I should warn you that this is going to be a long post -- I am stuck at home because of the monsoon rains and have nothing better to do, so go and get yourself a nice, hot cup of coffee before reading on.
All my troubles started when, in a fit of uncalled-for exuberance, I removed 128 megs of RAM from my Windows machine and tried to add them to my Linux box. I must have plugged the RAM the wrong way, because the motherboard started giving out a burning odour and the machine refused to boot. I quickly removed the offending memory card, but found that one of the three power supply wires for the CPU fan had been severed (not sure whether it was burnt or had given out earlier due to some other reason). I re-attached the wire, but restarting the PC resulted in the BIOS POST failing, with long beeps repeated in an infinite loop. I tried my usual trick of re-plugging and swapping the IDE cables, but this did not solve the problem. One suggestion I found after Googling for this error was to check the CPU fan wires. Thinking that my quick fix might not have been sufficient, I made a quick visit to the neighbourhood electronics store for a new CPU fan.
The problem was not with the fan, it turns out. The PC continued its long (and, at this point, incredibly irritating) beeps. I finally had to call for a hardware engineer, who quickly made the beep go away by resetting the CMOS jumper. The PC now booted, but now it reported that there was a CRC failure in the hard disk. I sent away the engineer, thinking that this was a software problem that I could handle on my own.
Repeated reboots resulted in different failures -- kernel panics midway through the boot process at different points. Finally, the dreaded long beep returned again.
At this point, I did what I must have done a long while ago: RTFM. I realised that the long beeps meant that the memory was not being recognised. After repeating the hardware engineer's trick and getting the machine to boot, I ran a memory test via the Suse installation CD: turns out that the memory had gone bad (why the BIOS POST didn't detect this is still a mystery to me).
Cut to me receiving a new 256 meg memory card from Fabmall three days later. I plug this new memory in, reboot, no go: same old story.
The story does not have a happy ending yet. I have managed to get the PC to boot and stay up as long as I don't try any networking activity (LAN or the internet).
The only thing that redeemed itself during the entire nasty episode is the eight-year old 2GB hard disk that showed it still had it in it when I substituted my regular hard disk with it in a bid to isolate the problem.
I am mulling over whether I should get a new motherboard or bite the bullet and treat myself to a long-overdue laptop.