IBM Hard Drive Autopsy

by Adrian Black Performed April 17, 2002

Today I took home two dead IBM hard drives from work to .... well, just to take them apart. :-)

One of the drives was a dead IBM GXP75 drive. Classic scratch-click, scratch-click, etc.

Basically the same thing that has happened to all of these drives. I had one at home do this and this particular drive was from a Macintosh. (30gig.) I know numerous people who've had this happen to their GXP75's as well.

The other drive was a 4gig Travelstar drive that was from my laptop. It was noisy and horrid -- and then it decided to give up the ghost. Click-chunk. Click-chunk. I lost everything on that drive.

Unfortunately for me, both of these drives are OEM and have no warrenty from IBM.

Here are two pictures of the drive before I touched them. I had already peeled off their labels as I know there are always screws hidden under them.

First I began on the laptop drive.

Nothing interesting other than it's pretty cool!! IBM Travelstar drives always have a sound like something is loose inside of them if you handle them -- turns out it is a lock mechanism for the heads and I assume is designed that way. It's the silver metal/bar thing closest to you on the first picture.

Next I started on the GXP75. The drive worked pretty well befor I began. Occasional scratch-click but it only takes one time to head that sound when I decide it's time to yank the drive.

Very similar to the travelstar except more "spacious" inside -- if you can call it that. Haha.

So I plugged some power into the drive. IMMEDIATELY, scratch-clunk, scratch-clunk. Wierd, I thought, since the drive was working pretty well before I removed the cover. So with the power on I started fiddling with the head assembly by trying to move it by hand while it was clunking around.

I found something interesting, if I kinda pulled up on the it, the drive would settle down and be happy!!

So time for an expriment. I put the cover back on and used only the single screw that goes into the head assembly.

With the screw loosely put on, it would just scratch-clunk without working at all. I found that as I tightened it the drive would work better. When I tightened it down as much as I could without having the correct Torx bit, the drive would work normally!! I botted up a linux CD and proceded to 'fdisk' and mke2fs it!! BLAM! Worked fine. So then I ran badblocks on it. It did end up having some bad blocks but DUHH ... I had been messeing around with it with the cover off for over 30 minutes and moving the heads around by hand.... I'm amazed that it worked at all.

But what I was amazed at was the lack of scartch-click and the correlation to how tight that screw was.

I yanked the cover and put just the srew in the hole. Then using my finger nails I pulled up on it while the drive was powered up. Scratch-click, scratch-click ... pull up ... NORMAL. Let go, scratch-click, scratch-click.

So may I be onto something here!? I don't remember that screw being very tight when I took it off (using needle nose plyers.) Maybe all it needed was to be tightened down and that would fix the drive!?!

The GXP75 I had at home a few months ago that would scratch-click all the time, but I threw it away. I wish I had that right now -- I would test my theory.

I may be way off base, but what if all that's required to fix a scratch-click GXP75 is tightening that screw down!? It will voice the warrenty as you have to peel off two stickers to get to it but for people with OEM drives, it may be worth a try!

Oh, well, that was all I could do with this drive. I wanted to see what was up with these "glass" discs that IBM uses.

Sure enough, they are glass! I put one in a bag and put it under my chair leg. I had to really jab it but it did smash!! I tried hitting it with some tools on the edge of my table first but it was STRONG and didn't break.

Next, I took the voice coil magnet from the laptop hard drive.

It's very small, as you can see ... but DAMN, is it strong!!

Here it is holding up an entire box of mostly full Cheerios on the fridge!! I'm going to extract the GXP75's magnet next -- I'm sure it's even stronger.

Nothing here can be reproduced without permission.
Copyright Adrian Black (C) 2002.