Computer freezes during different stages of the boot process

[gsbergman]

I'm using PC Doctor 6 and troubleshooting a problem with a Gigabyte motherboard GA-K8N51GMF-9 with and AMD Athlon 3200+ processor. The system worked well for years until about a week ago.

The problem started as a video issue with the video misbehaving and causing the system to lockup. I bypassed the internal graphics card and went to an external graphics card. The system's video is now stable but the computer is still freezing. I next replaced the RAM thinking that the internal graphic card used the system RAM and that seemed logical. No dice, it did not fix the problem. I have run the motherboard diagnostics and CPU diagnostics from PC Doctor 6 and both pass multiple times (varies) then will randomly lockup. I have yet been able to keep the machine up long enough to finish the RAM tests. I did try using MemTest+ and was able to complete that test successfully before a lockup.

I've had it lock up during the POST but the lockups vary anywhere between POST and full up to Windows XP, and anywhere in between.

Swapped out RAM to different sockets with no luck. I also pulled the CPU fan and cleaned off the old thermal compound and replaced with fresh compound.

I'm puzzled because I've used 3 different sets of RAM, including a brand new stick, with the same results.

PC doctor 6 passes the motherboard and the CPU at various times with no failures but then randomly locks up. I've reset the BIOS to the "Fail-Safe" settings and have also pulled the battery, with no luck.

I've stripped the motherboard from all peripherals and have even pulled the motherboard out of the system and had benchtested it "Bare-Bones" with nothing attached running the test from RAM Disk instead of a hard drive.

I've used a brand new 400w power supply with the same results.

Short of tossing the motherboard and CPU, does anyone have any suggestions to why PC Doctor passes the motherboard and CPU, and new RAM and a power supply don't seem to fix the issues.

I wanted to reflash the BIOS but cant seem to keep the machine up long enough to get that far.

I'm very puzzled? Please help.

Thank you.
Gregg

[fwilson]

gsbergman,

It sounds as if you may be having a thermal issue. If a dust bunny had taken up residence in the heat sink it is possible the thermal compound may have overheated and become ineffective.

Try removing the CPU heat sink, scraping the old thermal paste off of the CPU and heat sink, reapply new, high quality compound and retest.

I have also had success reviving systems that act as yours does by performing this same procedure on the Chipset heatsinks.  Note:  Only do this on chipset heatsinks with a mechanical attachment to the motherboard not the glued onto the chip type

-Fred

Note:  Only do the chipset heatsink fix on systems with a mechanical attachment from the chipset heatsink to the motherboard NOT the glued onto the chip type.  I wouldn't mess with the glued on type as more often than not the surface mount chip will come off the motherboard rather than the heatsink off the chip.

[gsbergman]

Guess what, your right it is thermal.

But it's thermal too cold. I heated it with a hair dryer and got it up to a temperature where it works fine.

What is likely running too cold?

Gregg

[fwilson]

gsbergman,

This is very rare but I have seen it before, but only twice.  One system I got working again and the other is in that great recycle bin in the sky...

I makes sense when you think about thermal expansion and contraction.  Motherboards, chips and sockets are all made of different materials with different thermal expansion characteristics.  If something is on the ragged edge with respect to making electrical contact, when one component expands faster or slower than another this can make or break a contact.

I am hoping this is the case with your system.  I would remove all components and reseat everything that has a socket on it, including the CPU.  Do the thermal compound replacement also.  Hopefully this will resore it to working order.

The other possibility is that a lead on a surface mount chip is lifting.  With above average soldering skills, this can be repaired, the hardest thing is to find it.

At some point a decision on whether it is cheaper to replace the motherboard or repair it has to be made but then again there is also the satisfaction of fixing something most people can't.

Good catch and good luck with the repair.

-Fred