Hi all,
Back in mid-march I set about building a new computer. The setup was as follows:
The first problem I had was the Ninja Rev B, which uses push-pins to attach itself to the motherboard. The only problem is that these were extremely hard to attach properly with the heatsinks around the CPU. At this point I couldn't get all 4 push-pins to attach correctly, and my computer would boot up, then crash after about 2 minutes, even just sitting in the BIOS setup.
Rebooting would result in memory-error type beeps and no other response, only attempting to boot after I had turned the computer off completely for a short while. I assumed that this was due to the not properly attached cooler, so I replaced it with a Zalman CNPS9700-NT nVidia Tritium CPU Cooler (which has a nice proper screw-on setup that the motherboard heatsinks didn't intefere with).
After replacing the CPU cooler, the computer still wouldn't boot for more than 2 mins. I set about watching the BIOS health readouts as the crashing was happening. These didn't indicate a major rise in temperature. I set about swapping out the RAM so that only 1 module was installed at a time, trying alternative modules, and keeping all non-essential devices detached.
Nothing I tried solved the issue. I even purchased a 40 quid Celeron (cheapest CPU I could get hold of at short notice, since all my other machines are AMD) and this had the same issue.
I replaced the motherboard with a Gigabyte N650SLI-DS4L which I got cheap from a local PC store (on the premise that it would be quick to return if I wanted to later).
This appeared to work perfectly, so I stuck with it. I had no other issues until last week.
I came home one day when it had been fairly warm to my computer screaming at me (long beeps). I assumed this was a simple overheat and hit the reboot button expecting it to return to normal practice (assuming it would've cooled between it dying whenever and me getting home). However, this was not to be. The computer booted to the BIOS Boot Block (iirc) and asked for either a hard disk or floppy disk with a BIOS image on.
What? I can't just use a USB drive (got 4 of those!)? What is this? 1990? I don't have floppy disks? I don't have a floppy disk drive! So off I went to the local computer store and timmidly asked if they had any floppy disk drives. It turns out they do still make them, so I happily walked home. Except I'd forgotten to buy some floppy disks. Fortunately I managed to dig about 5 out of the back of a cupboard. First one was dead (uh oh, I bet they're all dead I'm thinking, remembering the days when I'd carry around atleast 2 copies of files on seperate floppys). Second one worked. Hooray!
I got the BIOS loaded onto the motherboard successfully. The BIOS reinstall seems to cause the network device MAC address to change, which I correct. Networking then works fine in linux, but not in Windows. A dead network device is easily solved tho, so I get out a handy rtl8139too compatible PCI network card and stick it in, which works fine under both linux and Windows. Connectivity is restored.
A couple of days later windows crashes (I think I saw a flash of blue, so it may have been a bsod, but Windows annoyingly automatically reboots immediately - I've since rediscovered how to disable this). BEEEEP BEEEEP - black screen, my computer is screaming at me again (long beeps).
I reboot and check the even logs. The last entry is at the time of the crash and tells me that Windows has reached its TCP/IP connection limit. Odd. I certainly wasn't running anything network intensive (not gaming and torrents and email run on a separate box). This occurs twice more in the space of a week, even after uninstalling the nvidia network device and nvidia network allocator(?) from Device Manager (altho I'm not sure how much of the drivers would've been uninstalled, given that it's an nforce motherboard and thus still needs the rest of the nforce drivers).
Yesterday I decided to leave my PC running Linux to see if that had the same issues. Today I was browsing the web and viewing some videos - nothing intensive - and my PC suddenly dies (screen freezes, lock lights on keyboard won't toggle, ssh fails). I reboot and try to go back into linux. After selecting the entry in grub, the screen goes black and stays black.... Nothing. So I reboot and try again, same result.
Next I reboot and try my previous kernel. This works fine, so I mount /boot, cd /usr/src/linux and "make install" to reinstall the kernel (it's the right kernel). Reboot and try again. Same result - black screen, no kernel booting up. Another weird result!
I've just disabled the motherboard onboard network device in the BIOS setup (oops, maybe I should've tried that before) and linux now boots fine again.
BIOS temp alarms look fine. Cabling all looks fine, neat and tidy (no horrible thick IDE cables in the way \o/ ). The system has 1 SATA HDD (1TB).
This system is running through a UPS (APC model - lasted around 13 mins on my Athlon64 - haven't tried it with this system yet), so power fluctuations shouldn't be an issue. (And the UPS doesn't react in any way when the crash happens - no switching to battery even momentarily - it gives a definitive click when it does that so I usually notice - plus my speakers don't run through it, so I usually notice =D )
No metal filings or poorly stamped holes in the case.
Much blood has been sacrificed already in the creation of this computer - most of it mine. The Asus motherboard still has the blood stains to prove it (there's one on the side of the P/S2 connectors).
Am I just unlucky, or is there some common thread causing all these issues that I can't see? The only thing I haven't changed out is the PSU, which I unfortunately don't have a viable replacement for at the moment (it's a 400W with 20A rating on +12V, which a friend tells me won't be enough for the 800GTX - unfortunately without any onboard graphics on my mobo and this being my first PCI-E graphics card, I have no replacements I can use).
I've used AMD for years (since the K6-2) and I'm now beginning to regret not being patient enough to wait for the fixed version of the Phenom.
