All images on this page are clickable..

This is the actuator arm, which moves the drive heads around the disk surface. I've removed it from it's assembly, so that the carefully coiled wires can be clearly seen - when these wires have a current passed through them, the magnetic field created by the extremely powerful permanent magnets I mentioned before causes the coil to be pushed or pulled back and forth between the magnets - this is how the drive heads are positioned to read and write on the platters as they spin. The current that is passed through the coiled wires is exactly and precisely controlled by the drive electronics, so that the heads can be moved to exactly the right position on the disks when needed - and at high speed.

I told you the platters were shiny! You can very clearly see my digital camera, hands, the top of my head and the downlight in the ceiling above my desk in the reflection showing on the surface of the top platter. However, there appears to be an imperfection on the disk surface just above the reflection of my hand...

And oh what an imperfection it is. Somehow, a rather nasty scratch has "appeared" on the surface of the disk platter. I would say that there is a VERY good chance that this scratch was the cause of my drive's death - the scratch was there when I opened the drive - before I'd even had a chance to scratch it myself. I can't see how the scratch has occurred during the operation of the drive - when the drive is operating, the platters spin at 7,200rpm, and any scratch would be radial. When the drive is stopped, the heads are parked out of the way. Oh well - at least I know why it died now. Also noticable in that photo are a few specs of dust - it's amazing how quickly the specs of dust started to appear on the drive surface - and how noticable they are when they're sitting on such a smooth, reflective surface.

As this photo shows, it's rather easy to leave marks on the surface of the platters. The oil on your skin permanently damages the platter - even after wiping the oil off, there's a permanent "shadow" of the thumbprint still on the platter. Again, specs of dust are clearly visible on the surface of the platter.

This is what my IBM drive looks (looked) like underneath. Clearly visible is the PCB I mentioned earlier, containing all the drive's controlling electronics. On the left are the IDE connector (closest to camera), jumpers (for setting master, slave, etc) and power connector (furthest from camera). The orange "ribbon cable" (for want of a better term) leading off to the right provides power and control for the motor to spin the platters. I'm not sure what all the surface components on the PCB are - there are a few mosphets, I assume to regular voltage, and I think the big chip is the "controller", while the chip on the right side closest to camera (you can't really read the screenprint, but it has "Etrontech" printed on it) I believe to be the 2MB cache that the drive features. I removed the four screws holding the PCB in place, but expected it to still be rather firmly attached to the drive casing through some point where the electronic connections pass through the casing into the interior of the drive, but the PCB came completely free, revealing the following...

  

It appears that the sole form of electrical contact between the PCB and the components inside the drive is through a series of contacts which are literally "pushed" together - there's no soldering or anything other than the pressure the PCB exerts on the hard drive casing keeping the electronics on the PCB connected to the components inside. I'm no electrical engineer, but to me, that doesn't seem like a totally reliable way of making the connection required. I'm sure IBM had suitably qualified electrical engineers on staff to design these drives though, so I am probably wrong in my assessment of the reliability of contacts like this. Either way, that's the way IBM decided to do it, as you can clearly see. The contacts on the back of the PCB are really just very small "mounds" of solder, while the contacts on the drive casing seem to be "spring loaded", in that you can push them down and they sink back into the casing.

Well, basically, that's what's inside a modern day hard drive. My drive was a 20GB model, with two platters (none of the photos I took really show that well - you'll just have to trust me) - other models in the 75GXP range have more platters than my drive did - up to 5 platters in the 75GB model I believe. Not only have I managed to pull the drive apart, but I've also discovered the likely cause of this particular drive dying - I'm quite sure that the reportedly massive number of dead IBM hard drives recently haven't ALL died in the same fashion, but it does appear that my drive at least has suffered some sort of physical damage either during manufacturing, or during it's operational life (which, considering how the drive is designed, seems to be a little unlikely to me).