PDA

View Full Version : Bad motherboard or CPU?


Pages : [1] 2 3

CaptNKILL
03-18-10, 02:04 AM
I can't believe my luck. :\

I'm building a system for someone and something is seriously wrong with it.

The full specs are here:
http://www.nvnews.net/vbulletin/showpost.php?p=2204086&postcount=18

Basically, I'm having a problem with stability. It keeps freaking crashing! :headexplode:

I thought it was a fluke the first time it did it. It just rebooted itself right after I turned it on and when I switched my monitor over to it it was at the "windows did not shut down properly..." screen.

It got through the entire windows install process, it was totally up to date, everything was working great, no problems. I ran the OCCT stress tester to ensure that the CPU was stable and it ran fine for an hour with no errors and then it sat idle in windows for a couple hours. The next time I booted it the following day was the first time it rebooted on its own. It did it again tonight too. There are critical errors listed in the event log for both of these occurances. The system is just going dead.

The power is still on, the fans are still on, but the system just dies. I did notice once yesterday that the system sort of "froze" for several seconds after reaching the desktop and opening the control panel. I've never seen a system freeze like this running Windows Vista or 7 with multiple CPU cores and plenty of RAM. It was like windows XP on a single core, where everything just stops (cursor included) and then starts back up again several seconds later.

Now, if I run OCCT, it says "error detected on core 0" within a few minutes, or else the system just dies completely (with the power still on and the fans still running).

Now, I have it all stripped down to the basics. No cards, drives, extra fans or front panel (USB\audio) plugs are attached. The motherboard is still in the case though and I'm using onboard video (it was disabled before). Its actually freezing in the BIOS now. I've tried both sticks of ram individually and in both memory slots (there are only two). The system speaker will give a tiny click and then the system will be unresponsive. Strangely, the decorative colored bar at the top of the bios screen is still animating.

Something interesting just happened though. I was on the health monitoring screen waiting for it to freeze and it sat there normally for several minutes. When I started pressing keys the speaker let out a click, the arrow keys stopped moving the cursor and the fan speed monitor switched from 3000RPM to "NA" as if the monitoring died before the system froze. Its still sitting there in the BIOS right now, everything is frozen but the bar at the top of the screen is still animating.

I'm at a loss.

Does this sound like a defective motherboard or a defective CPU? I'm leaning toward the CPU, as uncommon as that might be. I have an older Athlon X2 5050e running on a 780G in my HTPC in the living room, so I guess I can swap CPUs tomorrow to see if the errors follow the motherboard or the CPU. I really hope its the CPU honestly. That'd be much easier. Though I think I want to reinstall windows on this thing regardless, since Microsoft Security Essentials seems to have been broken by one of the crashes\reboots. I'm sure other things have been borked as well.

AthlonXP1800
03-18-10, 03:26 AM
I think it the motherboard, why buy Foxconn? Try flash latest motherboard BIOS and also test the CPU on other motherboard see if CPU is faulty.

CaptNKILL
03-18-10, 12:45 PM
I think it the motherboard, why buy Foxconn? Try flash latest motherboard BIOS and also test the CPU on other motherboard see if CPU is faulty.

This board was highly rated at newegg, and everywhere else I looked. It should be pretty solid.

It already has the latest BIOS as well.

Imbroglio
03-18-10, 01:00 PM
trust me when i say, take the motherboard out of the case and assemble it on a wood block or wood table (anything that isn't conductive) and then start it up. I was putting together a system for a friend and the back panel i/o shield would cause a short from time to time... same exact symptoms as to what you are describing. after getting thoroughly pissed, i took everything out of the case and started with the basics, then after seeing it ran, added a component one by one and then finally put it back in the case (that's when I figured out what it was). oh and good luck... you'll need it :D

CaptNKILL
03-18-10, 01:14 PM
trust me when i say, take the motherboard out of the case and assemble it on a wood block or wood table (anything that isn't conductive) and then start it up. I was putting together a system for a friend and the back panel i/o shield would cause a short from time to time... same exact symptoms as to what you are describing. after getting thoroughly pissed, i took everything out of the case and started with the basics, then after seeing it ran, added a component one by one and then finally put it back in the case (that's when I figured out what it was). oh and good luck... you'll need it :D

Yeah, this is the next step after I test the CPUs.