Back in mid-march I set about building a new computer. The setup was as follows:
- Intel Q6600 Quad core CPU
- Asus P5N32-E Plus motherboard
- Asus Nvidia 8800 GTX 768MB PCI-E
- 4 GB RAM (2x1GB Crucial Ballistix DDR2 PC2-6400)
- Zalman ZM-600HP Heatpipe Cooled 600W Modular PSU
- Scythe Nija Plus Rev B CPU cooler
The first problem I had was the Ninja Rev B, which uses push-pins to attach itself to the motherboard. The only problem is that these were extremely hard to attach properly with the heatsinks around the CPU. At this point I couldn't get all 4 push-pins to attach correctly, and my computer would boot up, then crash after about 2 minutes, even just sitting in the BIOS setup.
Rebooting would result in memory-error type beeps and no other response, only attempting to boot after I had turned the computer off completely for a short while. I assumed that this was due to the not properly attached cooler, so I replaced it with a Zalman CNPS9700-NT nVidia Tritium CPU Cooler (which has a nice proper screw-on setup that the motherboard heatsinks didn't intefere with).
After replacing the CPU cooler, the computer still wouldn't boot for more than 2 mins. I set about watching the BIOS health readouts as the crashing was happening. These didn't indicate a major rise in temperature. I set about swapping out the RAM so that only 1 module was installed at a time, trying alternative modules, and keeping all non-essential devices detached.
Nothing I tried solved the issue. I even purchased a 40 quid Celeron (cheapest CPU I could get hold of at short notice, since all my other machines are AMD) and this had the same issue.
I replaced the motherboard with a Gigabyte N650SLI-DS4L which I got cheap from a local PC store (on the premise that it would be quick to return if I wanted to later).
This appeared to work perfectly, so I stuck with it. I had no other issues until last week.
I came home one day when it had been fairly warm to my computer screaming at me (long beeps). I assumed this was a simple overheat and hit the reboot button expecting it to return to normal practice (assuming it would've cooled between it dying whenever and me getting home). However, this was not to be. The computer booted to the BIOS Boot Block (iirc) and asked for either a hard disk or floppy disk with a BIOS image on.
What? I can't just use a USB drive (got 4 of those!)? What is this? 1990? I don't have floppy disks? I don't have a floppy disk drive! So off I went to the local computer store and timmidly asked if they had any floppy disk drives. It turns out they do still make them, so I happily walked home. Except I'd forgotten to buy some floppy disks. Fortunately I managed to dig about 5 out of the back of a cupboard. First one was dead (uh oh, I bet they're all dead I'm thinking, remembering the days when I'd carry around atleast 2 copies of files on seperate floppys). Second one worked. Hooray!
I got the BIOS loaded onto the motherboard successfully. The BIOS reinstall seems to cause the network device MAC address to change, which I correct. Networking then works fine in linux, but not in Windows. A dead network device is easily solved tho, so I get out a handy rtl8139too compatible PCI network card and stick it in, which works fine under both linux and Windows. Connectivity is restored.
A couple of days later windows crashes (I think I saw a flash of blue, so it may have been a bsod, but Windows annoyingly automatically reboots immediately - I've since rediscovered how to disable this). BEEEEP BEEEEP - black screen, my computer is screaming at me again (long beeps).
I reboot and check the even logs. The last entry is at the time of the crash and tells me that Windows has reached its TCP/IP connection limit. Odd. I certainly wasn't running anything network intensive (not gaming and torrents and email run on a separate box). This occurs twice more in the space of a week, even after uninstalling the nvidia network device and nvidia network allocator(?) from Device Manager (altho I'm not sure how much of the drivers would've been uninstalled, given that it's an nforce motherboard and thus still needs the rest of the nforce drivers).
Yesterday I decided to leave my PC running Linux to see if that had the same issues. Today I was browsing the web and viewing some videos - nothing intensive - and my PC suddenly dies (screen freezes, lock lights on keyboard won't toggle, ssh fails). I reboot and try to go back into linux. After selecting the entry in grub, the screen goes black and stays black.... Nothing. So I reboot and try again, same result.
Next I reboot and try my previous kernel. This works fine, so I mount /boot, cd /usr/src/linux and "make install" to reinstall the kernel (it's the right kernel). Reboot and try again. Same result - black screen, no kernel booting up. Another weird result!
I've just disabled the motherboard onboard network device in the BIOS setup (oops, maybe I should've tried that before) and linux now boots fine again.
BIOS temp alarms look fine. Cabling all looks fine, neat and tidy (no horrible thick IDE cables in the way \o/ ). The system has 1 SATA HDD (1TB).
This system is running through a UPS (APC model - lasted around 13 mins on my Athlon64 - haven't tried it with this system yet), so power fluctuations shouldn't be an issue. (And the UPS doesn't react in any way when the crash happens - no switching to battery even momentarily - it gives a definitive click when it does that so I usually notice - plus my speakers don't run through it, so I usually notice =D )
No metal filings or poorly stamped holes in the case.
Much blood has been sacrificed already in the creation of this computer - most of it mine. The Asus motherboard still has the blood stains to prove it (there's one on the side of the P/S2 connectors).
Am I just unlucky, or is there some common thread causing all these issues that I can't see? The only thing I haven't changed out is the PSU, which I unfortunately don't have a viable replacement for at the moment (it's a 400W with 20A rating on +12V, which a friend tells me won't be enough for the 800GTX - unfortunately without any onboard graphics on my mobo and this being my first PCI-E graphics card, I have no replacements I can use).
I've used AMD for years (since the K6-2) and I'm now beginning to regret not being patient enough to wait for the fixed version of the Phenom.
