Okay, I'm fairly experienced at troubleshooting, but this one has stumped me. A friend of mine has a computer that randomly turns off now and then. It did not start doing this until after a power outage a few weeks ago. He has everything on a surge protector. He first noticed it while playing games, so the first thing I decided to do was see if the video card was the culprit. I removed the video card and used the on-board video to run things, random shutdown still occured. From this, I ruled out that it wasn't the video card. I thought it might be a software problem, but I also experienced this random shutting off while I was in the bios, and while doing my next test, the memory. I ran the latest memtest86 and got a few weird behaviors. First, memtest went to an orange screen and there was hex everywhere. I was still able to access the menu in memtest, so I restarted and it seemed to go okay. Then, I got an error, a Gen Prot, which I deduce is a general protection fault. It told me it recieved an unexpected interrupt and had to halt. I restarted with memtest again and then it randomly shutoff while doing the tests. This is all in the first five minutes of running it, not hours into it or anything. So, I'm wondering if it is a memory problem, and I'm sure I've ruled out software seeing as it crashed in both the BIOS and memtest. I don't have a spare PSU to try in the machine, or another stick of ram. I figure it has to be the mobo, processor, psu or ram though, as the vid card is the only other hardware in the system that might be involved in the process of accessing the bios and I've already ruled it out. A few other things, one, sometimes, after a crash, we'll try to start it up again to try something new, but when we press the power button, nothing happens. I have to switch the PSU on the back to cut power to the whole system, then switch it back on before the power switch will work again. Second, his CPU temps are quite high, like 100 degrees celcius and I don't really know why. I think I need to reseat his HSF or something, as there is plenty of ventilation through the case. I'm afraid there may be an improper amount of thermal compound on it, maybe too little, I'm not sure. Any ideas or advice would be appreciated.

I had one that made a click and shut down as fast as a light goes out when the switch is thrown to the off position. The machine wasn't worth fixing so I tried another power supply and it did the same thing. The machine wasn't worth fixing so I didn't persue it any further but it sounded to me like it was a short in the mother board.

A hundred degrees celcius??? If that's true, something has to be done about it! Do yourself a favor BACKUP! Sorry, I do not check for private messages

Something may already have been done about it.

I hope it was done quickly before it burnt down the house. :-) But I would certainly check the heatsink to CPU installation and the thermal compound. Too much compound can be a problem. It's kinda hard to get too little if applied correctly. It's job is just to fill the microscopic imperfections in the surfaces. Do yourself a favor BACKUP! Sorry, I do not check for private messages

Bah! Bloomin crikey, I didn't even notice that in the preview. I meant Farenheit, not Celsius. Even then, is that normal? Seems a bit high to me as I've never noticed my own temps going that high. Maybe mine only monitors ambient case temperature, I'm not sure. I'll check his thermal compound and stuff sometime this week. I'm also going to check for any possible shorts that might be flaking it out. I need to figure out a way to test his PSU, don't they have some sort of test plug you can use to make sure the PSU is good? Hate has a way of closing the eyes, the ears, and the hearts of men.