Sorry about how long this is, I typed out the stuff below on the AMD Forums and you'll probably only be able to offer any useful ideas if you know the whole bizzare story. Thanks for any ideas you may have (I'm really quite baffled at the moment).
Cliffs:
-Building new system, boots to BIOS on first try
-Unplugged overnight, won't boot in morning (no beep, no POST, undisturbed overnight)
-Starts working again in evening, for a few hours
-RMA mobo back to Newegg, same problem
-RMA processor, same problem
-Test everything except mobo, CPU, and RAM (nobody else I know has AM2 or DDR2 to test with/against)
-Probably need to RMA mobo, CPU, or both, but not sure which is the problem
-HELP!!!!
Ok, here's the full story (copied from this thread I started at the AMD Forums: AMD Forums Thread):
I decided to upgrade to X2 when my Athlon64 3000's mobo died a month ago. Bought a Gigabyte M55SLI-S4, Athlon64 X2 3800, PNY 7600GT PCI-E, Zalman 9500, and a gig of Corsair XMS DDR2-800. I put it all together and the system booted to POST (I wiped the HDD, so no OS) on the first try. I explored the BIOS for a couple minutes (not changing things, just exploring) and shut it down because a storm was coming in and I didn't have it on a surge protector. I unplugged it for the night and when I got up and turned it on, the fans would start but there was no POST or beeps. Tried it again throughout the day when I had free time and same thing (tried RAM one stick at a time, none, a PCI video card, the usual things). Finally that evening it decided to boot, even though I didn't do anything differently to it. Installed XP and as I was working on it (rebooting as normal when installing drivers), it decided to stop working again (same symptoms).
I contacted Gigabyte and followed their instructions for how to test it (did everything except an isolation test outside the case) and concluded I had either a bad mobo or a bad CPU. RMAed the mobo and same thing. RMAed the processor (process of elimination, right?) ...and... same thing. Tried a friend's 500W PSU (known good, and I'm pretty sure my original 420W is still good too), same thing. Reseated everything (and yes, I do have both power connectors on the mobo plugged in), CPU is definitely oriented the right way and no bent pins (tightly clamped on by the Zalman heat sink), tested a different video card (both a PCI one and a borrowed PCI-E that is known good), and same problem.
When I did the isolation test (outside the case on a non-conductive surface with just mobo, CPU, 1 stick RAM, video card, and case speaker/power switch), it didn't work at first, but decided to work when I propped up the corner of the board that held my video card (didn't want to sit properly on floor because of the part of the card that sticks down). So I started putting it back in the case and started putting in components one at a time, testing along the way. For a while, it would only start when I was reaching in and lifting the upper-right corner of the board (by the RAM, flexing the mobo slightly), so I thought the board was shorting on a spare metal standoff pin (it didn't line up with any of the holes drilled in the case and board, but I left it in because the entire upper-right edge had no support otherwise and would flex whenever pressure was applied to seat RAM or connectors), so I took it out and was careful when connecting things. Seemed to start working then, so I put my components in and stood up the case while it was running to see if something would go wrong as the weight moved going to vertical. Still worked. As it was in Windows, I gently shook the case to see if it moving around would mess it up and nothing changed. Rebooted a few times, then decided to go to bed and work on it in the morning. In order to reproduce the original conditions that it stopped working, I shut it off and unplugged it over night.
When I got up and plugged it in, it wasn't working anymore (no POST, no beep). Keep in mind that this computer was sitting by itself, undisturbed and unplugged overnight and somehow decided to stop working in the meantime. Do I have another bad Mobo or CPU (could I have gotten two defective ones in a row?), if so, how do I tell (nobody I know has AM2, so I can't test them individually). My theories:
- Bad CMOS battery dying while unplugged overnight? (but I'm guessing it would just go to default settings, not stop POSTing)
- Bad CPU? (two defective ones in a row seems unlikely, especially because they have different stepping (CDB4F 0619UPAW vs. LDB4F 0622RPMW), although a friend who works with computers says he's had similar problems with AMD CPUs in the last week or two, so it could be a bad batch).
- Bad mobo? (seems most likely, because of the way it only wanted to work when flexed slightly (lifting up upper-right corner or putting something down by video card), but that could also be changing how the CPU is sitting in it's socket, but that doesn't explain the randomness of when it decides to stop working when sitting by itself and undisturbed).
Any ideas? Should I just RMA the mobo and the CPU together (Newegg is gonna be sick of me before long)? I've only got two more days that I can RMA the CPU, so I've got to decide fast (hate the 7-day policy they have).
Quote from reply by "Fizzbang":
Thanks Fizzbang. Yeah, I tried each of them individually and with none at all (although I wasn't able to try them on another machine or get any other DDR2, I don't know anybody using AM2 or DDR2). I could be wrong, but it seems that I'd be getting some response out of the computer if it were actually a RAM problem.