I'm updating the BIOS on my HTPC so it'll support the new CPU and then I'll throw it in that system. I'll try my old CPU in the other board while that's testing. If I get any more errors on the new computer I'll know its either the motherboard or a short (or the power supply, or the alignment of the planets :p).

Maverick123w
03-18-10, 01:21 PM
I had a similar problem with a foxconn build here in the shop. Ended up being a defective socket. Sometimes I could get it to run occt for hours, other times it would freeze within minutes.

Try swapping out the chip, that will tell you immediately.

CaptNKILL
03-18-10, 01:41 PM
I have the X3 in my other system now and everything seems to be fine.

I just took the board out of the new system and I've got it sitting on a box next to the case. I put my 5050e in there and it still does this strange "click click click" from the PC speaker as soon as it starts up, but so far its working. I'm just going to wait for now. If they both run like this for a while then I'll put the X3 back in this one with the board outside the case and continue testing that way. This is a really inexpensive case so its possible that something was shorting out.

I hope that's all it is. I can take care of that one way or another. Having to RMA a board for a computer that I said would be done this week certainly sucks though. I hate looking unprofessional with my work, and delays caused by defective components definitely don't help.

nekrosoft13
03-18-10, 02:53 PM
are you usre you didn't try to unlock the 4th core? double check your bios if there is anything related to unlocking 4th core on x3 processors enabled.

CaptNKILL
03-18-10, 05:11 PM
are you usre you didn't try to unlock the 4th core? double check your bios if there is anything related to unlocking 4th core on x3 processors enabled.

Its all set to defaults.

I'm testing it again now with the X3 installed and the motherboard outside the case. I have the hard drive plugged in and Windows seems to be pretty screwed up from the random crashes yesterday (most likely during windows update reboots). Its saying Windows is no longer genuine, but it was yesterday. :lol:

The strangest thing though is that OCCT keeps erroring at EXACTLY 1 hour, 0 minutes and 0 seconds. Its done it twice in a row now. I have it set to "infinite" test length, but it still just stops with an error on Core #1 or Core #2. I've switched to Prime95 with 3 threads to see if I can get a more reliable test. Nothing seems to be reporting temperatures correctly, but that isn't uncommon really.

The major problems seem to be gone though. It must have been shorting out on the case somewhere. Now I just have to figure out where, then fix it, and then reinstall windows all over again. :headexplode:

CaptNKILL
03-18-10, 06:46 PM
Well, its been running Prime95 with all three cores maxed for 2 hours without any errors.

The system should be alright, but I'm going to keep testing the bejesus out of this thing.

CaptNKILL
03-18-10, 09:45 PM
Ok, it ran Prime95 for about 3 1/2 hours with no errors.

I just inspected the case and found that there was an extra motherboard standoff screw where one wasn't needed (BAH!) and there was a tiny scratch on the motherboard tray near the back panel, which leads me to believe something must have been too close there. The tray is a bit flexible so I just pushed it back a little and removed the standoff. I put the board and hard drive back into the case and I'm running Prime95 again. Hopefully it stays working. I spent ALL day doing this crap.

nekrosoft13
03-18-10, 10:02 PM
great job capt, had that happen once, extra standoff

CaptNKILL
03-18-10, 10:35 PM
So far so good. Its still working.

I just hope Microsoft didn't do anything retarded with my Windows 7 Pro key. I think the registry got corrupted somehow during one of the crashes (it kept happening right at startup, during windows update installation most likely) and it keeps pestering me about windows not being genuine. I just hope it didn't send that back to the mother ship to invalidate my key or anything like that. I'll be PISSED.

CaptNKILL
03-19-10, 02:58 AM
Prime95 ran for 5 1/2 hours with no errors. Time to reinstall windows. :)

I kept all of the drivers and programs I downloaded the last time so it should be a pretty quick and easy process this time, assuming the motherboard doesn't decide to go tits up again for no reason.

