PC Tale Of Woe
All I wanted to do was test my newly burned Slackware 12.1 CD-ROMs.
So I got out an old PC with Pentium II processor, 448Mb RAM and an 8.5Gb hard disk. I've used it before for Slackware testing, and it had 12.0 on it. First thing, make sure 12.0 is OK, then blow it away with a fresh install of 12.1.
But I remembered as soon as I'd got the box out and plugged it in, that the power supply was tricky to get started last time I'd used it. It took a bit of coaxing and cajoling, but it did start up. I assumed that it was either a startup resistor going high in value, or a leaky capacitor. So, after it failed to switch on, I took out the PSU and opened it up. No sign of bulging capacitors, so I check a couple of likely-looking resistors, but they're OK. Darn. I didn't fancy my chances of fixing this, not with 300-odd volts DC on the PCB, so I reassembled it.
Then I realised that I had another PSU in an identical case, wired up to an Alpha motherboard (yes, the ill-fated DEC Alpha, a 64-bit RISC CPU, on an ATX motherboard). So, out with that and swap over the power supply to the Pentium II machine. Still won't switch on. Shorted the power-on pins with a pair of pliers, still no go. Go on-line, look up the ATX pinout and find the power-on pin on the connector. All I have to do is ground that pin, and the ATX PSU will switch on. Which it does, and off again when I remove the short.
Must be the motherboard itself that's failing to pull that pin low when I press the button. So, I pull all the boards out of the slots (it's got VGA, ethernet, sound and SCSI cards) and remove the three DIMMs and disconnect all the wires. Then, I take out the six screws and remove the motherboard from the case. Check for bulging capacitors again, but no, they're fine. Hmm... I wonder if the BIOS battery is OK... It's at 1.3V (should be 3V). Pop a fresh CR2032 lithium battery in, and sure enough the motherboard can power up the ATX power supply and all is well. Except it's not in the case any more.
So, back into the case goes the motherboard, DIMMs, I/O cards, cables and the original power supply. Of course, it works fine and I go through the BIOS setup from scratch (because it had lost all the settings when I changed the battery). But it won't auto-detect the hard disk. Oh, that's because it hasn't been used for a while and it has stiction. So, power off, remove hard disk, twist it in mid-air in the way that I've been doing since the days of the 105Mb SCSI disks in the Sparcstation 1, and put it back in (another one for the CV: skilled at shaking hard disks to make 'em spin). This time, the drive spins up OK and makes all the right noises.
But there's no screen display. The monitor goes from standby (yellow LED) to normal (green LED) but there's no BIOS memory check screen, nothing. Crap. Now what's gone wrong? Has the VGA (Tseng ET6000 on a PCI card) failed? Swap that for another, and then another. No display. Maybe I set up the BIOS settings wrongly when I re-did them from scratch after changing the battery. There's a "clear BIOS settings" jumper, so I flip that and try again. No display. Take out the RAM, and it gives a beep code for "no RAM". Put some back in, and -- aha -- two beeps, then one. Go on line, look for MSI 6117 beep codes, find some, but none that go two, then one. Even try a really old ISA bus VGA card, but just the same: monitor wakes up,
green LED, no display.
Out with all the PCI and ISA cards, out with the floppy disk and IDE cables, just a motherboard, processor, RAM and a VGA. Beeps, no display. I began to suspect that the monitor had died, but the built-in on-screen display was working, so it must be fine. Re-seat the BIOS ROM, re-seat the Slot-1 processor, no change, still no display. With more RAM installed, the beeps are delayed, but still there. Time to give up and use the other PC, so I dump the Pentium II in the bathroom and grab the Pentium Pro machine.
The Pentium Pro is a fine old pre-ATX motherboard with 128Mb of RAM in 72-pin SIMMs. But it's so old that I have to dig up a mini-DIN to full-size-DIN adaptor for the keyboard. It starts up, but there's no display. OK, time to test the monitor on the proper Linux box (Athlon XP 1800). This means crawling back under the desk to plug the VGA cable back in, at wich point at 15-pin D-connector falls on my head. The plug in the back of the monitor had loosened just enough to make the video stop working but not affect the sync. Hence the green LED and the working on-screen display. With the VGA cable plugged back in behind the monitor, the Pentium Pro is fine and boots up into LILO.
So now I go back and get the Pentium II and plug it all back in, and of course there's a perfectly fine display of the (rather slow) RAM test followed by two beeps, then one. The BIOS beeps like that when it has no valid configuration in NV-RAM, like when some geek has just flipped the jumper to erase it. So I go through the BIOS setup again and it's fine. Then I realise that I haven't plugged the drives back in, so I switch off and do that.
At last, after two hours of plugging in, unplugging and swearing, I have the Pentium II machine back at the LILO prompt. Which is where I wanted it i the first place, before pulling all the cards and cables out and plugging them all back in, twice. The power supply wasn't faulty, the battery was dead and for some reason the motherboard won't power up without it. The VGA output
hadn't died, the cable had fallen out of the back of the monitor.
Slackware 12.0 now starts up just fine. In with the CD-ROM and that boots up just fine. Without further ado, I ran the Slackware 12.1 installation program and set it going with a full install. The CD-ROMs appear to be fine.
