Layman question: When a laptop has a head crash, that by definition means it scratched stuff off the disk, kind of like a needle-recording? And that stuff - paladium, aluminum, glass, etc. - just gets scratched off as particles and blown into the air inside the computer? Then does that mean the data, or part of the hard drive, got scratched off physically? I have been Wikipedia-ing it and am totally confused.
Yes, under some circumstances (like excessive shock) the head(s) can hit the platter and this can happen. There are pictures on Wikipedia where you can see the damage. Ejected particles will remain inside the disc enclosure, where the head can hit them, too. If your hard disk suffered a head crash, just buy a new one.
At one point, I would manually run a command to park the heads before shutting off the power to my computer, just in case. These days, that should all happen automatically, and there is additional protection in place as well (e.g., in case you drop it). But for something like a laptop computer, you may as well abandon hard disks as last century’s technology and buy a storage medium with no moving parts.
OK thanks. I gave my 2 laptop computers several accidental hard jolts while they were on and running. I can’t detect hard-drive damage, but then again that’s hard to detect anyway. Wonder if laptops have physical-shock-protection? I know some are meant to immediately park the heads if they sense gravity’s freefall (sensing imminent impact,) but this was different, the hard bumps to the computers came in a way that their systems couldn’t have possibly sensed in advance.
The computer’s ventilation fan doesn’t blow them out the laptop’s vents?
A hard disc enclosure is sealed to keep stray dust particles out.
If particles of a hard drive platter could get out of the hard drive, then dust could get in. Since dust on your hard drive platter would be a very bad thing, the whole hard drive is sealed.
If you haven’t done so before, I suggest taking one apart. The magnets can be great fun if you avoid hurting yourself with them.
What? 
Most (all?) hard drive heads are controlled electromagnetically in a way that involves powerful rare-earth magnets. This magnets are really cool. The are also capable of crushing skin if you manage to maneuver them into a position where they can do so. This is very unpleasant, and such a position usually also means there is no way you can pry the magnets apart, especially not while a bit of the pad of one of your fingers is caught.
Oh, and taking apart a hard drive is a one way operation outside a lab. Don’t do it if there’s a chance you might want to use it again.
A “Hard bump” is unlikely to cause any damage. Laptop drives are designed to withstand 100’s of g’s of acceleration when operating. A “bump” might be a few g.
The head itself can be contaminated with stuff scraped off the surface during a bump. So even if the head is moved to a track without damage, it might no longer function properly.
Most HDs are very nearly sealed. There’s a tiny pinhole with a filter for balancing pressure. You don’t want the case expanding and contracting due to air pressure changes, esp. due to the air inside heating during operation. There might be a warning on the drive label telling you not to cover up a spot.
OTOH, there are helium filled hard drives*. The lower density of the gas helps improve all sorts of things resulting in denser medium and therefore higher capacity drives.
- How do you do a “sic” for a typo in a URL?
Nope. (That is only one example.)
I once bought an old dead 5 1/4 full height 10 MB drive for a few bucks as a novelty. I opened it up, and the top platter had a deep shiny curved track cut into the aluminum where a head crashed. (Modern drives, having much smaller and lighter parts, probably don’t have anything quite that dramatic happen.)
I had a boss about 25 years ago who had one of the old DEC hard drive platters with grooves on it from a head crash. This was back when the platters were around 18 inches across and the whole hard drive box was the size of a piece of furniture. His story was that some one came into his office saying the audible alarm on the hard drive was going off. There wasn’t an alarm it was just the shrieking noise of the head scraping along the platter.
I’d say that guy has a nice home lab. 
I worked for DEC back in those days. Many folks (myself included) had a deeply grooved platter in their cubes as a badge of honor.
Plus the on the web page the guy says the drive was from 1997. That is over 20 years old.
Well, naita is correct in the sense that taking a hard drive apart outside of a properly equipped clean room is in general a very bad idea. Sure, it might work, if the environment is exceptionally clean and you’re lucky. But then again, it might not. The flying height of a modern HDD head might be only a few nanometers – this illustrates the scale – and any disruption of those few nanometers of critical air cushion is going to cause a head crash.
Those were the days when most disk drives had removable platters. Fixed-head non-removable drives did exist, as well, but they were less common. The entire drive was indeed like an appliance, resembling a washing machine. The platters were stored in plastic cases that looked like those transparent plastic cake boxes, and the platters themselves resembled a stack of giant pancakes (the whole scene was enough to make a person hungry!). The platter was lowered into the unit with the top still attached, which was then unscrewed and the drive door closed.
It was all a minimal effort to keep contaminants off the platters but tolerances were much, much lower on those old drives. The heads moved laterally, rather than the radial movement of modern drives, and with considerable force. The big platters and the heads were spring-cushioned as a single unit, and when the heads were seeking you could see the whole assembly bopping around quite violently.
By programming the correct seek patterns you could walk an RL02 unit across the room.
I wrote test software for those drives for Memorex in 1979. My favorite part of the test software was a program called SCRATCHFILL. The program looked for a set of bad blocks arranged in a straight line on the platter, and assumed that this was due to a scratch that occurred during manufacturing. Any good blocks along that line would be marked bad, because they may have been “nicked” and now be marginal.
–Mark
IIRC several of the important data-center class drives marketed by DEC as the RP series were actually rebranded Memorex drives, notably the RP01, the venerable RP02 (which was a Memorex 660-1), and the RP06.
The RP02 had a capacity of 20 megabytes and (around 1968-70) cost about $25,000, which in today’s dollars would be about $167,000. IOW, if you had a 20 MB disk drive in your house at that time, it would have been worth a lot more than the house.