Oddly enough, it's decided to start working again after sitting by itself for the day. So...the only known cases of this thing working are in the evening/night, even when the BIOS has been reset and has no idea what the real time is. Maybe it needs to be at a certain temperature to work? (I know the electronics have a minimum temperature, but that's usually well below freezing and Indiana isn't nearly that cold at this time of year). Very strange, very strange indeed.
I guess my next step might be to flash the BIOS to the latest version (which I think I can do because it's working at the moment. If I'm reading the POST screen right, it's got version F3 and the latest version on the Gigabyte site is F4 (although they don't have F3 available for download and the change between F2 and F4 is "Improve CPU fan regulation method "). My other option is to leave it running overnight and see what happens if I try to reboot it in the morning.
Update (8/21/06 11:00pm):
It's still working, at least for the moment. I re-installed Windows (with one BSOD when installing the latest nForce4 drivers) and it seems to be working nicely. Strange, when it works, it seems to work pretty well, but there seems to be no rhyme or reason to when it feels like working.
Again, it would appear the test will be what happens tonight. My three options are to shut it off and unplug it (as in the past, likely to make it stop working), shut it off and not unplug it (leaving the PSU switch on), or leaving it on overnight. I'm inclined to try the middle one tonight and see if it'll power up in the morning. If it does, that would point to the CMOS battery as being bad. If it doesn't, then either the mobo or the processor is likely defective (I'm leaning slightly towards the mobo, but I'm not sure).
Should I go through the hassle and shipping costs of RMAing both the mobo and the CPU at this point or just the motherboard?
Update (again)(8/21/06 1:15am):
Well, scratch that idea (shutting it down and leaving it plugged in). It idling in Windows (default screen saver) while I was getting ready for bed (wanted to leave a few minutes for you guys to reply before I shut it down) and it spontaneously rebooted. I wasn't watching it, so I don't know if it actually shut down or just went straight to black, but it started trying to boot, got to the point of detecting IDE devices, and then starting the boot process over again in an (seemingly) endless loop. When I turned it off and back on to try to break the loop, it went back to the original symptoms (fans on, no beeps or POST). Absolutely no sign what could have caused it to spontaneously crash like that. I guess I'll leave it plugged in when I turn it off just to see what happens, but I definitely am going to need to RMA something (the only question is which part, or both).
Cliffs:
-Building new system, boots to BIOS on first try
-Unplugged overnight, won't boot in morning (no beep, no POST, undisturbed overnight)
-Starts working again in evening, for a few hours
-RMA mobo back to Newegg, same problem
-RMA processor, same problem
-Test everything except mobo, CPU, and RAM (nobody else I know has AM2 or DDR2 to test with/against)
-Probably need to RMA mobo, CPU, or both, but not sure which is the problem
-HELP!!!!
Ok, here's the full story (copied from this thread I started at the AMD Forums: AMD Forums Thread):
I decided to upgrade to X2 when my Athlon64 3000's mobo died a month ago. Bought a Gigabyte M55SLI-S4, Athlon64 X2 3800, PNY 7600GT PCI-E, Zalman 9500, and a gig of Corsair XMS DDR2-800. I put it all together and the system booted to POST (I wiped the HDD, so no OS) on the first try. I explored the BIOS for a couple minutes (not changing things, just exploring) and shut it down because a storm was coming in and I didn't have it on a surge protector. I unplugged it for the night and when I got up and turned it on, the fans would start but there was no POST or beeps. Tried it again throughout the day when I had free time and same thing (tried RAM one stick at a time, none, a PCI video card, the usual things). Finally that evening it decided to boot, even though I didn't do anything differently to it. Installed XP and as I was working on it (rebooting as normal when installing drivers), it decided to stop working again (same symptoms).
I contacted Gigabyte and followed their instructions for how to test it (did everything except an isolation test outside the case) and concluded I had either a bad mobo or a bad CPU. RMAed the mobo and same thing. RMAed the processor (process of elimination, right?) ...and... same thing. Tried a friend's 500W PSU (known good, and I'm pretty sure my original 420W is still good too), same thing. Reseated everything (and yes, I do have both power connectors on the mobo plugged in), CPU is definitely oriented the right way and no bent pins (tightly clamped on by the Zalman heat sink), tested a different video card (both a PCI one and a borrowed PCI-E that is known good), and same problem.
