More problems then you can shake a stick at

F

FiZi

n00b
Joined
Oct 2, 2004
Messages
25
My desktop:

CPU: AMD AthlonXP 2800+
Mobo: Soltek SL-75DRV5 (Bios 1.8T)
RAM: 1x Samsung DDR333 512MB, 1x Samsung DDR400, 1x Kingston DDR400
Video: ATi Radeon 8500DV All-In-Wonder
Audio: Audigy 2 Plat.
Hard Drivers: Maxtors 60gigs and 120gigs
PSU: Powerman 300w

I was playing Battlefield 2 today and during the game my system was restarting itself. I attributed this to a heating problem. I checked my temps in the BIOS and found the CPU was cooking itself at around 70C.

I took the computer outside and began cleaning. I removed the video card and cleaned it with compressed air, removed the fan and cleaned it as well from the card. I tried to remove the heatsink but it was stuck on. I then removed the IDE cables from the case and the FDD cable and finished blowing out all the dust from in there. Then I removed the heatsink and fan from the CPU. I took the fan off the heatsink and blew it out. I then took the heatsink and blew it our, ran it under hot water and wiped off all the thermal grease (Arctic Silver) and then dried it with a hair dryer making sure that there was no more water left. I then removed the CPU from the socket, blew out the socket, blew off the CPU and replace it on the motherboard. I then blew out the PSU. I put some Thermal Compound on the CPU Diode and then put the heatsink/fan back on it. I then took the IDE cables and played origami with them to fold them out of the way. So far so good right? I replaced the video card in the computer and closed everything up again. This is when things started going wrong.

I powered the computer up and the Windows XP screen would appear and then system would restart itself. I checked the BIOS and saw the CPU was a nice 80C and the BIOS is set to protect the computer at 85C. So my guess was the CPU would overheat trying to load Windows. I then took the case back off. Removed the heatsink/fan from the CPU and cleaned all the thermal paste off the heatsink again. I then used a knife to level out all the thermal paste on the diode and then replease the headsink/fan on the computer and started it up again. This time the BIOS read the CPU Diode as 45C, much better right? The computer started and I was back in Windows and happy. I started checking my e-mail and getting back onto mIRC. The computer was running for around 30min (during which Firefox and Thunderbird were randomly crashing) and then it restarted and didn't come back up again. I opened the case and took a look inside for anything I had changed. I removed one of the fan sensors and plugged my CPU fan (Volcano 7+) directly into the motherboard for power. Normally it uses a switch which runs through one of the larger 4pin power cables, the ones meant for HDDs and CD/DVD-Roms. The computer started up fine. I shut it down, put the CPU fan back onto the 4pin power cable and started the PC, the computer wouldn't get past the Windows XP boot screen. I then put it back directly on the mobo and the computer worked again. I figured the problem might be power. I unplugged the CD and DVD drivers from my system, put the CPU back on the 4pin and the computer still failed to start same as before. I put the CD/DVD back on and put the CPU back on the 4pin connector and now the whole system wouldn't start into Windows. I put in the Ultimate Boot CD and ran memtest86 and the ram came back clean. I then did the CPU Burn-in test and it kept looping through doing its thing but the keyboard wasn't responsive to end the program. Next I put the CPU fan on another 4 pin connector and the computer started right up. I started doing a few things on it and then it went out again and gave a long beep, a pause and another long beep and then i turned it off figuring the beeps were just going to repeat. I checked the Pheonix beep codes anyway and 1-1-x means either the bios needs to be replaced or the mobo is bad. I reflashed my bios with the latest verison and tried to boot the pc, no go. I couldn't get into Safe Mode either though ealier I had managed to get to Safe Mode with Command Prompt and do a chkdsk, the HDs are fine. On a random whim I removed the stick fo 256MB RAM from the mobo and the computer fired up fine. Since I'd been through this before I started opening up a bunch of programs to fill the ram and use the cpu to see what would happen, it loaded them all fine. I then downloaded 3DMark03 and set it to a loop and figured I would leave it all night to see if I'd fixed the problem. It got to the second test and then the system rebooted itself and came back up into Windows. I thought it could be the video card, so I started downloading PCMark05 to run just HDD and CPU benchmarks and during that process the system rebooted itself again.

I'm out of ideas. To me it could be the Video Card, CPU, Mobo, RAM or the PSU causing this problem. I've got a spare PSU at work I'm going to bring home tomorrow and give a shot to rule that out as the problem. I've also got a AGP Radeon 9200 at work I am going to bring home and pop into the machine if the PSU doesn't work to see if its the video card. I guess if those don't fix the problem its down to the mobo/ram/cpu. I guess I'll bring home a stick of ram from my system at work as well and see if it works. So that leaves the mobo and the cpu. I've got a spare Athlon XP 2600+ I can test to. So if after all that the computer still is having issues I guess its clear its the mobo. Does anyone else have any ideas or a clear answer to why this is happening? Before I took the computer out for its cleaning because of the heating problem today the system was stable.

Any feedback is much appreciated.