CaptNKILL
03-19-10, 04:18 AM
Grrrr.... :headexplode:

I'm getting ntfs.sys bluescreens. These are probably what was causing windows to crash before, I just never saw the bluescreens.

I think I may have had two different problems going on here. Probably a bad SATA cable or something. The new board only came with one so I had to use an old one for the CDROM drive. I don't even know if I have any more...

Maverick123w
03-19-10, 12:51 PM
Have you tested the ram?

CaptNKILL
03-19-10, 01:11 PM
Have you tested the ram?

Yeah, but I think I'm going to do it again.

I can't believe the customer hasn't called yet. They gave me the "down payment" on monday of last week and I said it should be ready by the end of this week.

Maverick123w
03-19-10, 01:42 PM
In all my experiences calling the customer and letting them know what is going on is much better than just praying they don't call. People are much more chill about stuff when they have been kept in the loop.

CaptNKILL
03-19-10, 01:48 PM
In all my experiences calling the customer and letting them know what is going on is much better than just praying they don't call. People are much more chill about stuff when they have been kept in the loop.

Yeah I'm going to call him today, I was just hoping to figure out exactly what the problem was first.

I just ran memtest 86+ and it actually came up with some errors after about 15 minutes. I don't know if the boards "auto" settings were correct for this memory though, since I've read about some issues with AMD boards in that respect.

I just downloaded a newer version of the Ultimate Boot CD so that I can get an up to date version of memtest 86+, so I'll try that again with the speed and timings set exactly as they should be in the BIOS.

Maverick123w
03-19-10, 02:08 PM
Keep us posted.

CaptNKILL
03-19-10, 02:19 PM
Ok, more errors with the memory set at the correct speed and timings, with memtest 86+ 4.0 reporting all the correct specs.

So, the RAM is most likely defective. GREAT. :headexplode:

I just put in some old crappy DDR2 533 that I had laying around (2x256Mb LOL) to run memtest again, just to varify that it isn't a problem with the memory controller or motherboard or something screwy like that.

EDIT: Actually, I just put some other ram in there (from another system that's waiting to be finished) since I recently ran memtest on THAT for 6 hours with no errors. If it errors in this system I know there are other problems.

Capt. Picard
03-19-10, 02:21 PM
In my African experience it most often is the motherboard that causes the problem.

I've got a problem with the fact that so many utilities and functions are integrated into the motherboard. South bridge, North bridge, Sound, PS2, USB, Firewire, BIOS, LAN, MEMORY ...

I would much rather buy components that are more expensive initially, as a whole, but that separates component failures to smaller units.

The real problem is the cascading effect componnent failures have on the whole system.

snowmanwithahat
03-19-10, 02:29 PM
Have you tested the ram?

Sounds likely.... I had very similar issues with mine making it look like it could have been the board or cpu but in the end it was definetly the memory.

Pull out the RAM and run memtest86 with them alone instead of doing it all at once, it helps narrow it down.

The reason I say memory is because you're having corrupt files, that usually points to a corrupt memory stick / slot

logan
03-19-10, 02:46 PM
It could also be incompatible ram, but that's probably less likely... I remember once pulling known working/tested memory from a server at work for my (awful, awful) work desktop and it started flipping out before Linux even finished booting, and I'm not even sure that memtest86+ even complained once. Every other configuration I tried was fine. 512M*2, 1G+512M, but once those 2*1GB sticks hit the motherboard it was over. It just happened to be a picky motherboard. I tried another brand and eventually managed to get that thing up to 2GB of ram.

CaptNKILL
03-19-10, 03:00 PM
It could also be incompatible ram, but that's probably less likely... I remember once pulling known working/tested memory from a server at work for my (awful, awful) work desktop and it started flipping out before Linux even finished booting, and I'm not even sure that memtest86+ even complained once. Every other configuration I tried was fine. 512M*2, 1G+512M, but once those 2*1GB sticks hit the motherboard it was over. It just happened to be a picky motherboard. I tried another brand and eventually managed to get that thing up to 2GB of ram.

It may have been one of those deals with "high density" memory. I remember that being a big issue toward the end of the DDR400 era.