rescue her entire “Documents” and “Library” folders, i.e. all her Office documents, all her e-mail, address book entries, etc. Not bad.

But I still had to try and attempt to rescue her iPhoto library, which contained about 10 gigabytes’ worth of pictures. At this rate, it was going to be painful indeed. She had a number of pictures on various CDs, so she was not totally desperate to rescue the entire thing, but I still figured I would give it a try, since there were some pictures she did not have copies of.

What was worrying, however, was that the noises coming from the laptop were getting ever more ominous, even after the laptop had rested for a while. And the periods during which I was still able to copy stuff where getting shorter and shorter, and sometimes I wouldn’t get anything at all. The volume wouldn’t even mount.

At that stage, I decided that it might be worth opening the machine up and extracting the hard drive. I have one of those USB 2.0 Universal Drive Adapter things, which lets you connect any PATA or SATA hard drive without an enclosure to a USB 2.0 port. I figure that, by extracting the hard drive, I would be able to verify that the noises were indeed coming from that particular device, and I would also have to spend less time trying to boot the laptop in target disk mode. I also had my sister-in-law’s blessing, since she said she would be grateful for anything I might be able to rescue.

Now, if you know anything about repairing PowerBook G4 laptops, especially the Aluminum generation, you know that replacing the internal hard drive is not the most pleasant experience, even for Mac tech support people like me, who have some experience dealing with the internals of Mac computers. It involves removing a large number of screws of various sizes from various places, and prying open a laptop with a tight design where every millimeter counts.

But I had some previous experience (described here) with replacing the internal hard drive of a PowerBook G4 12″, and I also found a number of on-line guides for the process.

It all started pretty well. The first big hurdle was to pry open the upper case, which involves accessing some pretty hard-to-see metal tabs and pulling the upper case off without breaking anything, and also preferably without cutting oneself on the sharp metal edges of the case in the process. After some fiddling and with some moral support from my wife, I managed to get it open, without cutting myself, much to my relief.

But then there was still the small matter of dislodging the hard drive itself. From my experience with the PowerBook G4 12″, I remembered both the finger cut on the aluminum edges and the very frustrating process of trying to unscrew the final two screws holding the hard drive in place. All the other screws had come off easily, but for some reason these two screws were screwed in incredibly tightly.

I had a lot of difficulty with those two back then, and sure enough I had a lot of difficulty with these two here again. The two screws were incredibly tight again, and my small screw driver did not have enough grip on the handle for me to apply enough force. I ended up using a pair of players to hold the handle of the screw driver and, with a combination of strong force from both hands, I was finally able to unscrew one of them.

But the other one simply wouldn’t unscrew. Worse still, as in the case of the PowerBook G4 12″, I ended up working so hard trying to unscrew it that the groove in the screw head became worn and the screw became effectively impossible to unscrew.

With the PowerBook G4 12″, I got lucky, because even though I could no longer unscrew the screw with no groove, I was able to loosen it by removing a metal spacer underneath it. No such luck in this case. The screw was not on an elevated plate. It was level with the other one, and held a metal bar tight above the two side screws of the hard drive itself, holding it in place.

I am afraid the only solution I had at that stage was to break the metal bar, which I did. I could lift it from the side of the screw that I had managed to remove, and then I just forced it out, breaking it around the edges of the screw that wouldn’t move.

The hard drive was finally out. I connected it with my USB 2.0 adapter and, sure enough, all the horrible noises were back, which confirmed that this was the defective part. I could now clearly tell that there was some pretty ominous rubbing of parts happening inside, which explained the humming and whining. It also told me that this hard drive did not have many spins left in it.

I was able to mount it on the desktop and copy a few pictures, but then the process hung again, with the noises from the hard drive intensifying as it was getting warmer.

That’s when I had a small flash of inspiration and, instead of simply unplugging the hard drive and waiting a while, I tried moving it around in the air. It was rather counterintuitive, because theoretically you are definitely not supposed to move a hard drive in all kinds of directions while it’s spinning, but for some reason, I saw that this was helping, by enabling the process of copying the files to continue instead of stalling.

Indeed, I soon realized that I was able to get an optimal performance from the defective hard drive by gently rocking it from side to side. This enabled me to copy up to a gigabyte of picture files in one go, whereas before the process would stall after 100 MB or so.

Even more intriguingly, if I stopped rocking the hard drive, the copying would stall. As soon as I started rocking again, the process would resume. Very strange indeed!

Still, it did not last forever. After a while, even with constant gentle rocking, the copying process stalled, and I was forced to unplug the drive and let it rest. But after half an hour I was able to again copy some more files, again with various gentle rocking movements from side to side, or even by holding the hard drive upside down and rocking it from side to side in this position!

Amazingly enough, after a little more of this, I was indeed able to rescue my sister-in-law’s entire iPhoto library. All in all, not a single file from her user folder was missing. I was able to rescue all of it, despite the fact that the hard drive was clearly on its last legs.

The only explanation that a non-engineer such as myself can come up with in this particular case is that the hard drive damage did not affect the magnetic media itself, or even the disk’s heads, but rather some other part of the mechanism that was more purely physical and somehow responded to physical movements in space. Still, when you think about how tiny things are inside these devices, the fact that such coarse, enormous (at that scale) physical movements helped at all is quite remarkable. But they did. And now I can order my sister-in-law’s new MacBook Pro and look forward to being able to rebuild her entire work environment without a single file having been lost.

The only outstanding question for me is what to do with the dead laptop. The hard drive is clearly defective, but I am still not 100% convinced that it was the only thing that was wrong with this machine. I mean, I don’t see how a defective hard drive can explain random shutdowns, or the fact that I wasn’t even able to boot from the system CD at some point. So I am a bit worried that, even if I get a new hard drive, I won’t be able to “refurbish” this machine for my sister-in-law or for someone else.

The fact that the power adapter itself is also clearly defective is another issue. I don’t know if electrical problems in the power adapter can have any impact on the laptop itself, but it certainly does not sound/look good when you can clearly hear sparkles and crackling inside the power adapter and the green/amber light around the connector to the laptop keeps going off and on.

I guess I’ll have to decide if I want to take a gamble and spend some money on a hard drive (and a replacement power adapter) or not.


5 Responses to “Defective hard drive: Rock me gently”

  1. apple4ever says:

    A good way to check the hard drive is SMART Utility. Then you can get an idea of what exactly is wrong with the HD. HD errors can cause weird things to happen.

  2. Pierre Igot says:

    Thanks. However, when problems are brought to my attention, it’s usually way too late and we’re way past the stage where a SMART warning might be useful. In addition, it looks like this utility only works for internal hard drives, not external (which I think is a limitation of the USB interface). I don’t think it’d work with FireWire either. But it’s a good link to have.

  3. AlanY says:

    Sounds like a failure of the bearings (or the mercury “bearing” system, if it was one of those drives), which would explain why rocking worked but leaving it steady didn’t.

  4. Andrew Aitken says:

    Reminds me of the days of Stiction and the “two inch drop test” we used to use to rescue dying drives. :)

  5. Pierre Igot says:

    Stiction does indeed sound a bit similar. Don’t know about all the ominous noises coming from my drive though. They don’t sound like a marginal lubricant problem to me. But who knows?

Leave a Reply

Comments are closed.