Edit: I just remebed one thing I changed. I put a USB Keyboard/Mouse on the system today. I just removed them and put the PS/2's back on. I ran 3Dmark03 again and the thing crashes now instead of causing the whole system to reboot.

Edit #2: For shits and giggles I tried 3DMark01 and it just crashed to desktop after 2 seconds of the first demo. Looking more and more like the video card right?

Edit #3: Tried to install PCMark05 and I got a few windows errors and then the computer restarted itself. I booted it up to post this and while starting explorer crashed and then came back. Maybe it isn't the video card?
 
As you indicated, this system is impossible to troubleshoot without reducing the number of variables.

Start off with just the mainboard placed on a non-conductive surface. Install the CPU, HSF, a single stick of RAM, the videocard, an external PC speaker if not onboard and the PSU. Reset the CMOS and turn on the system (by shorting the PWR_SW pins of the front I/O headers). You should be able to get into the BIOS now. Verify voltages (double check with DMM) and temperatures.

If things look good so far, add a single HDD (with the OS) and a NIC if not onboard. If you're able to successfully boot into Windows and run benchmarks and such without problems, start adding more components until things start going wrong again.
 
uhhh i dont know why might be the problem, but i can tell you, next time, dont use a knife to spread thermal goop on a processer :eek: :eek: credit card, or plastic baggy and card/finger





soulsaver_8229
 
You have a volcano 7+ CPU cooler... that might not have enough cooling capacity to shed the heat off of that 2800+. If all else fails, try a different CPU cooler as well...

EDIT: Also make sure that fan on your Volcano 7+ is spinning at its nominal speed. Might not be pushing enough air to cool properly.
 
Is the selector switch for that Volcano 7+ on low med or high? Try putting it on high if you haven't done so already. Its loud and annoying on high but it at least should be cooling your cpu.
 
mattgyver79 said:
Is the selector switch for that Volcano 7+ on low med or high? Try putting it on high if you haven't done so already. Its loud and annoying on high but it at least should be cooling your cpu.
Click to expand...

Yup when the CPU Fan was on the throttle for troubleshooting I had it on high.
 
l337zax said:
You have a volcano 7+ CPU cooler... that might not have enough cooling capacity to shed the heat off of that 2800+. If all else fails, try a different CPU cooler as well...

EDIT: Also make sure that fan on your Volcano 7+ is spinning at its nominal speed. Might not be pushing enough air to cool properly.
Click to expand...


It was on high and I've always had this cooler, it is good enough for atleast basic operation. I checked the BIOS and the temps are back to normal so I think I've managed to rule out a heating issue.
 
Socket7 said:
Run memtest86. Bad memory would account for your problems.
Click to expand...

Yup, I ran that one first because you're right, it seemed like a ram problem. RAM came back clean and working 100%. I'm at work right now, but I've got a spare PSU so I'm going to being going through the process that Elledan suggested. Thank you all though for your feedback so far. I will let you know how things turn out.
 
Consider that the proc was damaged by the heat and needs to be replaced...
 
Phoenix86 said:
Consider that the proc was damaged by the heat and needs to be replaced...
Click to expand...

Is there any clear way to see if the CPU was damaged by heat? Wouldn't the mobo or the cpu itself have scortch marks? Oh and I never saw or smelt smoke.
 
Well, good and bad news. I've got the new PSU in the system appears to be stable again. I was just running PCMark05 doing all the CPU/MEM/HDD tests and the system go through them all. I then ran 3DMark03 and it got through most of the tests but while I was eating the thing rebooted. I think I'm looking at a dual problem. Damaged PSU and a cooked video card. I'm going to swap out my Radeon with another we have here and see if I can get through 3DMark03. If I can then I know the answer to my problem.

Edit: Ok, so the new PSU seems to keep the system stable. I ran the PCMark05 tests twice and the system was gold. I then ran 3DMark01 on default settings and then with all the AA turned up and such and the system was fine. I then ran 3DMark05 and it either crashes to desktop after running for a bit or the system restarts. I think its pretty clear at this point the problem is my video card. I've got a spare identical card in another computer. I'm going to pop it in later and see how things go. All I'm hoping for now is the system will stay stable so I can turn it into a Linux server. No need for a good graphics card.
 
FiZi said:
Is there any clear way to see if the CPU was damaged by heat? Wouldn't the mobo or the cpu itself have scortch marks? Oh and I never saw or smelt smoke.
Click to expand...
Sometimes yes, sometimes, no on both counts.

Keeping the magic smoke in the CPU is good too. ;)

Good luck.
 
So I put a new video card into the system and the same problem occurs. It will run 3DMark01 or 03 for the most part and then crash to desktop or reboot the system. I think I'm looking at an OS issue now because I was able to defrag my HD last night as well as do a 1gig backup to another computer. The system stayed up all night, it rebooted itself in the morning when I went to run Spybot Search and Destroy. I'm going to dual-boot a fresh copy of XP and see what happens. If it fails then I will have to remove everything and try it with the bare minimum hardware installed and go from there.
 
soulsaver_8229 said:
uhhh i dont know why might be the problem, but i can tell you, next time, dont use a knife to spread thermal goop on a processer :eek: :eek: credit card, or plastic baggy and card/finger





soulsaver_8229
Click to expand...