When I did the isolation test (outside the case on a non-conductive surface with just mobo, CPU, 1 stick RAM, video card, and case speaker/power switch), it didn't work at first, but decided to work when I propped up the corner of the board that held my video card (didn't want to sit properly on floor because of the part of the card that sticks down). So I started putting it back in the case and started putting in components one at a time, testing along the way. For a while, it would only start when I was reaching in and lifting the upper-right corner of the board (by the RAM, flexing the mobo slightly), so I thought the board was shorting on a spare metal standoff pin (it didn't line up with any of the holes drilled in the case and board, but I left it in because the entire upper-right edge had no support otherwise and would flex whenever pressure was applied to seat RAM or connectors), so I took it out and was careful when connecting things. Seemed to start working then, so I put my components in and stood up the case while it was running to see if something would go wrong as the weight moved going to vertical. Still worked. As it was in Windows, I gently shook the case to see if it moving around would mess it up and nothing changed. Rebooted a few times, then decided to go to bed and work on it in the morning. In order to reproduce the original conditions that it stopped working, I shut it off and unplugged it over night.
When I got up and plugged it in, it wasn't working anymore (no POST, no beep). Keep in mind that this computer was sitting by itself, undisturbed and unplugged overnight and somehow decided to stop working in the meantime. Do I have another bad Mobo or CPU (could I have gotten two defective ones in a row?), if so, how do I tell (nobody I know has AM2, so I can't test them individually). My theories:
- Bad CMOS battery dying while unplugged overnight? (but I'm guessing it would just go to default settings, not stop POSTing)
- Bad CPU? (two defective ones in a row seems unlikely, especially because they have different stepping (CDB4F 0619UPAW vs. LDB4F 0622RPMW), although a friend who works with computers says he's had similar problems with AMD CPUs in the last week or two, so it could be a bad batch).
- Bad mobo? (seems most likely, because of the way it only wanted to work when flexed slightly (lifting up upper-right corner or putting something down by video card), but that could also be changing how the CPU is sitting in it's socket, but that doesn't explain the randomness of when it decides to stop working when sitting by itself and undisturbed).
Any ideas? Should I just RMA the mobo and the CPU together (Newegg is gonna be sick of me before long)? I've only got two more days that I can RMA the CPU, so I've got to decide fast (hate the 7-day policy they have).
Quote from reply by "Fizzbang":
Thanks Fizzbang. Yeah, I tried each of them individually and with none at all (although I wasn't able to try them on another machine or get any other DDR2, I don't know anybody using AM2 or DDR2). I could be wrong, but it seems that I'd be getting some response out of the computer if it were actually a RAM problem.
Oddly enough, it's decided to start working again after sitting by itself for the day. So...the only known cases of this thing working are in the evening/night, even when the BIOS has been reset and has no idea what the real time is. Maybe it needs to be at a certain temperature to work? (I know the electronics have a minimum temperature, but that's usually well below freezing and Indiana isn't nearly that cold at this time of year). Very strange, very strange indeed.
I guess my next step might be to flash the BIOS to the latest version (which I think I can do because it's working at the moment. If I'm reading the POST screen right, it's got version F3 and the latest version on the Gigabyte site is F4 (although they don't have F3 available for download and the change between F2 and F4 is "Improve CPU fan regulation method "). My other option is to leave it running overnight and see what happens if I try to reboot it in the morning.
Update (8/21/06 11:00pm):
It's still working, at least for the moment. I re-installed Windows (with one BSOD when installing the latest nForce4 drivers) and it seems to be working nicely. Strange, when it works, it seems to work pretty well, but there seems to be no rhyme or reason to when it feels like working.
Again, it would appear the test will be what happens tonight. My three options are to shut it off and unplug it (as in the past, likely to make it stop working), shut it off and not unplug it (leaving the PSU switch on), or leaving it on overnight. I'm inclined to try the middle one tonight and see if it'll power up in the morning. If it does, that would point to the CMOS battery as being bad. If it doesn't, then either the mobo or the processor is likely defective (I'm leaning slightly towards the mobo, but I'm not sure).
Should I go through the hassle and shipping costs of RMAing both the mobo and the CPU at this point or just the motherboard?
Update (again)(8/21/06 1:15am):
Well, scratch that idea (shutting it down and leaving it plugged in). It idling in Windows (default screen saver) while I was getting ready for bed (wanted to leave a few minutes for you guys to reply before I shut it down) and it spontaneously rebooted. I wasn't watching it, so I don't know if it actually shut down or just went straight to black, but it started trying to boot, got to the point of detecting IDE devices, and then starting the boot process over again in an (seemingly) endless loop. When I turned it off and back on to try to break the loop, it went back to the original symptoms (fans on, no beeps or POST). Absolutely no sign what could have caused it to spontaneously crash like that. I guess I'll leave it plugged in when I turn it off just to see what happens, but I definitely am going to need to RMA something (the only question is which part, or both).
Facebook
Twitter