What keeps killing my CPUs?

[PsychoCrazy]

OK. They're not actually my CPUs. A friend of mine brought me his computer because it just shut off while he was working on it, and it wouldn't come back on. I tested the CPU (Duron 700) and determined it was bad, so we ordered a replacement. The motherboard appeared to be OK because I popped a 1.2Ghz Athlon in it and it worked fine. We ordered a Duron 950 as a replacement, and I installed it for him. It seemed to work fine for a couple of weeks. The other day though, it locked up and he couldn't shut it down. He cut the power off on his surge protector and then tried to turn it back on, but he got nothing. I tested the CPU and it has gone bad.

What could be causing the CPUs to die? Would it be the power supply (250watt)? Would it be the motherboard (MSI K7T-Pro2 {KT133})? I can't understand why it would be running fine and then all of a sudden kill the CPU. I am going to get a replacement CPU, but I'm afraid to put it back into the system again if it's going to die again. Help?!

 

[ttn1]

You probably need to give some more specifics. How did you determine that the CPUs were "dead".

Have you monitored the temperature of the CPUs during operation? My first guess would be cooked processors. I highly doubt that a power supply would kill CPUs, except in a case of catastrophic failure. There are power regulators on the motherboard to prevent this.

 

[PsychoCrazy]

In order to determine the CPUs were bad, I tried them in another computer. I have an Iwill KK-266 with Athlon 1.2Ghz at work. I put the Duron CPU into the machine at work, and it wouldn't boot. The drives would spin up to speed, but nothing would ever post--the same behavior as in my friend's machine. But if I put the Athlon into my friend's machine, then it would boot fine. That's why I believe the CPU has died.

As for the temperature, when I first put in the 950, I monitored the temperature for a while. It seemed fine--it was less than the temperature on the Athlon in the other computer. I think it was running around 110°F. The heatsink/fan combo I put on it is supposed to be rated for up to 1.2Ghz CPUs. Here's a link to it: Heatsink/CPU. I even used Artic Silver II on it.

If the processers got cooked, would there be any evidence? Neither the original Duron 700 nor the more recent Duron 950 showed any evidence of being cooked. Is there something in particular to look for to find out if it got cooked?

 

[ttn1]

I believe arctic silver is conductive, and the resistors on top of those CPUs could have issue with that if it melted and flowed between them. This is just a wild shot in the dark.

I have never seen a CPU go bad, so I am just skeptical when I here about people "killing" them. Processors are actually quite tough to kill, when using the proper procedures. I have heard of AMD cpus getting cracked cores and frying, but usually this occurs when people are careless or in a hurry.

If the CPU temps are below 130F then I would say that you didn't fry it. Usually cpus only fry when they are run without a heatsink. There would probably be discoloration of the chip/core.

I would suspect the motherboard before I would think it was the power supply. Usually if it was the power supply, it would be unstable or just not run at all. I have never heard of an underpowered supply killing a CPU.

 

[PsychoCrazy]

ttn 1,

Thanks for the suggestions. There's no traces of artic silver anywhere else on the CPU except the core. I've never had a CPU go bad on me before, so I know they're usually pretty tough. The core isn't cracked at all, and I don't see any discoloration.

It almost seems like the chip somehow just got shorted out or something. If I leave it out of the motherboard and turn the computer on, nothing at all happens. But if the chip is in, the fans come on and the hard drive spins up, but nothing ever posts. It's like the motherboard can tell it's there, but the CPU just won't post.

I guess for the time being, we'll plan on ordering a new motherboard.

 

[ZZZZZZ]

Could be there is a bad regulator on the MB which isn't doing its job and providing overvoltage to the cpu which will eventially kill the cpu