And that, boys and girls, is why I hate working on PC hardware.
So I got out an old PC with Pentium II processor, 448Mb RAM and an 8.5Gb hard disk. I've used it before for Slackware testing, and it had 12.0 on it. First thing, make sure 12.0 is OK, then blow it away with a fresh install of 12.1.
But I remembered as soon as I'd got the box out and plugged it in, that the power supply was tricky to get started last time I'd used it. It took a bit of coaxing and cajoling, but it did start up. I assumed that it was either a startup resistor going high in value, or a leaky capacitor. So, after it failed to switch on, I took out the PSU and opened it up. No sign of bulging capacitors, so I check a couple of likely-looking resistors, but they're OK. Darn. I didn't fancy my chances of fixing this, not with 300-odd volts DC on the PCB, so I reassembled it.
Then I realised that I had another PSU in an identical case, wired up to an Alpha motherboard (yes, the ill-fated DEC Alpha, a 64-bit RISC CPU, on an ATX motherboard). So, out with that and swap over the power supply to the Pentium II machine. Still won't switch on. Shorted the power-on pins with a pair of pliers, still no go. Go on-line, look up the ATX pinout and find the power-on pin on the connector. All I have to do is ground that pin, and the ATX PSU will switch on. Which it does, and off again when I remove the short.
Must be the motherboard itself that's failing to pull that pin low when I press the button. So, I pull all the boards out of the slots (it's got VGA, ethernet, sound and SCSI cards) and remove the three DIMMs and disconnect all the wires. Then, I take out the six screws and remove the motherboard from the case. Check for bulging capacitors again, but no, they're fine. Hmm... I wonder if the BIOS battery is OK... It's at 1.3V (should be 3V). Pop a fresh CR2032 lithium battery in, and sure enough the motherboard can power up the ATX power supply and all is well. Except it's not in the case any more.
So, back into the case goes the motherboard, DIMMs, I/O cards, cables and the original power supply. Of course, it works fine and I go through the BIOS setup from scratch (because it had lost all the settings when I changed the battery). But it won't auto-detect the hard disk. Oh, that's because it hasn't been used for a while and it has stiction. So, power off, remove hard disk, twist it in mid-air in the way that I've been doing since the days of the 105Mb SCSI disks in the Sparcstation 1, and put it back in (another one for the CV: skilled at shaking hard disks to make 'em spin). This time, the drive spins up OK and makes all the right noises.
But there's no screen display. The monitor goes from standby (yellow LED) to normal (green LED) but there's no BIOS memory check screen, nothing. Crap. Now what's gone wrong? Has the VGA (Tseng ET6000 on a PCI card) failed? Swap that for another, and then another. No display. Maybe I set up the BIOS settings wrongly when I re-did them from scratch after changing the battery. There's a "clear BIOS settings" jumper, so I flip that and try again. No display. Take out the RAM, and it gives a beep code for "no RAM". Put some back in, and -- aha -- two beeps, then one. Go on line, look for MSI 6117 beep codes, find some, but none that go two, then one. Even try a really old ISA bus VGA card, but just the same: monitor wakes up,
green LED, no display.
Out with all the PCI and ISA cards, out with the floppy disk and IDE cables, just a motherboard, processor, RAM and a VGA. Beeps, no display. I began to suspect that the monitor had died, but the built-in on-screen display was working, so it must be fine. Re-seat the BIOS ROM, re-seat the Slot-1 processor, no change, still no display. With more RAM installed, the beeps are delayed, but still there. Time to give up and use the other PC, so I dump the Pentium II in the bathroom and grab the Pentium Pro machine.
The Pentium Pro is a fine old pre-ATX motherboard with 128Mb of RAM in 72-pin SIMMs. But it's so old that I have to dig up a mini-DIN to full-size-DIN adaptor for the keyboard. It starts up, but there's no display. OK, time to test the monitor on the proper Linux box (Athlon XP 1800). This means crawling back under the desk to plug the VGA cable back in, at wich point at 15-pin D-connector falls on my head. The plug in the back of the monitor had loosened just enough to make the video stop working but not affect the sync. Hence the green LED and the working on-screen display. With the VGA cable plugged back in behind the monitor, the Pentium Pro is fine and boots up into LILO.
So now I go back and get the Pentium II and plug it all back in, and of course there's a perfectly fine display of the (rather slow) RAM test followed by two beeps, then one. The BIOS beeps like that when it has no valid configuration in NV-RAM, like when some geek has just flipped the jumper to erase it. So I go through the BIOS setup again and it's fine. Then I realise that I haven't plugged the drives back in, so I switch off and do that.
At last, after two hours of plugging in, unplugging and swearing, I have the Pentium II machine back at the LILO prompt. Which is where I wanted it i the first place, before pulling all the cards and cables out and plugging them all back in, twice. The power supply wasn't faulty, the battery was dead and for some reason the motherboard won't power up without it. The VGA output
hadn't died, the cable had fallen out of the back of the monitor.
Slackware 12.0 now starts up just fine. In with the CD-ROM and that boots up just fine. Without further ado, I ran the Slackware 12.1 installation program and set it going with a full install. The CD-ROMs appear to be fine.
And that, boys and girls, is why I hate working on PC hardware.