Computer Technicians 203: Hard Disk Failures

In our last episode on Optical drives and firmware flashing, we discussed ways to upgrade your optical drive’s speed and features. While it’s a great finishing touch for both your own and your clients’ computers, it’s not exactly a common thing you’d be called up for. One of the most hazardous, annoying, and generally panic-inducing failures that happen on any computer are hard-drive failures. While they can be moderately to extremely hard to fix, diagnosing them comes first, so you could be able to determine just how much damage the disk sustained and can sustain before it finally dies.

The Anatomy of a Disk

Contrary to the most commonly used terminology, the hard drive and hard disk aren’t exactly the same thing. A hard drive is the entire unit, which consists of the metal chassis, the spindle motor, the disk platters (the actual hard disks, usually aluminium or glass substrate plates), the head actuators, the read-and-write heads and the drive electronics. Knowing this distinction is vital because of its accuracy – as you’ll read later on, faults may strike any of those components, or even more of them at the same time.
Today’s hard-drives are much more advanced than their predecessors – while the first hard-drive was approximately the size of a medium printer or fax-machine, it had a capacity of a measly 5 MB – something you could fit onto four floppy disks nowadays. The drive platters in those drives were about the size of a 12″ LP record, data density was smaller – however, the drives were sometimes enclosed in clear plastic, allowing both engineers and users to see how they work. Since you’re not likely to run into any of those, here’s a short explanation. In essence, the magnetic drive heads hover a couple of micrometers above the platters, never actually touching them. The head actuator moves it angularly over the portion of the disk covered by the drive’s arm length. The spindle drive keeps the disc rotating up to its maximum speed, either 5400, 7200 or 10800 revolutions per minute in most common drives, and allows the actuator arm full coverage of the disk surface. Much like the optical drive circuitry, the hard drive’s electronics read and write to the disk, handle the data and transport it over either the PATA or SATA cabling. Finally, the metal chassis seals it in to ensure there is no dust inside – this is vital because dust particles can be fatal to a drive’s overall health.

Broken lines and logical fallacies

So what exactly can go down on a disk? Like the similar “what could possibly go wrong”, that’s a question sometimes best not even uttered. It’s practically an invitation for something to happen to it. First of all, there’s the most common and most annoying case: one of the cables inadvertently slipped out of the drive or motherboard. That in itself can cause the hard drive to get intermittent power or connection with the motherboard. In that case, the less unseated the connector is, the better, as ironic as it seems. With a complete power or data cutoff, the worst that can happen is that the drive won’t show up during either POST, bootup, or inside the OS. If the cable is partially unseated, a lot of different effects can happen – the drive might write data without being able to read it, the drive might pop in and out of functionality, the drive motor could come under a constant spin-rest-spin-rest cycle which is uncontrolled and can damage the drive motor, and so forth. Fixing these is pretty simple, if the problem indeed is the cable – either unseating and reseating the cables or getting new ones can fix the problem. Rounded PATA cables used to be unreliable because of poor manufacturing quality and easily breakable wires, but modern-day rounded cables don’t suffer from it. SATA cables are notorious for being relatively easy to unseat due to the short and shallow contacts, so take special care when mounting or remounting them.
The second category includes physical errors in the drive media (the platters themselves). By far the most common of those is a bad sector – literally, a disk section that became unreadable due to mechanical damage and data stored on it is unreliable at best. These errors can be averted by teaching yourself (and the client) to scan the disk surface with ScanDisk, CHKDSK or fsck once in a while in order to check the filesystem’s integrity, as data may as well become corrupted. Note that the utilities don’t automatically performe surface scans, you usually have to check an option or a switch. At the same time, those utilities can pick up other errors such as inconsistent file-tables which can to drive space being incorrectly reported or files rendered inaccessible. This third category said programs can cover contains logical errors, which are almost always a combination of the operating system’s and the drive’s fault.