If there is too much paste or improper paste application the processor will get very hot and not transfer heat properly. I saw this issue before. Computer reboots at random, very hot temps shown from processor. :(

 
Shmuckety said:
If there is too much paste or improper paste application the processor will get very hot and not transfer heat properly. I saw this issue before. Computer reboots at random, very hot temps shown from processor. :(

Click to expand...


The temps on the CPU are low right now and way below the max running temp for the CPU. Under load I'm looking at 50C CPU Diode sensor, 43C CPU Socket seonsor and 33C Case. The Bios is set to reboot the computer at 75C (lowest setting). So its not heat.

I'm thinking it might be an OS issue at this point.
 
FiZi said:
I'm thinking it might be an OS issue at this point.
Click to expand...
No OS issue I have seen will cause sporadic restarts, that points firmly at the hardware. Same with "hard" locks where the mouse cursor locks.

Your idea of doing a parallel install will test this while preserving your orignial OS, so give it a spin. I just don't have high expectations it'll fix it.
 
If your proc was running at 70C then the OS isn't the issue. There's something wrong with the thermal transfer with your CPU. Is the CPU fan running full speed? (Or just plain running?) Heat sink seated correctly?

The comment about using a knife to spread the paste is suggesting to me that you've got too much goop on there. We're talking just a thin, thin layer of paste. No more than 1mm thick. All the paste is supposed to do is make sure there's even contact between the copper sink and the CPU.

I mean, it's not like spreading peanut butter on bread. And it's not like putting in a thermostat gasket in your car. Extra goop insures that water doesn't leak, but that's not what you're trying to achieve. It's about the most efficient way to get heat away from the CPU. To much and you end up trapping heat.

What's really funny is that I had that exact same CPU and it never got hotter than 45C under full load. Check your paste.
 
Torgo said:
If your proc was running at 70C then the OS isn't the issue. There's something wrong with the thermal transfer with your CPU. Is the CPU fan running full speed? (Or just plain running?) Heat sink seated correctly?

The comment about using a knife to spread the paste is suggesting to me that you've got too much goop on there. We're talking just a thin, thin layer of paste. No more than 1mm thick. All the paste is supposed to do is make sure there's even contact between the copper sink and the CPU.

I mean, it's not like spreading peanut butter on bread. And it's not like putting in a thermostat gasket in your car. Extra goop insures that water doesn't leak, but that's not what you're trying to achieve. It's about the most efficient way to get heat away from the CPU. To much and you end up trapping heat.

What's really funny is that I had that exact same CPU and it never got hotter than 45C under full load. Check your paste.
Click to expand...


Turns out its a 2200+ (1.8Ghz). I've cleaned off the CPU and Heatsink. Re-applied my Arctic Silver to it using a Credit Card this time. I've removed all non essential hardware. Its running at 50-53C in the bios. I am now trying to install XP (Was getting a shit load of STOP errors when I was trying ealier). I think the mobo/cpu is hosed. I can't even get a XP install going anymore.

Edit: Yanked all the ram but a 256 and its starting the XP install now, it got a bit farthjer in the intial load of devices and then crapped out again. I've got the bios set to warn me at 60C so its not getting that hot.

Edit #2: Just tried Windows XP Install with out the HDs in and it BSOD'd. Even if the issue were the HD's missing, Windows XP says no hds found, not BSOD.

So, final answer mobo/cpu? Anyone else thing so?
 
Mobo. Have damn near the same thing with my 'puter, ruled out memory from memtest, proc heat was fine, booted Knoppix OK, and randomly crashed. I'm talking between 5 minutes and 5 hours. I'd seriously look at that mobo kinda funny, and definetely work on a replacement.
 
Ya, I'm thinking the PSU issue might have also damaged the Mobo. I'm ordering a new Asus board tomorrow. Can't wait to drop kick this fucking Soltek board. If it still doesn't work guess I gotta pickup a Sempron too.
 
Wonder if the voltage regulators failed on the Mobo. It could be sending extra current to the CPU causing extra heat, the RAM to have problems and what not. It could also stem from the PSU.

At this point getting a new mobo and PSU is the right choice.
 
Torgo said:
Wonder if the voltage regulators failed on the Mobo. It could be sending extra current to the CPU causing extra heat, the RAM to have problems and what not. It could also stem from the PSU.

At this point getting a new mobo and PSU is the right choice.
Click to expand...


Ah good. The PSU I put in it during the troubleshooting process is new so that takes care of half the problem. I'm ordering a new desktop in the next day or two anyway since I need something to play games on (Radeon 8500DV and AGP just ain't cutting it). I'm throwing in a fresh Asus Socket A motherboard with the new system. If I can get the old one back up and running I'll make it my new linux server.
 
You must log in or register to reply here.
Back
Top