This list is compiled from my many years of being knee deep in x86 hardware, dealing with the public, etc. I know I'm missing some, but I just can't remember all of them.
The form of death most often seen by a video card is being inserted into the wrong slot, resulting in some spectacular forms of pyrotechnics.
Nearly all new machines come with an AGP slot for a video card. This slot is designed for high speed data transfers, lifting the burden that the PCI bus would have to deal with. With PCI being a shared bus, the video was just too demanding. In any event, users would most often buy a video card using this newer interface, ignoring pin counts on slots & inserting it into a PCI slot. The results usually are, melted plastic of the PCI slot, the surounding traces are scorched & raised from the motherboard. The surounding capacitors usually burst, the capacitors on the video card burst, the GPU is badly burned, the lettering printed on the ram on the card is usually a dull brown, & the persistant smell of burnt nastiness lingers for 20 minutes or so after the unfortunate events.
Ram chips are mounted to a PCB which is then inserted into various types of slots in a motherboard. The main types used for the last decade in order of introduction: 30 pin simms, 72 simms, EDO DIMMS, SDRAM, DDRSDRAM, RDRAM. These all have varying pin sizes, differing voltages, etc. I have seen many people who, when confronted with a slot their newly purchased ram wouldn't fit into, they decided to force it. This, in many cases nukes the board, or the slot at the very least, & most certainly the Ram Module.
You have different forms of processors used in the past, & today. Different generations have different interface designs, & each generation has sub generations, & sometimes more after that. The subtleties are often lost on the average joe, who just "wants to go faster!" w00t! lets hear it for ignorance, but anyways...
ZIF Socket based cpu's:
A cpu that uses a Pin interface using ZIF sockets usually has a marker on one corner. Making it impossible to incorrectly mount the cpu if you're paying attention. Sadly, many would just insert the cpu, paying no heed to the warnings that came with the board they mounted it to. This course of action insures the death of the cpu in roughly 85% of the cases, & in at least 65% of the cases, the board passes on, or shortly thereafter.
Some cpu's are OEM, some are Retail Box. The difference being retail box gives you a 3 year warranty with the respective manufacturer, & an authorized HS/F(Heat Sink/Fan) unit. OEM means, cheaper price, but no hs/f unit; find your own. Also, the warranty period varies from reseller to reseller. I've seen 3 years from some, to 3 hours from others. While kind of questionable, they are all perfectly legal. This factors into the situation in that, not all hs/f units are made for every processor.
Another form of death that the socketed CPU faces is core meltdown. Not exactly chernobyl, but close enough. Due to the designs of each respective core, some can last longer than others without anything removing the heat they produce. With todays higher clock frequencies, that time is shortening almost exponentially. Test example: An AMD Thunderbird Athlon running @ 1Ghz/1000Mhz, which has 64k of on die L1 cache, & 256k of on die L2 cache, running at the full clock speed, will incinerate itself in roughly 8 seconds from power on without the use of a proper HS/F unit.
Slot style cpu's:
With the high expense of On-die cache while still using a 0.25 micron process, both Intel & AMD were forced to add off-die cache to their cpu's. Intel used Slot 1, with 2 SRAM modules on the left side of the cpu core mounted on a PCB, while AMD used Slot A, with 2 srams, each opposite the cpu mounted on a PCB. The fate most often encounted by this cpu is the insertion of it into the slot incorrectly. There are plastic spacers to prevent this, but some people like to force them in, rather than double check, & some manufacturers used slots that didn't have any spacer...*poof*
For example, there are slight differences in the height of the cores on Socket-a Athlons/Durons, & Socket 370 P!!!'s/Celerons. Using a hs/f unit meant for another type of cpu can result in the following: crushed/fractured core. Easily detected by the silicon bits around the base of the core. Or worse, a slight space, while still filled by the thermal paste between the core & the hs/f unit, it dries out fast in those conditions, & the core has a meltdown.
Motherboards die a tragic death. They lay in wait of a good home, but a frightengly high percentage of their owners mistreat them.
Each motherboard has holes that line up to holes in a case. In those holes in the case, you install standoffs.A standoff is what the board sits on, to be mounted to the case. There are two kinds, those that screw in, & those that fit into little metel notches. Regardless of what kind are used, they should always be used! Some people will attempt to mount the board directly into the case. Circuitry mounted directly to a steel sheet. Some boards toast immediately, while others like ASUS, one of the greatest manufacturers, can last hours, sometimes days. Some even work after being removed, & mounted correctly; a testament to superior engineering.
Others only use half the standoffs, so it is a precarious balancing act. It's not grounding against the case now, but if it's bumped, you have sudden hardlocks that are hard to explain. The 1st thing I did when a customer came back with their machine, with a board they had just installed was check it for being properly mounted.
A properly built machine, should be able to be running, & held in the air at odd angles, & shaken, twirled, etc. Proper assembly is a must. Like their can be only one highlander, there is only one way to properly mount a motherboard.
Old boards of the AT standard had 2 connectors that supplied power. The simple rule, black in the middle! Red in the middle, & she's a toastie.
I have seen these used as cup holders. The plastic eventually fractures, & the drive needs to be replaced. Also, some people just shove cd's in...the cd is not aligned properly, it begins to be spun at a high rate, the balance is off, the cd hits the side of the drive, it shatters, sharp plastic bits go flying, damaging the internals of the drive. Imagine a glass shattering in your chest cavity. owe.
Floppy drives usually just go, they take a trip w/ Charon suddenly, but they usually take a few floppies & their file systems with them.
Power Supplies are always interesting. It is truly like performing an autopsy. You see, when PSU(Power Supply Unit) dies, it often takes a few things with it. In its last throws, the power regulation can be eratic, & some parts receive surges they were never designed for. The parts most often assasinated by this are Hard Drives, Motherboards, & sometimes cd roms. Though I have seen cases where everything went south, save for a $10 winmodem. A truly cruel fate for the owner of said machine.You spend a few hours testing each component. It is like laying a patients organs out on the table.
PSU's can die in different ways. A capacitor can blow off the PCB, & just go. There can be a short in the wire, in which you have a spectacular fire, & the plastic is melted all along the wire, & the drive that connected to it lays in melted glory. derobert reminded me of another one. If the cooling fan in the PSU dies, it can slowly bake & die from being run out of spec for so long, heat wise.
Hard Drives I damn near forgot, thanks Ground Control.
Hard drives can go south via a head Crash. Inside your hard drive, you have polished disks, which are read by read/write heads on either side. For various reasons, sometimes power loss or internal malfunction, the heads can slam into the platters/disks. The result is like the Stay Puft marshmellow man falling over. Data is trashed, the file system damaged, mass hysteria. Usually you can get most of the data off, but there is some damage, & I wouldn't trust the drive after that for anything.
Sometimes a drive will just go due to age. When you start to hear a rythmic *click*Click*click* that is not caused by normal activity, I would look into backing up the data & replacing the drive.
Ride the Lightning!
While a power strip with surge protection is good, it is in vain if used without a phone line surge protector as well. Sometimes it is just the modem that cooks, the motherboard as well, & sometimes the entire machine. You know it's an almost total loss when you remove the cover, & then gag. The smell of burnt circuitry fills the room. When confronted with a scene of that nature, I proceed to giving last rights before a cursery test of each component. Knowing it is in vain, but there may be a lone survivor; you pray for such.
VXO writes... "Conductive liquid + laptop" = dead laptop.
Fire up your experiences, & I'll slice n' dice'm into this Node.